Volcanic processes in

ore genesis
Volcanic processes in
ore genesis

Proceedings of a joint meeting of the Volcanic Studies Group of the

Geological Society of London and the Institution of Mining and
Metallurgy held in London on 21 and 22 January, 1976

The Institution of Mining and Metallurgy

The Geological Society of London
Special publication no. 7 of the Geological
Society of London

ISBN 0 900488 33 6

9 Institution of Mining and

Metallurgy and the Geological Society 1977

The ChameleonPress Limited, 5--25-Burr Road, London SW18 4SG

Contents
vi Preface
1 Introductory remarks on the transport problem
W. S. Fyfe
4 Model of hydrothermal ore genesis
J. W. Elder
14 Identification of ore-deposition environment from trace-element geochemistry of associated igneous host rocks
J. A. Pearce and G. H. Gale
25 Identification of the origin of ore-forming solutions by the use of stable isotopes
S. M. F. Sheppard
42 Hydrogen and oxygen isotope evidence for sea-water-hydrothermal alteration and ore deposition, Troodos complex,
Cyprus
T. H. E. Heaton and S. M. F. Sheppard
58 Hydrodynamic model for the origin of the ophiolitic cupriferous pyrite ore deposits of Cyprus
E. T. C. Spooner
72 Origin and emplacement of ophiolites
I. G. Gass
77 Hydrothermal alteration of the basaltic lavas of the Troodos Ophiolite Complex associated with l~heformation of the
massive sulphide deposits
G. Constantinou
78 Rare-earth element evidence for the genesis of the metalliferous sediments of Troodos, Cyprus
A. H. F. Robertson and A. J. Fleet
80 Modern submarine hydrothermal mineralization: examples from Santorini and the Red Sea
D. S. Cronan, P. A. Smith and R. D. Bignell
81 Mineralization at destructive plate boundaries: a brief review
M. S. Garson and A. H. G. Mitchell
98 Porphyry copper deposits
J. P. Hunt
99 Metallic mineralization affiliated to subaerial volcanism: a review
R. H. Sillitoe
117 Igneous geology and the evolution of hydrothermal systems in some sub-volcanic tin deposits of Bolivia
J. N. Grant, C. Halls, W. Avila and G. Avila
127 Occurrence, origin and significance of mechanically transported sulphide ores at Buchans, Newfoundland
J. G. Thurlow
128 Geological setting of the Skorovas orebody within the allochthonous volcanic stratigraphy of the Gjersvik Nappe,
central Norway
C. Halls, A. Reinsbakken, I. Ferriday, A. Haugen and A. Rankin
152 The Planes - San Antonio pyritic deposit of Rio Tinto, Spain: its nature, environment and genesis
D. Williams, R. L. Stanton and F. Rambaud
153 Kuroko deposits: their geology, geochemistry and origin
Takeo Sato
162 Stable isotope studies on Bougainville and in Matupi Harbour, New Britain, Papau New Guinea
J. H. Ford, D. C. Green, J. R. Hulston, I. H. Crick and S. M. F. Sheppard
163 Volcanogenic mineralization at Avoca, Co. Wicklow, Ireland, and its regional implications
J. W. Platt
171 Discussion
175 Name index
180 Subject index
PREFACE
Many geologists, particularly those in industry, have com-
mented adversely on the adequacy of university geology
courses as a realistic precursor to a vocational career in any
branch of the earth sciences other than in the universities
themselves. This volume, among other things, is evidence
that university staff are concerned about this situation and
are introducing the reality, stringency and urgency of indus-
try into areas where, until recently, they have been absent
or poorly represented. In this context the study of volcanic
processes belongs almost entirely to academics, whereas ore
genesis, although a fascinating intellectual subiect, is of prime
concern to the mining industry.
The contents of this volume are the proceedings of a
joint two-day meeting of the Volcanic Studies Group of the
Geological Society of London and the Institution of Mining
and Metallurgy held on 21 and 22 January, 1976. It is re-
levant to note that the proposal to hold this meeting came
from the dominantly academic Volcanic Studies Group,
whose members saw it as an opportunity to usefully marry
the rapidly increasing knowledge on volcanic processes to
the economically critical understanding of ore genesis. The
proposal to hold a joint meeting was enthusiastically sup-
ported by the Institution, and representatives from both
groups started to plan the meeting.
From the start it was evident the global tectonics would
be the framework and that sessions on volcanism and ore
genesis at (1) constructive margin, (2) destructive margin
and (3) within-plate settings would be a realistic subdivision.
There then came the task of selecting speakers who would
present authoritative papers ranging from global consider-
ations to the detailed study of one orebody. Also, the theme
was to be 'processes', so descriptive discourses on field,
petrographic, mineralogical and/or geochemical relation-
ships were to be discouraged. Conversely, as processes in
ore genesis meant hydrothermal processes, stable isotope
studies into the origin of ore-forming fluids would be
particularly welcome.
The reader will find that in some contributions the geo-
logical setting of mineralization is the prime concern
(Garson and Mitchell; Sillitoe; Grant etal.; Gass). In others
(Pearce and Gale; Sato) the use of trace-element geochemis-
try in identifying environments and processes is the main
theme. The role of hydrothermal fluids is, however, the
common denominator to virtually all papers, whether con-
sidered on the global scale by Fyfe, in a regional context by
Hunt and Constantinou or with respect to a single orebody
VI
by Platt and Thurlow. As well as variations in the scale of
study, there is a range in approach. Elder represents one
extreme in a largely theoretical discourse on hydrothermal
systems, whereas others (Sheppard, Ford et al.; Spooner,
Heaton and Sheppard; and Williams etal.) all use stable iso-
topes in the evaluation of the origin and role of ore-forming
fluids.
But what, other than a wealth of factual information and
an abundance of stimulating ideas, came out of the meeting?
What were the major conclusions and implications? This
assessment has, of course, to be personal, but I was particu-
larly impressed by the stable isotope evidence, which indic-
ated that the mineralized solutions were rarely magmatic in
origin being sea water in the case of constructive margins
- -
and meteoric waters in that of destructive margins. Seem-
ingly, magmatic processes provide the heat energy for the
hydrothermal systems and very little else. Equally impres-
sive were the semi-quantitative calculations on the time-
scales involved, which indicated that mineralization proces-
ses, at least at constructive margins, immediately followed
the magmatic event, were geologically short-lived, in terms of
103-104 rather than 106 years, and vigorous,
Those of us concerned with the organization of this meet-
ing consider it to have been a successful and productive
liaison between two groups of earth scientists that in the
past have only too rarely collaborated. The evidence is in
this volume that the studies and ideas of one group can be
of use to and stimulate the other. Hopefully, this is part of
a trend where fruitful collaboration between industrial
practicality and academic research is the rule and not the
exception.
C. Halls, M. J. Jones and J. M. Moore represented the
Institution and P. W. Francis, I. G. Gass and I. L. Gibson
the Volcanic Studies Group on the organizing committee.
M. J. Jones, the Institution's Secretary, was entirely
responsible for the preparation and processing of these
proceedings for publication. His efforts, and those of other
officers and members of both the Institution and the
Society, which helped to ensure the success of the meeting
are gratefully acknowledged.
I. G. GASS
Department of Earth Sciences
The Open Universi~/
Milton Keynes
January, 1977
Introductory remarks on the
transport problem
W. S. Fyfe
Department of Geology, University of Western
Ontario, London, Ontario, Canada
553.067
Synopsis
Subsurface aspects of the ore-forming process can be des-
cribed in terms o f standard chemical transport parameters
such as T - P gradients, mineral solubilities, solvent chemistry
and volume, diffusion parameters and energy flow. These
necessary factors are rather easily assessed by geochemical,
isotopic and structural studies. I f deposition is in the sur-
face environment, a complex array of factors will determine
the success of the natural concentration process and the
chances of preservation of a deposit.
Basically, most large-scale ore-forming processes involve
large fluid volumes and energy sources. Many environments
where these requirements are met, the modern ocean ridge
and the subduction environment, are still not well under-
stood. The need for remote-sensing techniques in the
submarine environment is stressed.
The science of geochemistry is concerned with the abund-
ance, distribution and behaviour of the elements and their
isotopes in the earth. As such, this science must include the
study of processes that lead to high concentrations of ele-
ments or ore deposits - perhaps the most complex of all
geochemical phenomena. During the last decades instru-
mental methods of describing the abundance of elements has
improved dramatically. Vast quantities of data are being
produced, but, despite such advances, our knowledge of the
exact geochemical behaviour of any given element and its
geochemical cycle is still remarkably inadequate. In part,
this is a result of the rethinking that has been demanded by
the new global tectonics, or what I would prefer to call our
new models of heat transfer by large-scale convective
motions that involve the crust and lithosphere. We are only
beginning to appreciate the geochemical linkages such as
atmosphere-rain-continent-ocean and mantle-basalt-
ocean-subduction. Only for a decade or so have we become
aware of the masses involved and the rate processes in the
creation and subduction of crust.
We are all aware of.the notions of renewable and non-
renewable resources (e.g. Skinner 11) and, previously, we
would have confidently placed a tree in the former class and
a massive sulphide deposit in the latter. But these distinc-
tions are based on our impressions of rate processes - rates
relative to our life span - and it is now clear that the time
constants of a redwood tree and a copper deposit may not
be so very different and both could be growing now.
Recent studies have shown that the understanding of
the geochemical behaviour of the elements may require an
integration of knowledge from many parts of earth science,
and few are unimportant. For example, Lowenstam 8 has
raised important questions regarding the significance of the
biosphere in mineral-forming processes (who would have
guessed that some mollusca secrete crystals of magnetite,
some arthropoda fluorite?). And as other papers presented
at this symposium show, there is much evidence that the
sulphur in many near-sea-floor deposits may come from
marine sulphate and it is difficult to conceive reduction
mechanisms other than those which involve bacteria. Work
in progress at the University of Western Ontario with our
department of bacteriology 2 has shown that the resistant
cell walls of bacteria, with their amino acid rich surface, can
absorb impressive quantities of specific metals.
Many other factors must be involved in ore formation.
Rock mechanical properties such as permeability, fracture
density, etc., control flow. Even the geomorphology of a
region may finally control the success or failure of a process
when all other factors are appropriate.
Chemical transport
Although there may be many approaches to ore-forming
processes, at some stage the chemical transport factors in
the system must be evaluated. All such processes require
chemical migration, Which enriches some given volume of
rock. Some chemical transport systems are quantitatively
well described and typical examples have been described
elsewhere, lo
The factors in such processes deserve careful considera-
tion if we are to make real progress in basic understanding,
or prospecting or in assessing reserves and resources. I do
not believe that we can scientifically estimate a resource if
we do not understand, with a high degree of certainty, the
processes that led to its formation.
To illustrate a very simple example of a transport pro-
cess we may consider the example of two containers at
temperatures T 1 and T 2 linked by a diffusion path at tem-
perature T 1 (see Sch~fer, TM p.22, and Fyfe and Henley 5).
The entire system is sealed, so motion is simply by diffusion.
The container at T 1 has a complex material A - X - Y - and
the entire system contains a gas (B) capable of forming a
gas molecule with the solid species (iA=ol~ + kBgas -~jCgas).
The quantity of A transported from T 1 to T 2 is given by
i Dqt
NA ~ -- 9
j IRT
where D is a diffusion coefficient, q the cross-section of the
diffusion path, / the length, RT the gas constant and tem-
perature T1, and ~P the difference in equilibrium pressure
of the gas species at T 1 and 7"2, which is a measure of the
'solubility' grad lent.
A fundamental feature of any transport process is that
for a complex system it is inconceivable that all species will
move at the same velocity or in the same quantities. Thus,
the final concentrations at source and sink will differ from
those initially present. Obviously, the most soluble material
will move (e.g. a quartz vein) and the temperature of de-
position will depend on solubility changes, flow velocity,
etc.
The transport equations for a geological situation will be
similar. To produce flow either a temperature or pressure
gradient must exist, and it is normally assumed that T
gradients are most important (but not always, e.g. growth
in a pressure shadow in a metamorphic rock). The factors
in the geologic case will include (a) the T gradient (and P),
(b) solubility and chemistry of all species in a given solvent,
(c) the fluid volume involved and the nature of flow (one
pass, multi-pass, etc.), (d) diffusion rates, (e) volume of
effective source and sink, (f) concentration at source and
(g) continuity and magnitude of the energy source produc-
ing the gradient.
 To these basic variables we must add all the restrictions
on a successful sink, which may be extreme, particularly
under near-surface conditions.
When the number of variables in the geologic case is con-
sidered, it is apparent that there will be no simple 'laws' of
ore deposition. But despite the real complexity, we can
consider certain required factors that may be easily observed
and may facilitate the search for ores. Among these one
might include the following.
(1) The presence of a high source concentration greatly re-
duces the requirements of a large permeable volume that
must be leached (this is clearly seen in some secondary en-
richment phenomena, and is perhaps important in the pro-
duction of some gold deposits by remobilization from rela-
tively gold-rich sediments).
(2) Access to an appropriate volume of fluid of good solvent
properties (this might imply proximity to the surface).
Today, emphasis is placed on chloride-rich (S-poor) fluids
and much work has shown that concentrations of such
metals as Cu, Pb and Zn are proportional to salinity,
0.01-10 ppm or more with CI" 100-100 000 ppm (see
Andrews and Fyfe 1 ).
(3) The presence of a large energy source, which will in-
crease the lifetime and vigour of the system (see Elder 3):
normally, the energy sources will be thermal or
gravitational.
(4) A steep thermal gradient should lead to focussed dump-
ing, but the precipitated material may be complex. Proxi-
mity of igneous material to the surface must provide steep
gradients. Shallower gradients may lead to cleaner precipi-
tation of a single species, but, at the same time, lower
grades, unless the solubility is very temperature-sensitive
(cf. gold 5).
(5) The source rocks must have pervasive permeability or
must be easily leached over large volumes (scoriaceous flow
tops, regions of high fracture density over rising plutons,
etc.).
Again, we could write down many geomorphological
constraints at the site of deposition, particularly if this site
is near to the surface. Considering this aspect of the prob-
lem, we should expect some types of deposits best shown in
younger rocks, and if we consider the modern sea-floor,
ophiolites or even porphyry copper deposits, there is some
validity in the age-preservation linkage. The preservation of
Archaean deposits may be related to volcanic burial pheno-
mena at a time of more intense and widespread centres of
volcanism.
Perhaps one of the most simple examples of ore forma-
tion by solution and transport is that involved in the forma-
tion of concentrated residuals in'lateritic weathering. The
solvent is rainwater and the energy a combination of solar
and gravitational energy. The sources are almost infinite
and the ore deposits (AI, Ni, Nb, etc.) large.
Prospecting
When the transport parameters'are considered, certain may
be rather easily assessed -- for example, by geochemistry,
by O, H isotopes and fluid inclusion studies, by stratigraphy
and metamorphic studies and by structural geology. The
important feature of the approach is that if any one condi-
tion is lacking, the chance for the presence of a large
hydrothermal ore depos!t must diminish rapidly.
Volcanism
When all the above factors are considered, the reasons
behind the title of this symposium - V o l c a n i c processes i n
ore genesis - become rational. Perhaps if we had attended
this symposium a few years ago, we might have expected to
hear much discussion of magmatic fluids and fumarolic de-
posits. But we have many papers dealing with downward
convective flow of surface waters in proximity to igneous
rocks. The fact that today we so often hear the magic
words volcanogenic-exhalative is a simple reflection of the
fact that thermal energy and water are available in the vol-
canic environment. In many cases, the water is sea water.
Magmatic fluids seem to be declining in importance and, at
the same time, experimental petrologists are working on
drier systems. We also know that as melts rise, although
initially rather dry, they tend to inhale fluids at depth from
their wallrocks and may later exhale as they crystallize. If
magmas were wet, and their water alone was the key to
transport, then it is difficult to see why so few plutons are
successful in, say, a porphyry copper situation. Naturally,
there are examples where this is the case (Sn, W, pegmatites);
but, in general, this is rare.
Scale
In the past two decades estimates on the mass of igneous
materials brought near to the surface have increased
dramatically. To account for ocean-floor spreading basaltic
magma must arrive near the surface (most intrusive) at a
rate in the order of 10 km 3 yr "1. Most must cool by driving
convective flow of sea water. 1 The energy available annu-
ally is large and, for example, would provide energy to heat
about 40 km 3 of sea water to 300~ - a temperature at
which metal extraction is feasible.
Direct observation of thermal systems and experiment
have shown that under these conditions such metals as Cu
and Zn have solubilities of several ppm in hot sea water. If
this moves to the water interface as a reduced (H 2 excess)
metal-rich fluid, sulp;,ides may be precipitated if the en-
vironment is appropriate. It should be noted that reactions
such as 3FeO (silicate) + H20 -+ Fe30 4 + H 2 will lead to a
low concentration of H 2 in a discharging fluid. This, in
turn, could generate the conditions necessary for sulphate-
reducing bacteria to flourish.
Ellis 4 observed copper concentrations near 4 ppm when
andesites were subjected to salt solution leaching at 400~
By use of this figure our 40 km 3 of sea water could bring
near the surface 1.6 x 105 tons of copper annually. Hence,
in the time to grow a large tree (100 years) we could grow a
large ore deposit. The Red Sea may be mined in the near
future. Perhaps soon, we will find ore deposits near the
ocean ridges. For the future, new submarine prospecting
techniques are needed. The dispersed deposits in nodules
are already well known. 9
The potential energy of a high-level pluton for fluid
transport is also impressive. Modern granite plutons 6 tend
to average about 500 km 3 in size. A single pluton could
heat about 100 km 3 of water to 300~ Copper deposits
in the million ton range could result, even on a single-pass
basis. If closed cellular convection occurred, the quantity
of fluid needed is much reduced.
The thesis of this introduction is that, given energy and
fluid, ore deposits may result. We know that the sub-
duction process returns vast quantities of water and other
elements to the mantle in rocks such as spilites and sedi-
ments. 7
Most of this water must return, or else we would have no
hydrosphere left. The subduction zone environment is one
of intense metasomatism and magmatism where certain
characteristic types of ore deposits occur.
But it seems that we do not fully understand the process
of the return flow of subducted fluids and metals. Do we
have more economic surprises in store? The exploration of
the sea-floor has certainly shown us that we did not know
all about metals in 1940.

References
1. Andrews A. J. and Fyfe W. S. Submarine metamor-
phism and massive sulphide generation. Geoscience, 3,
1976, 84-94.
2. Beveridge T. Murray R. G. and Fyfe W. S. Metal absorp-
tion on bacterial cell walls. Chem. Geol., in preparation.
3. Elder J. W. Model of hydrothermal ore genesis. In.Vol-
canic processes in ore genesis (London: Institution of
Mining and Metallurgy and Geological Society, 1977),
4-13.
4. Ellis A. J. Natural hydrothermal Systems and experi-
mental hot-water/rock interaction. Geochim. cosmochim.
Acta, 32, 1968, 1356-63.
5. Fyfe W. S. and Henley R. W. Some thoughts on chemi-
cal transport processes, with particular reference to gold.
Miner. Sci. Engng, 5, 1973, 295--303.
6. Fyfe W. S. The generation of batholiths. Tectono-
physics, 17, 1973, 273-83.
7. Fyfe W. S. Hydrosphere and continental crust: growing
or shrinking? Geoscience, 3, 1976, 82-3.
8. Lowenstam H. A. Impact of life on chemical and
physical processes. In The sea, volume 5 Goldberg E. D.ed.
(New York: Wiley, 1974), 715-96.
9. Deep sea searches: the story of the seabed assessment
program (Washington, D.C.: National Science Foundation,
1976), 22 p.
10. Sch~ifer H. Chemical transport reactions (New York:
Academic Press, 1964), 161 p.
11. Skinner B. J. Earth resources, 2nd edn (New York:
Prentice-Hall, 1976), 151 p.
Model of hydrothermal
ore genesis*
J. W. Elder
Department of Geology, University of Manchester,
Manchester, England
53.072:553.216.5
Synopsis
The dynamical nature of the mechanisms o f operation o f
hydrothermal systems and their key parameters are
described. A simple mode/suitable for a gross analysis o f
the role o f water circulation on the solution, transportation
and deposition o f ores is described in detail.
Ores are rocks which within localized volumes contain
minerals that are quite different in kind or amount from
those in most rocks.
From a global point of view the crust itself is a great
'ore' deposit, being exceptionally rich in minerals such as
quartz, and feldspars, which are believed to be largely absent
from the mantle. Although the idea is a useful one, it is
an overextension of the notion, since clearly an orebody is
recognized by its existence within a volume not only small
compared with the volume of the crust but all of whose
dimensions are small compared with the thickness of the
crust. On the other hand, neither do a few localized
grains of, say, rutile in a high-grade metamorphic rock
represent an orebody.
Whereas the great bulk of rocks, more than 99% by
volume, are composed of the elements O, Si, AI, Fe, Ca,
Na, K and Mg, and the relative abundance of all other
elements is generally much less than 1%, there are some
rocks for which they are rich. In comparison w i t h the
crustal average, relative enrichment by a factor of 102 to
103 is typical, t2
The central question is how these localized concentra-
tions were formed. Clearly, there are two extreme types
of such bodies.
(1) The segregation of material may occur essentially
locally within the ambient rock. I include here most of the
deposits closely associated with igneous, metamorphic and
sedimentary rocks, including sinter, pegmatites, small-
scale metasomatism deposits, placers, laterites and so on.
*This paper is dedicated to William of Ockham, 1285--1349, a
somewhat querulous Franciscan monk noted for the use of a form
of argument based on the notion of economy of ideas and con-
cepts that is known as 'Ockham's razor' and is an early version of
modern operational philosophy. This antipathy to the ad hoc was
put: 'non sunt multiplicanda entia praeter necessitatem' -- entities
are not to be multiplied beyond necessity. Some of the ideas were
particularly charming: for example, that motion is merely the re-
appearance of a thing in a different place. This particular concept
is even rather apt in describing ore genesis. It is fair to say that this
paper is about the motion rather more than the thing.
It is also worth noting that this paper is very much a sequel to
an earlier study. 2
These processes are in some f o r m operating more or less
continually throughout the crust and are most readily
described as variants of the petrogenesis of ordinary rocks.
Except where the t w o types of segregation processes work
together, systems of this type are not my present concern.
(2) The segregation of material may occur by its being
moved to a quite different part of the ambient rock. The
notable example of this type are the so-called hydrother-
mal deposits. The physical processes within such systems
are the concern of this paper.
There may well be systems that involve both of these
processes, but at different times or places. For example,
the chrome deposits found in layered ultrabasic intrusions
were presumably produced by transportation of Cr in a
more or less homogeneous magma, and subsequently, after
the intrusion was in place, becoming segregated as the
layering developed. Most of this paper is concerned with
water substance as the solvent and transporting agent, but
we should not overlook, even if it is only mentioned in
passing, the transporting role of basic magmas, particularly
for those elements such as Ni, Cr and Co, which they
readily carry.
The discussion proceeds under Formulation of the
problem with some elementary estimates of the order of
magnitude of the water flow mass rate and the scale of a
deep hydrothermal system. A reservoir model is outlined.
Next, the features that control its initiation and mainten-
ance are discussed. The main technical aspect of the paper
is the analysis of the conditions in the gas column; this is
followed by a consideration of the extraction of matter and
energy from the source. The main result is obtained. The
surface zone, driven by the power transported in the gas
column, is identified through its characteristic temperature
in the section entitled Flow in surface zone. The
transport of soluble matter -- in particular, the effects of
the variation of solubility with temperature - is briefly
treated in the penultimate section.
Although the model to be described is rather general,
for the sake of concreteness I shall present the results and
discussion for only one system for which the power and
mass transports are similar to those of the Taupo district,
New Zealand, today. The deep part of this system is most
like that responsible for a moderately large copper porphyry
deposit.
F o r m u l a t i o n of the problem
Whereas the mantle is rather dry, with water content
possibly less than 0.1%, water substance is a ubiquitous
constituent of the crust. This is most dramatically and
directly seen in the regions of surface mass discharge in
geothermal and volcanic areas. Among the studies of these
phenomena, those which describe the role of water sub-
stance in modern hydrothermal systems are the most de-
tailed. Freely circulating bodies of hot water substance
have been identified within the upper 10 km of the crust.
Mass transport at the rate of 1 ton sec] is typical (see,
for example, Eider3). These systems are the uppermost
parts of deep systems, which ultimately draw their energy
from the mantle. The mass and assumed chemistry of the
so-called meso- and hypothermal ore deposits strongly
suggest that they are formed by transportation in a deep-
seated system of water-substance circulation. In effect, an
ore deposit is a fossil hydrothermal system in which we
obtain a glimpse of a deeper part of the system. Whatever
the details, water is vital. Its existence in the crust is taken
for granted here. All I am concerned to demonstrate is
that the system can move enough of it.
The essential questions The localization of the primary deposit reservoir at the
There are two extreme possibilities for the mode of trans- critical boundary may at first sight seem rather arbitrary.
portation of the ore material in the hydrothemal fluid. It is very important to appreciate that it is not the change
of state of water substance and the possible consequent
(1) The ore material concentration is high: this is the least effect on its solvent capability that is dominant. By far
demanding on the reqLrired transport of water. Undoubtedly, the most powerful effect arises through the permeability.
this is the case for some pegmatitic deposits; for hydro-
thermal deposits, fluid inclusions, although they often show
///
high concentrations of, for example, sodium chloride, do
not give any indication of high ore material concentration.
(2) The ore material concentration is low, being of the
same order as that in the source rocks. T h i s situation is the
most demanding on the required transport of water. Never-
theless, even in this extreme case, ample water can be avail- Wet system
able - indeed, it is the main point of this paper to show in sub-criticol
some detail how this enormous flow of water is generated
and maintained.
Consider the very simplest situation in which an orebody
Primory
of volume V is produced from a source volume W as deposit
follows.* Let the concentration ratio of ore material in the
deposit to that in the source be n. If ore material is taken at
the concentration in the source and transferred to the de-
posit volume and all dumped there, then V = W/n. Also
this will require a total equivalent liquid water transport
T
rn, , GQS cotumn

of order volume W. If we also.assume that during this time,

t, the source region has been substantially cooled solely by
the heat transported out of it by the flow at mass rate rn,
then t ~ (total mass of the soume volume)/m. By way of
illustration consider: concentration o = 10 ppm; m = 1 ton
sec"1, t = 10 4 yr, n = 10 3. Then a source volume of
10 3 km 3, from which all the ore material is extracted and
deposited within a volume of 1 km 3, produces a deposit of
1% concentration and of ore material mass 3 x 10 7 ton.
The total mass transport of water is equivalent to a f l o w of
3 x 1012 ton through the source.
A simple calculation like this, which I am sure has been
done by everyone who has ever thought about this problem,
exposes the three key questions of solubility, flow rate and
time. Are these quantities at all reasonable and could such a
system work? As we shall see, these are all acceptable
values and a variety of suitable mechanisms can be
quantitatively identified.
A schematic model
A hydrothermal extraction system is envisaged as a cascade
of reservoirs interconnected by permeable conduits (pipes)
as illustrated in Fig. 1 in a manner analogous to that in an
elaborate distillation column. In the analysis presented
here only three reservoirs are considered.
(1) The source is the largest reservoir and here, for the
sake of concreteness, is considered as a granitic mass of
typical volume 10 3 km 3, partially molten at the time of
its emplacement.
(2) The primary deposit - a much smaller reservoir - is
the region in which most of the ore and associated gangue
material is deposited. Here it is considered to exist near
the main interface in the system, the boundary between
the super-critical deep system and the wet sub-critical
su rface zone.
Source
Fig. 1 Schema of reservoir-pipe model of hydrothermal
system
In the vicinity of the interfacial zone two phases of water
substance can be present. This has the effect of drastically
reducing the effective permeability of the medium. For
a single phase the permeability is solely determined by the
void geometry and not at all by the fluid. But when two
phases are present they restrict such other's passage.
The localization of the head of the mushroom is not so
sharply defined. It is limited strongly only in its relation
to the 'boiling point with depth' (BPD) relation (see
Appendix). If the heat flow is sufficiently high for temper-
atures to locally exceed the BPD, phreatic eruptions are
likely. The surface zone hydrothermal system becomes
(certainly for a time) a one-pass eruptive system without
the ability to create a deposit.
The models discussed here do not impose any strong
constraints on the source of the water substance, since the
bulk of it may be recirculated many times or it may be
derived directly from the surface zone.
In a rather schematic way the approach here recognizes
three main types of hydrothermal ore deposits, which are
called in the classification of Lindgren 11 hypothermal
(my primary deposit); mesothermal, here considered to be
deposited in the surface zone 'mushroom'; epithermal
and telethermal, virtually ignored in this paper, but thought
of as arising from the recharge-discharge part of the
surface zone system.

(3) The hot water mushroom: the hydrothermal system

in the surface zone has a temperature distribution shaped
like a mushroom, within the head of which is a rather dis-
crete body of hot water and rock of characteristic temper-
ature typically about 250~ for a vigorous system.
*Definitions of the symbols used in this paper are given on pages
12--3.
Source constraints
The initiation and maintenance of a hydrothermal extraction
system imposes three conditions on the source.
(1) In the mode o f emplacement the main concern here is
the time-scale of the emplacement. If the emplacement is
rapid, the analysis of the problem will be much more
complex. On the other hand, if the emplacement is
sufficiently slow, the analysis can be treated as a quasi-
steady model - indeed, this is the only case considered in
this paper.
(2) T w o distinct regions are envisaged for access o f the
hydrothermal f l u i d to the ore material within the body o f
the source. (a) While the source is sufficiently hot for it to
remain partially molten, the hydrothermal fluids presum-
ably have direct access only to the outermost portions of
the source region, otherwise being transported through the
source dissolved in the magma. Provided that there is
sufficiently rapid mixing within the source, this will provide
ready access to the interior. The time-scale of this situation
will be relatively short. (b) When the source is sufficiently
cooled for it to be substantially solid it behaves as a porous
medium - to the flow, apart from its permeability, it is
indistinguishable from the ambient rock. The time-scale
of this situation will be relatively long.
(3) With regard to the gas column, as we shall see, the
most effective mode of transportation of material from the
source to the deposit volume is in a column of super-
critical gas (of water substance). Only under certain condi-
tions can a gas column be initiated and maintained, and it
is during the life of the column that the bulk of the
material transport occurs.
Emplacemen t time-scale
Consider the rise of a molten body. W~ can estimate the
rate of rise for two extreme sets of circumstances.
If the rate of rise is controlled by a balance between
the buoyancy of the body and the viscous drag of its
su rrou nd i ngs, the vertical velocity
w "- 2/9 a 2 g 'Iv, with g '= (~o/p) g
where Ap/p is the proportional density contrast. This is the
well known Stokesian expression. For the typical values
a = 10 km, Ap/p = 0.1 and v = 10 ]6 m 2 sec"] we have
w = 0.07 m yr "l
This body would rise a distance equal to its diameter in
3 x 105 yr.
If the rate of rise is controlled by the rate at which the
advancing face can melt the ambient rock and refreeze
material at its base, it is readily shown that
w ~ (tO N/Ea)
with
N = (A/Ac) ]/3
and
E=L+cO
where L is the latent heat of freezing, c is the specific heat
and 0 is the temperature contrast. This expression is de-
rived by equating the rate of transfer of heat to the inter-
face by the internal convection to the rate of supply of
enthalpy to melt the interface. For the typical values used
above and a = 10 km, c = 103 J k.g 1 and 0 = 100K
we have
w = 30 m yr "]
This body would rise a distance equal to its diameter in
700 y r.
For this later case the time-scale of emplacement is com-
parable with that of the development of the hydrothermal
system. This suggests, even though we ignore this possibi-
lity in the subsequent discussion, that some hydrothermal
extraction systems will be dominated by the emplacement
of the source region. For somewhat smaller bodies, how-
ever, and if the interaction of the hydrothermal system on
the source region during its emplacement is also taken into
account, conditions will not usually be so extreme.
Circulation within the source
If the source region, at the time of its emplacement, is
largely molten, the time-scale r of the internal mixing
process is given by
r ~ ( A c / A ) I / 3 a2/K
with
A = 7gOa3/Kv
where a is a length scale, the mean radius for a blob-like
source, 7, r, v are, respectively, the coefficient of cubical
expansion, thermal diffusivity and kinematic viscosity of
the magma, 0 is the temperature difference between the
source interior and the immediate surroundings, A is the
Rayleigh number and A c ~ 10 a, an empirically determined
constant (see, for example, Eider, 6 pp. 6 0 - 7 7 ) . The
derivation of this expression assumes that the interior
motion creates a surface boundary region of extent
"" a (Ac/A.) 1/3 of time-scale r "" $2/~.
As an example consider a large granitic pluton with
a = 1 0 k m , 0 = 1 0 0 K , r = 1 0 " 6 m 2sec " 1 , v = 1 0 2 m 2sec "l
and 7 = 10"s K "] for which A = 10 ]4 and r = 102 yr. This
source is being rapidly m i x e d - indeed, motion will persist
until A ~" 103, when v "" 10 ]4 m 2 sec"1, which is well into
the freezing range for granite.
Subsequently, therefore, I will assume that the time-
scale of the mixing within the source region is small in
comparison with the time-scale of the hydrothermal
system - typically, 104 yr (see below). In other words,
mixing within the source is sufficiently rapid to allow
almost complete access of the hydrothermal fluid to the
material of the source throughout the life of the hydro-
thermal system.
Existence o f the gas column
The problem of the existence of the gas column can be
treated by means of.a pipe model (following the detailed
discussion given in Eider, 2 pp. 3 2 - 9 ) . Thus, if water
leaves the surface zone, of thickness d, with density Po,.
passes through the source volume at depth h, rises as a
gas, of density P l to heat the surface zone to a local density
P2, it can be shown that the mass rate of f l o w m is given
by
m/kSg = [ ( P o - P l ) -- (P2 - P l ) d / h ] / [ v l + (v2 - v l )d/h]
where k is the permeability, S is the pipe cross-sectional
area and, although they are readily accounted for, the
variation of permeability along the pipe is ignored and the
recharge resistance is taken to be zero.
We are here principally interested in the range of mass
transport. Consider the t w o extremes: (1) an erupting
system has d/h = 0 and m (erupt)/kSg = (Po - Pl )/vl ;
(2) a collapsed system has d/h = 1, namely no gas column,
when m (collapse)/kSg = (Po - P2)/v2. Thus, the mass
rate ratio
= m (collapse)/m (erupt) = (Po--Pl) v ; / ( P o - P 2 ) V ]
Taking representative values Po = 1 g cm "3, cold surface
zone water, P2 = 0.96 g cm -3, surface zone water at 100~
Pl = 0.3 g cm "3, near-critical gas, and v2/vl ~ 1, and we
have ~ -~ 1/20. In this case we can expect that the bulk of
the possible mass transport will have occurred before
collapse sets in.
Peculiarity o f a super-critical hydro thermal system
There is an important distinction between a wholly sub-
critical (wet) hydrothermal system and that which is in
part or wholly super-critical. In a wet system, even though
liquid water substance is compressible, the dominant effect
of the density arises from variations of temperature; Temperature distribution
moreover, these density variations are principally effective A typical set of profiles of T(z) in a gas column is shown
through their role in generating buoyancy forces. Further- in Fig. 2 for various values of m. For small values of m,
more, this is the basis of the analysis of such systems by here somewhat less than 0.1 ton sec"1, the gas column is
means of the Boussinesq approximation in which density nearly isothermal at the temperature of the base of the
variations are ignored, except insofar as they lead to the
sub-critical zone. For large values of m, here those greater
generation of buoyancy forces. In a super-critical (dry gas)
than about 1 ton sec"1, the gas column is strongly non-
hydrothermal system the working fluid is a very com- isothermal, especially in the region of expansion near the
pressible gas, so density variations arise from variations of
top of the column. High temperatures are required at
both temperature and pressure. We cannot therefore
depth to maintain the column, and these temperatures are
directly use data from a wet system to aid the analysis of a
strongly related to m. For this column the operating range
dry system; nor can we use the Boussinesq approximation,
but this is a very minor point in a pipe-like model.
rn
I000~ 30
The gas column
The key element in the extraction process is the gas
column, of super-critical water substance, which connects
the source region, of the magma body and its immediate
surroundings, to the surface zone of (generally) sub-critical
water substance. As has already been discussed, for an
interval the gas column is close to equilibrium. It is
therefore straightforward to describe its properties in
some detail. We shall first treat the gas column in isolation
and subsequently consider its effect on the source region.

Formulation
Field equations The mechanics of the gas column can
be described by the following relations:
p = PlaT (a)
= flT'/~ (b)
w = m/pS (c)
dP/dz = pg + IJw/k (d)
dT/dP = 1/p(Cp + a) (e)
The properties of the fluid are described by (a) and (b)
and are discussed elsewhere. The conservation of fluid
mass is expressed by (c), where m is the rate of mass flow-

ing through every section of the column of local cross-
sectional area S. The pressure gradient in the column,
represented by (d), arises from two effects: the weight of
the fluid itself, together with the resistance experienced by
the fluid in permeating the ambient ground, represented by
the relation of Darcy. Finally, (e) represents the energy
relation, in which it is assumed that there is a balance
between the rate of working of the decompressing gas
and the upward advection of heat. For the present dis-
cussion, unless stated to the the contrary, the following
quantities are taken as constants for a given gas column:
n,/~, m, S, g, k and Cp. There is, however, no difficulty in
treating the more general cases - for example, in which
S = S(z), k = k(z).
Boundary conditions The gas column is represented by
two first-order differential equations. Provided, therefore,
that two conditions are imposed, P, T are defined. For
the moment it is convenient (and realistic) to consider the
pressure and temperature given at the top of the column or,
rather, the base of the sub-critical surface zone. Thus, at
some level z -- d we take P -- Pd, T = Td. The equations
can the be integrated (numerically) downwards to obtain
P(z), T(z). Notice that there is some redundancy in the use
of parameters, since the effect of m, k, S is determined in
this model solely through the combination m/kS. It is
convenient to use and state these values separately, but it
should not be forgotten that they occur in combination.
Further, the relative importance of the flow resistance is
measured by (mpJ'gkSp 2 ).
1
o
5
0 10
Z, k m
Fig. 2 Temperature, T, ~ asa function of depth, z,
km, in gas column with d = 2 km, Pd = 0.2 kbar, Td =
400~ S = 0.01 km 2 and k = 1 darcy for various values
of rate of mass flow, m, ton sec"1
is about 0 - 3 0 ton sec"1, the upper limit being set by the
temperatures of basaltic magmas; the effective lower limit
is about 0.03 ton sec"1 when the column is barely working.
Given temperature at depth
So far the rate of mass flow, m, has been treated as a free
parameter. It is, of course, deterrhined by the system.
Because m is strongly related to the temperature in the
gas column at depth, which is itself controlled by con-
ditions in the source region, it is appropriate to specify a
temperature Th at the base of the gas column z -- h. We
then have boundary conditions imposed at two levels,
which are sufficient to determine T(z), P(z) and m. This
is performed iteratively by integrating the equations down,
wards from d to h with an assumed value of m, obtaining
the base temperature, adjusting m accordingly, and re-
peating until the base temperature is equal to Thwith the
required precision (-+ 0.1% used here).
Thus, the state of the column is determined by its
geometry, the quantity m/kS, the properties of water
substance and the two temperatures Td, Th.
Flow velocity
It is important to realize that the volumetric flow velocities
are very small. For the archetypal model described here w
lies generally in the range 10"4-10 -3 m sec"1 . The system
operates entirely within the range of validity of the Darcy
relation,
Role o f solute on flow
So far we have ignored the effect of the solute concentra-
tion on the flow by, in effect, considering only the case
of low solute concentration. Although this is a legitimate
first approximation that may well apply in some cases, it
is clearly not the case in many situations. Certainly, the
concentration of the economically important ore materials
may be low, but we cannot ignore the possibly very much
greater concentration of the other constituents of the
hydrothermal fluid such as SiO2 and NaCI. Inclusion
studies indicate that NaCI is an important constituent,
representing typically of the order of 10% of the fluid -
values of 50% are quite common. 14
The main effect of the solute on the flow arises from
its contribution to the density of the fluid. Hence, in
relations (a)-(e) we treat the fluid as a two-component gas,
one component being the gaseous water substance of
variable partial pressure and the other the solute of fixed
partial pressure.
No new qualitative effects arise, although this is a very
important feature in the details of the operation of a
system. In particular, much of the material thus trans-
ported will lead to the production of gangue.
Extraction from the source
Given the mechanics of the gas column, we have a method
of transporting matter to another place. But there remains
the essential problem of first extracting that matter from
the source and, even more to the point, whether enough
matter can be so extracted.
Provided that the time-scale of adjustments of the gas
column is rapid in comparison with all the other rate-
dependent processes in the system, we can make a quasi-
steady model of the extraction process.
Formulation
Consider a source volume V 1 of rock substance of density
po and specific heatc, in contact with fluid circulating at
the mass rate m at recharge temperature T' and discharge
temperature Th. Then we have the following relations:
F = mc (Th - T')
dTh/dt = F/poCo V 1
dM/dt = m/p, V 1
dO/dt = - o r e
Relations (f) and (g) express the conservation of energy as a
balance between the rate of loss of the total source energy
and the net rate F transported by the flow. For simplicity,
the recharge temperature T ' w i l l be taken as Td, the tem-
perature at the base of the surface zone, unless stated to the
contrary. Relation (h) merely gives the total mass of fluid
M as a proportion of the source mass which has been in
contact with the source. The total amount of 'salt', Q, in
the source is depleted at the rate om, as expressed in (i),
where o is the concentration of 'salt' in the discharge. It
is known that o = a (P, T) and, in fact, varies strongly with
P, T, but for the moment we will shut our eyes to this and
treat o as a constant, so the total amount of 'salt' extracted
is o(p, V 1M).
Collapse o f the gas column
The situation is actually somewhat more restrictive than has
so far been described. The gas column cannot persist in-
definitely. In order to maintain itself its internal pressure
must at every depth exceed the water pressure in the
adjacent rock - otherwise it will collapse and the gas phase
is over. The inclusion of this feature does not greatly alter
the total amounts transported for the more intense systems,
so a very simplified approach is adequate.
Let the adjacent ground He open and have an effec-
tive mean fluid density Pa. Then the adjacent ground
pressure is Pa = Pd+ Pa g (z--d). We then require simply
that the pressure at the base of the gas colume Ph > Pa.
If this condition is satisfied, in most cases the gas column
pressure will at every level exceed that in the adjacent
ground. Depending on the circumstances envisaged, we may
choose Pa in the range 0 . 3 - 3 g cm 3 - namely, from that
of water substance at the critical point to a lithostatic load.
Total mass flow
Data from a particular model shown in Fig. 3 give values
of the total mass flow ratio M(t). The initial mass extraction
is rapid. In this example it takes about 1 Myr to transfer
an amount M = 1, but already half of this is transferred in
5 x 104 yr. The ultimate rate for a long-life system without
collapse is M ~" log t -- a continually decelerating rate.
M
/ Co|iopse points
' I I
0 50 100
t, kyr
Fig. 3 Total mass flow ratio of water substance, M, passed
through gas column as a function of time, t, for case
described in text
Collapse points are indicated for various values of Pa.
Even for the extreme case of Pa = 1 g cm "3, so that collapse
is relatively early, the bulk of the possible mass flow has
been transferred.
(f)
A key result
(g)
(h) The mass of fluid available to flush out the source region is
(i) at most about equal to the mass of the source region. This
result applies to a very wide range of systems - basaltic
and granitic volcanism, extreme variation of source depth,
extreme variation of source volume and extreme variation
of permeability. Furthermore, this mass of fluid is largely
produced in a relatively early part of the life of the system
in an interval typically of the order of 104 yr. In other words,
the extraction process is geologically rapid.
We thereby deduce that the essential feature of the
production of an orebody is the extraction of the material
from a large volume and its deposition in a small volume.
This point of view is quite different f r o m the notion that
the orebody is produced from a highly concentrated solu-
tion w i t h i n the source region. T o put this into concrete
terms consider an orebody of volume 1 km 3 in which the
concentration of an ore material is 103 times normal. Then
that material was extracted f r o m a source of volume
103 km 3.
Power extracted
Although our immediate concern is the total mass trans-
port, it is of interest and provides a link with the surface
zone (see later) to consider the power extracted from the
source region (see Fig.4). The power level during the first
few thousand years is very high (!), being of the order of
105 -
1000 ~
104
0 systems
3 thermodynamic boundary conditions. This cannot be
I0
102
omic limit
10
I I I I I I
10 102 103 104 105 106
yr
Fig. 4 Power, MW, transmitted in gas column as a
function of time for initial source temperatures of 750 and
1000~
104 MW, falling to the power level found in, for example,
the Taupo systems after 5 0 0 0 - 1 0 000 yr, thereafter remain-
ing rather vigorous for an interval* of the order 10 s yr,
after which the power is rather small. Note, however, that
the initial high power levels are not available at the surface,
since the bulk of this power is absorbed in heating up the
ambient hot water and rock.t
Comment on the time-scale
The time-scale, of the order of 104 yr, suggested in this
work is for a single hydrothermal event. This must not be
confused with the time-scale of a geothermal or a volcanic
area. Within an orogenic belt, for example, volcanism and
plutonism may persist for times of the order of 100 Myr,
but individual elements of such a system have much
shorter lives. In the Taupo district, as observed at the
surface and at depth from aeromagnetic surveys, it is
possible to identify of the order of 102 rhyolitic intrusions
*Throughout which time geothermal power exploitation would be
economical with current technology.
tlO 4 MW is sufficient to heat a cylinder of rock of area 30 km 2
and depth 10 km in 103 yr.
produced over the last million years of activity, which
suggests a rate of production of about 10 .4 yr .] An exten-
sive area of mineralization will have been produced by a
multitude of individual short-lived events.
Comment on the gas column cross-section
It is important to point out that the model described here
is a phenomenological model. The discharge is, for a
given system and boundary conditions, determined by
m/kS. In principle, k could be estimated for a given
system by a detailed inspection of the deposit and its
surroundings to determine the void geometry. But until
such studies are make k is virtually unknown, and my
choice of 1 darcy is simply that found as an upper limit
for Taupo systems as determined by pumping tests on
bores about 1 km deep. Even so, there remains the cross-
sectional area S. I have been unable to devise a satisfactory
method of independently estimating S from the f l o w
mechanics (for a wet system S ~ (heated area)/40). A t
least S is more or less known from direct observation for
some orebodies. Fortunately, however, the gross
thermodynamics of the gas column is such that the total
mass transport ratio M is determined largely from the
said for the time-scale.
Flow in surface zone
Above the gas column, within the wet surface zone, there
will be a freely circulating body of liquid water. The
mechanics of these wet hydrothermal systems are reason-
ably well known. Here we are concerned with only the
grossest features in order to show how the deep and the
surface systems are linked together and thereby provide
further methods for identification of the whole system.
Gross properties
In summary, the t w o gross properties of an ideal free
system* (in which 3",v are independent of temperature) in
a state of vigorous convection, the power transferred, F,
and the rate of total volume flow, ~ , are found to be:
F/S 1 = k pc ~g (AT)2/vAc; A T = T d - T o
~J/S 1 = k ~(g Z~T/vA c
where $1 is the heated area and A c ~ 40 is the critical
Rayleigh number. The effective column cross-sectional
area of the hot water column is SI/A c, In the natural
situation the heated area S] is effectively unknown. If we
regard it as norminally of area d 2, we have
= Km d A m / A c ; Am = k?gdAT/Km v
F = pc~AT
These relations can be understood in the following way.
The differences of density between the h o t column and
the cold column produce a pressure difference P~'?pgOd.
From Darcy's relation the corresponding vertical velocity
w"kp/Iad. The power transferred F ~ p c w e d 2, where here
the area of the hot 'column' is taken as proportional to d 2.
Hence F ' k p c T g 8 2 d 2 / v and the rate of total volume f l o w
~J~wd2,,~k~gSd2 /v.
Let us consider a case d = 10 km, Rayleigh number
A = 4 x 103, Km = 1 0 - 6 m 2 sec"1. Then we have 9 =
1 ton sec"1 . These are values similar to those in the Taupo
hydrothermal systems.
*In convection in a permeable medium Km is the effective thermal
diffusivity of the fluid-saturated medium = K/pc, where K is the
thermal conductiviW of the rock and pc the specific heat capacity
of the fluid; u is the kinematic viscosity of the fluid.
Characteristic temperature o f surface zone
We have, of course, no direct evidence for any of the above
description of the deep system. It is, however, possible
to make a partial identification as follows. Let us regard
the surface zone as a kind of calorimeter into which the
heat transported in the gas column is deposited and from
which heat is lost by the convective process in the surface
zone. Then, by noting the amplitude and time-scale of
the surface zone temperatures and comparing these with
field information, we can check our description.
As was shown by Elder, 2 the power F" transferred by a
wet hydrothermal system, allowing for the variation of
density and viscosity of liquid water with temperature, is
related to the characteristic temperature ~ of the liquid
water (and rock) in the 'mushroom' of the hydrothermal
system by F'~. ~O4, where ~ is a fixed quantity defined
by the system geometry. For data from the Taupo hydro-
thermal systems we find ~ ~. 0.2 W K "4. Equating F ' t o the
power F transferred in the gas column allows us to calculate
8. Typical results are given in Fill. 5. The data suggest that
individual Taupo systems with F " 1 0 3 MW have been
active for of the order of 104 yr, the whole area having been
active very much longer, as is indicated by the most deeply
exposed ignimbrites of age at least 10 s yr. Older systems
reach temperatures less than 100~ in times of the order of
10 s yr, comparable with the system of hot springs in the
Hauraki area.
The evidence is flimsy, but is reasonably convincing.
,00F /0, / p . V i d q i / d t = m i _ ! (oi_ 1 - o i) + m i ( o i + 1 - o i)
falls little during this time. Smaller surface systems not
only achieve their maximum temperature earlier but also
reach a higher temperature. Not only do these temperatures
approach those at the top of the gas column but they also
approach BPD temperatures above which the surface zone
is unstable to an adiabatic decompression, and phreatic
eruptions arising from depths of as much as a few kilo-
metres are possible. Such eruptions are a feature of the
Taupo systems.
From the point of view of ore genesis our concern is to
enquire if such an eruption could also disperse a proto-ore
deposit. For the model studied here the answer is largely
no. The time-scale of onset of such an event is much shorter
than the time-scale of the ore deposition, so the bulk of
the material would not have been deposited.
Transport of soluble matter
If the solubility were fixed, the source inexhaustible and
all the material in solution were dumped in the primary
deposit, the mass of the deposit would be simply oMp3 V 3,
None of these conditions is likely to be met. It is therefore
necessary to look in some detail at the distribution of
potential ore material in the system.
Reservoir cascade
Consider a cascade of n reservoirs labelled i = 1, n. Let the
rate of mass discharge from reservoir i of volume V i and
ore material concentration qi at enthalpy ei and solute
concentration oi. Assume that the recharge to reservoir i
is equal to its discharge m i and comes from reservoir (i + 1).
Let the rock density be p.. Then for each reservoir the
conservation of energy and solute give
p.~dei/dt=mi_ 1 (el_ 1 -- e i) + m i ( e i + l - - e i) (k)
(I)

I//
1 0 0 ~ 0 . 3 ~..~..~
L I I I
102 103 10/" 105
t, yr
Fig. 5 Characteristic temperature, 0, ~ as a function
of time, t, yr (initial heating-up phase also represented
for various values of V 3 / V 1 )
Also shown in Fig. 5 are data for the heating up of the
surface zone for various values of V3/V 1, the ratio of the
volume of the sub-critical convection region to that of l~he
source. The heavy line represents the characteristic temper-
ature for V3/V 1 ,r 1, so that the surface zone is always in
equilibrium with the gas column. Otherwise, the energy
balance for the surface zone requires
p . c . V 3 dS/dt = F - t;84
During the early life of the surface hydrothermal system
the bulk of the energy supplied by the gas column is used
to heat up the surface zone 'calorimeter'. After an interval,
for example, of the order of 250 yr for the Taupo systems,
the behaviour is no longer dominated by the thermal
capacity of the surface zone and the characteristic temper-
ature approaches that for an equilibrium system.
A remark on the early vigour o f the surface zone
The time-scale of the early 'calorimeter' phase of the surface
zone system is roughly proportional to V 3 / V 1. Thus, for
the archetypal model, as is shown in Fig. 5, the time to
reach maximum temperature is about 104 V3/V t yr. F
Given the m i ( t ) and initial data for el, oi, these 2n
equations can be integrated to give el(t) , oi(t). A t the base
of the system we take m o = 0, corresponding to a source
isolated from depth. At the top of the system we take
I
106 en+ l = 0 and On+ ] = 0, corresponding to access of cold, non-
saline, meteoric water. Note that, in general, the ei, oi are
functions of P, T.
I have investigated a great variety of models of this type,
but here I shall describe only a very simple version of a
system with n = 3. In this case, as seen above, m l is deter-
mined by the gas column mechanics; m 2 is determined by
the strength of the circulation in the wet system; and m 3
is, in effect, abitrary, since it is determined by the details
of the surface topography and permeability in the opper-
most parts of the surface zone. Unless stated otherwise, I
shall restrict the discussion to m 3 = 0, that is, no surface
discharge, the condition for trapping the greatest ore
material mass in the upper reservoir.
(j) If the amount of available solute p, Vq diminishes to the
level at which it approaches the amount in the permeable
volume eVpo, where e is the porosity, typically a few
per cent, the body is vietually exhausted of solute.t At this
stage the concentration in the discharge will fall below the
saturated value. Ihave then taken 0 = q, but there is actually
negligible error in taking 0 -- 0.
All the solute equations are linear in solute concentra-
tion. Hence the data for the qi depend only on their ratios
and it is simplest to think of the values relative to the
t l f all the available solute were in the pore fluid as a saturated
solution
o l q = p , lpe ~ 10 3
for a porosity of 1%.

10
original concentration in the source. (This simplification
does not apply to the temperatures, since their relations
are non-linear, largely through the relation re(T).)
Role of solubility variation with temperature and pressure
The solubility of the material of interest is assumed to vary
strongly with temperature and pressure. This is certainly the
case for materials that have been measured in the laboratory,
but for the situation here we really have very little actual
information. What, for example, is the solubility and its
temperature variation of a heavy metal in a saturated
solution of (water + silica + sodium chloride + h~/drogen
sulphide)? We really don't know. I therefore represent
the variation of the solubility, s, by the empirical relation
s = s* exp (-G/7"), where G is a temperature scale, such that
if G is small the solubility varies little with temperature.
This relation can be interpreted as arising from the change
of free energy, kG, where k is Boltzmann's constant,
associated with the solution--dissolution reaction. The
complications that arise from changes of pressure are
ignored here. A good example is the solubility of NaCI for
which G ~ 500K provides a good fit over the temperature
range 0 - 8 0 0 ~ (three-phase boundary; data from Keevill~
Since NaCI is an almost ubiquitous and prominent solute
in hydrothermal fluids, as revealed by fluid inclusion
studies, I shall take its G value as typical. I have, in fact,
run models with G in the range 1 0 - 1 0 00OK.
Primary deposit
The amount of solute mass removed from V 1, V 2 as a
proportion of the maximum possible value as a function of
G is shown in Fig. 6. For values of G ~ 300K the bulk of
the solute material in the source is removed and deposited
1 -- Moss p r o p o r t i o n
10 -
ce
Primary ~ \ and 3000K
10'-
10-3 I I
102 103 10I*
G~K
Fig. 6 Proportion of mass taken from source and
primary deposit as a function of G
in the primary deposit and at most about 30% of this is
removed for deposition in V 3. For larger values of G
there is a rapid drop-off, since the solubility falls very
strongly with temperature.
Removal of a primary deposit
Under certain circumstances the primary deposit of
material dumped in V 2 at the top of the gas column may
be completely removed by the surface system. This
situation occurs readily if G1 >> G2, for then the solubility
in the gas falls rapidly with temperature: after the source
region has cooled only a little, the concentration of solute
in the gas column is small, whereas the solubility in the
GI=2000K
0,2
Primary deposit
for large G1
(GZ=100K)
2500K
~ 0'1
3000K
0 ~
0 50
t, kyr
Fig. 7 Proportion of mass in primary deposit as a
function of time for G2 = 100K and Gz = 2000, 2500
surface zone fluid is still relatively high - indeed, the sur-
face zone temperature may still by rising - so, after a
time of the order of ~, the concentration in V 2 begins to
fall. Some typical results are shown in Fig. 7, which shows
data for a system with G ] = 3000K, G 2 = lOOK, for
which the primary deposit is virtually removed after 5 x 104
yr.
This situation is probably rather unusual.
Ore material in modern hydrothermal systems
A characteristic of modern hydrothermal systems is their
conspicuous lack of or extremely low concentration of ore
materials. At first sight this is a surprise. Part of the
explanation is due to a misconception. In most cases only
a very small proportion of an ore deposit is represented by
11
the economically noticeable material. Otherwise the ex-
planation is due to sampling. Most of the hydrothermal
systems that have been explored, and only the uppermost
parts of those, are wet systems. These systems, such as
those of the Taupo area, operate close to single-pass
systems in which the amount of water recirculated is
very small - typically, a mere few per cent. Such systems
are continually flushing themselves out. On the other hand,
systems such as that at Lardarello with a dry surface zone
of low-pressure steam in which material has a very low
solubility just cannot transport material to the surface.
Final remarks
In order to present the results of this study in a simple
manner the description has been given for a single model
of an archetypal super-critical system. The specific quan-
tities mentioned are peculiar to this one system. Never-
theless, the numerical model itself, represented by relations
(a)-(/), is capable of a quantitative description of a very
wide range of systems.
The essential conclusion of this study is that the bulk of
the mass of water substance is transported during the super-
critical phase of the operation of the hydrothermal system,
and the integrated mass of water substance transported
in the deep part of the system is of the same order as the
mass of the source region, and, further, that the time-scale
of the process is typically of the order of 104 yr.
References
1. Barnes H. L. Geochemistry ofhydrothermal ore
deposits (New York: Holt, Rinehart and Winston, 1967),
670 p.
2. Elder J.W. Heat and mass transfer in the earth:
hydrothermal systems. Res Bull. N.Z. Dep. Sc~ Ind. 169,
1966, 115 p.
3. Elder J. W. Physical processes in geothermal areas.
Monogr. Ser. Am. geophys Union no. 8, 1965, 211-39.
4. Elder J. W. Steady free convection in a porous medium
heated from below. J. Fluid Mech., 27, 1967, 29-48.
5. Elder J. W. Transient convection in a porous medium.
J. Fluid Mech., 27, 1967, 609-23.
6. Elder J. W. The bowels of the earth (London: Oxford
University Press, 1976), 234 p.
7. Elder J. W. Quantitative laboratory studies of
dynamical models of igneous intrusions. In Mechanism of
igneous intrusion Newall G. and Rast N. eds (Liverpool:
Gallery Press, 1970), 245-60. (Geol. J., spec. Issue no. 2)
8. Gray D. E. ed. American Institute of Physics handbook
(New York: McGraw-Hill, 1957), v.p.
9. Helgeson H. C. Complexing and hydrothermal ore
deposition (Oxford, etc.: Pergam'on, 1964), 128 p.
10. Keevil N. B. Vapor pressures of aqueous solutions at
high temperatures. J. Am. chem. Soc., 64, 1942, 841--50.
11. Lindgren W. Mineral deposits, 4th edn (New York:
McGraw-Hill, 1933), 930 p.
12. Mason B. H. Principles of geochemistry, 3rdedn (New
York: Wiley, 1966), 329 p.
13. Park C. F. Jr and MacDiarmid R. A. Ore deposits
(San Francisco: Freeman, 1964), 476 p.
14. Roedder E. Fluid inclusions as sample of ore fluids.
Reference 1, 515-74.
12
Symbols
Symbols used in the text are listed here; also given are
nominal or typical values of some of the quantities. These
are the values used for the model calculations referred to in
the text, unless stated otherwise. S.I. units are generally
used, but note the use of length, area, volume in km, km 2,
km3; mass rate in ton sec1 = 103 kg sec'l; permeability in
darcy ~ 1012 m2; pressure in kbar; absolute temperature,
but quoted as ~ density in g c m a ; viscosity in poise, P =
0.1 N sec m'2; and various units of time.
a Source volume equivalent radius
c Specific heat: rock "~ 103 J kg "1 K "1
d Thickness of surface zone ~ 2 km
e Enthalpy of hydrothermal fluid
g Acceleration of gravity: g' = g~ko/P, reduced
acceleration
h Source depth: "~ 10 km
i Integer reservoir label
k Permeability: "" 1 darcy ~ 10"12 m 2
m Mass rate: ~" 1 ton sec1
n Concentration ratio: ~" 103
q Concentration in reservoir
$ Solubility
t Time elapsed
w Vertical velocity, upward
z Vertical coordinate, measured downwards from
the su rface z = 0
A Rayleigh number
cp Specific heat of gas
E Enthalpy change of magma on freezing
F Power
G Solubility temperature scale: "~ 300K
K. Thermal conductivity
L Latent heat of freezing of magma
M Integrated mass of water th rough sou rce/sou rce mass
N Nusselt number, dimensionless heat flow
P Pressure: Pd ~ 0.2 kbar
Q Total mass of deposit
S Area of cross-section: "~ 0.01 km 2
T Temperature, normally absolute, but quoted in ~
V Volume
W Source volume
G; Gas constant, p T/P
Viscosity coefficient,/~T "1/2
'7 Coefficient of cubical expansion: ~ 10-5 K"1
Length scale of boundary region
Porosity
0 Temperature difference or scale; characteristic
temperature of the surface zone
Thermal diffusivity: "~ 10-6 m 2 sec"1
Viscosity
Kinematic viscosity: magma 102 m 2 s e c l ; rock
10 ~6 m 2 sec-]
 Ratio used only in local working
P Density" rock p. ~ 3 g cm "3
(7 Concentration, mass proportion in fluid: "~ 10 ppm
T Time-scale (normally of total mass flow)" "" 104 yr
Volume rate o f f l o w
values of quantities such as permeability, all that is
required in the representation of the properties of water
substance is an adequate, but otherwise as simple as
possible, set of relations that describe the essential
features. The simple forms used here vary in accuracy,
but are typically + 10%, and this is more than adequate.

Subscripts
0 Unheated groundwater
1,2,3 Reservoir labels
8 Ambient
C Rayleigh number scale, A c
d Base of surface zone
h Depth of source
m Porous medium
Rock

Appendix

Properties of water substance

In the course of the work it is necessary to have a number
of convenient representations of the physical properties of
water substance. The relations used in this work are
collected here.
Equation of state An adequate approximation to the
equation of state of the super-critical gas is to represent it
as a perfect gas by taking P = opT, where a is a slowly
varying function of P and T, by choosing a to fit the ex-
pected range of conditions. For example, taking values
from the steam tables, at P "~ 1 kbar, T ~" 600~ gives
a ~ 300 J kg" ! .
BPD relation A particularly useful reference quantity is
the so-called 'boiling point with depth' (= BPD) relation,
which is the temperature as a function of depth in a static
column of liquid water in equilibrium and everywhere at
its boiling point. The BPD is the extreme temperature dis-
tribution possible in a wet hydrothermal system. In
100(50) 300~ the BPD is reached at the following depths
m: 0, 42, 162, 460 and 1064.
Gas viscosity Although, in general, it is somewhat simpler
to work in terms of y = IJ/p, the kinematic viscosity, we find
for super-critical water substance that the greater part of
the variation of u with P, T follows the variation of p.
Thus, it is better to work directly with an empirical .
representation of the viscosity/J.
There are elaborate and accurate representations of the
form of/J (P, T) available, notably those of Sutherland,*
but these involve unnecessary refinements. Thus, noting
the form of the Sutherland relationship, I take/~ = ~T "/2,
where/3 is a slowly varying function of P and T. At
P = 1 kbar over the range 200-1000~ an adequate fit
to the data of the steam tables is given with fl =
1.0 x 10"s P K "'/'. This gives nearly precise values near
250 and 600~ but is in error by about 20% near
400-1000~
Gas specific heat is about Cp = 2300 J kg"1 K "1 and is
nearly independent of P and T.
General remark about the representation of the properties
of water substance In view of the virtually unknown

*Gray D. E. ed. American Institute of Physics handbook (New

York: McGraw-Hill, 1957), v.p.

13
 deposition, at major ocean ridges, in marginal basins, in
Identification of ore-deposition island arcs and in oceanic islands. In many cases, however,
environment from trace-element particularly in the deformed and metamorphic rocks of
mountain belts, the geological evidence is ambiguous. Geo-
geochemistry of associated chemical techniques then have a useful application. Because
igneous host rocks the host rocks are commonly metamorphosed, stable trace-
elementgeochemistry 6 can be used to deduce their environ-
J. A. Pearce ment of eruption. Elements such as Ti, Zr, Y, Nb, 7.8'9 p,9
Cr 10,11 and the rare earths 12 all exhibit systematic
Department of Earth Sciences, The Open
differences between basalts from different tectonic settings
University, Milton Keynes, Buckinghamshire,
while remaining relatively immobile during weathering and
England metamorphic processes. In this paper we study the concen-
trations of these elements in the volcanic rocks related to
G. H. Gale mineralization in the Norwegian Caledonides, Newfound-
Department of Mines, Resourcesand Environmental land, Cyprus and Oman and in Precambrian terrain in
Management, Winnipeg, Manitoba, Canada Scandinavia and Canada.
Ore deposits related to continental magmatism can also
550.42:546:553.067 be studied by such a geochemical approach. It is fairly well
established that porphyry copper deposits are formed above
Synopsis destructive plate margins. TM The environment of formation
of tin deposits is much more enigmatic, however. The
The tectonic environment of formation of volcanogenic possibilities suggested 14-18 are post-orogenic, within-plate
massive sulphide deposits and porphyry tin and copper de- ('hot-spot') and destructive plate margin settings. Only
posits can be identified from the geochemical characteristics rarely can the precise setting of tin-bearing granites be
of the associated igneous rocks. The "stable" trace-element identified from geological criteria alone. Geochemical
geochemistry (involving Ti, Zr, Y, Nb, Cr and rare-earth methods are developed in this paper to distinguish between
elements) and geology of metabasalts related to 12 massive granitic rocks originating in 'within-plate' and 'volcanic arc'
sulphide deposits indicate that the deposits studied fall into settings. These methods are applied to some tin-bearing
four distinct classes. (1) Cyprus-type, including Cyprus, granitic rocks in Cornwall, Bolivia, Indonesia and Nigeria.
Oman and Betts Cove, possibly formed during the early
stages of back-arc basin development; (2) L r ken- type, Massive sulphide deposits
including Lr and York Harbour, possibly formed at
back-arc basin spreading centres; (3) Joma-type, including Magmatism in the oceanic environment
Joma, Rr and Bidiovagge, possibly formed in a small If (as is widely accepted) hydrothermal circulation of sea
ocean of Red Sea type; and (4) Gjersvik-type, including water through oceanic crust is the dominant process in the
Gjersvik, Buchans, Noranda and Lynn Lake, possibly formation of massive sulphide deposits, TM then, given a suf-
formed during an early stage of island arc evolution ficiently high geothermal gradient and permeable crust,
(Gjersvik), later during island arc evolution (the Buchans massive sulphides could form at almost any site of sub-
Kuroko-type deposit) or in a Precambrian setting (Noranda marine igneous activity. So, before studying the lavas re-
and Lynn Lake). Deposits related to major ocean ridge lated to these deposits, it is first necessary to geochemically
crests appear to be small and relatively uncommon, per- categorize lavas from major ocean spreading centres, margin-
haps because relatively few favourable sites for ore deposi- al basin spreading centres, continental margins, island arc
tion exist in such environment~ seamounts, ocean islands and seismic/aseismic ridges. To do
The environment of intrusion of acid-intermediate this, three main magma types are first defined: ocean-floor
igneous rocks can be deduced by use of diagrams based on
the element Nb (e.g. SiO2 versus Nb). On this basis, tin-
Table 1 Mean concentrations of trace elements (ppm) in
bearing granites can be classified either as within-plate
three main oceanic magma types (from literature survey
magmas (Nigeria) or as magmas from evolved volcanic arc
made in 1973 33); ocean island basalt column includes both
settings (e.g. Bolivia, Cornwall and Indonesia); the latter
tholeiitic and alkalic types
group can be further subdivided geologically into post-
orogenic and back-arc extensional settings. Although the Ocean-floor Island arc Ocean
continental crust may play a significant role in the forma- tholeiites tholeiites island basalt
tion of tin deposits, the geochemical data presented here
Ti 8400 5000 13 500
suggest that partial melting of tin-enriched mantle above a Zr 83 45 211
subduction zone may be the single most important genetic Y 28 18 27
factor. Nb 2.5 1.5 27
Cr 280 107 200
The purpose of this paper is to show how igneous geo- La 3.0 3.2 27
Ce 10 6.5 56
chemistry can be used to identify the tectonic environments Yb 2.5 2.0 2.3
in which volcanogenic ore deposits are formed. The paper K 1300 2700 5450
is divided into two parts: the first deals with deposits Rb 2.5 4.7 13
formed in a submarine environment (i.e. volcanogenic mass- Ba 8.5 60 176
Sr 120 175 415
ive sulphide deposits), and the second with deposits formed
U 0.09 0.15 0.64
in a continental environment - in particular, tin deposits. Th 0.16 0.37 0.77
Many works of synthesis have already been published on Cu 73 62 81
the relationship between massive sulphide deposits and plate Co 51 30 60
tectonics. 1-s These use mainly geological criteria to point V 23O 270 240
Ga 17 15 18 .
to a variety of possible environments of massive sulphide

14
basalts (which include tholeiitic and alkalic types), wit:bin-
plate basalts (which also include tholeiitic and alkalic types)
and island arc basalts (which include tholeiitic and calc-
alkaline types). Within-plate tholeiites are distinguished
from ocean-floor tholeiites by their higher concentrations of
most lithophile trace elements, the few notable exceptions
including Y, the heavy rare earths, Sc and Cr. Island arc
tholeiites are distinguished from ocean-floor tholeiites by
lower concentrations of small ion lithophile elements, in-
cluding Cr; their concentrations of large ion lithophile
elements are typically similar to or greater than those of
ocean-floor tholeiites. Table 1 illustrates the nature of
these geochemical differences.
Submarine Within -plate
basalt basalts
types
basin directly overlies a subduction zone, its magma type
might be expected to be intermediate between ocean-floor
and island arc type; similarly, magmas erupted during the
initial stages of sea-floor spreading could be intermediate
between ocean-floor and within-plate magma types. Fig. 1
illustrates the probable relationship between magma type
and tectonic setting. It emphasizes that geochemical studies
will often need to be supplemented by geological evidence
for an unambiguous interpretation to be made.
Geochemical characteristics o f the host lavas
Stable trace-element geochemistry, coupled with geological
information, is used here to characterize metavolcanic rocks
related to a variety of massive sulphide deposits. We deal
first with the geochemical information, which is summarized
in Table 2.
9 Joma type /
/
L~kken type /
9 Gjersvik type /
/
15000- o Cyprus type /
/

////

10 000

E
ck
/...'" "'',. _ _, . . . . . . , . . . . . , . , ..,. ,,.

5000 -
,*" 9 .."
- ~/,~ 9 .......

o~....... ..... . ,......,, ~ -"'"''"'""

Ocean- Inter-arc basin lsland
floor arc 0 r , , , I ' ' ' ' I . . . . I . . . . I
basalts basalts 0 50 100 150 200
Fig. 1 Distribution of three main magma types within Zr, ppm
ocean basins - illustrated by use of triangular diagram with Fig. 2 Mean analyses of metabasalts from various massive
main magma types at apices and transitional magma types sulphide deposit types plotted on T i - Z r discrimination
along sides diagram, s Dotted line denotes field of island arc basalts and
andesites, continuous line field of ocean-floor basalts and
In practice, lavas do not always fall neatly into one or dashed line field of within-plate basalts
other of these magma types. The main reason for this
appears to be the existence of transitional tectonic environ- Various discriminant diagrams can be used to identify
ments. For example, if a spreading centre in a marginal magma type. Fig. 2 is a plot of Ti against Zr 8 in which there

Table 2 Mean concentrations of 'stable' trace elements (ppm) in greenstones used in this study. Lavas from deposits 1, 2, 5,
7, 9 and 10 include samples from actual mines; remainder are samples from same geologic unit containing deposit, but not
from deposit itself
No. of
Deposit Location of lavas Ti Zr Y Nb Cr P2Os analyses

1 Cyprusdeposits 20, 21 Troodos Massif 11, 22 (axis Sequence) 5900 61 28 1 100 0.08 85
Troodos Massif 11, 22 (upper pillow lavas) 3270 31 15 1 405 0.06 45
2 Lasail,Oman 23 Lasail 23 3750 34 16 -- 109 0.05 15
3 BettsCove,24 Newfoundland Bale Verte 2s 1950 15 11 2 -- -- 3

4 York Harbour, Newfoundland 26 Bay of Islands complex 27 10700 125 32 -- 160 0.17 7
5 L~kken, Norway 2s, 29 L~kken 30 8550 93 26 4 221 0.09 38

6 RePros,Norway 31 Gjelsvik sequence 32 9850 129 32 - 114 0.14 18

7 Joma,Norway 28, 29 Joma 3o 11400 171 26 35 367 0.19 5
8 Bidjovagge,Norway Finnmark 33 8050 103 24 -- 171 0.09 18

9 Gjersvik,Norway 28, 29 Gjersvik 3o 8400 73 27 5 24 0.07 16

10 Buchans,Newfoundland Buchans 34 3900 75 -- -- 74 0.13 90
11 Lynn Lake, Manitoba 36 Rusty Lake greenstones 33,3s 5100 54 16 2 139 0.24 24
12 Noranda,Ontario 36 Noranda 37 7150 92 -- -- 157 0.13 78

15
 9 Joma type 100
a L~kken type
"7'5--
9 Gjersvik type
o Cyprus type

5.0-. 9 '
"~

.m

E
o 10
2.5- oo u

o
=Prate margin' \ Within-ptate o
basalt \ basalt 1 Joma
2 York Harbour
0 ' ' ' ' I ' ' ' ' I ' ' ' ' I ' ' ' ' I 3 Gjersvik
0 250 500 ?50 1000 &, L~kken
Ti/Y 5 Oman
Fig. 3 Mean analyses of metabasalts from various massive 6 Troodos
sulphide deposit types plotted on Z r / Y - T i / Y discrimination
diagram (analogous to T i - Z r - Y diagram of Pearce and
Cann 8). Diagram separates within-plate from other magma
types La C'e ' ' Sm
' Eu
' G'dT'b . . . . Tm V'b Lu
'

Fig. 5 Chondrite-normalized rare-earth patterns for typi-

is a general increase in Ti and Zr from island arc through cal samples of metabasalt f r o m various massive sulphide
ocean-floor to within-plate magma types. Fig. 3 is a plot of deposit types. Cyprus analysis from Kay and Senechal 38
T i / Y against Zr/Y. This distinguishes within-plate from
other magma types, making use of the distinctive enrichment
in Ti and Zr (but not Y) of within-plate basalts; Fig. 4 is a Cyprus-type (?early back-arc) deposits Deposits of this
plot of Ti and Cr 11 which separates the 'plate margin basalts' type were identified from the Troodos Massif, Cyprus, 2~ 21
of Fig. 3 into ocean-floor and island arc basalts, since for Oman 23 and Betts Cove, Newfoundland, 24 and all contained
Cu > Zn. Geologically, these deposits are situated in the
lava unit of f u l l y developed ophiolite complexes, from which
9 Joma type
it is implicit that they were formed at a constructive plate
a Lekken type
margin. Basaltic pillow lavas form the main host rocks.
9 Gjersvik type
Cyprus and Oman contain t w o lava units 21,22 _ the lower
o Cyprus type
basaltic, the upper containing rocks of komatiitic compo-
sition as well as basalts. The lower unit is thought to have
10 000 been erupted at the ridge axis; the upper off-axis outside
the zone of deep circulation of sea water. Ore deposits in
E 5OO0 these areas are mostly located at the u p p e r - l o w e r pillow
(3.
.._~
o o ~ Ocean-floor lava boundary, though some (such as Lasail, Oman 2 3 ) O c c u r
I-- _ ~ basalt
within the lower sequence and are cut by numerous feeder
Island arc
basalt dykes to the overlying lavas.
Geochemically, the lavas are characterized by low con-
1ooo centrations of the small ion lithophile elements, including Cr,
10 ' ' 's'o . . . lOO
. . . . . 5bd" boo '
and therefore classify as island arc tholeiites in Figs. 2, 3 and
Cr, ppm
4; the light rare-earth-depleted patterns in Fig. 5 are con-
Fig. 4 Mean analyses of metabasalts f r o m various massive sistent within this interpretation. Clearly, the geological and
sulphide deposit types plotted on log T i - l o g Cr 11 geochemical data are at variance. A consistent model can,
discrimination diagram

any given Ti concentration island arc basalts will nearly Table 3 Some important characteristics of Cyprus lavas
always contain lower Cr concentrations.
1 Two petrologically distinct lava units
Fig. 5 shows chondrite-normalized rare-earth patterns,
which provide useful additional information on magma type. 2 Concentrations of Ti, Zr and Y decrease up lava sequence 22
Ocean-floor tholeiites and island arc tholeiites show slight 3 Overall, Ti, Zr and Y concentrations are low tl, 22, and there are
strong light rare-earth element depletions 38
rare-earth element depletions; ocean-floor alkali basalts and
island arc calc-alkaline basalts show slight rare-earth element 4 Low Cr and MgO concentrations (at given Ti concentration) 11
enrichment; whereas within-plate basalts show moderate to 5 Basaltic komatiites present
strong enrichment in the light rare-earth elements. The 6 Orthopyroxene phenocrysts in some lavas tl, 22
massive sulphide deposits investigated in this paper are 7 8"/Sr/S6Sr ~ 0.7036 in rocks apparently unaffected by isotopic
associated with various basalt types listed above and can be exchange with sea water 39
divided into four categories according to the characteristics
8 Chromite deposits
of the host lavas. Each of these categories has been named
9 Underlying 'upper mantle' ultramafics very depleted 4o in their
after one of its member deposits, and its inferred tectonic
basaltic component
environment is given in parenthesis.

16
however, be built up by considering a transitional back-arc
environment where new oceanic crust is being formed at
spreading axes above subduction zones. There is some petro-
genetic evidence for this hypothesis. Table 3 lists some of
the ways in which lavas of the Cyprus type differ from rocks
dredged and drilled from major ocean ridges. Features 3, 5
and 9 in Table 3 can be explained in terms of high degrees
of partial melting under hydrous conditions; features 4, 6
and 8 can be explained by crystal fractionation under
hydrous conditions; 41 feature 7 can be explained by input
of water rich in radiogenic Sr derived from subducted
oceanic crust; features 3 - 7 are common characteristics of
island arc volcanic rocks. This particular type of back-arc
environment is therefore an attractive possibility, although
this is still at present only one of several hypotheses
proposed.
Lr (evolved back-arc) deposits L~kken,
Norway, 28,29 and York Harbour, Newfoundland,26, 27 fall
into this category. These deposits contain Cu > Zn and
Cu < Zn, respectively. The York Harbour deposit is found
in the lavas of the Bay of Islands ophiolite complex; the
Lr deposit occurs within an inverted lava sequence with
associated gabbros but no exposed ultramafics. Both
Lr and York Harbour apparently contain two basaltic
lava units, one above and one below the ore deposit.
Geochemically, the lavas contain higher concentrations
of all the small ion lithophile elements in comparison with
Cyprus-type lavas, although the upper lava units again appear
more depleted in these elements than the lower units. These
lavas classify as ocean-floor basalts in Figs. 2, 3 and 4 and
this is supported by light rare-earth element depletion In
Fig. 5. The question still remains whether the setting was in
a major or marginal ocean basin; lavas of similar character
have been found in both environments. The geological set-
ting of the L~kken lavas 42 and their high concentration of
large ion lithophile elements (which cannot be accounted
for solely by alteration) both favour a back-arc setting
similar to the Marianas basin at the present day. The situ-
ation for York Harbour is ambiguous in view of the conflict-
ing interpretations of Newfoundland geology, 43 and both
a major or marginal basin environment are possible.
Joma-type (continenta/ margin) deposits This category
includes Palaeozoic deposits from Joma, Norway, 26, 29
Rr Norway, 31 and the Proterozoic deposit at
Bidjovagge in Finnmark, Norway. 30 These deposits are all
related to greenschist facies metabasalts and more silicic
tuffs; metagabbros and serpentinites of uncertain association
are also found in these areas. Metamorphosed sedimentary
rocks overlie or are interbedded with the volcanics, and
these include argillaceous and arenaceous rocks in the
Rr area, and quartzites and limestones at Joma and in
Finnmark. Geologically, the environment appears to re-
semble that of 'Sullivan-type' 5 deposits.
Geochemically, the Joma samples contain high T i / Y and
Zr/Y ratios (Figs. 2 and 3) and light rare-earth-element-
enriched rare-earth patterns (Fig. 5), both indicative of a
within plate origin. The Rr and Finnmark lavas had
lower Ti and Zr concentrations, and this, coupled with
high Cr concentrations, points to an ocean-floor basalt
character. Geochemically, they closely resemble the
Lr type of greenstone. If the geology and the geo-
chemistry are combined, a model can be built up to suggest
that these deposits were formed at a continental margin, in
a small ocean of Red Sea type. At Joma both the geology
(shelf sedimentation) and geochemistry (the within-plate
character) support this. At R~ros and Finnmark the lavas
have an ocean-floor basalt character; but, as Fig. 2 shows,
they are chemically somewhat transitional towards within-
plate basalts; the presence of silicic tuffs and interbedded
sediments also argues against a normal oceanic ridge setting.
Their geochemistry and the geological settings are therefore
both consistent with their formation in a continental
margin environment.
Gjersvik-type (island arc) deposits This type of deposit,
defined by the peculiar geochemical characteristics of its
related basic lavas, encompasses several sub-classes of very
different geological character.
With regard to the Gjersvik type, the Gjersvik C u - Z n
deposit itself 28,29 lies within a pile of predominantly basic
metavolcanic rocks, although some andesitic and silicic
lavas, agglomerates and dykes are present. Geochemically,
the lavas contain moderate concentrations of Ti, Zr and Y
(Figs. 2 and 3), but very low Cr (Fig. 4), which classifies
them as island arc basalts; flat rare-earth patterns are con-
sistent with this interpretation. The nearby Z n - C u deposit
at Skorovas, described elsewhere in this volume, 44 has a
similar setting, but with more andesites and fewer basalts
in the volcanic pile; the geochemistry of the basic lavas
from Gjersvik and Skorovas is very similar. 44 The deposits
seem on this evidence to have formed during submarine
volcanism in an island arc setting.
Kuroko-type deposits have been described in detail
from Japan and elsewhere, 45,46 where the geological evi-
dence favours an island arc origin. There are, however,
several geological differences between these deposits and the
Gjersvik type of deposit. These differences include the
nature of the lavas (Kuroko depositsare mainly related to
acid pyroclastic rocks), and the ore mineralogy and texture
(Kuroko deposits contain Cu, Zn and Pb minerals with
barite and gypsum). Kuroko-type deposits appear to be
genetically related to small acid intrusions. Nevertheless,
published geochemical data on andesites from the Buchans
deposit, Newfoundland, 34 closely resemble data on the
andesites from the Gjersvik area.
The Noranda type of Z n - C u deposit occurs within
Precambrian greenstone terrain in association with acid
pyroclastic rocks. 47 Geochemical data on the more basic
rocks from the Noranda 37 and Lynn Lake 33,35 regions
revealed low to moderate Ti, Zr and Y concentrations
(Figs. 2 and 3) and low Cr (Fig. 4) - characteristics very
similar to those of basic lavas from the Gjersvik region. The
Lynn Lake lavas appeared to be tholeiitic in character, the
Noranda lavas more calc-alkaline. Thus, although we do
not yet know sufficient about tectonic processes in the
Archaean to interpret these data in terms of tectonic en-
vironment, we can say that the lava geochemistry most
closely resembles that from present,lay island arcs.
Tectonic setting of massive sulphide deposits
The main conclusion that can be drawn from this study is
that massive sulphides can form in a variety of environ-
ments. Most of the examples used here appeared to have
formed in a back-arc setting or during incipient formation
of an ocean basin; Fig. 6 illustrates the possible environ-
ments of formation of the various types of massive sulphide
deposit in a typical western Pacific arc-trench system.
The results also raise the important question: do all
settings in which submarine volcanism takes place have an
equal massive sulphide forming potential? The evidence
from this paper suggests that perhaps they do not. With
the possible exception of the small deposit at York Harbour,
no good examples have been found of massive sulphide
deposits related to the major ocean ridge types of volcanic
17
 Continental Marginal Third Inter-arc Island Arc --trench Trench
margin basin arc basin arc gap
. 9
~ ~ , a ~ j ~ Sea-Leve[

5
10
km
15
/" ~ I M
/, ,,~ s f 20
J
i

1 Joma type
2 Lc~kken type
3 Gjersvik type
4 Cyprus type
5 Kuroko type

Fig. 6 Possible distribution of massive sulphide deposit types in marginal basin setting. Section taken from Karig 8o

rock. This might be explained in terms of the absence of
true mid-ocean ridge ophiolites in the geologic record. A
few probable examples do exist, however, in Macquarie
Island, 4 the western Alps 49,50 and St~ren and Stavenes 30
in the Norwegian Caledonides. These do contain tiny
deposits, confirming that ore-forming processes do take
place in this environment; but there are no known deposits
of any major economic value.
If this apparent absence of deposits in major oceans is
real, it can be explained in several ways. Magmatic fluids
could be more important than has hitherto been considered;
perhaps magmatic fluids in a back-arc setting contain a
source of reduced sulphur derived from the breakdown of
pyrite to form pyrrhotite in subducted oceanic crust. Per-
haps the passage of the fluids through evaporites at some
continental margins yields an important source of sulphur.
Both these possibilities can be tested by thermodynamic
calculations. The most probable explanation may, however,
lie in the necessity to have a site for ore deposition. The
highly faulted ocean crust that exists during early spreading
and in back-arc settings could provide fault-bounded
troughs where low Eh and restricted sea-water circulation
would provide a favourable environment for ore deposition;
these conditions are less likely to be met in an evolved
ocean.
Tin deposits
Magmatism in the continental environment
Tin deposits are related to acid-intrusive rocks emplaced
into continental crust - at 'hot spots' within plates, at
destructive plate margins and as a result of c o n t i n e n t -
continent collision. These rocks belong to two main magma
types: volcanic arc magmas (which include rocks of the
tholeiitic, calc-alkaline and shoshonitic series) and within-
plate magmas (which include rocks of the tholeiitic and
alkalic series). There are substantial geochemical variations
both within and between these magma types, whose relation-
ship with tectonic setting is illustrated in Fig. 7.
Acid and intermediate within-plate magmas tend to have
high concentrations of both large and small ion lithophile
elements, with the exception of a few elements, e.g. Ti and
Sr, which have not behaved incompatibly during magma
genesis. The concentrations of the incompatible elements
increase from tholeiitic to alkali magma types. Volcanic
arc magmas also contain high conentrations of the incom-
patible large ion l ithophile elements - concentrations which
increase from tholeiitic through calc-alkaline to shoshonitic
types. By contrast, however, concentrations of small ion
lithophile elements are low, particularly in the island arc
tholeiitic series.
Continental magma types
Within- Thoteiitic~ ~ Atkatic
plate Continental rifts
magmas "hot-spot = volcanoes
Volcanic
I Tholeiitic ~ C a t c - a t k a l i n e ~ S h o s h o n i t e s ~
arc island a r c s Andeon-type continental
magmas arcs collision
ttMay include crustal melts
Fig. 7 Distribution of magma types within continental
environment. Magma types transitional between within-
plate and volcanic arc types do occur, but only in rare
tectonic settings
The most effective elements for identifying whether an
acid igneous rock has a within-plate or volcanic arc character
are therefore those with small highly charged ions. Nb is
particularly useful, and diagrams such as SiO2-Nb,
K 2 0 - N b and Z r - N b are all valuable in the identification of
magma type. The S i O 2 - N b diagram presented in Fig. 8
separates low Nb volcanic arc magmas from high Nb within-
plate magmas. An overlap occurs on this diagram between
tholeiitic within-plate magmas and high K calc-alkaline or
shoshonitic magmas from volcanic arc and post-orogenic
settings. K20 and Rb can be used to partly eliminate this
overlap, provided that K-silicate alteration has not taken
place. Acid within-plate magmas range from about 100 ppm
(at Nb = 15 ppm) to 300 ppm Rb (at Nb = 500 ppm),
whereas acid volcanic arc magmas range from about less
than 50 ppm Rb (at Nb = 5 ppm) to greater than 200 ppm
(at Nb :> 20 ppm). This rough guide suggests that acid
rocks which contain very high Rb and K20 concentrations
and which plot within the overlap field have a volcanic arc
origin.
Because so little is known about geochemical and mineral-
ogical variations in the lower crust and upper mantle it is
impossible to give a precise explanation for this diagram. A
few general points can, however, be made.
It is apparent from the magnitude of the Nb variation
( 5 - 5 0 0 ppm in acid rocks) that variations in the degree of

18
 500

9 Q 9 00 9

-- 9 9149 9 9 9 9 9 O 9 9 9
9 ~0 0 0 9 9 9 9
9 . . . 9 " 00

9 9 9 9 9 9 00
100 -- 9
9 9 9 00 9 000 9 9
_ 9 9 9 9 9 00 9 9 0 9 9 9
O O 9 9 0 9 9 9 9
9 9 9 9 9
0 9 9 9 O
5 0 - " - 9 9 "

9 9149149 9149 9 9 9 9 ~

9 9 9 9 1 4 9 194 99 1 4 99 9 9 ~ *
9 9 9 9 9
E --
9 9
0~
w+
9 00
+ + +
c2
9 9 9 9 + + + /

z 99 . 9 1 4 9 . . . 9 *
10_- 9 1 4 99 ++T + ~__________.,~. .

9** ~ **+ * * . * | l
5 "- . * .§ : § 2 4*****
7 § § . §247
I *
9+ * + I

++ * ++ , + .~. + + + + ~ I

9 + + ++ 4-++ + +
9 \ /
+ + + + + + % /
+ 4- * ++ + + + * +* + *

1 , , , + + + ++ + ++++ ++~+ .++

45 ' 50
' ' ' ' ' 55
' ' ' 60
' ' ' ' ' 65
' ' ' ' ' 710 ' ' ' ' 7'5

Si02, ~
9 Within-plate + Votcanic ,--,~ Estimated fietd of
magmas arc magmas "-- crustal melts

Fig. 8 Nb-SiO 2 diagram showing data from volcanic rocks of known tectonic settings. Single diagonal line gives lower
limit for within-plate magma type; double diagonal line gives upper limit for volcanic arc magma type. In general, Nb content
of within-plate magmas increases from tholeiitic to alkalic types, and Nb content of volcanic arc magmas increases from
island arc to shoshonitic types. Estimate of Nb-SiO 2 range of crustal melts derived from sediments metamorphosed up to
and including amphibolite facies superimposed on diagram

partial melting and/or crystal fractionation cannot alone cesses that affect K + and Nb 5+ to an approximately equal
explain why this element does discriminate between the extent. These could be caused by migration of small
various tectonic environments. The nature and composition amounts of incompatible rich melt 56 or represent variations
of the source rock may be the single, most important in the melting--recycling history of the mantle. In either
factor. case partial melting of mantle enriched in this way followed
Partition coefficients for Nb 33 between crystal and melt by crystal fractionation should produce acid melts rich in
are probably low (< 0.2) for olivine, orthopyroxene, Nb. There are also other factors that could also partially
garnet and plagioclase, moderately low (< 0.6) for clino- explain the Nb variations - in particular, the behaviour of
pyroxene, moderately high (> 1.0) for amphibole and high some of these Nb-bearing minor phases in the mantle. For
for micas ('> 2.5), magnetite, apatite and zircon. As a example, residual mica phases will reduce the Nb content
result, the presence of minor phases, particularly minerals of the melt in equilibrium with them.
such as phlogopite,52,,s3 in the source rock and their be- The arguments outlined above apply to magmas derived
haviour during melting will play an important part in from the mantle; but, of course, some acid magmas may
determining the concentration of Nb in the melt. Because have had a crustal source. A field for crustal melts has
of their high ionic potential, Nb 5+ ions will not be strongly therefore been superimposed on this diagram. This was
partitioned into aqueous fluids. Ions of low ionic potential, delineated by explaining the geochemistry of concordant
such as K +, will, however, be readily transported in these granitic veins in Precambrian terrains (see, for example,
fluids. Bowes 57 and Drury 58) and by carrying out petrogenetic
This last point may partially explain the geochemical calculations on the melting of metasedimentary rocks.
characteristics of volcanic arc magmas. The mantle wedge Nb-bearing phases, hornblende, biotite, muscovite or
above the subduction zone will be modified by aqueous phlogopite, are usually residual phases during partial melting
fluids derived from dehydration of oceanic lithosphere, s4, 55 of pelitic and siliceous sediments 59,6o,61 and this con-
These fluids should contain high K + concentrations and strains the melt to low Nb concentrations in most models.
relatively low Nb s+ concentrations. Thus, any acid magma There are, however, some exceptions to this: partial melting
derived from partial melting of this mantle followed by of eclogite and some granulites will leave pyroxene, garnet
crystal fractionation will not be enriched in Nb. By and plagioclase as residual phases, and high Nb melts can
contrast, heterogeneities in within-plate settings involve pro- therefore be produced.

19
 Most granites produced by melting of continental crust high concentrations of Nb, Zr, Y and the rare earths and
at depths where amphiboles and micas are residual phases plot in the within-plate magma field in Fig. 9. This interpre-
should, however, plot in or close to the area shown and will tation is consistent with geological evidence for a 'hot-spot'
geochemically overlap (and perhaps include) thole 9149 origin for these granites. 15,61,62 The other granites contain
within-plate magmas and magmas from volcanic arcs, par- much lower concentrations of these elements and, in Fig. 9,
ticularly as far as their Nb and SiO 2 concentrations are plot in the overlap zone between volcanic arc and within-
concerned. plate magmas; the very high Rb concentrations, however,
point to a volcanic arc magma type. Geological criteria are
Geochemical characteristics of tin-bearing magmas then needed to distinguish between an active volcanic arc
Table 4 summarizes some of the important geochemical and a post-orogenic setting. The Bolivian granites were
characteristics of tin-bearing granites from Nigeria, 62, 63, 64 clearly related to subduction processes; however, their
Cornwall,17,18 Bolivia 65 and Indonesia. 66 These data setting in the Eastern Cordillera of the Andes argues in
have been plotted on to the SiO2-Nb diagram (Fig. 9). favour of an origin associated with tensional stressesand
The common characteristic of these four granite pro- back-arc rifting rather than with the processes that produced
vinces is their very high concentration of K 2 0 and Rb. magmas in the Western Cordillera (the volcanic arc sensu
There are, however, significant differences in the concentra- stricto). The Cornish granites are geochemically almost
tions of other elements. The Nigerian granites contain very identical to the Bolivian granites, although geological opinion
is divided between a post-orogenic and a volcanic arc
Table 4 Mean concentrations (ppm) of some trace ele- setting.i7,18 The Indonesia granites contain concentrations
ments in tin-bearing granites (note that Nigeria data include of Rb, Zr and Nb similar to those of the Bolivian and
all Younger Granites; tin-bearing biotite granites have Cornish granites, but significantly higher concentrations of
Slightly lower concentrations of incompatible trace elements) Y and the heavy rare earths. This does not indicate any ob-
vious difference in tectonic environment (a back-arc rifting
Nigeria Bolivia Indonesia Cornwall setting is geologically likely 16), but it does suggest signifi-
cant differences in the role of garnet during the genesis of
SiO2, % 74.5 67.1 70.8 72.4
the magma.
Zr 650 173 130 78 To illustrate the composition of intrusive rocks related to
Y 220 19 64 13
Nb 170 21 15 15
porphyry copper deposits some analyses 60 from the region
of the Coed-y-Brenin porphyry copper deposit 67 in Wales
La 185 41 68 33
Sm 57 5.2 11.2 7.0
have been plotted in Fig. 9. They plot, as expected 68,69 in
Yb 34 1.6 4.7 0.9 the field of volcanic arc magmas, their very low Nb concen-
Rb 330 310 405 580 trations favouring an island arc rather than continental
margin setting.

500 [] Cornwall
H i g h Rb
9 Indonesia
m II
o Bolivia
u 9
II Nigeria
9 9 m 9
x Wales
100 9 9149 Ill

i 9 9

5O
E Within-plate _ ~ /
EL
EL

Z
Z
High Rb L w Rb 9
10

Volcanic arc ~x $
Crustal metts I
x x \\ /I
\

Low Rb

45 50 55 60 65 70 75
S i 02, %
Fig. 9 SiO 2 and Nb data from some tin-bearing granites plotted on to N b - S i O 2 diagram of Fig. 8. Also plotted for
comparison are analyses of intrusive rocks from Coed-y-Brenin p o r p h y r y copper district, Wales

20
 Source of tin-rich magmas
A number of hypotheses have been proposed to explain why
certain magmas should have high tin concentrations: these
include the melting of tin-rich detrital sediments 70 or
granitic rocks 71 in continental crust; melting of upper
mantle; 72 and evolution of fluorine-rich fluids derived from
breakdown of apatite in subducted oceanic crust followed
by mantle or crustal melting. 16 Here we briefly set out
some of the petrological constraints on such processes, em-
phasizing the deep-seated processes rather than the processes
that concentrate and deposit the tin at shallow crustal
levels. The following factors appear to be significant.
(a) As was stressed in the previous section, the only ob-
viously distinctive features of tin-bearing granites are that
they are emplaced into continental rather than oceanic crust
and that they are regional rather than local phenomena.
They can be formed in most tectonic settings.
(b) Examination of patterns of trace-element partitioning 73
and the few available Sn analyses of mineral phases suggests
that partition coefficients for Sn between crustal and melt
will be low for quartz, feldspar, pyroxene, garnet or olivine,
moderate for amphibole, but high (> 1) for micas and other
minor phases such as apatite, zircon and sphene. This means
that production of tin-rich magmas by partial melting is not
possible if micas and amphiboles make up a significant part
of the residue. Thus, partial melting of pelitic rocks or of
any rock metamorphosed in amphibolite or greenschist
facies is not likely to yield tin-rich magmas; partial melt-
ing of eclogite and granulite facies metabasic rocks or partial
melting of the mantle could yield tin-rich magmas, provided
that the tin-bearing phases, such as phlogopite, are complete-
ly melted and the source rocks themselves contain suffi-
ciently high tin concentrations.
(c) Although there are few thermodynamic data on tin
complexes, it is evident that tin has a particular affinity for
fluorine (see, for example, Mitchell and Garson 16 and
Barsukov 74) and that the partitioning of tin into a melt or
aqueous fluid will increase with increasing fluorine concen-
tration. The main fluorine-bearing minerals in the lower
crust and mantle are phlogopite and apatite (minerals which
also accommodate tin), so the stability of these minerals is
likely 88 be particularly important.
(d) It is well established 75 that fluorine-rich melts and
fluids can leach tin from mica lattices. Reaction between
fluorine-rich magma and wallrock may therefore be a potent
mechanism for tin enrichment.
(e) The small amount of available data shows that tin con-
centrations in mantle-derived basalts range from 0.2 ppm
(see, for example, Gill 76) in island arc tholeiites to > 5 ppm
in within-plate alkali basalts. In addition, a recent compila-
tion by Bailey and Macdonald 77 has shown that peralkaline
obsidians from continental settings contain significantly
higher F/CI ratios than those from oceanic settings.
The factors listed above have implications for the origin
of the four tin-bearing granite provinces discussed in this
paper. The Nigerian granites may give the simplest story,
since no subduction zone was involved in their genesis.
These granites appear to contain concentrations of Zr, Nb
and the rate earths that are much higher than those expected
for crustal melts, even when the proposed enrichment or
redistribution of these elements during wallrock alteration
has been taken into account. 76 Furthermore, their geo-
chemistry resembles that of acid rocks from ocean islands
where no continental crust existed. This evidence supports
a mantle origin for the granites - by partial melting and
subsequent crystal fractionation. Initial 87Sr/86Sr ratios
and lead isotope ratios 63 indicate that some crustal con-
tamination may have taken place, although without stable
isotope data this cannot be regarded as unequivocal evi-
dence. The tin itself could have both a mantle component
(see above) and a crustal component (tin-bearing pegmatites
have been recorded in the Pan-African basement 63, 64).
The other tin-bearing granites contain lower concentra-
tions of the small ion lithophile elements and some analyses
do overlap the field of crustal melts in Fig. 9; neither a
crustal origin nor an origin from mantle enriched by fluids
emanating from subducted oceanic crust can therefore be
ruled out. We suggest the following multistage model to
explain the origin of the Bolivian tin;bearing granites. This
model most closely resembles that of Mitchell and Garson, 16
although there are a number of important differences.
Stage one involves subduction of oceanic crust beneath the
South American continent. During the long time-interval
(Late Triassic-Tertiary) 79 between incipient subduction
and emplacement of tin-bearing magmas, the composition of
the mantle wedge beneath Bolivia is altered by input of
small amounts of melt/aqueous fluids derived from eclogite
in the subducted oceanic lithosphere. A mantle phase such
as phlogopite could accommodate both tin and fluorine
(and, of course, other elements, such as potassium) intro-
duced in this way.
Stage two involves partial melting of this enriched mantle in
a back-arc rifting environment. The tectonic process is
probably analogous to that which causes marginal basins to
form behind island arcs, but with one important difference
- the time<jap between incipient subduction and rifting is
much greater because of the great thickness of continental
crust. This time-gap may be necessary to permit sufficient
accumulation of tin in the mantle prior to partial melting.
Stage three involves ascent of magma to the surface. Fur-
ther concentration of tin can take place by crystal fraction-
ation and by wallrock alteration. Alternatively, this basic
magma may initiate partial melting of the lower crust; high
tin concentrations are likely if the source rocks are in
eclogite or granulite facies and are enriched in tin by volatile
transfer from the underlying basic magma.
The details of this model may need to be modified for
the Cornish and Indonesian provinces. For example, the
melting of enriched mantle in stage two might result from
continent-continent collision rather than back-arc rifting.
A similar model could also explain the Nigerian province;
the Nigerian Younger Granites could have been derived from
mantle that had been enriched in tin during the Pan-African
tectonothermal event but that had not been involved in
magma genesis until the Jurassic. It must be emphasized,
however, that this model is only one of several that could
explain the data available. Nevertheless, it does fit several
petrogenetic constraints and more and better petrogenetic
data will enable it to be rejected, modified or improved.
Conclusions
(1) Stable trace-element characteristics of metabasalts
related to massive sulphide deposits indicate that such de-
posits have been formed in a variety of different
environments.
(2) Geology and geochemistry together indicate that the
deposits studied formed at Red Sea and marginal basin
types of spreading centres and in island arcs: to our know-
ledge no large deposit was formed at a spreading axis
within a large ocean.
(3) We speculate that within oceanic crust really favour-
able sites for ore deposition may only exist in highly faulted
small ocean basins of Red Sea and marginal basin type.
(4) The geochemical characteristics and geological setting
of tin-bearing granites indicate that tin deposits of this type
can form within continental crust at 'hot spots', in exten-
21
sional regimes above subduction zones and in continental
collision environments.
(5) We speculate that tin-rich magmas can form by a
three-stage process that involves (a) a long period of enrich-
ment of the mantle above a subduction zone in tin and
other elements as a result of incipient melting of, or react-
ions within, subducted oceanic crust metamorphosed to
eclogite facies, (b) partial melting of this phlogopite-
bearing tin-enriched mantle as a result of back-arc spreading
processes, continental collision or 'hot-spot' activity and
(c) crystal fractionation, wallrock reactions and perhaps
crustal melting as this initially basic magma ascends to the
surface.
Acknowledgment
We gratefully acknowledge the assistance of the following:
Dr. A. O. Brunfelt, P-R. Graff, G. Marriner, Dr. P. J. Potts,
M. Sarre, B. Sundvoll and C. K. Winter for help and advice
on analytical techniques; Dr. D. Alderton, Dr. J. N. Grant
and Dr. R. H. Verschure for samples from tin provinces;
Martin Taylor (Prospection, Ltd.) for samples of drill core
from the Lasail prospect, Oman; Dr. W. R. A. Baragar, Tor
Grenne, Dr. W. J. Rea and M. Steeves for unpublished data;
Dr. J. R. Cann, J. Cramer, A. J. Fleet, Dr. C. Halls,
Dr. D. Roberts, Professor F. M. Vokes, J. B. Wright and
geologists from the various massive sulphide mines and pros-
pects for valuable discussions; and Professor I. G. Gass,
Dr. I. L. Gibson and J. B. Wright for constructive criticism
of the manuscript. Most of the rare-earth analyses were
carried out (by J.A.P.) at the Mineralogisk-Geologisk
Museum in Oslo during tenure of a Royal Society European
fellowship; X-ray fluorescence analyses were carried out at
the University of East Anglia, Norwich; Bedford College,
London; and NGU, Trondheim, Norway.
References
1. Sillitoe R. H. Formation of certain massive sulphide de-
posits at sites of sea-floor spreading. Trans. Instn Min.
Metal/. (Sect B: Appl. earth sci.), 81, 1972, B141-8.
2. Sillitoe R. H. Environments of formation of volcano-
genic massive sulfide deposits. Econ. Geol., 68, 1973,
1321-5.
3. Hutchinson R.W. Volcanogenic sulfide deposits and
their metallogenic significance. Econ. Geol., 68, 1973,
1223-46.
4. Sawkins F. J. Sulfide ore deposits in relation to plate
tectonics. J. Geol., 80, 1972, 377-97.
5. Sawkins F.J. Massive sulphide deposits in relation to
geotectonics. In press.
6. Cann J. R. Rb, Sr, Y, Zr and Nb in some ocean floor
basaltic rocks. Earth planet. Sci. Lett., 10, 1970, 7-11.
7. Pearce J. A. and Cann J. R. Ophiolite origin investi-
gated by discriminant analysis using Ti, Zr and Y. Earth
plant. Sci. Le~L, 12, 1971,339--49.
8. Pearce J. A. and Cann J. R. Tectonic setting of basic
volcanic rocks determined using trace element analyses.
Earth planet Sci. Lett., 19, 1973, 290-300.
9. Floyd P. A. and Winchester J. A. Magma type and tec-
tonic setting discrimination using immobile elements. Earth
planet. Sci. Lett., 27, 1975, 211-8.
10. Bloxam T. W. and Lewis A. D. Ti, Zr, and Cr in some
British pillow lavas and their petrogenetic affinities.
Nature, phys. Sci., 237, 1972, 134-6.
22
11. Pearce J. A. Basalt geochemistry used to investigate
past tectonic environments on Cyprus. Tectonophysics, 25,
1975, 41-67.
12. Herrmann A. G. Potts M. J. and Knake D. Geo-
chemistry of the rare earth elements in spilite from the
oceanic and continental crust. Contr. Mineral. Petrol., 44,
1974, 1-16.
13. Sillitoe R. H. A plate tectonic model for the origin of
porphyry copper deposits. Econ. Geol., 67, 1972, 184-97.
14. Halls C. Tin and plate tectonics. Ind. Eastern Eng.,
116, 1974, 493-4.
15. Sillitoe R. H. Tin mineralisation above mantle hot
spots. Nature, Lond., 248, 1974, 497-9.
16. Mitchell A. H. G. and Garson M.S. Relationship of
porphyry copper and circum-Pacific tin deposits to palaeo-
Benioff zones. Tran~ /nstn Min. MetalL (Sect B: Appl.
earth sci.), 81, 1972, B10-25.
17. Mitchell A. H. G. Southwest England granites: magma-
tism and tin mineralization in a post-collision tectonic
setting. Trans.Instn Min. Metall. (Sect. B: AppL earth sci.),
83, 1974, B95-7.
18. Bromley A. V. Tin mineralization of Western Europe:
is it related to crustal subduction? Trans. Instn Min. MetalL
(Sect- B: AppL earth sci.), 84, 1975, B28-30.
19. Spooner E. T. C. and Fyfe W. S. Sub-sea-floor meta-
morphism, heat and mass transfer. Contr. Mineral. Petrol.,
42, 1973, 287-304.
20. Hutchinson R. W. and Searle D. L. Stratabound pyrite
deposits in Cyprus and relations to other sulphide ores. In
Proc. IMA-IAGOD meetings "70, IAGOD volume (Tokyo:
Society of Mining Geologists of Japan, 1971), 198-205.
(Spec. Issue no. 3)
21. Constantinou G, and Govett G. J. S. Genesis of sulphide
deposits, ochre and umber of Cyprus. Trans. Instn Min.
MetalL (Sect. B: AppL earth sci.), 81, 1972, B34-46.
22. Smewing J. D. Simonian K. O. and Gass I. G. Meta-
basalts from the Troodos Massif, Cyprus: genetic implica-
tion deduced from petrography and trace element geo-
chemistry. Contr. Mineral PetroL, 51, 1975, 49-64.
23. Prospection Ltd. Personal communications, 1976.
24. Upadhyay H. D. and Strong D. F. Geological setting of
the Betts Cove copper deposits, Newfoundland: an example
of ophiolite mineralization. Econ. Geol., 68, 1973, 161-7.
25. Kidd W. S. F. The evolution of the Bale Verte line-
ament, Burlington Peninsula, Newfoundland. Ph.D. thesis,
University of Cambridge, 1974.
26. Duke N. A. and Hutchinson R.W. Geological relation-
ships between massive sulfide bodies and ophiolitic volcanic
rocks near York Harbour, Newfoundland. Can. J. Earth
Sci., 11, 1974, 53-69.
27. Williams H. and Malpas J. Sheeted dikes and brecciated
dike rocks within transported igneous complexes, Bay of
Islands, Western Newfoundland. Can. J. Earth Sci., 9, 1972,
1216-29.
28. Vokes F. M. Caledonian massive sulphide deposits in
Scandinavia: a comparative review. In press.
29. Vokes F. M. and Gale G. H. Metallogeny relatable to
continental drift in sothern Scandinavia. In press.
30. Gale G. H. Reconnaissance geochemical survey of
Caledonian volcanics and intrusives in central and south
 Norway: Part II - Analytical data. Rep. Norges. geol.
unders 1228A, 1974.
31. Rui I.J. Structural control and wall rock alteration at
Killingdal mine, Central Norwegian Caledonides. Econ.
Geol., 68, 1973, 859-83.
32. Grenne T. and Pearce J. A. Unpublished data.
33. Pearce J. A. Unpublished data.
34. Thurlow J. G. Swanson E. A. and Strong D. F. Geo-
logy and lithogeochemistry of the Buchans polymetallic sul-
fide deposits, Newfoundland. Econ. Geol., 70, 1975, 130-
44.
35. Steeves M. A. Geology of the Rusty Lake area.
Pub/. Manitoba Mines Brch. In press.
36. Hutchinson R.W. Ridler R. H. and Suffel G. G. Metal-
Iogenic relationships in the Abitibi Belt, Canada; a model
for Archaean metallogeny. CIM Trans., 74, 1971, 106-15.
37. Baragar W. R.A. Unpublished data.
38. Kay R. W. and Senechal R.G. The rare earth geo-
chemistry of the Troodos ophiolite complex. J. geophy&
Res., 81, 1976, 964-70.
39. Chapman H.J. Spooner E. T. C. and Smewing J. D.
87Sr enrichment of ophiolitic rocks from Troodos, Cyprus,
indicates sea water interaction. Trans. Am. geophys.
Union EeS, 56, 1975, 1074. (Abstract)
40. Menzies M. and Allen C. Plagioclase Iherzolite-residual
mantle relationships within two eastern Mediterranean
ophiolites. Contrib. Mineral. Petrol., 45, 1974, 197-213.
41. Nicholls I. A. and Ringwood A. E. Production of
silica-saturated tholeiitic magmas in island arcs. Earth
planet. Sci. Lett., 17, 1972, 243-6.
42. Gale G. H. and Roberts D. Trace element geochemistry
of Norwegian lower Palaeozoic basic volcanics and its tec-
tonic implications. Earth planet. Sci. Lett., 22, 1974,
380-90.
43. Strong D. F. Plate tectonic setting of Appalachian-
Caledonian mineral deposits as indicated by Newfoundland
examples. Trans. Am. Inst. Min. Engrs, 256, 1974, 121-8.
44. Halls C. eta/. Geological setting of the Skorovas ore-
body within the allochthonous metavolcanic stratigraphy of
the Gjersvik Nappe, central Norway. In Volcanic processes
in ore genesis (London: Institution of Mining and Metallurgy
and Geological Society, 1977), 128-51.
45. Horikoshi E. Volcanic activity related to the formation
of the Kuroko-type deposits in the Kosaka District, Japan.
Mineral. Deposita, 4, 1969, 321-45.
46. Clark L.A. Volcanogenic ores: comparison of cupri-
ferous pyrite deposits of Cyprus and Japanese Kuroko
deposits. In Proc. IMA-IAGOD meetings "70, IAGOD
volume (Tokyo: Society of Mining Geologists of Japan,
1971 ), 206-15. (Spec. Issue no. 3)
47. Sangster D. F. Precambrian volcanogenic massive sul-
phide deposits in Canada: a review. Pap. geol. Surv. Can.,
72-22, 1972, 44 p.
48. Varne R. and Rubenach M.J. Geology of Macquarie
Island and its relationship to oceanic crust. In Antarctic
oceanology I1: the Australian-New Zealand Sector Hayes
D. E. ed. (Washington, D.C.: American Geophysical Union,
1972), 251-66. (Antarct. Res. Ser. no. 19)
49. Bickle M. J. and Pearce J. A. Oceanic mafic rocks in
the Eastern Alps. Contrib. Mineral. Petrol., 49, 1975,
177-89.
50. Dietrich V.J. Evolution of the Eastern Alps: a plate
tectonics working hypothesis. Geology, 4, 1976, 147-52.
51. Delong S. E. Hodges F. N. and Arculus R.J. Ultra-
mafic and mafic inclusions, Kanaga Island, Alaska and the
occurrence of alkaline rocks in island arcs. In press.
52. Flower M. F.J. Schmincke H.-U. and Thompson R. N.
Phlogopite stability and the 87Sr/8esr step in basalts along
the Reykjanes Ridge. Nature, Lond., 254, 1975, 404-6.
53. Aoki K. Phlogopites and potassic richterites from mica
nodules in South African kimberlites. Contr. Mineral.
Petrol., 48, 1974, 1-7.
54. Ringwood A. E. The petrological evolution of island
arc systems. J/geol. Soc. Lond., 130, 1974, 183-204.
55. Gill J. B. Role of underthrust oceanic crust in the
genesis of a Fijian calc-alkaline suite. Contr. Mineral. Petrol.,
43, 1974, 29-45.
56. Green H.W. A CO2 charged asthenosphere. Nature,
phys Sci., 238, 1972, 2-5.
57. Bowes D. R. Geochemistry of Precambrian crystalline
basement rocks, North-West Highlands of Scotland. In
24th Int Geol. Congr. (Montreal: The Congress, 1972),
Sect. 1, 97-103.
58. Drury S. A. The chemistry of some granitic veins from
the Lewisian of Coil and Tiree, Argyllshire, Scotland. Chem.
Geol., 9, 1972, 175-93.
59. Brown G. C. and Fyfe W. S. The transition from meta-
morphism to melting: status of the granulite and eclogite
facies. In 24th Int. geol. Congr. (Montreal: The Congress,
1972), Sect. 2, 27-34.
60. Kilinc I. A. Experimental study of partial melting of
crustal rocks and formation of migmatites. In 24th Int
geol. Congr. (Montreal: The Congress, 1972), Sect. 2,
109-13.
61. Clifford T. N. Stumpfl E. F. and Mcllver J. R. A
sapphirine-cordierite-bronzite-phlogopite paragenesis
from Namaqualand, South Africa. Mineralog. Mag., 40,
1975, 347-56.
62. Bowden P. The geochemistry of some Nigerian igneous
rocks. Ph.D. thesis, Imperial College, London, 1961.
63. Bowden P. and van Breemen O. Isotopic and chemical
studies on younger granites from Northern Nigeria. Proc.
conference on African geology, 1970 (London: Common-
wealth Gological Liaison Office, 1971), 105-20.
64. Wright J. B. Controls of mineralization in the older and
younger tin fields of Nigeria. Econ. Geol., 65, 1970, 9 4 5 -
51.
65. Priem H. N. A. etal. Isotope geochronology in the
Indonesian tin belt. Geol. Mijnbouw, 54, 1975, 61-70.
66. Sillitoe R. H. Halls C. and Grant J. N. Porphyry tin
deposits in Bolivia. Econ. Geol., 70, 1975, 913-27.
67. Rice R. and Sharp G.J. Copper mineralization in the
forest of Coed-y-Brenin, North Wales. Trans. Instn Min.
Metall. (Sect. B: Appl. earth sci.), 85, 1976, B1-13.
68. Rea W. J. eta/. Igneous intrusion and metallogenesis
on and around the Harlech Dome. In Mineral exploitation
and economic geology 1974 (Cardiff: University College,
1974), pap. 8, 5 p.
69. Allen P. M. etal. Geochemistry of some igneous rocks
from the Harlech Dome and their relationship to mineral-
ization. Tran& Instn Min. Metal/. (Sect. B: Appl. earth sci.),
85, 1976, B100-8.
23
70. Schuiling R. D. Tin belts on the continents around the
Atlantic Ocean. Econ. Geol., 62, 1967, 540-50.
71. Bowden P. Origin of the younger granites of northern
Nigeria. Contr. Mineral. Petrol., 25, 1970, 153-62.
72. Noble J. A. Metal provinces of the western United
States. Bull. geol. Soc. Am., 81, 1970, 1607-24.
73. Jensen B. B. Patterns of trace element partitioning.
Geochim. cosmochim. Acta, 37, 1973, 2227-42.
74. Barsukov V. L. The geochemistry of tin. Geochemistry,
1957, 41-52.
75. Cramer J. Personal communication, 1975; Alderton D.
and Jackson N. Personal communications, 1976.
76. Gill J. B. Geochemistry of Viti Levu, Fiji, and its evo-
lution as an island arc. Contr. Mineral. Petrol., 27, 1970,
179-203.
77. Bailey D. K. and Macdonald R. Fluorine and chlorine
in peralkaline liquids and the need for magma generation in
an open system. Mineralog. Mag., 40, 1975, 405-14.
78. Alexiev E. I. Rare earth elements in younger granites
of Northern Nigeria and the Cameroons and their genetic
significance. Geochemiya, 1970, 192-8; Geochemistry
intn., 7, 1970, 127-32.
79. James D. E. A plate tectonic model for the evolution
of the Central Andes. Bull. geolo So~ Am., 82, 1971,
3325-46.
80. Karig D. E. Ridges and basins of the Tonga-Kermadec
island-arc system. J. geophys Re~, 75, 1970, 239-54.

24
Identification of the origin of ore-
forming solutions by the use of
stable isotopes
S. M. F. Sheppard
Isotope Geology Unit, Scottish Universities Research
and Reactor Centre, East Kilbride, Glasgow,
Scotland
539.155. 2:553.216.5: 553. 067
Synopsis
The four major different types of w a t e r - magmatic, meta-
morphic, sea water and/or connate, and meteoric w a t e r -
have characteristic hydrogen (D/H) and oxygen (180/I60)
isotope ratios. Applied to the analysis of isotopic data on
hydrothermal minerals, fluid inclusions and waters from
active geothermal systems, these ratios indicate that waters
of several origins are involved with ore deposition in the
volcanic and epizonal intrusive environment.
Water o f a single origin dominates main-stage mineraliza-
tion in some deposits: magmatic - Casapalca, Peru
( A g - P b - Z n - C u ) ; meteoric - Butte, Montana (Cu-Zn-Mn),
epithermal deposits, e.g. Goldfield, Tonopah, Nevada
(Ag-Au), Pachuca, Mexico (Ag-Au), San Juan Mountains
District, Colorado ( A g - A u - P b - Z n ) ; sea w a t e r - Troodos,
Cyprus (Fe-Cu), Kuroko, Japan (Fe-Cu-Pb-Zn). Solu-
tions of more than one origin are important in certain de
posits (magmatic and meteoric - porphyry copper and
molybdenum deposits) and are present in many.
In the porphyry Cu-Mo deposits the initial major ore
transportation and alteration processes (K-feldspar-biotite
alteration) are magmatic-hydrothermal events that occur at
750-500 ~C. These fluids are typically highly saline N a - K -
Ca-CI-rich brines (more than 15 vvt % equivalent NaCI). The
convecting meteoric-hydrothermal system that develops in
the surrounding country rocks with relatively low inte-
grated water/rock ratios (less than 0.5 atom % oxygen) sub-
sequently collapses in on a waning magmatic-hydrothermal
system at about 350-200~ These fluids generally have
moderate to low salinities (less than 15 wt % equivalent
NaCI). Differences among these deposits are probably in
part related to variations in the relative importance of the
meteoric-hydrothermal versus the magmatic-hydrothermal
events. The sulphur comes from the intrusion and possibly
also from the country rocks.
Deposits in which meteoric or sea water is the dominant
constituent of the hydrothermal fluids come from epizonal
intrusive and sub-oceanic environments where the volcanic
country rocks are fractured or well jointed and highly per-
meable. Integrated water/rock ratios are typically high, with
minimum values of 0.5 or higher (atom % oxygen) - the
magmatic water contribution is often "drowned out'.
Salinities are low to very low (less than 10 wt % equivalent
NaCl), and temperatures are usually in the range 3 5 0 -
150~C. The intrusion supplies the energy to drive the
large-scale convective circulation system. The sulphur comes
from the intrusion, the country rocks and/or the sea water.
Argillic alteration, which occurs to depths of several hun-
dred metres, generated during supergene weathering in many
of these deposits is isotopically distinguishable from
hydrothermal clay~
Introduction
A wide variety of hydrothermal ore deposits are intimately
associated with volcanic and sub-volcanic igneous processes.
This review examines the general principles of hydrogen and
oxygen isotope geochemistry as it applies to the determina-
tion of the origins of waters in such ore deposits. Although
the origin of both the metals and the hydrothermal fluids
responsible for transporting them have often been tied
closely to the genesis of the magma, emphasis is placed here
on the diverse origins of the fluids in volcanic and sub-
volcanic environments - largely as a result of the combined
hydrogen and oxygen isotope studies of the last 20 years on
active geothermal systems and post-magmatic hydrothermal
ore deposits. 1-6
Hydrothermal fluids are hot aqueous-rich solutions of un-
specified origin. Except for water, the origin of many of
the constituents in the fluid - magmatic, leached from the
country rocks, or from the source of the aqueous fluid (e.g.
sea w a t e r ) - is, in general, indeterminate. During the
passage of the fluid through the rocks the concentrations of
many of the elements are controlled by the solubility and
ion-exchange equilibria between the fluid and its wallrocks
at the prevailing temperature and pressure. The stable iso-
tope ratios of hydrogen (2H/1H or D/H) and oxygen
(180]160) vary in rocks, minerals and waters and also are
involved in f l u i d - w a l l r o c k exchange reactions. Neverthe-
less, in many situations the hydrogen isotope ratio is con-
trolled by the source of the fluid because there is relatively
so little hydrogen in the rock reservoir. This enables the
origin of the water to be determined, the oxygen isotopes
setting limits on how much water was involved in the hydro-
thermal system. Most of the other elements in the solutions,
with the notable exception of C, S, Sr and Pb, do not have
isotope ratios that vary measurably; therefore, they do not
have natural tracers that readily record their origin. Recent
reviews on the application of stable isotopes to problems in
ore deposits are by Doe and Stacey 7 on lead isotopes, Rye
and Ohmoto 8 on sulphur and carbon isotopes and Taylor 6
on hydrogen and oxygen isotopes.
Studies of the metal-bearing thermal brines discovered in
1963 at Satton Sea, California, excellently illustrate the
importance of stating nothing about the origin of the H20
in the absence of both hydrogen and oxygen isotope data.
When first reported, the brines, with 307 000 ppm of total
dissolved salts, 970 ppm Zn and 104 ppm Pb, were inter-
preted as being possibly of magmatic origin. 9 Later D/H
and 180]160 isotope studies demonstrated that at least 95%
of the water was of local meteoric origin. TM
Magmas and hydrotherma/ fluids
Before the stable isotope evidence for the origin of H20 in
hydrothermal fluids is explored, we briefly consider the con-
sequences of emplacing a water-bearing magma into an upper
crustal environment and the constraints that can be placed
on the generation and nature of hydrothermal solutions. A
rigorous discussion of water in magmas is outside the scope
of this review (see especially Burnham, 11 Davis and
Burnham, TM Hamilton and Anderson 13 and Jahns and
Burnham 14).
Most fresh basaltic and granitic rocks contain less than
0.6 and 1.0 wt % H20 , respectively. TM Although the solu-
bilities of the volatile components increase with increase in
pressure, there is a major solubility gap between water and
basaltic or granitic melts throughout crustal pressures. At
pressures greater than 400 bar the solubility of water in
magmas is greater than 1.0 wt %,16 Therefore, if magmas
are not dramatically undersaturated in H20 throughout
25
 much of their crustal history, they will lose a major I)art of
their original hydrogen content during their rise through the
crust.
In the volcanic and sub-volcanic (epizonal) environment
there is abundant evidence that some magmas become
saturated in volatiles - explosive volcanism, ash extrusion,
explosive intrusions, miacrolitic cavities, etc. Davis and
Burnham 12 have shown experimentally that the mechanical
energy released during boiling processes is approximately
equivalent to that consumed in volcanic eruptions
( " 1023 ergs km "3 of magma). Hence it is a potent energy
source for both intrusive and extrusive emplacement and
for fracturing rocks (development of 'crackled' texture in
porphyries). The separating fluid is a potential magmatic-
hydrothermal solution.
For a felsic magma to attain a volcanic or sub-volcanic
environment the initial water content cannot be greater than
about 3 wt %11 If such a magma becomes saturated in
H20, it does so at depths less than about 4 km. Magma
with a higher initial water content would become completely
crystallized, after boiling off most of its volatiles, at a depth
of several kilometres. Petrographical observations indicate
that a part, and a major part for some, of many epizonal
plutons has crystallized before the residual magma became
saturated in H20. Thus, 3 wt % H20 is probably sub-
stantially higher than the initial water content of most epi-
zonal plutons. The amount in excess of the final 0 . 6 - 1 . 0
wt % will form the magmatic-hydrothermal part of the sys-
tem unless it is lost directly to the atmosphere. The evolu-
tion of this hydrothermal fluid and its associated metaso-
matic activity will develop over a relatively narrow depth
range for a magma that initially is undersaturated in water.
Chemical studies of fluids from geothermal systems and
fluid inclusions in hydrothermal minerals give comparable
results and indicate that Na+, Ca2+ , K + and Mg 2+ are the
dominant cations and that c r , HCO~ and SO 2" are the
dominant anions, 17,18 and that water is usually, but not
invariably, the dominant constituent by weight. The total
solute concentrations of hydrothermal solutions range from
less than 10 "1 molal (about 0.6 wt % equivalent NaCI) to
3000 I I I I I of the magma and whether it became saturated in H20 but
Ab-Or-Qtz
~cotectic minimum
2000 -
_
&
lOOO_ \ _
I I I I
l, OI 500 600 700 800 900 1000
Temperature, oC
Fig. 1 Critical curve for system N a C I - H 2 0 and its re-
lationship to melting-point curve of granite. Numbers along
critical curve give NaCI content (wt %) of fluid. Modified
after Holland 21 and Sourirajan and Kennedy 22
26
greater than 6 molal (about 25 wt % equivalent NaCI).
Fluid inclusions with more than 50 wt % salts, essentially
hydrous saline melts, are known from high-temperature
parts of many porphyry deposits. 19,2~ When a magma be-
comes saturated in an H20-bearing fluid this may be highly
saline because such elements as Na, K, CI, etc., are strongly
partitioned into the aqueous phase coexisting with a silicate
melt.11,21
The system N a C I - H 2 0 can be used as a simplified model
of a hydrothermal fluid. The critical curve in the NaCI-
H20 system (Fig. 1 ) intersects the minimum melting curve
for granite at about 700~ and 1240 bar. 22,21 Aqueous
fluids separating from a magma at P < t240 bar can be
present as either one or two H20-bearing fluids. When two
aqueous fluids are present, one is a highly saline, high ~
density brine (> 50 wt % salt; > 1 g cm "3) and the other a
low-salinity, low-density vapour (< 5 wt % salt; "" 0.1 g
cm'3). 23 Fluid inclusion evidence for two H20 fluids
(boiling) is widespread, particularly from porphyry copper
deposits. 19 Because of the density differences, the two
f l u i d s could separate and develop into two different hydro-
thermal systems with different salinities. Apart from the
N a C I - H 2 0 system, there are no extensive data available
on two-fluid systems that are chemically more similar to
natural hydrothermal fluids, and, in particular, on how ele-
ments and isotopes are partitioned between the coexisting
high- and low-density fluids. These processes are probably
fundamental in the sub-critical region of Fig. 1 - the vol-
canic and sub-volcanic environment. In the absence of data,
we assume that the hydrogen and oxygen isotope fraction-
ations are small or negligible during such two-phase separa-
tion processes, in the H20 liquid-vapour system hydrogen
and oxygen isotope fractionations are very small, exc.ept for
temperatures less than 200~ 24
The country rocks of an epizonal pluton are probably
permeable water-bearing rocks; in fact, the very emplacement
processes may generate fracture permeability. 12, 25 Hence,
we have to consider the nature of these fluids, their origins
and chemistries, and how they have interacted with the
igneous system. The results of the emplacement of an epi-
zonal igneous system do not just depend on the composition
also critically on the relationships between the igneous sys-
tem and its environment, country rocks and water, at both
the magmatic and post-magmatic stages. Knowledge of the
origins and chemistries of the hydrothermal fluids and the
nature of their involvements with the igneous system during
the different stages of its development is fundamental to any
theory of ore genesis.
Some principles of hydrogen and oxygen isotope
geochemistry
Isotopic notation Variations in the isotope ratios of
hydrogen and oxygen are measured mass-spectrometrically
on H 2 and CO 2 gases, respectively, extracted quantitatively
from the minerals and waters. Because changes in the iso-
tope ratio are of more interest than knowledge of the
absolute isotope ratio, variations are measured relative to an
arbitrary, convenient standard. Ratios are reported in the
notation in parts per thousand (per mil, %o) where
(~X = r Rsample 1] 1000
L Rstandard J
For oxygen Sx = $180 R = 180/160
For hydrogen Sx = ~ D R = D/H
The standard for both hydrogen and oxygen is standard
mean ocean water or SMOW. 26 In nature the D/H ratio is
about 1/7000 and the 1 8 0 / 1 6 0 ratio is about 1/500. Varia-
tions in these ratios are typically measured with an accuracy
of about 1% o forD/Hand0.1% 0 for oxygen. Thus, a
value of - 1 0 % o means that the sample is 10 parts per
thousand or 1% depleted in deuterium or 180 relative to
SMOW.
The 6D value of a hydrous mineral refers to the hydroxyl
hydrogen, which is not readily exchangeable with water
vapour in the atmosphere. This is measured after the re-
moval of all labile absorbed and interlayer water that may
have been present in the sample. The ~ 180 value of minerals
refers to the structurally bound oxygen (e.g. silicate-oxygen
plus hydroxyl oxygen), which again is measured after re-
moval of all absorbed water.
The partitioning of two isotopes between two species, X
and Y, is described by the fractionation factor, (~XY, where
OIxy = RX/R Y
able. In such cases the fractionation curves have been
derived by combining experimental data on other minerals
with data derived from natural mineral assemblages where
there is evidence for the preservation of isotopic
equilibrium.27, 33
Equilibrium hydrogen and oxygen isotope fractionation
curves for several different gases and minerals were sum-
marized in Figs. 1 - 4 of Taylor. 6 The uncertainties associ-
ated with some of these mineral systems introduce problems
in detail, particularly in geothermometry. For the calcula-
tion of the isotopic composition of a water in equilibrium
with a mineral that is of immediate interest, however, these
problems are less critical because high precision is not
usually essential.
0 i
Wl / A /
-50 - ~ ""--- "--. z~,, -

9 " ~ " "- .......j

"0 0oC / /4'
and R is the isotope ratio D/H or 1 8 0 / 1 6 0 in the species.
From the definition of 6 it can be shown that

W2/ B _ ///
1000 In eXY ~" $ x - 6 y -1oo - ~-.-~.~ . ~ 200o c / -

This is a very good approximation for 6 < 10. I

, "0 " " ~"~"
10C
-- .~ // C) Kaolinite
Equilibrium isotope fractionations Although the isotopic L ~"T'XJ [] Woter
composition of either thermal waters or fluid inclusions can -- 150 II -

be measured directly, there are many situations in which it

is necessary to calculate the isotopic compositions of water
in equilibrium with the selected mineral. Taylor 6 and
Bottinga and Javoy 27 have reviewed the hydrogen and
oxygen isotope fractionation factors as a function of tem:
perature in mineral-water-gas systems. The general princi-
ples are illustrated here by consideration of the system
kaolinite--water.
At equilibrium, analogous to the partitioning of chemical
species among coexisting minerals, the isotopic species dis-
tribute or fractionate themselves among the available sites in
the coexisting minerals and fluid in a well defined manner
dependent on their vibration frequencies. For the system in
which kaolinite-water is in isotopic exchange equilibrium,
the following exchange reaction can be written:
Kaolinite (h__)+ water (1_) = kaolinite (1_) + water (h_.)
where hrefers to the species with all the hydrogen and oxy-
gen as D or i s 0 , and 1refers to the species with all the
hydrogen and oxygen as 1H or 160. At equilibrium the
hydrogen or oxygen isotope fractionation factor,
(](hydr~-~=n
kaolinite--water or ~oxygen
~kaolinite--water, is related to the equilibrium
constant for the reaction. These (~ are temperature-depend-
ent, but independent of pressure because of the essentially
identical volume of the heavy and light isotopes. They form
the basis of the isotope geothermometers. Fractionation
factors may also be sensitive to the chemical composition of
the mineral; this effect is particularly important for hydro-
gen in such minerals as biotite or amphiboles. 28
For gases, and to a limited extent for minerals, the
equilibrium isotope constants for an exchange reaction can
be calculated by the methods of statistical thermodynam,
ics.29,30 For mineral systems in general, these temperature-
dependent equilibrium fractionation factors are determined
experimentally in the laboratory or from studies of active
geothermal systems. 4,31,32 For several important rock-
forming minerals there are no such experimental data avail-
-10 0 10 20
6180
Fig. 2 Hydrogenand oxygen isotope fractionations be-
tween kaolinite-H20 at 10 and 200~ for waters of two
different isotopic compositions, WI and W2. Dashed lines
join kaolinite-H20 pairs. M L and KL are meteoric water
line 26 and kaolinite weathering line,34 respectively (see
text)
As an example of mineral-water isotopic fractionation
the system kaolinite-water is considered in Fig. 2. Two
different waters, initially with hydrogen and oxygen isotope
compositions W1 and W2, are shown in equilibrium with
kaolinite at 10 and 200~ At 10~ the kaolinite concen-
trates 180 relative to the water by about 27%o, but the
water concentrates deuterium relative to the kaolinite by
about 30O/o0.34 Although the kaolinites K1 and K2 have
different ~D and $180 values, the kaolinite-water hydrogen
and oxygen isotopic fractionation factors are the same at
10~ Therefore
~ D w - 1 --(SDK-1 ~ ( ~ D w - 2 - (~DK-2
(~180K_1 --(~18Ow_ 1 " ~ 1 8 O K _ 2 --~18Ow_ 2
because 1000 In O~K_ w ~ (~K -- (~W
If we now raise the temperature of our two kaolinite--
water systems to 200~ a situation such as that shown in
Fig. 2 might be observed. The isotopic ratios of kaolinite
and water have changed because of hydrogen and oxygen
isotope exchange and equilibration. Although the ~D and
~180 values of systems 1 and 2 are different, the differences
between the ~ values for kaolinite and water are essentially
identical for both hydrogen and oxygen isotopes because we
have equilibrium at the same temperature, 200~ This
illustrates how we can either determine the temperature, if
we can analyse co-genetic minerals and water, or, if we
know the temperature, how we can determine the isotopic

27
composition of the water in equilibrium with the mineral.
Other mineral-water systems behave in a similar way. A t
some temperatures within the hydrothermal range minerals
such as biotite and chlorite are depleted in both 180 and D
relative to the coexisting water. This reflects differences in
the characteristics of mineraI-H20 fractionation factors
and is not a difference in principle.
Water/rock ratios Fig. 2 also illustrates how limits can be
set on the water/rock ratio - the amount of hydrogen or
oxygen in water relative to the amount of hydrogen or oxy-
gen in rock. Let us consider the kaolinite-water systems
1 and 2 at 200~ system 1 having a smaller water : kaolinite
ratio than system 2. Comparing systems 1 and 2, W1 has
undergone a larger '180 shift' away from the meteoric
water line to point A than has W2 to point B. The magni-
tude of the negative '180 shift' for kaolinite is reversed,
K1 experiencing a smaller change in 6180 than K2. The
size of the '180 shift' is related to the amount of oxygen in
the water relative to that in the kaolinite and the tempera-
ture. For the simplest closed-system model at equilibrium
these '180 shifts' can be expressed as water/rock ratios: 4
w$ i +r i +r f
H20 $rock= W $ f 2 0 $rock
where w is atom % oxygen in water, r is atom % exchange-
able oxygen in rock, i is initial value and f is final value.
Then
f i
w $ rock -- $ rock
r ~'.~o-~.=o
At 200~ for system 1 the water/rock ratio is 0.7 (atom %
oxygen) and for system 2 it is 3.6. The water/rock ratio by
weight is approximately twice the atom % oxygen value.
This measure of the integrated water/rock ratio gives a
minimum value only because (1) a lot of water may pass
through the rocks without equilibrating and (2) the water
entering the volume of rock under consideration may have
experienced an '180 shift' of unknown magnitude from the
original isotopic composition it had before entering the rock
system as a whole. The larger the extent of this shift, the
greater will be the difference between the measured water/
rock ratio and the 'true' ratio. 4
A similar water/rock ratio expression can be written in
terms of hydrogen isotopes. Because the hydrogen content
of rocks is usually so low (rock with 1 w t % water contains
"<).11 w t % hydrogen), this isotopic measure of water/rock
ratios is only useful for systems in which the ratio is
extremely small (less than about 0.05 wt % oxygen).
The concept of water/rock ratio is very important for
our models of hydrothermal systems. Some care has to be
taken in interpreting its meaning, however, particularly be-
cause water/rock ratios may vary on a small scale, in a
system with well defined veins o~ ch~nnelways the ratio in
a unit of rock at the wall of the channelway may be sub-
stantially larger than that in a similar unit of rock located a
few centimetres away from a vein.
Isotopic characterization of natural waters
There are two ways of determining the isotopic composition
of a hydrothermal water: (1) directly by measuring the
D/H and 180/160 ratios in a water sample, a hot spring
water or fluid inclusion water, and (2) indirectly by calcu-
lating the hydrogen and oxygen isotope composition of a
water in equilibrium with an analysed hydrous mineral at
the appropriate temperature by use of predetermined
28
mineraI-H20 fractionation data. In a few cases both
methods have been applied to the same material. Such data
have demonstrated that the methods are equivalent within
the experimental uncertainties. 38 In practice both methods
have their advantages and disadvantages.
In the direct method knowledge of the temperature and
mineraI-H20 fractionation factors is not required. Because,
usually, several grammes (often 3 0 - 4 0 g) of a mineral have
to be crushed in a vacuum line to extract sufficient water for
analysis, there are often major problems associated with the
selection of suitable material. Samples with both primary
and secondary inclusions and recording a multistage
mineralization history are common. In the analysis of such
samples co-genetic fluid inclusions cannot be selectively
crushed. Fluid inclusions from some higher-temperature
assemblages where boiling occurred are extremely complex
on a microscopic scale and hydrated daughter minerals may
be present. 19 Oxygen isotope exchange between the H20
in the inclusion and an oxygen-bearing host mineral prevents
meaningful interpretation of the oxygen isotope analysis
from such samples. 38,3~ Leakage, especially of hydrogen,
which has large H 2 0 - H 2 fractionation factors at hydro-
thermal-type temperatures, may be critical in some
inclusions.
In the indirect method knowledge of the temperature and
mineraI-H20 fractionation factors is required. Either or
both of these may be known less precisely than is desirable.
Selection of suitable material, particularly from the igneous
and high-temperature part of the system, may be less of a
problem. Retrograde isotopic exchange may occur during
the often complex cooling history.
In hydrothermal systems with multistage mineralization
there may be problems of relating the isotopic composition
of the solution, derived by either method, to a specific
mineralization event. Gangue minerals that coexist with ore
minerals are not necessarily co-genetic.
Types of waters
Various kinds of waters can be defined in terms of their
genesis. Sea water is of the seas or oceans at any time or
derived from the ocean with little discernible modification.
Connate water is fossil water trapped in the pores of the
sediment at the time of formation. This generally, but not
invariably, is used to refer to trapped sea water. If a pore
water is of non-sea water origin, we cannot in general re-
cognize whether the water was trapped contemporaneously
with sedimentation (connate, sensu stricto) or was flushed
into the pores at some later time (non-connate origin).
Therefore the term is best restricted to fossil sea water.
Meteoric water is non-marine water - rain, snow, ice, river,
lake and most low-temperature groundwaters - of any age
that was derived from ocean water through atmospheric
circulation processes. Magrnatic water is that which has
equilibrated with or has come from magma, regardless of its
ultimate origin. 4 Metamorphic water has equilibrated with
or has come from metamorphic rocks undergoing dehydra-
tion. Juvenile water is that which is being released from the
mantle and which has never been part of the hydrosphere.
It is not considered further here because it has not been
knowingly recognized.
There are two other water terms that by themselves are
best used descriptively only. A suitable prefix can be added
when there is genetic evidence for its origin. Fo~mation~
water is water of no specified origin that is found in the
pores of sediments. Hydrothermal water is hot aqueous
fluids of any origin. When the dominant source of the fluid
has been identified (is0topically in general) the following
are used: magmatic-hydrothermal, meteoric-hydrothermal,
sea-water-hydrothermal, or metamorphic-hydrothermal.
Isotopic characteristics of major water types
Sea water Knowledge of the isotopic composition of the
oceans and its variation with time is critical for two imme-
diate reasons: (1) the isotopic compositions of meteoric
waters are related to the ocean water composition, and (2)
for problems related to the interaction of sea water with
oceanic rocks.
, , , peratures are about 25~
0- Modern meteoric waters s~'//~JSMOW
Most met. W ~
. /b'\Ocea~
waters
Mediterronean--------7///
SO= 861%+10
,,';"/
-
//J Evaporation170---100~
</i-'"
-100 / - ,/~-" - i,.Exchange . - 0o c
-150 f" I I I I
-2 -15 --10 -5 0 +5
6180, %=
Fig. 3 Relationship between D and 180 variations in
ocean waters and various meteoric waters. 28, 37, 38, 39, 42, 43
Approximate relationship between mean annual air tempera-
ture and mean annual isotopic composition of meteoric
waters observed at many localities is shown. 42 Isotopic
trends are given for meteoric waters undergoing exchange
with 18a-rich rocks, and evaporation in acid hot spring
type environment
The present hydrogen and oxygen isotope composition
of ocean water is relatively uniform with 8D = +5 to - 7 ,
8180 = +0.5 to --1.0 (Fig. 3) and with our standard SMOW
very closely approximating to the average ocean water
value. 37,38,39 Larger variations do exist, but only in very
special situations. At very high latitudes a large glacial melt
contribution depletes the ocean water in D and 180. Seas
with restricted access to open oceans, such as the
Mediterranean and Red Sea, are modified by evaporation
processes and are enriched in D and 180 and salinity
relative to 'SMOW'.
Before the growth of the ice caps, particularly Antarctica,
which are strongly depleted in D and 180, in the Miocene
the mean isotopic composition of the oceans was about
8D = - 7 , 8180 = - 1 . 40 From our present knowledge of
the oceans and their evolution, at least since the late Pre-
cambrian, there is no compelling evidence for variations
larger than about 1%o for 8180 from the present mean
ocean water value. 6,41
Although these variations are quite small relative to those
discussed below, it is important, however, to bear in mind
that this is an exceedingly complex problem, which requires
closer study.
Meteoric waters The hydrogen and oxygen isotope com-
positions of present-day meteoric waters (Fig. 3) vary in a
very systematic way.26, 37,42, 43
8D = 8 8 1 8 0 + 1 0 % o primary magmatic water range as defined above. (The D/H
Most meteoric waters that have not undergone extensive
evaporation plot within a band up to -+ 1 ~ in 8180 wide
of this line. The recently proposed Craig-Epstein meteoric
water line (SD = 8 8180 + 5) lies within the band mentioned
above. Meteoric waters in the Japan and Mediterranean
regions typically are depleted in 180 by 1 to 1.5 relative to
most other meteoric waters that have the same 8 D value.
These small variations reflect differences in the isotopic
composition of the source water and in other meteorological
processes.
Most of the water vapour held in clouds originates from
the equatorial regions of the oceans, where average air tem-
Hydrogen and oxygen isotope
fractionations are associated with the evaporation-conden-
15~ sation processes. Rain leaving a cloud is preferentially en-
riched in D and 180. Therefore, as the air mass moves to
higher latitudes, the precipitation becomes progressively more
--10o j depleted in D and 180 (Rayleigh distillation process). The
meteoric waters that are most depleted in D and lSO are
observed in samples from polar regions (SD ~" - 4 4 0 ;
8180 --- -55). 44 Although there are variations in detail and
.~ local exceptions, owing to geophysical, topographical and
seasonal effects, the annual mean isotopic compositions of
meteoric waters are broadly related to the latitude-eleva-
tion or, alternatively, to mean annual air temperature
=E (Fig. 3). The meteoric water relationship provides us with
a very powerful natural tracer.
-10~ The meteoric water relationship is critically related to the
isotopic composition of the major source for the water
vapour. Insofar as the isotopic composition of the oceans
has been relatively uniform (+ 1%o for 8180) since at
least the late Precambrian, we can assume that ancient
meteoric waters followed a relationship very similar to that
of today.
Magmatic waters Waters of undisputed magmatic origin
cannot be directly sampled with certainty. Nevertheless,
by use of the experimentally determined mineraI-H20
fractionation factors the isotopic compositions of H20 in
equilibrium with magmas of different composition can be
calculated at temperatures of about 700-1200~ The
great majority of fresh volcanic and plutonic rocks have
very uniform 8D values (-50 to - 9 0 % o ) and 8180 values
(+5.5 to +10.0% o) (see Figs. 10 and 11 in Taylor s), and
these values are referred to as 'normal' igneous values.
Waters in equilibrium with 'normal' magmas have composi-
tions in the range 8D "" - 4 0 to - 8 0 % o , 8180 "" +5.5 to
+9.5%0. Such waters are defined as primary magmatic
water, without any implication concerning their ultimate
origin.4, 6
Insofar as 'normal" basaltic and granitic magmas typic-
ally have a more restricted range of 8180 values within the
'normal" igneous range, 5.5-7.5 and 7.0-10.0, respectively,
it is sometimes convenient to use a more restricted range of
values for primary magmatic water. For example, in oceanic
environments where granitic magmas are typically absent
primary magmatic waters will have a more restricted range
of oxygen isotope compositions (8 180 -- +5.5 to +8.048).
Both magmas and fresh igneous rocks are known which
have compositions that lie outside these 'normal' values.
For example, some recent basaltic flows on Iceland have
8180 values as low as +2%o, 3.5 to 4% odepleted relative
to average basalt; 46 The Tuscan igneous rocks north of Rome
have 8180 values in the range +11.2 to §176 47'48 gran-
ites from the Cornubian batholith, southwest England, have
values as high as +13~ 49 Waters in equilibrium with
these magmas will have compositions that lie outside the
29
ratios from both the Cornubian batholith and Tolfa district often saline (up to 400 000 ppm total dissoved solids) and
of the Tuscan province have 'normal' values.) Such 1SO-rich hot (usually < 100~ the hottest waters usually being the
calc-alkaline rocks may be volumetrically more abundant most 1SO-rich. Because of the geographical dependence of
than was previously appreciated. 49 Rather than redefine the low 180 end of the individual fields, formation waters
'primary magmatic water' in such situations, it is more con- must contain a significant meteoric water component. 51
venient to define the isotopic composition of magmatic Simple mixing of original connatesea water with later
waters for a given geological situation. meteoric groundwaters cannot account for the shapes of the
Metamorphic waters There have been no direct analyses fields because many of the waters are more 1SO-rich than
of aqueous fluid inclusions from metamorphic rocks so, sea water. Other processes must also be active, such as iso-
similarly to magmatic waters, the isotopic compositions of topic exchange with the sedimentary host rocks and mem-
metamorphic waters are calculated from the isotopic com- brane filtration, s2-55 and also the possibility of mixing with
positions of the rocks at the relevant temperatures. Al- uprising metamorphic waters.
though the total amount of hydrogen isotope data is still Despite these complexities, the results obtained indicate
rather limited, metasediments typically have ~ D = - 4 0 to that the most commonly encountered waters in sedimentary
- 1 0 0 % o , ~ 180 = +8 to +26%o and meta-igneous rocks rocks are not of connate sea-water origin. Thus, when a
have oxygen isotope values extending down to +30/00. 6 pluton is intruded into unmetamorphosed sedimentary rocks,
Metabasalts from oceanic dredge samples have ~ D = - 3 5 to it is meteoric or formation waters that may most commonly
-600/00. 5~ The calculated hydrogen and oxygen isotope interact with it. Such saline formation waters, which are
field for metamorphic waters is shown in Fig. 4. In certain often associated with hydrocarbons, are probably intimately
situations (e.g. marble-rich terrains), metamorphic waters involved with the genesis of Mississippi Valley and Pennine
could have ~ 180 values up to 250/o0. The relatively large type Pb-Zn ore deposits. 56, 57, 58

I I I I ~J--= -- ~ I I I Origin o f hydrothermal fluids

Sedimentary basins /-- GO -- I
_ GC Gulf COaSt X// /.~. / The isotopic characteristics of waters of different genesis -
0 C California ~/l~ .-7,x" .~ I
t minois ./~"~'~.'~ - T'~ i
sea water, meteoric water, primary magmatic water and
M Michigan io ,~.;-" . . . . . ' J metamorphic water - are summarized in Fig. 5. The kao-
o Po,and /~:~::>.-d Metomorph,c,20 / /
+ Northeast X~".S"" '" I ,30^~;~;'--'I'---7 . . . . . . linite line defines the isotopic variations of kaolinites from
Engiond /..-~"r M ~.._I.-..;~ ,, ,. . . . , I I
-~o- + ~ " ......;t:-~-.
/.. i I: i // I I I /1 I I I I [
...... 7j ' " /i , Sediments 0 - /~,~ . . . . . Sea-water-hydrothermal waters -
_-" ,~/'~+ +,.+ ..- ! / Sea w o t e r s - - ~ J
- / i'"
o " P/ /"~lber to i" , /
- / F .............. "I -
#'I / i " B<,-~,7,-.../
...i ..... i - . . . . . . . . . . . . . I Metamorphic I /
-,~ + ~ i .l" i.. ''i _/~ - -4o- ~ ! ~.2o i / -
/ i" i'" /<-~
/"
t"
..--'+
..i-"
4-
t,,c"
+r .... . . . . . . . .
/
,- /oo
....... i,lt~ -80 ,.o/I -
-150 I I I I l i~' I I I l~ I Primary /k~
-20 -10 0 +10 +20 +30 ~"~/ Mete'oric- ma.gmotic /~,r
~)18 0 , *1.. c~" "r hydrothermal HzU ~r -
uo / w~oters /~_~
Fig. 4 Isotopic composition and fields for formation -,20 / /
waters, including oil-field brines, from various sedimentary le_____p_~ 1=
basins, s2-55 Fields representing isotopic composition of
common sedimentary rocks and metamorphic waters are -160
i1+~176
shown I I I I I / I I I I
-20 -10 0 6180,%, 10 20 30

180/160 variations reflect the wide range of compositions of Fig. 5 Summary diagram of isotopic composition of
sedimentary and igneous source rocks and the wide range of waters of different origin. '180 shift' trends due to w a t e r -
temperatures attained in this environment. rock interaction and exchange are shown for sea water and
Oceanic crust that has been through a sea-water-hydro- meteoric waters of composition A and B
thermal metamorphic process 45,50 will undergo dehydra-
tion during subduction. The isotopic compositions of such surface weathering environments. 34 Isotopic trends for
metamorphic waters are estimated to have 6 D "~ --20 to (1) sea-water-hydrothermal fluids and (2) meteoric-hydro-
--60%0 and 6180 ~ -I-4 to +12%o. These waters are more therma, fluids at two different geographic localities A and B
deuterium-depleted than sea water which was involved with are shown for geothermal systems where neither mixing of
the major hydration processes because dehydration reactions waters of different origin nor evaporation processes are im-
tend to take place at higher temperatures than the initial portant. The constant D/H ratio of the hydrothermal water
hydration reaction and hydrogen isotope fractionations in each example implies that the hydrogen isotope com-
decrease with increasing temperature. position of the water source overwhelmingly dominates the
Our recognition of metamorphic waters at the earth's D/H ratio in the water-rock systems as a whole. The vari-
surface is very limited. In the Coast Range, California, some able '18a-shifts' are a result of water-rock exchange pro-
geothermal springs have been interpreted to be of meta- cesses in which the amount of exchangeable oxygen in the
morphic origin because their D/H ratios are distinctly rock is a major part of the total oxygen content of the
different from local meteoric waters, sl water-rock system. These results are applied now to the
Formation waters The hydrogen and oxygen isotope com- analysis of the isotopic composition of directly and in-
position of formation waters from different sedimentary directly measured hydrothermal fluids and limits are set on
basins are summarized in Fig. 4. Waters from a given basin their origins.
plot within a fairly well defined field, the most 180-depleted Geothermal waters The isotopic results for a large number
samples lying by the meteoric water line. These waters are of active geothermal systems are summarized in Fig. 6.

30
 I I I ~ ~ i
Sea~arnnge
0 -
/ ?
. . . .
? Reykjanes
Saudi Arabian springs\ // ........ (225o)
Wairakei (268~ "~i/ JO Shimogamo (100~
Larderello (240~ ~'"~-/~=" _ _
Broadlands (2980)-~-e~'~";'O----El The . . . . . . . .
_ D^_~;. . . . _.C----~'f. . . . . . . . n _~eysers~z~u ; I-,,,,,u~)~
-5i icelandlR=y~j~, ,== _,L'_. . . . ---------u- -o " Imagmatir
"'Hekla.~- --
-10( t/ / ~ a m b o a t
~__/--~---~176 Springs
Park (187
......~ * SMOW , l
J 9 Meteoric water
J o Hot water
,4~__ _OV=il~t.=ton e [3 Steam
-15( I I I demonstrated the dominance of either a local meteoric water
-20 -15 -10 -5 0 +5 +10
~>180,"/.
Fig. 6 Isotopic composition of near-neutral chloride-rich
geothermal waters compared with composition of their
local meteoric waters. 10,59-63 Dashed lines illustrate trends
for geothermal waters; for clarity only most 18a-rich
water or steam is plotted with temperature for each system
These near-neutral waters are discussed in two groups. Ex-
cluding the Red Sea brines, Reykjanes, Iceland, and
Shimogamo, Japan, geothermal systems, the D/H ratios in a
given system are essentially identical to that in the local
meteoric waters - demonstrating the overwhelming domin-
ance (> 95%) of local meteoric groundwaters in these
thermal waters. 59 These waters have a wide variety of
chemical compositions and total salinities and occur at tem-
peratures UP to 320~ (Salton Sea). At a given locality the
constancy of the D/H ratio implies that mixing of waters of
different origins is unimportant in all of these examples. The
water/rock ratios are sufficiently large that the hydrogen
isotope composition of the system is controlled by the
meteoric water reservoir. The size of the oxygen isotope
shift is variable - negligible at Wairakei, but large at Salton
Sea - indicating that water/rock ratios are larger in the
former. The brines at Salton Sea are more saline (up to
307 000 ppm total dissolved solids) than those at Wairakei
(up to 4700 ppm). 64 Transportation and deposition of ore
metals are occurring in several of these geothermal systems.
At Salton Sea the >300~ brines have D/H and 180/160
ratios that are similar to primary magmatic water values.
Nevertheless, magmatic water is considered not to be pres-
ent because all of the waters in the system have D/H ratios
identical to those of the local meteoric waters. Waters at
intermediate temperatures plot about the dashed line of
Fig. 6 connecting the meteoric waters with the 320~ water.
This system clearly illustrates how meteoric-hydrothermal
waters can acquire isotopic characteristics similar to those of
primary magmatic waters when meteoric and primary mag-
matic waters have the same ~D values. Meteoric waters with
~D values in the range --40 to - 8 0 % 0 are very common.
Reykjanes is a narrow penensula surrounded by the sea.
Even with the absence of ~180 data, sea water must be the
dominant source of these thermal waters. 6~ Shimogamo,
on the Izu peninsula, Japan, is also close to the coastline.
Isotopically it is neither local meteoric nor sea water, nor
does it lie on a simple mixing line between these. Additional
processes, such as enrichment in 180 through water-rock
interaction, must also occur. 62 The hot brine pools in the
Red Sea have isotopic compositions like those of evaporated
sea waters that occur near the southern sill of the Red Sea. 1~
Because of the serious hydrological problems of Craig's
model, however, White 63 has proposed that mixing of local
highly evaporated sea water with coastal meteoric waters of
/ Saudi Arabia is more plausible and supported by the isotopic
data (Fig. 6).
Acid geothermal waters are also dominantly of meteoric
origin. They undergo hydrogen and oxygen isotope moeifi-
cation because of the evaporation processes. These waters
typically plot about a line of slope three (Fig. 3). Although
Iw~er - important metalliferous deposits are not often associated
directly with these acid waters, these processes are locally im-
portant in some ore deposits, such as in the El Salvador
porphyry copper deposit. 65'66 Some commercially import-
ant kaolinite-alunite deposits were formed in acid hot
spring environments. 48
Isotopic studies of active geothermal systems have
or sea-water source, o r their mixture, for the hydrothermal
fluids. In most cases it is the D/H ratio that more clearly
identifies the origin of the waters. Many of the thermal
waters have experienced major chemical and oxygen isotope
modifications relative to their source composition. No hot
spring is known where magmatic water is a recognizable
component of the hot water or steam.
Porphyry copper-molybdenum deposits
The porphyry copper-molybdenum deposits are a particu,
larly important class of ore deposit that have formed in a
sub-volcanic environment throughout at least the
Phanerozoic period. 67'68,69 They have been, and continue
to be, a major source for copper, molybdenum, silver and
gold. Although there is great variety in detail, most deposits
are centred about a 1- to 2-km diameter polyphase porphy-
ritic stock, which is intensely shattered and which displays
a systematic concentric pattern of alteration-mineralization
in the stock, in the adjacent country rocks, or in both.
Early hydrothermal alteration is characteristically associated
with K-silicate alteration assemblages (quartz-K-feldspar-
biotite-anhydrite plus low sulphur sulphide mineralization)
in the centre surrounded by a broad propylitic alteration
zone. K-feldspar and biotite destructive alteration (phyllic)
formed later and collapsed in on the K-silicate altered zone.
The major alteration-mineralization associations have
been isotopically studied in six porphyry deposits from
North and South America and Australasia. 4,5,66, 70, 71 Data
for Butte, Montana, are also considered here. 4, 72, 73 For
clarity, the results are summarized in Fig. 7 by showing the
calculated fields, rather than the actual data points, for
hydrothermal fluids associated with alteration biotite from
K-silicate assemblages and sericites from phyllic alteration.
The field for hydrothermal waters from Climax is based, in
part, on the direct determination of the D/H ratios of fluid
inclusions in quartz. 7~
Waters in equilibrium with K-silicate alteration assem-
blages at temperatures of 550-700~ are isotopically in-
distinguishable from primary magmatic waters. This applies
to all six porphyry copper-molybdenum deposits and re-
connaissance data from Sar Cheshmeh, Iran, TM but not for
Butte. Waters associated with pervasive sericitization and
sericites from well defined haloes adjacent to veins that
were formed at temperatures of about 300 + 100~ are
typically more 1SO-depleted than primary magmatic water
and have D/H ratios that vary from deposit to deposit. Re-
connaissance analyses of hypogene clays and sericites from
Safford, Mineral Park and Copper Creek, Arizona, s indicate
that Santa Rita is probably isotopically representative of the
large number of Laramide porphyry copper deposits in
Arizona, New Mexico and Sonora, Mexico.
Sheppard and co-workers 5 have shown that for the
Tertiary North American deposits the D/H ratios of argillic
31
 I 1 ~SMOW
Porphyry Cu-Mo deposits Porphyry copper deposits
Atteration biotite H20
Sericites (~ 300~ (,. 550-- 700~
/\.
El Salvador ,,~._.~..... /
Santa Rita--~~,V;""- . "",, '\.
i-'anguna~
A/
, ,,~
", rt',
,' 31 ~" /7-/r 4
,"1
Pa nguna
-5(
Bingham....~ ~ If///

04
Or /.-,-.~z 4, 4 ,
"1- Santa Rib
/i"L'=. ~
/'~-""
E3 / "j Primary
cO // ./ ' / . /
/..//I watermagmatic
i /" /.-
-100 / ./ /~.......~'" :" I
Bingham ./" . .But te. . .i/" I ! Butte
.~'~.--~Atteration
Ely /' Main Stage (-300~ ,'/f'~_/~,'" i biotite H20
/ i './//~/ ',i(-400~
[ ........../ ..~--
..
/ /~...."'~' .... ...I~
~'-'~ D~

~.. Climax (-600~

Climax . ~ ~ -- -- ----""
-15C I I I I I I
-20 -15 -10 -5 0 *5 o10
81eOH~o, %0, SMOW
Fig. 7 Isotopic composition of hydrothermal waters associated with alteration biotite
from K-silicate alteration assemblages, sericites from phyllic alteration from porphyry
copper deposits. 4,5,66, 71 Hydrothermal water fields for Butte and Climax include both
phyllic and hypogene argillic alteration. 4, 70, 73 Isotopic composition of present-day
meteoric waters is indicated
and phyllic alteration correlate systematically with inferred argillic alteration. This is shown schematically in Fig. 8.
Tertiary meteoric water values (function of latitude and Magmatic-hydrothermal solutions, initially under l itho-
elevation). The isotopic data for hydrothermal advanced static pressure, evolve from the crystallizing magma. These
and intermediate argillic alteration from these same deposits solutions react with the surrounding porphyry rocks and,
show an identical pattern to the phyllic alteration data. to varying degrees, with the surrounding country rocks
These patterns are even more clearly emphasized by the (Fig. 8(a)). K-silicate alteration and the introduction of
data for Butte and Climax. The only common process that much of the metals and sulphur accompany this stage.
we know of to generate such a pattern is the presence of a Further out into the country rocks convective circulation of
dominant meteoric water component in the hydrothermal the indigenous waters, meteoric or connate, is established.
fluids responsible for phyllic, intermediate and advanced As the complex cools and evolves, the meteoric-hydro-

K-silicate ~X,,~ Phyltic ~Ore-shetl Pr= P r o p y t i t i c

r ~ (~,~,.~.. /, Hot ~t^r'~"('~'S
-~rr .... " " -;nTI/t

/
1 / ~ 1 ~ ; Pr t Pr Pr /
Pr Pr PI

p + Pr Pr r

Fig. 8 Schematic cross-section through porphyry copper deposit showing two types of
hydrothermal fluids and their postulated interaction at two different stages during their
evolution: a, early stage; b, late stage (see text)

32
thermal system, under approximately hydrostatic conditions,
encroaches in on the waning magmatic-hydrothermal part of
the system. Reactions occur to give pyritic mineralization
and feldspar-destructive alteration, phyllic with usually
lesser argillic alteration (Fig. 8(b)). In many porphyry
copper deposits much of the chalcopyrite mineralization is
centred about the boundary zone between the meteoric and
magmatic-hydrothermal systems. This may be controlled
by the gradients in pH, pressure, temperature, salinity, etc.,
which undoubtedly occur across this evolving interface.
In Fig. 7 the isotopic compositions of present-day
meteoric waters are also shown. The D/H ratios of the
dominantly meteoric-hydrothermal fluids associated with
sericitization are generally more D-rich than present-day
local meteoric waters. For the North American deposits
this demonstrates that hydrothermal sericites (and clays)
have not re-equilibrated with present-day waters.
Excluding Butte and Climax, sericites or their calculated
waters (Fig. 7) from more typical porphyry deposits have a
narrow range of 1 8 0 / 1 6 0 ratios, and these are only a few
per mil depleted relative to primary magmatic water values.
This lack of a dramatic 180 depletion implies that meteoric
water/rock ratios during the development of phyllic and
argillic alteration were not particularly large. The variations
of the hydrogen isotope compositions of the fluids during
sericitization in any given deposit may be reflecting the mix-
ing processes between the magmatic and meteoric-hydro-
thermal solutions as the meteoric system collapsed in on
the waning magmatic stage. In contrast, the dramatic 180
depletions of many of the hydrothermal fluids at Butte and
Climax dictate that enormous quantities of heated meteoric
waters circulated through parts of these systems.
The hydrogen isotope composition of alteration biotites
from Butte contrast with all other data on biotites from
K-silicate alteration assemblages. From the present data
there are three possible interpretations of their low D/H
ratios: (1) retrograde hydrogen isotope exchange occurred
with later meteoric-hydrothermal solutions; (2) meteoric-
hydrothermal waters had already collapsed in on and over-
whelmed the magmatic-hydrothermal system; or
O--
t C Casapalca,
I Peru I I
P Providencia
x Nevada, epithermal Au-Ag
9 Thermal waters 4,-
~-~-~ /
Tui, N.Z.
Broodlands, N.~ " z ~ -~
.\
-50 / 0- . . . . . . . . . . . . .
Salton Sea V i
Pachuca
X
-100
Ou rayz~ X
-Steamboat
\
x
x x
Comstack
-1501d" I I
-20 -15 -I0 -5
818OH20, %o
Fig. 9 Isotopic composition of hydrothermal water associated with various epithermal
A u - A g deposits, base-metal, tin and tungsten deposits. 2, 4, 5, 49, 63, 75-82, 84 Meteoric
water line and primary magmatic water box given for reference
(3) meteoric waters had interacted with magma and modified
its hydrogen isotope composition so that subsequently re-
leased magmatic waters were more D-depleted than primary
magmatic water.
Sheppard and Taylor 73 have shown that although the
major part of the Boulder batholith as exposed had 'normal'
D/H ratios, locally meteoric water had interacted with the
batholith. Much of this interaction was probably subsolidus.
Nevertheless, part could be at higher temperatures because
D-depleted rocks were observed that had no textural evi-
dence for post-magmatic alteration. If pre-Main Stage
alteration biotite formed at 550--600~ (see Sheppard and
Taylor, 73 p. 938) and retrograde exchange is unimportant,
then (3) above is most probable.
Other epizonal ore deposits
In addition to the porphyry copper-molybdenum deposits,
a number of other hydrothermal ore deposits formed in a
continental epizonal igneous environment have been studied
isotopically.2, 4, 5, 49, 75-82 These deposits are typically
associated with intermediate to acid plutonic and volcanic
rocks. They include the 'epithermal' gold-silver deposits,
base-metal, tin and tungsten deposits. Also, several active
hot spring systems in volcanic rocks are associated with t h e
transportation and deposition of metal such as those in New
Zealand and at Steamboat Springs. 9,63 only the major
conclusions of the isotopic results are discussed here more -
detailed accounts of the individual deposits can be obtained
from the original papers.
The isotopic composition of hydrothermal fluids, determ-
ined by direct and indirect methods, for a large number of
these deposits and hot spring systems are summarized in
Fig. 9. The water compositions typically lie between the
primary magmatic water field and the meteoric water line.
The D/H ratios vary from one deposit to another (e.g.
Cornwall versus Tonopah) and also within a single deposit
complex (e.g. Comstock Lode). Remarkably, some of the
hydrothermal waters lie on or extremely close to the
meteoric water line (e.g. Bodie, Tonopah). Such low 180
, S MOWF . . . . L i _ _ _
/
I Cornwall
~'~'~"~-~ .~.
,jm
," Carrock Fell Primary magmatic'
, '1
Creede !
i
Comstock
0
. . . . . . . . . . . . . . V---- _.all
X
I I
0 +5 +10
33
 waters imply that at least during certain stages in the evolu-
tion of these deposits meteoric waters were overwhelmingly
the dominant source for the hydrothermal fluids. Fluid
inclusion studies on minerals from several of these "low 1SO'
hydrothermal systems, which typically characterize the
epithermal A u - A g deposits, indicate that salinities are low
(0.3-4 wt % equivalent NaCI) and deposition temperatures
were in the range 150-300~ 83 These isotopic character-
istics, salinity and temperatures values are directly compar-
able with those observed in many hot spring systems.63
Studies on the Comstock Lode district are particularly
illuminating. 7s,77 As is shown in Fig. 9, one sample is en-
riched in D and leO and lies close to the primary magmatic
water field; other waters are depleted in D and 180 and lie
towards the meteoric water line. The D-lSO-rich sample
comes from the 1200-ft level of the Con Virginia mine in the
high-grade 'bonanza-type' silver ore. Salinities are about
3.5 wt % and homogenization temperature is 300~ A usually dominantly submarine lavas and pyroclastics of
major magmatic water contribution is strongly indicated, 77
Thus, although low-salinity meteoric-hydrothermal solutions
must play a dominating role in transportation and deposition
processes in the epithermal Au--Ag deposits, the develop-
ment of an early (?) magmatic-hydrothermal stage may be
critical for the initiation of this type of ore deposit. In the
majority of these deposits it appears that the large scale of
the meteoric system and its associated voluminous zone of
propylitic alteration have overwhelmed the earlier magmatic-
hydrothermal stages.
Meteoric-hydrothermal waters also appear to be a domin-
ant or a major component of the fluids in many base-metal
vein and pipe deposits, e.g. Creede 79 and Ouray-
Silverton, 4,s4 Colorado, and during greisenization associ-
ated with tin or tungsten deposits, e.g. Cornwall, s,49 Pasta
Buena, Peru,81 and probably Carrock Fell, England. The
D/H ratios of the fluids at Creede and Ouray, which are both
in the San Juan mountains, are very different. Like the
present, Tertiary meteoric waters in the eastern and western
San Juans were different from each other, reflecting differ-
ences in their weather patterns. 84 Fluid inclusion studies
on the base-metal deposits indicate that salinities are typic-
ally higher (2 to " 4 0 wt % NaCI) than those in the epi-
thermal A u - A g deposits.
-20
/o,,
-4o- i \ "-..~ ,
Hydrothermal H20 / '&"~"
I ~
- . Greisen (500--400 o) , C~
9 Early vein (350--200 ~ :~ ~".:--2.
~ . . . . t ,(
"~=-60- o Late vein (260--200 o)
9 Vugh (240-150 o)
-- I
?. . . . k
-80--
/
I ,
Modern meteoric water
-100 I , , I i I I I i =
-5 0
6180,'1..
Fig. 10 Plot of ~D versus ~ 180 for hydrothermal waters
during evolution from greisen to vein and to vugh stages
(arrow indicates path). Data taken from Landis and Rye 81
In the Pasta Buena W - C u - P b - A g deposit 81 (Fig. 10)
and the Casapalca A g - P b - Z n deposit 8o (Fig. 9) the isotopic
evidence indicates that magmatic-hydrothermal waters were
dominant at certain stages in their development. At Pasta
34
Buena there were major fluctuations in the isotopic com-
position of the fluids as they evolved from greisen through
vein to vugh stage (Fig. 10), reflecting the mixing processes
between magmatic and meteoric waters. Important deposi-
tion of wolframite occurred with episodes of influx of
meteoric waters. In contrast, at Casapalca the influx of
meteoric water appears only to be important during late
post-sulphide stages.
In many of the base-metal, tin and tungsten deposits the
salinity of the fluids fluctuated, typically in the range 4 to
" 4 0 wt %, with time. No simple relationship between
salinity and isotopic composition is, however, usually
observed.
Submarine volcanogenic deposits
A variety of stratiform massive sulphide deposits are inti-
mately associated with volcanic rocks. The volcanics are
basaltic, andesitic or rhyolitic composition. The massive
sulphide ore usually grades down into a stockwork zone
consisting of disseminated, vein and veinlet mineralization
in the altered igneous rocks. This class of deposit can be
subdivided into different types, depending on the composi-
tion of the affiliated igneous rocks and dominance or
absence of one of the metals copper, lead and zinc. s5,86
Silver and gold may also be important. Two different types
of volcanogenic deposits have been studied isotopically:
(1) the cupriferous massive sulphide deposits related to
ophiolite igneous complexes s7 and (2) the F e - C u - P b -
Z n - A g - A u Kuroko deposits, which are closely associated
with submarine felsic volcanism. 8s
Cyprus sulphide deposits The cupriferous pyrites deposits
in the Troodos complex, Cyprus, are an excellent example
of the massive sulphide deposits associated with an ophio-
lite sequence.89,9~ The deposits are relatively underformed
and have not been metamorphosed. The isotopic character-
istics of the fluids involved with the metamorphism of the
oceanic basaltic host rocks and the hydrothermal stockwock
mineralization are discussed in detail elsewhere in this
volume. 45 The results are summarized in Fig. 11, and in
Fig. 9 of Heaton and Sheppard. 4s
The hydrothermal system was dominated by waters of
sea-water origin. The brines penetrated down through the
upper 3 - 5 km of the oceanic crust. They were involved
with the metamorphic reactions and pervasive alteration
before rising through the stockwork zone at temperatures
_ up to about 300~ and issuing on the sea-floor as thermal
~"'~,,
a
~~,
ii t I t
' springs. The massive sulphide deposits were formed from
,~' ~l ~1 I Ii chemical precipitates generated during the interaction of the
) + ;~ M
P
sea-water-hydrothermal fluids with the 'cold' sea water.
v i w They accumulated above the thermal springs within the
9 /
I caldera-like depressions that developed during an inter-
volcanic period - after the production of the spreading-axis
// m',-
/ volcanic episode and before the commencement of the off-
axis volcanism. No evidence has been found for either a
magmatic-H20 or metamorphic-H20 (from the dehydra-
I 1 , A i I tion of subducted crust) contribution to the ore-depositing
+5 +1(
fluids.
These results raise the question of whether there are
massive sulphide deposits in the present oceanic crust. Mass-
ive sulphide deposits comparable with the Troodos deposits
have not yet, to the writer's knowledge, been recorded from
the oceans. In Fig. 11 isotopic results are included for
dredge samples of basalt and dolerite from the North
Atlantic which had been metamorphosed to zeolite and
greenschist facies, and for comparable rocks from
Troodos. 5~ They are isotopically indistinguishable. The
calculated D/H ratios of waters in equilibrium with the
oceanic dredge samples, by use of the fluid inclusion homo-
genization temperatures of Jehl, 91 are identical to sea-water
values and demonstrate that sea water was involved during
the metamorphism of the oceanic crust, s~
Seo woters
Meteoric woters {Most-
9 Japan
-2,
/
o~ Green Tuff type //
thermG[ waters /
1:[~"- 41
/
/
bO /
//
-6~
-8~
~ i
I I
-10
Fig.11 Plot of ~ D versus ~ 180 for metabasalts and metadolerites from Troodos, Cyprus,
and North Atlantic dredge samples, and altered stockwork basalts from Troodos, and
fluid inclusion waters from two Kuroko deposits, Japan. 4s, 50,98 Also shown are com-
positions of gypsums (water of crystallization) from same Kuroko deposits and isotopic
composition of Green Tuff type thermal waters of Japan 62
There is additional evidence for sea-water-hydrothermal
activity. The formation of F e - M n hydroxyoxide rich
oceanic sediments is considered to be related to discharging
thermal sea-water springs. 92,93,94 Sea-water-hydrothermal
metasomatism of carbonate ooze, which overlies basalts,, has
also been recognized in DSDP samples from the Indian
Ocean. 95 These data on oceanic basalts, Troodos and the
sediments emphasize the widespread nature, importance
and occurrence of sub-sea-floor metasomatism resulting
from the interaction of convective circulating sea-water-
hydrothermal solutions with oceanic crust.
Does the apparent absence of massive sulphide deposits
imply that the coincidence of the necessary factors required
to produce a deposit has not taken place during the recent
generation of oceanic crust, or is our sampling of the ocean
floor still too incomplete to locate such a small target as a
massive sulphide deposit?
A number of factors appear to be critical for the genera-
tion of a volcanogenic massive sulphide deposit in basaltic
rocks in addition to establishing a sea-water-hydrothermal
convection system: (1) chemical modification of the hydro-
thermal brine at temperatures greater than 300~ by deep
penetration through the oceanic crust; (2) development of
a volcanic quiet period immediately following the growth
of layer 2 of the oceanic crust; (3) the issuing of hot springs
on the sea-floor in depressions; (4) the preservation of the
depression at the site of the thermal spring for a time
period that is at least comparable with the life of the con-
vective system so that the sulphide precipitates formed from
the interaction of the thermal waters with sea water are
localized and not dispersed widely; and (5) the depressions
may enable gravitationally stable thermal brine pools to
develop so that chemical precipitation takes place under Eh
and pH conditions different from those in more open ocean
water.
1 I
Cyprus
Stockwork H20 (200--350~
/
/
"
_ . \
" \
T'-l---
|
~-J
Submarine
W I bosolts and
I
greenstones,
-- North Attantic
Kuroko~ gypsum
I I
-5 0 *5 .10 .15
8180, %0
In addition, it is possible that a two-stage process is
necessary in which subduction takes place underneath a
submarine oceanic spreading centre. 87
Kuroko deposits The Kuroko deposits of Japan occur in
the Miocene Green Tuff sequences composed of submarine
sediments and pyroclastics that overlie basement. 88 Hydro-
gen and oxygen isotope data for two Kuroko deposits from
northeast Honsh-G, Matsumine and Shakanai mines, are
summarized in Fig. 11.62,96 The data are from the strati-
form part of the deposits: water in fluid inclusions from
the lower part of the stratiform pyrite ores, the p y r i t e -
chalcopyrite yellow ore, and water of crystallization of
gypsum, the gypsum occurring between and in the under-
lying rhyolitic tufts and breccias and the overlying pyrite
ore.
The D/H ratios of the fluid inclusion waters, with salini-
ties in the range 0.5-2.5 molal NaCI equivalent, are
depleted by 1 1 - 2 6 % o relative to Cenozoic ocean waters
(the isotopic composition of the ocean changed from about
- 7 % 0 to 0%0 during the Miocene (see above)). Ohmoto
and Rye 98 concluded that sea water was the dominant
source of the hydrothermal fluid, but probably not to the
exclusion of a minor (< 25%) meteoric and/or magmatic
H20 contribution to help account for the small D-shift
away from the ocean water value. A meteoric water com-
ponent is not out of the question in this submarine environ-
ment because ore deposition probably occurred within a few
kilometres of the coastline. This model is schematically
illustrated in Fig. 12. We note that a mixed sea w a t e r -
35
 /
Sea W a t - l

k rocks ~lli.~v/~,/ "v v "v~v/~ ~' Massivespudhlei /

~vll f ;:~tlte MeteoricH20?

Fig.12 Schematic cross-section of Kuroko deposit. Model shows convective circulation
of sea water and mixing with a possible meteoric and/or magmatic water contribution.
Modified after Ohmoto and Rye 98 and Lambert and Sato 88
meteoric water origin is favoured as the source of the
metalliferous brines in the Red Sea. Reconnaissance isotopic
results on the Kuroko-type deposit on Vanua Levu, Fiji, 97
indicate that the hydrothermal fluids were of sea-water
origin. 9s Mixing of water of different origin is not required
by the data. Thus, mixing of meteoric waters with sea
water does not appear to be an essential feature of Kuroko
deposits.
Much of the gypsum in Kuroko deposits was deposited
originally as anhydrite. Water in equilibrium with gypsum
from Matsumine and Shakanai mines plot by the meteoric
water line and within the field for present-day thermal
waters ( T " - 50-60~ that occur in the Green Tuff
sequence. 62 The age of the meteoric hydration process is
not known, but the results indicate that secondary processes
take place after uplift of the deposits.
Supergene versus hypogene clays
The kaolinite-water system as discussed above (Fig. 2) can
be directly applied to distinguish, between kaolinite, mont-
morillonite or other clay minerals that were formed at earth
surface type temperatures during weathering or low-
temperature supergene processes from those which formed
0 i i i
-20 ,,...Q.~
~.~0
-40
'Magmatic H20?"
hydrothermally. In certain environments it is not meaning-
ful to distinguish between supergene and hypogene minerals
and solutions. For example, near the surface of an acid hot
spring system where convective overturning is occurring the
descending solutions (supergene sensu stricto) may be hot re-
cycled fluids, which, immediately before, were rising hypo-
gene solutions. Hence, following Sheppard and co-workers, 4
supergene is used here to refer to the weathering and deep
alteration processes that~ake place at temperatures up to
about 40~ or so (i.e. non-hydrothermal conditions). Such
temperatures are readily attained in the supergene altera-
tion environment because a large amount of energy is re-
leased during the generation of supergene acids by the oxida-
tion of such minerals as pyrite. Similar temperatures are
observed in leaching dumps at porphyry copper mines.
Data for dickite, kaolinite-rich and montmorillonite-rich
samples are given in Fig. 13 for some selected major hydro-
thermal ore deposits where isotopic criteria have been used
to distinguish between hydrothermal and supergene clay
deposition.4, 34, 48, 49, 66, 70 Montmorillonites formed in a
continental weathering environment also plot in a band
about the kaolinite line because their fractionation factors
with water are similar to those for kaolinite. 99
For porphyry copper deposits such as Santa R ita and El
Salvador there is a clear isotopic distinction between hydro-
, i , I !
thermal and supergene kaolinite and montmorillonite. At
Butte, although searched for, kaolinite of supergene origin
has not been detected. 73 In Cornwall, the major high-
Bohemia
,~6", C o r n w a l l .'
quality china clay deposits are isotopically supergene clays. 49
',',,
\/, The hydrothermal kaolinite comes from an altered granite

-6o ~.~
SantaRita
'" . . . . g f~
',~'"3r
~,._~. / sheet adjacent to a major vein in the metasediments. It is
broadly consistent with the data on sericites from Cornwall
(Fig. 9) and the Bohemian hydrothermal dickite from a
o~ -80 El ~al~a.or (o-. ',.,.~;v
/o ..... , 9 Sn-W greisen that is geologically comparable with
mineralization in Cornwall. 49
uo ~ O~
Bingham I,O'L " " - - , f%_ ~
Where more detailed studies have been carried out, e.g.
-120 / , j
Santa Rita, supergene clays have been detected to depths of
500 m below the present surface (see Fig. 2). 4 For the
~ o~
Ouray -~" ", Clays deposits in the western United States information is avail-
-140 0 ~'" A ,~ , /.~ .
,,~ Butte o~ K,D M
- ----'" /%..,C" Hypogene 0 ~|
4 able on the isotopic composition of present-day meteoric
-160 C[imax ( o o:' / CO Supergene 45 waters. Also, there are some independent data of Tertiary
I i . . . . ,l L /~'~ I i I I meteoric waters. 6 Thus, it can be shown that most of the
-10 0 10 20
8180,%o supergene (and hydrothermal) clays are not in equilibri'um
Fig. 13 Isotopic composition of hypogene and supergene with present-day meteoric waters (see Fig. 13 for Santa
clay minerals from some selected ore deposits. 4,34,49,66, 70 Rita, and recall that kaolinite is about 30%o depleted in D
Present-day meteoric water at Santa Rita is shown. K, relative to co-genetic water). Clay minerals do not appear
kaolinite-rich; D, dickite; M, montmorillonite-rich to undergo major retrograde isotopic exchange without re-

36
 crystallizing. The data for the porphyry copper deposits are
consistent with major supergene processes following on
shortly after hydrothermal activity has essentially ceased. At
El Salvador, uplift had probably commenced before the
main development of supergene alteration, but alteration
had ceased prior to the major uplift of the Andes in this
region. 66
Water/rock ratios in hydrothermal systems
In the discussion in Fig. 2 the concept of water/rock ratio
was introduced. It was shown how the hydrogen and/or
oxygen isotope data can be used to set a limit on the
minimum amount of water of external origin - meteoric,
formation or sea water - that was involved with the hydro-
thermal reactions. This requires knowledge of the initial
isotopic composition of the source water and its final com-
position after exchange and equilibration at a specific tem-
perature, and the initial and final isotopic composition of
the rock that underwent exchange. In practice, some of
these data usually have to be estimated - for example, the
initial isotopic composition of the meteoric water or rock
before it undergoes exchange.
I I ' I ' I ' I'1
Sonto Rito, phyllic
Butte {phyllicK-Silicate?
-o--
Climax, phytlic
Comstock, Pochuco
Goldfield, Tonopah
O u r a y - - Silver t o n
Cyprus
Kuroko
I I , I t I,IJl
0'1 0"2 0 " 4 0 ~ (~8 1
Meteoric or sea w a t e r / r o c k r a t i o ( a t o m % o x y g e n )
Fig. 14 Minimum i n t e g r a t e d m e t e o r i c or sea water/
rock ratios for some selected hydrothermal ore
deposits 4,5,6,45,70,73,77~4 ~98
To date, most applications have been concerned with
setting an order of magnitude limit on the total, bulk or
integrated water/rock ratio for the life of the system as a
whole - an important first step. Some results for a variety
of meteoric-hydrothermal and sea-water-hydrothermal stages
of ore deposition are summarized in Fig. 14.
A wide variety of integrated water/rock ratios are ob-
served (Fig. 14), ranging from near zero (<<0.1) to greater
than one. For systems with values greater than 1 this implies
that large to extremely large volumes of water are involved,
recalling that the ratios are minimum values only. These
amounts are geologically quite reasonable when they are
compared with some active geothermal convection systems
(e.g. Wairakei). Also, as Taylor and Forester lOO have
pointed out, only a few per cent of normal amounts of
precipitation needs to be added to the deep convective circu-
lation system during its active life. If formation water
rather than meteoric water is the dominant source of the
water, the values given in Fig. 14 are too small because
formation waters are characteristically more 180-rich than
meteoric waters (Fig. 4).
For the dominantly sea-water-hydrothermal Cyprus and
Kuroko deposits, where the initial isotopic compositions of
the sea water and rocks are relatively well known, extremely
large quantities of brine have passed through and interacted
with these rocks. In some of the A u - A g epithermal deposits,
e.g. Bodie, Goldfield, Tonopah, the ~ 1sO values of the
hydrothermal fluids are so similar to inferred local meteoric
water values that again extremely large quantities of water
flushed through these deposits. Salinities of these meteoric-
hydrothermal fluids are generally quite low (< 4 wt %
equivalent NaCI). At some other A u - A g and base-metal
deposits water/rock ratios were not so uniformly high.
At Butte and Climax during sericitization water/rock
ratios were quite variable from one place to another. The
interpretation of the data at Climax and some other deposits
is more complex because there is isotopic evidence for
mixing of magmatic waters with the meteoric waters. Dur-
ing the early pre-Main Stage mineralization at Butte
(K-silicate alteration), if modified magmatic waters were not
dominant, meteoric water/rock ratios were very small. In-
ferred meteoric water/rock ratios for porphyry copper d e -
posits are typically quite low. As some of these deposits
were emplaced into relatively undeformed Palaeozoic and
Mesozoic sediments, however, formation waters might be
an important source for the fluids. Formation water/rock
ratios would be considerably larger than the values shown
for Santa Rita in Fig. 14.
In most hydrothermal systems with through-going veins
water/rock ratios will vary on a local scale. The rocks form-
ing the walls immediately adjacent to a vein will 'see' much
more of the fluid relative to a unit of rock that is some dis-
tance from the nearest open fracture. Hence, more detailed
interpretation of integrated water/rock ratios in terms of
mass transfer must await more detailed studies where there
is good control of sampling. This does not, however, detract
from the importance of setting qualitative limits on the
minimum amount of water that has reacted with the rocks
in the system as a whole.
Diverse origins of hydrothermal fluids - implications
I , I , I,I,J
2 4 6 8 10 Combined hydrogen and oxygen isotope studies of post-
magmatic hydrothermal ore deposits have provided com-
pelling evidence for the dominance of waters of non-
magmatic origin in some deposits and the importance of
mixing of waters of different origin in others. In the epi-
zonal continental environment, the upper 5-.10 km of the
crust, meteoric and/or formation waters are very common
sources for the hydrothermal fluids. In the submarine en-
vironment the large-scale interaction of sea water with the
upper layers of the oceanic crust is a general phenomenon
in the region of the spreading-a~(is. Similar sea water-rock
interaction occurs during submarine volcanism (basaltic to
felsic) in island arcs such as Japan and Fiji. Hence, it is now
well established that the interaction of waters of external
origin and history with volcanic and sub-volcanic rocks is a
common process. The influx of waters of external origin
into the igneous system is not, however, always prevalent
and its importance also decreases with increase in depth of
emplacement as the roles of magmatic and metamorphic
waters increase in dominance.
Magmatic fluids, derived from the crystallizing magma,
are undoubtedly major ore-depositing fluids in many sub-
volcanic hydrothermal ore deposits, as, for example, in the
porphyry copper deposits. Nevertheless, in most deposits
that initially were dominated by magmatic-hydrothermal
solutions, fluids of external origin - meteoric, formation,
connate or sea water - became progressively more domin-
ating as the system evolved. Some systems were completely
overwhelmed by the meteoric-hydrothermal stage (e.g. epi-
thermal A u - A g deposits) or sea-water-hydrothermal stage
(e.g. Cyprus), leaving little or no record of any magmatic-
hydrothermal event.
37
 Excluding the submarine volcanic processes for the
moment, the importance of an initial magmatic-
hydrothermal stage may be completely out of proportion to
its possible small volumetric contribution to the hydro-
thermal fluids relative to the meteoric or formation water
contribution. There are a number of consequences of a
magma becoming saturated with water and other volatiles
that may be fundamental to the nature of later develop-
ments. (1) Chlorine has a strong preference for the aqueous
phase relative to the silicate melt 11,21,101 (experimental
data are for granitic and granodioritic compositions).
Earlier-separating aqueous fluids will probably be enriched
in chlorine relative to later-separating fluids. In the sub-
volcanic environment at P < 1240 bar (Fig. 1 ) two water-
bearing fluids, one with low salinity and low density and
the other highly saline and dense, may coexist with the
silicate phase. Because a chloride solution is a potent solvent
for base metals, the initial water and chloride content of
the magma, the nature and occurrence of the boiling pro-
cesses and how the elements are partitioned between the
two possible aqueous fluids may play a critical role in the
base-metal concentration and redistribution processes.
(2) The nature of the processes that take place when the
magma becomes saturated in H20 and other volatiles, re-
lated to the geological setting of the magma body and its
crystallization history, may play a key role in determining
whether the aqueous fluid is released passively or explo-
sively. Explosive volcanism is an effective mechanism for
the generation of fracture permeability and, hence, prepar-
ing the ground, igneous and/or country rocks, for
subsequent meteoric-hydrothermal processes.
This emphasis on the magmatic-hydrothermal stage is
not to deny the important role of later meteoric-
hydrothermal processes in transporting, reworking and/or
depositing ore metals. Mixing of waters of different origin
and chemistry, and the removal of thermal energy from the
magmatic stage by the meteoric convection system, either
singly or in combination, will tend to promote ore deposi-
tion. These processes may, however, be ineffective in
generating an ore deposit if the earlier magmatic-
hydrothermal solutions, perhaps independent of their volu-
metric contribution to the fluids as a whole, are not
generated in a fairly specific way.
Continental volcanic-plutonic basaltic complexes typi-
cally are not accompanied by well developed hydrothermal
ore deposits, although many of these complexes were
associated with large-scale meteoric-hydrothermal convec-
tion systems. For example, in the Tertiary Brito-Arctic
igneous province very many of the epizonal igneous intru-
sions have interacted with meteoric groundwaters on a
grand scale.6, 46,100,102 In addition, at the Lilloise intru-
sion, east Greenland, there was an efflux of magmatic water
out into the contact metamorphosed basalts, which pre-
viously had been involved with meteoric waters. 1~ Hydro-
thermal o r e deposits of any consequence are, however,
essentially unknown in the Brito-Arctic province.
Most hydrothermal ore deposits related to basaltic vol-
canic and plutonic rocks are associated with submarine
volcanism, like the Cyprus deposits. For basic systems this
prevalence of hydrothermal ore deposits with submarine
processes suggests that the chemistry of the sea-water
source of the fluids plays a critical role in ore genesis. For
Cyprus, the sea water is a major source for sulphur, 45 and
perhaps also provides the necessary chlorine for base-metal
leaching.
In both the continental (subaerial) and oceanic epizonal
environment basic intrusions were typically emplaced into
volcanic rocks, which, because of their well jointed or frac-
38
tured nature, supported meteoric or sea-water convective
circulation. These hydrothermally altered and hydrated
rocks may become involved in melting, assimilation or ex-
change processes in the roof zone to the magma chamber. 6
If the scale of the processes is large enough, the upper parts
of the magma chambers could be both chemically and iso-
topically modified. The low 180 magmas of Iceland could
be generated in this way. 46 The end results of these pro-
cesses will be different in the oceanic environment from the
continental, because the chemistries of the waters involved
are so profoundly different - sea water versus meteoric
water. Not only may the genesis of the soda-rich trondh-
jemites in the oceanic rocks relative to the potassic granites
and rhyolites in the continental basic complexes be related
to these processes 45 but they may also affect the nature of
any subsequent post-magmatic mineralization. The nature
and consequences of the proposed magma-altered rock in-
teraction requires further study.
The major sources for sulphur and the metals in ore de-
posits are not considered in detail here. The present status
of our knowledge of lead, the only ore metal to have iso-
topic variations that can be used to trace its origin, and sul-
phur have been reviewed recently. 7,8 In the dominantly
magmatic-hydrothermal systems probably most of the sul-
phur and metals were derived, along with the water, from
the crystallizing magma. The magmatic processes that are
necessary to generate the sulphur and metal anomaly (a
magmatic ore deposit) are, however, not well understood.
When water whose origin is external to the igneous system
is dominant either the source of the fluid provides much of
the sulphur (e.g. Cyprus), or sulphur and the metals are
picked up by the fluid as it undergoes either mixing with
magmatic waters or during water-rock exchange reactions.
Meteoric-hydrothermal fluids, particularly if they are saline,
may just as readily leach out the metals from the country
rocks as from the igneous system that provides the energy
to drive the convecti'on cells.
Conclusions
The D/H and 180/160 isotope ratios of sea water, meteoric,
primary magmatic and metamorphic waters have relatively
well defined ranges of values with few exceptions since at
least the late Precambrian (Fig. 5). When a water, of what-
ever origin, passes through a rock that is out of equilibrium
with it, chemical and isotopic exchange reactions occur.
Unless extremely large quantities of water exchange with
the rocks, the 180/160 ratio of the water and its chemistry
will undergo modification. Because the hydrogen content
of rocks is usually so small (1 wt % H20 + = ~ 0.11 wt %
hydrogen) relative to the amount in the water, however,
the D/H ratio of the water is usually not significantly modi-
fied by the exchange reactions. Although it is impracticable
in many situations to measure the isotopic composition of
the fluid directly, it can generally be calculated satisfactor-
ily. The isotopic composition of the fluid is related to the
temperature of equilibration and the isotopic composition
of the exchanging minerals (Fig. 2).
These observed systematics form the foundation of the
isotopic tracer approach to the determination of the origin
of waters in hydrothermal ore deposits. Because there are
fewer detailed studies of Precambrian systems at present
and our knowledge of, in particular, the isotopic composi-
tions of these early oceans is less specific, there are greater
uncertainties in the application of this 'fingerprinting'
method to Precambrian ore deposits. This situation is
likely to be significantly improved in the next few years.
 In addition, to make isotopic studies more rewarding,
ore deposits should be sought in localities where there is
evidence that the hydrogen isotope composition of the
contemporaneous meteoric waters did not overlap the pri-
mary magmatic water range of values. We must not assume,
just because an ore deposit is up at high elevation today
where meteoric waters may be more D-depleted than pri-
mary magmatic water values, that at the time of
mineralization it was necessarily in a favourable locality
isotopically for distinguishing between meteoric-
hydrothermal and magmatic-hydrothermal solutions.
The formation of an ore deposit requires the coincidence
of a number of favourable factors. Our understanding of
present ore deposits and our capacity to generate more
powerful models to aid prospecting for new deposits are
critically dependent on our being able to enumerate the
important factors and to assesstheir role in producing an
ore deposit. To improve our understanding of volcanic
processes and ore genesis further studies of essentially un-
mineralized or poorly mineralized hydrothermal systems
such as those in the Brito-Arctic province are required. In
this way we may be able to evaluate the role of the different
factors that need to coincide to generate an economic ore
deposit.
Acknowledgment
I am grateful to H. P. Taylor, Jr., in particular, and many
colleagues and students for beneficial discussions and for
introducing me to their problems. The Isotope Geology
Unit of the Universities of Scotland is supported by a
N.E.R.C. research grant.
References
1. Craig H. Boato G. and White D. E. Isotopic geo-
chemistry of thermal waters. In Proc. 2nd Conf. Nuclear
Processes in Geologic Settings. Rep. natn. Res. Council,
nucl. Sci. Ser. 19, 1956, 29-38.
2. Rye R. O. The carbon, hydrogen, and oxygen isotopic
composition of the hydrothermal fluids responsible for the
lead--zinc deposits at Providencia, Zacatecas, Mexico. Econ.
Geol., 61, 1966, 1399-427.
3. Sheppard S. M. F. Nielsen R. L. and Taylor H. P. Jr.
Oxygen and hydrogen isotope variations in minerals from
porphyry copper deposits. Abstr. geol. Soc. Am. 1967, 203.
4. Sheppard S. M. F. Nielsen R. L. and Taylor H. P. Jr.
Oxygen and hydrogen isotope ratios of clay minerals from
porphyry copper deposits. Econ. Geol., 64, 1969, 755-77.
5. Sheppard S. M. F. Nielsen R. L. and Taylor H. P. Jr.
Oxygen and hydrogen isotope ratios in minerals from por-
phyry copper deposits. Econ. Geol., 66, 1971, 515-42.
6. Taylor H. P. Jr. The application of oxygen and hydro-
gen isotope studies to problems of hydrothermal alteration
and ore deposition. Econ. Geol., 69, 1974, 843-83.
7. Doe B. R. and Stacey J. S. The application of lead iso-
topes to the problems of ore genesis and ore prospect
evaluation: a review. Econ. Geol., 69, 1974, 757-76.
8. Rye R. O. and Ohmoto H. Sulfur and carbon isotopes
and ore genesis: a review. Econ. Geol., 69, 1974, 826-42.
9. White D. E. Anderson E. T. and Grubbs D. K. Geo-
thermal brine well: mile-deep drill hole may tap ore-bearing
magmatic water and rocks undergoing metamorphism.
Science, N. Y., 139, 1963, 919-22.
10. Craig H. Isotopic composition and origin of the Red
Sea and Salton Sea geothermal brines. Science, N. Y., 154,
1966, 1544-8.
11. Burnham C.W. Hydrothermal fluids at the magmatic
stage. In Geochemistry of hydrotherrnal ore deposits
Barnes H. L. ed. (New York: Holt, Rinehart and Winston,
1967), 34-76.
12. Davis N. F. and Burnham C.W. Energies involved in
crystallization of water-bearing magmas. Trans. Am.
geophys. Union, E~S, 50, 1969, 340. (Abstract)
13. Hamilton D. L. and Anderson G. M. Effects of water
and oxygen pressures on the crystallization of basaltic mag-
mas. In Basalts Hess H. H. and Poldervaart A. eds (New
York: Interscience, 1967), 445-82.
14. Jahns R. H. and Burnham C.W. Experimental studies
of pegmatite genesis: I. A model for the derivation and
crystallization of granitic pegmatites. Econ. Geol., 64, 1969,
843-64.
15. Nockolds S. R. Average chemical compositions of
some igneous rocks. Bull. geol. Soc. Am., 65, 1954,
1007-32.
16. Hamilton D. L. Burnham C. W. and Osborn E. F. The
solubility of water and effects of oxygen fugacity and water
content on crystallization in mafic magmas. J. Petrology,
5, 1964, 21-39.
17. Roedder E. Data of geochemistry, chapter J J, Com-
position of fluid inclusions. Prof. Pap. U.S. geol. Survey
440-J J, 1972, 164p.
18. White D. E. Mercury and base-metal deposits with
associated thermal and mineral waters. In Geochemistry of
hydrotherrnal ore deposits Barnes H. L. ed. (New York:
Holt, Rinehart and Winston, 1967), 575-631.
19. Roedder E. Fluid inclusion studies on the porphyry-
type ore deposits at Bingham, Utah, Butte, Montana, and
Climax, Colorado. Econ. Geol., 66, 1971,98-120.
20. Grant J. N. etal. Igneous geology and the evolution of
hydrothermal systems in some sub-volcanic tin deposits of
Bolivia. In Volcanic processes in ore genesis (London:
Institution of Mining and Metallurgy and Geological Society,
1977), 117-26.
21. Holland H. D. Granites, solutions, and base metal
deposits. Econ. Geol., 67, 1972, 281-301.
22. Sourirajan S. and Kennedy G. C. The system H 2 0 -
NaCI at elevated temperatures and pressures. Am. J. Sci.,
260, 1962, 115--41.
23. Fournier R.O. The importance of depth of crystal-
lization on the character of magmatic fluids. In 24th Int.
geol. Congr. (Montreal: The Congress, 1972), pt 10, 214
(Abstract).
24. Bottinga Y. and Craig H. High temperature liquid-
vapour fractionation factors for H 2 0 - H D O - H 2 0 1 8 . Trans.
Am. geophys. Union, E~S, 49, 1968, 356--7. (Abstract)
25. Phillips W. J. Mechanical effects of retrograde boiling
and its probable importance in the formation of some por-
phyry ore deposits. Trans. Instn Min. Metall. (Sect. B:
Appl. earth sci.), 82, 1973, B90-8.
26. Craig H. Isotopic variations in meteoric waters.
Science, N. Y., 133, 1961, 1702-3.
27. Bottinga Y. and Javoy M. Oxygen isotope partitioning
among the minerals in igneous and metamorphic rocks.
Rev. Geophys. Space Physics, 13, May 1975, 401-8.
39
28. Suzuoki T. and Epstein S. Hydrogen isotope fraction-
ation between OH-bearing silicate minerals and water.
Geochim. cosmochim. Acta, 40, 1976, 1229-40.
29. Urey H. C. The thermodynamic properties of isotopic
substances. J. chem. Soc., 1947, 562-81.
30. Bottinga Y. Calculated fractionation factors for carbon
and hydrogen isotope exchange in the system calcite-
carbon dioxide-graphite-methane-hydrogen-water
vapor. Geochim. cosmochim. Acta, 33, 1969, 49-64.
31. O'Neil J. R. and Taylor H. P. Jr. The oxygen isotope
and cation exchange chemistry of feldspars. Am. Miner.,
52, 1967, 1414--37.
32. Blattner P. Oxygen isotopic composition of fissure-
grown quartz, adularia, and calcite from Broadlands geo-
thermal field, New Zealand with an appendix on quartz-
K-feldspar-calcite-muscovite oxygen isotope geothermo-
meters. Am. J. Sci., 275, 1975, 785-800.
33. Wenner D. B. and Taylor H. P. Jr. Temperatures of
serpentinization of ultramafic rocks based on O18/O 16
fractionation between coexisting serpentine and magnetite.
Contrib. Mineral. Petrol., 32, 1971, 165-85.
34. Savin S. M. and Epstein S. The oxygen and hydrogen
isotope geochemistry of clay minerals. Geochim. cos-
mochim. Acta, 34, 1970, 25--42.
35. Rye R. O. and O'Neil J. R. The 018 content of water
in primary fluid inclusions from Providencia, North-
Central Mexico. Econ. Geol., 63, 1968, 232-8.
36. Sugisaki R. and Epstein S. Unpublished data, 1968.
37. Epstein S. and Mayeda T. Variation of 0 TM content of
waters from natural sources. Geochim. cosmochim. Acta,
4, 1953, 213-24.
38. Friedman I. Deuterium content of natural waters and
other substances. Geochim. cosmochim. Acta,4,1953,89-103.
39. Craig H. and Gordon L. I. Deuterium and oxygen 18
variations in the ocean and the marine atmosphere. In
Symposium on marine geochemistry, 1964. Rhode Island
Univ. Narraganset Marine Lab. occas. Pub/. 3, 1965, 2 7 7 -
374.
40. Shackleton N. J. and Kennett J. P. Paleotemperature
history of the Cenozoic and the initiation of Antarctic
glaciation: oxygen and carbon isotope analyses in DSDP
sites 277,279 and 291. Initial Rep. Deep Sea Drilling
Project, 29, 1975, 743-55.
41. Knauth L. P. and Epstein S. Hydrogen and oxygen
isotope ratios in nodular and bedded cherts. Geochim. cos-
mochim. Acta, 40, 1976, 1095--108.
42. Dansgaard W. Stable isotopes in precipitation. Tellus,
16, 1964, 436-68.
43. Friedman I. et al. The variation of the deuterium con-
tent of natural waters in the hydrologic cycle. Rev. Geo-
physics, 2, 1964, 177-224.
44. Epstein S. Sharp R. P. and Gow A. J. Antarctic ice
sheet: stable isotope analyses of Byrd station cores and
interhemispheric climatic implications. Science, N. Y., 168,
1970, 1570-2.
45. Heaton T. H. E. and Sheppard S. M. F. Hydrogen and
oxygen isotope evidence for sea-water-hydrothermal
alteration and ore deposition, Troodos complex, Cyprus.
In Volcanic processes in ore genesis (London: Institution
of Mining and Metallurgy and Geological Society, 1977),
42-57.
46. Muehlenbachs K. Anderson A. T. Jr. and Sigvaldason
G.E. Low-O TM basalts from Iceland. Geochim. cosmochim.
Act'a, 38, 1974, 577--88.
4O
47. Taylor H. P. Jr. and Turi B. High-180 igneous rocks
from the Tuscan magmatic province, Italy. Contrib.
Mineral. Petrol., 55, 1976, 33-54.
48. Lombardi G. and S~heppard S. M' F. Petrographic and
isotopic studies of the altered acid volcanics of the T o l f a -
Cerite area, Italy: the genesis of the clays. Clay Minerals,
in press.
49. Sheppard S. M. F. The Cornubian batholith, south-west
England: D/H and 180/160 studies of kaolinite and other
alteration minerals. J/geol. Soc. Lond., 133, pt 6 1977, in
press.
50. Jehl V. Poty B. Weisbrod A. and Sheppard S. M. F.
In preparation.
51. White D. E. Barnes I. and O'Neil J. R. Thermal and
mineral waters of nonmeteoric origin, California coast
ranges. Bull. geol. Soc. Am., 84, 1973, 547-60.
52. Clayton R. N. etal. The origin of saline formation
waters: I. Isotopic composition. J. geophys. Re~, 71,
1966, 3869-82.
53. Hitchon B. and Friedman I. Geochemistry and origin
of formation waters in the western Canada sedimentary
basin - I. Stable isotopes of hydrogen and oxygen.
Geochim. cosmochim. Acta, 33, 1969, 1321--49.
54. Kharaka Y. K. Berry F. A. F. and Friedman I. Isotopic
composition of oil-field brines from Kettleman North Dome,
California, and their geologic implications. Geochim.
cosmochim. Acta, 37, 1973, 1899-903.
55. Langley K. M. and Sheppard S. M. F. Hydrogen and
oxygen isotope evidence for the origins of the sub-surface
waters in the Coal Measures and Permian of North East
England. Unpublished manuscript.
56. Hall W. E. and Friedman I. Composition of fluid in-
clusions, Cave-in-Rock fluorite district, Illinois, and Upper
Mississippi Valley zinc-lead district. Econ. Geol., 58,
1963, 886-911.
57. Heyl A. V. Landis G. P. and Zartman R. E. Isotopic
evidence for the origin of Mississippi Valley-type mineral
deposits: a review. Econ. Geol., 69, 1974, 992-1006.
58. Pinckney D. M. and Rye R. O. Variation of O18/O TM,
C13/C12, texture, and mineralogy in altered limestones in
the Hill mine, Cave-in-Rock district, Illinois. Econ. Geol.,
67, 1972, 1-18.
59. Craig H. The isotopic geochemistry of water and car-
bon in geothermal areas. In Nuclear geology on geothermal
areas, Spoleto 1963 Tongiorgi E. ed. (Pisa: Consiglio
Nazionale della Ricerche, Laboratorio de Geologia
Nucleate [1967] ), 17-53.
t
60. Arnason B. and T6masson J. Deuterium and chloride
in geothermal studies in Iceland. Geothermics, Special
Issue 2, 1970, 1405-15.
61. Giggenbach W. Isotopic composition of waters of the
Broadlands geothermal field. N.Z.J. Sci., 14, 1971, 9 5 9 -
70.
62. Sakai H. and Matsubaya O. Isotopic geochemistry of
the thermal waters of Japan and its bearing on the Kuroko
ore solutions. Econ. Geol., 69, 1974, 974-91.
63. White D. E. Diverse origins of hydrothermal ore fluids.
Econ. Geol., 69, 1974, 954-73.
64. Ellis A. J. The chemistry of some explored geothermal
systems. In Geochemistry of hydrothermal ore deposits
Barnes H. L. ed. (New York: Holt, Rinehart and Winston,
1967), 465-514.
 65. Gustafson L. B. and Hunt J. P. The porphyry copper
deposit at El Salvador, Chile. Econ. Geol., 70, 1975,
857-912.
66. Sheppard S. M. F. and Gustafson L. B. Oxygen and
hydrogen isotopes in the porphyry copper deposit at El
Salvador, Chile. Econ. Geol., 70, 1976, 1549-59.
67. Lowell J. D. and Guilbert J. M. Lateral and vertical
alteration-mineralization zoning in porphyry ore deposits.
Econ. Geol., 65, 1970, 373-408.
68. Rowe R. B. Porphyry deposits of the Canadian
Cordillera. Can. Min. J., 94, 1973, Nov., 3 5 - 9 ; Dec., 37-41.
69. Laznicka P. Porphyry copper and molybdenum de-
posits of the U.S.S.R. and their plate tectonic settings.
Tran~ Instn Min. Metall. (Sect. B: Appl. earth sci.), 85,
1976, B14-32.
70. Hall W. E. Friedman I. and Nash J.T. Fluid inclusion
and light stable isotope study of the Climax molybdenum
deposits, Colorado. Econ. Geol., 69, 1974, 884-901.
71. Ford J. H. eta/. Stable isotope studies on Bougainville
and in Matupi Harbour, New Britain, Papua New Guinea.
In Volcanic processes in ore genesis (London: Institution of
Mining and Metallurgy and Geological Society, 1977), 162.
(Summary).
72. Garlick G. D. and Epstein S. The isotopic composition
of oxygen and carbon in hydrothermal minerals at Butte,
Montana. Econ. Geol., 61, 1966, 1325-35.
73. Sheppard S. M. F. and Taylor H. P. Jr. Hydrogen and
oxygen isotope evidence for the origins of water in the
Boulder batholith and the Butte ore deposits, Montana.
Econ. Geol., 69, 1974, 926-46.
74. Etminan H. and Sheppard S. M. F. Unpublished data.
75. Taylor H. P. Jr. O18/O 16 evidence for meteoric-
hydrothermal alteration and ore deposition in the Tonopah,
Comstock Lode, and Goldfield mining districts, Nevada.
Econ. Geol., 68, 1973, 747-64.
76. O'Neil J. R. etal. Stable isotope and chemical relations
during mineralization in the Bodie mining district, Mono
County, California. Econ. Geol., 68, 1973, 765-84.
77. O'Neil J. R. and Silberman M. L. Stable isotope rela-
tions in epithermal A u - A g deposits. Econ. Geol., 69, 1974,
902 - 9 .
78. Robinson B.W. The origin of mineralization at the
Tui mine, Te Aroha, New Zealand, in the light of stable
isotope studies. Econ. Geol., 69, 1974, 910-25.
79. Bethke P. M. Rye R. O. and Barton P. B. Jr. Hydro-
gen, oxygen and sulfur isotopic compositions of ore fluids
in the Creede district, Mineral County, Colorado. Geol.
Soc. Am. Abstr., 5, 1973, 549.
80. Rye R. O. and Sawkins F.J. Fluid inclusion and
stable isotope studies on the Casapalca A g - P b - Z n - C u
deposit, Central Andes, Peru. Econ. Geol., 69, 1974,
181-205.
81. Landis G. P. and Rye R.O. Geologic, fluid inclusion,
and stable isotope studies of the Pasto Bueno tungsten-
base metal ore deposit, Northern Peru. Econ. Geol., 69,
1974, 1025-59.
82. Taylor H. P. Jr. Oxygen isotope evidence for large-
scale interaction between meteoric ground waters and
Tertiary granodiorite intrusions, western Cascade Range,
Oregon. J. geophys. Res., 76, 1971, 7855-74.
83. Nash J. T. Fluid-inclusion studies of some gold
deposits in Nevada. Prof. Pap. U.S. geol. Surv. 800-C,
1972, C15-9.
84. Taylor H. P. Jr. Oxygen and hydrogen isotope evidence
for large-scale circulation and interaction between ground
waters and igneous intrusions, with particular reference to
the San Juan volcanic field, Colorado. In Geochemical
transport and kinetics Hoffman A. W. et al. eds
(Washington, D.C.: Carnegie Institution, 1974), 299-324.
(Publ. 634).
85. Anderson C.A. Massive sulfide deposits and volcan-
ism. Econ. Geol., 64, 1969, 129-46.
86. Hutchinson R.W. Volcanogenic sulfide deposits and
their metallogenic significance. Econ. Geol., 68, 1973,
1223-46.
87. Gass I. G. Origin and emplacement of ophiolites. In
Volcanic processes in ore genesis (London: Institution of
Mining and Metallurgy and Geological Society, 1977),
72-6.
88. Lambert I. B. and Sato T. The Kuroko and associated
ore deposits of Japan: a review of their features and
metallogenesis. Econ. Geol., 69, 1974, 1215-36.
89. Constantinou G. and Govett G. J. S. Genesis of sul-
phide deposits, ochre and umber of Cyprus. Trans. Instn
Min. Metall. (Sect. B: Appl. earth sci.), 81, 1972, B34-46.
90. Constantinou G. and Govett G. J. S. Geology,
geochemistry, and genesis of Cyprus sulfide deposits. Econ.
Geol., 68, 1973, 843-58.
91. Jehl V. Le m~tamorphisme et les fluides associ~s des
roches oc~aniques de I'atlantique nord. Th~se, U niversit~
de Nancy, 1975, 242p.
92. Bostrom K. and Peterson M. N. A. Precipitates from
hydrothermal exhalations on the East Pacific Rise. Econ.
Geol., 61, 1966, 1258-65.
93. Corliss J. B. The origin of metal-bearing submarine
hydrothermal solutions. J. geophys. Res., 76, 1971,
8128-38.
94. Robertson A. H. F. Cyprus umbers: basalt-sediment
relationships on a Mesozoic ocean ridge. Jlgeol. Soc.
Lond., 131, 1975, 511-31.
95. Easton A.J. Hamilton D. Kempe D. R. C. and
Sheppard S. M. F. Low-temperature metasomatic garnets
in marine sediments. Phil. Trans. R. Soc., in press.
96. Ohmoto H. and Rye R. O. Hydrogen and oxygen
isotopic compositions of fluid inclusions in the Kuroko
deposits, Japan. Econ. Geol., 69, 1974, 947-53.
97. Colley H. and Rice C. M. A Kuroko-type ore deposit
in Fiji. Econ. Geol., 70, 1975, 1373-86.
98. Colley H. Rice C. M. and Sheppard S. M. F. Un-
published data.
99. Lawrence J. R. and Taylor H. P. Jr. Deuterium and
oxygen 18 correlation: clay minerals and hydroxides in
Quaternary soils compared to meteoric waters. Geochim.
cosmochim Acta, 35, 1971, 993-1003.
100. Taylor H. P. Jr. and Forester R.W. Low-O 18 igneous
rocks from the intrusive complexes of Skye, Mull, and
Ardnamurchan, western Scotland. J. Petrology, 12, 1971,
465-97.
101. Kilinc I. A. and Burnham C. W. Partitioning of
chloride between a silicate melt and coexisting aqueous
phase from 2 to 8 kilobars. Econ. Geol., 67, 1972, 231-5.
102. Sheppard S. M. F. Brown P. E. and Chambers A. D.
The Lilloise intrusion, east Greenland: hydrogen isotope
evidence for the efflux of magmatic water into the contact
metamorphic aureole. In Fourth European Colloquium of
geochronology, cosmochronology and isotope geology,
Amsterdam, 1976, 84. (Abstract).
41

Hydrogen and oxygen isotope
evidence for sea-water-hydro-
thermal alteration and ore
deposition, Troodos complex,
Cyprus
T. H. E. Heaton
Grant Institute o f Geology, Edinburgh, Scotland
S. M. F. Sheppard
Scottish Universities Research Reactor Centre,
East KiJbride, Glasgow, Scotland
539.155. 2 : 5 5 3 . 2 1 6 . 5 : 5 5 3 . 4 3 5 ( 5 6 4 . 3 )
Synopsis
A mode/is presented for the genesis of the Cyprus massive
cupriferous sulphide deposits, their associated stockwork
mineralization and the metamorphism of the Troodos
ophiolite complex based on D/H, 180//160 and 13Cp2C
ratio analyses, where applicable, of 94 whole rocks and
minerals, chemical, mineralogical and sulphur iso tope data.
The upper 3 - 5 km of the hot Troodos c r u s t - pillow
lavas, Sheeted Intrusive Complex, trondhjemites and upper
gabbros- were metamorphosed to zeolite facies, tempera-
tures of about 0 to 250~ and greenschist-amphibolite
facies, temperatures of about 250 to 450--550~ during
interaction with a deep circulating sea-water geothermal
system. Minimum isotopic temperatures of 320--400~ are
given for the epidote-bearing vein-halo assemblages in the
upper gabbros, trondhjemites and Sheeted Intrusive Complex.
/ f temperatures at the base of the Sheeted Intrusive Complex
were higher than " 500~ either a magmatic or metamor-
phic water component must have been present in the hydro-
thermal fluids. The magmatic water could be of sea-water
origin, being derived from the melting or assimilation o f
sea-water-hydrothermally altered roof rocks overlying the
gabbroic magma chambers Minimum water/rock ratios
during pervasive alteration decrease from > I by weight in
zeolite facies pillow lavas to "~ O. 1 in the underlying green-
schist facies rocks, but were an order of magnitude larger in
the veins and fractures.
In the stockwork zones, which underlie the sea-floor sites
of massive sulphide orebodies, isotopic temperatures de-
clined from initial values greater than 300~ (with water/
rock ratios > I - 5 , by weight) to less than 220~ during
post-sulphide later-stage vein quartz formation; these tem-
peratures were higher than temperatures in the predomi-
nantly zeolite facies country rock~ The stockwork zones
are interpreted to represent the discharge zones of the sea-
water-hydrothermal fluids responsible for the regional-scale
mass transfer and the mineralization. A t least some o f the
sulphide, sulphur in the stockworks and orebodies is of sea-
water origin. Formation of massive sulphide deposits
requires the coincidence of several factors, including the
penetration of a brine to deep levels, where temperatures
were greater than 300 ~ in an active spreading centre, leach-
ing of ore metals, and discharge of the hot springs into
basins where both sulphide deposition and localization could
occur.
Massive sulphide deposits are closely associated with the
basic volcanic rocks of many ophiolite sequences (e.g.
42
Cyprus; Oman; Bay of Islands, Newfoundland; Buena Vista
Islands, British Solomon Islands; Balabac Island, Philippines;
Kure, Turkey). These ophiolites are considered to represent
uplifted fragments of oceanic crust. 1 The genesis of the
massive sulphide deposits is related to the development and
evolution of the crust in the sub-sea-floor environment, and
the role of hydrous fluids in such an environment is of par-
ticular importance. To understand the magmatic and post-
magmatic processes in such an environment we require
knowledge of the origins of these fluids.
Recent studies have demonstrated that, in general, a com-
bined hydrogen and oxygen isotope approach is necessary to
distinguish between waters of different origin. 2 This stable
isotope study attempts to set limits on the origins of waters
and the nature of their interaction with different rock types
during the metamorphic and hydrothermal processes that
occurred in a section of oceanic crust, and to give informa-
tion about the environment of ore deposition of massive
sulphide deposits.
Although currently there are several active hydrogen and
oxygen isotope studies of ophiolite and oceanic dredge and
core material, few data are widely available. Isotopic evi-
dence for the interaction of sea water with hot oceanic
igneous rocks was first proposed by Sheppard and Epstein 3
from studies of the hornblende peridotites and serpentines
from the St Paul's Rocks, mid-Atlantic Ridge. Additional
evidence for the interaction of sea water with hot igneous
rocks comes from serpentinized ultramafic rocks, 4 recon-
naisance data from the Troodos complex, s, 6 ophiolites in
the Franciscan, California, 7 and Kuroko base-metal
deposits. 8
In a number of studies on dredged oceanic crust and
ophiolite samples oxygen isotope data on their own have
been interpreted as being consistent with the interaction of
sea water. 9, lO, 11,39 Because waters of more than one
origin (primary magmatic water, sea water, etc.) could be
involved, however, the oxygen isotope data alone cannot be
so simply interpreted, as was pointed out by Muehlenbachs
and Clayton. TM
The Troodos ophiolite complex, Cyprus, was selected for
this study for the following reasons: (1) the igneous complex
and associated orebodies are, geologically, one of the better
known examples; (2) this fragment of inferred oceanic crust
has experienced minimal horizontal deformation during up-
lift; and (3) interpretation of the nature of the metamorphic
and hydrothermal processes that occurred in the submarine
environment has not been complicated by subsequent meta-
morphic and/or magmatic processes.
Geological setting
Genera/
It has been suggested that the Troodos massif, like many
ophiolites, may have formed in a behind-arc spreading axis
situated above a subducting plate in a marginal ocean basin
environment. 1,87 As we are concerned with determining
the origins of the fluids, waters of three potential origins
must be considered - primary magmatic water, sea water
and water released during dehydration of a subducted plate
(hereafter termed metamorphic water).
Troodos massif A general ized map of the geology of the
Troodos complex showing the location of the more import-
ant massive sulphide deposits is given in Fig. 1. The follow-
ing summary of the geology is based on a wide variety of
detailed studies, many of them recent. 12-21
The centre of the Troodos annular structure, the Mount
Olympus region (approximately 2000-m elevation), is in
harzburgites, now largely tectonized and heavily serpent-
inized, and interpreted to represent depleted mantle. Above
this, geologically, lies a thick layer of ultrabasic and basic
layered cumulates (peridotites, gabbros, etc.) with trond-
hjemite bodies locally at the top. The term trondhjemite
(see Streckeisen 22) is used here to describe collectively
rocks also termed granophyres, quartz diorites, plagiogran-
ites, 23 etc. These plutonic rocks pass up into the basic
Sheeted Intrusive Complex (S.I.C.) - a unit composed al-
most entirely of near-vertical dykes and estimated to be
about 2 km thick. The dyke complex merges into an
approximately 1-km thick pile of basic volcanic rocks of
tholeiitic parentage, which is dominated by pillow structures
- the Lower Pillow Lavas (L.P.L.). The Troodos complex
is Upper Cretaceous in age, and was uplifted during Miocene
times.
Most of the massive sulphide deposits, consisting of
pyrite with varying amounts of chalcopyrite and rarer
sphalerite, are interpreted to have formed in tectonically
controlled basins or depressions in the sea-floor at the sea
water-L.P.L, interface. The Skouriotissa orebody lies on
the Upper Pillow Lavas (see below), but has been over-
thrusted from its inferred position at the top of the L.P.L.
These massive sulphide deposits are underlain by a zone of
highly altered pillow lavas, dominantly a quartz-chlorite-
pyrite assemblage, which are cut by a ramifying network of
quartz-chalcedony-pyrite-chalcopyrite veins and veinlets,
called the stockwork zone.
The Lower Pillow Lavas, Sheeted Intrusive Complex and
at least the upper parts of the gabbros have been affected
by a' pervasive low-pressure, low- to high-temperature type
zeolite-greensch ist-amphibolite facies metamorphism.
The mineral assemblages given below do not distinguish
between the relict igneous and new metamorphic minerals.
Lower Pillow Lavas are dominantly in the zeolite facies,
with, typically, the assemblage plagioclase-pyroxene-
altered glass-smectite-celadonite- zeolites-chalcedony-
quartz--calcite--magnetite.
The transition with depth to greenschist facies - the con-
version of an essentially plagioclase (An80_90), pyroxene
and magnetite assemblage to a plagioclase (Ans_as)-
pyroxene--quartz-chlorite-actinolite-epidote-ilmenite-
sphene-leucoxene-pyrite assemblage - very approximately
coincides with the Lower Pillow Lava-Sheeted Intrusive
Complex transition. With increase in depth in the S,I.C. and
the upper parts of the gabbros, chlorite becomes rare, actin-
olite becomes less abundant, and the abundance of meta-
morphic hornblende progressively increases; there is also an
increase in the anorthite content of altered plagioclase.
This suggests the development of an amphibolite facies, al-
though we appreciate the possibility that in a greenschist
facies environment Ca-rich plagioclase may be stable. 24 In
a trondhjemite outcrop very close to gabbros the assem-
blage garnet (~95% andradite) + plagioclase (An86_89) +
quartz + magnetite was observed, sug0esting the develop-
ment, at least locally, of temperatures higher than 600~
(2.0 kbar fluid). 2s,26 This may be a result of contact
metamorphism by the gabbro.
Within the lower parts of the L.P.L., and particularly in
the S.I.C., trondhjemites and upper parts of the gabbros,
quartz-rich veins with well developed epidotic haloes cut the
pervasively altered rocks (epidote also occurs in the deeper
parts of some stockworks). These veins, like the stockwork
zones, contain assemblages different from those of the
surrounding rocks. They are interpreted to represent the
major channelways for the hydrothermal solutions.
The stockwork zones in the Lower Pillow Lavas have
usually been considered to be pipe-like in form, but be-
cause of the widespread prevalance of normal faulting their
shapes are not well known. 2~ Outcrops of Lower Pillow
Lavas are, however, overwhelmingly in the zeolite facies.
The stockworks therefore must have a restricted horizontal
extent, although not necessarily as restricted as that of the
massive sulphide deposits (typically 104 m 2 in area).
The massive sulphide deposits and metamorphosed Lower
Pillow Lavas are overlain, generally unconformably, by the
Upper Pillow Lavas (U.P.L.). These basalts and olivine
basalts are interpreted to represent off-axis magmatic
activity; they subsequently underwent zeolite facies meta-
morphism. Fe-Mn-rich sediments (umbers) lie immediately
above and locally within the U.P.L. These Upper Pillow
Lavas and sediments, which are later than the development
of the major part of the Troodos ophiolite sequence and its
metamorphism, are of less concern here.
Experimental
Sample preparation
Whole-rock samples were crushed to less than 250/~m, the
development of too much fine power being avoided. Miner-
als were separated by standard density and magnetic tech-
niques with a purity of better than 95%. To separate quartz
from the finely intergrown quartz-chlorite-pyrite ground-
mass assemblage of stockwork samples chlorite was removed
chemically (after density separation of pyrite) by a method
outlined by Syers eta/. 27 In thin section this groundmass
quartz is optically homogeneous. The oxygen isotope com-
position of a groundmass quartz, which was not chemically
separated, because of the absence of chlorite, is isotopically
similar to our chemically separated quartz values. Therefore,
we conclude that the chemical separation method does not
affect the isotopic composition of our groundmass quartz
samples.
Isotopic analysis
Well established analytical methods 2s, 29, 30 were used to
extract hydrogen, carbon and oxygen from minerals and
whole rocks after removal of all absorbed or interlayer
water. Isotopic ratios were determined on CO2 and H2
gases in McKinney-Nier type mass spectrometers.
Isotopic notation
The isotopic data are presented as the enrichment of the
heavy isotope in the sample, compared to a standard in per
mil
6~ sample = ( Rsarnple. 1~ x 1000
\ Rstandard /
where R = D/H, i80/160, 13C/12Cand 34S/32S. For hydro-
gen and oxygen the standard is Standard Mean Ocean Water
(SMOW); for carbon PDB; and for sulphur Ca~_on Diablo
troilite.
Exchange experiments
Exchange experiments were conducted on a variety of
samples with D- and 1SO-enriched waters to determine
whether hydrogen and/or oxygen isotope exchange with
water - for example, atmospheric water vapour - is so
rapid as to prevent meaningful isotopic analysis. 21 The
results obtained showed that hydrogen isotope analyses of
many of our zeolite facies rocks are generally of limited use
geologically because hydrogen exchange occurs within a few
days at room temperature.
43
Mineral-H20 fractionation factors
The hydrogen and oxygen fractionation curves summarized
by Taylor 31 are used here to calculate the isotopic compo-
sition of water in equilibrium with our minerals. Additional
data on chlorite, actinolite, hornblende and epidote are
derived from the experimental studies of Suzuoki and
Epstein, 32 studies on epidote from active geothermal sys-
tems (S. M. F. Sheppard, unpublished) and chlorites from
oceanic dredge samples with known fluid inclusion
homogenization temperatures. 33 Significantly, both the
epidote and chlorite data are for temperatures directly
comparable with those of our Troodos samples. In addition,
when the composition effect on hydrogen isotope fraction-
ations is taken into account, the chlorite data broadly
support the empirical curve given by Taylor. 31 Because of
the critical influence of mineral chemistry on the hydrogen
isotope fractionation factors, 32 representative minerals were
chemically analysed (Table 2) and these data are used in our
calculations.
Sample localities and isotopic results
The distribution of our sample localities is shown in Fig. 1.
Fig.1 Generalizedgeological map of Troodos complex, Cyprus, showing locations of
major massive sulphide orebodies and sample localities
The ~180 and ~D data on whole rock and mineral separates
from these localities are given in "t'abl~ 1 and in Figs. 2, 3
and 4. Because of mineral separating problems, not all
samples could be analysed for both D/H and 180/160 ratios.
Data for serpentinized samples and for the umber sediments
are not considered here.
Discussion
The whole-rock ~ 180 data for the Troodos samples are dis-
cussed first and compared with data from other ophiolite
complexes and from oceanic samples. The isotopic charac-
teristics of the individual minerals are then outlined. The
Troodos ~ 180 whole-rock data discussed are for non-vein
associated samples, with mineralogies summarized above.
44
Whole,rock samples- 180/160 ratios
Fig. 2 compares the ~180 values of fresh sub-oceanic basalts
with data for basalts, dolerites and gabbros from oceanic
dredge and drill-cores, and from ophiolites -Troodos, which
has suffered no syn- or post-emplacement metamorphism,
and Pindos and Liguria, which have. Fresh oceanic basalts
with H2 O+ ~< 0.45%, both from "normal' and 'marginal'
ocean basins, have a restricted range of ~ 180 values,
+5.5 to +6.2%o, compared to which (data of this study):
(1) Five Troodos gabbros have similar values, ranging
+5.3 to +6.5%o, average +6.0%o; they include two fresh
gabbros with ~180 values +5.7 to +6.5%o. The Trood(';
gabbros are, on average, 1.2%o heavier than four oceani~
gabbros, which average +4.8%o (Fig. 2). The Troodos
gabbro samples, however, contained 60-70% plagioclase,
most of which was strongly twinned and appeared fresh,
and for one sample plagioclase ~ 180 = +6.2%o; the remain-
ing 30-40% of the gabbros were essentially pyroxene or
amphibole (below). In contrast, for the two 180-depleted
oceanic gabbro samples on which mineral separates were
made, the plagioclase was considered to be isetopically al-
tered, yielding ~ 180 values +4.0 and +4.9O/oo.37' 39 This
slight difference between Troodos gabbro and some oceanic
gabbro whole-rock ~ 180 values may be a consequence of
differences in the temperature and relative masses of the
exchanging reservoirs during alteration, and is not necessarily
due to differences in the initial oxygen isotope compositions
of the magmas.
(2) The majority of the S.I.C. metadolerites analysed have
~180 values in the range +4.4 to +6.0%o (6 samples; average
+5.4%o). These values are slightly lower than those reported
for oceanic dolerites (Fig. 2). This probably reflects
differences in modal mineralogy: the Troodos S.I.C. rocks
commonly contain more of the relatively 180-poor chlorite
(up to 35%) and less of the relatively 180-rich quartz than
the analysed oceanic dolerites (mineralogy in Aumento and
 Table 1 Hydrogen, oxygen and carbon isotope analyses Table 1 Hydrogen, oxygen and carbon isotope analyses
of minerals and w h o l e rocks of minerals and w h o l e rocks (continued)
6180, 6180, 6180, 6180,
Sample no.* Mineralogyt 6D:]: whole rockw mineral** Sample no.* Mineralogyt 6Dr whole rock w mineral**

Upper Pillow Lavas Sheeted Intrusive Complex

3--3 WR -66 + 12.8 +31.3c v 5 pg, ch, q --43 + 9.7
10--2 (c) +31.6(c) w 16--3 pg, ch, px --38ch + 8.7 + 2.7ch
29--2 WR --54 +10.6pg
61 --2 WR -58 +29.1 c x 23 pg, ch, px, q --39ch + 3.9ch
40/202 WR --54 24 [q, ep] -- 3[ep]
Sk. 1 - 4 WR --78 + 16.0 32 pg, ch, q, --38ch + 5.8
Sk.2--4 WR +15.8 [ep] + 2[ep] + 0.9 [ep]
Ms.30 WR -43 33--1 [q, ep] + 3.4[ep]
435 WR -61 +11.8 + 9.5[q]
33--2 pg, ch, q --37ch + 5.6 + 3.4ch
Lower Pillow Lavas 34--2 pg, ac, hd --43 + 6.0
66 WR --54 +11.4 68p pg, ch, q, -- 9(ep) + 0.2(ep)
67--1 rim WR -40 + 7.7 (q, ep, c) --34pg, ch +16.0c z
6 7 - 4 core WR -40 70 pg, ch, q -43 + 6.0
92 (q) +31.7(q) 89 [ep] + 7[ep] + 0.1 [ep]
95-1 [q] +30.9 [q] Cy42 pg, ac --47ac + 5.6 + 3.8ac
40/258 WR -63 + 9.5 Cy67 pg, ac, q --49 + 4.4
401362 [c] +30.4 [c] Y
401532 WR -63 +10.3 Trondhjemites
401652 WR -64 6 pg, q, am + 4.4
409 WR +12.5 11--1 [q, ep] + 6.4 [q]
+ 4[ep] 0.0[ep]
Stockworks 20--1 pg, q, [ep] -- 3[ep]
Ms.21 q, ch --37 49 pg, ac, q, ch --52ac + 5.3
Ms.31 ch, q --39 + 2.5 + 5.5q Cy92 pg, q, am + 5.0
Ms.3/109 q, ch, [q] +10.7 [q]
Ms.3/113 q, ch, [q] +14.1 [q] Gabbros
Ko.1 q, ch --33 7--1 [q, ep] + 3[ep]
Ko.2 q + 6.9q 7--3 ac --54ac + 4.3ac
Mat.6 q, ch, [q] +14.1 [q] 13 pg, ac, hd --51ac + 5.7ac
Mat.7 ch, q --37 + 4.1 + 7.0q 15--1 pg, hd, ac --44am + 5.3am
Li.3 q, ch -41 + 3.7 + 6.0q +10.5pg
Li.Q q, ch, [q] +10.5[q] 15--2 pg, hd, ac, --17 + 6.2
Sk.11 --2 q, ch + 2.9 ep, q + 5ep + 3.8ep
Sk.74/1425 q, ch, (ep) --13(ep) --45am + 4.8am
Sk.74/1475 ch, q, (ep) --39q, ch 35--2 [q, ep, ac] -- 3[ep] + 1.6[ep]
--18 (ep) + 0.2 (ep) --47[ac] + 1.5[ac]
121266 q, ch, [q] +11.0 [q] + 6.8[q]
12/624 ch, q --40 + 1.4 + 5.7q 35--4 pg, ac, hd, [ep] + 0.8 am
12/665 ch, q, (ep) -- 5(ep) 45 pg, px + 5.7
75 pg, px + 6.5
79 pg, q, hd, --51hd + 7.6q
ac, px
107 pg, ac, hd -47am + 6.4
121 pg, hd, ac --51 hd + 5.2hd
I T - 84 ' I ' I ' I
129 pg, ac, hd --51am + 5.3 + 4.6am
Sub-oceanic Basalt s
28 + 6.2pg
Fresh
Metabasolts IIIi I
Met o d o l er ites *See Fig. 1 for sample localities. Abbreviations: Ms. = Mousoulos,
Metagabbros Ko. = Kokkinoyia, Mat.-- Mathiati, Li. = Limni, Sk. = Skouriotissa.
Troodos
tAbbreviations: WR = whole rock, ac = actinolite, ~am= amphibole,
PilLow lavas c = calcite, ch = chlorite, ep = epidote, hd= hornblende, pg = plagio-
S.I.C o o,p~l~ clase, px = pyroxene, q -- quartz. Order of mineralogy refers to
Gabbros o~ eqb relative modal abundance of major minerals given in decreasing
sequence -- excludes magnetite, pyrite, ilmenite, sphene, leucoxene.
Pindos
u [], () = minerals in veins, in vesicles, respectively.
E Basalts I ~ I
o
.~ Metadoler ites I 56 D values refer to whole-rock analysis, except where specified.
r' ~ Metagabbros H Most values represent means of duplicate analyses with a precision
.~'o E. Liguria of -+1 or 2~
~. Basalts ~ ~ ~' ~ ~
wMost values represent mean of duplicate analyses with a precision
I L L- L I 1 , [ of +0.1 or 0.2~
0 +4 +8 -.12 +16
*'513C values for calcites: v = +2.6, w = +2.0, x = +0.7, y = +2.6,
8~80, %0 z = +2.1, per mil.

Fig. 2 S u m m a r y of w h o l e - r o c k oxygen isotope data for

oceanic and ophiolite basalts, dolerites and gabbros. Ocean-
ic data f r o m Taylor, 34 Garlick and D y m o n d , 35
Muehlenbachs and Clayton, 1~ 36, 39 Pineau eta/., 37 co-workers 86). T w o Troodos S.I.C. samples gave high
Easton etal., 38 Jehl etal. 33 Troodos data: crosses, 1 8 0 values, +8.7 and +9.7%o; the heavier sample, f r o m
Margaritz and T a y l o r ; s open circles, Spooner etal. 111 solid close t o the t o p o f the S.I.C., had a significant a m o u n t of
circles, this study; Pindos data: Javoy; 9 East Liguria data: quartz, and both samples contained at least 50% plagioclase
Spooner et al. 11 A n 5 _ 3 o , which f o r the lighter sample was +10.6%o.

45
(3) Zeolite facies U.P.L. and L.P.L. samples (Figs. 2 and
u PL I I
4) show a wide range of $180 values: U.P.L. + 1 1.8 to Pillow IQVQS
L~L I II I I
+16.0%o (four samples; average +14.1%o), L.P.L. +7.7 to
+12.5%o (five samples; average +10.3%o). These are within Stockworks I If { E)E~
9 e~ ~
the range for altered oceanic basalts (Fig.2). Nevertheless,
I IIIII I
they are more 180-rich than many of the altered oceanic S.I.C. oo ~
&
o
basalts. The two most 180-rich samples come from the &
9 Chlorite
l_ &-' 9 F
U.P.L. very close to Skouriotissa mine and may have been I II 9 Epidote
involved in supergene processes in the subaerial environment. Tr'ondhjemites r | 9 Actinolite

There is no evidence for subaerial processes affecting the E3 H o r n b l e n d e

II III
IB Act=-Hnbd.
other pillow lava samples, two of which were borehole
Gabbros ID ~ LL Jr | Quartz
samples from depths of 80 and 200 m. H2 O+ content of Jr FeldFDar
the pillow lavas ranged 1.2-2.3%, with no apparent correla- IT---+ @
E3 I Whole rock
tion between H2 0+ and whole-rock $180 value. The lowest Fresh oceanic basa[ts Jo~
~180 value (+7.7%o; 1.8% H2 0+) was for the glassy rim of I I I I I I I I I I
-2 0 -*2 +4 +6 +8 +10 .12 +14 +16
a L.P.I. pillow.
8'80, %0
Whole-rock samples - D/H ratios Fig. 4 Troodos whole-rock and mineral separate ~ 180
data. Tie-lines connect coexisting minerals. Fresh oceanic
Fig. 3 compares ~ D data for Troodos samples with zeolite
basalts from Fig. 2
and greenschist facies oceanic dredge samples. For many of
our samples (e.g. those lacking epidote) the D/H ratio of
the hornblende and actinolite are petrographically distinct,
their chemical compositions are less distinct (actinolite-
1 r I ~ - - - ~ i I ' I
aluminous actinolite; Table 2). We have no evidence for
41i 9
Stockwork 9 9 9 9 9

i Lavas UPL I II I II I Table 2 Compositions of hydrous minerals

L PL III I II
O--
Sample
S. IC. ~ 9149 9
=J Mineral no. , Composition
I l 9

Trondhjem~tes 9 Amphibole 6 Na0.13Cat.89(Mg2.2s Fe2.6sA10.21)

9
ID
ID
(SiT.s 1A10.49)O22(OH)2
GQbbros mm l 9 13 K0.03 Nao.20Ca1.71 (Mg3.s4Fe1.47A10.24)
9 Chlorite
[] (Si 7.S4AI0.46) 022 (OH)2
9 Epidote
St. Paul's Rocks O [] 15--1 K0.07 Na0.3sCa1.,~ (Mg2.6t Fe2.36AI0.27)
9 Act i n o l i t e
Sub-oceonic (Si 7.30AI0.70)O22(OH)2
bosalts and dolerites H H I I D Hornblende
Fresh continental ,=-90 Act~-Hnbd
15--2 Ko.03 Na0.26Ca1.82(Mg2.37Fe2 .S9AIo. 17)
bosolts
I Whole rock
(Si7.33A10.67)O22(OH)2
1
-80
,~
-60
I ,
-40
I , I
-20
, I 35--4 Ko.ot Nao.t 3Cat .6s (Mg3.09Fa1.77A10.29)
0
(SiT.ssAI0.4s)O22 (OH)2
8D,%o 79 Nao.osCal.88 (Mg4.2s Feo.TsA10.08)
(Si 7.8sAIo.ls )022 (OH)2
Fig. 3 Troodos whole-rock and mineral separate 6D data 121 Ko.07Nao,29Ca1.80(Mg2.98Fe2.03Alo.ts)
(whole-rock values for majority of stockwork, S.I.C., (Si7.37A10.63)022(OH)2
trondhjemite and gabbro samples that lack epidote are 121 K0.04 Nao.09Ca1.74(Mg3.29Fe2.47AI0.08 )
equivalent to the chlorite--amphibole mineral separate (Si7.64AI0.36)O22 (OH)2
values); compared with data from St. Paul's Rocks, 3 sub-
oceanic metamorphosed basalts and dolerites 33 and fresh Chlorite 5 Mg4.99Fe4.66AI2.24 (Sis.85AI2.15 )020
(OH)16
continental basalts (summarized by Taylor 31 )
16--3 Mg4.36Fe4.s2AI2.76(5i5.92AI 1.98)020
(OH) 16
32 Mg5.60Fe4.07AI2.18 (Si6.o2AI 1.98)020
the mineral closely represents the whole-rock ratio. There- (OH)t6
fore, the values observed for the stockworks, S.I.C., trond- Ms31 Mgs.soFe3.63AI2.69(Siso69AI2.31)02o
hjemites and gabbros are within the range of values observed (OH)16
for the small number of sub-oceanic metabasalts and meta- Li 3 Mg3.37Fes.70AI2.84(Sis.35AI2.6S)020
dolerites, many in the greenschist facies, that have been (OH)16
12/624 Mg6.48 Fe2.46AI2.77 (Sis .76 A I 2 . 2 4 ) 0 2 0
analysed. The ~D values of the sub-oceanic samples are
(OH)16
known to be due to exchange with hot sea water, the range
reflecting variations in the temperature and mineral Epidote 11--1 Ca1.99 (Al2.21 Feo.sT)Si3.010 12 (OH);
chemistry. 33 28% pistacite
The hydrogen and oxygen isotope compositions of 15--2 Ca 1.98 (AI2.24 Feo'69)Si3.00 O 12 (O H);
24% pistacite
Troodos mineral separates are discussed below and
summarized in Figs. 3 and 4. 20--1 Ca2.00 (AI 2.14 Feo. 74 )Si3.02 O 12 (O H );
26% pistacite
32 COl .98 (AI2. IS Fe0.83)Si2.98 O 12 (OH);
Amphiboles 28% pistacite
Isotopically analysed amphibole separates include blue-green 24 29--32% pistacite (refractive index
hornblende and fibrous actinolite, or, more commonly, determination)
intergrown mixtures of the two, since hornblende invariably 68 27--32% pistacite (refractive index
determination)
shows some retrograde alteration to actinolite. Although

46
significant differences in the isotopic characteristics of
Troodos hornblendes and actinolites; they are thus
considered together here.
Amphiboles from the gabbros, trondhjemites and lower
parts of the S.I.C. (excluding veins) have ~D values of
- 4 3 to -54%o (twelve samples; average -48%o), and ~ 180
values of +3.8 to +5.7%o (seven samples; average +4.8%o).
An actinolite from an epidote vein in the gabbros has
~D = -47%o, which is identical to our other actinolites, but,
together with another actinolite in epidote-veined gabbros,
is depleted in 180 (+1.5 and +0.8%o).
Chlorites
Chlorite samples come from the S.I.C. as part of the green-
schist facies assemblage, and from the stockworks. S.I.C.
chlorites have similar Fe/Fe + Mg ratios (pycnochlorite 4o),
whereas stockwork chlorites, although chemically similar
to S.I.C. chlorites, have a wider range of Fe/Fe + Mg ratios
(Table 2). Chlorites from the greenschist facies assem-
blages from the top to the bottom of the S.I.C. have ~ D =
- 3 4 to --43%0 (seven samples; average -39%0) and ~ 180 =
+2.7 to 3.9%0 (three samples; average +3.3%o). Chlorites
from the mineralized stockworks in the L.P.L. and upper
parts of the S.I.C. have ~D = - 3 3 to -41%o (seven samples;
average -38% o), which are indistinguishable from our other
metamorphic chlorites. In contrast, however, the stockwork
chlorites are depleted in 180 (approximately - 0 . 9 to
+1.4~ as calculated from whole-rock quartz-chlorite
data corrected by use of quartz separate data.
Epidotes
Epidotes have chemical compositions in the range 24-32%
pistacite (Table 2). Although vein, vesicle and groundmass
epidotes in a particular rock unit have similar compositions,
there is a progressive increase in the pistacite content from
the gabbros to the base of the L.P.L. Vein epidotes from
the S.I.C., trondhjemites and gabbros are strongly enriched
in deuterium relative to all other analysed minerals (+7 to
-3%o: seven samples; average +2%o) and have ~ 180 ---
+0.0 to +3.4%o (five samples; average +1.2%o). An epidote
from a gabbro that is not obviously related to a vein has a
similar D/H ratio (+5%o), but is enriched in 180 (+3.8%o)
relative to the vein epidotes. We also note that epidotes
from vesicles in deep borehole samples in stockwork from
beneath Skouriotissa mine at borehole depths of approxi-
mately 430 and 450 m, in an upper S.I.C. stockwork and in
an upper S.I.C. inter-pillow cavity have slightly lighter ~D
values ( - 5 to -18%o; four samples; average -10%o), but
similar ~ 180 values (two samples = +0.2%o) relative to the
vein epidotes.
Quartz
Quartz with epidote from veins in trondhjemite and gabbro
have ~ 180 values of +6.4 and +6.8%0, respectively, and
quartz from a gabbro pegmatite +7.6%0. Quartz from the
altered basalts in the stockworks (excludes veins), hereafter
termed 'groundmass quartz', has ~ 180 values of +5.5 to
+7+0%0 (five samples from different stockworks; average
+6.3%0). In contrast, vein quartz from the stockworks has
~180 values of +10.5 to +14.1%o (five samples; average
+12.1%o). Drusy quartz from a large vesicle from the lower
parts of the L.P.L. and a chalcedonic vein quartz from the
L.P.L. yielded ~180 values of +31.7 and +30.9~/oo,
respectively.
Calcites
Calcite pseudomorphous after olivine and as an inter-pillow
cavity filling in the U.P.L., and as a vein in a L.P.L. hyalo-
clastite, has ~ 13C values of +0.7 to +2.6%o (four samples;
average +2.000o), and ~ 180 +29.1 to +31.6%o (four samples;
average +30.6%o). An upper S.I.C. inter-pillow cavity
calcite has a ~13C value of +2.1%o and ~180 +16.0%o.
Isotopic composition of hydrothermal waters
The hydrogen and oxygen isotopic compositions of waters
in equilibrium with our samples are calculated here by use
of the mineraI-H20 fractionation factors discussed above.
These calculations require knowledge of the temperature.
Our temperature data are based on a few isotopic tempera-
tures and phase equilibria relations (and also on some fluid
inclusion data mentioned later). Zeolite facies metamor-
phism of the pillow lavas developed in the approximate
temperature range 0-250~ 17 and the greenschist facies
from about 250 to 450-550~ 33 The upper stability limits
of chlorite, actinolite and epidote-bearing assemblages are
dependent on pressure, composition and the oxygen fugacity
From thickness estimates of the Troodos crust (e.g.~Vine
and Moores 41 )pressures of about 1.5--2.0 kbar may be
expected for the upper gabbros. At these pressures the
breakdown of chlorite, increase in amphibole aluminium
content (actinolite to aluminous actinolite, Table 2) and an
increase in plagioclase anorthite content may be expected
over the temperature range 450-550~ (QFM buffer data
of Liou and co-workers 42), Epidote plus quartz breaks
down at about 500~ at low fo2 (QFM), whereas epidote
is stable up to 630~ at higher fo2 (H.M. buffer) at 2 kbar
fluid. 42
In several cases we calculate a range of isotopic composi-
tions of waters for a possible range of temperatures. Be-
cause waters more 1SO-depleted than sea water are geo-
logically unrealistic for development of our greenschist
facies samples, a minimum temperature can be derived for
these assemblages. In interpreting the origin of the waters,
three types of water must be considered: (1) primary mag-
matic water, (2) sea water and (3) metamorphic water
released from a subducted lithosphere that is undergoing
dehydration. We have no evidence that meteoric waters
in the subaerial history are important for the samples in this
study, and waters of this origin are not considered further.
For rocks of basaltic composition primary magmatic
waters are defined to have ~D -- - 4 0 to -80%o and ~180 =
+5.5 to +8.00/00.2
The average oxygen isotope composition of present-day
ocean water is about 0%o. Shackleton and Kennett 43 have
estimated that prior to the formation of the currently exist-
ing ice sheets in the Miocene the oxygen isotope composi;
tion was about -1.0%o. We assume that Cretaceous sea
water had ~D = -7%0 and ~ 180 = -1.0%o.
With regard to metamorphic waters, waters released
during dehydration of subducted crust are considered to
have compositions in the range ~D -- --20 to -60%o,
~180 = +4 to +12~ 2,21 It is possible that the ~180 value
might be a few per mil heavier if the size of the sediment
component in the crust that is undergoing dehydration is
very large.
In applying the combined hydrogen and oxygen isotope
approach to determining the origin of water, it is important
to appreciate that the D/H ratio tends to be more diagnostic
of the origin than the 180/160 ratio unless water/rock
ratios are extremely large. Most rocks contain very much
less chemically bound hydrogen relative to the large amount
of hydrogen in the hydrothermal fluids with which they
exchange. The D/H ratio of the water, compared to the
180/160 ratio, is therefore much less affected by water-
47
rock exchange processes, unless water/rock ratios are
extremely small (< 0.1 by weight).
In discussing the isotopic composition of the Troodos
hydrothermal waters we consider first the samples from the
pervasive metamorphic assemblages, followed by the epidote
vein and stockwork assemblages.
Zeolite-greenschist-amphibolite facies assemblages
The calculated isotopic composition of waters in equilibrium
with our chlorites, actinolites and hornblendes from the
greenschist and amphibolite facies assemblages are shown
in Fig. 5 for a range of temperatures. Because of the rapid
rate of exchange of hydrogen in many of our zeolite facies
rocks a field for waters in equilibrium with these assemblages
Iwater in e q u i l i b r i u m 300 or
with quartz J J
' I ' I
"1"20
9~ ~ - -
/ ~
n
co
-2o
-~0
-60
I , I I
-2 0 +2
Fig. 5 Calculated hydrogen and oxygen isotope composi-
tion of waters in equilibrium with Troodos S.I.C. and gabbro
chlorites and amphiboles and quartz at 200-800~ Fields
for ocean water (S.W.), primary magmatic water, meta-
morphic water and meteoric water (M.W.) line are given for
reference. Asterisk represents SMOW
is not given. The primary magmatic water field, representing
our best estimate of the isotopic composition of waters in
equilibrium with basaltic magmas that have not recogniz-
ably interacted with waters from the country rocks, and a
field for metamorphic waters are given for reference.
For greenschist facies temperatures of 250--450~ the
hydrogen isotope data demonstrate unequivocally that sea
water was the dominant source of the water during this
stage of hydrothermal metamorphism (Fig. 5). For a given
temperature the hydrogen and oxygen isotope composition
of the waters calculated from the actinolite data is indis-
tinguishable from that calculated from the chlorites when
the uncertainties in the mineraI-H20 fractionation data are
taken into account. This would suggest that both the actino-
lite and the chlorite have preserved their equilibrium fract-
ionations. Because there is no depletion in deuterium rela-
tive to the probable sea-water range, there is no isotopic
evidence for a significant magmatic and/or metamorphic
water contribution during the major hydration processes
that occurred under greenschist facies conditions. These
data also set a minimum isotopic temperature of about
200~ for these greenschist facies assemblages in the S.I.C.
and gabbros.
Waters in equilibrium with the amphiboles, from near
the base of the S.I.C. and upper parts of the gabbros, have
5D values that are depleted in deuterium relative to sea
48
water if temperatures were higher than 400-500~ Many
of the Troodos hornblendes show evidence for a metamor-
phic origin, which suggests that they formed under amphi-
bolite facies temperatures of perhaps 500-800~ If iso-
topic equilibrium is preserved at temperatures greater than
500~ a metamorphic or magmatic water component is
probably present, the amount of this component increasing
with increase in temperature. In all the hornblende-bearing
samples studied, however, some degree of pervasive retro-
grade alteration of hornblende to actinolite was visible,
ranging from a very small amount to total replacement. It
is possible that, if the hornblende, formed at temperatures
much higher than 500~ or so, they underwent retrograde
isotope exchange during formation of the actinolite. The
apparent similarity between the isotopic compositions of
hornblende and actinolite may result, therefore, from iso-
topic equilibration of the amphibole (hornblende plus
' 4000 500 ~ ' 800 ~ Pegmatite
I J C) q u a r t z actinolite) as a whole with the same sea-water fluids, near
' t . I ' I 9Cmor,'e the prograde amphibolite-greenschist facies boundary,
9 Act i n o l i t e
I--I H o r n b l e n d e under retrograde greenschist facies temperature conditions.
Actinolite-
IR hornblende We note that hornblende is characteristic of the upper parts
- Chl lColcu,ated of the gabbros and the lower S.I.C. rather than the fresh
cumulate-layered lower parts of the gabbros. Although the
stability relations for amphiboles are generally complex -
(e.g. Holloway 44), this occurrence of hornblende at the top
I ~- 7">. Metamorphic
of the gabbros may imply that (1) if it is of igneous origin,
the water content of the magma may have been low, or (2)
J 8 0 0 ~ w a t e r s if it is of metamorphic origin, water most probably entered
s-tc. ~'A""-A--,~,
\.... I.,~ ............ the system from above (sea water) and not from below
(metamorphic or magmatic water) because of the large
Gabbros i(,IIL._..::.-.-r~".magrnat'ic
thickness of amphibole-free, pyroxene-rich, cumulate-layered
L __ _ _ ! _ w a t e r
l i [ : I i = rocks.
+/. +6 +8
8180, %0 Data for the rock units that underlie the zeolite facies
pillow lavas demonstrate that the metamorphic fluids were
overwhelmingly of sea-water origin. Large volumes of sea
water must be expected during zeolite facies alteration of
the pillow lavas. Insofar as the D/H ratios of the zeolite
facies pillow lava samples are diagnostic of their metamor-
phic waters, the data also indicate that sea waters were
dominant during zeolite facies metamorphism. The 8180
values of the pillow lavas are, on average, more leO-rich
than the greenschist facies assemblage rocks of the SoI.C.
The presence of relict phenocrysts (e.g. anorthite-rich
plagioclase and pyroxene) in the pillow lavas indicates that
the whole-rock samples have not totally equilibrated during
metamorphism. If, however, approximate estimates are
made of the percentage of alteration minerals in the whole-
rock samples (from mod~il mineralogy and H20 + content),
the whole-rock 5180 data are found to be consistent with
alteration in the presence of sea water at zeolite facies tem-
peratures (0-250~ The 5180 values of the Troodos
pillow lavas do not vary systematically with H2 O+ content
- in contrast to the patterns observed for oceanic basalts,
'weathered' at low temperatures on the sea-floor. 3s, 36
Isotopic fractionations between carbon-bearing species
in solution and in minerals are largely a function of tempera-
ture, pH and fo2"45 Assuming HCO~ to be the dominant
carbon-bearing species in the pillow lava calcite-forming
fluids, the pillow lava calcite 513C data indicate a sea-water
source for the fluid; and the 5180 data suggest calcite form-
ation temperatures of 10-15~ Since sea-water fluids have
been demonstrated for the S.I.C., the upper S.I.C. calcite
isotopic data imply relatively high fo2 values for near-
neutral solutions, with temperatures of about 100~ (for
sea water 5180 = -1%o). The 5180 data for the L.P.L.
vesicle and chalcedonic vein quartz (+31.7 and +30.9%o)
suggest, if they formed from sea water with 5180 = -1%o,
temperatures of about 20~ (cf. Knauth and Epstein 46).
Epidote veins from gabbros, trondhjemites and S.I.C. Stockworks
Four epidotes from veins in the upper parts of the S.I.C., The hydrogen and oxygen isotope compositions of'whole-
trondhjemites and gabbros are shown in Fig. 6, together rock' stockwork samples are shown in Fig. 7 (~Vhole rock'
with quartz, which coexists with one of the vein epidotes, refers to the altered basalt that has been separated from the
and actinolite'and quartz, which coexist with another vein vein material that is so characteristic of stockwork samples).
epidote. Prior to analysis, the disseminated pyrite was physically
separated from the quartz and chlorite, which are dominant
I I Waterl in equiliblr ium {
minerals in the 'whole-rock' samples.
with quartz The 'whole-rock' samples in Fig. 7 are from parts of the
300 i 600~ I 5000 I ~ JD Vein quartz
~ ~O L.P.L. stockworks of Mousoulos and Mathiati, inferred to
, t , I.--~'"..-t"~ ~~ I be approximately 25-20 m below the base of the massive
+2(
ore, and from the Limni stockwork (depth below massive
Veins .~" ~ 9 Actinolite _
ore not known) and a stockwork in the upper S.I.C. (with
_ no overlying massive ore). The whole-rock ~D values are
, .- ~,
oI.o,oo,.1
~ Veto
uniform, reflecting the chlorite ~iD values, whereas the
+.Y_ /---""~",~"~'~~ ~ | ', ~ caveo ~ t ~ ....
/ SW" , /IgJ , ~ , amp ~> H20
~7/~./% ,,/ O~'~00~ O Epidote l wide range in ~ 180 reflects, at least in part, variations in
f ~'~//~/t~./.~r//r 500 ~ ~t f-lArnphibolite
, J the quartz and chlorite contents of the different samples.
d \ (~ Exchanged 5.W
t W/R -0.1 (400~ To apply oxygen isotope fractionations to the data the
proportions of quartz and chlorite in each sample were cal-
i
" ri 1I culated by comparing the AI content of the chlorites
I~I In Primary
(Table 2) to that of the whole-rock samples (analysis by
I mogmatic J
I M. J. Saunders). Both the quartz from the groundmass of
t wat~ i
[
_I2 ' I , 1 , I , I I , I , these samples (together with a Kokkinoyia stockwork
0 § +/~ +6 "+8
6'80, %0
Fig. 6 Calculated hydrogen and oxygen isotope composi- I , I i
o
I
~ ~ ~,
tions of waters in equilibrium with epidote, actinolite and Ou~hrtz in equihbrium
wi sea water
quartz from veins in S.I.C., trondhjemites and gabbros, and ~'1 I I ! /
500" 350* 250~ 200~ 180~
with non-vein associated gabbro epidote and amphibole at , I , f , t , I , I , I , I , I , I
400-500~ Dashed lines connect coexisting minerals. I ~ ~ Water/rock ratio (350~
Stockwork .samples
S.W., M.W. and primary magmatic water as for F ig. 5 9 Whole rock unvelned
__~
.~ 350~ Whole rock talc H20
t Qtz groundmoss
5000 E) Qtz vein
9 Exchanged sea water
The isotopic composition of the waters are calculated
for 400 and 500~ These results indicate that 320-400~ -~0
is a minimum isotopic temperature of formation of vein I Primary I
Imog ..... :
epidote. Fluid inclusion homogenization temperatures on I water I
-60 J _ 2I i I i I i I ~ I I
the vein-halo assemblage in a trondhjemite sample (11-1 ) 0 +2 +L. +6 +B +12
are in the range 340--400~ (unpublished data with B. S'"O,~
Poty) and support the isotopic temperatures as minimum
Fig. 7 Calculated hydrogen and oxygen isotope composi-
values. For the coexisting quartz-epidote and quartz-
tions of waters in equilibrium with Troodos stockwork
actinolite-epidote the hydrogen (epidote and actinolite
'whole-rock' samples ('groundmass' quartz and chlorite),
only) and oxygen isotope compositions of the waters are
'groundmass' quartz and vein quartz at 180-500~ S.W.,
more nearly identical at 350--400~ than at 500~ Al-
M.W. and primary magmatic water as for Fig. 5, asterisk
though the uncertainties in the mineraI-H20 fractionations
represents SMOW
for epidote and actinolite preclude precise geothermometry,
the 350-400~ formation temperature is favoured. At
these temperatures sea water is overwhelmingly the domin-
ant component of the hydrothermal fluids. groundmass quartz) and the whole-rock data set minimum
Fig. 6 also gives the isotopic composition of waters isotopic temperatures of about 300~ for their formation
associated with epidote and amphibole, disseminated in a (Fig. 7). Thus, temperatures in the stockwork zones were
gabbro sample, which are not obviously associated with higher than those in the enclosing predominantly zeolite
any vein. At 400 and 500~ the hydrogen isotope compo- facies lavas. Fluid inclusion homogenization temperatures
sitions of the waters are identical to the vein epidote fluids for a 485-m depth borehole sample from Skouriotissa in
and sea water, but they are enriched in 180 by 3-5%o rela- mineralized basalts range 340-400~ (unpublished data
tive to sea water. This indicates, quite plausibly, that sea with B. Poty), and are consistent with our minimum isoto-
water/rock ratios were lower in the gabbros than in the pic temperatures.of 300~ based on higher-level stockwork
Veins (Fig. 6). We also note that the implied isotopic samples. At temperatures between 300 and approximately
compositions of the waters in this epidote-bearing gabbro 400~ the hydrogen isotope data indicate a dominantly
sample are identical to those derived from the greenschist sea-water source for the fluid. Like the fluids related to the
facies assemblages at similar temperatures in Fig. 5. epidote veins, the fluids associated with the whole-rock part
If formation temperatures were possibly as high as of the stockworks have 180/160 ratios essentially identical
500~ close to the upper stability limit of epidote, 25 to sea-water values. This indicates in both situations that
either a magmatic or metamorphic water component must the oxygen isotope compositions of the minerals are being
also be present in the epidote-forming fluids. The data as a largely controlled by the sea-water reservoir and their tem-
whole are, however, more consistent with lower tempera- perature of formation.
tures (~'400~ Vein quartz from the upper parts of the Mousoulos and

49
Mathiati stockworks (about 5 to 10 m below the massive workers 48 showed that the water/rock ratio (W/R) is given
ore), from the Limni stockwork and from the upper S.I.C. by
stockwork (from approximately 100 m above the S.I.C.
stockwork whole-rock sample) has 6180 = +10.5 to +14.1%o
2W 6 r i - 6rf
-- values significantly higher than the groundmass quartz
(Fig. 7). If we assume that the vein quartz also formed from R 6wf-6wi
essentially isotopically unmodified sea waters, formation
temperatures of about 180-220~ are indicated. where W/R is water/rock ratio by weight, the factor of 2
There are several possible explanations for the tempera- assumes that water contains typically twice as much
ture differences between vein and groundmass quartz. It exchangeable oxygen as rock by weight, 6ri is the initial
could be a sampling problem - groundmass quartz was isotopic composition of the rock (taken as +6%o for
from deeper stockwork samples. More substantial retrograde Troodos rocks), 6wi is the initial isotopic composition of
isotopic exchange of vein quartz relative to groundmass water (--1%o for sea water), 6rf and 6wf are final isotopic
quartz is possible, although quartz is notoriously one of the composition of the rock and water, respectively, and are
most resistant minerals to exchange once formed. 47 Alter- related by 6rf - 6wf--- ~rw, which is the r o c k - H 2 0
natively, and considered more likely, vein quartz formed fractionation (see below).
later than much of the groundmass quartz at lower tempera- Therefore, for sea-water circulation in the Troodos crust
tures. Textural evidence, such as vein crosscutting relations we have
and quartz growth zoning, suggests that there are several
generations of vein quartz; chalcedonic quartz is common in W +6-6wf-Arw
some stockwork veins. Constantinou and Govett 16 sug-
gested that vein quartz commonly formed later than, and in R 2 6wf+ 2
some cases replaced, earlier vein sulphides (pyrite with chal-
copyrite and sphalerite as inclusions and interstitially). It is emphasized that water/rock ratios calculated in this
The stockwork veins contain a considerable amount of way give minimum values. The actual amount is critically
silica, and pyrite in the veins contains chalcopyrite and related to the previous exchange history of the water prior
sphalerite; in contrast, for the altered basalt there is no to its 'seeing' the rock of interest. 8,31,48
evidence for a significant increase in silica, and disseminated The D/H ratios for all chlorites, amphiboles and epidotes
chalcopyrite and sphalerite are absent. This, together with in Troodos rocks have quite restricted ranges of values. In
the very sharp contacts between the veins and the altered contrast, however, 1 8 0 / 1 6 0 ratios for these minerals have
basalt, suggests that the physico-chemical conditions during wider ranges of values. In particular, chlorites from the
formation of the altered basalt assemblages (essentially S.I.C. (non-vein) rocks have 6180 values on average 3%o
quartz-chlorite-pyrite) may have been different from those higher than stockwork chlorites, and epidote and amphiboles
during precipitation of the vein minerals. from non-vein assemblages have 6180 values up to 3-4%o
These data suggest that the alteration of the basalts oc- higher than those from vein or vein-halo assemblages. Such
curred when sulphur-bearing fluids of sea-water origin at a pattern could result from either low water/rock ratios for
temperatures > 300~ passed through the stockwork the more 1SO-enriched assemblages or possibly, though
fractures (temperatures in the fractures must be equal to, geologically less plausible, lower equilibration temperatures
or higher than, temperatures in the rock). Insofar as present for the 180-rich minerals in the bulk of the rock relative to
sampling is adequate, the data suggest that vein quartz did the 1SO-poor vein and vein--halo minerals. For example,
not form from these high-temperature fluids. Later, possibly a temperature difference of about 100~ would be implied
during waning stages, vein quartz was precipitated when the for rock epidote versus vein epidote. Although there are no
fluids became saturated in SiO 2 at lower temperatures
(180-220~ The pyrite-chalcopyrite-sphalerite vein
assemblages, which are texturally similar to the primary Water I r or k rat io (weight)
co 5 _...i ...,., _ _ - o,1
sulphides in the massive ore, 16 formed at some time before
the vein quartz; if they formed from the high-temperature ~.. ~l . ~ / "//~f/, / i/" :-7
(> 300~ fluids, conditions were such that chalcopyrite ~oc ~/ ,~\\%~ ~ _~ /
and sphalerite only formed on the fracture walls and not in
the rock between the fractures.
S.I,C, and
I~" " / , ~ " ~ . ~ /" gabbro rock H20
Water~rock ratios I//_ / / . . . . . .. Y
~2OOl r ~ " / ..... / i
Figs. 5, 6 and 7 show that some of our calculated water I
6180 values are significantly different from the unmodified
PlagiodQse--
sea-water value. These '180 shifts' could be caused by a I I
II Con I
Exchanged sea water
LAmphibotite
number of processes, but if we take into account the lack
of variability in the calculated D/H ratios of the waters,
0I
-2
~1
0
' I
+2
I I I
+4
=
+6
I

variation in the mineral-water isotopic exchange reaction
is the most probable dominant process. The isotopic com-
position of a mineral forming in a hydrothermal system is
related to the temperature of crystallization and to the
isotopic composition of the fluid. The oxygen isotope
composition of the fluid is particularly sensitive to its pre-
vious exchange history, which is conveniently expressed in
terms of water/rock ratios - the amount of oxygen in the
water relative to the amount of exchangeable oxygen in the
rock. By use of a closed-system model Sheppard and co-
6'80,%o,Water
Fig. 8 Calculated oxygen isotope compositions of waters
from S.IIC. and gabbro rocks, vein and vein-halo assem-
blages, and stockwork 'whole rocks', compared with calcu-
lated oxygen isotope composition of sea water (initial
6180 = -1.0%o) that has undergone isotopic exchange with
basalts (initial 6180 = +6.0%o) at temperatures of 2 0 0 -
500~ and water/rock ratios of 0.1 to infinity by weight,
with plagioclase and amphibole for the exchange model.
See text for discussion

50

geothermometric data to rule this out, variations in water/
rock ratios are considered more probable.
The ~ 180 values of sea-water fluids (initial ~ 180 = --1%o)
that have undergone isotopic exchange with oceanic basalt
(initial ~180 = +6%0) are shown in Fig. 8 as a function of
water/rock ratio (by weight) and exchange temperature.
The waters are calculated by use of the equation given above.
If we assume total r o c k - H 2 0 oxygen isotope equilibration,
the modal mineralogy - for example, of S.I.C. rocks -
suggests that r o c k - H 2 0 fractionations are approximately
equivalent to plagioclase (An6o),H20, For many samples,
however, particularly the gabbros, the total rock may not
have equilibrated during metamorphism; amphibole is the
dominant metamorphic mineral and r o c k - H 2 0 oxygen iso-
tope fractions may approximate to amphibole-H20. In
constructing Fig. 8 both plagioclase (An6o)-H20 and
amphibole-H20 fractionations are used; the majority of
Troodos r o c k - H 2 0 fractionations may be expected to lie in
this range. The calculated oxygen isotope compositions of
the Troodos hydrothermal fluids for the different assem-
blages as a function of temperature are also shown in Fig. 8
(from Figs. 5, 6 and 7). For the S.I.C. and gabbro rocks
the results given in Fig. 8 indicate that minimum water/rock
ratios were about 0.1 (by weight) 9 In contrast, stockwork
alteration and S.I.C. and gabbro vein or vein-halo forma-
tion occurred in the presence of relatively isotopically un-
modified sea-water fluids, which were part of a system with
high water/rock ratios (minimum water/rock ratios between
"~1 and 5, by weight). For the development of the high
pillow lava whole-rock ~ 180 values at temperatures less
than' 200~ water/rock ratios must have been > 1 (by
weight).
From reasonable estimates for the geothermal gradient +
and crack width and spacing for rocks in Troodos order of
magnitude calculations may be made for the fluid dynamic
parameters that control sea-water circulation in the Troodos
crust 21 (see methods of Lister 49). For those rocks in
which bulk permeability results solely from cooling contract-
ion (i.e. non-veined, non-tectonically fractured parts of the
S.I.C. and gabbros), the convection parameter exceeds the
minimum critical value by at least one or two orders of
magnitude and free convection of fluid may be expected 9
In the heavily fractured (veined) parts of the S.I.C. and
gabbros, in the stockworks (heavily fractured) and in the
pillow lavas (large amount of inter-pillow cavity space),
however, the calculated permeabilities and permitted fluid
flow rates are several orders of magnitude higher than in
those rocks in which permeability results solely from cool-
ing contraction. The pattern for water/rock ratios suggested
by the isotopic data (Fig. 8) is therefore consistent with a
model in which large volumes of sea water passed through
the pillow lavas, along veins in the S.I.C. and gabbros, and
through the stockwork.zones, whereas smaller amounts of
fluid penetrated the rocks away from veins and fractures.
Sulphur isotopes
34S//32Sratios in sulph ides
The sulphur isotope compositions for Troodos orebody
pyrites reported by a number of workers 50-53 are sum-
marized here, and limits on the source of the sulphur are
placed in light of the chemical, mineralogical and other
isotopic data. The sulphur isotopic data are not considered
in detail because, among other things, there are possibly
some small inter-laboratory differences among data on
samples that are probably equivalent. The 34S/32S ratios of
pyrites have been reported from Troodos stockwork dissem-
inations (+4.1 to +7.3%o; 9 analyses; average +5.1%o), veins
0 0
(+2.0 to +6.5/0o; 12 analyses; average +4.8 ~/oo),the massive
sulphide ores (+2.8 to +7.5%o; 22 analyses; average +4.3%o)
and from the 'ochre' sediments that overlie some of the
massive sulphide deposits (+1.9 to +5.0%0; two samples;
average +3.4%o).5~ Although there are slight variations
between individual deposits, at a single mine the sulphur
isotopic compositions of pyrite from the different para-
genetic types - stockwork disseminations, veins, massive
sulphide ore - - are quite similar. Detailed arguments have
not been presented, but the source of the sulphur has been
interpreted as magmatic 51, 53 and Cretaceous sea water. 52
To interpret the sulphur isotope ratios, however, knowledge
of the temperature, pH and fo2 is required 45, 54 before
limits can be placed on the nature of the dominant sulphur
species in solution and its isotopic composition; In this dis-
cussion it is assumed that pyrite precipitated in isotopic
equilibrium with the hydrothermal fluid (although this is
not always so 55).
Physico-chemical conditions
The sulphur isotope chemistry of the stockworks is empha-
sized here because, unlike much of the massive ore, primary
assemblages are preserved. On the basis of the above hydro-
gen and oxygen isotope data, temperatures of about 250~
(intermediate between the suggested basalt alteration and
the vein quartz formation temperatures) are considered. The
use of higher (350~ or lower (150~ temperatures would
make little difference to the general conclusions.
The chlorite alteration assemblage associated with the
stockwork sulphides suggests, by analogy with modern geo-
thermal systems, pH values within 1 or 2 units of neutrality
at 200-300~ 56-59 This is consistent with sea water-
basalt hydrothermal experiments. 6~ In the system
F e - C u - S - O (for the stockworks) pyrite and chalcopyrite
were stable, magnetite was unstable and pyrrhotite and
bornite did not form as primary minerals. 16 The fluid fo2
was therefore within the pyrite + chalcopyrite stability
field (see Fig. 12 55). More precise limits on fo2 do not
appear possible at present. Sulphide-free jasper veins and
lenses have been observed in a number of Troodos stockwork
samples - the majority are probably late-stage. Their pres-
ence suggests that fo2 was relatively high during the late-
stage history of the stockworks. Hematite has been observed
as a daughter mineral in stockwork fluid inclusions (un-
published data with B. Poty); if leakage of hydrogen is not
important, this implies that fo2 values were relatively high.
Magmatic source for sulphur
Pyrite with (~34S = +2.0 to +7.3%o will only form from
fluids with total sulphur (~34S = 0 % when the pH is higher
and the f02 lower than that suggested by the mineral
stability data (see Fig. 5 45). Therefore, magmatic sulphur,
including sulphur leached from igneous rocks ((~34S ~"
+1~ 62), cannot be the only source of sulphur in the
stockwork flu ids.
Cretaceous sea-water source for sulphur
Upper Cretaceous sea-water sulphate has a sulphur isotope
composition in the range +15 to -I-20~ 63,64,65 Sea-water
sulphate could have been the dominant source of sulphur
in the stockwork pyrites (+2.0 to +7.3%o) if the stockwork
fluid fo2 were relatively high and the major sulphur species
in the fluid were both SO~ and H2S. The narrow range for
9 0
stockwork pyr=te (~34S values ( ' 4 ~o) would also require
that the S04/H2S ratio remained relatively constant during
pyrite precipitation (see Figs. 3 and 5 45); this is quite
51
plausible in geothermal systems, where water/rock ratios.
were large and the buffering capacity of the fluid is
considerable.
The stability of pyrite in such a fluid would require the
sulphur concentration in the fluid to be t> 10 "3 mole/kg
H20 (250~ Fig. 5 48). Therefore, assuming that Cre~
taceous sea water contained a sulphur concentration similar
to that of present-day sea water (~'0,03 mole/kg H20), the
sea-water-hydrothermal fluid could lose up to 95% of its
initial sulphur during its passage through the Troodos crust.
Because the solubility of anhydrite decreases with increasing
temperature 66 we would expect a large part of the initial
sulphur to be precipitated as the sea-water-hydrothermal
solutions became heated up during their descent through
the crust. 59,60 Although, to the best of our knowledge, an-
hydrite has not been reported from the metabasites of
Cyprus, this does not mean that it was not precipitated. We
note that the recognition of the importance of anhydrite in
the porphyry copper environments is relatively recent. 67
Mixed igneous-sea water source for sulphur
The observed sulphur isotope ratios could also be obtained
from a mixed igneous (magmatic and/or leaching of basaltic
rocks)-sea water source for the sulphur. The lower the
value for f02 during sulphide deposition, the greater is the
inferred igneous sulphur contribution. Thus, more precise
limits on f02 are critically required to assess the importance
of the igneous contribution to the sea-water sulphur source.
For a mixed sea water-igneous sulphur model the data re-
quire that the sea water/igneous sulphur ratios be very simi-
lar from one deposit to another. As the igneous sulphur
contribution increases, the isotopic composition of pyrite
becomes less sensitive to variations in pH and/or f02"
Discussion of the pyrite sulphur isotope data for the
massive sulphide ores is complicated by their having under-
gone textural and mineralogical changes that suggest
secondary submarine supergene alteration, ls, lS Since, how-
ever, the stockwork zones represent the channelways for
the ascending hydrothermal fluids (see below), the source of
sulphur in the massive sulphide ores may be expected to be
largely the same as the sulphur source for pyrite in the
stockworks.
Chemical changes
In comparison with analyses of fresh oceanic basalts,
basalts and dolerites from Troodos have a very wide range
of chemical compositions that suggest large-scale mobiliza-
tion and redistribution of many of the elements. Calculation
of the major chemical gains and losses that occurred during
alteration requires knowledge of the initial composition of
the fresh Troodos basaltic rocks prior to alteration. This is
not known precisely. Nevertheless, an estimate of whether
a given element was gained or lost duringthe hydrothermal
metasomatism can be made by comparing the averaged
chemical compositions of the L.P.L. and S.I.C. - 44
analyses 12,14, 68-71_ with the chemical composition of
fresh oceanic basalt. 72, 73 T h e results, with use of a con-
stant-volume model, are given in Table 3. Oxidation .of these
rocks is also a major process. Taking as a reference a
Troodos glassy basalt analysis 68. 74 gives a similar pattern,
except for Na in the L.P.L. It is emphasized that the gains
and losses in Table 3 are averages for a large number of
samples; for any particular sample the changes may be differ-
ent from those shown.
The data for the zeolite--greenschist facies rocks from
Troodos are compared in Table 3 with similarly calculated
gains and losses for altered oceanic rocks collected by
52
dredging. The alteration of the dredge samples may be
multistage insofar as low-temperature alteration on the
Table 3 Gains and losses of some principal chemical
constituents during hydrothermal alteration of oceanic
crust
Troodos* Oceanict
Zeol. G reen. Zeol. G reen.
Si - 0 -?
AI -- -- --?
Fe -- --
Mg -- -- +
Ca . . . .
Na + + + +
K + + --? --
H20 + + + +
+ = addition to rock; -- = loss from rock.
*lsovolumetric model. Zeol. = zeolite facies (L.P.L. samples,
density ~,2.4 g/cm3); Green. -- greenschist facies (S.I.C. samples,
density *-2.7 g/cm3) (seetext).
tSummary of zeolite and greenschist facies oceanic alteration
data.24,75, "/6,85,86
sea-floor (halmyrolysis) may have been superimposed on
some of the samples. For a given element the gains and
losses are qualitatively similar between Troodos and the
oceanic rocks, except for Mg and possibly K. These varia-
tions could reflect differences in the alteration processes or,
possibly, be a result of the initially fresh Troodos basaltic
rocks having higher K and lower Mg contents than fresh
oceanic basalts; basalts from island arc environments have
higher K and lower Mg contents.
The gains and losses summarized in Table 3 emphasize
not only the general similarities in the alteration processes
that affect ophiolite complexes like Troodos and current
oceanic crust but also the large-scale nature and widespread
occurrence of sea-water-hydrothermal metasomatism.
General hydrothermal model
The following model is proposed for the development and
evolution of the metamorphism and mineralization of the
Troodos ophiolite complex. The model is based on com-
bining the available geological and mineralogical information
with our interpretation of the isotopic data, in terms of the
origins of the fluids and the limits that can be set on the
temperatures of development of the different mineral
assemblages and water/rock ratios.
The Troodos igneous complex developed at an active
spreading centre in a sub-oceanic environment. Subsequent
to the emplacement of at least the L.P.L. and Sheeted
Intrusive Complex, large-scale convective circulation of sea
water was established through the 3 - 5 km of these upper
oceanic layers (Fig. 9). Permeabilities were generally high
and locally very high, in part because the development of
normal faulting during the spreading processes generated
fracture permeability, and in part because of the large
amount of inter-pillow cavity space in the pillow lavas.
During the sea-water-hydrothermal alteration processes the
L.P.L. and S.I.C. mineral assemblages were metamorphosed
to zeolite and greenschist facies assemblages.
By analogy with active geothermal systems 77 downward
migration of sea water into the hot oceanic basalts occurred
over a large area, and upward discharge was localized. The
initiation of the influx of sea water deep into the oceanic
crust must be critically related to the initiation of fracture
permeability. The massive cupriferous pyrite deposits are
interpreted to have formed in tectonically controlled basins
at the pillow lava--sea water interface, their underlying
stockwork zones of altered and fractured pillow basalts re-
presenting the localized upward discharge zones of the sea-
water-hydrothermal solutions within the pillow lava
sequence.
Ocean 6D- - 7%o
~ - ~ I
0--250 ~ S2t~0Cok
3;r0k
(W/R>1) I;J~'~' '
/ \ \ ,w,R01
).,,,JJ'I
~ ~ ~-300 ~ - - -- -- .-- ~ / ~ "
\ I
>/'50~
Fig. 9 Diagrammatic model for sea-water circulation
through Troodos crust leading to metamorphism (F =
facies) and development of sulphide orebodies. W/R =
water/rock ratios, by weight, for pervasive alteration; ratios
are locally much higher in vein-fracture zones. 300~ iso-
therm illustrates thermal anomalies associated with localized
discharge of sea-water convective cell. See text for
discussion
During the downward flow of the brines (Fig. 9) they
became progressively hotter. The pervasive zeolite meta-
morphism of the L.P.L. occurred at approximately 0-250~
and the greenschist facies metamorphism, dominantly of
the S.I.C., at temperatures of about 250~ to 450-550~
Isotopic and preliminary fluid inclusion thermometry indi-
cate that temperatures were higher than 300~ in the stock-
work zones, and were therefore significantly higher than
temperatures in the surrounding pervasively metamorphosed
(dominantly zeolite facies) country rocks. These localized
rising thermal plumes support our evidence that the stock-
works represent the channelways for the upward discharg-
ing hydrothermal solutions (Fig. 9). Average regional geo-
thermal gradients in the L.P.L. and S.I.C. must have been
about 100-150~ whereas thermal gradients in the
stockwork zones must have been two to five times, or more,
higher and increased as the sea water-pillow lava interface
was approached. During regional metamorphism water/
rock ratios are high in the pillow lavas (> 1, by weight),
whereas the minimum or integrated water/rock ratios for
the underlying greenschist facies rocks are relatively low
(~),1, by weight). The vein-vein halo alteration and
stockwork assemblages indicate that water/rock ratios were
very large along fractures (> 1--5, by weight). Therefore,
the circulation of most of the sea water through the crust
must be fracture controlled.
During the initiation of the sea-water-hydrothermal con-
vection process in the newly created upper layers of the
oceanic crust it is probable that some of the underlying
gabbro plutons were still undergoing magmatic differentia-
tion and development of their cumulate layering. The
upper parts of the magma chambers could be modified by
interaction with their roof rocks, particularly because these
were undergoing major hydration and chemical modification
through reaction with the circulating brine solutions. Melt-
ing or assimilation and exchange processes could occur here,
any one of which could both chemically and isotopically
modify the upper parts of the plutons. Coleman and
Peterman 23 and H. P. Taylor Jr. (personal communication,
1976) have emphasized that the trondhjemites or 'oceanic
plagiogranites' from ophiolite complexes are enriched in
sodium and depleted in potassium relative to what are con-
sidered to be equivalent rocks from spreading centres that
developed in a non-sub-oceanic environment with fluids of
/Massive meteoric origin, such as Iceland and Jabal Turf, Saudi
/_sulphide Arabia. The chemical compositions of the hydrothermal
l" deposi_t._t
solutions - sea water versus meteoric waters - that have
Pillow t. . . . modified the roof rocks could be playing, among other
(Zeolite F) things, a critical role in the genesis of these soda-trondhjem-
ites relative to the potassic granites and rhyolites. We note
that the granophyric textures of the trondhjemites suggest
that these magmas may have become saturated in volatiles
~- S.l.C during their crystallization. Such a process could generate
hydraulic fracturing in the overlying S.I.C. and L.P.L. At a
later stage the very upper layers of the gabbro cumulates
and trondhjemites support fractures that develop epidote-
Trondhjemites bearing veins and sea-water-hydrothermal solutions pene-
Gobbros
trated directly to these depths. Isotopic and fluid inclusion
thermometry indicates that vein temperatures were 3 2 0 -
400~ or higher and that the solutions were brines. Meta-
morphic assemblages that developed regionally are upper
greensch ist to amph ibolite facies.
If primary magmatic waters were present during the
crystallization of the gabbros and the development of the
amphibolite facies metamorphism, their presence is over-
whelmed by the influx of the sea-water convective system.
It is possible, however, that water, initially of sea-water
origin, was introduced into the magmas during the magma--
roof rock processes by assimilation, exchange with or melt-
ing of hydrous roof rocks with typically ~D = - 4 0 to
-50%o. Such water, although ultimately of sea-water
origin, is magmatic in thesense that it has equilibrated with
the magma system, and might be termed sea-water-mag-
matic water to distinguish it from primary magmatic water.
We have noted for Troodos rocks that if isotopic equilibra-
tion temperatures were higher than 500~ (e.g. during
hornblende formation), a more deuterium-depleted com-
ponent than sea water has to be inferred; this could be
primary magmatic water, metamorphic water or sea-water-
magmatic water.
During the sea-water-hydrothermal alteration of the
Troodos rocks major changes occurred in their chemical
and isotopic characteristics. There are:marked changes in
the major-element compositions of the rocks. Also, the
87Sr/86Sr ratios of many of the L.P.L. and S.I.C. rocks are
higher than those for fresh unaltered oceanic rocks. 23, 78, 79
As was emphasized by Deffeyes, s~ Spooner and Fyfe 81 and
others, not only is the oceanic crust being chemically modi-
fied but also the chemistry of the discharging hydrothermal
solutions must be fundamentally different from their
initial sea-water composition. Changes in fluid chemistry
have been demonstrated in experimental sea water--basalt
hydrothermal interaction studies by Bischoff and Dickson 6o
and Hajash;61 there is a preferential partition of copper into
solution. Because of the scale of these sea-water-hydro-
thermal systems, it is probable that they control certain as-
pects of both the chemical and isotopic compositions of the
ocean.21,76 For example, a column of oceanic crust only
2.5 km thick contains the same amount of oxygen as does
a column of ocean water with an equal area of average
depth 3.8 km.
The development of massive cupriferous pyrite deposits
and their underlying stockworks are another facet of the
sea-water-hydrothermal system and scale of mass transfer.
53
 The isotopic compositions of the stockwork fluids are only
slightly different from normal sea-water values and demon-
strate that very large volumes of heated sea water passed
through these channelways. Although these fluids are
dominantly of a single origin, they may have had different
sub-crustal histories. Because temperatures in the stockworks,
which were within a few tens of metres of the ocean floor,
were initially at least 300~ and therefore substantially
higher than those in the surrounding country rocks, the
great mass of these fluids must have had a deep crustal his-
tory. Similar temperatures on a regional scale are observed
at depths of 3 or 4 km. Therefore, the larger part of the
fluid passing through the stockwork zones must have had a
history within the S.I.C., and became strongly chemically
modified. Although the sea water is a source of at least
some of the orebody sulphur, the importance of an igneous
sulphur contribution cannot be accessed at present; we re-
quire greater knowledge of fo2 values, and more informa-
tion on the changes in rock sulphur content during sub-
oceanic metamorphism.
The localized areas of stockwork formation in the u p p e r
parts of the L.P.L. represent zones where ascending, hot,
chemically modified sea-water fluids passed through an en-
vironment containing large volumes of cool relatively un-
modified sea water (zeolite facies pillow lavas); these zones
are therefore characterized by sharp temperature and chemi-
cal gradients. The factors responsible for the specific loca-
tion of discharge zones are not yet known. Underlying
thermal anomalies (e.g. a pluton intruded later than, or to a
higher level than, surrounding plutons), specific fracture
zone patterns resulting from oceanic ridge tectonics, or geo-
metrical constraints imposed by the preferred dimensions of
sea-water convection cells may be responsible; Lister 82 has
suggested that upward movement of sea water in the oceanic
crust may be favoured below topographic 'highs' on the
sea-floor.
The shape of the massive sulphide orebodies suggests
that they formed where the warm fluids discharged into
depressions on the sea-floor above the stockwork zones. If
the fluids discharged at > 300~ with salinities in the range
0-25%, water depths had to be greater than 1100 to 700 m,
respectively, for boiling not to have occurred (or 1 7 0 -
110 m for 200~ fluids). Although 300~ is considered to
be a good estimate of stockwork temperatures, the thermal
gradients in this environment will be high and the formation
of the massive sulphides is likely to have occurred at con-
siderably less than 300~ The ascending hydrothermal
fluids in the stockworks must have been less dense than the
low-temperature fluids in the surrounding country rocks,
but, to form stable brine pools on the sea-floor, must have
become more dense than cold sea water; this density in-
crease, unless it is the result of an increase in the hydro-
thermal fluid salinity, is assumed to result from a drop in
fluid temperature. These large thermal gradients probably
play a critical role in the precipitation of the ore metals.
During spreading of the oceanic crust away from the
axis major cooling of the crust has occurred through con-
vective heat transfer. This is accompanied by deposition of
minerals in the fractures and cavities. If the intensity of
fracturing also decreases, permeabilities are reduced until
convective flow finally ceases. Evidence for this waning
stage may be indicated by the stockwork vein quartz data,
which suggest temperatures of 180-220~ subsequent to
the higher-temperature stockwork basalt alteration stage,
and also possibly by the upper S.I.C. calcite ("100~ and
pillow lava calcite (~15~ and quartz (~"20~ data.
Subsequent to the metamorphism and mineralization of
the L.P.L. and S.I.C., the upper pillow lavas were formed.
54
The U.P.L. were also hydrothermally metamorphosed up to
zeolite facies grade (Gass and Smewing 17 suggested tem-
peratures of about 0-100~ Although hydrothermal
channelways within the U.P.L. have been observed, 83 major
deposition of sulphides is absent. Instead, F e - M n hydroxy-
oxide-rich 'umber' sediments were formed as chemical pre-
cipitates related to the discharging thermal sea-water springs.
Thus, although both sulphide deposition and umber forma-
tion are related to sea-water-hydrothermal convection sys-
tems in hot oceanic crust, the end-products are quite differ-
ent. The combination of deep penetration of much of the
sea water into the S.I.C. and below, the attainment of re-
latively high temperatures (> 300~ and the discharge of
the metal-bearing fluids into tectonically controlled basins
where redox conditions were lower than those in open sea
water may be critical factors necessary for the generation of
massive sulphide deposits. In the U.P.L. the fluids did not
attain such high temperatures and were probably less
chemically modified from their initial sea-water composition.
Sulphide orebodies are characteristic of a number of ophio-
lite complexes and, although the Red Sea is, to our know-
ledge, the only oceanic environment known to possess
'economic sulphide orebodies', further detailed oceanic
sampling may demonstrate sulphide mineralization to be
characteristic of the oceanic crust; minor mineral ization has
been reported for rocks associated with the Iceland spread-
ing centre. 84 The broad similarity between the chemistry,
metamorphic mineralogy and isotopic properties of Troodos
and other ophiolite and oceanic dredge and drill-core samples
suggests that the processes of sea-water-hydrothermal circu-
lation described above are major and very widespread.
Differences in details of metamorphism and mineralization at
different localities may simply reflect variations on a theme.
Acknowledgment
We gratefully acknowledge discussion with and assistance
in sample collection by Y. Hadjistavrinou, G. Constantinou
and other members of the Cyprus Geological Survey; the
management and staff of the Cyprus Mines Corporation,
the Hellenic Mining Company and Limni Mines; C. M.
Graham and J. D., Smewing. Additional samples were kindly
donated by J. R. Cann, G. Constantinou, A. H. F. Robertson,
and C. R. Allen. K. R. Gill gave advice on petrologic
analysis and mineral separation. We are particularly grate-
ful to J. Borthwick for his assistance in the laboratory.
T. H. E. Heaton gratefully acknowledges receipt of a
Natural Environment Research Council Studentship grant.
The Isotope Geology Unit of the Universities of Scotland
is supported by a N.E.R.C. research grant.
Referez,ces
1. Gass I. G. Origin and emplacement of ophiolites. In
Volcanic processes in ore genesis (London: Institution of
Mining and Metallurgy and Geological Society, 1977), 72-6.
2. Sheppard S. M. F. Identification of the origin of ore-
forming solutions by the use of stable isotopes. In Volcanic
processes in ore genesis (London: Institution of Mining and
Metallurgy and Geological Society, 1977), 25-41.
3. Sheppard S. M. F. and Epstein S. D/H and 180//160
ratios of minerals of possible mantle or lower crustal origin.
Earth planet. $ci. Lett., 9, 1970, 2 3 2 - 9 .
4. Wenner D. B. and Taylor H. P. Jr. Oxygen and hydrogen
isotope studies of the serpentinization of ultramafic rocks in
oceanic environments and continental ophiolite complexes.
Am. J. $cL, 273, 1973, 2 0 7 - 3 9 .
5. Margaritz M. and Taylor H. P. Jr. Oxygen and hydrogen
isotope studies of serpentinization in the Troodos ophiolite
complex, Cyprus. Earth planet. Sci. Lett., 23, 1974, 8--14.
6. Heaton T. H. E. and Sheppard S. M. F. Hydrogen and
oxygen isotope evidence for the origins of the fluids during
metamorphism of oceanic crusts. (Troodos complex,
Cyprus). Paper presented to NATO Advanced Study Insti-
tute on Volatiles in Metamorphism, 1974.
7. Margaritz M. and Taylor H. P. Jr. Oxygen, hydrogen and
carbon isotope studies of the Franciscan formation, Coast
Ranges, California. Geochim. cosmochim. Acta, 40, 1976,
215-34.
8. Ohmoto H. and Rye R.O. Hydrogen and oxygen iso-
topic compositions of fluid inclusions in the Kuroko de-
posits, Japan. Econ. Geol., 69, 1974, 947-53.
9. Javoy M. Utilisation des isotopes de I'oxyg~ne en mag-
matologie. Th~se de Doctorat d'Etat des Sciences Physiques,
Facult~ des Sciences de Paris, 1970.
10. Muehlenbachs K. and Clayton R. N. Oxygen isotope
ratios of submarine diorites and their constituent minerals.
Can. J. Earth Sci., 8, 1971, 1591-5.
11. Spooner E. T. C. etal. 018 enriched ophiolitic meta-
basic rocks from E. Liguria (Italy), Pindos (Greece), and
Troodos (Cyprus). Contrib. Mineral. Petrol., 47, 1974,
41-62.
12. Wilson R. A. M. The geology of the Xeros-Troodos
area. Mere. geol. Surv. Cyprus no. 1, 1959, 1-135.
13. Gass I. G. and Masson Smith D. The geology and gra-
vity anomalies of the Troodos Massif, Cyprus. Phil. Trans.
R. Soc., A255, 1963, 417-67.
14. Moores E. M. and Vine F.J. The Troodos Massif,
Cyprus and other ophiolites as oceanic crust: evaluation and
implications. Phil. Trans. R. Soc., A268, 1971,443-66.
15. Constantinou G. and Govett G. J. S. Genesis of sul-
phide deposits, ochre and umber of Cyprus. Trans Instn
Min. Metal/. (Sect B: Appl. earth sci.), 81, 1972, B34-46.
16. Constantinou G. and Govett G. J. S. Geology, geo-
chemistry and genesis of Cyprus sulfide deposits. Econ.
Geol., 68, 1973, 843-58.
17. Gass I. G. and Smewing J. D. Intrusion, extrusion and
metamorphism at constructive margins: evidence from the
Troodos Massif, Cyprus. Nature, Lond., 242, 1973, 26-9.
18. Menzies M. and Allen C. Plagioclase Iherzolite-residual
mantle relationships within two eastern Mediterranean
ophiolites. Contrib. Mineral. Petrol., 45, 1974, 197-213.
19. Robertson A. H. F. and Hudson J. D. Pelagic sediments
in the Cretaceous and Tertiary history of the Troodos
Massif, Cyprus. Spec. Pubis int. Ass. Sediment., 1, 1974,
403-36.
20. Constantinou G. Hydrothermal alteration of the
basaltic lavas of the Troodos Ophiolitic ComPlex associated
with the formation of the massive sulphide deposits. In
Volcanic processes in ore genesis (London: Institution of
Mining and Metallurgy and Geological Society, 1977), 77.
21. Heaton T. H. E. A hydrogen and oxygen isotope study
of the metamorphism and mineralisation of the Troodos
complex, Cyprus. Ph.D. thesis Edinburgh University, 1976.
22. Streckeisen A. To each plutonic rock its proper name.
Earth Sci. Rev., 12, 1976, 1--33.
23. Coleman R. G. and Peterman Z. E. Oceanic plagio-
granite. J. geophy~ Res., 80, 1975, 1099-108.
24. Miyashiro A. Shido F. and Ewing M. Metamorphism
in the mid-Atlantic ridge near 24 ~ and 30~ Phil. Tran& R.
Soc., A268, 1971, 589-603.
25. Liou J. G. Synthesis and stability relations of epidote,
Ca2AI2FeSi3012(OH). J. Petrology, 14, 1973, 381-413.
26. Gustafson W. I. The stability of andradite, hedenbergite,
and related minerals in the system C a - F e - S i - O - H .
J. Petrology, 15, 1974, 455-96.
27. Syers J. K. etal. Quartz isolation from rocks, sediments
and soils for determination of oxygen isotopes composition.
Geochim. cosmochim. Acta, 32, 1968, 1022-5.
28. McCrea J. M. On the isotopic chemistry of carbonates
and a paleotemperature scale. J. chem. Phy~, 18, 1950,
849-57.
29. Godfrey J. D. The deuterium content of hydrous
minerals from the east-central Sierra Nevada and Yosemite
National Park. Geochim. cosmochim. Acta, 26, 1962,
1215-45.
30. Clayton R. N. and Mayeda T. K. The use of bromine
pentafluoride in the extraction of oxygen from oxides and
silicates for isotopic analysis. Geochim. cosrnochim. Acta,
27, 1963, 43-52.
31. Taylor H. P. Jr. The application of oxygen and hydro-
gen isotope studies to problems of hydrothermal alteration
and ore deposition. Econ. Geol., 69, 1974, 843-83.
32. Suzuoki T. and Epstein S. Hydrogen isotope fraction-
ation between OH-bearing silicate minerals and water.
Geochim. cosmochim. Acta, 40, 1976, 1229-40.
33. Jehl V. Poty B. Weisbrod A. and Sheppard S. M. F.
Unpublished data.
34. Taylor H. P. Jr. The oxygen isotope geochemistry of
igneous rocks. Contrib. Mineral. Petrol., 19, 1968, 1-71.
35. Garlick G. D. and Dymond J. R. Oxygen isotope ex-
change between volcanic materials and ocean water. Bull.
geol. Soc. Am., 81, 1970, 2137-42.
36. Muehlenbachs K. and Clayton R. N. Oxygen isotope
studies of fresh and weathered submarine basalts. Can. J.
Earth Sci., 9, 1972, 172-84.
37. Pineau F. et al. Oxygen isotope variations in marginal
basin and ocean-ridge basalts. Earth planet. Sci. Lett., 28,
1976, 299-307.
38. Easton A. J. etal. Low-temperature metasomatic
garnets in marine sediments. Phil. Tran& R. Soc., in press.
39. Muehlenbachs K. and Clayton R. N. Oxygen isotope
geochemistry of submarine greenstones. Can. J. Earth Sci.,
9, 1972, 471-8.
40. Deer W. A. Howie R. A. and Zussman J. Rock forming
minerals, vol. 3: Sheet silicates (London: Longmans,
Green and Co., 1962), 270 p.
41. Vine F. J. and Moores E. M. A model for the gross
structure, petrology, and magnetic properties of oceanic
crust. Mere. geol. Soc. Am. 132, 1972, 195-205.
42. Liou J. G. Kuniyoshi S. and Ito K. Experimental studies
of the phase relations between greenschist and amphibolite
in a basaltic system. Am.J. Sci., 274, 1974, 613-32.
43. Shackleton N. J. and Kennett J. P. Paleotemperature
history of the Cenozoic and the initiation of Antarctic
glaciation: oxygen and carbon isotope analyses in DSDP sites
277, 279, and 281. Initial Rep. Deep Sea Drilling Project,
29, 1975, 743-55.
55
44. Holloway J. R. The system pargasite-H20-CO2: a
model for melting of a hydrous mineral with a mixed
volatile fluid - I. Experimental results to 8 kbars. Geo-
chim. cosmochim. Acta, 37, 1973, 651-66.
45. Omoto H. Systematics of sulfur and carbon isotopes in
hydrothermal ore deposits. Econ. Geol., 67, 1972, 551-78.
46. Knauth L. P. and Epstein S. Hydrogen and oxygen iso-
tope ratios in silica from the JOIDES Deep Sea Drilling
Project. Earth planet. Sci. Lett., 25, 1975, 1-10.
47. Clayton R. N. Muffler L. J. P. and White D. E. Oxygen
isotope study of calcite and silicates of the River Ranch
No. 1 Well, Salton Sea geothermal field, California. Am. J.
Sci., 266, 1968, 968-79.
48. Sheppard S. M. F. Nielsen R. L. and Taylor H. P. Jr.
Oxygen and hydrogen isotope ratios of clay minerals from
porphyry copper deposits. Econ. Geol., 64, 1969, 755-77.
49. Lister C. R. B. On the penetration of water into hot
rock. Geophys. J. R. astr. Soc., 39, 1974, 465-509.
50. Johnson A. E. Textural and geochemical investigation
of Cyprus pyrite deposits. Ph.D. thesis, University of
Western Ontario, 1970.
51. Johnson A. E. Origin of Cyprus pyrite deposits. In
24th Int. Geol. Congr. (Montreal: The Congress, 1972),
Sect.4, 291-8.
52. Clark L. A. Volcanic ores: comparison of cupriferous
pyrite deposits of Cyprus and Japanese Kuroko deposits.
In Proc. IMA-IAGOD meetings, "70, IAGOD volume
Takeuchi Y. ed. (Tokyo: Society of Mining Geologists of
Japan, 1971), 206-15. (Spec. Issue no. 3)
53. Hutchinson R. W. and Searle D. L. Stratabound pyrite
deposits in Cyprus and relations to other sulphide ores. In
Proc. IMA-IAGOD meetings, "70, IAGOD volume Takeuchi
Y. ed. (Tokyo: Society of Mining Geologists of Japan, 1971),
198-205. (Spec. Issue no. 3)
54. Sakai H. Isotopic properties of sulfur compounds in
hydrothermal processes. Geochem. J. (Jap.), 2, 1968,
29-49.
55. Rye R. O. and Ohmoto H. Sulfur and carbon isotopes
and ore genesis: a review. Econ. Geol., 69, 1974, 826-42.
56. Helgeson H. C. Solution chemistry and metamorphism.
In Researches in geochemistry Abelson P. H. ed. (New York:
Wiley, 1967), vol. 2, 362--404.
57. Steiner A. Clay minerals in hydrothermally altered
rocks at Wairakei, New Zealand. Clays Clay Minerals, 16,
1968, 193-213.
58. Browne P. R. L. and Ellis A. J. The Ohaki-Broadlands
hydrothermal area, New Zealand: mineralogy and related
geochemistry. Am. J. Sci., 269, 1970, 97-131.
59. T6masson J. and Kristmannsd6ttir H. High temperature
alteration minerals and thermal brines, Reykjanes, Iceland.
Contrib. Mineral. Petrol., 36, 1972, 123-34.
60. Bischoff J. L. and Dickson F. W. Seawater-basalt in-
teraction at 200~ and 500 bars: implications for origin of
sea-floor heavy-metal deposits and regulation of seawater
chemistry. Earth planet Sci. Lett., 25, 1975, 385--97.
61. Hajash A. Hydrothermal processes along mid-ocean
ridges: an experimental investigation. Contrib. Mineral.
Petrol., 53, 1975, 205-26.
62. Schneider A. The sulfur isotope composition of
basaltic rocks. Contrib. Mineral. Petrol., 25, 1970, 95-124.
56
63. Thode H. G. and Monster J. Sulfur isotope geochemis-
try of petroleum, evaporites and ancient seas. In Fluids in
subsurface environments - a symposium. Mere. Am. As~
Petrol. Geol. 4, 1965, 367-77.
64. Nielsen H. Sulphur isotopes and the formation of
evaporite deposits. In Geology of saline deposits: Proceed-
ings of the Hanover Symposium 1968 (Paris: UNESCO,
1972), 91-102.
65. Holser W. T. and Kaplan I. R. Isotope geochemistry of
sedimentary sulfates. Chem. Geol., 1, 1966, 93-135.
66. Holland H. D. Gangue minerals in hydrothermal
deposits. I n Geochemistry of hydrotherrnal ore deposits
Barnes H. L. ed. (New York: Holt, Reinhart and Winston,
1967), 382--436.
67. Hunt J. P. Porphyry copper deposits. In Volcanic
processes in ore genesis (London: Institution of Mining
and Metallurgy and Geological Society, 1977), 98.
(Summary).
68. Bear L. M. The geology and mineral resources of the
Akaki-Lythrodondha area. Mern. geol. Surv. Cyprus
no. 3, 1960, 122 p.
69. Bear L. M. and Morel S. W. The geology and mineral
resources of the Agros-Akrotiri area. Mere. geol. Surv.
Cyprus no. 7, 1960, 88 p.
70. Carr J. M. and Bear L. M. The geology and mineral
resources of the Peristerona--Lagoudhera area. Mere. geol.
Surv. Cyprus no. 2, 1960, 79 p.
71. Gass I. G. The geology and mineral resources of the
Dhali area. Mere. geol. Surv. Cyprus no. 4, 1960, 116 p.
72. Engel A. E.J. Engel C. G. and Havens R. G. Chemical
characteristics of oceanic basalts and the upper mantle. Bull.
geol. Soc. Am., 76, 1965, 719-34.
73. Cann J. R. Major element variations in ocean-floor
basalts. Phil. Trans. R. Soc., A268, 1971,495-505.
74. Gass I. G. Is the Troodos Massif of Cyprus a fragment
of Mesozoic ocean floor? Nature, Lond., 220, 1968,
39-42.
75. Cann J. R. Spilites from the Carlsberg Ridge, Indian
Ocean. J. Petrology, 10, 1969, 1-19.
76. Hart R.A. A model for chemical exchange in the
basalt-seawater system of oceanic layer II. Can. J. Earth Sci.,
10, 1973, 799-816.
77. Elder J. W. Physical processes in geothermal areas. In
Terrestrial heat flow Lee W. H. K. ed. (Washington, D.C.:
American Geophysical Union, 1965), 211-39. (Geophys.
Monogr. no. 8)
78. Peterman Z. E. Coleman R. G. and Hildreth R. A.
87Sr/S6Sr in mafic rocks of the Troodos Massif, Cyprus.
Prof. Pap. U.S. geol. Surv. 750-D, 1971, 157-61.
79. Chapman H. J. Spooner E. T. C. and Smewing J. D.
87Sr enrichment of ophiolitic rocks from Troodos, Cyprus,
indicates seawater interaction. Am. geophys. Union, E~S
Trans, 56, 1975, 1074. (Abstract)
80. Deffeyes K. S. The axial valley: a steady state feature
of the terrain. In The rnegatectonics of continents and
oceans Johnson H. and Smith B. eds (New Brunswick,
N.J.: Rutgers UniversiW Press, 1970), 194-222.
81. Spooner E. T. C. and Fyfe W. S. Sub-sea-floor meta-
morphism, heat and mass transfer. Contrib. Mineral. Petrol.,
42, 1973, 287-304.
82. Lister C. R. B. On the thermal balance of a mid-ocean
ridge. Geophys. J. R. astr. Soc., 26, 1972, 515-35.
83. Robertson A. H. F. Cyprus umbers: basalt-sediment
relationships on a Mesozoic ocean ridge. Jlgeol. Soc. Lond.,
131, 1975, 511--31.
84. Jankovic S. The origin of base-metal mineralization on
the Mid-Atlantic Ridge. In 24th Int. Geol. Congr. (Montreal:
The Congress, 1972), Sect. 4, 326-34.
85. Humphris S. E. Elemental fluxes during hydrothermal
alteration of basalts. Am. geophys. Union EoS Trans., 56,
1975, 469. (Abstract)
86. Aumento F. Loncarevic B. D. and Ross D. I. Hudson-
geotraverse: geology of the mid-Atlantic ridge at 45~
Phil. Trans. R. Soc., A268, 1971,623-50.
87. Pearce J. A. and Gale G. M. Identification of ore-
deposition environment from trace-element geochemistry of
associated igneous host rocks. I n Volcanic processes in ore
genesis (London: Institution of Mining and Metallurgy and
Geological Society, 1977), 14-24.

57
Hydrodynamic model for the
origin of the ophiolitic cupriferous
pyrite ore deposits of Cyprus
E. T. C. Spooner
Department of Geology and Mineralogy, University
of Oxford, Oxford, England
553.065:553.661.21 (564.3)
Synopsis
The information presently available - in particular, that
derived from the use of the isotopes (87Sr/86Sr, 6 180 and
~341) as geochemical tracers - indicates that a model for
mineralization and metamorphism of the Upper Cretaceous
ophiolitic rocks of the Troodos MasSif, Cyprus, which in-
volves hydrothermal convection of sea water, is reasonable.
During water-rock interaction in the recharge part of the
cycle of convective heat and mass transfer, the pillow lavas,
Sheeted Dyke Complex and uppermost gabbros were hydro-
thermally metamorphosed at zeolite facies to amphibolite
facies conditions, and base metals (e.g. Fe, Mn, Cu, Zn, Pb,
Ni, Co, Ag and Au) were leached by the fluid from the
rocks. A t the positions of discharge of hot fluid at the sea
water--rock interface sulphide ore deposits or ferro-
manganoan hydroxyoxide sediments (umbers) were
precipitated.
By calculation of a Rayleigh number for the conditions
of hydrothermal activity it is shown that the fluid dynamic
requirements for vigorous free convection were satisfied. It
is deduced that the ore deposits developed within the active
zone o f a spreading oceanic ridge, and took " 105 yr to
form. Available geochemical evidence is discussed to demon-
strate that the hydrothermal convection mode/is chemically,
as well as fluid-dynamically, reasonable.
As discussed by Fyfe and Henley, 1 the essential require-
ments for the successful formation of a hydrothermal ore
deposit may be simplified to the following.
(a) A working fluid which has transferred mass in solu-
tion. Stable isotope geochemistry has shown that this hot
water has had diverse origins. 2
(b) A source rock from which metals and other elements
have been derived.
(c) Rock that was sufficiently permeable, in bulk, to
permit hot water either to flow through it or through frac-
tures in it.
(d) A source of energy, which drove a hydrodynamic flow
system, by means of which mass was transferred from a
dispersed source to a position at which it was deposited in
concentrated form. The energy source must have been suf-
ficiently large to maintain flow for a sufficient length of
time to form a large deposit, and the pattern of flow must
have been geometrically persistent to precipitate ore con-
tinuously at a specific position. Energy sources have nor-
mally been in the form either of heat, which has driven
some form of hydrothermal convection, or of gravitational
potential energy, which has driven water flow in aquifers.
(e) A mechanism of producing persistent, localized and
selective mineral precipitation so that a deposit, which was
sufficiently large and concentrated to be of economic
interest, was formed.
58
It is the object of this paper to show how these require-
ments, which have been stated above in general terms, may
be specifically identified in regard to the origin of the
cupriferous pyrite ore deposits in the ophiolitic rocks of
the Troodos Massif, Cyprus.
Within the sequence of metamorphosed basaltic pillow
lavas, sheeted basic dykes, trondhjemites, gabbroic rocks and
ultramafic rocks, which form the ophiolitic complex of
Troodos, 3,4 the ore deposits occur usually as massive sulp-
hide lenses underlain by approximately pipe-shaped
9stockworks. 5 - 9 The lenses of ore are semi-conformably
intercalated within the pillow lava sequence, and, from the
following three observations, it is known that stockworks
penetrate into the sheeted dyke complex: (1) one small ore
deposit at Alestos has been opened in Basal Group material;
(2) in the southeast footwall of the Limni opencast mine
strongly pyritized sheeted complex has been faulted up
against the exploited cupriferous stockwork that was formed
in pillow lavas 10 (author's personal observation); and (3)
small gossans developed on uneconomic pyritized material
are ubiquitously developed in sheeted complex 5 (G.
Constantinou, personal communication).
As ore deposits, those of the Troodos Complex are not
particularly large; however, some have been quite rich in
copper. Typical ore tonnages of the more important de-
posits before development were between 1 000 000 and
10 000 000. 5 The largest and richest deposit was
Mavrovouni, which contained "~ 15 000 000 tonnes ore
grading between 3.5 and 4.5% Cu. 5 Limni, which is the
largest working mine at present, produces "~ 450 000
tonnes ore per annum at a mean grade of ~ 0.85% copper 11
The current status of the cupriferous pyrite mining
industry of Cyprus has recently been reviewed by
Spooner. 12
Since the Troodos Complex, in particular, and ophio-
litic rocks, in general, are considered to be tectonically
emplaced fragments of oceanic crust and upper
mantle, 3,4,13-24 a discussion of mineralization of the
Troodos ophiolitic rocks is also, probably, a discussion of
mineralization at sea-floor spreading ridges, and, therefore,
of the oceanic crust in general. It must be emphasized,
however, that it is uncertain how representative of the
oceanic crust ophiolitic sequences are.
Geochemical evidence for sea-water interaction during
down-flow
Essentially all the material of the sequence of pillow lavas
and sheeted dykes has been strongly hydrothermally
metamorphosed. 3,18 Metamorphic assemblages charac-
teristic of the zeolite, greenschist and amphibolite facies
were developed.IS Metamorphism affected the complex
transition zone between the sheeted dykes and the gab-
broic rocks, but did not significantly affect the deeper
gabbros. Secondary minerals occur filling veins and
vesicles, and as pseudomorphic replacements of the primary
igneous minerals. With respect to most major elements,
many trace elements, iron oxidation ratios and water
contents metamorphism was strongly non-
isochemical. 24~5~6 All of these features are characteristic
of hydrothermal metamorphism and metasomatism, which
was a consequence of water-rock interaction at elevated
temperatures. The clearest indication of this is the observed
increase in water content, which reflected the generation
of secondary hydrous minerals. For example, the H20 +
contents of nine metabasalts and metamorphosed dykes
reported by Spooner and co-workers 26 range from 0.6
to 4.5 wt % (average, 2.5 wt %). These data compare
with a mean water content for fresh deep-sea basalt of
"~ 0.30 wt %27,28 and suggest an increase in water content
of "" 2 wt %.
This type of metamorphism is the same sort of process
that has been observed happening within geothermal
systems in the basaltic terrain of Iceland. 29,3~ The
mineral assemblages observed in Icelandic geothermal
systems, in the ophiolitic rocks of the Troodos Complex
and in ocean-floor material 32-37 are all closely similar.
This inference of water interaction with the sequence

I i I I I I
- Modern sea water 1 Q
Cam p o n i a n - M aestr ichtian 2
- Sea w a t e r

Opl~iotitic rocks, # Interstitiat -

Troodos~ Cyprus z e o l i t e (JSI)
ZeoLite facies
9 9 9 metabasatts -
Greenschist facies
metaba.so[ts -
Greenschist facies
9 ~ 9 ~ o metodoterites -
A m p h i b o t i t e facie.,
metadoterite -
Trondhjemites
9 []e 8oo (variably attered )"

[] # ~ 0 Metagabbros -

Fresh 9abbros
~/ll~I Plagioctase fetdspar
concentrates from
fresh 9abbros -

Comparative data Means oP vatues,

and totat range
Thoteiitic basatts,3
lcetand (13my. B.P.- -
Recent )
Reykjanes r i d g e L ' 5 -

3
E) M.A.R. {/,.6~N)/'
/,
M.A.R. ( 6 ~ 1 7 6

I~ EP.R. and J.ER./' -

Leg 34 (D.S.D.P.),Nazr
plate (15-/,0 my. B.P. )

M.I.O.R./,

I- ~ M.I.O.R .6

j_~ Back arc basin~

M a r i a n a s arc
I I I I I I I I
0.7020 0.7030 0.70/.0 0.7050 0.7060 0.7070 0.7080 O-70go
87Sr/86Sr ratios

Fig. 1 Strontium isotope ratios of (1) fresh and altered mafic material from ophiolitic
complex of Troodos, Cyprus, (2) modern and Campanian-Maestrichtian sea water,
(3) fresh deep-sea basaltic rocks, and (4) fresh tholeiitic basalts from Iceland. Altered
material from "l'roodos Complex is variably strontium isotopically enriched relative to
fresh material. All results quoted relative to an Eimer and Amend (E & A) SrCO 3
standard value of 0.70800 and age-corrected, where necessary. Numbers near data
points indicate numbers of samples analysed to produce range and mean. Solid diamonds,
new data; open squares, Peterman and co-workers; 4~ open circles, Greenbaum; 41
open diamonds, Coleman and Peterman. 141 Sources of additional data: 1, Veizer and
Compston; 142 2, Peterman and co-workers; 42 3, O'Nions and Gr6nvold; 143 O'Nions
and Pankhurst; 144 4, Hart; 145 5, O'Nions and Pankhurst; 14e 6, Subbarao and Hedge; 147
and 7, Hart and co-workers. 148 M.A.R., Mid-Atlantic Ridge; E.P.R., East Pacific Rise;
J.F.R., Juan de Fuca Ridge; M.I.O.R., Mid-Indian Ocean Ridge

59
of pillow lavas, basal group and sheeted dykes, which
represents a thickness that may be approximately estim-
ated at "" 2 kin, 38 poses three important questions:
(1) what was the source of the hot water?; (2) in what
direction was the water flowing?; and (3) how much
water flowed through the sequence?
The most direct evidence to suggest that this fluid was
of sea-water origin has been obtained from the use of
strontium isotopic ratios as geochemical tracers. 26,39
From analysis of exceptionally fresh gabbro samples
and of plagioclase feldspar concentrates from fresh gabbros
initial magmatic 87Sr/86Sr ratios from 0.70338 _+0.00010
to 0.70365 -+0.00005 have been obtained. 26,3g These
values, and all subsequent quoted values, have been age-
corrected and are quoted relative to an Elmer and Amend
standard SrCO 3 value of 0.70800. The errors quoted
for individual determinations are 2o. This range of stron-
tium isotope ratios overlaps the top end of the known
spread of ratios, which have been obtained from fresh
deep-sea basalts (Fig. 1).
Seventeen 87Sr/86Sr determinations of variably meta-
morphosed basaltic pillow lavas and dykes 26,40,41 range
from values comparable with the initial magmatic range to
considerably enriched ratios as high as 0.7069. The mean
87Sr/S6Sr ratio of the 17 determinations is 0.7049
-+0.0009 (lo). It is clear, therefore, that hydrothermal
metamorphism was accompanied by strontium isotopic
contamination. This indicates that the source of the
hydrothermal fluid was sea water, since sea water was the
only available reservoir of higher 87Sr/86Sr ratio material in
the submarine environment in which the Troodos Complex
was formed. The 87Sr/S6Sr ratio of Campanian-
Maestrichtian sea water has been determined as 0.7076
+0.0003(10)
9 from analysis of 25 marine carbonate rocks. 42
This interpretation is confirmed by the strontium isotopic
composition,which has been found for a zeolite (laumontite,
Ca(AI2Si4Olo).4H20) sample obtained from ~ 800 m
below the top of the pillow lavas. The zeolite
occurred filling an interstice between pillows and, therefore,
the strontium in it should be a sample of the strontium
contained in the fluid that flowed through the pillow lava
pile. The 87Sr/86Sr ratio of this zeolite (0.70760
+0.00003) proved to be statistically indistinguishable from
the sea water that coexisted with the ophiolitic rocks of
the Troodos Complex in Campanian (Upper Cretaceous)
time.26,39
Spooner and co-workers 26 have shown that the
amount of strontium isotopic contamination requires that
the bulk water/rock ratio was greater than "" 15:1. Since
this ratio is greater than a water/rock ratio of "~ 0.01, which
would be implied by a reasonable estimate for the porosity
of the sequence of "~ 1%,43 a flow system must have
operated. The depth of sea-water penetration in this
flow system was approximately 2 km, since metagabbros
and altered trondhjemitic rocks from the top of the gabbro
sequence were also strontium isotopically
contain i nated. 26,39
This interpretation of sea-water interaction is supported
by the whole-rock oxygen isotopic composition of
metabasalts and metadolerites. Relative to a 6180 value
of -" + 60/00 for fresh basaltic rocks, 44.45,46 the hydro-
thermally metamorphosed rocks show strong modifications.
Strong enrichments in the upper part of the sequence
change to depletions at depth (Fig. 2 (a)). 47,48 These
changes in the ~ 180 values of the altered rocks indicate
interaction with large volumes of hot water and are consis-
tent with interaction with heated sea water. 26,47,48
60
I I I i I 1 I I
Troodos~Cyprus
0.5 _ i "I
E
.! o2
e5
Z~.o
1:3
1.5
.6
L b)
2-0 2"! , ,
+6 +6 +8 +10 +12 0
-
I /"~1 I i
0-2 0.4 0.6 O'8 1.0
~S18O, %~ FezOs/FeO"
Fig. 2 Variation with depth of oxygen isotope ratios
(6180) and mean oxidation ratios (Fe203/FeO*) irr
sequence of ophiolitic metabasalts and metadolerites of
Troodos Complex
The model of sea-water circulation has been further
reinforced and developed by separate mineral combined
hydrogen and oxygen stable isotope studies. 48 Heaton and
Sheppard 48 have shown, in particular, that the 6D value
of the hydrothermal fluid in equilibrium with greenschist
facies mineral assemblages at 200-450~ was close to the
sea-water value of 0~
The direction of fluid flow has been deduced from the
shape of the iron oxidation profile of the altered rocks
(Fig. 2(b)). 26 Relative to an iron oxidation ratio
(Fe203/FeO*, where FeO* = total iron as FeO) of ~ 0.18
for fresh deep-sea basaltic material, rocks from near the
top of the sequence are highly oxidized. Values as high as
0.9 have been reported by Spooner and co-workers. 26
The degree of oxidation decreases downwards (Fig. 2(b)).
This information shows, first, that the hydrothermal fluid,
which was responsible for metamorphism of the ophio-
litic rocks in the submarine environment, was oxygenated
and, secondly, because rocks near the top are most oxidized,
that the fluid was introduced through the upper surface
boundary layer. These points supply excellent confirma-
tory evidence that the hydrothermal fluid was parentally
sea water, since sea water contains 2 - 6 ml (STP) kg"1
of dissolved oxygen, 49-55 and that the fluid motion
through the hydrothermally metamorphosed rocks of
the Troodos Complex possessed a dominant downward
compo'~ent in areas away from the cupriferous pyrite
orebodies.
For a hydrothermal flow system the parameter of
greatest significance that expresses the total amount of
mass transfer is the water/rock ratio integrated over the
period of hydrothermal activity. Increases in 87Sr/86Sr
ratios and 6180 values can only yield minimum estim-
ates for the size of this parameter. The iron oxida-
tion profile can,however, be used to obtain an absolute
water/rock ratio, because it is a function of the total
amount of oxidizing agent introduced by the flowing
fluid. 26 The water/rock ratio may either be quoted as a
bulk ratio relative to the whole size of the sequence, or as
a value standardized to a unit volume of rock. In the
latter case it is numerically equal to the fluid flux. 26 It
has been shown 26 that this method, when applied to the
ophiolitic rocks of the Troodos Complex, gives a bulk
water/rock ratio of "- 103:1 (for water at STP), and a
water/rock ratio relative to 1 cm 3 of rock of 108-109:1.
These figures, which were obtained for the case when
oxygen dissolved in sea water was the important oxidizing
agent, suggest therefore that the interstitial fluid was
replaced approximately 105 times, for a porosity for the
sequence of ~ 19o.
Hence, the three important questions posed at the
beginning of this discussion have now been answered with a
reasonable degree of confidence. (1) The hydrothermal
fluid that produced the hydrothermal metamorphism and
metasomatism of the basaltic and doleritic ophiolitic
rocks was of sea-water origin; (2) the principal component
of fluid flow was downward; and (3) the quantity of water
that flowed downward was ~" 108-109 cm (cm3/cm2).
Geochemical evidence for formation of sulphide ore
deposits during discharge of modified sea water
The large quantity of sea water that flowed down through
the pillow lava and dyke sequence must also have discharged
from the system. It has been suggested that the sites of
discharge are represented by the cupriferous pyrite ore
deposits.26,47, 56
Separate mineral combined hydrogen and oxygen
=sotope data 48 support this model. From analysis of the
~D and ~ 180 values of chlorite mineral separates and of
the ~ 180 values of quartz concentrates obtained from
mineralized stockwork material it has been shown that, for
reasonable estimates of stockwork temperatures
(250-350~ the ~D value of the fluid was close to that
of sea water (~" 0%o).
The sulphur isotopic composition of the pyrite in the
ore deposits may also be examined to see if it shows any
evidence for the presence of reduced sea-water sulphate
as well as remobilized magmatic sulphide.
As Ohmoto 57 has shown, the (~34S value for a solid
sulphide (~34S i) is governed by five main variables:
(~34S i = f (T, pH, f o 2 , / , (~34S~s)
where T = temperature, pH = Ioglo {H ~ } , fo2 = oxygen
fugacity, / = ionic strength and (~34ST~s = ~34S value of
total dissolved sulphur.
The reason for the dependence on temperature and
(~34S~.s is that sulphur isotopes, like hydrogen, carbon
and oxygen isotopes, fractionate among solid sulphur-
bearing phases and dissolved sulphur species. These
fractionations are temperature- and phase-dependent.
Unlike oxygen isotopes, however, sulphur isotopes may be
distributed among several possible oxidized and reduced
sulphur species in solution. Hence, the ~34S value of a
given solid sulphide is-also dependent on the relative pro-
portions of HSOz, SO4, H2S, HS- and S-- in solution.
These relative proportions are controlled by equilibria that
are affected by changes in T, pH, f02 and/.
The temperatures of stockwork mineralization may be
approximately estimated at 250-350~ 48 At these tern-
peratures the pH of the solution may be estimated from
the calculated pH values for hot sea water that is interact-
ing with basaltic material within the Reykjanes geothermal
system, Iceland, 3~ and from the calculated pH values
found in laboratory experiments on basalt-sea-water
interaction. 58 Bj~rnsson and co-workers 30 calculated pH
values of 5.2 and 6.1 at 221 and 277~ respectively.
Bischoff and Dickson 58 computed a pH of ~" 5 for syn-
thetic basalt--sea water interaction at 200~ These
values are close to the neutral pH value for pure water at
similar elevated temperatures (neutral pH at 250~ =
5 . 5 - 6 ) , 59 or are slightly acid. Hence, a pH for the
aqueous solution in the stockwork of 5 - 6 may be estim-
ated. With existing mineralogical information it is not
possible to set reasonable constraints on oxygen fugacity
during mineralization, since the recorded mineral assem-
blage of pyrite-chalcopyrite-sphalerite-galena 9 is not
particularly sensitive to variations in fo2. The ionic
strength (/) may be estimated at ~-" 1 by comparison with
sea water at 0.7 and (~34S~s may be treated as a variable to
be estimated. Ohmoto 57 has shown that for the above
conditions and for an indefinable oxygen fugacity
(~34Spyrite -- (~34S~ s ~ ~ 4. 1%o
The sulphur isotopic compositions (h34S) of pyrite
from the Cyprus ore deposits have been reported as rang-
ing from +0.43 to 7.00/oo. 6'60 In both papers attention was
drawn to the fact that there does not appear to be any
significant difference between massive ore and stockwork
pyrite (~34S values. Hutchinson and Searle 6 reported a
mean value of +4.8%o from 19 samples and Johnson 6o
reported a mean value of +3.75 o/oo from 24 analyses.
These data would suggest that during ore deposition
(~34S~s ~ -- 0.57 to + 6.0 %o
Therefore, since the (~34S value of primary magmatic
sulphides in the basaltic and doleritic rocks was, by com-
parison with modern deep-sea basalts, ~" + 1%o.61 this
information clearly indicates that the dissolved sulphur
in the ore-forming solutions was variably contaminated by
a certain amount of higher $34S sulphur. Since Upper
Cretaceous sea water, with a 634S value of ~" + 16%o for
dissolved sulphate, 62 was the only readily accessible source
of high (~34S sulphur, this implies a sea-water sulphur
component in the sulphide of the ore deposits greater than
~- 10%. If the oxygen fugacity were near that for the
magnetite-hematite stability curve, the observed pyrite
$34S values could have been produced from a fluid that
contained reduced sea-water sulphate only (see Ohmoto 57).
It has been inferred that this was the case for the origin
of sulphur in the Kuroko-type deposits of Japan. 63
Hydrothermal convection
The geochemical evidence summarized above strongly
suggests that chemically modified sea water was the fluid
which, during down-flow, hydrothermally metamorphosed
the pillow lava and dyke sequence and which, during up-
flow, precipitated the sulphide ore deposits. This model
of fluid circulation corresponds geometrically to hydro-
thermal convection in a permeable medium heated from
below, and which had an open upper boundary through
which hot water was free to discharge. The energy source
that drove convection was heat produced during generation
of the lithosphere by sea-floor spreading, and associated
with the introduction of basaltic magma into the base of
the oceanic crust.
Modern examples of hydrothermal convection systems
are provided by the geothermal systems of Iceland, New
Zealand and elsewhere. 64,65,66 Past examples of high-level
magma intrusions that were cooled by convecting water,
often of meteoric origin, have been widely documented. 2
This qualitative model of fluid circulation in the ophio-
litic rocks of the Troodos Complex is schematically shown
in Fig. 3(a). For comparison, a photograph (Fig. 3(b)) is
shown of an analogue of convection in a permeable
medium. 65 The inferred pattern of circulation corresponds
with that which has been produced experimentally.
The model of hydrothermal convection has, so far, been
constructed purely in order to explain the geochemical and
61
mineralogical evidence. It is the next object of this dis-
cussion to shown that it can be independently p r o v e d
that the fluid dynamic conditions necessary for vigorous
free convection in the permeable medium of the ophiolitic
rocks of the Troodos Massif were satisfied.
Fig. 3 Schematic diagram (a) of inferred mode of hydro-
thermal convection, metamorphism and mass transfer
within ophiolitic rocks of Troodos Complex. General
metamorphism 18 occurred in delocalized zones of re-
charge flow, whereas formation of cupriferous pyrite
orebodies 5-10 occurred in localized zones of
discharge. 26,47.56 Precipitation of ferromanganoan
'umber" sediments was also probably a consequence of
discharge of hot metalliferous solutions. 56,128-131 a,
gabbro, metagabbro and variably altered trondhjemite;
b, sheeted dyke complex; c, pillow lavas; d, umbers
(Fe/Mn enriched sediments); e, massive sulphide ore;
f, mineralized stockwork. Dashed lines approximate to
metamorphic boundaries between zeolite facies (z),
greenschist facies (g) and amphibolite facies (am).
Continuous lines are approximate flow lines for hydro-
thermal convection; dotted line is approximate form of
isotherm produced by convection (see Elder 85). Photo-
graph (b) of free convection in Hele--Shaw cell with
free water discharge. Open upper boundary condition was
simulated by withdrawing water above rising jet of hot
fluid through siphon. Constant volume was maintained by
replenishing cell at same rate as water was withdrawn.
Pattern of localized discharge and delocalized recharge is
similar to inferred pattern of convective flow within ophio-
litic rocks of Troodos Complex (Fig. 2(a)). Reproduced, by
kind permission of American Geophysical Union and
Professor J. W. Eide r, from Fi 9. 19 65
For flow through permeable media the equation of
motion is given by Darcy's law. 67 This relationship is a
transport equation in which the flux and the pressure
gradient,through a permeable medium are directly porpor-
tional. In simplest form it is given by
k dP
O F = ~. ~ 9 ~ .
/~ dx
The constant of proportionality is subdivided into two
62
factors. One constant is relevant to the properties of the
travelling fluid, and the other to the hindrance to motion
imposed by the solid material. The former is the viscosity
(dynamic viscosity (~u) for a volume flux; kinematic
viscosity (9) for a mass flux), and the latter the permea-
bility (k). The cgs unit of permeability is the millidarcy
(= 0.987 x 10"11 cm 2).
Darcy's law holds for most of the flow regimes that are
likely to be encountered in natural, undisturbed subsurface
situations. For example, at the lower limit for flow rates
in materials of the lowest permeabilities (10"6-10 .7 md),
Brace and co-workers 68 have shown that Darcy's law is
obeyed. Fluxes as low as 5 x 10 -8 cm sec"1 were measured.
The materials used were unfractured samples of igneous
and metamorphic rocks, which were examined at pressures
up to 4 kb. 68,69 At high flow rates through highly per-
meable materials Darcy's law is replaced by a power law
when inertial effects become important in the upper part
of the laminar flow regime, but before the transition to
turbulent flow in the void space.7~ This occurs above a
Reynolds number (RM) of ~" 1--10 for granular porous
materials,71, 72 where the Reynolds number is defined
Qv " '~
RM -
1)
The length scale (A) is defined in different ways for
different situations involving both unhindered flow of
viscous fluids, and flow through permeable media: for
example, it is taken as the mean grain diameter for granular
porous media. For flow rates of the order of 11~6 cm/sec,
which can be encounted in natural convective systems, 73,74
a kinematic viscosity (9) of 2 x 10 .3 cm -2 sec"1 109 and an
average grain diameter of 0.1 cm, the Reynolds number is
0.5 x 10"4 - a value well below the critical value of 1-10.
Another situation that may be encounted naturally involves
concentrated flow through open fractures. For example,
most of the boreholes that produce large amounts of hot
water and steam from the Wairakei geothermal field tap
fractures. 75 Also, the occurrence of hydrothermal ore
deposits as lodes filling fissures is well known. In this case
is defined as the channel width, and the Reynolds
number for the transition from laminar to turbulent flow
between rough parallel-sided walls is ~ 2 x 103. 70 For
high flow rates of the order of 10 .3 cm sec"1 , a channel
width of 100 cm and a kinematic viscosity of 2 x 10 .3 cm 2
sec"1 the Reynolds number is 50. The value is, again, well
below the critical value.
When a permeable medium is heated from below, the
density variations produced by the non-uniform distribu-
tion of temperature generate buoyancy forces. These
forces to cause hot fluid to rise. At low intensities of heat-
ing the buoyancy forces are weak and the rate of conduc-
tive heat transfer is sufficient to prevent the build-up of
significant density differences. Hence, convective
motion 79,80 does not occur until the temperature difference
across the permeable layer becomes large enough to over-
come the restraints imposed by the kinematic viscosity of
the fluid, and the thermal diffusivity (KM) and permeability
of the porous material.
By use of the Boussinesq approximation that density
variations are only important in generating buoyancy
forces, 76 the balance between the activating and restraining
forces may be expressed by a dimensionless Rayleigh
number (R a) :77
k.~.Ae.g.H
Ra- - -
KM.v
where k = permeability,/3 = coefficient of cubical expansion v = 1.5 x 10-3 cm 2 sec"1 lO9,11o,111
of the fluid, A0 = temperature difference across the per- For this approximate evaluation of a Rayleigh number it
meable layer, g = acceleration due to gravity, H = thickness has been assumed that the properties of a dilute saline
of the permeable medium, KM = thermal diffusivity of the solution are comparable with those of pure water. The
saturated medium temperature- and pressure-dependence of the coefficient of
KM cubical expansion, the heat capacity and the kinematic
KM - viscosity have been neglected. The values of the parameters
p.Cp used are those applicable to the means of the estimated
where K M = thermal conductivity of the permeable temperature and pressure variations (200~ 250 bar).
medium, p = density of the fluid and Cp = specific heat of The thickness of the permeable medium (H) may be
the fluid, and v = kinematic viscosity of the fluid(= p/p), taken as the thickness of the pillow lava-dyke sequence in
where/1 = dynamic viscosity. which the bulk of sea-water circulation occurred. This
If the Rayleigh number exceeds a certain critical value, corresponds to a depth of "" 2 km. 38 At the base of this
which is dependent on boundary conditions, 78 convection sequence amphibolite facies metamorphic assemblages
occurs. The amount by which the Rayleigh number have been recorded, which would suggest a temperature
exceeds the critical value is directly related to the vigour difference across the permeable layer of "" 450~ 112
of convection. For a uniformly permeable medium, which
is heated uniformly and constantly from below, and which Table 1 Measured and calculated bulk permeabilities of
is enclosed above and below by two isothermal parallel natural materials
plates, Lapwood 8o deduced that the critical value of the
Rayleigh number should be 47r2 (= 39.48). Despite early Material Bulk permeability, md
experiments, which produced quantitative disagreement
Postglacial basaltic lavas in 106_107
with this value, e2,e3"84 because initial but not steady-state
Iceland 66
processes were examined, subsequent experimental work
Impermeable rock with vertical Vertical permeability =
has established its validity. 81~5 For different boundary 2x 106
open fractures 0.3 cm wide
conditions, however, different critical Rayleigh numbers spaced at 1-m intervals (see
apply. 7a,8~ For example, the critical value is 27.1 when the text)
upper boundary is open and permits unimpeded discharge Loose sand ?0 104_10 s
of hot fluid and the lower boundary is isothermal. 7e It is Impermeable rock with Vertical permeability = ~" 104
17.65 for the same configuration but when heat flux is vertical open fractures 89cm
maintained constant along the lower boundary. If the wide spaced at 1-km intervals
temperature distribution along this lower boundary is non- (see text)
uniform, fluid flow occurs for all values of the Rayleigh Horizontal permeability for 35
number. buried flood basalts in
Reykjavik geothermal
The theory of convection in permeable media has found system, Iceland.66
particular applications in geophysical problems associated Permeability measured in
with hot groundwater flow in geothermal areas, and, hence, bulk over depth-interval
with the production of geothermal power. Following early from 300 m below surface
studies, 77,86,87 Elder 65,71 analysed the thermal structure to 1300 m
of the Wairakei geothermal system in New Zealand and Main Waiora volcaniclastic 14 (5--27)
showed that free convection of water below the surface was aquifer, Wairakei geothermal
system, New Zealand 7s
required to maintain the thermal output. A quantitative
pipe model for heat and mass transfer was developed. 65,71 Wairakei geothermal system 73 Vertical permeability = ~" 11
Donaldson 73,74,97,98 and others 88,9o,91,92 have continued Aquifer cemented by < 1; e.g. 0.6 (hole 25)
to develop further applications of the theory. hydrothermal minerals,
Wairakei geothermal
Recently, qualitative ideas based on convection in porous system, New Zealand 75
media have been used to explain other geological and geo-
Broadlands geothermal "0.1--1
physical phenomena. One of the more important develop-
system, New Zealand 74
ments has been the recognition that sea water circulates
convectively within spreading oceanic ridges on a massive Oceanic crust between 5 "~ 0.5
and 30 km from Galapagos
scale.47,56,99-102 Particular attention has also been paid spreading axis and down to
to aspects of the origin of hydrothermally metamorphosed depth of "~3.5 km 1o2
rocks associated with high-level igneous intrusive bodies and Unfractured diabase, gabbro 10-6_10-7
of certain types of hydrothermal ore deposit. 2,47,103,104,105 and granite 68,69
In order to evaluate the Rayleigh number for the con-
ditions of hydrothermal metamorphism and mineralization
The final parameter that requires to be estimated is the
of the ophiolitic rocks of the Troodos Complex the follow-
permeability. The values of measured and calculated per-
ing values of some of the variables in the Rayleigh number
meabilities given in Table 1 show that this is the most
expression may be taken:
critical variable to assess in calculating a Rayleigh number,
/3 = 10"3 ~ lO7 because the permeabilities of natural materials vary con-
g = 103cmsec "2 siderably. For example, the permeabilities of postglacial
KM= 5 x 10.3 cm sec"1 basaltic lava flows in Iceland and of unfractured diabase or
The last value was obtained by assuming that the thermal gabbro differ by 12-14 orders of magnitude.
conductivity of the saturated medium is approximately A relatively high permeability might be expected for the
equal to the thermal conductivity of the rock ('~" 0.005 cal pillow lava sequence, because of the existence of vesicles,
cm-1 sec-1 Oc-1 ),108 and by taking values of density and porosity between pillows, and drainage cavities within
isobaric heat capacity from Helgeson and Kirkham. 1~ pillows. Also, pillows show a radiating pattern of cooling

63
joints. The most useful measurement that may be used to
estimate a permeability for pillow lava is that given by
P~lmason 66 for the bulk horizontal permeability of a
lO00-m section of interbedded lava flows and tuffs
(~" 35 md; see Table 1). Hence, a bulk permeability of
"~ 40 md would seem to be a reasonable estimate.
The permeability of the sheeted dyke complex was con-
trolled by a ramifying system of planar and curved fractures,
which now appear filled with secondary minerals (author's
personal observation). The original fractures may have been
produced by thermal contraction during cooling of the
dykes (cf. Lister loo). To show how effective the develop-
ment of cracks can be in developing permeability the re-
lationship given below between permeability (k) and the
width (b) and spacing (a) of a parallel series of cracks may
be used. 113
b3
k- - - . 1011 ( f o r k in millidarcies)
12(a + b)
For example, if basaltic rock cools by 600~ and cracks
develop at a spacing of "~ 1 m, then, for the coefficient of
linear thermal contraction of basalt (~ 5 x 10-6 ~ ),1o8
cracks 0.3 cm wide develop. Such a system of cracks pro-
vides a high permeability of 2 x 106 md. Personal observa-
tions of the sheeted dyke complex of Cyprus suggest that
this sort of spacing for mineral-filled fractures is quite rea-
sonable. It is unlikely, however, that such high permeabili-
ties were reached, because of the precipitation of secondary
minerals.
Another effect to be investigated, which would provide
permeability, is that of faulting, which is ubiquitous in the
ophiolitic rocks and the oceanic crust. For example, a spac-
ing of fractures only 89cm wide every 1 km can provide a
permeability of "~ 104 md. Hence, before extensive pre-
cipitation of secondary minerals in porosity and fractures
high-bulk permeabilities in the range 4 0 - 1 0 4 and might be
expected. For such permeabilitiesthe Rayleigh number
would have been between 6 x 103 and 1.2 x 106. Since this
range of values far exceeds the critical values of the Rayleigh
number (27.1 and 17.7 for an open top and uniform lower
boundary condition, see above), it can be concluded that the
conditions necessary for vigorous convection of sea water in
the ophiolitic rocks of the Troodos Massif, Cyprus, were
satisfied. These values compare with Rayleigh number
estimates for the Wairakei and Reykjavik geothermal
systems of "~ 104 65 and 2 x 102,66 respectively.
The precipitation of secondary minerals in porosity and
fractures is probably one of the most efficient ways of
reducing the vigour of convection. This effect has been
widely documented. For example, Grindley 75 recorded
that the permeabilities of hydrothermally cemented aquifers
in the Wairakei system are reduced to ~" 0.6 md relative to
a bulk permeability of the uncemented aquifer of "" 14 md.
The permeability of the aquifer in the Broadlands system,
which is a mixture of massive rhyolite flows interbedded
with fine-grained tufts, is "~ 0.1-1 md. 74 These low values
are in good agreement with an estimate of 0.5 md calculated
by Ribando and co-workers lO2 for the upper 3 89km of
the oceanic crust outside the active zone of spreading on
the Galapagos Ridge, and suggest that hydrothermal
cementation may reduce permeabilities from 50--104 md
down to "~ 0.5 md, and, presumably, lower. This would
correspond to a reduction of the Rayleigh number from
6 x 103 - 1.2 x 106 down to a low value of 60. Hence,
this would suggest that the main factor that may limit
vigorous convection of sea water to a fairly narrow zone at
the spreading ridge crest is the sealing of the 'plumbing system'
64
by 'scale' formed by the precipitation of hydrothermal
minerals. Weak convection can, and does, persist outside
the zone of greatest geothermal activity. 96,1~
The main conclusion from this discussion of the value
of the Rayleigh number for the conditions of sea-water
interaction with the ophiolitic rocks of Troodos, Cyprus,
is that the requirements necessary for vigorous hydro-
thermal convection were clearly satisfied. Therefore, the
qualitative model of convection, which was formulated
on the basis of the geochemical evidence, has been justi-
fied. An implication of this discussion is that the pre-
cipitation of secondary minerals was probably a very
important factor in limiting vigorous convection to a
fairly narrow zone near the spreading ridge axis.
Further aspects of hydrothermal convection
It has now been established, first, that the geochemical
evidence requires that a sea-water circulation system pro-
duced the hydrothermal mineralization and metamorphism
of the Troodos ophiolitic rocks and, secondly, that the
fluid dyanmic conditions necessary for free convection
were fully satisfied. Hence, further aspects of the hydro-
thermal convection model may now be explored.
Geometry
As is shown in Fig. 4, the major ore deposits distributed
along the northern pillow lava outcrop of the Troodos
Complex have an imperfectly regular geometrical distrib-
ution. Hence, accepting that the positions of these ore
deposits represent the original positions of rising hot water,
it can be said that the three-dimensional geometry of hydro-
thermal convection approximated to an arrangement of
polygonal 'cells', the locations of which were fixed. At
the centres of the cells occurred rising buoyant jets of
hot fluid, which were approximately axially symmetric
in cross-section. The mean half-distance between
the rising jets of 2.6 +_1.4 (lo) kin, which has been measured
from the 21 spacings between ore deposits marked in Fig. 4,
was comparable with the height of the permeable layer
('~ 2 km): this observation is qualitatively consistent with
convection, s9 The sizes of the ore deposits are all com-
parable. For example, the mean tonnage of the 11 ore
deposits for which data are available in Bear 5 and
Constantinou and Govett 9 is 3 600 000 +_4 000 000 (lo).
This observation is, also, qualitatively consistent with a
convection model, since it suggests that a similar total
quantity of fluid was discharged at each rising jet position.
These observations might, at first, seem to be consistent
with the simplest model of a permeable layer heated uni-
formly from below. It has been shown experimentally, how-
ever, that at Rayleigh numbers greater than ~ 2 4 0 - 2 8 0
a stable pattern of axially symmetric polyhedral convection
cells, 93,94 or two-dimensional rolls, gives way to a fluctuat-
ing convective state in which the positions of rising and
descending fluid vary with time. 95 It has also been shown
that at Rayleigh numbers greater than ~ 380 a three-
dimensional geometry of convection is stable relative to
two-dimensional rolls or axially symmetric cells. 89 The
inferences that can be made from the mode of occurrence of
the Cyprus ore deposits are not consistent with these
observations because, at the high Rayleigh numbers con-
sidered to have applied during hydrothermal convection in
the ophiolitic rocks (> 6 x 103; see above), the geometry of
convection was not fluctuating through time and involved
"axially symmetric cells'. This suggests that there was an
additional mechanism, first, of generating axially symmetric
rising jets of hot fluid, and, secondly, of maintaining their
 Fig. 4 Distribution of cupriferous pyrite ore deposits along northern pillow lava out-
crop of Troodos Complex. 1, Mavrovouni; 2, Lefka; 3, Apliki; 4, Skouriotissa;
5, Alestos; 6, Memi; 7, Kokkinoyia; 8, Kokkinopezoula; 9, Agrokipia; 10, Klirou
district; 11, Kapedhes; 12, Kambia; 13, Mathiati; 14, Lythrodondha, 15, Sha
positions through time. The simplest mechanism of pro-
ducing this effect would have been to have had approximate-
ly axially symmetric thermal anomalies on the lower boun-
dary layer of the permeable layer. The discrete gabbro
intrusions, which would have corresponded to discrete
magma bodies, and which have been identified in the
Troodos Complex by Allen, T M could have provided this
requirement.
Quan titles o f circulating fluid
The mean sulphur content of the ore deposits along the
northern pillow lava outcrop, as calculated from data for
11 deposits in Bear 5 and Constantinou and Govett, 9 was
1 400 000 tonnes. If a value for the original sulphur
content of the fluid can be estimated, an approximate
estimate can be obtained for the total quantity of fluid
required to form an ore deposit in the Troodos ophiolitic
rocks.
In the Salton Sea geothermal system in California hot
water at 300-350~ coexists with sulphides ( p y r i t e -
chalcopyrite--sphalerite-galena) and the sulphide content
of the solution is 15-30 ppm. 1~ Similarly, it can be cal-
culated from the data in Helgeson 115 that the sulphide
content of a solution formed by the stoichiometric dis-
solution of pyrite at 350~ is ~" 20 ppm S. Hence, taking
a value of "~ 20 ppm initial dissolved sulphur, a minimum
quantity of ~" 7 x 1016 g of fluid was required to form an
average Cyprus ore deposit. This estimate is minimal,
because it assumes that all the sulphur that was in solution
was precipitated as sulphide, and because it neglects the
quantities of sulphide in uneconomic stockworks.
This value may now be compared with the amount of
fluid that could have been supplied by circulating sea water,
and with a hypothetical amount of primary magmatic
fluid that could have been released during crystallization
of underlying gabbroic magma.
The rectangular area outlined in Fig. 4 contains the 14
most important known ore deposits, and one strongly
mineralized area (Klirou district), along the northern
pillow lava outcrop. This rectangle measures 57 km x 8 km
and consists of a 456 km 2 area. Hence, the drawdown
region per orebody would have been "~ 30 km 2. Taking
the figure of 108--109 cm calculated by Spooner and co-
workers 26 for the approximate quantity of sea water that
flowed down, the total amount of sea water available to
flow up was ~ 1019-1020 g per ore deposit. This value
compares favourably with the minimum value of ~ 1017 g
fluid calculated above to have been necessary to form an
ore deposit.
The amount of magmatic fluid that could have been
available can be calculated on the basis of the unlikely
assumption that the total amount of water in the basaltic
magma, from which the gabbroic rocks formed, was released
in a focussed form. Knowing from the rare-earth element
abundances of the ophiolitic rocks of the Troodos Complex
that the basaltic magma which formed the igneous rocks
was chemically comparable with deep-sea basalt, 24 an
initial water content of ~ 0.30 wt % can be estimated for
the gabbroic magma. 27,28 Since the thickness of gabbroic
rocks is "~" 2 km, a volume of ~ 6 x 1016 cm 3 of gabbro per
orebody could have contributed magmatic water. Assum-
ing complete release of "" 0.3 wt % water, the maximum
quantity of primary magmatic fluid available was
~" 6 x 1013 g. This maximum value is three orders of
magnitude lower than the minimum estimate of the quan-
tity of fluid required per ore deposit ('~ 7 x 1016 g).
Therefore, this order of magnitude calculation suggests that
the theoretically available amount of primary magmatic
fluid was too small to have been important in formation of
the ore deposits.
Time required to form an ore deposit, and fluid flow rates
A characteristic of the sulphide ore deposits in the ophio-
litic rocks of the Troodos Complex is that they are fre-
quently overlain by unmineralized and relatively unaltered
lavas and crosscut by unmineralized dykes. 6-1~ This in-
formation indicates that ore deposition had ceased before
the end of volcanism. Hence, if a spreading oceanic ridge
model 20 for the origin of the ophiolitic rocks of the
Troodos Massif is accepted, this leads to the important
deduction that the formation of the ore deposits, and,
presumably, the associated hydrothermally metamorphosed
rocks, occurred within the active volcanic zone of a ridge.
Detailed near-bottom observation of magnetic anomalies
near active spreading ridges in the Pacific region have shown
that at least 90% of the upper ~" 0.5 km of the volcanic
part of the oceanic crust forms between 1 and 6 km of the
ridge axis. 116,117 The width of the active zone appears to b~
independent of spreading rate. 117 This range of values is in
good agreement with the observations of Needham and
Francheteau 118 on rifted ridges, which suggest that the
active volcanic zone is limited to the inner floor of the
central graben. Needham and Francheteau 118 reported
widths of inner floors of between 2 and 9 km, which sug-
gests that active volcanism occurs within 1 - 5 km of the
spreading axis. Accepting that the ophiolitic rocks of the
Troodos Massif were formed by a smooth ('fast') spreading
ridge, since no evidence is present for very large-amplitude
normal faults (~" 1.5 km 1 18), which characterize rough
('slow') spreading ridges, a spreading rate between 2 and
9 cm/yr/limb may be estimated. 119 For the maximum width
of a volcanically active zone and minimum spreading rate,
65
and vice versa, this indicates that the time spent by a given
section of oceanic crust within the active volcanic zone is
between ~" 104 and ~" 3 x 105 yr. This suggests that the
time taken to form an ore deposit in the ophiolitic rocks was
~- 104-10 s yr. Since estimates for the times required for
the formation of hydrothermal ore deposits are few and
far between, this range of values can only be compared with l
that for the actively forming 120 Atlantis II deep deposit,
Red Sea. 121 This deposit, which, on a dry basis, approxi-
mates to a 200 000 000-tonne orebody grading 0.34% Cu
and 1.58% Zn (calculated from data in Hackett and
Bischoff 122), has formed within the last 1.3 x 104 yr. 122
Although the chemistry of the fluid that is forming this
deposit is unusual, 121 this information shows that a
deposit an order of magnitude larger than any of the
Cyprus deposits has formed in a time comparable with
the minimum estimate for the time of formation of
these Cyprus deposits.
For a formation of ~ 10 s yr and for a required
fluid quantity of "~ 7 x 1016 g a vertical fluid flow rate of
"- 2 x 104 g sec"1 may be calculated. Taking a value for
the diameter of a stockwork of ~" 500 m, as calculated
from information in Trennery and Pocock 123 and
Constantinou and Govett, 9 this is equivalent to a vertical
fluid flux of ~" 3 x 10 .6 g cm "2 sec"1 . The value is entirely
reasonable for a geothermal system because it compares
with flow rates of 10"s-10 "7 g cm "2 sec"1 calculated for
the Wairakei and Broadlands geothermal systems in New
Zealand.73, TM
One final comment to be made is that the ore deposits
did n o t form synchronously. The reason for this is that
they are distributed in a direction at right angles to the
strike of the dolerite dykes in the sheeted complex, and the
sheeted complex formed by a process of sea-floor
spreading. 4,13,2~ Since the distance between the
Mavrovouni (furthest west) and Sha (furthest east) mines is
"" 50 km (Fig. 4), there was a difference of "~ 2 . 5 - 0 . 6 x
106 yr between the times of formation of the two deposits.
Since the estimated formation time for a deposit is ~-" 105yr
(see above), the older deposit, which may be deduced to
have been Sha from the direction of one-way chilling in the
sheeted dykes, 2~ had finished forming before the other
had started. At any given instant in time ore deposits were
only actively forming within the volcanically active zone of
a ridge crest (see above). After formation they were
carried laterally away from the ridge axis while new ore
deposits developed.
Chemical aspects of sea water-basalt interaction and of
ore deposition
The final question to consider is: can circulating sea water,
which is a relatively dilute solution (total dissolved solids
= "~ 3.5 wt %; chloride ion content = 2 wt %), produce ore
deposits?
Experimental work on basalt-sea water interaction at
temperatures between 200 and 500~ pressures between
500 and 800 bar and run-times of up to two years 58,124,125
has shown that sea water changes from an oxidizing, slightly
alkaline Na"i" - M g + + - C I - S O 4 = solution to a reducing,
slightly acidic Na+-Ca++-CI solution characterized by
high concentrations of dissolved Si, Fe and Mn. These
experimentally observed changes are closely similar to the
differences that have been found between normal and
heated sea water which reacts with basaltic material in the
Reykjanes geothermal system, Iceland. 29,3~
66
Two results of the experimental work would seem to be
particularly pertinent to the problem of ore deposit form-
ation by sea water circulating in basaltic rock.
(a) It was observed that sea-water sulphate was inorgani-
cally reduced to dissolved sulphide-bearing species and solid
sulphide phases. In all experiments H2S was released when
the experimental capsules were broken open after a run,
and Hajash 125 reported the synthesis of chalcopyrite and
pyrrhotite. This shows that the process of inorganic sea-
water sulphate reduction, which is suggested to have
happened from the sulphur isotopic composition of the
Troodos sulphide deposits (see above), is feasible.
(b) The concentrations of dissolved iron and manganese
increased considerably. For example, in the long experi-
ments of Bischoff and Dickson 58 at only 200~ iron and
manganese concentrations of N 5 ppm were observed. In
shorter experiments at 500~ Hajash 125 reported iron
concentrations up to 1500 ppm and manganese concentra-
tions up to 190 ppm. As noted by the authors, these data
show that the process of metal leaching by sea-water circu-
lation, which has been suggested as responsible for the
formation of aluminium-poor metalliferous sediments on
oceanic ridges, 56,126,i27'128 is feasible. The equivalents
of such sediments, called umbers, locally occur immediately
overlying, or intercalated with, the pillow lava sequence in
Cyprus.129,13o, 131
The results for other transition metals show that they
were not leached to nearly the same extent as Fe and Mn,
which conveniently explains why F e - M n hydroxyoxides are
the main components of submarine metalliferous sediments.
Absolute concentrations cannot be quoted because of sus-
pected valve-block contamination in the experiments of
Bischoff and Dickson 58 (j. L. Bischoff, written communi-
cation) and because the experiments of Hajash 125 were
carried out in gold capsules, which absorb transition ele-
ments - in particular, copper (A. Hajash, written communi-
cation). Hajash 125 did, however, observe that ~" 75% of
copper in the solid basaltic material was leached at
400-500~
Hence, although experimental data do not yet prove that
transition elements other than iron and managanese can be
significantly leached, they do suggest that this process is
highly probable.
That modified sea water can produce base-metal ore
deposits is, however, shown by the Miocene
C u - P b - Z n - A g - A u Kuroko deposits of Japan (see
Lambert and Sato 132 for a recent summary). Ohmoto
and Rye 133 have shown from stable isotopic analysis
(~ D and ~lSo) of fluid inclusions that the dominant com-
ponent of the ore fluid was of sea-water origin. Of greater
interest, however, are the results of microthermometric
investigation of fluid inclusions (summarized by Lambert
and Sato 132), which have shown that the ore fluids were
not particularly saline and contained, on average,
~" 6 wt % NaCI equivalent. This salinity is only twice as
concentrated as sea water (~" 3.5 wt % NaCI equivalent), and
falls exactly in the range of salinities recorded for fluid
inclusions of sea-water origin in greenschist facies meta-
morphic rocks dredged from the ocean floor. 134
Hence, these data indicate, first, that the salinity of sea
water increased during interaction with the oceanic
crust, 134 and, secondly, that such a solution ('~ 1 molal
NaCI) has formed ore deposit~*
*The Kuroko deposits of Japan range in size from 2 000 000 to
35 000 000 tonnes ore and, in 1972, supplied "" 45% of Japan's
domestic production of Cu, 26% of its Zn, 33% of its Pb, 31% of
its Ag and 20% of its Au. ]32
 If sulphide deposition is now considered, the following on the basis of the foregoing discussion, that the following
observations have been used to infer that the stratiform specific identifications can be made for the case of the
lenses of sulphide ore, which occur in the Cyprus ophio- cupriferous pyrite ore deposits in the ophiolitic rocks of
litic rocks, actually formed on the ocean floor, e-9 the Troodos Massif, Cyprus.
(a) The lenses consist of pure sulphide ore and are not (a)Working fluid Sea water
replacement deposits.
(b) Source rock The ~- 2-km thick sequence of
(b) The lenses occur semi-conformably intercalated from which basaltic pillow lavas passing
within the pillow lava pile. metals and gradationally down into basic
(c) The ore is often altered at the top to a goethitic other elements dykes and underlying gabbros
material, which shows primary sedimentary structures were leached
(ochre). This material has been interpreted as a submarine (c) Sufficiently Same sequence as in (b); per-
weathering product of ore. 7 permeable rock meability was a consequence of
(d) The massive ore is often overlain by unmineralized to permit fluid space between pillows, joints
pillow lavas and/or crosscut by unmineralized dolerite flow and faults. Estimates of per-
dykes. meability are as high as 104 md
Hence, the problem that arises is why a reduced sulp- and greater than "" 0.5 md
hide precipitate formed at the sea water-basalt interface and (d) An energy The energy source was heat
not an oxidized precipitate, which would have been in equili- sou rce which supplied by the magma, which
brium with sea water and comparable with the umber. A maintained a cooled to form the gabbros.
reasonable qualitative explanation would seem to be that geometrically The hydrodynamic system
at high fluid discharge rates the hydrothermal fluid may persistent involved hydrothermal convec-
locally control the chemical environment immediately in hydrodynamic tion in a permeable medium with
the region of discharge. This would permit fluids to cool system an open upper boundary layer.
before mixing with sea water and allow a sulphide precipi- Geometrically persistent flow
tate to form. The viability of such a mechanism is shown was maintained for "~ 105 yr,
by the fact that the volcaniclastic sediments that floor the probably by thermal anomalies
Baia di Levante on the east side of the island of Vulcano in associated with specific magma
the Tyrrhenian Sea off southern Italy are actively being intrusions
cemented by pyrite and marcasite just below the sea water-
(e) A mechanism Sulphide precipitation was pro-
sediment interface in the discharge zones of fumaroles. 135
of maintaining bably principally a result of ex-
The temperature and chemistry of sea water are severely localized tremely sharp cooling when the
modified as much as 10 m away from any particular area
sulphide hot fluid at 250-350~
of fumaroles. In contrast, at slow rates of discharge hydro-
precipitation approached the ocean floor,
thermal fluid may mix with sea water as it cools. This
which was permanently main-
would result in the oxidized iron-manganese hydroxyoxide
tained at "" 2~ by flowing
precipitates often observed to form around the vents of
ocean-bottom water
submarine hot springs. 136-140
It is considered that this list is unusual for its complete-
This interpretation would therefore suggest that the ness.
sulphide ore deposits were formed at high fluid discharge The role of submarine volcanism in the formation of
rates, whereas the umbers were formed at low discharge the Cyprus sulphide deposits would appear to have been
rates. T h e geological relationships of ore and umber to (a) the provision of a sequence of permeable rocks of
the ophiolitic pillow lavas show that mineralization ended basaltic composition through which sea water could circu-
before the end of volcanism, whereas umber precipitation late convectively and transfer material and (b) the provision
persisted after the end of volcanism. Hence, this inform- of the energy source - heat.
ation suggests that the transition from sulphide to umber
formation corresponded to a waning of hydrothermal Acknowledgment
activity during the time taken for a given segment of
oceanic crust to migrate out of the active volcanic zone of a I am particularly grateful to the following people for
spreading oceanic ridge. This inference is consistent with helpful discussion: Dr. E. M, Parmentier, Dr. T. H. E.
interpretations (see above) that very vigorous hydrothermal Heaton, Dr. J. D. Smewing, Dr. G. Constantinou, Professor
activity is restricted tO the spreading zone itself. The W. S. Fyfe, Professor J. W. Elder, R. D. Beckinsale, Miss
reasons why hydrothermal activity declined are, first, that H. J. Chapman and Dr. R. G. W. Kidd. I am also very
the Large quantities of heat removed by the process of grateful to Mrs. Sheila Cotton for typing the manuscript.
hydrothermal convection itself would also have reduced
the energy source and, secondly, that the permeability of References
oceanic crust would have been significantly reduced by 1. Fyfe W. S. and Henley R.W. Some thoughts on chemical
precipitation of secondary minerals (see above). transport processes, with particular reference to gold.
Hence, the observed transition from sulphide to umber Miner. Sci. Engng, 5, 1973, 295-303.
formation is consistent with the model of hydrothermal 2. Taylor H. P. Jr. The application of oxygen and hydrogen
convection at a spreading oceanic ridge. isotope studies to problems of hydrothermal alteration and
ore deposition. Econ. Geol., 69, 1974, 843-83.
Conclusions 3. Gass I. G. and Masson Smith D. The geology and gravity
Returning to the list given in the introduction of the anomalies of the Troodos Massif, Cyprus. Phil. Trans. R.
generalized factors believed to be necessary for the success- Soc., A255, 1963, 417-67.
ful formation of a hydrothermal ore deposit, it is considered, 4. Moores E. M. and Vine F. J. The Troodos Massif,

67
Cyprus, and other ophiolites as oceanic crust: evaluation
and implications. Phil. Trans. R. Soc., A268, 1971, 443--66.
5. Bear L. M. The mineral resources and mining industry
of Cyprus. Bull. geol. Surv. Dep. Cyprus no. 1, 1963, 208 p.
6. Hutchinson R. W. and Searle D. L. Stratabound pyrite
deposits in Cyprus and relations to other sulphide ores. In
Proc. IMA-IAGOD meetings, "70 IAGOD volume (Tokyo:
Society of Mining Geologists of Japan, 1971), 198-205.
(Spec. Issue no.3)
7. Constantinou G. and Govett G. J. S. Genesis of sulphide
deposits, ochre and umber of Cyprus. Trans. Instn Mino
Metal/. (Sect. B: Appl. earth sci.), 81, 1972, B32-46.
8. Searle D. L. Mode of occurrence of the cupriferous
pyrite deposits of Cyprus. Trans. Instn Min. Metal/. (Sect.
B: Appl. earth sci.), 81, 1972, B189-97.
9. Constantinou G. and Govett G. J. S. Geology, geo-
chemistry, and genesis of Cyprus sulfide deposits. Econ.
Geol., 68, 1973, 843-58.
10. Adamides N. G. Geological history of the Limni con-
cession, Cyprus, in the light of the plate tectonics hypo,
thesis. Trans. Instn Min. Metall. (Sect. B: Appl. earth sci.),
84, 1975, B17-23.
11. Petropoulos P. G. Annual report of the Senior Mines
Officer for the year 1974 (Nicosia: Government Printer,
1975).
12. Spooner E. T. C. Cyprus pyrite to-day. Sulphur, no. 121,
1975, 23-7.
13. Gass I. G. Is the Troodos Massif of Cyprus a fragment
of Mesozoic ocean floor? Nature, Lond., 220, 1968,
39--42.
14. Moores E. M. Petrology and structure of the Vourinos
ophiolitic complex of northern Greece. Spec. Pap. geol. Soc.
Am. 118, 1969, 74 p.
15. Coleman R. G. Plate tectonic emplacement of upper
mantle peridotites along continental edges. J. geophys.
Res., 76, 1971, 1212-22.
16. Dewey J. F. and Bird J. M. Origin and emplacement of
the ophiolite suite: Appalachian ophiolites in Newfoundland.
J. geophys. Res., 76, 1971, 3179-206.
17. Pearce J. A. and Cann J. R. Ophiolite origin investi-
gated by discriminant analysis usingTi, Zr and Y. Earth
planet. Sci. Lett., 12, 1971, 339-49.
18. Gass I. G. and Smewing J. D. Intrusion, extrusion and
metamorphism at constructive margins: evidence from the
Troodos Massif, Cyprus. Nature, Lond., 242, 1973, 26-9.
19. Pearce J. A. and Cann J. R. Tectonic setting of basic
volcanic rocks determined using trace element analyses.
Earth planet. Sci. Lett., 19, 1973, 290-300.
20. Kidd R. G. W. and Cann J. R. Chilling statistics indi-
cate an ocean-floor spreading origin for the Troodos
complex, Cyprus. Earth planet. Sci. Lett., 24, 1974,
151-5.
21. Hynes A. Comments on "The TroOdos ophiolitic com-
plex was probably formed in an island arc", by
A. Miyashiro. Earth planet. Sci. Lett., 25, 1975, 213-6.
22. Moores E. M. Discussion of "Origin of Troodos and
other ophiolites: a reply to Hynes", by Akiho Miyashiro.
Earth planet. Sci. Lett., 25, 1975, 223-6.
23. Gass I. G. et ai. Comments on "The Troodos ophiolitic
complex was probably formed in an island arc", by
A. Miyashiro and subsequent correspondence by A. Hynes
68
and A. Miyashiro. Earth planet. Sci. Lett., 25, 1975,
236--8.
24. Kay R. W. and Senechal R. G. The rare earth geo-
chemistry of the Troodos ophiolite complex. J. geophys
Re~, 81, 1976, 964-70.
25. Smewing J. D. Simonian K. O. and Gass I. G. Metaba-
salts from the Troodos Massif, Cyprus: genetic implication
deduced from petrography and trace element geochemistry.
Contrib. Mineral. Petrol., 51, 1975, 49-64.
26. Spooner E. T. C. Chapman H. J. and Smewing J. D.
Strontium isotopic contamination and oxidation during
ocean floor hydrothermal metamorphism of the ophiolitic
rocks of Troodos, Cyprus. Geochim. cosmochim. Acta, in
press.
27. Moore J. G. Water content of basalt erupted on the
ocean floor. Contrib. Mineral. Petrol., 28, 1970, 272-9.
28. Moore J. G. and Schilling J. G. Vesicles, water, and
sulfur in Reykjanes Ridge basalts. Contrib. Mineral. Petrol.,
41, 1973, 105-18.
29. Tdmasson J. and Kristmannsd6ttir H. High tempera-
ture alteration minerals and thermal brines, Reykjanes,
Iceland. Contrib. Mineral. Petrol., 36, 1972, 123-34.
30. Bj~rnsson S. Arn6rsson S. and Tdmasson J. Economic
evaluation of Reykjanes thermal brine area, Iceland. Bull.
Am. Ass. Petrol. Geol., 56, 1972, 2380-91.
31. Kristmannsd6ttir H. and T6masson J. Hydrothermal
alteration of basaltic rocks in Icelandic geothermal areas.
Paper presented at the Second United Nations symposium
on the development and use of geothermal resources, San
Francisco, U.S.A., May 20-29, 1975. (Reykjav~k, Iceland:
National Energy Authority, 1975).
32. Melson W. G. and van Andel T. H. Metamorphism in
the Mid-Atlantic Ridge, 22~ latitude. Mar. Geol., 4,
1966, 165-86.
33. Cann J. R. Spilites from the Carlsberg Ridge, Indian
Ocean. J. Petrol., 10, 1969, 1-19.
34. Cann J. R. Petrology of basement rocks from Palmer
Ridge, NE Atlantic. Phil. Trans. R. Soc., A268, 1971,
605-17.
35. Matthews D. H. Altered basalts from Swallow Bank, an
abyssal hill in the NE Atlantic, and from a nearby seamount.
Phil. Trans. Ro Soc., A268, 1971,551-7.
36. Miyashiro A. Shido F. and Ewing M. Metamorphism
in the Mid-Atlantic Ridge near 24 ~ and 30~ Phil. Trans.
R. Soc., A268, 1971,589-603.
37. Bonatti E. etal. Metagabbros from the Mid-Atlantic
Ridge a~ 06 ~N: contact-hydrothermal-dynamic meta-
morphism beneath the axial valley. J. Geology, 83, 61-78.
38. Kidd R. G. W. Modelling of processes at spreading
plate boLrhdaries. Ph.D. thesis, University of East Anglia,
1976.
39. Chapman H.J. Spooner E. T. C. and Smewing J. D.
87Sr enrichment of ophiolitic rocks from Troodos, Cyprus,
indicates sea water interaction. Am. geophys. Union
E~S, Trans, 56, 1975, 1074. (Abstract)
40. Peterman Z. E. Coleman R. G. and Hildreth R. A.
SrS7/Sr86 in mafic rocks of the Troodos massif, Cyprus.
Prof. Pap. U.S. geol. Surv. 750-D, 1971, 157-61.
41. Greenbaum D. Geology and evolution of the Troodos
plutonic complex and associated ore deposits. Ph.D. thesis,
Leeds University, 1972.
42. Peterman Z. E. Hedge C. E. and Tourtelot H.A. Iso-
topic composition of strontium in sea water throughout
Phanerozoic time. Geochim. cosmochim. Acta, 34, 1970,
105-20.
43. Davis S. N. and DeWiest R. J. M. Hydrogeo/ogy (New
York: Wiley, 1966), 318-45.
44. Taylor H. P. Jr. The~oxygen isotope geochemistry of
igneous rocks. Contrib. Mineral. Petrol., 19, 1968, 1-71.
45. Muehlenbachs K. and Clayton R. N. Oxygen isotope
studies of fresh and weathered submarine basalts. Can. J.
Earth Sci., 9, 1972, 172-84.
46. Muehlenbachs K. and Clayton R. N. Oxygen isotope
geochemistry of submarine greenstones. Can. J. Earth Sci.,
9, 1972, 471-78.
47. Spooner E. T. C. eta/. 018 enriched ophiolitic
metabasic rocks from E. Liguria (Italy), Pindos (Greece),
and Troodos (Cyprus). Contrib. Mineral. Petrol., 47, 1974,
41-62.
48. Heaton T. H. E. and Sheppard S. M. F. Hydrogen and
oxygen isotope evidence for sea-water-hydrothermal
alteration and ore deposition, Troodos complex, Cyprus.
In Volcanic processes in ore genesis (London: Institution
of Mining and Metallurgy and Geological Society, 1977),
42-57.
49. Craig H. and Weiss R. F. The Geosecs 1969 inter-
calibration station: introduction, hydrographic features,
and total CO2--O 2 relationships. J. geophys. Res., 75,
1970, 7641-7.
50. Craig H. The deep metabolism: oxygen consumption
in abyssal ocean water. J. geophys. Res., 76, 1971,
5078-86.
51. Craig H. Chung Y. and Fiadeiro M. A benthic front in
the South Pacific. Earth planetSci. Lett., 16, 1972, 50-65.
52. Duedall I. W. and Coote A. R. Oxygen distribution in
the South Atlantic. J. geophys. Res., 77, 1972, 496-8.
53. Kroopnick P. Weiss R. F. and Craig H. Total CO2, 13C,
and dissolved oxygen - 180 at Geosecs II in the North
Atlantic. Earth planetSci. Lett., 16, 1972, 103-10.
54. Spencer D.W. Geosecs II, the 1970 North Atlantic
Station: hydrographic features, oxygen and nutrients.
Earth planet. Sci Left., 16, 1972, 91-102.
55. Roether W. and MLinnich K. O. The 1971 transatlantic
section of F/S "Meteor" near 40~ Earth planet. Sci. Lett.,
23, 1974, 91-9.
56. Spooner E. T. C. and Fyfe W. S. Sub-sea-floor meta-
morphism, heat and mass transfer. Contrib. Mineral. Petrol.,
42, 1973,287-304.
57. Ohmoto H. Systematics of sulfur and carbon isotopes
in hydrothermal ore deposits. Econ. Geol., 67, 1972,
551-78.
58. Bischoff J. L. and Dickson F. W. Seawater-basalt
interaction at 200~ and 500 bars: implications for origin
of sea-floor heavy-metal deposits and regulation of seawater
chemistry. Earth plane~ Sci. Lett., 25, 1975, 385-97.
59. Helgeson H. C. and Kirkham D. H. Theoretical pre-
diction of the thermodynamic properties of aqueous electro-
lytes at high pressures and temperatures. II1. Equation of
state for aqueous species at infinite dilution. Am. J. Sci.,
276, 1976, 97-240.
60. Johnson A. E. Origin of Cyprus pyrite deposits. In
24th Intn. geol. Congr. (Montreal: The Congress, 1972),
Sect. 4, 291-8.
61. Kanehira K. etal. Sulphide globules and sulphur iso-
tope ratios in the abyssal tholeiite from the Mid-Atlantic
Ridge near 30~ latitude. Geochem. J., 7, 1973, 89-96.
62. Hoefs J. Stable isotope geochemistry (Berlin:
Springer-Verlag, 1973), 77.
63. Rye R. O. and Ohmoto H. Sulfur and carbon isotopes
and ore genesis: a review. Econ. Geol., 69, 1974, 826-42.
64. B6dvarsson G. Physical characteristics of natural heat
resources in Iceland. J~kull., 11, 1961, 29-38.
65. Elder J.W. Physical processes in geothermal areas.
Am. geophys. Union Geophys. Monogr. 8, 1965, 211-39.
66. P~lmason G. On heat flow in Iceland in relation to the
Mid-Atlantic ridge. In Iceland and Mid-ocean ridges,
S. Bj6rnsson ed. (Reikjavik: Prentomidjan Leiftur H. F.,
1967), 111-27. (Soc. Sci. Islandica, 38, 1967).
67. Darcy H. P. G. Les fontaines publiques de la ville de
Dijon (Parie: Victor Dalmont, 1856), 647 p.
68. Brace W. F. Orange A. S. and Madden T. M. The effect
of pressure on the electrical resistivity of water-saturated
crystalline rocks. J. geophys. Res., 70, 1965, 5669-78.
69. Brace W. F. Walsh J. B. and Frangos W. T. Permeability
of granite under high pressure. J. geophys. Res., 73, 1968,
2225-36.
70. Scheidegger A. E. The physics of flow through porous
media (Toronto: The University Press, 1957), 248 p.
71. Elder J. W. Heat and mass transfer in the earth: hydro-
thermal systems. Res. Bull. N.Z. Dep. Sci. Ind. 169, 1966,
115p.
72. Palm E. Weber J. E. and Kvernvold O. On steady con-
vection in a porous medium. J. Fluid Mech., 54, 1972,
153-161.
73. Donaldson I. G. The flow of steam water mixtures
through permeable beds: a simple simulation of a natural
undisturbed hydrothermal region. N.Z.J. Sci., 11, 1968,
3--23.
74. Donaldson I. G. The estimation of subsurface flows
and permeabilities from temperature and pressure data.
Geothermics, Special issue 2, 1970, 677-83.
75. Grindley G. W. The geology, structure and exploitation
of the Wairakei geothermal field, Taupo, New Zealand.
Bull. geol. Surv. N.Z., 75, 1965, 76.
76. Boussinesq J. Th~orie analitique de la chaleur, volume
2 (Paris: Gauthier-Villars, 1903), 172.
77. Wooding R.A. Steady state free thermal convection of
liquid in a saturated permeable medium. J. Fluid. Mech., 2,
1957, 273-85.
78. Nield D. A. Onset of thermohaline convection in a
porous medium. Water Resour. Res., 4, 1968, 553-60.
79. Horton C. W. and Rogers F. T. Jr. Convective currents
in a porous medium. J. appl. Phys., 16, 1945, 367-70.
80. Lapwood E. R. Convection of a fluid in a porous
medium. Proc. Cambridge Phil. Soc., 44, 1948, 508-21.
81. Katto Y. and Masuoka T. Criterion for the onset of
convective flow in a fluid in a porous medium. Int. J. Heat
Mass Transfer, 10, 1967, 297--309.
82. Morrison H. L. Rogers F. T. Jr. and Horton C. W.
Convective currents in porous media, I1. Observation of
conditions at onset of convection. J. appl. Phys., 20, 1949,
1027-9.
69
83. Rogers F. T. Jr. and Morrison H. L. Convection currents
in porous media, II1. Extended theory of the critical
gradient. J. appl. Phys., 21, 1950, 1177-80.
84. Rogers F. T. Jr. Schilberg L. E. and Morrison H. L.
Convection currents in porous media, IV. Remarks on the
theory. J. appl. Phys., 22, 1951, 1476-9.
85. Elder J. W. Steady free convection in a porous medium
heated from below. J. Fluid. Mech., 27, 1967, 29-48.
86. Wooding R. A. Convection in a saturated porous
medium at large Rayleigh number or P~clet number.
J. Fluid Mech., 15, 1963, 527-44.
87. Donaldson I. G. Temperature gradients in the upper
layers of the earth's crust due to convective water flows.
J. geophys. Res., 67, 1962, 3449-59.
88. Hoist P. H. and Aziz K. A theoretical and experimental
study of natural convection in a confined porous medium.
Can. J. chem. Engng, 50, 1972, 232-41.
89. Straus J. M. Large amplitude convection in porous
media. J. Fluid Mech., 64, 1974, 51-63.
90. Mercer J. W. Pinder G. F. and Donaldson I. G. A
Galerkin-finite element analysis of the hydrothermal system
at Wairakei, New Zealand. J. geophys. Res., 80, 1975,
2608-21.
91. Ribando R. J. and Torrance K. E. Natural convection
in a porous medium: effects of confinement, variable
permeability, and thermal boundary conditions. J. Heat
Transfer, 95, 1976, 42-8.
92. Wooding R. A. An experiment on free thermal con-
vection of water in saturated permeable material. J. Fluid
Mech., 3, 1958, 582-600.
93. Bories S. Sur les m~canismes fondamentaux de la
convection naturelle en milieu poreux. Revue g~n. Therm.,
9, 1970, 1377-401.
94. Combarnous M. Convection naturelle et convection
mixte dans une couche poreuse horizontale. Revue g~n.
Therm., 9, 1970, 1355-75.
95. Caltagirone J.-P. Cloupeau M. and Combarnous M.
Convection naturelle fluctuante dans une couche poreuse
horizontale. C.R. S~anc. Acad. Sci. Paris, B273, 1971,833--6
96. Williams D. L. eta/. The Galapagos spreading centre;
lithospheric cooling and hydrothermal circulation.
Geophy~ J. R. astr. $oc., 38, 1974, 587-608.
97. Donaldson I. G. A possible model for hydrothermal
systems and methods of studying such a model. In Proceedings
of the third Australasian conference on hydraulics and
fluid mechanics, 1968 (Sydney: Institution of Engineers,
1970), 200-4.
98. Donaldson I. G. The simulation of geothermal systems
with a simple convective model. Geothermics, special issue
2, 1970, 649-54.
99. Lister C. R. B. On the thermal balance of a mid-ocean
ridge. Geophys. J. R. astr. Soc., 26, 1972, 515-35.
100. Lister C. R. B. On the penetration of water into hot
rock. Geophy~ J.R. astr. Soc., 39, 1974, 465-509.
101. Lowell R. P. Circulation in fractures, hot springs, and
convective heat transport on mid-ocean ridge crests.
Geophy~ J. Ft. astr. Soc., 40, 1975, 351-65.
102. Ribando R.J. Torrance K. E. and Turcotte D. L.
Numerical models of hydrothermal circulation in the
oceanic crust. J. geophys. Res., 81, 1976, 3007-12.
70
103. White D. E. Environments of generation of some base
metal ore deposits. Econ. Geol., 63, 1968, 301-35.
104. Sheppard S. M. F. Nielsen R. L. and Taylor H. P. Jr.
Oxygen and hydrogen isotope ratios of clay minerals from
porphyry copper deposits. Econ. Geol., 64, 1969, 755-77.
105. Sheppard S. M. F. Nielsen R. L. and Taylor H. P. Jr.
Hydrogen and oxygen isotope ratios in minerals from
porphyry copper deposits. Econ. Geol., 66, 1971, 515-42.
106. Henley R.W. Some fluid dynamics and ore genesis.
Tran~ Instn Min. Metall. (Sect. B: Appl. earth sci.), 82,
1973, B1-7.
107. Helgeson H. C. and Kirkham D. H. Theoretical pre-
diction of the thermodynamic behaviour of aqueous electro-
lytes at high pressures and temperatures. I. Summary of the
thermodynamic/electrostatic properties of the solvent. Am.
J. Sci,, 274, 1974, 1089-198.
108. Clark S. P. Handbook of physical constants. Mere.
geol. Soc, Am. 97, 1966, 587 p.
109. Dudziak K. H. and Franck E. U. Messungen der
Viskositat des Wassers bis 560~ und 3500 bars. Ber.
Bunsengesellschaft phy& Chem., 70, 1966, 1120-8.
110. Keenan J. H. eta/, Steam tables (New York: Wiley,
1969).
111. United Kingdom Committee on the Properties of
Steam. U.K. steam tables in S.I. units (London: Edward
Arnold, 1970), 161 p.
112. Liou J. G. Kuniyoshi S. and Ito K. Experimental
studies of the phase relations between greenschist and
amphibolite in a basaltic system. Am. J. Sci., 274, 1974,
613-32.
113. Bear J. Dynamics of fluids in porous media (New
York: American Elsevier, 1972) 764 p.
114. Allen C. R. The petrology of a portion of the Troodos
plutonic complex, Cyprus. Ph.D. thesis, University of
Cambridge, 1976.
115. Helgeson H. C. Thermodynamics of hydrothermal
systems at elevated temperatures and pressures. Am. J.
Sci., 267, 1969, 729-804.
116. Atwater T. and Mudie J. D. Detailed near-bottom
geophysical study of the Gorda Rise. J. geophys. Re~, 78,
1973, 8665-86.
117. Klitgord K. D. eta/. An analysis of near-bottom mag-
netic anomalies: sea-floor spreading and the magnetized
layer. Geophy~. J. Ft. astr. Soc., 43, 1975, 387-424.
118. Needham H. D. and Francheteau J. Some character-
istics of the rift valley in the Atlantic Ocean near 36o48 '
North. Earth planet Sci. Lett., 22, 1974, 29-43.
119. Anderson R. N. and Noltimier H. C. A model for the
horst and graben structure of midocean ridge crests based
upon spreading velocity and basalt delivery to the oceanic
crust. Geophy~ J. R. astr. Soc., 34, 1973, 137--47.
120. Schoell M. and Hartmann M. Detailed temperature
structureof the hot brines in the Atlantis II Deep area
(Red Sea). Mar. Geol., 14, 1973, 1-14.
121. Degens E. T. and Ross D. A. eds. Hot brines and
recent heavy metal deposits in the Red Sea (New York,
etc.: Springer, 1969), 600 p.
122. Hackett J. P. Jr. and Bischoff J. L. New data on the
stratigraphy, extent, and geologic history of the Red Sea
geothermal deposits. Econ. Geol., 68, 1973, 553-64.
123. Trennery T. O. and Pocock B.G. Mining and milling G. C. Amstutz and A. J. Bernard eds (Berlin, etc.: Springer,
operations at Limni mine, Cyprus. Trans. Instn Min. MetalL 1973), 227-245.
(Sect A: Min. industry), 81, 1972, A1-12~ 141. Coleman R. G. and Peterman Z. E. Oceanic plagio-
124. Mottl M.J. Corr R. F. and Holland H. D. Chemical granite. J. geophys. Res., 80, 1975, 1099-108.
exchange between sea water and mid-ocean ridge basalt 142. Veizer J. and Compston W. 87Sr/86Sr composition of
during hydrothermal alteration: an experimental study.
sea water during the Phanerozoic. Geochim. cosmochim.
GeoL Soc. Am. Abstracts with program, 6, 1974, 879-80.
Acta, 38, 1974, 1461-84.
125. Hajash A. Hydrothermal processes along mid-ocean 143. O'Nions R. K. and Gr~nvold K. Petrogenetic relation-
ridges: an experimental investigation. Contrib. Mineral.
ships of acid and basic rocks in Iceland: Sr-isotopes and
PetroL, 53, 1975, 205-26.
rare-earth elements in late and postglacial volcanics. Earth
126. Corliss J. B. The origin of metal-bearing submarine planet. Sci. Lett., 19, 1973, 397-409.
hydrothermal solutions J. geophys. Re~, 76, 1971,
144. O'Nions R. K. and Pankhurst R. J. Secular variations
8128-38.
in the Sr-isotope composition of Icelandic volcanic rocks.
127. Dymond J. etal. Origin of metalliferous sediments Earth planet. Sci. Lett., 21, 1973, 13-21.
from the Pacific Ocean. Bull. geol. Soc. Am., 84, 1973,
145. Hart S. R. LI L- element geochemistry, Leg 34 basalts.
3355--72.
In Initial report of the deep sea drilling project, vol. 34
128. Spooner E. T. C. Sub-sea-floor metamorphism, heat (Washington: U.S. Govt Printing Office, 1976), 283+288 p.
and mass transfer; an additional comment. Contrib.
Mineral. Petrol., 45, 1974, 169-173. 146. O'Nions R. K. and Pankhurst R. J. Petrogenetic
significance of isotope and trace element variations in
129. Robertson A. H. F. and Hudson J. D. Cyprus umbers: volcanic rocks from the Mid Atlantic. J. PetroL, 15, 1974,
chemical precipitates on a Tethyan ocean ridge. Earth 603-34.
planet. Sci. Lett., 18, 1973, 93-101.
147. Subbarao K. V. and Hedge C. E. K, Rb, sr and
130. Robertson A. H. F. and Hudson J. D. Pelagic sedi- 87Sr/86Sr in rocks from the Mid-Indian oceanic ridge.
ments in the Cretaceous and Tertiary history of the Earth planet. Sci. Lett., 18, 1973, 223-28.
Troodos Massif, Cyprus. Spec. Pubis int. Ass. Sediment.,
1, 1974, 403-36. 148. Hart S. R. Glassley W. E. and Karig D. E. Basalts and
sea floor spreading behind the Mariana island arc. Earth
131. Robertson A. H. F. Cyprus umbers; basalt sediment planet. Sci. Lett., 15, 1972, 12-8.
relationships on a Mesozoic ocean ridge. JlgeoL Soc. Lond.,
131, 1975, 511--31.
132. Lambert I. B. and Sato T. The Kuroko and associated
ore deposits of Japan: a review of their features and
metallogenesis. Econ. GeoL, 69, 1974, 1215-36.
133. Ohmoto H. and Rye R.O. Hydrogen and oxygen
isotopic compositions of fluid inclusions in the Kuroko
deposits, Japan. Econ. GeoL, 69, 1974, 947-53.
134. Jehl V. Le m~tamorphisme et les fluides associ~s des
roches oc~aniques de I'Atlantique Nord. Th~se pr~sent~e
I'Universit~ de Nancy pour I'obtention du titre de
Docteur-lng~nieur, Universit~ de Nancy, France, 1975.
135. Honnorez J. eta/. Present day formation of an ex-
halative sulphide deposit at Vulcano (Tyrrhenian Sea),
pt. I1: active crystallization of fumarolic sulfides in the
volcanic sediments of the Baia di Levante. In Ores in
sediments G. C. Amstutz and A. J. Bernard eds (Berlin,
etc.: Springer, 1973), 139-66.
136. Zelenov K. K. Underwater volcanism and its role in
the formation of sediments. Tr. geoL Inst. Akad. Nauk
S.S.S.R. 81, 1963, 30-86. (Russian text).
137. Zelenov K. K. Iron and manganese in exhalations of
the submarine Banu Wuhu volcano (Indonesia). Dokl. Akad.
Nauk SSSR, 155, 1964, 1317-20; DokL Acad. Sci. USSR,
Earth Sci. Sect-, 155, 1965, 94-6.
138. Stanton R. L. and Bass-Becking L. G. M. The forma-
tion and accumulation of sedimentary sulphides in seaboard
volcanic environments. Proc. K. ned. Akad. Wet-, Ser. B.,
65, 1962, 236-43.
139. Ferguson J. and Lambert I. B. Volcanic exhalations
and metal enrichments at Matupi Harbour, New Britain,
T.P.N.G. Econ. GeoL, 67, 1972, 25-37.
140. Pucheldt H. Recent iron sediment formation at the
Kameni islands, Santorini (Greece). In Ores in sediments

71
Origin and emplacement of Society of America's Penrose conference on ophiolites.
This described an ophiolite as an assemblage of basic and
ophiolites ultrabasic rocks, a complete assemblage being as indicated
in Table 1.
I. G. Gass Increasingly, ophiolites are being regarded as slices of
Department of Earth Science, The Open oceanic lithosphere that have been tectonically emplaced as
University, Milton Keynes, Buckinghamshire, thrust masses on to continental margins and island arcs. 1-9
England In Table 2 the environmental and environment indicators of
ophiolites are given. These factors strongly suggest that
ophiolites are submarine. Furthermore, this environment
552.11:552.31 "551.3 must generally be extensional, to explain the presence of
sheeted dyke complexes, and one where large masses of
Synopsis basic magmas can occur close to the surface. There are few
who would question the oceanic setting for ophiolite
By definition, an ophiolite suite is a particular sequence of
genesis. And many, pointing to the need for an extensional
basic and ultrabasic igneous rocks. This sequence seemingly
venue, an abundance of near-surface basaltic magma and the
coincides with that deduced for the oceanic lithosphere
repeated occurrences of the same sequence, see only the
from geophysical and geochemical evidence. This compari-
oceanic lithosphere produced at constructive plate margins
son is reviewed and the ways in which fragments o f ancient as fitting these requirements. 1, 3, 6, 8,10 Fig. 1 gives several
sea-floor may have been emplaced upon continental
margins are discussed. Oceanic
lithosphere
Lo~,er 1 Ophiolites -0
The currently most widely accepted definition of an Layer 2 ;r
ophiolite was that erected in 1972 by the Geological Vp 5.1kin see1 Vourinos, Papua New Troodos,
Greece Caledonia Cyprus -2
Oceanic Basalt
Loyer 3 /
JPiliow Lavas -L
Table 1 Idealized complete ophiolite sequence

Basic voicanics
Vp 6? km sec-1 Gabbro ( ~

Moho
Gobbro +
pyroxenite -
dunite
Basa

Gabbr~ ~
i•,• I ~o~[rt
ISheeted
|Dyke Complex
I~o~o~'~ -6
km
(commonly pillowed) Pyroxenite
Layer L
Merging into Upper -8

Sheeted dyke complex Mantle

Peridotite Peridotite Peridotite i10
VpS.lkm sec-1
( Trondjemites
( Gabbros
Layered cumulates ( Olivine gabbros
( N Peridotites
( * Dunites (with chromite)

Fig. 1 Comparison of various ophiolite sequences with

N Horzburgite (sometimes Iherzotite)
with tectonized fabric that deduced from refraction seismology for oceanic crust.
Modified from Fig. 1, Coleman 5 (T = trondjemite,
Commonly serpentinized. S = serpentinite)

Table 2 Environment and geochemical characteristics of ophiolites and subsequent events affecting them
Rock type Sediments Basic Lovos Sheeted Dyke Cumulate Tectonized
(commonly rodiolorion) (commonly pi(towed) Complex Ptutonics Peridotite

Genetic
en vironmen t
oI-oqo.oo,_,]
mostly bothyo[
marine
Sub-aqueous
Ix,.o
I ,ooo,.LI Passive
plutonic
I J Zone of
deep shearing

I I
JLack of intercalated
I I
JRhythmic layering
I
Deformed and elongated

J J
] sediments indicates and igneous spinel and olivine
Environment High Si needed ldistonce from land Lack of siotic sediment structures indicate past
indicators for rodiotorio
bloom
Imass, rapid extrusion
Jar no land. Little or
.oO:,o'm0"es suggest presence
of torge magma
deformationa[ history.
Absence of tow Ti phases
Jno vesiculority envtronment body suggests partial fusion
lindicates deep water

Geochernica/ Depleted in incompatible elements. Experimenta| data

characteristics suggest magma genesis or equilibration at low pressure
(less than 5 kbar)

Si leached from |JLavas, dyke rocks and basic ptutonics affected by

taros dispersed tJhigh-temperaturelt~ metamorphism in
Jinto overlying 1 zeolite--greenschist--amphibotite facies by reaction
sea water Jbetween basaltic rocks and sea water at elevated
| temperatures
Subsequent
events
Possible origin |- J- Fe, Mn, Cu, Au, Ag and Ni removed
of sutph
Cu massive
des ~ from [ovos by circulating brines,
|precipitated at sea--lava interface
Jos metal-rich sediments

72
well-documented ophiolite sequences and compares them
with that of the oceanic lithosphere deduced largely from
geophysical data.
Detailed investigation of well-established ophiolite
terrains has shown that the igneous rocks involved are
commonly impoverished in the incompatible elements with
large ionic radii, such as K, Rb, Sr, U, Th and the light rare
earths.11,12 Characteristically, a horizontally disposed low-
pressure/high-temperature metamorphism ranging
downwards from the zeolite to the greenschist and
amphibolite facies has been impressed upon the extrusive
and dyke rocks of the assemblage and thermal gradients as
high as 150~ km "1 have been inferred. 13 Stable isotope
studies on these metamorphosed basalts indicate 1 6 0 / 1 8 0
ratios compatible with reactions between the basalt and
sea water at elevated temperatures. 14
Ophiolites as oceanic lithosphere
The idealized sequence given in Table 1 is rarely present,
and the thicknesses of the lava, dyke and layered plutonic
units vary from ophiolite to ophiolite (Fig. 1). 5 There is,
however, a common sequence from top to bottom of lava,
dykes, cumulate plutonics and tectonized peridotite, and it
is the production of this sequence that has to be explained.
The field relationships, and the chemical and mineral
composition of the tectonized peridotite, suggest that it is
material from which a liquid basaltic fraction has been fused
and removed to higher levels. The magma body (or bodies),
some 2 to 3 km deep and perhaps 25 km long, collects at
shallow depth ( 1 - 2 km) below the sea-floor. In an exten-
sional environment magma is tapped off, injected as dykes
and extruded as pillow lavas. At the same time, the magma
remaining within the chamber fractionates on cooling to
give a layered sequence ranging upwards from dunites with
chromite, through peridotites and olivine gabbros to
gabbros and minor Na-rich granites. 3,15
These events produce the ophiolite sequence. They are
quite independent of any interpretation that equates an
ophiolite sequence with ocean floor. Such a sequence does,
however, resemble the oceanic crust in that the composition
of the ophiolite lavas is very similar to that of lavas dredged
from present-day oceanic ridges; both are commonly low-
potassium tholeiites. Similarly, ophiolite lavas, which have
not been intensely deformed or regionally metamorphosed,
display a low-pressure/high-temperature zeolite-greenschist
facies metamorphism that also characterizes volcanics from
the oceanic ridges. 16,17
Apart from the geochemical/petrological data on
dredge specimens, however, the bulk of the information on
the oceanic lithosphere comes from geophysical, particu-
larly seismic refraction, investigations. These are
interpreted in terms of horizontal layering in which layer
1 (Vp = 1.5-3.4 km sec"1 ; average thickness 0.3 km) is
sediments, layers 2 (Vp = 4.2 to 6.0 km sec"1 ; average
thickness 1.7 km) and 3 (Vp = 6.7 km sec"1 ; average
thickness 4.8 km) are the basement and oceanic layers,
respectively, and layer 4 ( Vp = 8.1 km sec"1 ) is the upper
mantle, which is separated from layer 3 by the Mohorovicic
discontinuity. High heat flow is recorded in a narrow belt
over the seismically active axial zone of the oceanic ridges
and the gravity profile over the ridge suggests the existence,
at depths, of low-density mantle. The model of processes
at constructive margins deduced from the geophysical
data is that of mantle material partially fused within a
narrow vertical sheet beneath the axial zone of the oceanic
rises. Magma separates from this hot mantle and ac-
cumulates in the crust, perhaps only 1 - 2 km below the
ocean floor, and is injected presumably as dykes and
extruded as lava flows as ocean-floor spreading takes
place. 18
The models erected to explain the formation of
ophiolites and that deduced from geophysical studies for
processes at constructive margins are so similar that the
observational ophiolite data have been used to clothe the
skeletal geophysical model for present-day constructive
margin processes. The only interpretative step necessary
is the correlation of the peridotitic substratum of ophiolite
complexes with the upper mantle of the oceanic litho-
sphere. The argument is that ophiolites seem to have been
produced by processes similar to those at present-day
constructive margins, and therefore they will probably
continue to be interpreted as oceanic lithosphere until a
better hypothesis is suggested, e Irrespective of these
genetic implications and deductions, however, an
ophiolite is an ophiolite purely on the grounds of the rock
types and sequence it contains. With this proviso strictly
in mind, a tentative correlation of ophiolite structures and
rock types with those of the ocean floor and processes at
constructive margins is given in Fig. 2.
Oceanic [deatized Constructive J
seismic ocean- ftoor/ophiotite p l a t e margin
!.oyers se( uence
La Ye r 2T Dyke
~;lnh~'e{le'~ . . . . . . llI l llr[11I[
i ....
omp,ox [ltilillliHill .... rti
,~ Layer 3 I - [ ~ f a y . . . . . . . ~1'~11 ~'1! )! ember
9 . [.cumulate .... L::-- - -
M o n a - ~uq~o-:-m-:aEc- E-::: -- :__:
o Lo er /.I/:umutozes F_-.Ls :
= t~ectooi
Lper,.otitezed ...................
0 5 10kin
I = = = ' I . . . .
Approximate vertical and horizonto[ scale
Fig. 2 Proposed correlation of ophiolite processes,
structures and rock types with those of ocean floor. After
Greenbaum 3
These are oversimplified models. Detailed studies of
undeformed ophiolites indicate the presence of several
magma chambers, each of relatively small dimensions
(< 10 km in diameter; ca 2 km in depth) overlapping both
laterally and vertically. 15,19 This, in turn, implies erratic
spreading processes. Recently, detailed seismic studies of
the East Pacific Rise north of the Galapagos 2o and the
Mid-Atlantic Ridge (M. Fowler, personal communication)
suggest that discrete magma bodies exist beneath the fast-
spreading (ca 10 cm yr "1 ) East Pacific Rise 20 and no dis-
crete magma body is under the slower (ca 0.8 cm yr "1 ) Mid-
Atlantic Ridge. It could be that discrete magma bodies
only exist beneath fast-spreading ridges and that if spread-
ing is slower than 1.0 cm yr "1, no permanent magma masses
exist.
It is inherent in the correlation of ophiolites with
oceanic lithosphere produced at constructive margins that,
so long as the processes there produce a four-layer
sequence of lavas, dykes, cumulate plutonics and tectonized
peridotite, the location of the margin within the ocean
basin is irrelevant. Indeed, the size of the ocean basin is
irrelevant. Ophiolites could equally well be produced at
major constructive margins such as the Mid-Atlantic Ridge
or East Pacific Rise, in minor 'oceans'such as the Red Sea
or Gulf of Aden, at a behind-arc spreading axis like that
in the Scotia Sea or in marginal seas similar to those of the
73
Western Pacific. By the same reasoning, ophiolites could be descending African plate. 24 The necessary water could
produced near to or far from continents and island arcs have come from dehydration of hydrous minerals in the
from which they can be separated by a destructive plate sediments and igneous rocks of the downgoing slab. The
boundary. Indeed, most, if not all, of the ophiolites in the hydrated mantle converts to serpentinite, in this case a
Alpine-Himalayan Chain seem to have originated in small 10-km diameter diapir, which, as it rose, elevated the over-
ocean basins. These oceans may have been like the Red lying crust. This mechanism for the uplift of the Troodos
Sea and Gulf of Aden, being unassociated with a nearby massif ophiolite not only satisfies geophysical and
descending slab, or, more likely, they were behind-arc structural criteria but can be quantified on geomorpho-
ocean basins forming contemporaneously with the nearby logical evidence. 25 The most elevated part of the massif is
consumption of a descending oceanic plate. 21 So far, the overlain by a mature Pontion (Upper Miocene, ca 7 m.y.)
petrochemical-/mineralogical correlation between some erosion surface. Accepting the reasonable assumption t h a t
ophiolites and some ocean ridge dredge specimens has not this surface was formed at or near sea-level, ca 2100 m of
been entirely satisfactory. 1~ In theory, many of the uplift has occurred in 7 m.y. - a n uplift rate of 0.03 cm yr "1.
discrepancies in geochemistry and mineralogical parageneses Although there is no evidence for continuing subduction
could be more satisfactorily explained if partial fusion of under the island, the presence of a large negative Bouguer
,:
hydrous mantle in a back-arc situation were invoked. In
practice, however, no difference seems to exist between
present day back-arc and open-ocean crustal rocks.
As they migrate away from a constructive plate margin
and become part of a plate, the igneous layers of the ocean
floor are seemingly little affected. Other than the
metamorphic imprint imposed upon them by hydrothermal
reaction between sea water and hot oceanic crust, TM the
principal change is that the water depth increases above the
ocean floor. The increase is attributed to thermal contrac-
tion of the cooling oceanic lithosphere rather than to
significant changes in the thicknesses of the oceanic crust
and immediately underlying oceanic mantle.

Emplacement processes
The precise mechanism(s) whereby part of an oceanic
plate is emplaced on to a continental margin or island arc
is (are) not known. What should normally go down goes
up. The special term o b d u c t i o n , the converse of sub-
duction, has been coined to describe the overall process. 5
The least deformed ophiolite sequences appear to be
vertically uplifted pieces of ocean floor that have not yet
undergone much horizontal deformation. Two such
sequences are Macquarie Island, 23 south of New Zealand,
and the Troodos massif in Cyprus. 1,3,10,19 Most other
ophiolites are much more deformed and their gross
structure is commonly that of a series of complexly
deformed tectonic sheets stacked one on top of the other.
The autochthon is usually continental material. In many
cases olistostromes and other mass flow phenomena
suggest the action of gravity-sliding just before the arrival
of the tectonic sheets at a particular locality. When
reconstructed, the sedimentary facies within the
allochthonous slices indicate that they came from oceanic
crusts near the former continental margins. This, in turn,
implies that obduction occurs at the beginning of destructive
margin processes before the oceanic 'self-destruct'
mechanism of subduction took over.
The processes of ophiolite emplacement include the
fracturing of ocean floor; the sinking or underthrusting of
that floor into the mantle; the uplifting of the ocean floor
which is to be emplaced as ophiolites, its slicing up into
sheets and their emplacement on to the edge of a former Fig. 3 Schematic models illustrating possible mechanism
continent or island arc. Although most of these processes of ophiolite emplacement: (a) Troodos massif, Cyprus
are poorly understood, they probably involve hydration model; uplift of overlying oceanic plate by serpentinite
reactions in the mantle, gravity-sliding and/or tectonic diapir originating from subduction zone; (b) regional uplift
pushing, together with other processes as yet unknown, and due to extensive serpentinization of mantle overlying sub-
are worthy of further speculation. ducting plate; gravity-sliding of uplifted oceanic lithosphere;
The initial stage in the emplacement of the few un- (c) movement of oceanic lithosphere upwards along sub-
deformed ophiolites appears to be vertical uplift. In duction zone inclined away from continental mass;
Cyprus part of the uplift of the Troodos massif seems to (d) elevation of oceanic plate before subduction; gravity-
have been caused by hydration of the mantle overlying the sliding therefrom

74
gravity anomaly means that serpentinite-generated uplift is
not finished. It is unrealistic to extend the one quantified
example to all ophiolites; nevertheless, many masses, such
as the Semail Nappe of the Oman, 26 have not been com-
pressionally deformed during emplacement; some kind of
gravity-slide mechanism must be invoked. If, as work on
Greek ophiolites 27 in particular has indicated, the
ophiolite slices are derived from oceanic lithosphere near
a continental margin, regional uplift of oceanic lithosphere
by hydration of its own peridotitic mantle could produce
the necessary elevation. This mechanism is all the more
realistic now that the possibility of serpentinite poly-
morphs occurring at depths of up to 500 km along
subduction zones has been realized. 28'
Uplift of oceanic lithosphere by serpentinite diapirs, as
in the case of Cyprus, is shown in Fig. 3(a); that of regional
serpentinization in Fig. 3(b). Other tectonic solutions are
also illustrated in Fig. 3. Fig. 3(c) shows a slip zone dipping
away from a continent, rather than toward and underneath
it. Such zones are rare at the present time. A possible
example occurs in New Britain and the Solomon Islands,
where the slip zone dips under the island arcs toward the
Pacific Ocean. In nearby Papua and New Caledonia there
are large exposures of ophiolites emplaced in Tertiary times,
possibly along similarly dipping slip zones. Fig. 3(d) shows a
slip zone dipping underneath a continent. Here the
ophiolites are interpreted as slivers removed from the
descending slab and pushed upward on to the continental
margin. The model has been applied to the Papuan and
Mesozoic Franciscan ophiolites of California.
Conclusions
(1) The term 'ophiolite' is widely used to define a
repeatedly occurring basic-ultrabasic assemblage. It should
be retained in this purely descriptive role, despite the fact
that the assemblage resembles that produced by tectono-
magmatic processes at constructive plate margins.
(2) Most ophiolites so far studied in detail seem to
represent allochthonous fragments of the crust of small
oceans or marginal seas. No convincing representative
from a major ocean has been identified. Petrographic and
geochemical differences between ophiolitic rocks and
suites collected from major oceanic rises are, as yet, un-
explained. More data are required on minor constructive
margins in small oceans. Despite many hypotheses, the
emplacement of ophiolites on to continental margins is not
clearly understood, although serpentinization of mantle
above subduction zones is envisaged as a powerful uplift
mechanism.
(3) If ophiolites are ocean-floor remnants, they are
clearly important in plate tectonic interpretations of the
past. As ocean floor they were created at former
extensional plate margins. As ophiolites emplaced near
compressive plate margins, they mark the proximity of such
margins.
Acknowledgment
This paper is largely based on a previous paper 8 with the
same title written in collaboration with Dr. A. G. Smith and
Professor F, J. Vine. I am most grateful to my former co-
authors for allowing me to use our work in this way.
References
1. Gass I. G. Is the Troodos massif of Cyprus a fragment
of Mesozoic ocean floor? Nature, Lond., 220, 1968, 39-42.
2. Moores E. M. Petrology and structure of the Vourinos
ophiolitic complex of northern Greece. Spec. Pap. geol. Soc.
Am. 118, 1969, 74 p.
3. Greenbaum D. Magmatic processes at ocean ridges: evi-
dence from the Troodos massif, Cyprus. Nature, phys Sci.,
238, 1972, 18--21.
4. Bailey E. H. Blake M. C. Jr. and Jones D. L. On-land
Mesozoic oceanic crust in California coast ranges. Prof. Pap.
U.S. geol. Surv. 700-C, 1970, 70-81.
5. Coleman R. G. Plate zectonic emplacement of upper
mantle peridotites along continental edges. J. geophy& Res.,
76, 1971, 1212-22.
6. Dewey J. F. and Bird J. M. Origin and emplacement of
the ophiolite suite: Appalachian ophiolites in Newfoundland.
J. geophy~ Re~, 76, 1971, 3179-206.
7. Coleman R. G. and Irwin W. P. Ophiolites and ancient
continental margins. In The geology of continental margins
Burk C. A. and Drake C. L. eds (New York: Springer-Verlag,
1974), 921-31.
8. Gass I. G. Smith A. G. and Vine F. J. Origin and em-
placement of ophiolites. In Geodynarnics today (London:
The Royal Society, 1975), 54-64.
9. Hutchinson C. S. Ophiolite in southeast Asia. Bull. geol.
Soc. Am., 86, 1975, 797-806.
10. Gass I. G. etal. Comments on 'The Troodos ophiolitic
complex was probably formed in an island arc', by
A. Miyashiro and subsequent correspondence by A. Hynes
and A. Miyashiro. Earth planet. Sci. Lett,, 25, 1975, 236-8.
11. Pearce J. A. and Cann J. R. Ophiolite origin investi-
gated by discriminant analysis using Ti, Zr and Y. Earth
planet ScL Let~, 12, 1971, 339-49.
12. Smewing J. D. Simonian K. O. and Gass I. G. Meta o
basalts from the Troodos massif, Cyprus: genetic implica-
tion deduced from petrography and trace element geo-
chemistry. Contrib. Mineral. Petrol., 51, 1975, 49-64.
13. Gass I. G. and Smewing J. D. Intrusion, extrusion and
metamorphism at constructive margins: evidence from the
Troodos massif, Cyprus. Nature, Lond., 242, 1973, 26-9.
14. Spooner E. T. C. etal. 018 enriched ophiolitic meta-
basic rocks from E. Liguria (Italy), Pindos (Greece), and
Troodos (Cyprus). Contrib. Mineral. Petrol., 47, 1974,
41-62.
15. Allen C. R. The petrology of a portion of the Troodos
Plutonic Complex, Cyprus. Ph.D. thesis, University of
Cambridge, 1975.
16. Cann J. R. Spilites from the Carlsberg Ridge, Indian
Ocean. J. Petrol., 10, 1969, 1-19.
17. Miyashiro A. Shido F. and Ewing M. Metamorphism
in the Mid-Atlantic Ridge near 24 ~ and 30~ Phil. Trans
R. Soc., A268, 1971, 589-603.
18. Cann J. R. New model for the structure of the ocean
crust. Nature, Lond., 226, 1970, 928-30.
19. Moores E. M. and Vine F. J. The Troodos massif,
Cyprus and other ophiolites as oceanic crust: evaluation and
implications. Phil. Trans R. Soc., A268, 1971, 443-66.
20. Orcutt J. A. Kennett B. L. N. and Dorman L. M.
Structure of the East Pacific Rise from an ocean bottom
seismometer survey. Geophys J. R. Astron. Soc., 45, 1976,
305-20.
21. Smith A. G. The so-called Tethyan ophiolites. In
75
Implications of continental drift to the earth sciences, vol-
ume 2 Tarling D. H. and Runcorn S. K. eds (London, etc.:
Academic Press, 1973), 977-86. (NA TO Adv. Study Inst,
1972)
22. Miyashiro A. The Troodos ophiolitic complex was prob-
ably formed in an island arc. Earth planet Sci. Let~, 19,
1973, 218-24.
23. Varne R. and Rubenach M. J. Geology of Macquarie
Island in relation to tectonic environment. In The Western
Pacific: island arcs, marginal seas, geochemistry Coleman P. J.
ed. (Nedlands: University of Western Australia Press, 1973),
535-41.
24. Gass I. G. and Masson Smith D. The geology and
gravity anomalies of the Troodos massif, Cyprus. Phil. Trans
R. Soc., A255, 1963, 417-67.
25. Vaumas E. de. Further contributions to the geo-
morphology of Cyprus. Ann. Rep. geol. Surv. Cyprus, 1960,
1961, 24-34.
26. Reinhardt B. M. On the genesis and emplacement of
ophiolites in the Oman mountains geosyncline. Schweiz.
miner, petrogr. M i ~ , 49, 1969, 1-30.
27. Hynes A. J. etal. Spreading and emplacement ages of
some ophiolites in the Othris region (eastern central Greece).
Z. dt geol. GeL, 123, 1972, 455--68.
28. Ringwood A. E. Phase transformations in descending
plates and implications for mantle dynamics. Paper pre-
sented at International Union of Geodesy and Geophysics
16th general assembly, 1975, 1A.

76
Hydrothermal alteration of the Field, chemical and mineralogical relationships suggest
that the stockwork zones represent the roots of the massive
basaltic lavas of the Troodos ores and the channels through which the ore-bearing fluids
have ascended to form the massive ores and the sea water-
Ophiolite Complex associated pillow lava interface. The persistent presence of illite and
with the formation of the absence of montmorillonite as a hydrothermai alteration
product in the stockwork zone, combined with experimental
massive sulphide deposits data and studies of the active geothermal areas, suggest that
the temperature of the ore-forming fluids was 250-260~
and the pH 6 . 6 - 6 . 8 . Recent detailed stable isotope studies
suggest that these fluids were mostly heated marine waters,
G. Constantinou
which leached the ore metals from the underlying Sheeted
Geological Survey o f Cyprus, N icosia, Cyprus
Complex.

553.24.065: 553.43(564.3)

Summary
The Cyprus massive sulphide deposits are restricted to and
constitute an integral part of the pillow lava succession of
the Troodos Ophiolite Complex. Extensive geophysical,
geological and geochemical data collected in the last 20 years
suggest that this complex represents an uplifted fragment of
oceanic lithosphere and underlying upper mantle, which
evolved in the Mesozoic at divergent plate boundaries.
The massive sulphide deposits consist of the basaltophilic
elements S, Fe, Cu and, less commonly, Zn. They have been
formed at the original sea water-Lower Pillow Lavas or
Basal Group interface and are characterized by distinct ver-
tical zoning comprising, in downward succession, a zone of
massive ore, a zone of sulphide with silica and the stockwork
zone. The latter consists of veins and disseminations of
pyrite in brecciated and/or pillow lavas, which have under-
gone extensive silicification, chloritization and argillization.
The veins represent fillings of fractures in the lavas by
sulphides, whereas the disseminated pyrite is believed to have
been essentially produced by sulphurization of the titano-
magnetite of the basaltic lavas and it often contains inclusions
of euhedral leucoxene. Part of the disseminated pyrite
formed from the iron present in the glassy mesostasis and the
ferromagnesian minerals of the lavas.
Detailed optical, chemical and X-ray diffraction studies
indicate that the sulphide ore-forming solution produced
extensive mineralogical and chemical changes in the stock-
work lavas. These are distinctly different from the miner-
alogical and chemical changes in the unmineralized lavas pro-
duced by regional hydrothermal metamorphism. Mineralogi-
cal changes in the stockwork lavas include the complete
pseudomorphism of the primary minerals plagioclase and
pyroxene by quartz, the sulphurization of magnetite and the
alteration of the glassy mesostasis to a mixture of chalced-
ony, chlorite, illite and, less commonly, kaolinite.
The chemical changes in the stockwork lavas involved the
dramatic decrease of Ca, Na and, to a lesser extent, AI and K,
and considerable increase of S and Fe. Magnesium decreases
considerably in the intensely mineralized lavas, whereas it
increases in the weakly mineralized lavas. The stockwork
lavas are invariably intensely silicified, but there was no appre-
ciable removal or introduction of silica in these lavas.
Silicification of these lavas was probably produced by the
selective leaching of the bases from the primary minerals and
the glass, leaving behind residual silica in the form of quartz
and chalcedony. In the stockwork lavas Cu, Zn and Co
increased, whereas Ni and Mn decreased in comparison with
the unmineralized lavas.

77
Rare-earth element evidence
for the genesis of the metalliferous
sediments of Troodos, Cyprus*
A. H. F. Robertson
Department of Geology, University of Cambridge,
England
A. J. Fleet
Department of Earth Sciences, The Open
University, Milton Keynes, Buckinghamshire,
England
550.42:546.65:553.068.1:553.43(564.3)
Summary
The origin of metalliferous sediments associated with active
ocean ridges continues to be debated. The various types of
metalliferous sediment that occur on and within the lavas of
the Troodos massif, Cyprus, which is accepted by many
workers as a slice of Mesozoic ocean crust formed at a
constructive plate margin, enable the field relationships of
such sediments to be studied at first hand. Umbers, which
are chemically and mineralogically comparable with the
metalliferous sediments known as ochres occur intimately
associated with the massive sulphide ores of Troodos.
The rare-earth element (REE) contents and, hence,
relative fractionation of samples of these metalliferous
sediments and of some associated lavas and sulphides have
been determined by instrumental neutron activation
analysis in an attempt to identify the sources of these
materials and/or the different igneous and sedimentary
geochemical processes that give rise to these materials. All
the metalliferous sediments are enriched in the REE relative
to chondrites and in the light relative to the heavy REE. In
the case of umbers this is to a greater degree than normal
pelagic clays and the bentonitic clays, which overlie the
umbers and are thought to be continental in origin, are
enriched.
All these sediments also exhibit a marked negative Ce
anomaly, the size of which decreases with the increasing
lithogenous content of the sediments. Such a marked
negative Ce anomaly is characteristic of the fractionation
undergone by the REE in sea water and is a feature of the
East Pacific Rise metalliferous sediments, but contrasts with
the positive anomaly of manganese nodules. The REE
contents of the lavas analysed are difficult to interpret, but
are enriched in La relative to the other REE, probably
because of alteration. Preliminary analyses show that, as
would be expected, the REE contents of the sulphides are

2 I 2

O IS Ipkm
t li AIA~
Upper Pillow Lovos (UP.L)
Lower Pillow Lovos(LP.L)
iooom On-axis extrusions (L.P.L)
i

Off-axis extrusions (U.PL.)

---~ Circulating seo water

Fig. 1 Reconstructed cross-section of spreading Troodos ocean ridge to show postu-

lated sites and conditions of formation of massive sulphides, ochres and umbers: 1, pre-
cipitation of massive sulphides under reducing conditions within median valley slightly
away from areas of lava extrusion; 2, formation of ochres prior to, during and soon after
uplift of median valley to flanks of ridge under redox condition causing iron but not
manganese deposition; block-faulting of Lower Pillow ;_avas; 3, extrusion of Upper
Pillow Lavas on ridge flanks; 4, precipitation of umbers under typical oxidizing oceanic
conditions on further-removed flanks of ridge where Upper Pillow Lava volcanism came
to an end; 5, retrog:ade metamorphism of lavas with widespread deposition of calcite and
pyrolusite in veins

metalliferous sediments of present-day active ocean ridges,
overlie the uppermost lavas of the massif. Beneath the
umbers t,here are extensive zones of veined, brecciated and
chemically altered lavas, which are interpreted as channel-
ways for ascending hydrothermal solutions, and within
which metalliferous sediments occur. Manganese-poor
*Paper published in Earth planet. ScL Letto, 28, 1976, 385--94.
very low (generally less than 1 ppm), and therefore, as yet,
no data comparable with those for the metalliferous
sediments have been obtained.
The REE contents of the Troodos metalliferous
sediments appear to have been incorporated into the
sediments from sea water. This supports the idea that the
major components of the sediments, which are iron and,
in the case of the umbers, some manganese oxides and

78
hydroxides, were derived by the pervasive leaching by sea
water of the igneous rock sequence. The REE contents of
the interstitial sediments imply deep penetration by sea
water far into the pillow lava succession, in agreement with
the results of oxygen and hydrogen isotope analyses.* All
the differing types of metalliferous sediment may be
interpreted as related to differing redox conditions and
events in the evolution of the Troodos ocean ridge (Fig. 1).
In contrast, evidence from present-day ocean ridges and
DSDP investigations indicates that the reducing conditions
necessary to bring about sulphide deposition are not usually
found in present-day median valleys and have not generally
existed in the past; moreover, the redox conditions that lead
to the deposition of iron but not manganese are only found
locally around the mouths of channelways carrying
hydrothermal solutions.t
Finally, it may be noted that the umbers from above the
massive Troodos sulphides apparently have anomalously
high copper contents: such copper-enriched umbers may
therefore be indicators of sulphides at depth.

*Spooner E. T. C. Hydrodynamic model for the origin of the

ophiolitic cupriferous pyrite ore deposits of Cyprus. In Volcanic
processes in ore genesis (London: Institution of Mining and
Metallurgy and Geological Society, 1977), 58--71.
Heaton T. H. E. and Sheppard S. M. F. Hydrogen and oxygen
isotope evidence for sea-water-hydrothermal alteration and ore
deposition, Troodos complex, Cyprus. Op. cir., 42--57.
tCronan D. S. Smith P. A. and Bignell R. D. Modern submarine
hydrothermal mineralization: examplesfrom Santorini and the Red
Sea. Op. cit.. 80.

79
Modern submarine hydrothermal The significance of these observations in regard to metal-
Iogenesis on mid-ocean ridges in general has been discussed
mineralization: examples from elsewhere. 3A~ Metalliferous sediments found to date on
Santorini and the Red Sea the World Mid-Ocean Ridge System are of much lower grade
than those in the Red Sea, but these may, in part, represent
the analogues of the widely dispersed Fe and Mn oxides
around the Atlantis II Deep. 3,4 On this basis, Cronan 3A
D. S. Cronan concluded that higher-grade deposits, including metal-rich
P. A. Smith sulphides, might occur within the upper part of the oceanic
R. D. Bignell crust and within the deep fracture zones that cut mid-ocean
A p p l i e d G e o c h e m i s t r y Research G r o u p , ridges in all the oceans. Furthermore, studies on the geo-
chemical relationship between metalliferous sediments and
D e p a r t m e n t o f Geology, Imperial College, L o n d o n ,
ferromanganese oxides on mid-ocean ridges suggest that the
England
two are genetically related in that they receive their metals
from the same sources, but in different proportions and at
different rates.4,5
Ancient volcanogenic orebodies sometimes show features
5 5 3 . 0 6 8 . 2 (267.5)
similar to those described here, indicating that in the modern
environment we may be observing a not uncommon volcanic
ore-forming process in operation.

Summary Acknowledgment
Submarine hydrothermal mineralization is known to occur The work summarized here was supported by the Natural
on parts of the World Mid-Ocean Ridge System and in Environment Research Council, to which the authors are
association with island arc volcanism. most grateful.
Two areas of submarine hydrothermal activity have been
examined in detail - one in the Atlantis II Deep of the Red References
Sea 1# and the other off the volcano of Santorini in the 1. Bignell R. D. Cronan D. S. and Tooms J. S. Red Sea
Cyclades Volcanic Arc, eastern Mediterranean. 6 metalliferous brine precipitates. In Metallogeny andplate
In each case hydrothermal solutions are mixing with tectonics Strong D. F. ed. Spec. Pap. geol. Ass. Can.
sea water, resulting in the fractionation and precipitation of no. 14, 1976, 147-79.
a variety of metal-bearing phases as the physico-chemical
properties of the solution change. 2. Bignell R. D. Cronan D. S. and Tooms J. S. Metal
Sulphides occur locally as a minor component of the dispersion in the Red Sea as an aid to marine geochemical
newly precipitated hydrothermal sediments off Santorini, exploration. Trans. Instn Min. Metall. (Sect. B: Appl. earth
which are principally composed of iron oxides. Iron pre- sci.), 85, 1976, B274-8.
cipitates close to the hydrothermal outlets and decreases in 3. Cronan D.S. Basal metalliferous sediments from the
concentration away from the volcanic source. By contrast, Eastern Pacific. Bull. geol. $oc. Am., 87, 1976, 928-34.
manganese is low in the sediments richest in iron, but in-
4. Cronan D.S. Riches of the ocean floor. Spectrum
creases in concentration more than twenty-fold away from
no. 137, 1976, 10-2.
the hydrothermal outlets and iron-rich zone. 6 The distribu-
tion of certain minor elements in the sediments follows that 5. Cronan D. S. Manganese nodules and other ferro-
of Mn (e.g. Zn) and Fe (e.g. Mo), indicating that scavenging manganese oxide deposits. In Chemical oceanography,
may be a control on their distribution. volume 5 Riley J. P. and Chester R. eds (New York, etc.:
Early-formed precipitates from the Atlantis II Deep Academic Press, 1976), 217-63.
brines also include sulphides, principally of Cu, Zn and Fe. 6. Smith P. A. and Cronan D. S. The dispersion of metals
Subsequent precipitates include iron silicates, after which associated with an active submarine exhalative deposit. In
the remaining iron is oxidized and precipitated as ferric Proc. Oceanology intn. Conf., Brighton 1975 (London:
oxide and hydroxide. Finally, Mn 2+ is oxidized to Mn 4+ BPS Exhibitions, 1975), 111-4.
and can scavenge considerable quantities of trace metals out
of solution on precipitation. Manganese precipitated in the
waters over the Deep is probably redissolved on falling back
into the brine. By contrast, manganese and those minor
elements that escape from the Deep can precipitate up to
10 km from it, giving rise to well-developed geochemical
dispersion haloes. 1
The selective fractionation and precipitation of hydro-
thermally introduced metals on mixing with sea water in
these two environments results in a geochemical zonation
away from the metal source in each case. The principal
fractionation, other than the early precipitation of metal
sulphides, is between Fe and Mn, but some other metals also
show distinctive concentration gradients. These observa-
tions may be of great potential value in geochemical explor-
ation for submarine hydrothermal orebodies, as the extent
of the dispersion of the most mobile elements can be many
times the extent of the potential orebody itself.

80
Mineralization at destructive plate
boundaries: a brief review
M. S. Garson
Institute of Geological Sciences, London, England
A. H. G. Mitchell
United Nations Development Programme,
Rangoon, Burma
553.061.11:551.3
Synopsis
Three main types of destructive plate boundary are recog-
n i z e d - island arc type, continental margin type and colli-
sion type - each displaying distinctive mineralization.
The first, characterized by an ensimatic island arc mag-
matic belt developed within oceanic crust, has associated
deposits, which include diorite-type porphyry copper and
gold, Kuroko-type zinc-copper-lead sulphides, mercury,
native sulphur-pyrite and Besshi-type cupriferous sulphides.
The second, characterized by a continental margin type
magmatic belt developed within continental crust, includes
associated porphyry copper and molybdenum and, locally,
tin-tungsten deposits. Where comple.te, mineral zonation
away from the plate margin into continental crust com-
prises copper-molybdenum, silver-lead-zinc, tin-tungsten
and antimony. A modification of this type of plate margin
results from the formation of a marginal basin by rifting and
outward migration of incipient continental margin mountain
belts bordered on the ocean side by a subduction zone.
Granite plutons with associated tin, tungsten and fluorite are
emplaced in this setting.
Centres of porphyry copper mineralization may be con-
trolled locally by transform faulting at both types of plate
margin.
Tectonic wedges of oceanic crust and upper mantle are
scraped up with overlying sediments and emplaced above
subduction zones that border magmatic arcs. These wedges
may contain mineral deposits previously formed in ocean-
floor spreading-related settings, including Cyprus-type
massive sulphides, podiform chromite, platinum, magnesite
and asbestos.
The collision type of destructive plate boundary is
formed during and following the final stage of subduction of
ocean floor between two continents, between two island
arcs or between a continent and island arc. Slices of tec-
tonically emplaced oceanic crust and upper mantle, forming
ophiolites, are obducted or thrust on to the continental
crust or inactive island arc on the subducting plate to form
elongate belts at the junction of the two plates. Collision
belts contain mineral deposits, interpreted as having formed
at oceanic ridges and magmatic arcs, and may also include
tin-tungsten, iron-titanium ores associated with anortho-
sites in the lower continental crust, native silver-nickel-
cobalt arsenides, gemstones and strata-bound uranium-
vanadium-copper mineralization in molasse derived from
the associated mountain belts.
The concept by Hess 1 and Dietz 2 of sea-floor spreading
raised the problem of the disposal of the actively growing
oceanic plates. Bullard has pointed out that there are only
two possibilities: either sea-floor is being destroyed or the
earth is expanding. 3 The most widely accepted mechanism
is that of plate destruction along the broad zones of earth-
quakes associated with island arcs. Early papers on sea-
floor spreading suggested that the plates descend vertically,
and it was not until 1967 that it could be shown that the
plates bend down beneath or adjacent to oceanic trenches
and subduct, commonly at angles around 45 ~ beneath
continents and island arcs along seismically defined Benioff
zones. 4 An obvious inference from this was that belts of
igneous rocks aligned parallel to the subduction zones must
in some way be related to processes of melting of subducted
oceanic crust, partial melting of either continental crust or
upper mantle above the subducting plate, or both.
Although it was only a short step from the recognition of
the relationship between various igneous belts and sub-
duction zones, it was not until 1971 that publications
appeared 5-9 with the first attempts to link the genesis of
ore deposits with plate tectonics. More detailed papers
soon followed, in which specific mineral deposits, such as
porphyry copper and stratiform massive sulphide ore-
bodies associated with volcanic rocks, and also circum-
Pacific tin, tungsten and fluorite deposits, were related to
palaeo-Benioff zones; 1~ 11,12 subsequently, there has been
a flood of geological literature on metallogenic processes
at plate boundaries. In general, the mineralization described
in these publications can be grouped into three categories:
mineralization at destructive plate boundaries, m ineraliz-
ation at accreting margins (oceanic rises) and mineralization
postulated to be related to hot spot activity. This paper is
primarily concerned with the first of these categories, but
deposits formed in the other two are commonly involved in
subduction zones and collision belts.
Destructive plate boundaries can be conveniently classi-
fied broadly into three main types - island arc type, conti-
nental margin type and collision type - each with character-
istic associated mineralization. Modifications of the first
two types may occur if marginal basins are developed, and
collision margins or sutures may also vary, depending on
whether the approach and eventual collision of island arc or
continental crust on the subducting crust is with another
island arc or continent on the overriding plate.
Island arc type plate boundaries
Essential elements of island arcs are a belt of calc-alkaline
magmatic rocks, with, in some cases, an outer arc closer to
the subduction zone or trench; the outer arc consists largely
of scraped-up ocean-floor rocks and sediments, and is separ-
ated from the magmatic arc by an outer arc trough. In many
cases older magmatic and outer arc rocks occur beside or
beneath the active volcanic arc.
The mode of formation of different types of island arcs
is still not entirely clear. The most straightforward, termed
ensimatic arcs, developed within the oceans and are under-
lain solely by oceanic crust. Others, termed ensialic arcs,
probably formed originally in a continental margin setting
similar t o that of the present-day Andes, and, after rifting,
migrated away from the continent with the development of
a marginal basin (Fig. 1 (c)); 13 arcs of this type, e.g. Japan,
may include a segment of older continental crust. In the
western Pacific some island arcs have been rifted parallel to
their lengths, and the faulted portions of the arc have then
become detached during the development of an inter-arc
basin analogous to marginal basins between island arcs and
continents (Fig. l(a)).
Both marginal and inter-arc types of back-arc marine
basins are underlain by crust apparently identical to that of
the oceans. Although their development is undoubtedly re-
lated to subduction beneath the arc, less certain is whether
the new crust is invariably created at a spreading ridge
analogous to that of the oceanic basins.
81
 Fig. 1 Schematic cross-sections through destructive plate margins. After Mitchell and
Garson 22

Magmatic belts in present-day island arcs of the Pacific,
the eastern Indian Ocean, the Caribbean and the
Mediterranean are among the most mineralized portions of
the earth's crust (Figs. 1 (a) and 2); ancient arcs now within
continents are also rich in mineral deposits. The type of
mineralization is related in part to the stage of evolution of
the arc, TM and depends partly on whether or not continental
crust is present, 1~, 16 and perhaps also on the position of the
arc relative to the continental margin. 1~
Outer arc troughs lying between outer and magmatic arcs
are important environments for coal and petroleum, but not
for magmatic hydrothermal mineralization.
Porphyry copper deposits
Porphyry copper deposits, often with associated gold and
less commonly with molybdenum, are the most important
forms of mineralization in island arcs. Examples of Tertiary
age, with grades mostly above 0.4% copper, occur in mag-
matic arcs in the Philippines, in Bougainville in the Solomon
Islands, in Taiwan and in the Ryuku and Burma arcs, and in
Puerto Rico.

Fig. 2 Schematic representation of mineralization in an island arc with well-developed

outer arc. After Mitchell and Bell 14

82
 Compared with their cratonic or cordilleran counterparts,
there is generally less potassium in the island arc intrusions,
which comprise mainly quartz-diorites and, less commonly,
some syenites and quartz-monzonites that lack K-feldspar. 18
Island arc intrusions seem also to be depleted in lead and
rubidium and to exhibit 87Sr/86Sr ratios lower than those
measured in intrusives from cratonic settings. 18 Hollister 16
has pointed out, moreover, that the hydrothermal alteration
haloes differ between the cratonic and dioritic island arc
types, the phyllic zone often being absent in the mineralized
diorites.
The island arc dioritic stocks commonly intrude thick
piles of lavas and volcaniclastic rocks that only slightly pre-
date the plutons; a few examples, however, are intruded into
deformed sediments of the outer arc, e.g. Kinabalu in
Sarawak. The cordilleran or continental margin type of
porphyry deposits, in contrast, are associated with stocks
which mostly have penetrated continental crust. 16 Hollister
considered, therefore, that the gold associated with island
arc deposits is derived from lavas and volcaniclastic rocks,
whereas the molybdenum in the cratonic porphyry deposits
has originated in the continental crust. 16 Similarly, Kesler
and co-workers 18 indicated that the island arc intrusions
are compositionally more primitive than their cratonic
counterparts and that an epicrustal source is unlikely for
major or minor constituents of porphyry copper-related
island arc magmas.
The depth of erosion in porphyry deposits has been esti-
mated by Lowell 19 and Sillitoe 20 by comparison of the al-
teration zones and mineral assemblages exposed at the pre-
sent ground level with an assumed model of vertical
zonation. This technique, together with stratigraphic studies,
would suggest that mineralization probably occurred in
island arcs at depths less than 5 km. The occurrence of
many deposits at a high elevation indicates that uplift of the
pre-ore host rocks, during or prior to intrusion, promotes
ore formation, perhaps by facilitating fractionation of the
rising magma. There is some evidence that mineralization
occurs at a late stage in the evolution of a volcanic arc, poss-
ibly during times of increase of convergence of plates, 21
shortly before cessation of volcanism.
Tectonic control of the distribution of some porphyry
orebodies in island arcs is suggested by their location in the
Philippines 22 along or near to branches of a major trans-
current fault of possible transform type. A possible relation-
ship between the initiation of inter-arc spreading and
mineralization has also been noted in the Philippines, 23 but
in other arcs there is little direct evidence for tectonic con-
trol of either location or time of formation of porphyry
copper deposits.
Relics of ancient magmatic arcs now lying within conti-
nents are mainly deeply eroded, and porphyry ores, if they
ever existed, have mostly been removed. In some early
Palaeozoic arcs, however - for example, Coed y Brenin in
North Wales - porphyry-type deposits are preserved, and
recent research 24 suggests that the diorftic host rock at this
locality is associated with an island arc environment similar
to that of porphyry copper deposits in the Caribbean. 25
The Coed y Brenin deposit, however, shows some character-
istics of both the 'diorite' 16 and the 'Lowell and
Guilbert' 26 cratonic types of deposit - indicating the diffi-
culty of rigidly classifying porphyry copper types. 24
Mercury
A link between some types of mercury mineralization and
porphyry deposits has been proposed by White and co-
workers, 27 who considered that mercury could be deposited
in vapour-dominated reservoirs above boiling brine zones in
which porphyry copper ores are formed. Possible examples
of this are cinnabar and quicksilver ores in active or slightly
eroded volcanoes within some of the complex modern island
arcs, such as the Philippines, New Zealand and Japan.
Another possible example is the cinnabar ore at Pinchi-Lake
mercury mine in western Canada, which forms veins in
dolomitic rocks and carbonated serpentinite in the Pinchi
Fauff; 28 the mercury may be derived from a volcanic arc of
Permian or older age in the Pinchi Geoanticline.
Native sulphur-pyrite deposits
Sulphur-pyrite deposits on Honshu Island in Japan are
associated with calc-alkaline volcanic rocks within the
tholeiitic belt of volcanoes. These deposits comprise well-
bedded sedimentary pyrite and associated native sulphur in
adjacent clastic andesitic rocks interpreted as steam explo-
sion breccias. 15 A tectonic control of this type of mineral-
ization is indicated by the location of the major native
sulphur-pyrite deposits of Japan above seismically defined
offsets, or transform faults, in the descendJng oceanic
lithosphere, although not all sulphur deposits are associated
with subduction-related volcanoes.
Sulphur deposits at Vanua Lava in the New Hebrides 22
and at various localities within the calc-alkaline volcanic
belt in the Chilean Andes occur in broadly similar settings.
Kuroko-type zinc-lead-copper stratiform sulphides
Kuroko-type deposits of stratiform massive sulphides,
named from the well known deposits of Miocene age in the
Green Tuff belt of Japan, 29 and associated with rhyolitic
lava domes and volcaniclastic rocks (Figs. 3 and 4) within belts
of calc-alkaline volcanoes, are characteristic of modern island
arcs. Vein-type deposits of copper, lead and zinc occur in
[ f ~ , '~. _.~"
~ ..~ Steam explosion
/';p t ; ~ 1 4 , 1 / a ~ -~
Basement rocks /)'f
13 Uplift of dacite,
formation tuff breccia
,~-~ ,~..~ Steam exptosion.
IT[ ..:" .a "=/."~.~ ~.~.,~ hydrothermol
~~,~'4~;~;~..~ 9,, minerotizotion
i~'-~'~o~q, 4''" .:':~;,-'. . . . . . .: ,. _..
B A Deposition of
I-v ..~ mineral deposits
"~ "'. C
Fig. 3 Schematic figure showing history of development
of a typical Kuroko deposit. After Horikoshi 29
root-zones and stockworks below the sedimentary sulphides,
and also alone within the Kuroko belt. Similar ores of
modern arcs occur in the Philippines and on Vanua Levi in
Fiji, 3~ and at Tirebolu in northern Turkey barytic-
polymetallic stratiform sulphide deposits of late Meso-
zoic to early Caenozoic age in a submarine volcanic environ-
83
Fig. 4 Aerial photograph of Kosaka mining district in northern Honshu, Japan - an area of Kuroko mineralization in an
island arc setting of middle Miocene age.29 Tableland consists of pumice flows from the Quaternary Towada volcano.
1, Dacite lava dome; 2, Motoyama open-pit in first Kuroko deposit, now mined out, but Kosaka mine still produces copper
there by bacterial leaching; 3 and 4, Furutobe and Ainai mines working a group of orebodies; 5, Uchinotai mining off.ices
and plant. Photograph (kindly provided by Professor E. Horikoshi) by courtesy of Dowa Mining Co., Ltd., Japan

ment are probably also Kuroko-type ores. 31
In Japan the Kuroko mineralization occurred towards
the close of a major episode of calc-alkaline volcanism at a
late stage of arc evolution, and was emplaced in a shallow
marine environment within a very narrow stratigraphic
interval of perhaps 0.2 m.y. 32 The deposits are distributed
in a belt about 500 km long and less than 100 km wide
parallel to the present subduction zone. 33 The mineraliz-
ation, dated as late Miocene, coincided with the onset of
tholeiitic volcanism in Honshu, which occurred both in the
belt of mineral-bearing calc-alkaline volcanic rocks and
closer to the trench. The wide distribution of the magma
suggests a regional tectonic control, perhaps by incipient
but aborted opening of a back-arc basin (Figs. l(a) an.d 2).
Horikoshi 34 has suggested that the period of extensional
tectonics related to the opening of the marginal basin of the
Sea of Japan resulted in high heat flow and increased
porosity, which favoured the circulation and storage of
metal-rich thermal waters. These metalliferous brines were
then forced out in a later compressional period ( 1 7 - 1 3 m.y.)
to produce the metallogenic epoch of Kuroko deposits.
Much of the Kuroko belt is underlain by older metamorphic
rocks, which may have influenced the mineralization.
The apparent lack of porphyry deposits in Japan has led
to early speculations that Kuroko ores are genetically re-
lated to porphyry ores and form instead of them when the
sulphides and magma reach the surface. 10~2-35 Kuroko
ores, however, are of submarine origin, whereas porphyry
copper deposits have probably formed at depths of less
than 5 km in elevated volcanic settings. The concept of a
genetic relationship between the two types of deposit has
recently been renewed by Colley 30 in a reclassification of
Kuroko-type deposits based on occurrences in Fiji. Five
types of deposit are proposed (Fig. 5), the first four con-
forming broadly with those of Tatsumi and Watanabe. 36
Colley's type 1 deposit forms within a submarine volcanic
conduit emanating from a diorite pluton with porphyry-style
mineralization, and consists of breccia-pipe and stockwork
mineralization (quartz, pyrite and chalcopyrite). The second
type of mineralization, produced where a volcanic conduit
reaches the sea-bottom on the slopes of the volcanic pile,
consists of siliceous Keiko ore and massive pods of Kuroko
Z n - C u - P b ore and copper-rich Oko ore within pipelike
exhalative centres. Type 3 and 4 deposits are near-horizon-
tal stratiform, chemical precipitate and partial replacement
orebodies in unconsolidated sediment, with some reworking

84
 Emergent

Type]Il
~,iql
strato-volcano
Subaerial/submarine
facies boundary ] Caldera

Sediments ~Volcanic
. rocks
_ ' ~ Lateral
' vent. " .~
dominate dominate r Exhalahve centre J _~.~=.~__--~4"=~

T e m I ".,_:_.~

...... ~ ~ ~ - ~ - ~,o.~'"
......... _ .....+
'~.. !.:~. :~..~:
__
+::,: ~ ~....
-__--- _ ... + +
.p

limestones, volcanictastic rocks

Fig. 5 Schematic diagram by Colley 3o of inter-relationships of five proposed types of

Kuroko deposit (based on occurrences in Fiji), and a proposed link with porphyry-
style mineralization in an island arc environment

downslope progressively further from the volcanic centre. sulphides in Newfoundland (Fig. 6), 41 the Palaeozoic Avoca
Colley's theoretical type 5 deposit, which has not yet deposit in an island arc setting in Eire (Fig. 7), 42 the similar Parys
been found in Fiji, is believed to form distant from the vol- Mountain deposit in Anglesey, North Wales, 43~'4 and the
canic centre by fine ore precipitation, perhaps partly from
a brine-ore solution lighter than sea water, as suggested by
Sato. 37 This would show sedimentary structures and only
minor evidence of volcanic activity. Old examples of this
category, although perhaps not in an island arc setting,
could be the stratiform barite-rich Z n - P b - C u deposit of
Devonian age at Rammelsberg in Germany, 38,39 and also
the strata-bound Z n - C u - P b horizon in Dalradian meta-
sediments at McPhun's Cairn, Loch Fyne, Scotland. 4~ At
Rammelsberg, however, the presence of a nearby volcanic
centre may be suggested by the occurrence, stratigraphi-
cally below parts of the basal pyrite-chalcopyrite-rich por-
tions of the deposit, of siliceous stringer or stockwork ore
known as Kneist, which comprises siliceous and chloritized
mudstone ramified by cracks carrying quartz, carbonate,
pyrite and chalcopyrite.
Probable Kuroko-type ores in ancient island arcs, thrust Fig. 7 Quartz-chlorite stockwork carrying pyrite
on to or incorporated in continental blocks during collision and small amounts of chalcopyrite, sphalerite and
processes, include the Lower Palaeozoic Buchans stratiform galena; this forms a root-zone in altered pyroclastics
below horizon of massive Kuroko-type ore in island arc
setting of Palaeozoic age, Cronebane open-pit, Avoca,
Eire. 42 Photograph, M. S. Garson

Fig. 6 Banded Kuroko-type ore in an island arc setting

of Lower Palaeozoic age, Oriental orebody, Buchans mine,
Newfoundland. 41 Speckled barite layer overlies dark-
coloured chalcopyrite-rich layers, which, in turn, overlie
layers of l ight-coloured chalcopyrite--galena-sphalerite--
barite (Kuroko ore) with distinct sedimentary features.
Photograph, J. G. Thurlow
Fig. 8 Pyritite ore showing graded bedding within
Kuroko-type deposit in an island arc setting of lower
Carboniferous age. Fil6n Norte open-pit, Tharsis,
Spain. 47 Photograph, M.S. Garson

85
stratiform pyrite deposits of Lower Carboniferous age of the
Iberian belt (Fig. 8). 45
The deposit at Anglesey, consisting now largely of sili-
ceous ore carrying disseminated pyrite and chalcopyrite,
has been compared by T. Sato (personal communication)
with the 'chimney-type' deposits in the Uchinotai area. It
is possible that pods of bluestone ore, now worked out,
corresponded to Colley's type 2 deposit, considerably
tectonized and reworked during later orogenies.
The cupriferous pyrite deposits of the Iberian belt in
southern Spain, which include the well known mines of Rio
Tinto 46 and Tharsis 47 (Fig. 8), superficially resemble
Cyprus-type sulphide ores. They are, however, commonly
associated with submarine acid volcanism and display other
features more typical of Kuroko-type ores, such as well-pre-
served stratification, slump structures and localized across-
layer zonation with an increase in copper towards the base of
the deposits, especially where there are underlying cupriferous
stockworks. 45 These deposits occur directly above or close
to underlying feeder stockworks, 48 and are believed by
Williams and co-workers 46 to be Kuroko-type ores that
have formed at a late stage of arc evolution during a period
of relative quiescence within the waning episodes of sub-
marine explosive felsic volcanism.
The copper-poor pyritic stockworks, such as those below
the Planes pyritite deposit, 45 usually occur in sericitized
pyroclastic rocks, whereas the cupriferous pyritic stock-
works, such as those at Cerro Colorado 48 and La Zarza, 47
Fig. 9 Aerial photograph of Vidlin district, Shetland, Scotland - an area of possible Besshi mineralization associated with
metabasic volcanic rocks (V) of Dalradian age.52 S, outcrops of massive sulphide; xl, marble; CS, calc-silicate granulite.
Geophysical anomalies outlining sulphide mineralization shown by heavy pecked line. Photograph by Dennis Coutts
86
and at Avoca in Eire, 22, 42 are typically associated with
zones of more intense silicification and chloritization, which
may occupy discrete zones within older and more wide-
spread pyritic stockworks. This is believed to be due to the
convective circulation of juvenile-meteoric solutions, which
also enriched the lower parts of the stratiform sulphide
bodies. 45
Other possible Kuroko-type deposits, believed to be of
Upper Proterozoic age, occur at Asmara in Ethiopia and
Umm Samiuki in Egypt within an ancient geosynclinal tract
that extends through Ethiopia, Sudan, Saudi Arabia and
Egypt. 49 These deposits consist of Kuroko-style poly-
metallic sulphide ores (with associated barite at Asmara)
within a sequence of pelites, thin limestones, tuffs and
calc-al kaline volcanics.
Besshi-type stratiform massivesulphides
Several stratiform orebodies are associated with intermedi-
ate to basic volcanic rocks, together with carbonaceous
mudstones, clastic carbonates or quartzites, mostly of
fairly deep-water facies, and the depositional environment
is therefore different from that of both oceanic ridge strati-
form ore of Cyprus type and island arc ores of Kuroko
type. In the absence of Caenozoic examples, Mitchell and
Bell 14 have adopted the term Besshi-type deposit used by
Kato 5o to classify cupriferous iron sulphides of Mesozoic
age at Shikoku in Japan. The Shikoku orebodies, and simi-
lar orebodies now within continents, have been interpreted
in terms of early island arc volcanism adjacent to a conti- Mesozoic-Caenozoic orogen
Oceanic Trench Coast
nent, but other settings are possible. A tensional tectonic crust , ~ Fe ,(Au)-Cu~ A g - P b - Z n !Sn-Mo
environment, for example, is favoured by Sawkins, 51 where- 0 Lithosoheric --~ ...... Zone of calc-.'alkqline m.trusive ana
plate r~ovlng away ~ / ...... vo~tc.o.nlc rocks . . . . . . . . Moho--
as Bernard and Soler 6 considered that Besshi deposits from ocean rise ~ ,~ 9 .... ,~
formed in a submarine trench.
~00
A possible Besshi-type stratiform deposit of Dalradian Asthenosphere ~ ~-Oc~e/~ | Or c^
.9 o .~ "~ ,,) c,
age at Vidlin in the Shetland Islands consists of massive
pyrrhotite and some chalcopyrite associated with amphi- 200 ",,',~' et ]"6,., o'e~'

\ "%Z,>,~
bolites which probably represent basic tufts and flows
(Fig. 9). 52 Some of the massive to disseminated polymetallic km
'~ ~ eo;~ l
sulphide deposits of Fe, Cu, Zn and Pb in volcanic and volcano- 300
sedimentary Lower Palaeozoic sequences in the metamorphic
nappes of the central Caledonides of Scandinavia 53,55 also
have affinities with Besshi-type deposits. 400

Gold
500
Gold deposits are found in three main settings in magmatic 0 100 200 300 400 500 600 700
km
arcs.
First, gold mineralization occurs within and at the mar- Fig. 10 Diagrammatic representation by Sillitoe of
gins of quartz-diorite plutons at Guadalcanal in the Solomon generalized mineral zonation in western South America
Islands, and deposits associated with similar intrusions are in context of a plate tectonic-subduction modei 67
known from older arcs. Secondly, gold- and sulphide-
bearing quartz veins occur in thick successions of meta-
andesites and volcaniclastic rocks - for example, in the like to emphasize here that the distribution of certain metals,
Hauraki Peninsula of New Zealand 54 and in the Eastern such as porphyry copper-molybdenum and granitic t i n -
Desert of Egypt. 49 These types of deposit are of interest tungsten, is essentially related to the genesis, differentiation
as primary ores, and, following erosion, as placers. Thirdly, and intrusive history of the source magmas of their charac-
auriferous sulphides and tellurides are mined at Vatukoula teristic host rocks; it is the zonal distribution of these host
in Fiji, in veins associated with a caldera fault cutting andes- rocks in relation to the Benioff zone during their emplace-
ites intruded by a Pliocene monzonite. 56 Deposits of ment that requires explanation as much as the mineral
Tertiary age similar to those at Vatukuola occur in t h e zonation. Moreover, in relating mineral deposits to different
Philippines. depths to the underlying Benioff zone, it is essential to de-
monstrate that the deposits are of the same age within, at
Magmatic arc type plate boundaries at continental margins most, 5 m.y.
Belts of calc-alkaline volcanism either comprising or at the
Porphyry copper deposits
summit of mountain chains near continental margins are
formed, in a similar manner to magmatic belts of island arcs, Porphyry copper deposits are the most abundant manifest-
as a result of tectonic movements, magmatism and crustal ations of mineralization in continental margin type belts,
thickening above Benioff zones (Fig. 1 (b)). The type area and numbers of new porphyry deposits are being prospected
of these mountain belts is taken to be the Andes, although each year in the Andes and North American Cordillera, 25
the pattern there is far from simple because of the complex special attention now being focussed on the Mexican part of
Mesozoic and Caenozoic history, and the modern Andean this belt 59 and southwards in Central America (Fig. 12).
chain is perhaps typical of continental margins only at a Numerous porphyry copper-molybdenum deposits similar
late stage of development. to those in the classic area of the southwestern United
In an analogous fashion to the island arcs, where there is States also occur in ancient continental margin type,
a general variation in igneous composition with distance
from the subduction zone from island arc tholeiite to calc-
alkaline volcanic rocks, within the continental margin mag-
matic belts there is a tendency for an increase in alkalinity
from the margin inwards from tonalite--granodiorite to mica-
granite and alkaline granite, and, finally, locally to under-
saturated rocks, including nepheline-syenite and associated
carbonatite, e.g. Bolivia.
The distribution in the Peruvian Andes of mineral
deposits (iron, gold-copper, lead-zinc-silver and t i n -
tungsten-silver--molybdenum) led Sillitoe '57 to suggest
that mineral zonation is typical of Andean-type belts, and
that the source of the metals in each zone is related to pro-
gressively greater depths to the underlying Benioff zone
iF ig. 10). An alternative explanatio~n, allowing for age differ-
ences in the mineral zones, requires that mineralization is
related to a heat source at approximately constant depth
along the Benioff zone, with lateral migration of this source
beneath crust of varying thickness and perhaps composition Fig. 11 Quartz stockwork with copper sulphides cutting
resulting from changes in inclination of the Benioff zone. 58 porphyry intrusion related to Andean plate margin,
Both these ideas are somewhat speculative, and we would Michiquillay, Peru. 6t Photograph, V. F. Hollister
 , 9
O 0
ADANAC

LI IkBRITISHCOLUMBIA MOLYBDEN
I)- DRY CREEK ~ . _ I aGRANISLE-HUDSON BAY MTN-BERG",~_
PYRAMID
oO - ~I~)ENDAKO
~ , O /BETHLEHEM-VALLEY COPPER I,~
ISLAND CO;TPF~C~
COPPER A ~ I I NBRENDA
RAIGMONT
~PPER
GLACIER K IHEDDLESTON-BIG B
NORTH FORI~I ~BUTTE 0
EAR/Lg
BINGHAM CANYON ~.t
) i ~ URAD-HENDERSON
YERI NGT~h
THAC/
I
BAG ~ Y9 CLIMAX
D 9
SHEEP ~ C MORENCI
SANTA RITA l
ANAN EA , . . - J ' ~ - ~
ACOZ~kI-LA CARID%

UTUADO

ROSITA

#
CHAUCHA U~I"

' MICHIQUILLAY
ANTAMINA
t MOROCOCHA-CASAPALCA
FERRO | ,MB~A~ T' NTAYA
CERRO ERD~IGUELLAVECO
TOQt PALA" ~CUAJONE
CHUQUICAMATA- C6TIC1 9
9 EL SALVADOR-POTk~E

RIO BL ~DA
~ADENI

O PORPHYRYCOPPER-MOLYBDENUM DEPOSITS OR DISTRICTS

Fig.12 Distribution of main porphyry copper deposits in North and South America. After Hollister, is Armstrong and
co-workers 25 and Sillitoe 59

88
volcanic-plutonic belts located on continental or micro-
continental margins in the U.S.S.R.; in contrast to the
deposits of the western America, a large number of these
deposits are of Palaeozoic age.6~
Hollister has neatly summarized the characteristic fea-
tures of the cratonic continental margin or Lowell and
Guilbert type porphyry copper model. 16 This comprises a
porphyry copper deposit associated with a calc-alkaline
pluton, generally lying within the granodiorite-quartz-
monzonite range; hydrothermal alteration zones usually in-
clude a potassic core (orthoclase-biotite) with phyllic
(quartz-sericite-pyrite), argillic (clay-pyrite) and propy-
litic (chlorite--epidote) assemblages arranged in concentric
but incomplete shells around the orebody, and ore minerals"
include copper and molybdenum sulphides. Where the in-
trusion is silica-rich and iron-poor with high sulphur poten-
tial, quartz, sericite and pyrite are developed in abundance,
and there are extensive stockwork patterns of q u a r t z -
sulphide veinlets (Fig. 11 ),61 and pervasive pyrite forms a
typical halo around the centre of copper mineralization. In
contrast, in the dioritic island arc model the sulphides tend to
occur more frequently as disseminations than as veinlets. 16
Other distinctive features were noted briefly in the previous
section.
Porphyry-type deposits associated with quartz-monzonite
stocks of Cambrian-Ordovician and Devonian--
Mississipian age in the Appalachian area are probably related
to a continental margin plate boundary, but they differ in
certain respects from the Lowell and Guilbert model. 26
These differences, believed by Hollister and co-workers 62
to be due to the greater age and depth of erosion of the
Appalachian porphyry-type deposits, include either a reduc-
tion in size or virtual absence of the phyllic zone, the presence
locally of microcline in addition to, or in place of, ortho-
clase in the potassic core, and a less extensive pyrite develop-
ment. The copper mineralization associated with diorites
and porphyry dykes at Tom na Dashan and Fortingall in the
Scottish Highlands 4o may belong to the Appalachian model,
with characteristics roughly intermediate between those of
the island arc and Andean-type (Lowell and Guilbert)
models. It is tempting to suggest that a reason for this
could be that the diorite stocks in Scotland and the
Appalachians were intruded through great thicknesses of
geosynclinal metasediments with associated metavolcanics in
places, and the nature of the upper crust is therefore perhaps
intermediate between that beneath the island arc diorite type
and the Lowell and Guilbert cratonic type.
Difficulties in interpreting North American porphyry de-
posits of different ages located along narrow structural
zones in terms of subduction processes have been discussed
by Lowell. 19 In particular, several copper porphyries in
western North America are located more than 500 km east
of an inferred late Mesozoic continental margin. Suggested
sources for the associated igneous rocks include both a
shallow-dipping Benioff zone and sliding of oceanic litho-
sphere beneath the continent and descent in a sub-
continental subduction zone lying several hundred kilo-
metres east of the continental margin. 63 In the south-
western U.S.A. clusters of porphyry copper deposits form-
ing an elongate belt are considered by Livingston to have
been intruded during movement of the continent relative to
a hot spot in the Laramide orogen. 64
A considerable amount of information is now accumula-
ting to indicate that in many cases there has been a tectonic
control of porphyry deposits in addition to the fundamental
relationship to the palaeo-Benioff zone recently re-
emphasized by Sillitoe. 65 Types of tectonic control include
transverse tectonic zones (transform faults) similar to those
recorded in the Red Sea area, 49 and transcurrent or strike-
slip faults parallel to the subduction zone which were active
or reactivated during the subduction process. In addition,
Sillitoe 66 has suggested that the Andean orogen, and by
anology other comparable orogenic belts above subduction
zones, is divided into tectonic segments by transverse bound-
aries, as in Japan, at which there may be a change in the
characteristics of the belts of magmatic rocks and the type,
age and size of oredeposits.
Examples of porphyry copper deposits associated with
strike-slip faults include the near-linear deposits of the
Chuquicamata--CopaquirE area of Chile, considered by
Hollister and Bernstein to have formed in a fracture system
that penetrated a zone of partial melting associated with a
palaeo-Benioff zone. 67 An even closer tectonic control of
porphyry mineralization has been described in the Highland
Valley, British Columbia, where simultaneous movements on
the Lornex and Highland Valley faults during magmatic
differentiation are believed to have created the tensional
fractures that were filled with both dykes and mineral com-
ponents that now make up the orebodies. 68 The significance
of possible tectonic control of porphyry mineralization has
also been recognized in Iran, where lineaments derived by
satellite and aerial imagery have been used for porphyry
copper exploration. 69
Tin and tungsten deposits
The tin and tungsten mineralization of the Peruvian Andes,
south Bolivia and probably part of the Bolivian Cordillera
Real is of late Tertiary age and its genesis has been related to
a Benioff zone near the position of the present one. 67 The
Miocene tin-bearing Windy Fork granites in Alaska occur in
a broadly similar setting. 7~
The tin mineralization in north Bolivia and Alaska is
associated with granodiorities, mica-granites or alkali granites,
but in south Bolivia, where bismuth and silver are locally
important ores, the host rocks are dacitic quartzoporphy-
ries (Figs. 13 and 14). 71 In Bolivia the deposits are separated
from the belt of porphyry copper deposits and associated
batholiths to the west by the Altiplano plateau, and are under-
lain by very thick continental crust above a shallow-dipping
Fig. 13 Kami dacite stock at edge of Altiplano, west of
Cochabamba, Bolivia, with quartz veins carrying cassiterite
and tungsten, is related to Andean plate margin. Access
to quartz lodes is by means of a precipitous road circling the
mountain. Photograph, M. S. Garson
89
 Benioff zone at a depth of several hundred kilometres. The
Altiplano itself is an arcuate trough infilled by great thickness
.of continental and lacustrine sediments and pierced by
Fig. 14 Orthogonal network of veins carrying pyrite,
cassiterite and boulangerite (dark vein at tip of hammer);
this cuts rhyolite-porphyry in the Andean plate margin
setting, Gardner Vein, Oruro, Bolivia. Photograph,
M. S. Garson
occasional calc-alkaline and alkali granite stocks and volcanic
rocks. It can be interpreted as a broad zone of incipient, but
aborted, marginal basin fracturing.
Back-arc continental margins, back-arc basins and outer arcs
Tin-tungsten-fluorite
In Southeast Asia and along the northwestern Pacific, conti-
nental margins are mostly bordered by island arcs and by
marginal basins of late Mesozoic to Tertiary age. Some of
these margins in East and Southeast Asia exhibit belts of late
Mesozoic to early Tertiary granitic plutons with associated
deposits of tin, tungsten and minor molybdenum and bis-
muth. Fluorite is also commonly present, generally as
quartz-fluorite bodies at granite-limestone contacts. Ex-
amples are the late Mesozoic-early Caenozoic mineralized
granites of peninsular Thailand and Burma, and the South
China tungsten province of late Mesozoic age.
Evidence that island arcs bordering many marginal basins
were previously adjacent to the continent indicates that
they once formed part of incipient continental margin type
mountain belts bordered on the ocean side by a subduction
zone (Fig. l(b) and (c)). The late Mesozoic tin- and
tungsten-bearing granites of Japan and the Asian mainland
were probably emplaced either in these continental margin
magmatic arcs, or during incipient marginal basin develop-
ment. For example, the deposits of eastern China and Japan
were possible related to subduction of an ocean rise, the
Kula-Pacific Ridge, which subsequently resulted in develop-
ment of the Japan Sea 72 and South China marginal basins.
Antimony
Antimony ores occur on the continental side of the t i n -
tungsten belts (Fig. 1(c)) in China, Bolivia and Burma.
These deposits are probably similar in age to the tin-
bearing granites and therefore were probably emplaced on
the continental side of an Andean-type belt possibly prior
to the formation of a marginal basin.
90
Epithermal gold-silver deposits
Vein-type epithermal gold-silver deposits in the Great
Basin of Nevada 73 are possibly related to a back-arc ensialic
environment. The mineralization is mainly of late Miocene
age and coincided with a volcanic episode that comprised
andesites, dacites and rhyolites, this slightly predated the
start of Basin and Range faulting. Isotopic data indicate
that the mineralizing fluids were diluted magmatic brines,
probably related to the associated igneous rocks. 73 Associ-
ation between pre-Caenozoic epithermal gold-silver veins
and back-arc magmatism has not yet been recognized.
Gold deposits in outer arcs and inter-arc troughs
The only typesof mineralization of economic interest
characteristically formed within outer arcs are gold-bearing
quartz veins within deformed flysch-like sediments of some
partly eroded outer arcs (Fig. 2). Although the veins are
generally not extensive enough to comprise orebodies,
erosion may result in minor placer deposits within the outer
arc trough.
Tectonically emplaced ocean-floor deposits
Wedges of oceanic crust and upper mantle that form rocks
of the ophiolite suite are sometimes scraped up with the over-
lying sediments and tectonically emplaced in outer arcs.
These wedges could contain mineral deposits previously
formed in the ocean, e.g. Cyprus-type massive sulphides,
chromite and platinum.
Deposits of Cyprus-type sulphides in this setting occur
at Betts Cove 74 and Ming mine in Newfoundland, and in
the ophiolite complexes of the northern Apennines of
Italy. 75 The large copper orebody in the Semail ophiolite
nappe of Oman is also possibly a Cyprus-type deposit tec-
tonically emplaced during late Cretaceous continental'colli-
sion. Mercury ore associated with magnesite in altered set-
pentinites of the Coast Ranges of California and in the
Ordovician of Newfoundland (R. Stevens, personal com-
munication, 1974) apparently formed in outer arcs, although
whether the serpentinites are ophiolites or deformed mantle
diapirs (Fig. 1(c)) is uncertain.
Ultrabasic rocks, generally interpreted as upper mantle
of the overriding plate, but perhaps including diapiric intru-
sions, are commonly thrust over or within the outer arc
flysch (Figs. 1(c) and 2); examples are the ultramafic bodies
along the eastern edge of the outer Sunda Arc emplaced in
the early Tertiary, and the Lizard and Unst 76 ophiolite
complexes, emplaced in the Palaeozoic. The ultramafic
rocks of Unst in the Shetlands and the outer Sunda Arc in
Burma contain numerous minor podiform chromite deposits
of economic interest.
Source of metals
In magmatic arcs related to subduction processes, the metals
are usually considered to have originated with the magmas,
which are interpreted variously as partial melts from sub-
ducted oceanic crust and sediments, 16,57 from overlying
upper mantle, and from melted lower continental crust, with
or without significant differentiation or contamination by
continental crust.
In the first case, the distribution of porphyry coppers
could be related to the metal distribution in subducting
crust, 77 the mineralization depending on the presence of
metal-rich muds, Cyprus-type massive sulphides, or copper-
rich tholeiites descending along the Benioff zone. Other
possible sources of metals proposed by Laznicka 6o include
old subducted island arcs with massive sulphide deposits,
and local crustal segments anomalously enriched in copper,
such as cupriferous 'red beds', which have reacted with
Benioff zone-generated magmas. The direct physical trans-
fer of material from partially melting, metamorphosed and
dehydrated subducted crust to form magmas 78 and sul-
phides emplaced in the upper crust is, however, debatable,
and neither Sr nor Pb isotope data provide convincing evi-
dence of such recycling.
Many petrologists favour an origin of calc-alkaline mag-
mas by partial melting of upper mantle, or upper mantle
plus some lower crust above the Benioff zone. 79 Lowell 19
has suggested that the Laramide porphyry copper ores of
central Arizona were derived from copper-rich parts of the
underlying continental crust in which the Precambrian
Jerome massive sulphides occur. Extraction of copper from
volcanic rocks intruded by plutons has also been proposed,
but the absence of pre-ore volcanic rocks around many
porphyry deposits makes this hypothesis seem unlikely.
Whatever the origin of the calc-alkaline magmas, there is
perhaps no necessity to postulate that the associated con-
centrations of metals were derived by reworking of original
enrichments either in subducted oceanic crust, or within
continental crust; comparatively inefficient leaching pro-
cesses of either the country rocks or at a late stage of the
magmas themselves can provide ample quantities of metals
to form ore deposits.
Stable isotope 8o and fluid inclusion studies sl indicate
that the initial main ore metal transportation and alteration
agencies in the porphyry copper deposits were highly saline
alkaline brines of magmatic-hydrothermal origin (also see
Sheppard 82). Deposition and additional alteration t o o k
place when the magmatic-hydrothermal brines encount-
ered the lower-temperature convective system of less saline
meteoric-hydrothermal water produced within the surround-
ing country rocks. The base-metal enrichment noted in
volcanic sublimates at Vesuvius and Vulcano 83 and in
Andean volcanoes 11 may be the visible volcanic by-product
of these reactions. Sheppard considers that differences in
deposits are probably in part related to the variations in the
relative proportion of the meteoric-hydrothermal and mag-
matic-hydrothermal events. 82 Isotopic evidence of the
source of the metals is corroborated by data from structural
and petrographic studies of porphyry mineralization at
Highland Valley, British Columbia, which indicate that the
metals were derived from a differentiating magma. 68
In Japan, island arc tholeiite magmas within the Kuroko
ore belt probably resulted from partial melting of upper
mantle, with the addition of water solely from the subduct-
ing crust, 84 but it is unlikely that the felsic magmas associ-
ated with the Kuroko mineralization are derived directly
from subducting crust. 17,85 It is still uncertain if the metals
in the Kuroko ores were leached by brines from underlying
rocks or were of magnlatic-hydrothermal origin. Sato,
however, has pointed out that the lava dome rhyolites with
which Kuroko-type deposits are associated are typically
highly differentiated. 86 He suggests that the metal-rich
fluids are therefore dominantly of magmatic origin and
have separated at the last stage of magmatic differentiation.
Postulations regarding the origin of tin are complicated
by the uncertainty as to whether the tin mineralization
coincides with or post-dates the final stages of crystallization
of the associated granite. It is possible also that many tin
deposits in Southeast Asia, Cornwall and the circum-Pacific
portion of Siberia have been produced during successive
episodes, which have increasingly concentrated tin grades in
the upper parts of the granites. There is considerable evi-
dence that volatiles, and especially fluorine, play a vital role
in this process of cyclical concentration. 1~ The relation
of these episodes to subduction is difficult to determine, as
most tin granites are pre-late Caenozoic, and so their location
cannot be related to present subduction systems; but it
seems likely that a long history of subduction may be
necessary to produce some types of tin deposit.
Subducted oceanic crust has been invoked as a source of
tin for some of the Andean mineralization, 57 but the very
low trace content of tin in oceanic crust and ocean-floor
sediments makes this suggestion seem unlikely. Tin and
tungsten are absent from most island arcs that lack exposed
pre-Mesozoic rocks, which suggests that thick continental
crust, perhaps with a lower layer of high-grade metamorphic
rocks, is necessary for the generation of tin-bearing magmas.
The distribution of fluorite and tin deposits within back-
arc continental margins in Southeast Asia has been related
to a rise of fluorine together with other volatiles from deep
levels of an outwardly migrating Benioff zone during margin-
al basin development. 1~ It is suggested that rising volatiles
could extract tin from the deeper levels of already em =
placed, but possibly still hot, granitic bodies and deposit it
around the upper levels of the intrusions.
Experimental studies of the role of fluorine and chlorine
in late-stage magmatic processes show that in hydrous
granitic melts the alkalis and base metals may concentrate
quite strongly in a CI-bearing aqueous 'vapour' phase,
whereas alkaline-earth elements, Sn, Zr and Nb tend to re-
main in the melt in the presence of F. 8s In terms of
Benioff zone-related magmas, this may support the suggest-
ion that base metals would be concentrated in a CI-rich
magmatic-hydrothermal phase to form porphyry deposits at
a certain distance above a Benioff zone, whereas more dis-
tant from the subduction zone the deposition of tin and
associated metals would be favoured from a F-rich melt. A
possible source of the fluorine and H20 would be phlogo-
pite either in the partially hydrated downgoing lithosphere
slab 89 or at the base of the continental crust.
The source of the metals in the mineralized zones of the
Peruvian Andes, as was mentioned previously, has been re-
lated by Sillitoe 57 to progressively greater depths to the
Benioff zone. This relationship is not strictly analogous to
that between potassium content of late Caenozoic arc rocks
and depths of the Benioff zone, 90,91 as not all the mineral
zones of the Andes are of the same age. Age differences are
allowed for in alternative suggestions that changes in position
and inclination of the Benioff zone have been responsible for
the rock and associated mineral zonation (see p.87). 58
Recently, a zonal distribution of a parallel metamorphic
belt, tungsten belt and molybdenum belt of different ages
has been noted in Japan. Differences in K - A r ages between
the two mineralized belts indicate that separate magmatic
cycles were possibly associated with different subduction
zones. 92
The tin-bearing granites in the Andes, Alaska and eastern
Burma are bordered on the ocean side by tonalitic rocks
with porphyry copper deposits, giving rise to suggestions
that the presence of a thick continental crust is required for
the formation of tin-rich granites, with heat and volatiles
available from a shallow-dipping Benioff zone lying no more
than 200 km beneath the deposits. In central California the
distribution of potassium- and sodium-rich granitic rocks
with initial 87Sr ratios in excess of 0.706 in areas underlain
by Precambrian crust indicates that granites similar in com-
position to many tin-bearing plutons originate at least in
part in the lower crust. 79
Suggestions of primitive concentrations of tin in the
continental crust, 19 or, alternatively, in the upper mantle, 93
have been made to explain tin mineralization in Bolivia and
western North America, respectively, but there is little
91
unequivocal evidence to support these ideas.
Collision-related plate boundaries
Magmatic belts associated with continental collisions
Tin--tungsten deposits The post-tectonic tin- and
tungsten-bearing granites of southwest England have
recently been interpreted as having originated in the under-
thrusting continental plate following continental collision
(Figs. 1 (d) and 15), in a setting analogous to that of the late
Tertiary Malarkachung granite southwest of the Indus
suture in the Himalayas. 94 The southwest England collision
belt extends through the Hercynian tin beff of the
Erzgebirge.95,96 A collision origin has also been invoked
for the late Triassic tin belt of the Main Range in Malaysia. 97
In alternative explanations, however, the southwest
England plutons are interpreted as a magmatic arc related to
northward subduction 98, 99 and as collision-related granites
in an overriding northern plate. 1~176
a lithology that normally carries very low trace amounts of
tin. According to the collision hypothesis, the sediments
were probably deposited largely on oceanic crust, but were
underthrust by continental crust prior to intrusion b'y gran-
ites and so were not present in depth to form a source for
the magma or metals. The granites and tin were probably
derived from the lower continental crust, but it is uncertain
whether the tin was concentrated from trace amounts in the
lower crust through magmatic processes, or whether there
was already a preexisting concentration of tin in the crust
further concentrated by the magma. Minor tin ores do occur
in volcanic host rocks of pre-Variscan age in the Erzgebirge, 1~
but there is no evidence that Variscan tin was derived from
these or sknilar, now unexposed, intrusive rocks.
Native silver-nickel-cobalt arsenide deposits Vein de-
posits of this association had been interpreted by Badham
in terms of explacement in Andean-type magmatic belts. 1~
Some examples, however, including those of Cornwall and
the Erzgebirge, are associated with granites interpreted above

o) Middle Devonian
North Rheic Ocean South
Devon Cornwall
Paratic *Bathya! Gramscatho Magmatic ,,
sediments lull' beds
01
EIO "_=
Upper mantle ,~
30

I(b) Middle Carboniferous

I North South South
I Devon Devon Cornwall
J Paralic Crackington Wildf!ysch Lizard
sediments formation over gramscatho
beds ~ & ~ < ~ thrust
0 ~z~' o x \

12o
30

(C) End Carboniferous

South North South South Engtish

~176
Wales Devon Devon Cornwall Channel
Paralic Bude Tin Lizard
sediments formation granite

o , ,,,"'"

20

Fig. 15 Diagrammatic cross-sections by Mitchell 94 through Hercynian arc systems and

collision belts of southwest England

In Europe several of the post-collision Hercynian granites as collision-related magmas.

are potassic, and have greater quartz contents and lower
albite-orthoclase ratios than granitic rocks of calc-alkaline
magmatic arcs, and were intruded in belts of higher geo-
chemical gradient, lol, lO2 Most of these granites were em-
placed in highly folded and thrusted flysch-type sediments-
Iron-titanium deposits in anorthosites It is considered
by Green lO5 that some large bodies of anorthosite and
layered ultramafic rocks are derived by partial melting of
calc-alkaline rocks at the base of the crust, yielding alkalic
magma and an anorthositic residue. Partial melting could

92
occur above subducting oceanic crust, but more probably
occurs following continental collision, resulting in the em-
placement in the lower crust of these rock types with
associated iron~-titanium orebodies and chromitite sheets.
Deposits of this type are dominantly of Precambrian age,
perhaps because erosion has not yet revealed significant
areas of younger deep crustal rocks.
Nickel-copper mineralization in banded gabbroic and
ultramafic rocks Dewey and Pankhurst considered
that the main orogeny and prolonged accompanying
thermal history of the metamorphism in the Scottish
Highlands can be related to rise of magmas from a basaltic
root above a Benioff zone inclined northwards from the
Ballantrae area. 1~176 The orogeny probably began during
tectonic emplacement in the Arenig of ophiolitic rocks,
remnants of which now occupy faulted wedges along the
later Highland Boundary and Southern Upland Faults; 76 we
suggest that this emplacement could be interpreted in terms
of island arc-continental collision of the Grampian orogeny
following southward subduction. During this period the
'younger' basic intrusions of northeast Scotland were emplac-
ed in hot country rocks; these intrusions were mainly banded
gabbroic and noritic rocks and some peridotites 108 with
associated copper-nickel mineral ization. 109
There are probably many other examples of N i - C u
mineralization in this setting throughout the world, but the
Scottish Highlands example is the only one so far known to
the authors where the host rocks can be dated as having
been intruded concomitantly with the obduction of
ophiolitic rocks.
Gemstone deposits Several deposits of ruby, sapphire and
related precious and semi-precious gemstones are associated
with undersaturated alkaline rocks intruding areas of
carbonates and calc-silicate rocks. Examples include the
major gemstone deposits of Caenozoic age in Pakistan, and
those of probable Caenozoic age in the Mogok area in
Burma, 11~ which are related to intrusions of alaskites and
nepheline-syenites into marbles and charnockitic rocks.
In the Mogok area there are adjacent peridotite plugs from
which precious peridot is obtained.
It is suggested that the alkali magmas in these areas are
related to continental collision, although it is not proposed
that the formation of gemstones is necessarily either res-
tricted to, or characteristic of, collision belts.
Outer and magmatic arcs within continental collision belts
Magmatic and outer arcs, together with their mineral
deposits, may be preserved almost in their original tectonic
setting if subduction beneath them ceases prior to collision
with a continent. The early Devonian Sidlaw volcanic belt
and the Southern Uplands outer arc in Scotland is believed
to comprise an arc system of this type, preserved between
the Scottish Highlands and Baltic Shield continental
masses. 111 The tectonic setting in Scotland appears to be a
mirror image of the Burman orogen, and it is therefore postu-
lated that porphyry copper and gold deposits analogous to
the Monywa porphyry and gold mineralization in the vol-
canic arc of the Central Lowlands of Burma could possibly
be present in the similarly situated Lower Old Red Sand-
stone andesite of the Sidlaw anticline in the Midland Valley
of Scotland.
With continued subduction beneath a volcanic arc until
continental collision occurs, the arc system is underthrust
by the continent on the subducting plate, and the outer arc and,
commonly, the magmatic arc are thrust on to the continent.
Considerable variations in the tectonic and metamorphic
events among collision belts are indicated by differences
among the Caenozoic Alpine, Himalayan, North
Australian-Papuan and Taiwan-China belts. These varia-
tions may be related partly to rate of plate convergence and
partly to irregularities in the colliding continental 112 or
island arc margins.
The most mineralized types of rock emplaced in these
settings and tectonically transported on to the under-
thrusting continent are those of island arc magmatic belts
with associated porphyry copper, Kuroko-type massive
sulphides and gold deposits. Examples of these subduction-
related orebodies, now found in or near to intra-continent
'suture zones', are given above. Outer-am flysch belts with
minor ophiolite wedges may also be emplaced in the same
setting, and these are likely sources for Cyprus-type
massive sulphides, chromite bodies and auriferous quartz
veins.
Obducted ophiolites in collision settings
Slices of oceanic crust and upper mantle can be obducted
on to the subducting continental margin 113,114 during the
final stages of subduction of ocean floor between two conti-
nents, between two island arcs, or between a continent and
island arc (Fig. l(d), (e) and (f)). These slices, forming
ophiolites, either occur in elongate belts at the junction of
the two plates in a suture zone (Fig. l(d)) or are thrust on to
the subducting plate together with the magmatic arc rocks
(Fig. l(f)). The most extensive ophiolites - for example,
the Semail Nappe in Oman 115 and the Troodos Complex
in Cyprus 116 occur where a major outer arc was appar-
_
ently absent.
Late Mesozoic to Caenozoic examples of this type occur
in the Alpine-Himalayan belt extending through the Alps,
Eastern Europe, Turkey, Iran and Pakistan, and the
Himalayas to the Sunda Arc. 117'118 Other Caenozoic ex-
amples occur in New Caledonia and New Guinea, 113 and
examples of older ophiolites, which indicate the site of past
collisions, include the Lizard Complex emplaced in the
Variscan, late Palaeozoic to Permian ophiolites along the
Llralide suture zone in the U.S.S.R., 119 ophiolites at Unst
in the Shetlands, Ballantrae in south Scotland 120 and the
Bay of Islands Complex in Newfoundland 121 obducted
during the Caledonian orogeny, and Proterozoic ophiolites
in the Red Sea region. 49
Like the ophiolitic wedges within outer arcs, ophiolites
at collision margins contain mineral deposits interpreted
as having formed at ocean ridges or within ocean basins.
In addition to the examples of orebodies described above
as having formed in these settings, deposits of ocean-floor
manganese nodules could be emplaced and exposed in
ophiolite belts - for example, the manganese deposits of the
Oman ophiolite complex. High-pressure metamorphic rocks
associated with serpentinites in deeply eroded suture zones
of Caenozoic age and older contain important deposits of
gem-quality jadeite, e.g. north Burma 111 (Fig. l(d)).
Interior of underthrusting continents
Late Caenozoic mineralization is scarce in presently under-
thrusting continents adjacent to the collision belt, such as
India south of the Himalayas, Australia south of New
Guinea and Arabia south of Zagros, perhaps largely because
of the thick sedimentary cover. There is, however, increasing
evidence that it may be favourable for some types of strata-
bound mineralization.
Strata-bound lead-zinc-copper mineralization This type
of mineralization is normally associated with tensional con-
93
ditions related to incipient rifting and opening of oceanic
basins.22,122 Recently, Russell pointed out that strata-
bound lead-zinc-copper mineralization in north-south-
trending geosutures in central Ireland and also local volcanic
activity coincided both with the early opening of the
Atlantic and with the start of collision and southward under-
thrusting of the southern margin of the continent beneath
the latePataeozoic flysch of southwest England. 123 He
considers that hydrothermal systems related to the rift-
associated high thermal gradient resulted in mineralization
only where the continent was depressed below sea level
because of underthrusting along the southern margin.
Strata-bound uranium-vanadium-copper mineralization
Late Caenozoic strata-bound deposits of this type occur in
the southern Himalayan molasse,124,125 a syn-to post-
collision continental facies in the Siwalik Hills directly re-
lated to the uplift and southward thrusting within the
Himalayas on the Indian plate. 118 The uranium is possibly
derived from collision-related granites of this belt to the
north (Fig. l(d). Similar deposition of intra-continental
molasse with associated uranium mineralization occurs in
back-arc ensialic or foreland basins, e.g. Khorat Plateau in
Thailand, 126 but these sedimentary basins are probably
related to back-arc thrusting rather than to collision, and
there are no collision-related potash-rich plutons.
Post-collision magmatic belts
The uranium mineralization of the Rome Province provides
an example of mineralization adjacent to, but younger than,
a collision belt 127 (Fig. l(e). In this area Quaternary
alkaline volcanism was accompanied by uranium mineraliz-
ation in volcaniclastic sediments within grabens adjacent to
trachytic vents. Deposits of lacustrine sediments in a
similar setting carry large contents of very fine-grained
fluorite, 128 also almost certainly associated with this alkali
volcanism. It is believed that the igneous rocks forming the
source of the mineralization were emplaced in a rift-type
zone of tensional faulting parallel to, and possibly related
to, earlier subduction and continental collision, which built
the Apeninnes, but the genetic relationship of these rocks
to plate movements is not yet fully understood.
Other post-collision igneous rocks include the Permian
potassium-rich lavas of southwest England, which were
intruded through the underthrust north European plate
following late Carboniferous collision, but no characteristic
mineral deposits have yet been recognized in this area.
Acknowledgment
The authors are grateful to the Director of the Institute of
Geological Sciences in London and to the United Nations
for permission to publish this paper.
Thanks are also due to Professor D. K. Bailey of Reading
University and Dr. V. F. Hollister of Duva] International
Corporation for useful discussions, respectively, on the role
of fluorine in subduction processes and the relationship of
porphyry copper mineralization to major faults, and to
Dr. T. Sato of the Geological Survey of Japan and
J. G. Thurlow of Memorial University, Newfoundland, for
providing information on Kuroko-type deposits in Japan
and Newfoundland, and for helpful discussions on compari-
sons between these deposits and those at Avoca and Parys
Mountai'n.
Acknowledgment is due to B. Lister, Mineralogy Unit,
Institute of Geological Sciences, for the preparation of
black and white prints from colour transparencies, and to
the Institution of Mining and Metallurgy and the Drawing
94
Office of the Institute of Geological Sciences for the
preparation of text figures.
References
1. Hess H. H. History of ocean basins. In Petrologic studies
- volume in honor ofA. F. Buddington (Washington, D.C.:
Geological Society of America, 1962), 599-620.
2. Dietz R. S. Continent and ocean basin evolution by
spreading of the sea floor. Nature, Lond., 190, 1961, 854-7.
3. Bullard E. The emergence of plate tectonics: a personal
view. Ann. Rev. Earth Planet. Sci., 3, 1975, 1-30.
4. Oliver J. and Isacks B. Deep earthquake zones, anoma-
lous structures in the upper mantle, and the lithosphere.
J. geophys. Res, 72, 1967, 4259-75.
5. Pereira J. and Dixon C. J. Mineralisation and plate tec-
tonics. Miner. Deposita, 6, 1971, 404-5.
6. Bernard A. and Soler E. Sur la Iocalisation g~o-tectonique
des amas pyriteux massifs du type Rio Tinto. C. R. S~anc.
Acad. Sci. Paris, D273, 1971, 1087"90.
7. Brousse R. and Oyarzun J. Les complexes calco-
alkalines et la province cuprifere circumpacifique. In Les
roches plutoniques dans leur rapports avec les g[tes min~raux
(Colloque E. Raguin 1971) Morin P. ed. (Paris: Masson, 1973),
309-14.
8. Guild P.W. Metallogeny: a key to exploration. Min.
Engng, N. Y., 23, Jan. 1971, 69-72.
9. Snelgrove A. K. Metallogeny and the new global tec-
tonics. Bull. Miner. Res. Explor. Inst., Ankara, no. 76,
1971, 130-49.
10. Mitche!l A. H. G. and Garson M.S. Relationship of
porphyry copper and circum-Pacific tin deposits ~o palaeo-
Benioff zones. Trans. Instn Min. Metal/. (Sect. B: Appl.
earth sci.), 81, 1972, B10-25.
11. Sillitoe R. H. A plate tectonic model for the origin of
porphyry copper deposits. Econ. Geol., 67, 1972, 184-97.
12. Sawkins F.J. Sulfide ore deposits in relation to plate
tectonics. J. Geol., 80, 1972, 377-97.
13. Karig D. E. Origin and development of marginal basins
in the western Pacific. J. geophys. Res., 76, 1971, 2542-61.
14. Mitchell A. H. and Bell J. D. Island-arc evolution and
related mineral deposits. J. Geol., 81, 1973, 381-405.
15. Horikoshi E. Development of Late Caenozoic petrogenic
provinces and metallogeny in Northeast Japan. Spec. Pap.
geol. Ass. Can. no. 14, 1976, 121-42.
16. Hollister V. F. An appraisal of the nature and source of
porphyry copper deposits. MineralsSci. Engng, 7, 1975,
225--33.
17. Mitchell A. H. G. Tectonic settings for emplacement of
subduction-related magmas and associated mineral deposits.
Spec. Pap. geoL Ass. Can. no. 14, 1976, 3-21.
18. Kesler S. E. Jones L. M. and Walker R. L. Intrusive
rocks associated with porphyry copper mineralization in
island arc areas. Econ. GeoL, 70, 1975, 515-26.
19. Lowell J. D. Regional characteristics of porphyry
copper deposits of the Southwest. Econ. Geol., 69, 1974,
601--17.
20. Sillitoe R. H. The tops and bottoms of porphyry copper
deposits. Econ. GeoL, 68, 1973, 799-815.
21. Titley S. R. Geological characteristics and environments
of some porphyry copper occurrences in the southwestern
Pacific. Econ. Geol., 70, 1975, 499-514.
22. Mitchell A. H. G. and Garson M.S. Mineralization at
plate boundaries. Minerals Sci. Engng, 8, 1976, 129-69.
23. Wolfe J. A. Tectonic fingerprint in Philippine porphyry
deposits. Society of Mining Engineers, AIME, preprint
73-S-35, 1973, 31 p.
24. Allen P. M. etal. Geochemistry and relationships to
mineralization of some igneous rocks from the Harlech
Dome, Wales. Trans. Instn Min. Metall. (Sect. B: Appl.
earth sci.), 85, 1976, B100-8.
25. Armstrong R. L. Harakal J. E. and Hollister V. F. Age
determination of late Caenozoic porphyry copper deposits of
the North American Cordillera. Trans. Instn Min. Metal/.
(Sect. B: Appl. earth sci.), 85, 1976, B239-44.
26. Lowell J. D. and Guilbert J. M. Lateral and vertical
alteration-mineralization zoning in porphyry ore deposits.
Econ. Geol., 65, 1970, 373-408.
27. White D. E. Muffler L. J. P. and Truesdell A. H. Vapor-
dominated hydrothermal systems compared with hot-water
systems. Econ. Geol., 66, 1971, 75-97.
28. Stevenson J. S. Mercury deposits of British Columbia.
Bull. Dep. Mines Br. Columb. no. 5, 1940, 93 p.
29. Horikoshi E. Volcanic activity related to the formation
of the Kuroko-type deposits in the Kosaka District, Japan.
Miner. Deposita, 4, 1969, 321-45.
30. Colley H. Classification and exploration guide for
Kuroko-type deposits based on occurrences in Fiji. Tran~
Instn Min. Metall. (Sect B: Appl. earth sci.), 85, 1976,
B190-9.
31. Egin D. Mineralization in the vicinity of the Harsit-
Koprubasi Mine, Tirebolu, Turkey. Paper presented to
Mineral Deposits Study Group, Session 6: Stratiform mineral
deposits, Leicester, Dec. 1975.
32. Ueno H. Duration of the Kuroko mineralization episode.
Nature, Lond., 253, 1975, 428-9.
33. Tatsumi T. Takagi Y. and Otagaki T. Geology of the
Kuroko deposits. Joint Meeting MMIJ--AIME 1972, pre-
print T1 bl, 13 p.
34. Horikoshi E. Metallogenic epoch of Kuroko-stage
deposits in the northeast Japan arc. In press.
35. Sillitoe R. H. Formation of certain massive sulphide
deposits at sites of sea-floor spreading. Trans. Instn Min.
Metall. (Sect- B: Appl. earth sci.), 81, 1972, B141-8.
36. Tatsumi T. and Watanabe T. Geological environment
of formation of the Kuroko-type deposits. In Proc. I M A -
IAGOD Meetings "70, IAGOD volume Takeuchi Y. ed.
(Tokyo: Society of Mining Geologists of Japan, 1971),
216-20. (Spec. Issue no. 3).
37. Sato T. Behaviour of ore-forming solutions in sea
water. Min. Geol. (Tokyo), 22, 1972, 31-42. (Japanese
text)
38. Bornhardt W. Die Entstehung des Rammelsberger
Erzvorkommens. Arch. Lagersta"ttForsch. , 38, Preussische
Geologische Landesanstalt no. 4, 1939, 61 p.
39. Kraume E. Stratigraphie und Tektonik der Rammels-
berger Erzlager unter besonderer Berucksichtigung des
neuen Lagers unter der 10. Sohle. Z. Erzbergb. Metal/-
h~ttenw., 13, 1960, 7-12.
40. Wilson G. V. and Flett J. S. The lead, zinc, copper and
nickel ores of Scotland. Spec. Rep. Mineral Resources
GtBr., 17, 1921, p. 76.
41. Thurlow J. G. Swanson E. A. and Strong D. F.
Geology and lithogeochemistry of the Buchans polymetallic
sulfide deposits, Newfoundland. Econ. Geol., 70, 1975,
130-44.
42. Platt J. W. Volcanogenic mineralization at Avoca, Co.
Wicklow, Ireland, and its regional implications. In Volcanic
processes in ore genesis (London: Institution of Mining and
Metallurgy and Geological Society, 1977), 163-70.
43. Wheatley C. J. V. Aspects of metallogenesis within the
Southern Caledonides of Great Britain and Ireland. Trans.
Instn Min. Metal/. (Sect. B: Appl. earth sci.), 80, 1971,
B211-23.
44. Ixer R. A. and Gaskarth J.W. Parys Mountain - a
possible Kuroko-style deposit. Paper presented to Mineral
Deposits Study Group, Session 6: Stratiform mineral
deposits, Leicester, Dec. 1975.
45. Williams D. Stanton R. L. and Rambaud F. The
Planes-San Antonio pyritic deposit of Rio Tinto, Spain:
its nature, environment and genesis. Tran~ Instn Min.
Metall. (SecL B: Appl. earth sci.), 84, 1975, B73-82. (See
also summary in Volcanic processes in ore genesis (London:
Institution of Mining and Metallurgy and Geological Society,
i977), 152.
46. Rambaud Perez J. El sinclinal carbonrfero de Rio Tinto
(Huelva) y sus mineralizaciones asociadas. Mere. Inst- geol.
min. Esp., 71, 1969, 253 p.
47. Strauss G. K. and Madel J. Geology of massive sulphide
deposits in the Spanish-Portuguese pyrite belt. Geol. Rdsch.,
63, 1974, 191-211.
48. Pryor R. N. Rhoden H. N. and Villal6n M. Sampling of
Cerro Colorado, Rio Tinto, Spain. Trans. Instn Min. Metall.
(Sect A: Min. industry), 81, 1972, A143-59.
49. Garson M. S. and Shalaby I. M. Precambrian-Lower
Paleozoic plate tectonics and metallogenesis in the Red Sea
region. Spec. Pap. geol. Ass. Can. no. 14, 1976, 573-96.
50, Kato T. Geology of ore deposits (Tokyo: Fuzambo,
1937). (Japanese text).
51. Sawkins F.J. Massive sulphide deposits in relation to
geotectonics. Spec. Pap. geol. Ass. Can. no. 14, 1976,
221-40.
52. Garson M. S. and May F. Copper mineralization at
Vidlin, Shetland. Tran~ Instn Min. Metall. (Sect. B: Appl.
earth sci.), 85, 1976, B 153--7.
53. Vokes F. M. Caledonian massive sulphide deposits in
Scandinavia. I n Handbook of strata-bound and stratiform
ore deposits, Vol. 6: Cu, Zn, Pb, and Ag deposits
Wolf K. H. ed. (Amsterdam, etc.: Elsevier, 1976), 79-127.
54. Denholm L.S. Geological exploration for gold in the
Tavua Basin, Viti Levu, Fiji. N.Z.J. Geol. Geophys., 10,
1967, 1185-6.
55. Vokes F. M. and Gale G. H. Metallogeny relatable
to global tectonics in southern Scandinavia. Spec. Pap.
geol. Ass. Can. no. 14, 1976, 413-41.
56. Robinson B.W. The origin of mineralization at the Tui
mine, Te Aroha, New Zealand, in the light of stable isotope
studies. Econ. Geol., 69, 1974, 910-25.
57. Sillitoe R. H. Relation of metal provinces in western
95
America to subduction of oceanic lithosphere. Bull. geol.
Soc. Am., 83, 1972, 813-7.
58. Mitchell A. H.G. Metallogenic belts and angle of dip of
Benioff zones. Nature, Phy~ Sci., 245, 1973, 49-52.
59. Sillitoe R. H. A reconnaissance of the Mexican por-
phyry copper belt. Trans. Instn Min. Metal/. (Sect. B: Appl.
earth sci.), 85, 1976, B170-89.
60. Laznicka P. Porphyry copper and molybdenum deposits
of the U.S.S.R. and their plate tectonic settings. Trans. Instn
Min. Metall. (Sect- B: Appl. earth sci.), 85, 1976, B14-32.
61. Hollister V. F. and Sirvas E. El porfido de cobre
Michiquillay. Boln Soc. geol. Peru, 44, 1974, 11-27.
62. Hollister V. F. Potter R. R. and Barker A. L. Porphyry-
type deposits of the Appalachian orogen. Econ. Geol., 69,
1974, 618-30.
63. Lipman P.W. Prostka H. J. and Christiansen R. L.
Cenozoic volcanism and plate-tectonic evolution of the
western United States. Phil. Trans. R. Soc., A271, 1972,
217-48.
64. Livingston D. E. A plate tectonic hypothesis for the
genesis of porphyry copper deposits of the southern Basin
and Range Province. Earth planet- Sci. Let'L, 20, 1973,
171--9.
:65. Sillitoe R. H. Subduction and porphyry copper deposits
in south-western North America - a reply to recent object-
ions. Econ. Geol., 70, 1975, 1474-7.
66. Sillitoe R. H. Tectonic segmentation of the Andes:
implications for magmatism and metallogeny. Nature, Lond.,
250, 1974, 542-5.
67. Hollister V. F. and Berstein M. Copaquire, Chile: its
geologic setting and porphyry copper deposit. Trans. (Soc.
Min. Engrs) Am. Inst- Min. Engrs, 258, 1975, 137-42.
68. Ho!lister V. F. etal. Structural evolution of porphyry
mineralization at Highland Valley, British Columbia. Can. J.
Earth Sci., 12, 1975, 807-20.
69. Forster H. Lineaments derived by satellite and aerial
imagery for porphyry copper exploration in central I ran.
Paper presented to Meeting of European Geological Societies,
Reading University, Sept. 1975, in press.
70. Reed B. L. and Lanphere M.A. Alaska-Aleutian Range
batholith: geochronology, chemistry and relation to
circum-Pacific plutonism. Bull. geol. Soc. Am., 84, 1973,
2583-610.
71. Turneaure F. S. The Bolivian tin-silver province. Econ.
Geol., 66, 1971,215-25.
72. Uyeda S. and Miyashiro A. Plate tectonics and the
Japanese islands: a synthesis. Bull. geol. Soc. Am., 85, 1974,
1159-70.
73. O'Neil J. R. and Silberman M. L. Stable isotope
relations in epithermal A u - A g deposits. Econ. Geol., 69,
1974, 902-9.
74. Upadhyay H. D. and Strong D. F. Geological setting of
the Betts Cove copper deposits, Newfoundland: an example
of ophiolite sulfide mineralization. Econ. Geol., 68, 1973,
161-7.
75. Bonatti E. etal. Metalliferous deposits from the
Apennine ophiolites: Mesozoic equivalents of modern
deposit from oceanic spreading centers. Bull. geol. Soc. Am.,
87, 1976, 83-94.
76. Garson M. S. and Plant J. Alpine type ultramafic rocks
96
and episodic mountain building in the Scottish Highlands.
Nature, Phy~ Sci., 242, 1973, 34-8.
77. Corliss J. B. Sea water, seafloor spreading, subduction
and ore deposits. Spec. Pap. geol. Ass. Can. no. 14, 1976,
145-6. (Abstract).
78. Stern C. R. Melting products of olivine tholeiite basalt
in subduction zones. Geology, 2, 1974, 227-30.
79. Kistler R. W. and Peterman Z. E. Variations in St, Rb,
K, Na and initial Sr87/Sr86 in Mesozoic granitic rocks and
intruded wall rocks in Central California. Bull. geol. Soc.
Am., 84, 1973, 3489-512.
80. Sheppard S. M. F. Nielsen R. L. and Taylor H. P. Jr.
Oxygen and hydrogen isotope ratios of clay minerals from
porphyry copper deposits. Econ. Geol., 64, 1969, 755-77.
81. Roedder E. Fluid inclusion studies on the porphyry-
type ore deposits at Bingham, Utah, Butte, Montana, and
Climax, Colorado. Econ. Geol., 66, 1971,98-120.
82. Sheppard S. M. F. Identification of the origin of ore-
forming solutions by the use of stable isotopes. In Volcanic
processes in ore genesis (London: Institution of Mining and
Metallurgy and Geological Society, 1977), 25-41.
83. Angus J. G. and Davis G. R. Base metal enrichment in
volcanic sublimates and secondary alteration products from
Vesuvius and Vulcano. Mineralog. Mag., 40, 1976, 481-6.
84. Ringwood A. E. The petrological evolution of island
arc systems. JI geol. Soc. Lond., 130, 1974, 183-204.
85. Lambert I. B. and Sato T. The Kuroko and associated
ore deposits of Japan: a review of their features and metallo-
genesis. Econ. Geol., 69, 1974, 1215-36.
86. Sato T. Characteristics, origins and classifications of
submarine exhalative-sedimentary ore deposits. 25th Int.
Geol. Congr. Sydney 1976, in press.
87. Barsukov V. L. and Kuril'chikova G. Y. On the forms
in which tin is transported in hydrothermal solutions.
Geokhimiya, 1966, 943-8; Geochem. Int., 3, 1966, 759-
64.
88. Hards N. J. The role of fluorine and chlorine in late-
stage magmatic processes: an experimental study. Mineralog.
Mag., in press.
89. Wyllie P.J. The dynamic earth (New York: Wiley,
1972), 416 p.
90. Kuno H. Lateral variation of basalt magma type across
continental margins and island arcs. Bull. volcan., 29, 1966,
195-222.
91. Hatherton T. and Dickinson W. R. The relationship
between andesitic volcanism and seismicity in Indonesia, the
Lesser Antilles and other island arcs. J. geophys. Res., 74,
1969, 5301-10.
92. Shibata K. and Ishihara S. K - A r ages of the major
tungsten and molybdenum deposits in Japan. Econ. Geol.,
69, 1974, 1207-14.
93. Noble J. A. Metal provinces in the western United
States. Bull. geol. Soc. Am., 81, 1970, 1607--24.
94. Mitchell A. H. G. Southwest England granites: mag-
matism and tin mineralization in a post-collision setting.
Tran~ Instn Min. Metall. (Sect B: Appl. earth sci.), 83,
1974, B95-7.
95. Gabelman J. W. and Krusiewski S. V. The metallo-
tectonics of Europe. In 24th Int. Geol. Congr. (Montreal:
The Congress, 1972), Sect. 4, 88-97.
96. Lorenz V. Formation of Hercynian subplates, possible
causes and consequences. Nature, Lond., 262, 1976,
374-7.
97. Mitchell A. H. G. Tectonic settings for emplacement
of tin-bearing granitic rocks in the Burma-Thai-Malaya
region. IGCP Circum-Pacific Plutonism Project, 5th
Meeting, Kuala Lumpur, Nov. 1975. (Abstract).
98. Badham J. P. N. and Halls C. Microplate tectonics,
oblique co!lisions and evolution of the Hercynian orogenic
systems. Geology, 3, 1975, 373-6.
99. Bromley A. V. Tin mineralization of Western Europe:
is it related to crustal subduction? Trans. Instn Min. Metall.
(Sect. B: Appl. earth sci.), 84, 1975, B28-30.
100. Riding R. Model of the Hercynian foldbelt. Earth
planet. Sci. Lett., 24, 1974, 125-35.
101. Hall A. Geothermal control of granite compositions in
the Variscan orogenic belt. Nature, Phys. Sci., 242, 1973,
72-5.
102. Brown G. C. Evolution of granite magmas at destruc-
tive plate margins. Nature, Phys. Sci., 241, 1973, 26-8.
103. Baumann L. Tin deposits of the Erzgebirge. Trans.
Instn Min. Metall. (Sect. B: Appl. earth sci.), 79, 1970,
B68-75.
104. Badham J. P. N. Orogenesis and metallogenesis with
reference to the silver-nickel, cobalt arsenide ore association.
Spec. Pap. geol. Ass. Can. no. 14, 1976, 559-71.
105. Green T. H. High-pressure experimental studies on
the origin of anorthosite. Can. J. Earth Sci., 6, 1969, 4 2 7 -
40.
106. Dewey J. F. Evolution of the Appalachian/
Caledonian orogen. Nature, Londo, 222, 1969, 124-9.
107. Dewey J. F. and Pankhurst R. J. The evolution of
the Scottish Caledonides in relation to their isotopic age
pattern. Trans. R. Soc. Edin., 68, 1970, 361-89.
108. Stewart F. H. The 'younger' basic igneous complexes
of north-east Scotland, and their metamorphic envelope.
Scottish Jo Geol., 6, 1970, 3 - 6 .
109. Dunham Sir Kingsley. Geological setting of the useful
minerals in Britain. Proc. R. Soc., A339, 1974, 273-88.
110. Searle D. L. and Haq B.T. The Mogok Belt of Burma
and its relationship to the Himalayan orogeny. In 22nd
Int, Geol. Congr. (New Delhi: The Congress, 1964), part XI,
132-61.
111. Mitchell A. H. G. and McKerrow W. S. Analogous
evolution of the Burma orogen and the Scottish Caledonides.
Bull. geol. Soc. Am., 86, 1975, 305-15.
112. Oxburgh E. R. Flake tectonics and continental
collision. Nature, Lond., 239, 1972, 202.
113. Coleman R. G. Plate tectonic emplacement of upper
mantle peridotites along continental edges. J. geophy& Res.,
76, 1971, 1212-22.
114. Dewey J. F. and Bird J. M. Origin and emplacement
of the ophiol ite suite: Appalach Jan ophiolites in Newfound-
land. J. geophys. Res., 76, 1971, 3179--206.
115. Welland M. J. P. The Oman ophiolite - some critical
stratigraphic and structural relationships and a possible
method of emplacement. Geol. Ass. Can./Min. Ass. Can.
Joint Annual Meeting. St. Johns, Newfoundland, 1974.
(Abstract).
116. Gass I.G. Is the Troodos Massif of Cyprus a fragment
of Mesozoic ocean floor? Nature, Lond., 220, 1968, 39-42.
117. Dixon C. J. and Pereira J. Plate tectonics and mineral-
ization in the Tethyan region. Miner. Deposita, 9, 1974,
185-98.
118. Powell C. M. and Conaghan P. J. Plate tectonics and
the Himalayas. Earth planet. Sci. Lett., 20, 1973, 1-12.
119. Hamilton W. The Uralides and the motion of the
Russian and Siberian platforms. Bull. geol. Soc. Am., 81,
1970, 2553-76.
120. Church W. R. and Gayer R. A. The Baliantrae
ophiolite. Geol. Mag., 110, 1973, 497-510.
121. Stevens R. K. Cambro-Ordovician flysch sedimenta-
tion and tectonics in west Newfoundland and their possible
bearing on a proto-Atlantic ocean. Spec. Pap. geol. Ass. Can.
no. 7, 1970, 165-77.
122. Russell M. J. Structural controls of base metal mineral-
ization in Ireland in relation to continental drift. Trans.
Instn Min. Metall. (Sect. B: Appl. earth sci.), 77, 1968,
B 117-28.
i23. Russell M. J. Incipient plate separation and possible
related mineralization in lands bordering the North
Atlantic. Spec. Pap. geol. Ass. Can. no. 14, 1976, 339-49.
124. Moghal M. Y. Uranium in Siwalik Sandstones,
Sulaiman Range, Pakistan. In Formation of uranium ore
deposits (Vienna: International Atomic Energy Agency,
1974), 383-403.
125. Basham I. R. and Rice C. M. Uranium mineralization
in Siwalik Sandstones from Pakistan. In Formation of
uranium ore deposits (Vienna: International Atomic
Energy Agency, 1974), 405-18.
126. Shawe D. R. Hite R. J. and Inthuputi B. Potential
for sandstone-type uranium deposits in Jurassic rocks, Khorat
Plateau, Thailand. Econ. Geol., 70, 1975, 538-41.
127. Locardi E. Uranium deposits in alkaline volcanites of
Central Italy. In Volcanism and metailogenesis, Proc. Symp.
Bucharest, 1972, in press.
128. Spada A. II giacimento di fluorite e baritina esalativo-
sedimentario in facies lacustre, intercalato nei sedimenti
piroclastici della zona di Castel Giuliano, in prov. di Roma.
Industria rain., Roma, 20, 1969, 5 0 i - 1 8 .
97
Porphyry copper deposits

J. P. Hunt
P. O. Box 954, La Jolla, California, U.S.A.

553.067:553.43

Summary

Porphyry copper deposits are large crustal sulphur anomalies

containing copper and other metals. Much of the sulphur
is oxidized and, along with a major portion of the copper,
deposited from fluids evolved from certain calc-alkaline
melts of intermediate composition. These melts are em-
placed with characteristic textures as dykes and stocks in
and above the cupolas of small granitic batholiths. The
depth of emplacement is shallow, and the intrusions are
comagmatic with overlying andesite-rhyolite volcanics.
Intrusion and mineralization are relatively rapid events and
occur towards the end of previous episodes of vulcanism
and intrusion. Porphyry copper deposits are clustered in
both space and time on regional, continental and global
scales. More may have formed in the past 65 million years
than in earlier geologic periods, especially in certain appar-
ently favourable continental regions. Sources of sulphur,
copper and of the associated magmas are unknown. Much
additional mapping of continental and oceanic margins is
needed to explain the abnormal, or unusual, coincidence of
several igneous and tectonic processes probably responsible
for the formation of porphyry copper deposits.

98
Metallic mineralization affiliated to
subaerial volcanism: a review
R. H. Sillitoe
Department of Mining Geology, Royal School of
Mines, Imperial College, London, England
553.067:553.3/.4
Synopsis
A considerable number of metallic ore deposits are genetic-
ally related to subaerial volcanism. Here four ore types are
reviewed: epithermal precious- and base-metal deposits are
usually o f vein type, areassociated with either propylitic
(+ adularia) or advanced argillic alteration, and cut mid- to
late Caenozoic volcanic sequences, commonly of andesitic
composition. Examples from western Mexico, central
America, the central Andes, the western U.S.A., Japan, the
North Island of New Zealand, the Philippines, Fiji and Spain
are described. They are commonly generated less than 3 m.y.
after their host volcanic rocks, and are localized at depths
shallower than 1000 m by pre-existing faults; in the San
Juan Mountains, Colorado, in particular, earlier fractures re-
lated to caldera collapse were utilized. Epithermal ores
were deposited from low-temperature (200-300 ~C), low-
salinity (normally < 12 equivalent wt % NaCI for base-metal
ores and < 2 equiv, wt % NaCI for precious-metal ores)
meteoric water, but isolated ~D values suggest a magmatic
contribution to the ore fluid. Various lines of evidence sug-
gest subjacent stocks as a principal, but not exclusive,
source o f ore components.
Manto-type copper deposits are best known from Chile
and southern Peru. They consist o f stratiform concentra-
tions of disseminated and vesicle-filling chalcocite-bornite-
chalcopyrite in the tops of Jurassic to early Tertiary andes-
ite flows and felsic ash-flow ruffs, and in associated volcanic-
iastics and organic-rich limestone~ Deposits were generated
by the circulation of meteoric water through hot volcanic
rocks immediately following their accumulation. Copper
and subordinate silver were leached from the volcanic rocks
and deposited in volcanics or related sediments in restricted
lagoons, perhaps aided by bacterial action. Volcanogenic
tin deposits are minor accumulations of cassiterite, wood
tin and hematite in felsic volcanic rocks (+ possible rhyolite
plugs) in western Mexico, the western U.S.A. and Bolivia
and adjoining northwest Argentina. Their origfn may l~e
similar to that of the manto-type copper deposit~
Volcanogenic magnetite (+ hematite, apatite and actino-
rite) deposits occur as flows, plugs and pyroclastic beds at
Cerro de Mercado, Mexico, and El Laco, Chile. The map
netite magma separated from its ca/c-alkaline parent because
of liquid immiscibility and was erupted because of a high
volatile content and a lowering of its melting point by the
con rained phosphorur
These ore deposits of subaerial volcanic affiliation were
generated at convergent plate boundaries, many o f them
while subduction was active. Volcanogenic tin and some
epithermal precious-metel deposits in the western U.S.A.
and an epithermal gold deposit in Fiji, however, were em-
placed following the cessation of subduction, in extensional
tectonic regimes, characterized in the western U.S.A. by a
bimodal basalt-rhyolite suite and in F i j i b y a basaltic suite
o f ocean-island affinity. Nevertheless, metallic mineraliz-
ation appears to be absent from subaerial volcanic rocks
accumulated in extensional environments at incipient
divergent plate boundaries, as i n the East African rift
system and the Scottish Hebride~
In this volume, Volcanic processes in ore genesis, a series of
papers considers both massive sulphide deposits related to
submarine volcanism, and copper and tin deposits of por-
phyry type related to sub-volcanic intrusions. Metallic
mineralization hosted by subaerial volcanic sequences is,
however, essentially ignored. In order to redress the balance,
therefore, the editorial committee requested the writer to
prepare a contribution to review salient aspects of metallic
ore deposits emplaced in subaerial volcanic rocks.
The principal ore type highlighted in this review is the
vein deposit of epithermal type, about which much new
information has recently become available. The term
'epithermal' is employed in the sense of Schmit 15 to des-
cribe mineralization emplaced at shallow depth and nor-
mally in volcanic rocks. Besides the epithermal deposits,
stratiform, manto-type copper deposits, volcanogenic mag-
netite deposits and volcanogenic tin deposits are also
briefly described. For the sake of brevity, attention is con-
fined to deposits in post-Palaeozoic volcanic sequences.
Only a brief mention of certain important aspects of these
ore types is presented here; the references cited should pro-
vide guidance for more detailed reading.
General characteristics of epithermal deposits
Employing definitions and examples of epithermal ore de-
posits that have appeared in compilations over the years, 1-7
the following features appear to be characteristic.
(1) The deposits are commonly emplaced in volcanic se-
quences ranging from andesite to rhyolite in composition,
and mid- to late Caenozoic in age. In many instances no
observable relationship to intrusive rocks is evident.
(2) Veins occupy pre-existing tension fractures, including
faults, and occur in branching patterns, which are common-
ly increasingly complex upwards. Veins are more persistent
laterally than vertically; ore shoots within these veins con-
stitute the so-called 'bonanza' orebodies. Stockworks and
pipes may also be present.
(3) Propylitic alteration affects volcanics in the vicinity
of ~veins, with more restricted alteration haloes of sericitiz-
ation and silicification, commonly with adularia, bounding
the veins. Propylitization is, however, commonly a regional
phenomenon not directly related to vein formation. Alter-
natively, zones or vein haloes consist of advanced argUlic
and argillic alteration, with clays, alunite, pyrophyllite,
diaspore, etc., accompanied by widespread silicification;
the advanced argillic alteration type was originally termed
'solfataric', 6 but is commonly generated at depths greater
than those that could be affected by strictly solfataric
alteration.
(4) The deposits contain principally precious metals in
various proportions, but also base metals. Silver occurs as
argentite (or acanthite) and arsenic- or antimony-bearing
sulphosalts, gold in the native state, but usually argenti-
ferous, or as tellurides, and lead and zinc as galena and
99
sphalerite. Copper is less plentiful, but commonly occurs production is difficult to estimate, but its magnitude is indi-
as enargite where advanced argillic alteration is developed. cated by figures of more than 1000 000 000 oz of silver
Stibnite, cinnabar and selenides can also occur. The main from both Pachuca and Guanajuato. Furthermore, quite a
gangue is quartz, commonly as aphanitic chalcedony, number of the mines have yielded more than $1000 000 000
accompanied by varying quantities of pyrite, calcite, dolo- worth of silver and gold at present metal prices.
mite, barite, adularia, rhodochrosite, rhodonite and The belt is broadly coextensive with the Sierra Madre
fluorite. Occidental and the Trans-Mexico volcanic belt, but in the
(5) Drusy cavities, filling, comb, colloform and crusti- northwest it impinges on the Sonoran Basin and Range
form textures and inter-mineral breccias are all features of province and in the southeast, south of the Trans-Mexico
veins suggesting low confining pressure. Ore and gangue volcanic belt, on the Sierra Madre Oriental. The northern
minerals are typically fine-grained. half of the belt was the subject of a comprehensive
synthesis by Wisser. 7
(6) The shallow depths of emplacement, perhaps from Most of the veins occur in andesitic volcanic rocks, al-
100 to 1000 m beneath surface, are emphasized by the though some cut more felsic volcanic units and others the
abrupt downward termination or 'bottoming' of mineraliz- sub-volcanic basement, consisting of sedimentary and intru,
ation or the downward increase of base-metal sulphides at sive rocks. Post-mineral volcanic units, particularly felsic
the expense of gold and silver. Ore shoots in veins have tufts, are also widespread. Details of the volcanic rocks and
rarely been worked for more than 600 m vertically. their ages are sparse and their relationships to volcanic
centres are largely unknown. The lower half of the volcanic
Typical regions of epithermal deposits
pile is, however, dominantly andesitic and is Lower to
Western Mexico Middle Tertiary in age, whereas the upper half consists of
rhyodacitic to rhyolitic pyroclastic materials erupted, at
In north and central Mexico a longitudinal belt of latitude 24~ 32 to 23 m.y. ago. 8 Much of the younger
Caenozoic volcanic rocks and epithermal mineral deposits sequence comprises ash flows derived from caldera com-
extends for 2400 km from the U.S. border to south of plexes comparable with those in the San Juan Mountains,
Mexico City (Fig. 1). This is the greatest known concen- Colorado (see below). Uplift of fault blocks occurred along
tration of epithermal silver deposits in the world and is the central portion of the Sierra Madre Occidental, at least
responsible for a major contribution to Mexico's current partly during volcanism, as attested by conglomerates inter-
position as the world's leading silver producer. Total past bedded in the volcanic sequences. 7 Fractures, commonly

117~ 113~ 109 ~ 105 ~ 101 ~ 970 93 ~ 89 ~ 85 ~

I I I I I I
32 ~

~......
!

0 200 400 600 km /

28 o I I I I I I I -t 28 ~

t..J ,~
/~.~. /]~
I s"
k"~,
f' ,.'~.
24 c ~~osat~~v \ ~'~dr~i
..... ~
'IL~. ,.~
~-
.,..J 14~
~,, .
rr~o oe Mercaoo ..,
/ )
~,.
(
...P

~x \&_ ~.....

~t
~ Precious-metal belt,
including e p i t h e r m o l d e p o s i t s 'l."v ~ . . . i ~ . \ ./''"
20 ~ :0o
,,4" Belt of v o l c a n o g e n i c
~ . j, tin d e p o s i t s '\i/'f

O Volconogenic magnetite deposi,

9 ,, j.,,.. !.
9 Principal epithermat silver deposit t r') --'~. ~..............
,*\...s S t a t e boundaries .... ) ' . ,

16~ ,' ._...'5. 16~

...."
I I 1 I 1
113~ 109 ~ 105 ~ 101 ~ 97 ~ 93 ~ 89 ~

Fig. 1 Limits of precious-metal and volcanogenic tin belts in Mexico, and location of the volcanogenic magnetite deposit..:
Note clear offset of mid-Tertiary precious-metal belt at about latitude 21~ caused by sinistral strike-slip fault now con-
cealed by Trans-Mexico volcanic belt, Small area of precious-metal mineralization on southern tip of Baja California has
been rifted away from mainland between latitudes 20 ~ and 21~ during opening of Gulf of California since 5 m.y.

100
longitudinal in direction, activated during uplift of the
Sierra or during localized uplift within it, are the sites for
subsequent vein formation and closely related dyke em-
placement. Stocks are known from only a few districts.
In the Mexican epithermal deposits argentite (and acan-
thite) and the ruby silvers are always deposited later than
the base-metal sulphides, and gold is concentrated in the
final phases of mineralization. 7 Wisser 7 recognized that
deposits close to the interface with the sub-volcanic base-
ment are rich in silver, carry abundant base-metal sulphides,
but only minor quartz, and are narrow, whereas veins
higher in the volcanic pile are richer in gold and quartz,
poorer in base metals and broader (Fig. 2). All veins en-
dowedwith abundant base metals lie within the basement
or within 500 m of it, and gold veins with a Ag : Au ratio
exceeding 200 (i.e. Au -- 1) lie higher than 500 m above the
basement (Fig. 2).
1500
E
.,.Y
~ooo
E
o
~ 500
a than 20% of the vein structures. 11 Chalcopyrite, galena and
] iI 9 [[ I sphalerite dominate at the lowest mine levels.
I 0
3000
25()0 20=00 15J00 10'00 5(~0
[ Vertical range of
ore in vein
Fig. 2 Silver:gold ratios of precious-metal vein deposits
in northwest Mexico plotted against height above sub-
volcanic basement. 7 Note higher Ag: Au ratios in deeper
deposits
Pachuca (Fig. 1) has been responsible for the production
of 1200 000 000 oz, or 6% of the world's silver, commenc-
ing in the sixteenth century, in addition to about 6 200 000
oz of gold, and important amounts of lead, zinc and copper.
According to an analysis of the district undertaken by
Geyne and co-workers, 9 the volcanic rocks that host the
veins range, on the basis of fossil plant remains, from early
Oligocene to Pliocene in age, are largely andesitic and
dacitic in composition, and originally attained a thickness
of 2600 m. Faulting was widespread in late Miocene times
when rhyolite and dacite porphyry dykes were intruded,
and was followed by renewed faulting. These workers
= " ~trt" ~ -I
~-~ Alluvium J'Veins ~Andesitic to rhyolitic dykes + plugs
[---1 Dominantly andesitic + < rhyolitic volcanics
Fig. 3 V~n-and dyke systems in Pachuca district,
Mexico 123
postulated vein formation on the faults in the Middle
Pliocene (Fig. 3). Mineralization spans a vertical interval of
700 m and is associated with albitization, silicification,
sericitization and carbonatization fringing the veins.
At Tayoltita, San Dimas district (Fig. 1), volcanic rocks
and an intercalated conglomerate were intruded by a series
of dykes, followed by stocks of granodiorite, diorite and
andesite, prior to gold-silver mineralization. 1~ The
granodiorite was dated by P. E. Damon at 36.9 + 1.8 m.y.
or late Eocene-early Oligocene. Production to date has
totalled 12 000 000 tons assaying 820 g/ton Ag and
16 g/ton Au. 10
The Guanajuato district (Fig. 1) is characterized by at
least 700 m of extrusive rocks, mainly rhyolitic and andesi-
tic flows and tufts, ranging in age from earliest Oligocene
(37.0 -+ 0.3 m.y.) to Middle Oligocene (32.1 -+ 1.0 m.y.). 11
They overlie Mesozoic rocks, including intrusives, one of
which was dated at 69 -+ 1.2 m.y., and an Eocene conglom-
erate. Silver and gold mineralization occurs in veins and
[ stockworks filled by quartz, calcite and adularia. Three
northwest-trending vein zones are recognizable, the most
persistent Veta Madre structure extending for 25 km.
Silver, mainly as acanthite, aguilarite and polybasite, and
gold, as electrum, 12 occur in restricted shoots as large as
90 m wide and 300 m long, which in total occupy no more
1 Gulf of Fonseca area, E1Salvador-Nicaragua-Honduras
In this area (Fig. 4) 15 epithermal vein deposits have been
important producers of silver and gold. 13,14 The veins are
emplaced in mid- to late Tertiary volcanic rocks, domin-
antly of andesitic composition, and cannot be related to
outcropping intrusive rocks; one of the few spatially related
intrusives, a quartz latite porphyry dyke at the San
Sebastian deposit, El Salvador (Fig. 4), is clearly post-
mineral (personal observations, 1976). Quartz is the main
gangue, pyrite is abundant and silicification of the vein
walls usual. Copper, lead and zinc values augment with
depth, in contrast to the Ag: Au ratios, which decrease.
Veins, as at Montecristo, El Salvador (Fig. 4), attain up to
2 km in length, and occupy northwest-trending faults and
shear zones that are believed to be related to the formation
of the Nicaraguan graben (which contains Lakes Managua
and Nicaragua; Fig. 4), possibly at the end of the Miocene,
prior to major ignimbrite eruption. TM
Comparable epithermal veins in andesitic volcanic rocks
of similar age are known from the Aguacate cordillera,
Costa Rica. They resemble those of San Sebastian, El
Salvador, in that their gold contents predominate over those
of silver.
Altiplano-Puna region, Peru-Bolivia-Chile-Argentina
In several regions of the central Andes, most notably on
the Altiplano--Puna block of western Bolivia, northwestern
Argentina, northeastern Chile and southern Peru, epi-
thermal deposits of minor commercial importance are
known to occur, but have been little studied.
Recently, the writer 15 described a small, fault-
controlled silver-lead deposit from the El Queva district of
northwest Argentina. The deposit is located in an extensive
zone of feldspar-destructive alteration in the eroded central
part of a complex Quaternary strata-volcano constructed on
Palaeozoic basement. Mineralization is accompanied by
chalcedonic quartz, dickite, alunite and barite that have
partly replaced the host-rock dacite. The El Queva mineral-
ization is particularly instructive because it clearly demon-
101
 88 ~
14~
I.
@ Q u a t e r n a r y alluvium and
marine s e d i m e n t s
r~//x~;;~
[] Quaternary volcanic rocks
D Middle-Upper Tertiary votcanic rocks
[] Mainly pre-volcanic ' b a s e m e n t '
Mainly n o r m a l f a u l t s , downthrow shown
12~ --
ee Epithermal s i l v e r - - g o l d deposits
~ Lakes
0 100
I 1
88 ~
Fig. 4 Locations of epithermal silver and gold deposits in Gulf of Fonseca area, in El
Salvador, Honduras and Nicaragua, Central America. Relationship of precious-metal de-
posits to mid- to late Tertiary volcanic rocks, and faults bounding graben in Nicaragua
and El Salvador is evident. Two sizes of symbols for precious-metal deposits roughly
indicate their importance. From Dengo eta/. 124
strates epithermal ore formation within the upper, central
parts of a volcanic superstructure that has not undergone
caldera collapse. In fact, it has been remarked that few
calderas accompany base-and precious-metal mineralization
in late Caenozoic volcanic rocks in central Peru, the de-
posits having formed in the vicinities of central volcanic
vents. 16 A number of small, epithermal deposits on the
Altiplano-Puna block exhibit characteristics similar to
those of El Queva inasmuch as they are located within iso-
lated andesitic to dacitic volcanic edifices and are commonly
associated with alteration zones; intrusive rocks are minor
or absent. Little-known examples include Antofalla in
Catamarca province, northwest Argentina, Laurani,
Carangas and San Antonio de Lipez, Bolivia, 17,18
Choquelimpie, Chile, 19 and mines in the Condoroma-Palca
district in the departments of Puno and Cuzco, Peru. 2~
Epithermal silver-lead--zinc ores occur north of the
Altiplano throughout Peru 20 and include Huachocolpa,
nearby Caudalosa and Finlandia, all mentioned below.
At Choquelimpie the andesitic volcanics, intruded by
plug- and dyke-like bodies of diorite porphyry and hydro-
thermal breccia, are located within a caldera of late
Caenozoic age, surrounded by ash-flow tuffs. Mineralization
consists of galena, sphalerite, acanthite, sulphosalts, realgar
and orpiment, and is accompanied by widespread chalce-
donic quartz, clays, alunite, pyrite, barite and calcite
(personal observations, 1972). At Carangas, pyrargyrite,
tennantite, sphalerite and galena, with barite, rhodochro-
site and calcite fill open-space in a breccia body, apparently
within a volcanic edifice. 18 Ahffeld 18 considered mineral-
ization on the Bolivian Altiplano to be related to acid
differentiates that have intruded the andesite-dacite vol-
canic piles, but at least in some areas, as at Laurani, these
are absent. At Laurani, the base of a volcanic edifice,
10 km x 11 km in area, is clearly marked by outward-
dipping andesitic and dacitic volcanic rocks. Veins cut the
102
14~
86 ~
12~
Lake
Nicaragua
200 km
I
86 ~
central part of the edifice and contain chalcedonic quartz,
fine-grained pyrite, enargite, boulangerite, calcite and
barite as the principal minerals.
San Juan Mountains,. Colorado, U.S.A.
The San Juan volcanic field in Colorado is the erosional
remnant of a huge volcanic field that covered the southern
Rocky Mountains in the Oligocene 21,22 (Fig. 5). Tertiary
volcanic activity was initiated about 35 m.y. ago and for
the next 5 m.y. was dominated by the construction of
strato-volcanoes by andesitic to rhyodacitic rocks. About
30 m.y. ago activity changed to the eruption of silicic ash-
flow tuffs derived from 17 collapsed calderas during a
4 m.y. interval (Fig. 5). Intermediate-composition lavas
were erupted during and subsequent to caldera formation.
The calderas were resurgently domed, during which time
felsic intrusives were emplaced around the caldera margins.
About 25 m.y. ago intermediate-composition volcanism
gave way to a bimodal mafic-silicic assemblage 23 and
between 25 and 5 m.y. ago the Lake City caldera developed,
and silicic alkalic minor intrusives were emplaced, some
controlled by earlier caldera structures (Fig. 6).
Lipman et al. 24 noted the emplacement of minor vein
deposits in the cores of the 3 5 - 3 0 m.y. old strato-
volcanoes (cf. El Queva) and also emplacement associated
directly with resurgence of the calderas, including the later
Lake City caldera, in the western San Juan Mountains.
Some 95% of mineral production in the western San Juan
Mountains has, however, been derived from caldera ring
faults, outward-extending radial fractures, and graben
faults along resurgent domes, all produced during subsi-
dence and resurgence of the 3 0 - 2 6 m.y. old calderas;
mineralization was, however, considerably later than caldera
development and was related to silicic magmas pertaining to
the bimodal suite. Notable ore deposits, many of which
possess epithermal characteristics, although base metals are
 108 ~ 107 o 106 ~ veins, including the Camp Bird deposit, trending northwest-
39 ~ I wards from the Silverton caldera and cutting flows, volcanic
I Colorado breccias and mud flows erupted during the strato-volcano
stage 26 (Fig. 7).
Area
map 107o45 '

38o0 '
9

San Juan volcanic fie d

38 ~

~C f~
i
# t
, a SM

3?

0 50 km
I i i 1 I !
.Miocene quartz porphyry intrusives Ma.jor
vein

(~ Catdera ~] Oligocene felsic intrusives

BCaldera hosting major mineralization
,~, Buried calderas
Fig.5 Limitof San Juan volcanicfield,southern
Colorado,U.S.A.,showingcalderas27 (Bz,Bonanza,C,
Creede, LC, Lake City;S, Silverton;SM, Summitville;UN,
Uncompahgre)
widespread, include breccia pipes associated with small
quartz latite porphyry stocks in the Red Mountain district,
in the Silverton caldera ring fault zone, which contain
copper and silver sulphosalt ores accompanied by silicifica-
tion and advanced argillic alteration; 2s and P b - Z n - A g - A u
10 000 .Silver ton
I /caldera
8000[
Red Mountain
~60001 ~ ~
Major
[~ Oligocene volcanic rocks fau It
Pre-volcanic rocks
Fig. 7 Location of Pandora-Camp Bird and other vein
systems and Red Mountain breccia-pipe district on north-
west side of Silverton caldera, Colorado 24
Elsewhere in the San Juan Mountains, mineralization is
spatially related to other caldera structures 27 (Fig. 5), but
again may be somewhat later in age in some districts and be
a product of post-25 m.y. magmatism. Well-known ex-
amples include the following: post-subsidence igneous
activity around the periphery of the Summitville caldera
includes, in the Summitville district, a quartz latite lava
dome, which hosts A u - A g - C u deposits in zones of ad-
vanced argillic alteration. 28 Graben faults radial to the
Creede caldera formed after the accumulation of volcanic
ash, stream and lake sediments and travertine of the Creede
l Ages of
ore-related Formation within the collapsed caldera, and were the locus
~ intrusives
of A g - P b - - Z n - A u mineralization of vein type valued at
1
rages of ore $100 000 0 0 0 . 29 Adularia from the veins was dated at
Camp 24.6 m.y., which suggests that they were products of the

E 2000 I-
t
n
.Lake City
Icaldera
Bird
calc-alkaline and not the later bimodal activity. 3~

Great Basin, Nevada, U.S.A.

40 35 30 25 20 15 10 5 Nearly all epithermal precious-metal deposits in the Great
Age, m.y.B.I~ Basin were formed in the mid- to late Tertiary. Many de-
Calc-atkaline
suite { ~] ntermediate lavas
Ash flows
posits in central Nevada formed from 16 to 14 m.y. ago,
during a period of volcanism lasting from 16 to 10 m.y.,
Bimodal suite ~ Basalt+rhyolite comparable with the bimodal alkali rhyolite-basalt suite
of the San Juan Mountains (see above), but deposits of
Fig. 6 Summary of relationships between petrologic lesser importance are also associated with a previous 3 8 -
type, volume and age of volcanism in western San Juan 20 m.y. period of normal calc-alkaline volcanism. 31 As in
Mountains, Colorado, along with ages of epithermal ores the San Juan Mountains, this pulse of calc-alkaline volcanism
and ore-related intrusives in region. 24 Note relation of many was initially (38--33 m.y.) andesitic, but changed to
dated deposits to low-volume, bimodal basalt--rhyolite dominantly rhyolitic to quartz latitic ash-flow eruption
suite until about 20 m.y. In western Nevada, mineralization

103
 apparently accompanied mixed bimodal and calc-alkaline
activity from 20 to 5 m.y. 31
As reviewed by Albers and Kleinhampl, 32 precious-metal
deposits in the Bullfrog, Goldfield and Silver Peak districts
may be spatially related to caldera ring fractures, whereas
other districts are apparently unrelated to calderas; but
nearby minor intrusives have been recognized that mark
andesite-dacite volcanic centres. Bodie, for example,
occurs in graben-faulted andesitic and dacitic volcanic rocks
and is spatially and probably genetically related to dacitic
plugs. 33 Furthermore, Silberman and McKee 31 concluded
that the preeminent precious-metal deposits in western
Nevada and adjoining California, at Bodie, Aurora, Tonopah,
Comstock Lode and Goldfield, all seem to have been em-
placed during the final stages of activity at andesitic-
dacitic volcanic centres, but rhyolitic centres also host
smaller deposits.
At Goldfield, a producer of 100 000 000 oz of silver and
4 200 000 oz of gold, the epithermal veins, emplaced
20 m.y. ago, cut Miocene trachyandesitic-rhyodacitic
volcanic rocks accumulated from 22 to 20 m.y. ago, but are
structurally controlled by caldera ring fractures formed
during an Oligocene ( 3 1 - 3 0 m.y.) episode of silicic vol-
canism, thereby demonstratingno genetic tie to caldera
development. 34 In the Goldfield district an alteration zone
40 km 2 in area consists of advanced argillic and argillic
alteration and structurally controlled silicification, a small
part of the last alteration type hosting much of the precious-
metal mineralization.34, 3~ The silicified vein structures
carry irregular patches and pipes of sulphides, sulphosalts,
including famatinite, and native gold.
The Tonopah district has a recorded production of
100 000 000 oz of silver and 1 900 000 oz of gold from
replacement veins characterized by quartz-sericite-
adularia alteration within andesitic flows and mud-flow
breccia. Argentite, polybasite, sphalerite, galena, chalco-
pyrite, pyrargyrite and pyrite are the principal hypogene
minerals. 36
Japan
Miocene volcanism in Japan was submarine on the inner
(Japan Sea) side of Honshu and Hokkaido, but essentially
subaerial in Kyushu. The two northeastern islands are the
locus of the well-documented Kuroko-type massive sulphide
deposits - an ore type absent from Kyushu because of the
lack there of submarine conditions during the M iocene.
Epithermal A u - A g , Pb-Zn and Mn vein deposits are, how-
ever, found in all three islands; they are associated with
the submarine Green Tuff formations and many of them
are believed to have been emplaced about 13 m.y. ago at
the same time as the Kuroko deposits, but mainly at lower
stratigraphic levels in the volcanic pile, 37 and they also
occur in the subaerial volcanic sequences. 38,39 K - A r ages
of 8.4 and 3.7 m.y. were, however, recently obtained on
adularia from two epithermal A u - A g veins, 4~ so not all
members of this ore type in Japan were emplaced con-
currently with the Kuroko ores. Vein and stockwork-vein
deposits seem to be comparable, whether or not their host
rocks are andesitic or rhyolitic, submarine or subaerial.
The veins are commonly finely banded, with black, silver-
rich 'ginguro' stringers carrying most of the ore. Pyrite,
sphalerite, galena, tennantite-tetrahedrite, argentite, silver
sulphosalts and gold--electrum are accompanied by quartz,
calcite and adularia as the most commonly reported vein
assemblage. On the basis of metal content and mineralogy,
Nishiwaki and co-workers 39 divided the veins into the
following types: base-metal; chalcopyrite; argentite; man-
ganese; sulphosalt with high silver content; sulphosalt with
104
low silver content; a variety intermediate between these
two sulphosalt types; and telluride. Pipe-like orebodies
of gold-bearing pyrite accompanied by advanced argillic
alteration may also be included in the epithermal category,
another ore type present in both the subaerial and sub-
marine Miocene environments of Japan; 37 examples include
Kasuga in Kyushu and Date in Hokkaido. Lambert and
Sato 41 suggested that Kuroko massive sulphide deposits
and stockworks were transitional to contemporaneous
fissure-filling vein deposits.
Hauraki region, New Zealand
Epithermal A g - A u and some base-metal vein deposits
follow north or northeast fractures in Upper Caenozoic
andesitic volcanic rocks, 800 m thick, and to a lesser extent
in 200 m of overlying rhyolites, pumice breccias and ig-
nimbrites, in Hauraki province of North Island, New
Zealand. 42 The region from Great Barrier Island in the
north through the Coromandel peninsula to Te Aroha
160 km to the south has accounted for the production of
about 44 000 000 oz of bullion. The principal mine, at
Waihi, produced 4 000 000 oz of gold from a number of
quartz veins up to 30 m wide, few of which outcrop, in
dacitic country rocks. 42
Recently described 43 from Great Barrier Island are
quartz veins cutting andesitic volcanics, an intrusive quartz
porphyry and a sinter deposit. The veins are enclosed by
haloes of silicification and argillic and propylitic alteration,
and carry early pyrite and marcasite, followed by base-
metal sulphides and then by sulphosalts, of which the ruby
silvers are dominant, selenides and stibnite. Electrum was
also an important late mineral. Precious-metal mineraliz-
ation on Great Barrier Island, as elsewhere in the region,
was probably late Miocene-Pliocene in age, but could be
continuing today - as indicated by the high silver contents
of thermal-spring waters. 43 In contrast to the Great
Barrier Island veins, those at the Tui mine, Te Aroha, are
essentially base-metal-bearing quartz veins. 44
Philippines
Gold-bearing, quartz--calcite vein deposits of epithermal
character occur in the Philippines, especially in the Baguio
district, northern Luzon. 45 The veins are on the western
side of a dioritic pluton in late Caenozoic, largely andesitic
rocks, at least partly emplaced as dykes. At Acupan mine,
bodies of hydrothermal breccia have been described, the
youngest of them carrying native gold. Comparable epi-
thermal gold veins have also been described from Papua
New Guinea. 46
Fiji
Gold mineralization of epithermal type is known from the
Vatukoula area, the Tavua goldfield, of Viti Levu, Fiji. It
has accounted for a production to date of some
4 000 000 oz of the metal. 47 In the Vatukoula area a
basaltic strato-volcano was constructed, followed by
caldera collapse, with accompanying and subsequent
andesitic eruption. Augite andesite was emplaced as radial
and tangential dykes and cone sheets during and after
collapse, followed by biotite andesite dykes, plugs of mon-
zonite, and finally by gold mineralization. Volcanism in
the Tavua area took place between 5 and 4 m.y. ago. 48
The caldera ring fault zone is marked by alteration and
pyritization, but most of the gold, present as gold-silver
tellurides and auriferous sulphides, fills veins where north-
west-trending faults intersect the southwest portion of the
ring fault zone. 47
 Spain
Epithermal, gold-bearing veins have been worked in the
Rodalquilar area, near Almerla, southeastern Spain. 49 The
veins occur within a zone of pyrite-rich, advanced argillic
and argillic alteration, some 5 km x 3 km in surface dimen-
sions, within an andesitic volcanic pile of probable Miocene
age. Dacite and ignimbrite units occur in the upper part of
the pile. Gold tellurides and native gold are closely related
to north-trending, chalcedonic quartz veins.
Isotopic and fluid-inclusion data for epithermal deposits
Age relations with volcanic rocks
At El Queva, northwest Argentina, the Pleistocene age of
the host volcanic rocks confirms that less than 2 rn.y.
elapsed before the onset of silver-lead mineralization, and
in a number of districts radiometric evidence has been ad-
vanced to show that epithermal mineralization followed
closely after the eruption of the volcanic host rocks (Fig. 8).
Goldfield,
Nevada
Tonopah,
Nevada
Bodie,
Catifornia
Aurora,
Nevada
Surnrnitville,
Co[orodo
Creede, Pre
Colorado I !
Guanajuato, Pre
Mexico I 11
Huachocolpa,
Peru
Tui,
New Zeatand
40 30
Fig. 8 K - A r ages of pre-mineral volcanic rocks (Pre), vein minerals (M), mainly adularia
or alunite, and post-mineral volcanic rocks (Post) from a number of epithermal precious-
and base-metal districts
At Guanajuato, Mexico, volcanism ended 32 -+ 1.0 m.y. ago
and the principal vein-type mineralization was dated by
means of associated adularia at between 27.4 -+ 0.4 m.y.
and 30.7 Jr 3.0 m.y. ago. 11 In the Huachocolpa district,
south-central Peru, the age of P b - Z n - A g veins is bracketed
by 10.4 to 8.2 m.y. old volcanic host rocks and 4.0 m.y.
old post-mineral dacite-rhyodacite dykes, thus demonstrat-
ing mineralization a maximum of 4 m.y. later than the
cessation of volcanism. 5~
At Goldfield, Nevada, host volcanic units yielded K - A r
ages of 21.5 -+ 0.5 to 20.6 -+ 0.4 m.y. - closely similar to the
ages of about 20 m.y. for vein and wallrock alunite that
accompanied vein filling.34, 51. At Tonopah, Nevada, pre-
mineral andesite yielded a K - A r age of 20.4 m.y., adularia
from one of the veins yielded 19.1 m.y., and post-mineral
rhyolite ash-flow tuff yielded 17 m.y. 52 Similarly, in the
Bodie district, California, the volcanic host rocks accumu-
lated from 9.4 to 8.6 m.yo ago and were immediately
followed by A u - A g mineralization from which adularia and
sericite gave ages of 8.6 to 7.2 m.y., suggesting that the
hydrothermal system had a life-span of about 1.5 m.y. 53 At
Summitville, Colorado, pre-mineral quartz latite was dated
at 22.9 +-0.6 and 22.8 +- 0.6 m.y., alteration alunite at
22.4 -+ 0.5 and 22.3 -+ 0.5 m.y., and post-mineral rhyolite
at 20.2 + 0.8 m.y. 54
*More recent results demonstrate generation of hydrothermal
alunite and K-mica 20--21 m.y. ago. 129
In the western San Juan Mountains, Colorado, epi-
thermal ore deposition took place during essentially the
same time-span of 30 to 10 m.y. as the associated volcanism
(as noted above), but the richest deposits spatially coinci-
dent with the calderas, particularly the Silverton, were gen-
erated 5 - 1 5 m.y. later than the termination of caldera de-
velopment 24 (Fig.6). For instance, adularia from the Camp
Bird vein outside the Silverton caldera (Fig. 7) was dated at
10.5 -+ 0.5 m.y., some 17 m.y. younger than the age
accepted for propagation of the host fracture 24 (Fig. 6).
Elsewhere in the San Juan Mountains the 22 m.y. old
alunite accompanying mineralization in the Summitville
district was introduced 6.4 m.y. after subsidence of the
Summitville caldera (Fig. 5) was completed. 24 Similarly, as
was noted above, the mineralization at Goldfield, Nevada,
was controlled by caldera ring fracturing, which took place
some 10-11 m.y. earlier. 34
In contrast, at Creede, Colorado, however, a time-gap of
only about 2 m.y. between caldera collapse and mineratiz-
Reference
Pre M
34
er~ M Post
IJ I I 52
Pre M
53
Pre H
= = 33
Pre Post
54
N
30
Pre Post
50
Pre M
i i , 55
20 10 0
Age, m,y
ation was quoted by Steven and co-workers, 22, 30 although
even there a genetic connexion between the two events
cannot be positively demonstrated.*
A situation comparable with that in many parts of the
western San Juan Mountains appears to occur at the Tui
mine in the Hauraki region of New Zealand. There host-
rock andesites yielded K - A r ages of 17 m.y., whereas
hydrothermal alteration related to vein formation is re-
flected by ages of only 7 - 2 . 5 m.y. 55 (Fig. 8).
Fluid-inclusion studies
Fluid-inclusion studies of vein minerals from a number of
epithermal districts confirm that ore deposition was from
low-salinity, low-temperature (200-300~ fluids (Fig. 9).
This temperature range of 200-300~ is somewhat greater
than the figure of less than 200~ originally proposed for
epithermal ore formation by Lindgren. 3 The fluid-inclusion
filling temperatures generally have not had a pressure
correction applied to them, but this is likely to be small for
such shallow ore deposits. A correction of +25~ was cal-
culated for deposits in the western San Juan Mountains, 56
which are some of the deepest examples considered here. A
number of examples of results are given here.
*Recent K--Ar ages determined for vein-related sericite and adularia
indicate mineralization 24.6 -+ 0.3 m.y. ago, 1--2 m.y. later thanthoe
youngest pre-mineral volcanics which post-date caldera collapse.
105
(1) Precious-metal deposits in Nevada contain quartz and quartz and fluorite from Hosokura mine, Honshu. 61,62
lesser fluorite deposited between 330 and 200~ generally Lambert and Sate 41 pointed out that these temperatures
around 250~ from fluids containing less than 2.1 equiv. are the same as those indicated for the formation of quartz
wt % NaCI.s7 By use of oxygen isotope data a tempera- and sphalerite in Kuroko deposits.
ture of ore formation of 245-215~ was calculated for
nearby Bodie, California. 5s Nash 57 concluded that the (7) Inclusion-filling temperatures of 305-202aC determ-
wide spread of temperatures demonstrates that th is par- ined in barite from advanced argillic-altered, enargite-gold-

M0st hose metals+field

Reference
Sitverton district
veins. Colorado 9. = - - - - Ouar t z § sphaLeritu - - - - - = . - , 9
---P0st-sulphide quartz+ fluorite 56
Red Mountain
pipes. Colorado
I 56
Nain-stalle quartz + sphuteritu + barite
I
Sunnyside mine.
SiLverton district. lluse end precious metals Past-ore llanllue minerals 71
Colorado
Creede district. v
Colorado Ore stage 59
Precious- metal
deposits. Nevada Pro-, oyo- ~ d post-ore quartz
5?
Chinkuoshih.
Toiwan i,-, e . i o I 6~
Most deturmiMtwns on ore-stulle barite
Guonajuoto. t, . 4
Mexico 9==rtz +calcite (pro-main ore)
11
Pachuca,
Mexico Sulphide depouitio~ around tGlloC 60
FinLandio mine, I Leadl+zinc
central Peru .,=-Precious
9 ~t0Ls-~ 65
Caudalosa, I
central Peru Silver + base-metal solpkides Late-stage sphaLerite 66
BQse- and precious-
metaL deposits, =-,,--Nest determinations ~ ' 61.62
Japan
Toyoho, , t'~--Spkaluritu =u
Japan Quort;
63

Cosapalco. o . i e i o
Peru roMelu-|tele o o i r t z § -'~-%ate-s=tl~|e q e a r t z - H . . J ~ - - - Post-ore calcite -=-'
( sph~lorrtu ~uowaro J
370 350 300 250 200 150
T, ~

Fig. 9 Fluid-inclusion filling temperatures from epithermal ore deposits

ameter is not critical for the formation of epithermal bearing breccia pipes of probable Pleistocene age at
precious-metal deposits. Chinkuashih, northern Taiwan, 64 are very similar to those
from comparable breccia pipes at Red Mountain, Colorado
(2) At Creede, Colorado, ore and gangue minerals were (see above).
deposited at 265-190aC from fluids with salinities of
4 - 1 2 equiv, wt % NaCI, which were boiling at high levels in (8) In the 10 m.y. old Finlandia vein, central Peru,
some veins.59 precious-metal ores were deposited at 270 -+ 20~ whereas
the paragenetically later galena, sphalerite and quartz were
(3) At Guanajuato, Mexico, vein quartz and calcite were deposited from 270 ~ to about 140aC.6s
deposited at temperatures ranging from 260aC at 2400 m
above sea level to 320~ some 700 m lower in the vein (9) At the late Tertiary Caudalosa vein deposit, south-
system. 11 central Peru, main-stage silver sulphosalts, galena and
sphalerite were deposited at 325-270~ whereas tempera-
(4) At Pachuca, Mexico, vein filling began at 310-300~ tures declined from 250 to 170~ during late sphalerite pre-
and ended at 200aC, with major quartz and sulphide deposi- cipitation and dropped below 100aC during the final barite
tion around 250aC. Salinities averaged about 1 equiv, wt % stage.66 Salinities fluctuated from 4 to 18 equiv, wt % NaCI,
NaCI and boiling occurred near the tops of veins.6~ the highest values being attained during late sphalerite
deposition. 66
(5) A fluid-inclusion study of the vein deposits, including
Although temperatures of precious-metal- and base-metal-
the Camp Bird, northwest of the Silverton caldera, and the
bearing epithermal veins appear to be roughly comparable,
Red Mountain breccia pipes (Fig. 7), in the western San
there is a suggestion that salinities determined for veins that
Juan Mountains, provided evidence for ore deposition at
carry base metals are somewhat higher (> 2 equiv, wt %
315--249aC from fluids carrying 7.9 to 0.1 equiv, wt %
NaCI) than those which carry precious metals (< 2 equiv.
NaCI (with base-metal and gold stages near 290--280~ and wt % NaCI). In the Finlandia and Caudalosa veins, Peru,
salinities less than 2 wt %), and 308-216~ at sal inities of
this has been clearly documented within single paragenetic
1.6 to 0.2 wt %, respectively, s6 Unlike the model presented sequences.6s,66 There may also be a tendency for some
for Guanajuato by Gross,11 no systematic variation in tem-
base-metal-bearing epithermal veins to be generated at
peratures or salinities was documented over the 1100-m
somewhat higher temperatures (e.g. western San Juan
vertical interval examined in the,Jeins of the Silverton area.
Mountains); this is certainly the case if veins such as those at
(6) Fluid-inclusion studies of Japanese epithermal Casapalca, central Peru, are considered to belong to the epi-
precious- and base-metal vein deposits have shown that ores thermal category, 3 because there the main-stage ore was de-
were deposited mainly at 240--180aC from fluids with posited at temperatures of 370--320~ 67 At Finlandia and
salinities of commonly less than 3 equiv, wt % NaCl. 61,62,63 Caudalosa, however, Pb-Zn mineralization occurred at
Salinities as high as 9 wt % were, however, encountered in lower temperatures than the silver sulphosalt stage.65,66

106
Stable-isotope studies
Recently, several stable-isotope studies have been undertaken
on epithermal vein deposits, such as those in the Great
Basin of Nevada and adjoining California and including the
famous Bodie, Tonopah, Comstock Lode and Goldfield
districts, 68,68,69 the San Juan Mountains, including
Creede,59,70, 71 Pachuca, Mexico, 68 and Finlandia and
Caudalosa, Peru. 65,66 In studies by O'Neil and co-
workers,68, 68 the majority of vein-mineral samples examined
possess low (light) ~ 1sO values - indicating a substantial
meteoric water component in the ore-depositing fluids;
somewhat heavier values are the result of isotopic exchange
(the 1sO shift) with host rocks in systems with lower water/
rock ratios. Deuterium analyses of fluid inclusions in vein
quartz and adularia from the same veins are isotopically
light, ~D ranging from - 9 0 to - 1 3 9 %o, thereby confirming
the dominance of meteoric water in the ore fluids.
Taylor's 69 studies were carried out on quartz and altered
rocks from Tonopah, Comstock Lode and Goldfield, and
his calculated ~ 1sO and ~D values accord with meteoric
water as the principal ore-fluid component. ~180 and ~ D
determinations for Creede also show major involvementof
meteoric water in ore formation, 59 and elsewhere in the
San Juans meteoric-hydrothermal systems were associated
with the ore-bearing magmatic rocks. 7~
In contrast to the most common results, however, a
single sample of inclusion fluid in quartz from a deep
(1200 ft) level of the Comstock Lode gave a ~D value of
- 6 8 . 5 %o - indicative of a large magmatic component in the
ore fluid.68 Similarly, a single analysis of fluid from pri-
mary fluid inclusions in galena at Tui mine, Hauraki region,
New Zealand, yielded a ~ D value of - 7 0 %o - strongly
suggestive of a magmatic derivation. 55
In the two stable-isotope studies carried out on late
Tertiary Peruvian epithermal deposits, both meteoric and
magmatic fluids were identified. At both Caudalosa 66 and
Finlandia 65 some paragenetic stages, including early A g - A u
ore and quartz and barite gangue at Finlandia, were de-
posited from meteoric fluids, whereas galena-sphalerite
mineralization was apparently deposited from pulses of
dominantly magmatic fluid with ~ D values of - 5 0 -+ 5%o
at Finlandia and - 6 0 to -70%o at Caudalosa. In the
Finlandia vein only two of the seven identified mineraliz-
ation stages show evidence of a magmatic component. A t
Caudalosa the fluids interpreted as of magmatic origin also
possess the highest salinities and, perhaps rather surprisingly,
at both localities magmatic fluids became more evidently
involved during the late paragenetic stages.
Therefore, although the majority of available ~ 1so and
~D values clearly show that meteoric water played a major
role in the ore-depositional process, up to about 5% of mag-
matic water could be mixed with it and remain undetected
with currently available analytical techniques 72 a
_
possibility enhanced by the magmatic fluids recorded from a
number of deposits, most notably those in Peru.
Genesis of epithermal deposits
The close spatial relationship of epithermal ore deposits and
the late structural stages of caldera development has been
widely emphasized, particularly for the western
U.S.A., 27,73,74,75 but also elsewhere. 76 It would appear, how-
ever, that although this is an important locale for epithermal
ore deposition in some regions, it is not the sole one, since un-
collapsed volcanic cones also host this ore type. 15 Paren-
thetically, it should be mentioned that the relationship of
ore deposition to volcanic landforms and processes is un-
known in many epithermal districts. It seems likely that an
appropriate environment for epithermal mineralization is
created where the upper parts of volcanic piles have under-
gone suitable structural preparation and are intruded at
depth by felsic stocks, ls,24 This situation commonly per-
tains in the vicinity of volcanic centres of various types.
Structural preparation may entail ring and radial fracturing
during caldera collapse, graben formation perhaps related to
resurgent doming, faulting induced by stock emplacement,
or regional faulting during caldera collapse or the construct-
ion of strato-volcanoes. Some of the concentric and radial
fracture patterns localizing epithermal ore deposits analysed
by Wisser 77 can now be related to caldera-created features.
Ore deposition commonly follows from less than 1 m.y. to
3 m.y. after accumulation of the host volcanic rocks
(Fig. 8), presumably being related to cooling of the sub-
jacent magma bodies that fed the volcanism, but can occur
much later in regions of recurrent magmatism. It is indeed
common for mineralization to utilize structural channel-
ways produced up to 15 m.y. earlier, commonly during
episodes of caldera development 24,78 (Fig. 6).
The composition of the host volcanic rocks does not
appear to be a crucial control of epithermal mineralization.
Andesitic-dacitic host rocks are more common than
rhyolitic varieties, but this situation probably only reflects
the dominance of rocks of intermediate composition in
calc-alkaline volcanic belts. It would appear, however, that
epithermal deposits are scarce in ash-flow tuffs, which may
indicate that the caldera collapse mechanism is not con-
ducive to processes of mineralization. 24, 31,79 Recent
studies 24, 34 strongly suggest, however, that in many
districts calderas act as structural traps for later epithermal
ore deposits with which they possess no direct genetic
connexion, although in Nevada only two out of 30 caldera-
type structures are significantly mineralized, 79 and in epi-
thermal districts in the central Andes late Caenozoic
calderas are not widespread.
Minor intrusive rocks are spatially, and possibly genetic-
ally, related to epithermal deposits in some districts, but in
many districts intrusive rocks are unknown, though com-
monly inferred at depth. At Creede, Colorado, 3~ and
Goldfield, Nevada, 34 for instance, underlying intrusive
stocks are predicted on the basis of gravity and magnetic
data. Minor felsic intrusives are common in many parts of
the San Juan Mountains, Colorado, and have been related
to the epithermal deposits by a number of workers, most
recently by Lipman etal. 24 Moreover, some of the minor
intrusives are probably digitations on the roof of a com-
posite batholith that underlies the volcanic field -- a
situation required by its coincidence with a marked negative
gravity anomaly. 22
From the reviews of the fluid-inclusion and stable-
isotope studies presented above it is clear that epithermal
ore deposition is largely from low- to moderate-temperature
and low-salinity fluids dominated by meteoric water. Con-
firmation of meteoric water as a major ore fluid component
vindicates Schmitt's 1, 2 earlier hypothesis. A t Creede,
Colorado, an assessment of the hydrological setting of ore
deposition suggests that it took place in a surficial zone of
interaction between upward- and downward-moving
meteoric water controlled by variable rock permeabilities
near the wall of the Creede caldera. 3~
In the Green Tuff region of Japan, however, the ore fluid
for Kuroko deposits and epithermal veins is thought, on the
basis of similar mineral assemblages and sulphur and lead
isotope ratios, to be the same. 37,8~ Since it has been shown
to consist of at least 75% sea water in the Kuroko ores,81 a
sea-water-dominated fluid, instead of a meteoric one, may
be assumed for some of the epithermal vein deposits as well.
107
 These conclusions prompt an enquiry into the source of
the ore components transported by this meteoric (sea
water)-hydrothermal fluid. Are they derived from the intru-
sive stocks inferred to underlie many epithermal districts
(Fig. 10(c)), or are they derived from the country rocks,
either the volcanic sequences themselves (Fig. 10(a)) or the
sub-volcanic basement (Fig. 10(b)), the stocks merely
acting as "heat engines'? Derivation of the ore components
from sub-volcanic basement has been specifically proposed
for the Tui deposit, New Zealand, 55 and for the
Guanajuato district, Mexico, 11 on the basis of lead and
sulphur isotopes and other considerations.* In contrast,
however, isotopic data for the Japanese Kuroko ores and,
by analogy, the spatially related epithermal veins, preclude
/ ; I/// Epithermol veins
/Volcanic t i~ t \
/ rocks ',,.7",..J ~
- / / / / / / / / / / / / / / S o b - v o l c a n ibasement
c thermal ore deposition. Geological evidence strongly sug-
/ Colorado, 56 but only 300 m 30 or 500 + 150 m 59
I\ I
\.../ "~ --7
- y
// / / / / lit!ik // / / / /
Fig. 10 Three possible models for generation of epi-
thermal vein deposits in volcanic rocks: (a) involves meteoric
water circulating through volcanic pile; (b) meteoric water
circulating through sub-volcanic basement; and (c) meteoric
water circulation combined with contribution of magmatic
fluid
a basement source,41, 81 but permit the leaching of com-
ponents from the Green Tuff volcanic sequence or their
direct separation from underlying bad ies of cooling magma.
At Bodie, California, available isotopic data also favour an
intrusive source. 58 In regions such as ~he western San
Juan Mountains, where mineralization extends downwards
from the volcanics into the underlying basement, 82 a
source beneath the volcanic sequence is strongly suggested.
As was noted above, however, intrusive rocks constituted
more than just a heat source, at least in some districts, as
demonstrated by the ~ D values indicative of a magmatic
water component. If magmatic water were, in fact, in-
valved in epithermal ore deposition in many districts, al-
though with its presence commonly masked by the meteoric
(or sea-water) contribution (Fig. 10(c)), sulphur, carbon
*A basement source for metals and strontiumhas also been proposed
recently for epithermal mineralization at Sunnyside mine, Colorado.131
108
and metals may also possess at least a partial magmatic
source - a hypothesis favoured on geological grounds by a
number of workers. 15,24,82 The near-zero (~34S values
from some districts (e.g. Creede 59 and Sunnyside 71 ) also
support this origin, although a non-magmatic sulphur com-
ponent is clearly important elsewhere (e.g. Tui 55). ~ 13C
values, where determined (e.g. Bodie 58), are also indicative
of a magmatic origin.
An additional line of argument can also be mustered in
favour of a magmatic contribution to the ore fluid and the
inadequacy of scavenging, heated meteoric water alone. In
the Tertiary igneous province of the Scottish Hebridesthe
high-level emplacement of felsic stocks into co-magmatlc
volcanic rocks is known to have induced extensive circula-
tion of meteoric waters 83 in a manner closely comparable
with that in the San Juan Mountains magmatic province, 7~
yet ore deposits are notably absent from the Scottish
Hebrides. Perhaps the stocks in the Scottish Hebrides
failed to contribute a metal-charged fluid to the circulating
meteoric-hydrothermal systems (see below).
Recent work has confirmed the shallow depth of epi-
gests a depth of 4 0 0 - 5 0 0 m, and even shallower for minor
volumes of ore, at El Queva, Argentina, 15 3 0 0 - 1 0 0 0 m at
Pachuca, Mexico, 9 and 100 m for the shallowest veins at
Great Barrier Island, New Zealand. 43 Geological evidence
combined with fluid-inclusion results suggests a somewhat
greater depth of about 1200 m in the Silverton district,
at Creede, Colorado.
Since epithermal ore deposition occurs at such shallow
depths and is commonly very young, its surface manifesta-
tions should be evident. White 84 considered thermal
springs to be the active surface equivalents of the hydro-
thermal systems that generated epithermal vein deposits.
The salinity, and therefore density, of epithermal fluids is
sufficiently low (see above) to allow their discharge at sur-
face. At Steamboat Springs, Nevada, quartz-chalcedony-
calcite veins carrying pyrargyrite have been encountered
and the meteoric-hydrothermal fluids attain temperatures
exceeding 200~ at depths greater than 100 m 72 _ within
the epithermal range defined above. At Steamboat Springs,
White 72,85 favoured a volumetrically small magmatic con-
tribution to the system to account for the provision of
elements of low crustal abundance, such as Li, B, As, Sb and
Cs, for more than 2.5 m.y. (the duration of the system's
active life). Similarly, at Broadlands, New Zealand, 86
pyrite, sphalerite and galena, accompanied by propylitic
alteration, including adularia, have been described from
drill core, and high Ag, As, Sb, Hg and Au contents are
known from surface siliceous sinter. 87 At Broadlands the
meteoric-hydrothermal fluids possess 0.3 equiv, wt % NaCI
and temperatures of 2 9 8 - 1 2 0 ~ in the vicinity of base-
metal sulphides, 86 but older fluids in inclusions from
sphalerite, quartz and adularia are more saline. 88
It seems reasonable to correlate epithermal vein deposits
carrying carbonates and adularia and flanked by propylitic
alteration with modern alkaline hot-spring systems, but
epithermal deposits related to zones of advanced argillic
alteration and silicification were derived from much more
acid solutions, of a different nature and perhaps unrelated
origin. It has been suggested 15 for El Queva, Argentina,
that this type of deposit was generated beneath solfatara
fields that marked the late-stage activity at volcanic vents.
At El Queva the silver--lead mineralization is transitional
upwards to a surficial deposit of native sulphur. Such ad-
vanced argillic alteration is commonly present inthe eroded
cores of strata-volcanoes and may be comparable with that
noted above at Summitville and Red Mountain, Colorado,
Goldfield, Nevada, and Rodalquilar, Spain.
Travertine aprons and exotic manganese oxide deposits
may also form in the vicinities of precious- and base-metal
deposits of epithermal type, from heated bicarbonate-rich
waters according to White. s5 They were deposited within
2 m.y. of epithermal sulphide deposition around the peri-
phery of ~he strata-volcano at El Queva, 15 but at Ouray hot
springs in the northern part of the Silverton caldera 89 some
10 m.y. after base-metal deposition.
The writer 90 proposed that some zones of advanced
argillic and argillic alteration and silicification, possibly con-
taining epithermal base- and precious-metal deposits, are
generated in volcanic piles above porphyry copper-bearing
stocks. Advanced argillic alteration has now been docu-
mented above the El Salvador, Chile, porphyry copper ore-
body, 91 and, similarly, at Red Mountain, Arizona, feldspar-
destructive altered vo!canic rocks carrying alunite, enargite
and sphalerite are located up to 1200 m above a buried
porphyry copper deposit associated with K-silicate
alteration. 92
Geotectonic setting of epithermal deposits
It is clear that most of the epithermal deposits mentioned
above were generated above subduction zones active during
mid- to late Caenozoic times at convergent plate boundaries,
either along continental margins or island arcs. Many de-
posits are integral parts of calc-alkaline magmatic activity -
the hallmark of subduction. The water 67 and ore compo-
nents 93 of magmatic origin could have been derived from
the vicinity of the subducted slab - in common with the
calc-alkal ine magmas.
Cu, Pb,
Zn,Ag Sn
Western Eastern
Peru -Chile
trench
Coast
CordiLLera Cordillera
i AI.tiplano !
' - ~ d" -
Arr gap (~ ~ ~f
plate I - ~ ~ 4 ~ c
[ntruswe rocks ~ ~ I I
Porphyry copper and ~ .~"~>~,~ I
9 tin deposits . ~ "~.~.~;~
,, Epithermat vein deposit
A Votcanogenic magnetite deposit
,~ VoIcanogenic tin deposit
Fig. 11 Diagrammatic cross-section of central Andes at
approximately latitude 20~ during late Caenozoic to depict
locations of porphyry copper and tin deposits, volcano-
genic magnetite deposits and volcanogenic tin deposits, with
respect to magmatic loci and eastward subduction.
Note well-defined polymetallic and tin provinces
In the case of the central Andean Altiplano-Puna region,
eastward subduction of the Nazca plate is known to have
been active during the last 1 0 - 1 5 m.y., the interval during
which epithermal mineralization took place. As described
elsewhere,93 the epithermal deposits are assignable to a belt
of polymetallic ( C u - P b - Z n - A g ) ore deposits generated
during an eastward expansion of the locus of magmatism
about 16 m.y. ago (Fig. 11). The expanded magmatic belt
associated with the polymetallic deposits has a width of
200 km in the Altiplano-Puna region and is underlain at a
depth of from about 170 to 250 km by a shallow-dipping
(20--25 ~ subduction zone. The rare-earth element system-
atics and initial strontium isotope ratios (0.7048-0.7058)
of felsic lavas and tufts genetically associated with central
Peruvian base- and precious-metal deposits, including epi-
thermal examples, indicate derivation from sub-crustal
levels. 16 Similarly, the deposits in the Gulf of Fonseca area
of Central America were emplaced about 2 0 0 - 2 5 0 km
above the subduction zone that consumes the Cocos plate
and dips northeastwards from the Middle America trench,
assuming that the position of the trench has not varied
significantly. In this regard, it should be noted that the
Quaternary volcanic chain is located a little nearer the
trench than the precious-metal deposits (Fig. 4).
During the mid-Tertiary, when most of the Mexican epi-
thermal deposits (e.g. Guanajuato and Tayoltita) in the
Sierra Madre Occidental were emplaced, eastward subduction
of the Farallon plate was taking place in a manner similar to
that beneath the present-day central Andes, as adduced
from ocean-floor magnetic data. 94 Transition to a transform
plate boundary may have commenced about 23 m.y. ago
west of the Sierra Madre Occidental. 8 The silver deposits in
@
f
J
/
Z~2~ (/li ~ 1 ~ 1
!t ,
&
i'
',l/" f
E'|tl..
oo
//p
36~ - ~,~ l Tokyo \
0 100 200 km
I , I , I
\
........ Boundaries of petrographic volcanic provinces
Depth contour of subduction zone
Boundaries between tectonic segments
9 Kuroko deposit
9 Epithermal deposit
Fig. 12 Distribution of late Caenozoic Kuroko and epi-
thermal precious-metal deposits in northeast Japan 96 in
relation to petrographic provinces (tholeiitic, high-alumina
and alkali basalt), 95 depth to subduction zone 125 and
tectonic segment boundaries 126
the Sierra Madre Occidental are located on the western edge
of a belt of polymetallic ( C u - P b - Z n - A g ) deposits of Oligo-
cene age. Further south in Mexico, subduction at the
109
Middle America trench has remained active to the present
and has given rise to the Trans-Mexico volcanic belt. If the
Pachuca deposit is as young as Pliocene, 9 it pertains to this
later episode of subduction, along with at least some of the
central American deposits.
In northeastern Japan most of the Kuroko deposits and
associated epithermal base-metal veins were generated
130-155 km above the underlying subduction zone, mainly
within Kuno's 95 tholeiitic petrographic province 4~
(Fig. 12). The trench appears to have remained static during
the last 15 m.y. or so because the Miocene and Recent
volcanic fronts are essentially coincident. The major
precious-metal deposits, however, are located northwest of
the Kuroko province in both the high-alumina and alkali
basalt petrographic provinces of Kuno 9s (Fig. 12). A num-
ber of the epithermal precious-metal deposits lie in the
vicinity of transverse boundaries that separate segments of
the oceanic lithosphere that are subducted together but as
discrete entities, suggesting that such boundaries may aid in
their location 96 (Fig. 12).
As was noted above, epithermal base- and precious-metal
deposits in the western U.S.A. are related to two distinct
petrographic suites - an early calc-alkaline suite and a later
bimodal suite. Those deposits of calc-alkaline affinity were
generated during eastward underthrusting of the Farallon
plate on either an imbricate 97 or very shallow-dipping sub-
duction zone; K20/SiO 2 ratios determined for mid-Tertiary
calc-alkaline volcanic rocks suggest that the depth to the
subduction zone in western New Mexico and Colorado was
in excess of 300 km. 98 Those deposits of bimodal affinity,
reported from Nevada, Colorado and New Mexico, were
generated after subduction at the continental margin had
ceased as a result of interaction between the trench and the
oceanic ridge, leading to the transform fault margin. The
Steamboat Springs, Nevada, thermal-spring system is also
a product of continuing bimodal activity. The onset of the
volumetrically minor bimodal magmatism was roughly coin-
cident with the commencement of Basin-Range block
faulting, therefore being related to an extensional tectonic
regime. 23 The transition of mineralization of calc-alkaline
parentage to that of bimodal parentage took place at
25 m.y. in southern Colorado, 24 at 3 1 - 2 9 m.y. in southwest
New Mexico, TM and between 20 and 16 m.y. in central
Nevada, although in western Nevada calc-alkaline activity
accompanied the bimodal suite until 5 m.y. 31
For the past 5 m.y., following the inception of intra-arc
spreading, Fiji has been a remnant arc, active calc-alkaline
volcanism being restricted to the Tonga--Kermadec island
arc to the east; 5 m.y. ago Fijian volcanism changed from
intermediate, calc-alkaline to basaltic, akin to that character-
istic of ocean islands. 99 Since the Tavua volcanic centre
was constructed from 5 to 4 m.y. ago, 4s the associated epi-
thermal gold deposits are affiliated not with calc-alkaline
but with ocean-island-type basaltic volcanism and its final
differentiates.
It can be concluded, therefore, that although epithermal
ore deposits in continental-margin or island arcs are gener-
ally related to subduction, they can also be generated in
additional volcano-tectonic regimes that are created follow-
ing the cessation of subduction. The regimes recognized
here are that formed following the change from a conver-
gent to a transform plate margin involving plate margin ex-
tension I~erhaps classifiable as incipient intra-arc spreading,
and that following the actual splitting of an island arc by
intra-arc spreading resulting in the creation of a remnant
arc. Although volcano-tectonic settings other than the
subduction-related type have been recognized here as locales
110
for epithermal ore deposition, it would appear that not all
subaerial volcanic environments are conducive to the genera-
tion of epithermal ore deposits. Caenozoic extensional
regimes such as the East African rift system and the
Scottish Hebrides magmatic province, the formation of the
latter heralding opening of the North Atlantic, contain no
known base- or precious-metal ore deposits related to the
widespread volcanic and intrusive activity, which suggests
that the parent magmas were incapable of generating
mineralization, even though the volcanic centres were sub-
jected to the activity of major meteoric-hydrothermal
systems (see above).
Manto-type copper deposits
Veins, veinlets or disseminations of copper or copper-iron
sulphides in calc-alkaline volcanic rocks are widely distri-
buted, but rarely do they constitute commercial orebodies.
In Chile and southern Peru, however, a number of copper
deposits of this type have attained considerable importance,
Buena Esperanza, Antofagasta province, Chile, for example,
possessing a recorded production of 2 500 000 tons of 3%
Cu. Silver is also a common by-product from this ore type.
Their stratiform nature has resulted in their designation as
'manto-type' deposits.l~176 It seems reasonable to conclude
that economic copper concentrations of this type will be
found in calc-alkaline belts elsewhere in the world as
exploration continues.
The manto-type deposits lOl, ~o2 occur as stratiform
bodies, up to several tens of metres thick and several km
along strike, in volcanic rocks or volcanogenic sediments
ranging in age from Jurassic to Lower Tertiary. Andesitic
flows or breccias or felsic ignimbrite flows act as volcanic
hosts. The sulphides, dominantly hypogene chalcocite,
bornite and chalcopyrite, occur as disseminations, vesicle
fillings and short fracture fillings, commonly in the upper
parts of the volcanic horizons. The tops of deposits are
abrupt, whereas the bottoms are gradational. In volcano-
genic clastic sediments, sulphides occur in the fragments
and in the matrix, whereas in organic-rich limestones they
are finely disseminated throughout the rock. Intrusive
rocks are normally difficult to relate to manto-type deposits,
hydrothermal alteration is weak, and the gangue minerals
calcite, pyrite and quartz are scarce.
A good example of the ore type is provided by Buena
Esperanza lOl, lO2 (Fig. 13), where 28 mineralized horizons
ranging from 2 to 25 m in thickness occur in a 270-m thick
andesitic sequence of Jurassic age. Mineralization consists
of chalcocite and bornite in a ratio of 4: 1 in vesicular flow
tops and in intercalated lenses of sedimentary breccia and
sandstone. The presence of mineralized fragments in these
sedimentary intercalations testifies to the close temporal
connexion between lava eruption and mineralization.
Westwards the mineralized volcanic sequence ends abruptly
against a barren, presumably post-mineral, diorite intrusive
(Fig. 13).
From latitudes 30 ~ to 34~ in Aconcagua and Valparaiso
provinces, disseminated mineralization, dominated by
bornite, occurs in the tops of Lower Cretaceous andesite
flows and in the basal portions of overlying calcareous
sandstones, as at Guayac~n and El Salado, 1~176 lOl and at the
El Jardln deposit in Atacama province the uppermost 2 m
of a Lower Tertiary ash-flow tuff, 75 m thick, and the im-
mediately overlying lacustrine sediments carry chalcocite--
bornite-tennantite-native silver mineralization for 1 km
along strike.l~ Somewhat different from the manto-type
deposits, but apparently genetically related to them, are
irregularly shaped bodies of disseminated sulphides, such as
that in Lower Cretaceous andesites at Lo Aguirre, Santiago.
province, which contains reserves of 3 400 000 tons of 2%
Cu. 101
Many of the manta-type deposits were apparently
W rocks during meteoric-hydrothermal leaching. The relatively
Liven i
cooling volcanic units or in sediments deposited on top of
the cooling volcanics in temporary lagoons. The metal-
depositing fumaroles active after the deposition of ash-flow
tuffs in the Valley of Ten Thousand Smokes, Alaska, 1~
confirm that metals can be extracted from cooling volcanic
low temperature of the circulating waters is confirmed for
Buena Esperanza by inclusion-filling temperatures of 195 ~ to
112~ from quartz and calcite, although higher temperatures
of 4 3 0 - 2 5 0 ~ are indicated for El Salado. 1~
The manta-type deposits in the central Andes are an
integral part of the copper metallogenic belt, where they are
commonly located west of the major porphyry copper de-
posits. Their andesitic host rocks and copper content
could well have originated on the subduction zone that
underlay the region from Jurassic through Lower Tertiary
time. 93

Volcanogenic magnetite deposits

Three Meso-Caenozoic examples of volcanogenic magnetite
deposits are known to the writer: El Laco, northern
Chile,lOl, lO8 Cerro de Mercado, Durango, Mexico, 1~ and
Chilghazi and Pachin Koh, Baluchistan, Pakistan. 11~ Only
El Laco and Cerro de Mercado, related to subaerial volcanic
sequences, should be included here since the small
Pakistani deposits are submarine in origin. Both El Laco
and Cerro de Mercado contain roughly 1000 000 000 tons
of iron ore, assaying 50% Fe at El Laco.
+ + The bulk of the ore, magnetite and lesser hematite, at
v v v v"~IF/~L//P~,
El Laco and Cerro de Mercado was extruded in the form of
lava flows, although small amounts also occur as nearby
plugs and dykes and, at Pefia Morada near to Cerro de
Mercado, as a bed of hematite tuff. Apatite is a notable
Level 210 associate of the iron oxides and actinolite is also present.
Five separate iron oxide bodies in an area of 7 km x 3 km
0 50m are known at El Laco, the thickest up to 60 m. 1~ The El
I I I I 1 J
Laco flows were derived from parasitic vents around the
Vezicular flanks of a Quaternary andesitic strata-volcano, and overlie
andesitir breccia I Sandstone
flat-lying andesitic flows. 1~ According to Frutos and
~l Microporphyritir Oyarz~n,111 the edifice underwent caldera collapse followed
andesite ~ Augite diorite by the central emplacement of a rhyodacite plug. The
Cerro de Mercado deposits are associated with 30 m.y. old
~ Porphyritic
andelite I Copper
mineralization rhyolitic domes and flows in a large caldera. 1~
It is interesting to note that both these deposits were
~'% Fault formed during active subduction, as isolated centres within
Fig. 13 Section through a number of copper- the metallogenic belts that host the epithermal deposits of
bearing horizons in Buena Esperanza manto-type the Altiplano-Puna and Sierra Madre Occidental provinces
copper deposit, Antofagasta province, Chile 102 (Figs. 1 and 11). El Laco is associated with andesitic vol-
canoes similar to those that host the epithermal deposits,
generated as facets of subaerial volcanism, although, at as at El Queva, whereas Cerro de Mercado is related to the
least in some cases, related to local, short-lived lagoonal en- caldera development, which seems to be generally later
vironments. Some of the deposits could, however, be re- than the accumulation of andesitic volcanics with which
lated to volcanism under shallow, submarine conditions many of the Mexican epithermal silver deposits are
(e.g. Ra61, southern Peru 104). Mineralization would appear affiliated.
to be essentially coeval with the host volcanic or volcanic- It seems reasonable to conclude that the magnetite melts,
lastic rocks and is believed to be relatable to hydrothermal containing an appreciable phosphorus content, became
systems active during cooling of the spatially related immiscible with their parent calc-alkaline magmas during
volcanics. Volcanic heat would drive the hydrothermal cir- differentiation, as discussed by Philpotts. 112 The phos-
culation, the water being essentially meteoric in origin, at phorus content of the magnetite magma would lower its
least in some cases provided by the overlying lagoons. Sub- melting point (perhaps to 800-1000~ sufficiently to
lagoonal, organic-rich environments would provide the permit eruption, which would also be facilitated by the high
necessary reducing conditions, probably produced by bac- volatile content, the presence of which in the El Laco flows
terial action. 1~ Metals are thought to have been leached is indicated by gas-escape tubes. 1~
from the underlying, cooling volcanic sequence, and pre- It was suggested recently, 111 on the basis of V, Cr and Ti
sumably include the easily soluble, metal-bearing condens- contents, that the iron in the El Laco deposit was derived
ates from volcanic gases, described by Taylor and Stoiber. 1~ by anatexis of sedimentary iron formations in the under-
Sulphide precipitation was either in the upper parts of the lying Lower Palaeozoic basement. The known Palaeozoic

111
ferruginous strata in this region are, however, of such res-
tricted development that it seems unreasonable to cite them
as the source material for such a huge tonnage of magnetite.
A sub-crustal source, in common with the parent andesite,
would seem to be a more plausible hypothesis of origin on
purely volumetric grounds. Furthermore, the low initial
strontium isotope ratios of calc-alkaline rocks in this region
preclude widespread crustal anatexis.
Volcanogenic tin deposits
Volcanogenic or, as they have also been called, Mexican-
type 113 tin deposits constitute a minor but interesting ore
type. Their most widespread development is in a 900-km
long belt in western Mexico, where more than 1000 occurr-
ences are known in the states of Durango, 7acatecas, San
Luis Potos;, Aguascalientes, Jalisco and Guanajuato (Fig. 1 ),
but annual production is restricted to only a few hundred
tons of tin concentrate.
The occurrences in western Mexico 113,114,115 take the
form of short, narrow fissure veins, low-grade disseminations
and breccia fillings in mid-Tertiary rhyolitic rocks. The
precise ages of the host rocks are not known, but probably
fall in the 32 to 23 m.y. interval defined for rhyolitic vol-
canism in the Sierra Madre Occidental. 8 Veins occur in
steeply flow-banded plugs of vesicular rhyolite in which
minor disseminated ore may be present. The plugs probably
represent the sites of eruptive centres. Tin-bearing breccias
are developed in the upper parts of certain plugs. Horizont-
al flows of dense rhyolite and later basalt cap the tin de-
posits. Tin is present as incrustations of cassiterite and wood
tin and is ubiquitously accompanied by specularite. Mime-
tite, tridymite, quartz, chalcedony, opal, sanidine, white
mica, zeolites, topaz, fluorite, native bismuth, magnetite and
limonite also occur. Montmorillonite is the chief alteration
mineral. The tin was deposited close to eruptive vents from
low-temperature (< 150~ 113, 116) hydrothermal fluids
during the waning stages of rhyolitic volcanism. Meteoric
fluids may well have played an important role in the ore
fluid.
Similar volcanogenic tin deposits occur in mid- to late
Tertiary rhyolitic volcanic rocks at Taylor Creek, southwest-
ern New Mexico, 117 Lander County, Nevada, 118 and Cerro
Pululus, northwest Argentina, and Cordillera de Los Frailes,
Bolivia. 119 In contrast, it has been proposed, on the basis
of the presence of pseudobrookite and bixbyite, ~~u that the
Taylor Creek cassiterite mineralization was generated from
a gas phase at temperatures in excess of 500~ immediately
after deposition of the host rhyolites.
The geotectonic settings of this ore type are still far
from clear. In southwest New Mexico, however, the tin is
associated with post-caldera, high-silica alkali rhyolites
erupted 24 to 22 my. ago 127 and pertaining to the bimodal
suite, which also carries important epithermal precious-
metal deposits in this region; TM the Nevadan occurrences
also appear to be related to the bimodal suite. The western
Mexican deposits are, however, apparently of calc-alkaline
affinity if recent concepts 8 are valid. Nevertheless, the
presence of soda-rich alkali feldspar and fayalite as pheno-
crysts in rhyolites hosting the Mexican tin concentrations 115
is a characteristic feature of the bimodal suite rather than of
rhyolitic differentiates of the calc-alkaline suite. 23 Whether
or not further work confirms the calc-alkaline affinity of
the Mexican tin deposits, the minor showings of this type
in northwest Argentina and Bolivia certainly occur in calc-
alkaline volcanic rocks that pertain to the southern, late
Tertiary part of the Bolivian tin belt. The ultimate source
of the tin in the Bolivian volcanogenic deposits is presum-
ably the same as that related to the co-genetic stocks and
112
could well be sub-crustal and subduction related. 93 A sub-
crustal origin is also suggested for at least some of the high-
silica rhyolites in the Sierra Madre Occidental, Mexico, by
initial strontium isotope ratios of 0.704 and 0.705.128
Concluding remarks
Magmatic activity in continental-margin or island arcs is
recognized as one of the principal geotectonic settings for
the generation of metallic ore deposits. Previous syntheses
of ore types in this setting have tended to dwell on mineral-
ization within or peripheral to intrusive rocks, either batho-
liths or high-level stocks, or mineralization related to sub-
marine volcanic activity. It is hoped that this brief review
has re-emphasized the fact that subaerial volcanic rocks can
also host significant ore deposits (cfl Goossens 121 ). Plu-
tonic and volcanic rocks are considered commonly to be de-
rived from the same magma source, and the transition be-
tween the intrusive and extrusive environments takes place
at the level appropriate for the exposure of sub-volcanic
stocks. It is such high-level stocks that commonly host
porphyry-type deposits, and in a number of porphyry
copper 90 and porphyry tin 122 deposits the erosional
remnants of centripetally dipping, co-magmatic lavas can
still be seen around ore-bearing stocks.
Although the four ore types considered in this review
are all associated with subaerial volcanism, their geneses are
significantly different. The epithermal precious- and base-
metal deposits are generated in temporally related volcanic
rocks because of the presence of underlying, coeval stocks.
Heat, contributions to the ore fluid and probably most of
the ore components were provided by the stocks during
late-stage differentiation, although ore deposition in'the
near-surface, epithermal environment was dominantly from
convectively circulating meteoric waters. Some of the
underlying stocks, particularly those beneath epithermal
deposits associated with zones of advanced argillic alteration,
could host porphyry-type deposits, which signifies that the
epithermal ore components could be the distal equivalents
of deeper-level porphyry-type mineralization. The con-
clusion that most epithermal mineralization is generated
during the waning stages of activity at volcanic centres, but
not during ash-flow eruption caused by caldera collapse,
could imply a comparable condition for porphyry deposits;
this tentative suggestion is supported by the presence of
andesitic to quartz latitic lavas and pyroclastics in the
vicinity of temporally related, ore-bear ing porphyry stocks,
but the apparent absence of nearby felsic ash-flow tufts.
The manto-type copper deposits are probably also pre-
cipitated from meteoric waters, but without a direct mag-
matic contribution, the ore components being supplied by
the volcanic flow or sequence that underwent meteoric-
hydrothermal leaching. Therefore the copper in many
manto-type deposits possesses a comparable source to that
in porphyry copper deposits. Instead of being concentrated
and fixed in sub-volcanic stocks during the final phases of
their consolidation, however, the copper was erupted in dis-
persed form, probably at near-background concentrations,
in flows and pyroclastics, to be concentrated by meteoric-
hydrothermal systems debouching in lagoonal environ-
ments not necessarily close to their parent volcanic centres.
Comparable meteoric-hydrothermal systems active im-
mediately after the eruption of a volcanic sequence could
also account for at least some of the volcanogenic tin de-
posits, the spatially related rhyolitic plugs in western
Mexico providing a proximal heat source. The manto-type
deposits can be genetically related to ash-flow tufts -
unlike the stock-related epithermal deposits.
 The volcanogenic magnetite deposits differ markedly
from the other three ore types insofar as their contained
metal was not concentrated during the final stages of con-
solidation of an intrusive magma body, nor following vol-
canic eruption of magma, but much earlier during the mag,
ma's pre-consolidation differentiation. The differentiated
product was then erupted in a similar manner to a normal
andesitic magma, followed by cooling and consolidation.
This review of ore types related to subaerial volcanism
has confirmed calc-alkaline (and related tholeiitic and
alkaline) magmatism as the principal generator of base-
and precious-metal mineralization, but has also stressed
that other magmatic suites characteristic of extensional en-
vironments, which immediately followed the cessation of
subduction and concomitant calc-alkaline activity, can also
be metal-bearing. Thus, a bimodal basalt-rhyolite suite in
the western U.S.A. hosts precious- and subordinate base-
metal deposits of epithermal type and volcanogenic tin de-
posits, and a basaltic suite akin to the ocean-island suite in
Fiji hosts epithermal gold telluride deposits. Therefore
exploration efforts should perhaps be extended beyond the
strictly calc-alkaline volcanic environments, but should not
incorporate all volcanic settings: quite clearly extensional
volcano-tectonic environments characteristic of the early
development stages of divergent plate margins are devoid
of significant metallic mineralization of the types discussed
above, as are intra-plate ocean-island settings unrelated to
preceding calc-alkaline activity.
Acknowledgment
Thanks are due to the geologists, too numerous to mention
individually, who have accompanied the writer during the
examination of ore deposits, mentioned above, in Chile,
Argentina, Bolivia, Peru, Costa Rica, El Salvador, Mexico,
Spain and Pakistan.
References
1. Schmitt H. Origin of the 'epithermal' mineral
deposits. Econ. Geol., 45, 1950, 191-200.
2. Schmitt H. The fumarolic-hot spring and 'epithermal'
mineral deposit environment. Q. Colo. Sch. Mines, 45,
1950, 209-29.
3. Lindgren W. Mineral deposits, 4th edn (New York:
McGraw-Hill, 1933), 930 p.
4. Nolan T. B. Epithermal precious-metal deposits. In
Ore deposits of the western States (New York: AIME, 1933),
623-40.
5. Burbank W.S. Epithermal base-metal deposits. InOre
deposits of the western States (New York: AIME, 1933),
641-52.
6. Burbank W. S. Problems of wall-rock alteration in
shallow volcanic environments. Q. Colo. Sch. Mines, 45,
1950, 287-319.
7. WisserE. The epithermal precious-metal province of
northwest Mexico. Rep. Nevada Bur. Mines no. 13, pt C,
1966, 63-92.
8. McDowell F. W. and Clabaugh S. E. Relations of ignim-
brites in the Sierra Madre Occidental to the tectonic history
of western Mexico. Abstr. Progr. geoL Soc. Am., 8, 1976,
609--10.
9. Geyne A. R. eta/. Geolog(a y yacimientos minerales
del distrito de Pachuca - Real del Monte, Estado de Hidalgo,
M~xico. Publ. Con~ Recursos naturales no renovables
M~xico 5E, 1963, 222 p.
10. Smith, D. M. Jr. Geologi'a de la Mina de Tayoltita,
distrito de San Dimas, Durango, Mexico. In Algunos
yacimientos rninerales de la Repbblica Mexicana. Geolog[a,
m~todos de extraccibn y de beneficio (Mexico City: AIME,
1973), 163-7.
11. GrossW. H. New ore discovery and source of silver-
gold veins, Guanajuato, Mexico. Econ. Geol., 70, 1975,
1175-89.
12. Petruk W. and Owens D. Some mineralogical char-
acteristics of the silver deposits in the Guanajuato mining
district, Mexico. Econ. Geol., 69, 1974, 1078-85.
13. Roberts R. J. and Irving E. M. Mineral deposits of
Central America. Bull. U.S. geol. Surv. 1034, 1957, 205 p.
14. Levy E. La metalog6nesis en America Central. Publ.
geol. Inst. Centraomericano de Invest. y Tecnologfa
Industrial no. 3, 1970, 17-57.
15. Sillitoe R. H. Lead-silver, manganese, and native sulfur
mineralization within a stratovolcano, El Queva, northwest
Argentina. Econ. Geol., 70, 1975, 1190-201.
16. Noble D. C. and Bowman H. R. Geochemical and
field constraints on the depth of differentiation and eruptive
history of silicic calc-alkalic magmas related to base and
precious metal mineralization, Andes of central Peru. Abstr.
Progr. geol. Soc. Am., 8, 1976, 615.
17. Ahlfeld F. Die Metallprovinz des Altiplano (Bolivien).
Neues Jahrb. Min. Abh., A85, 1953, 1-58.
18. Ahlfeld F. Metallogenetic epochs and provinces of
Bolivia. Miner. Deposita, 2, 1967, 291-311.
19. SalasO. R. et al. Geologl'a y recursos minerales del
departamento de Arica, provinc{a de Tarapac~. Bol. Inst.
Invest. geol. Chile no. 21, 1966, 114 p.
20. Bellido B. E. and de Montreuil D. L. Aspectos
generales de la rnetalogen/a de/Per5 (Lima, Peru: Serv. de
Geol. y Mineri'a, 1972), 149 p. (Geol. Econbmica no. I).
21. Steven T. A. Middle Tertiary volcanic field in the
southern Rocky Mountains. Mern. geol. Soc. Am. 144,
1975, 75-94.
22. StevenT. A. and Lipman P. W. Calderas of the San
Juan volcanic field, southwestern Colorado. Prof. Pap. U.S.
geol. Surv. 958, 1976, 35 p.
23. Christiansen R. L. and Lipman P. W. Cenozoic
volcanism and plate-tectonic evolution of the western
United States. I1. Late Cenozoic. Phil. Trans. R. Soc., A271,
1972, 249-84.
24. Lipman P. W. etal. Multiple ages of Mid-Tertiary min-
eralization and alteration in the western San Juan Mountains,
Colorado. Econ. Geol., 71, 1976, 571-88.
25. Fisher F. S. and Leedy W. P. Geochemical characteris-
tics of mineralized breccia pipes in the Red Mountain
district, San Juan Mountains, Colorado. Bull. U.S. geol.
Surv. 1381, 1973, 43 p.
26. Mayor J. N. and Fisher F. S. Middle Tertiary replace-
ment ore bodies and associated veins in the northWest San
Juan Mountains, Colorado. Econ. Geol., 67, 1972, 214-30.
27. StevenT. A. Luedke R. G. and Lipman P. W.
Relations of mineralization to calderas in the San Juan
volcanic field, southwestern Colorado. J. Res. U.S. geol.
Surv., 2, 1974, 405-9.
28. Steven T. A. and Ratt~ J. C. Geology and ore deposits
of the Summitville district, San Juan Mountains, Colorado.
Prof. Pap. U.S. geol. Surv.343, 1960, 70 p.
113
29. Steven T. A. and Ratt~ J. C. Geology and structural
control of ore deposition in the Creede district, San Juan
Mountains, Colorado. Prof. Pap. U.S. geol. Surv. 487, 1965,
87 p.
30. Steven T. A. and Eaton G. P. Environment of ore
deposition in the Creede mining district, San Juan
Mountains, Colorado: 1. Geologic, hydrologic, and geo-
physical setting. Econ. Geol., 70, 1975, 1023-37.
31. Silberman M. L. and McKee E. H. Ages of Tertiary
volcanic rocks and hydrothermal precious-metal deposits in
central and western Nevada. Rep. Nevada Bur. Mines Geol.
no. 19, 1974, 67-72.
32. Albers J. P. and Kleinhampl F. J. Spatial relation of
mineral deposits to Tertiary volcanic centers in Nevada.
Prof. Pap. U.S. geol. Surv. 700-C, 1970, C1-10.
33. Kleinhampl F. J. etal. Aeromagnetic and limited
gravity studies and generalized geology of the Bodie Hills
region, Nevada and California. Bull. U.S. geol. Surv. 1384,
1975, 38 p.
34. Ashley R. P. Goldfield mining district, Nevada. Rep.
Nevada Bur. Mines Geol. no. 19, 1974, 49-66.
35. Harvey R. D. and Vitaliano C. J. Wall-rock alteration
in the Goldfield district, Nevada. J. Geol., 72, 1964,
564-79.
36. Nolan T. B. The underground geology of the Tonopah
mining district, Nevada. Bull. Nevada Univ., 29, no. 5,
1935, 49 p.
37. Sato T. Distribution and geological setting of the
Kuroko deposits. In Geology of Kuroko deposits Ishihara
S. et al. eds (Tokyo: Society of Mining Geologists of
Japan, 1974), 1-9. (Spec. Issue no. 6).
38. Tatsumi T. Sekine Y. and Kanehira K. Mineral de-
posits of volcanic affinity in Japan: metallogeny. In
Volcanism and ore genesis Tatsumi T. ed. (Tokyo:
University of Tokyo Press, 1970), 3-47.
39. Nishiwaki C., Matsukuma T. and Urashima Y.
Neogenegold-silver ores in Japan. In Proc. IMA - IAGOD
Meetings, "70: IAGOD volume Takeuchi Y. ed. (Tokyo:
Society of Mining Geologists of Japan, 1971), 409-17.
(Spec. Issue no. 3).
40. Yamaoka K. and Ueda Y. The mineralization age of
gold-silver deposits. In Proc. Joint Autumn Meetings
(Japan: Akita University, 1973), C29 (Japanese text).
41. Lambert I. B. and Sato T. The Kuroko and associated
ore deposits of Japan: a review of their features and
metallogenesis. Econ. Geol., 69, 1974, 1215-36.
42. Williams G. J. Tertiary propylitic mineralization. In
Economic geology of New Zealand (Melbourne: Australasian
Institute of Mining and Metallurgy, 1965), 87-126. (Proc.
8th Commonw. Min. Metal/. Congr. vol. 4).
43. Ramsay W. R. H. and Kobe H.W. Great Barrier
Island silver-gold deposits, Hauraki province, New Zealand.
Miner. Deposita, 9, 1974, 143-53.
44. Wodzicki A. and Weissberg B. G. Structural control
of base metal mineralisation at the Tui mine, Te Aroha,
New Zealand. N. Z. J. Geol. Geophys., 13, 1970, 610-30.
45. Bryner L. Ore deposits of the Philippines - an
introduction to their geology. Econ. Geol., 64, 1969,
644 -66.
46. Rebek R.J. Edie Creek and Wau gold lodes. In
Economic geology of Australia and Papua New Guinea I.
114
Metals. Knight C. L. ed. (Parkville, Victoria; Australasian
Institute of Mining and Metallurgy, 1975), 867-72.
(Mongr. Set. no. 5).
47. Denholm L.J. Lode structures and ore shoots at
Vatukoula, Fiji. Proc. Australas. Inst. Min. Metall. 222,
1967, 73-83.
48. Gill J. B. and McDougall I. Biostratigraphic and
geological significance of Miocene--Pliocene volcanism in
Fiji. Nature, Lond., 241, 1973, 176-80.
49. Martrn Vivaldi J. L. Sierra J. and Leal G. Some aspects
of the mineralization and wall-rock alteration in the Rodal-
quilar gold-field, S. E. Spain. In Proc. IMA - IAGOD
Meetings "70, Joint Symposium volume (Toyko: Society of
Mining Geologists of Japan, 1971), 145-52. (Spec. Issue
no. 2).
50. McKee E. H. etal. Chronology of late Tertiary
volcanism and mineralization, Huachocolpa district, central
Peru. Econ. Geol., 70, 1975, 388-90.
51. Silberman M. L. and Ashley R. P. Age of ore deposi-
tion at Goldfield, Nevada, from potassium-argon dating
of alunite. Econ. Geol., 65, 1970, 352-4.
52. Bonham H. F. Jr. Garside L. J. and Silberman M. L.
K - A r ages of ore deposition at Tonopah, Nevada. Isochron/
West no. 4, 1972, 5 - 6 .
53. Siiberman M. L. eta/. K - A t ages of volcanic rocks
and gold-bearing quartz-adularia veins in the Bodie mining
district, Mono County, California. Econ. Geol., 67, 1972,
597-603.
54. Mehnert H. H. Lipman P. W. and Steven T. A. Age of
mineralization at Summitville, Colorado, as indicated by
K - A r dating of alunite. Econ. Geol., 68, 1973, 399-401.
55. Robinson B. W. The origin of mineralization at the
Tui mine, Te Aroha, New Zealand, in the light of stable
isotope studies. Econ. Geol., 69, 1974, 910-25.
56. NashJ.T. Fluid inclusion studies of vein, pipe, and
replacement deposits, northwestern San Juan Mountains,
Colorado. Econ. Geol., 70, 1975, 1448-62.
57. Nash J.T. Fluid-inclusion studies of some gold
deposits in Nevada. Prof. Pap. U.S. geol. Surv. 800-C,
1972, C15-9.
58. O'NeiI.J.R. etal. Stable isotope and chemical re-
lations during mineralization in the Bodie mining district,
Mono County, California. Econ. Geol., 68, 1973, 765-84.
59. Bethke P. M. Rye R. O. and Barton P. B. Jr.
Hydrogen, oxygen, and sulfur isotopic composition of ore
fluids in the Creede district, Mineral County, Colorado.
Abstr. Progr. geol. Soc. Am., 5, 1973, 549.
60. Dreier J. E. Jr. Fluid inclusion geothermometry of
the Pachuca-Real del Monte silver district, Hidalgo, Mexico.
In Base Metal and Precious Metal Districts of New Mexico
andArizona (Silver City, New Mexico: Western New Mexico
University, 1975), 7. (Abstracts.)
61. Miyazawa T. eta/. Formation temperatures of veins
in Japan. In Proc. IMA-IAGOD Meetings, "70: IAGOD
volume Takeuchi Y. ed. (Tokyo: Society of Mining
Geologists of Japan, 1971 ), 3 4 0 - 4 (Spec. Issue no. 3).
62. Enjoji M. Studies on fluid inclusions as the media of
ore formation. Sci. Rep. Tokyo Kyoiku Daigaku, sect. C,
11, 1972, 79-126.
63. El Shatoury H. M. Takenouchi S. and Imai H. Fluid
inclusion studies on the Toyoha mine, Hokkaido. In 4th
IA GOD symposium, Varna, abstracts of papers (Sofia,
Bulgaria, 1974), 275-7.
54. Folinsbee R. E. eta/. Chinkuashih - a golcl-pyrite-
enargite-barite hydrothermal deposit in Taiwan. Mere.
geol. Soc. Am. 135, 1972, 323-35.
65. Kamilli R. J. and Ohmoto H. Isotopic and chemical
evolution of the hydrothermal fluid responsible for the
P b - Z n - A g - A u mineralization of the Finlandia vein,
Colqui district, Department of Lima, Peru. Econ. Geol.,
70, 1975, 1322 (Abstract).
66. Sawkins F. J. and Rye R.O. Fluid-inclusion and
stable isotope studies indicating mixing of magmatic and
meteoric waters, Caudalosa silver deposit, central Andes,
Peru. In 4th IAGOD symposium, Varna, abstracts of
papers (Sofia, Bulgaria, 1974), 237-8.
67. Rye R. O. and Sawkins F.J. Fluid inclusion and
stable isotope studies on the Casapalca A g - P b - Z n - C u
deposit, central Andes, Peru. Econ. Geol., 69, 1974,
181-205.
68. O'Neil J. R. and Silberman M. L. Stable isotope
relations in epithermal gold-silver deposits. Econ. Geol.,
69, 1974, 902-9.
69. Taylor H. P. Jr. O18/O 16 evidence for meteoric-
hydrothermal alteration and ore deposition in the Tonopah,
Comstock Lode, and Goldfield mining districts, Nevada.
Econ. Geol., 68, 1973, 747-64.
70. Taylor H. P. Jr. The application of oxygen and
hydrogen isotope studies to problems of hydrothermal alter-
ation and ore deposition. Econ. Geol., 69, 1974, 843-83.
71. CasadevallT. Ohmoto H. and Rye R. O. Sunnyside
mine, San Juan County, Colorado: results of mineralogic,
fluid-inclusion, and stable-isotope studies. Econ. Geol., 69,
1974, 1178 (Abstract).
72. White D. E. Diverse origins of hydrothermal fluids.
Econ. Geol., 69, 1974, 954-73.
73. Smith R. L. and Bailey R.A. Resurgent cauldrons.
Mere. geol. Soc. Am. 116, 1968, 613-62.
74. Elston W. E. Rhodes R. C. and Erb E. E. Controls of
mineralization by Mid-Tertiary volcanic centers, south-
western New Mexico. I n Base metal and precious metal
districts of New Mexico and Arizona (Silver City, New
Mexico: Western New Mexico University, 1975), 11-3.
(Abstract)
75. McAnulty W. N. Jr. Resurgent cauldrons and
associated mineralization, southwestern Trans-Pecos, Texas.
Abstr. Progr. geol. Soc. Am., 8, 1976, 608-9.
76. Stokl J. and KoneEn~ V. Bansk~ Stiavnica ore deposit
and its position within caldera. In 4th IAGOD symposium,
Varna, abstracts of papers (Sofia, Bulgaria, 1974), 355-6.
77. Wisser E. Relation of ore deposition to doming in the
North American cordillera. Mere. geol. Soc. Am. 77, 1960,
117p.
78. Morris H. T. Volcanism and ore in the Tintic mining
district, west-central United States. In 4th IAGOD
symposium, Varna, abstracts of papers (Sofia, Bulgaria,
1974), 158-9.
79. McKee E. Ash-flow sheets and calderas: their relation-
ship to ore deposits in Nevada. Abstr. Progr. geol. Soc. Am.,
8, 1976, 610-1.
80. Sato T. Model for ore forming solutions and ore form-
ing environments; Kuroko vs. veins in Miocene Green Tuff
region of Japan. Bull. geol. Surv. Japan, 23, no. 8, 1972,
13-22.
81. Ohmoto H. and Rye R. O. Hydrogen and oxygen
isotopic compositions of fluid inclusions in the Kuroko
deposits, Japan. Econ. Geol., 69, 1974, 947-53.
82. Burbank W. S. and Luedke R. G. Geology and ore
deposits of the western San Juan Mountains, Colorado.
In Ore deposits of the United States, 1933-1967. The
Graton-Sales volume Ridge J. D.ed. (New York: AIME,
1968), 714-33.
83. Taylor H. P. Jr. and Forester R. W. Low-O TM
igneous rocks from the intrusive complexes of Skye, Mull,
and Ardnamurchan, western Scotland. J. Petrology, 12,
1971,465-97.
84. White D. E. Thermal springs and epithermal ore
deposits. Econ. Geol. 50th anniversary volume, 1955,
99-154.
85. White D. E. Mercury and base-metal deposits with
associated thermal and mineral waters. In Geochemistry of
hydrothermal ore deposits Barnes H. L. ed. (New York:
Holt, Rinehart and Winston, 1967), 575-631.
86. Browne P. R. L. Mineralisation in the Broadlands
geothermal field, Taupo volcanic zone, New Zealand. In
Proc. IMA-IAGOD Meetings "70, Joint symposium volume
(Tokyo: Society of Mining Geologists of Japan, 1971),
54-75. (Spec. Issue no. 2).
87. WeissbergB. G. Gold-silver ore-grade precipitates
from New Zealand thermal waters. Econ. Geol., 64, 1969,
95-108.
88. Browne P. R. L. Roedder E. and Wodzicki A.
Comparison of past and present geothermal waters from a
study of fluid inclusions, Broadlands field, New Zealand.
E| Tran~ Am. geophys. Union, 55, 1974, 456.
89. Hewett D. F. Fleischer M. and Conklin N. Deposits
of the manganese oxides, supplement. Econ. Geol., 58,
1963, 1-51.
90. Sillitoe R. H. The tops and bottoms of porphyry
copper deposits. Econ. Geol., 68, 1973, 799-815.
91. Gustafson L. B. and Hunt J. P. The porphyry copper
deposit at El Salvador, Chile. Econ. Geol., 70, 1975,
857-912.
92. Corn R. M. Alteration-mineralization zoning, Red
Mountain;Arizona. Econ. Geol., 70, 1975, 1437-47.
93. Sillitoe R. H. Andean mineralization: a model for
the metallogeny of convergent plate margins. Spec. Pap.
Geol. As~ Can., 14, 1976, 59-100.
94. Atwater T. Implications of plate tectonics for the
Cenozoic tectonic evolution of western North America.
Bull. geol. Soc. Am., 81, 1970, 3513-36.
95. Kuno H. Lateral variation of basalt magma type
across continental margins and island arcs. Bull. volcan.,
29, 1966, 195-222.
96. Ishihara S. Igarashi T. and Nishiwaki C. A re-examin-
ation of the regional distribution of the late Cenozoic ore
deposits in the east Japan arc. Bull. geol. Soc.Am., 86,
1975, 292-4.
97. Lipman P. W. Prostka H. J. and Christiansen R. L.
Cenozoic volcanism and plate-tectonic evolution of the
western United States. 1. Early and Middle Cenozoic.
Phil. Trans. R. Soc., A271, 1972, 217--48.
115
98. Damon P. E. Mid-Tertiary ash flows of the southern
Basin and Range province: time, space, chemistry and
genesis. Abstr. Progr. geol. Soc. Am., 8, 1976, 581-2.
99. Gill J. B. From island arc to oceanic islands: Fiji,
southwestern Pacific. Geology, 4, 1976, 123-6.
100. Carter W. D. Origin of "manto-type" copper deposits
of the Cabildo mining district, central Chile. In Rep. 21st
Int. Geol. Congr. (Copenhagen: The Congress, 1960), pt
16, 17-28.
101. Ruiz F. C. et aL Geologfa y yacimientos metallferos
de Chile (Santiago, Chile: Inst. Invest. Geol., 1965), 305 p.
102. Ruiz C. et aL Strata-bound copper sulphide deposits
of Chile. In Proc. I M A - I A G O D Meetings, "70: IAGOD
volume Takeuchi Y. ed. (Tokyo: Society of Mining Geolo-
gists of Japan, 1971 ), 252-60. (Spec. Issue no. 3).
103. Lortie R. B. and Clark A. H. Stratabound fumarolic
copper deposits in rhyolitic lavas and ash-flow tufts, Copiap0
district, Atacama, Chile. In 4th IAGOD symposium, Varna,
abstracts of papers (Sofia, Bulgaria, 1974), 154-5.
104. Ripley E. M. Mineralogy and sulfur isotopes of the
cupriferous sulfide ores at the Rail mine, Peru. Econ. GeoL,
70, 1975, 1328 (Abstract).
105. Taylor P. S. and Stoiber R. E. Soluble material on
ash from active Central American volcanoes. Bull. geol. Soc.
Am., 84, 1973, 1031-42.
106. Zei~ E. C. The Valley of Ten Thousand Smokes, 1.
The fumarolic incrustations and their bearing on ore deposi-
tion. Natn. Geograph. Soc. Contrib. Tech. Pap., Katmai
Series, 1, no. 4, 1929, 1-61.
107. Nisterenko G. V. etaL Temperaturas y presiones de
formaci6n de algunos yacimientos cuprlferos de Chile.
RevistageoL Chile no. 1, 1974, 74-80.
108. Park C. F. Jr. A magnetite " f l o w " in northern Chile.
Econ. GeoL, 56, 1961,431-6.
109. Lyons J. I. Jr. and Clabaugh S. E. Pyroclastic and
extrusive iron ore at Durango, Mexico. Abstr. Progr. geoL
Soc. Am., 5, 1973, 723.
110. Sillitoe R. H. Metallogenic evolution of a collisional
mountain belt in Pakistan: a preliminary analysis. JlgeoL
Soc. Lond., in press.
111. Frutos J. J. and Oyarzt~n M. J. Tectonic and ~geo-
chemical evidence concerning the genesis of El Laco
magnetite lava flow deposits, Chile. Econ. GeoL, 70, 1975,
988-90.
112. Philpotts A. R. Origin of certain iron-titanium oxide
and apatite rocks. Econ. Geol., 62, 1967, 303-15.
113. Pan Y. -S. and Ypma P. J. M. The Mexican type tin
deposits - its occurrence, chemistry and physical conditions
of deposition. Abstr. Progr. geol. Soc. Am., 5, 1973, 762.
114. Foshag W. F. and Fries C. Jr. Tin deposits of the
Republic of Mexico. Bull U.S. geol. Surv. 935oC, 1942,
99-176.
115. Ypma P. J. M. and Simons J. H. Genetical aspects of
the tin mineralization in Durango, Mexico. In A Second
technical conference on tin, Bangkok 1969 Fox W. ed.
(London International Tin Council, 1970), 179-91.
116. Pan Y. -S. and Ypma P. J. M. Heating and recrystalli-
zation of wood-tin: an indication of low temperature
genesis of cassiterite and wood-tin veins in volcanics. Abstr.
Progr. geoL Soc. Am., 6, 1974, 62-3.
116
117. Fries C. Jr. Tin deposits of the Black Range, Catron
and Sierra Counties, New Mexico - a preliminary report.
Bull. U.S. geol. Surv. 922-M, 1940, 355-70.
118. Fries C. Jr. Tin deposits of northern Lander County,
Nevada. Bull. U.S. geol. Surv. '931-L, 1942, 279-94.
119. Ahlfeld F. and Schneider-Scherbina A. Los yacimien-
tos minerales y de hidrocarburos de Bolivia (La Paz,
Bolivia: Dept. Nac. Geol., 1964), 388 p. (Bol. no. 5).
120. Lufkin J. L. Oxide minerals in miarolitic rhyolite,
Black Range, New Mexico. Am. Mineralog., 61, 1976,
425-30.
121. Goossens P. J. L'apport m~tallif~re "direct" du
volcanisme continental. Miner. Deposita, 10, 1975, 43-5.
122. Sillitoe R. H. Halls C. and Grant J. N. Porphyry tin
deposits in Bolivia. Econ. Geol., 70, 1975, 913-27.
123. Geyne A. R. Las rocas volc,~nicas y los yacimientos
argent~feros del distrito minero de Pachuca-Real del Monte,
Estado de Hidalgo. In 20th Int. GeoL Congr., excursiones
A - 3 y C-1 (Mexico: The Congress, 1956), 47-57.
124. Dengo G. et aL Metallogenic Map of Central America
1:2,000,000 (Guatemala: Inst. Centroamericano de Invest.
y Tecnolog~a Industrial, 1969).
. _
125. Sugimura A. Zonal arrangement of some geophysical
and petrological features in Japan and its environs. J. Fac.
Sci. Univ. Tokyo, sect. 2, 12, 1960, 133-53.
126. Carr M. J. Stoiber R. E. and Drake C. L. Discon-
tinuities in the deep seismic zones under the Japanese arcs.
Bull. geoL Soc. Am., 84, 1973, 2917-30.
127. Elston W. E. etal. Tertiary volcanic rocks, Mogollon-
Datil province, New Mexico, and surrounding region:
K - A r dates, patterns of eruption, and periods of mineraliz-
ation. Bull. geoL Soc. Am., 84, 1973, 2259-74.
128. Bagby W. C. et aL Petrology of Tertiary ash-flow tufts
and related volcanics from the Barranca de Batopilas and
Creel regions of the Sierra Madre Occidental, Chihuahua,
Mexico. Abstr. Progr. geoL Soc. Am., 8, 1976,566.
129. Ashley Ro P. and Silberman M. L. Direct dating of
mineralization at Goldfield, Nevada, by potassium-argon
and fission-track methods. Econ. GeoL, 71, 1976, 904-24.
130. Bethke P. M. etal. Environment of ore deposition in
the Creede mining district, San Juan Mountains, Colorado:
I1. Age of mineralization. Econ. GeoL, 71, 1976, 1006-11.
131. Casadevall T. Sunnyside mine, Eureka mining dis-
trict, San Juan county, Colorado: geochemistry of gold and
base metal ore formation in the volcanic environment. In
25th Int. GeoL Congr., abstracts (Sydney: Tha Congress,
1976), 3, 805.

Igneous geology and the evolution
of hydrothermal systems in some
sub-volcanic tin deposits of Bolivia
J. N. Grant
Department o f Geology, Royal School o f Mines,
Imperial College, London, England
C. Halls
Department of Geology, Royal School of Mines,
Imperial College, London, England
W. Avila
Servicio Geol6gico de Bolivia ( G E O B O L ) , La Paz,
Bolivia
G. Avila
Servicio Geol6gico de Bolivia ( G E O B O L ) , La Paz,
Bolivia
5 5 2 . 3 : 5 5 3 . 2 1 6 . 5 : 553.45(84)
Synopsis
The tin deposits o f the southern part o f the Cordillera
Oriental o f Bolivia consist of complex vein systems and
pervasive disseminated mineralization within zones of
strong hydrothermal alteration in, or at the margins of,
eruptive centres o f late Tertiary age. A t some centres
erosion has removed the volcanic superstructure completely
and only intrusive rocks are preserved (Llallagua, San Pablo
de Morococala). A t others, both the intrusives and coeval
volcanics are preserved (Chorolque, Oruro, Potosi), whereas
in some cases only the partially eroded volcanic super-
structure is exposed and no unequivocally intrusive rocks
are known (Chocaya, Tatasi).
In general, the geological relationships at all these de-
posits suggest that the mineralization is the product of
hydrothermal systems generated in the inner, deeper regions
of terrestrial strato-volcanoes.
The igneous rocks most closely associated with minerali-
zation are strongly altered quartz porphyries. Their original
composition was probably rhyodacite. Where stocks are
preserved, they are complex bodies, often conical in vertical
section, narrowing with depth. They show evidence of
several phases of intrusion, igneous brecciation and explosive
hydrothermal brecciation, which pre-date the formation of
the major vein systems.
Silicate alteration a.ssemblages include quartz-tourma-
line, tourmaline-sericite and sericite-clay, which often
show rough concentric zonation in that order outward from
the centre. An outermost aureole of propylitic alteration
is sometimes present where there are remnants o f the
volcanic superstructure. Alteration is pervasive and not
controlled by the major veins. Low-grade cassiterite minera-
lization is dispersed throughout the inner zones.
Fluid inclusion studies at Chorolque show that the hydro-
thermal system was initiated by the separation o f a highly
saline brine, or melt, of complex chemistry, which formed
at temperatures above 500~ The pervasive quartz-tourma-
line alteration, initial fracturing of the igneous rocks of the
vent and the hornfelsing of the adjacent sedimentary rocks
were accomplished while temperatures fell to below 450~
and intermittent boiling took place. Widespread dissemin-
ated cassiterite was deposited during this phase. The earliest
stages of growth of the major quartz-cassiterite veins took
place at temperatures of around 400~ from a fluid the com-
position o f which was similar to that o f the earliest fluid,
though substantially diluted (about 40 wt % NaCI equivalent)
Vein growth continued while temperatures fell to below
200 ~C and salinity decreased. Most cassiterite deposition
seems to have taken place in the temperature range 3 0 0 -
250~ and may have been accompanied by a major decrease
in the salinity o f the fluids.
The data obtained from the fluid inclusion studies, taken
in conjunction with the gross geometry of the mineralization,
suggest that hydrothermal processes at the volcanic centres
were initially controlled by failure of the balance between
confining fithostatic pressure and the pressure o f the hydrous
fluid residuum in the differentiated parts of the rhyodacite
magma, leading to pervasive hydraulic fracturing, brecciation
and alteration, o f which the early generation o f cassiterite
mineralization was an integral part. The upper levels of the
mineralized volcanic structures appear to have stabilized
following the hydrothermal exhaustion of the apical portions
of the sub-volcanic stocks. The focus of hydrothermal
activity must then have been relocated at greater depths from
which mineralization was controlled by the interplay of
tectonic stresses in the volcanic infrastructure and surround-
ing basement and the build-up of hydrothermal fluid
pressure in parallel vein-fault systems.
This concept of a bimodal style of mineralization related
to a retreating hydrothermal focus can lead to a clearer
understanding of the xenothermal or "telescoped" type o f
mineral and temperature zonation in sub-volcanic systems.
Temperature data obtained from fluid inclusion studies
show, however, that mineral paragenesis is not an unam-
biguous criterion o f temperature zonation and that the two
have to be considered independently before a combined
synthesis becomes meaningful
There are scattered tin deposits at intervals on the continent
side of the cordillera of the western Americas from Alaska
to Argentina. The most important section is the Bolivian
tin belt, which extends for about 800 km through the
Eastern Cordillera of the Central Andes. 3 The complete
explanation of the confinement of the tin deposits to this
restricted landward belt in the Eastern Andes remains a
subject of active discussion. 5 There is, however, a striking
parallelism between the main metallogenic belts and the
morphotectonic divisions of the Central Andes within
which they lie. 1,2,3 This led Sillitoe 21 to propose that the
laterally distinct metal provinces were produced as a result
of progressive liberation of the various metals from the
oceanic lithosphere in the underlying zone of subduction.
Systematic release of the metals was believed to be due to
dehydration, phase changes and partial melting governed by
the increase in pressure and temperature at increasing
depths. Various models have been proposed for the tem-
perature distribution in a downgoing plate, 12~7,28 and the
precise nature of phase changes that are taking place at
advanced depths is still under discussion, 33 but it follows
that if the source of the tin lies in the downgoing litho-
spheric plate and if the present pattern of seismic activity
beneath the Eastern Cordillera 14 can be accepted as similar
to that prevailing during late Mesozoic and Tertiary times,
the tin province must be explained in terms of Special phase
changes initiated only when the downgoing lithospheric
plate reaches depths of the order of 300 km. The break-
down of refractory phases such as phlogopite could be in-
voked, 7 supplying potassium as well as tin to the rising
magma. The tin in the calc-alkaline eruptives may also
117
have originated in the mantle wedge overlying the Benioff
zone, being progressively concentrated by a scavenging or
'zone-refining' process, 6,13 which would abstract tin to-
gether with other incompatible elements and concentrate
them in a rising magma or a partial melt at the base of the
continental crust.
It is difficult to avoid the conclusion that the generation
of Cordilleran tin deposits of the Bolivian type is linked in
some way to the subduction process. At present there are
insufficient petrochemical or isotopic data with which to
evaluate the problem, but whatever the hypothesis invoked,
account must be taken of the wide range of geological
factors that influence the tectonic framework of this
complex segment of the Andean orogen. 2
Bolivia produces the second largest amount of tin among
the non-communist countries of the world. Its economy
and the livelihood of a substantial section of the population
depend on the tin industry. The history of major mining
activity dates back to the Spanish conquest in the sixteenth
century, when the colonial empire derived much of its
wealth from the silver mines of the region. Most of the
important tin deposits were discovered during that period,
and were mined initially for the fabulously rich silver ores,
which had formed, partly by supergene enrichment pro-
cesses, in the upper parts of the vein systems. Some of
these deposits are still being mined today, mainly for the
tin ores found beneath the silver-rich zones. Fig. 1 Regional geology of the Central Andes
The Bolivian tin belt contains some of the greatest hard-
rock tin deposits of the world. The Llallagua deposit, for
out most of the orogeny it was emergent, bounded on the
example, is estimated to have produced more than 500 000
west by the Altiplano molasse basin, and on the east by the
tons of tin since production started at the beginning of
marine basin, the folded sedimentary rocks of which now
this centl,ry. 3 Considering their size and economic impor-
form the sub-Andean foothill zone. Within the Cordillera
tance, the geology of these deposits is not particularly well
there are some remnants of epicontinental sedimentary
known. Some classical studies were carried out, mainly by
formations of Cretaceous age, which once probably covered
American geologists, in the period when they were operated
the region but were largely removed by subsequent erosion 2
by private mining corporations prior to nationalization in
The remainder of th~ Cordillera is made up of shales and
the early 1950s. This work was naturally concerned mainly
sandstones of Cambrian to Devonian age and igneous rocks
with the structure and mineralogy of the vein systems, with
of the Andean orogeny. The sediments are strongly folded
less emphasis on understanding the overall relationships
and faulted, but metamorphism is generally weak or absent.
and evolutionary sequence of igneous and hydrothermal
The igneous rocks are acid to intermediate plutons and
activity. The progressive exhaustion of the high-grade vein
ores is now an urgent problem for the Bolivian tin industry. high-level stocks and eruptive complexes, which range in
age from Triassic to Upper Miocene. 5'9.11 The tin deposits
It is, however, apparent that some of the presently known
deposits contain enormous reserves of potentially economic are associated with these. There are also some extensive
low-grade tin mineralization, the distribution of which is not post-mineral ignimbrite sheets of Pliocene age.
related to the geometry of the well-known vein systems. The tin belt extends throughout the Eastern Cordillera
More comprehensive geological concepts than those pre- in Bolivia, from near the Peruvian border in the north to
viously applied should prove useful in the proper evaluation northern Argentina in the south (Fig. 2). To the west there
of this potential resource. is a well-defined province of polymetallic deposits in the
Altiplano, which are associated with Miocene-Pliocene high-
Geology of the tin belt level dacitic and andesitic igneous rocks. 1 The tin belt has
The Eastern Cordillera of the Central Andes is an elongated been divided into two sections. 31 In the north (Cordilleras
massif, made up mainly of a great thickness of marine Real and Quimsa Cruz) the deposits are mainly tin and
clastic sedimentary rocks of Lower Palaeozoic age (Fig. 1).2 tungsten veins associated with granodiorite batholiths. The
The sedimentary basin in which they were formed extended batholiths range in age from Triassic to Miocene and those
right across the width of the Andes almost to the present of the Miocene occur in the southern part of this section
coastal region. It was underlain by Precambrian metamor- and are the most highly mineralized. 9 These deposits are
phic rocks, the extension of the Brazilian and Argentina numerous but, with one or two exceptions (notably
cratons, and these are now exposed in the coastal reg.ions of Caracoles), they are individually small.
southern Peru and northern Chile lO and in the Eastern In central and southern Bolivia the deposits are mostly
Cordillera in Peru and northern Argentina. 4,32 The pre- associated with acid eruptive complexes of Tertiary age.
Andean tectonic history of the region has been partly ob- They are complex vein systems carrying tin, tin-silver or
scured by later events, but there were important periods of tin with base metals and silver. In some cases there is also
regional :folding, and there are granitic plutons in the widespread disseminated mineralization in the altered
Eastern Cordillera in Peru and Argentina that give Lower igneous rocks. There are also a few important deposits
Palaeozoic ages.24~6 formed by vein systems in sedimentary rocks with which
The Cordillera owes its morphotectonic identity to the there is little or no direct evidence of igneous activity (e.g.
Andean orogeny, which began in the Triassic. 14 Through- Huanuni and Colquiri). These are probably related to high-

118

Fig. 2 The Bolivian tin belt: location map
level stocks that failed to reach the surface and, hence, are
genetically similar to the deposits within eruptive complexes.
The sub-volcanic deposits of the central and southern part
of the tin belt are fewer, but, in general, individually larger
than those associated with the batholiths of the north, and
they account for the greater proportion of Bolivian tin
production.
Radiometric age dating currently in progress suggests that
this twofold subdivision of the tin belt should be replaced
by a threefold subdivision on the basis of the ages and some
of the geochemical and morphological characteristics of
the deposits. The Cordillera Real appears to form a dis-
tinct tin-tungsten sub-province associated with batholiths
of Triassic to Lower Jurassic age; the somewhat higher-level
plutons of the Cordillera Quimsa Cruz and Santa Vera
Cruz, together with the eruptive complexes of central
Bolivia as far south as Colquechaca, form a Lower Miocene
age province and the deposits include almost all the major
tin producers; the eruptive complexes of the southern part
of the belt, from Potosi southward, form a distinct Upper
Miocene age group and the deposits are characterized by
high silver and base-metal contents. The difference in metal
content displayed by these sub-provinces is, however, pro-
bably a function of the depth of erosion rather than a
change in the chemistry of the parent magmas with time. 11
Character of the eruptive complexes
A range exists from purely volcanic complexes, at which
no intrusive rocks other than dykes are recognizable,
to small stock-like bodies of intrusive porphyry and breccia.
The essentially volcanic types, such as Chocaya, Tatasi and
Colquechaca, consists of acid lavas, lava domes, pyroclastics
and volcanic breccias, which rest unconformably on the
older sedimentary rocks. They cover tens of square kilo-
metres. The rocks are generally quartz porphyries, ranging
in composition from rhyolite to dacite. They all have an
extensive zone of hydrothermal alteration in the centre,
which may measure several kilometres across, and the
mineralized veins occur within this. At Chocaya and Tatasi
the major veins extend downward into the sedimentary
rocks beneath the volcanics. Pervasive hydrothermal alter-
ation at these centres post-dates all eruptive activity.
Fig. 3 Potosi: schematic cross-section through the Cerro
Rico
At other centres, such as Chorolque, Oruro and Potosi,
volcanic rocks surround and are in contact with a central
stock-like body of intrusive porphyry. The porphyries
apparently fill the vents from which the volcanics were
extruded. Erosion has partly removed the overlying vol-
canic superstructure at these centres. The stocks are
usually about 1 km in diameter at the surface and conical
in vertical section, narrowing with depth. The Cerro Rico
stock at Potosi (Fig. 3) is an extreme example. It narrows
down to two small dyke-like bodies in the deepest levels of
the mine. The stock appears to have been emplaced at
about the level of the regional erosion surface. The lower
part cuts Palaeozoic sedimentary rocks and the upper is
in contact with Miocene sediments, tuff and breccia. 3~ The
Cerro Rico stock is not crowded with xenoliths, or
brecciated, as are most of these bodies; but, like the others,
it is pervasively altered. In the upper levels of the mine the
mineralized veins are within the stock, but they dip more
steeply than the stock contact and pass out into the sedi-
ments with depth. It is postulated that there is a larger
igneous body at depth. 22,3~
At some centres, such as Llallagua and San Pablo
(Morococala), erosion has completely removed any volcanic
superstructure that may have been present, leaving only a
porphyry stock in contact with the Palaeozoic sedimentary
rocks. The Salvadora stock at Llallagua is the most impor-
tant of these (Fig. 4). It is about 1 km in diameter at the
surface and narrows with depth. It is made up of xenolithic
and highly brecciated porphyry. The original texture and
rock composition (probably quartz-latite) are obscured by
' 119
alteration, but there appear to have been several pulses of
porphyry intrusion and these were followed by explosive
hydrothermal brecciation, which is an outstanding feature
of this stock in particular, as well as some of the others. 29
Fig. 4 Llallagua: Salvadora stock and vein system,
Cancaniri level
At Llallagua almost all the mineralized veins are within
the stock, whereas at San Pablo they are in the sedimentary
rocks outside the stock margins. Both stocks, however,
contain widespread disseminated and veinlet mineralization.
The hydrothermal breccias consist typically of rounded
clasts of igneous and sedimentary rocks, up to tens of centi-
metres in diameter, set in a somewhat vughy clastic matrix
of very fine-grained rock flour cemented by quartz and
tourmaline (see Figs. 7 and 8). At Llallagua a substantial part
of the Salvadora stock consists of breccia of this type, and
in the remainder of the stock the porphyry is laced with a
network of dykes and irregular patches of breccia, which
range in width from tens of metres down to narrow dyke-
lets. These breccias are interpreted as the product of explos-
ive hydrothermal activity. In part, they may be related to
the sudden release of the residual aqueous phase of the
crystallizing magma, but it is equally likely that they were
produced by steam generated as meteoric water from the
country rocks was drawn into the high-temperature stock
by convective circulation. There is a distinct tendency for
these breccias to be concentrated in the marginal zones of
thestocks. They are of particular economic importance
because they produced an increased permeability of the
stocks exploited by later hydrothermal fluids, which give
rise to patterns of alteration and mineralization indepen-
dent of the geometry of the later vein systems.
120
Hydrothermal alteration and mineralization
Pervasive hydrothermal alteration is a ubiquitous feature
of these deposits. In all cases the alteration seems to have
begun before the formation of the major vein systems, but
neither the alteration zones nor the veins are cut by
younger igneous rocks. The intrusive stocks, such as
Llallagua, San Pablo and Potosi, are altered almost through-
out, only minor patches of relatively fresh rock having been
preserved at their contacts. At the other complexes alter-
ation is typically intense in a central zone of perhaps 1 km
or so in diameter, and decreases outward so that fresh
volcanics make up most of the larger, purely volcanic com-
plexes such as Chocaya, Tatasi and Colquechaca. Silicate
alteration assemblages include quartz-tourmaline, sericite-
tourmaline and sericite-clay. Carbonate may be present
in the latter assemblage. At some centres there is a roughly
concentric zonation of alteration in that order outward
from the centre, with an outermost aureole of chloritic or
propylitic alteration at some of the larger complexes. Dis-
seminated sulphide and cassiterite mineralization generally
accompanies the sericitic alteration assemblages, frequently
as a partial replacement of biotite and feldspar phenocrysts.
Fig. 5 Schematic cross-section through the Cerro
Chorolque showing major geological features and lateral
zonation of hydrothermal alteration around central vent
The Chorolque deposit shows a particularly well-
developed zonation of alteration and mineralization (Fig. 5).
The central 'stock', which is about 1 km in diameter at the
surface, is made up of a complex of igneous and probably
hydrothermal breccia, and some unbrecciated porphyry.
Intense alteration has converted it to a quartz--tourmaline
rock. The original texture is largely obscured and the only
recognizable primary minerals are relict quartz phenocrysts
overgrown by secondary quartz. This rock is devoid of
sulphides or cassiteri'te except where mineralization has
been introduced subsequent to the pervasive alteration.
All the important tin-bearing veins are within this core
zone. They are quartz-cassiterite veins with virtually no
sulphides, and they formed later than the quartz-
tourmaline alteration. In the volcanics surrounding the
core there is a well-defined annular zone of intense
sericite-tourmaline alteration. It is about 400 m wide and
the intensity decreases outward, sericite giving way to
chloritic alteration about 600 m from the 'stock' margin.
It is accompanied by strong veinlet and disseminated sul-
phide mineralization, which extends downward into the sedi-
mentary rocks underlying the volcanics, although the sedi-
ments do not show much silicate alteration. Fig. 6 shows
 Sn ppm
IOO0 Ag ppm

8OO

6OO

400

200

. ~ O ~ NORTH

#71/../ L 71111/-I_'l_'l.'/'L'l'l_l . . . . . ; D-:_-~-"_/-/J-/-I'/ .....

LAMAS ..... ALTERATION ~ ; - / ~ - ~ /' ................ FiiruUnuAdl ~ LAVAS ...... AL TERATIONI ''' ~ '~. " ...............
~'~tGNEOUS BRECCIA ~ OUARTZ-TOURMALINE ~'Veta Colon O $OOm ~ [ ~ IsGHNEOUS B RECC i A ~ 7EU~;c]TZ; ~pOyU~TMt)Li N E ......... 0 500m
[77] SHALE ~ SERICITE (PYRITE)

600

Bi ppm
10( 500

400

300
50
200

I0C

SOUTH
B- ;-'- : - > ' - ; +VV; -:.'.

,, ......... ~ . . . . . . . . . ,..~-~ + : +i..~+ ~ C h ~ m b o , . = o ,,s,, . . . . . . . . .

.#. '.T. ./ , ',,/ Z. . . . . . ~ YIIII/.X-/.Z'/.'/_
. ' /./ . . . . . . . /'/l~ " :": "+ .: : . +' : .K +- ~: : ' ~ ~ . ". . . . . . . . " " 9". . .' . ". . ". . .' . ". . . . . . .

LAVAS ...... ALTERATION ~ ~ ................ LAVAS ...... AL TERATION'~ " " ' * ~ . " ...............
Veta Colon 500
IGNEOUS BRECCIA ~ OUARTZ-TOURMALINE Wta Colon o SoOm P ' ~ IGNEOUS BRECCIA ~ QUARTZ-TOURMALINE 0 m
V ~ J SHALE ~ SERICITE (PYRITE) SHALE ~ SERICITE (PYRITE)

Fig. 6 Geochemical profiles across the Chorolque deposit for tin, silver, bismuth and copper. The distribution of
disseminated and veinlet mineralization with respect to central vent and alteration zones is also shown. Semiquantitative
emission spectroscopic analyses smoothed by three-point moving average

rock geochemical data for samples taken on an underground and consists predominantly of pyrite, with lesser arsenopy-
traverse across the Chorolque deposit. They illustrate the rite, chalcopyrite, bismuthinite, stannite and cassiterite.
distribution of disseminated and minor veinlet mineraliz- The metals are not zoned with respect to one another.
ation. The data do not include samples from the major At the other stock-like centres alteration is mainly of the
quartz-cassiterite veins located in the central quartz- sericite-tourmaline type. It is typically pervasive almost
tourmaline alteration zone. throughout the stock and is accompanied by dispersed
It is particularly notable that this central zone, which sulphide and cassiterite mineralization, the distribution of
contains the economic veins, contains virtually no disper- which shows no control by the detailed geometry of the
sed tin mineralization. The profiles illustrate the close major veins. These deposits have been termed 'porphyry
association between sericitic alteration and dispersed metal tin deposits', 22 and in some cases the grade of dispersed
deposition. This dispersed mineralization is sub-economic tin reaches what may prove to be an economic grade for

Fig. 7 Breccia types, Salvadora stock: (a) (left) igneous breccia - fragments of igneous (arrows) and sedimentary (dark)
rocks in crystalline igneous (quartz-latite porphyry) matrix; (b) (right) hydrothermal breccia -- fragments of igneous
(white) and sedimentary (dark) rocks in clastic matrix. Scale in cm

121
bulk mining operations. The dispersed.mineralization and
alteration may be traversed by a system of major ore-bear-
ing veins, as at Llallagua (Fig. 4) and Potosi (Fig. 3), or
else these may be absent, as at San Pablo. The mineralogy
of these ore veins comprises an early quartz-cassiterite
stage with a later complex polymetallic sulphide minera-
lization. 3~ It is from these veins that the greater part of
Bolivia's tin production has been derived.
At the high-level volcanic centres the pervasive alteration
is generally more argillic, disseminated mineralization is not
well developed and the vein mineralization is characteris-
tically sulphide-rich with minor quartz gangue.
Fig. 8 Hydrothermal breccia microtextures: (a) (top)
breccia composed of igneous rock fragments replaced by
quartz (1) with clastic quartz-rich 'rock flour' matrix
cemented by fine-grained tourmaline (black); thin-section,
plane-polarized light, field width, 3 cm; (b) contact between
quartz-latite porphyry (left) and clastic hydrothermal
dyke composed of the same rock-flour material as matrix
of breccias; thin-section, plane-polarized light, field width,
2 cm
The Tatasi deposit produces no tin - only silver, lead
and zinc; Chocaya is a major silver producer, but there are
tin-rich ores also; and Colquechaca produces lead and zinc
with only minor tin, and much silver was mined in the past.
Evolution of the hydrothermal systems
Fluid inclusion studies are being used to investigate the
evolution of the hydrothermal system at some of these
deposits. Earlier work by Kelly and Turneaure 16 was con-
cerned mainly with the tin and tungsten vein deposits of the
northern part of the tin belt, though they did provide some
122
general thermometric data on one or two of the sub-
volcanic deposits (Llallagua, Oruro). Our most important
results to date come from the study of the Chorolque de-
posit. In the core of this system the quartz of the pervasive
quartz-tourmaline alteration assemblage contains primary
coexisting vapour-filled inclusions and brine-filled inclusions
(Fig. 9). The brine-filled inclusions are so packed with solid
daughter salts that the liquid phase is usually indistinguish-
able at room temperature. On heating in the microscope
heating stage, all the daughter salts dissolve in the inclusion
fluid, which expands to fill the inclusions at temperatures
ranging from 500~ to more than 800~ (homogenization
temperatures). The presence of primary coexisting vapour-
phase and liquid-phase inclusions indicates that the fluid was
boiling during this alteration stage. Under these conditions
it is possible that both liquid and vapour may be trapped in
an inclusion at its time of formation. Such inclusions will
give anomalous homogenization temperatures (higher than
the true trapping temperatures). Thus, the minimum tem-
perature of about 500~ is the best estimate of the forma-
tion temperature of the alteration assemblage. Homogen-
ization temperatures of fluid inclusions are normally regard-
ed as minimum possible formation temperatures, as they
are subject to an upward correction to compensate for the
confining pressure that prevailed during trapping. 19 The
correction can only be estimated from the inferred depth
of formation of the mineral and an estimate of whether
lithostatic or hydrostatic load prevailed. In the case of a
boiling fluid such a correction is not required as, by defini-
tion, the vapour pressure of the fluid must have been equal
to, or greater than, the confining pressure. Thus, the
homogenization temperatures discussed here can be taken
as formation temperatures where boiling is inferred. Where
there is no evidence for boiling, they are minimum possible
formation temperatures; however, the pressure correction
would be small because of the shallow depth inferred for
these deposits.
The salinity of these fluids can be estimated by compar-
ing the temperature of solution of their daughter salts with
the known temperature/solubility behaviour of pure NaCI
solutions. 15 This indicates a salinity in excess of 60 wt %
NaCI equivalent, although this is only an approximation
since the large number of different daughter salts that are
present at room temperature shows that they are complex
multicomponent fluids rather than pure NaCI solutions.
Fluid inclusion assemblages such as this, comprising
coexisting vapour and high-salinity liquid inclusions, seem
to be characteristic of the central", high-temperature alter-
ation zones of sub-volcanic hydrothermal ore systems.
Such assemblages have been described, in particular, from
several porphyry copper deposits as well as from sub-
volcanic polymetallic deposits. 17,18~5 It is notable, how-
ever, that at Chorolque no sulphide or cassiterite deposi-
tion took place in this inner zone during this initial stage
of high salinity and temperature.
In the sericite-tourmaline zone surrounding the central
core it is not possible to study primary fluid inclusions be-
cause of the nature of the mineral assemblage. A study of
secondary inclusions (in quartz phenocrysts) believed to be
related to the alteration strongly suggests that this zone of
alteration and disseminated sulphide and cassiterite miner-
alization was formed under conditions of lower tempera-
ture and distinctly lower salinity.
The outermost (earliest) quartz in the quartz-cassiterite
veins that cut the Chorolque core zone contain primary
inclusions of lower salinity and temperature than those of
the pervasively altered wallrocks (Fig. 9). They give
 Fig. 9 Primary and pseudo-secondary fluid inclusions in
quartz from Chorolque showing progressive decrease in
salinity of hydrothermal fluid from early high-temperature
quartz-tourmaline alteration stage (A), through progressive
growth of quartz-cassiterite veins (B-E). (,4) Coexisting
primary vapour-filled inclusions (1) and extremely saline
brine inclusions packed with daughter minerals (2). Th =
500~ In secondary quartz from quartz-tourmaline
alteration zone. Largest inclusion diameter is 10/~m.
(B) Extremely saline brine inclusion (Th = 480~ showing
multiple daughter minerals. In secondary quartz from
quartz-tourmaline alteration zone. Length of inclusion is
50 pro. (1), Vapour bubble; (2), NaCI or NaKCI daughter
mineral; (3), hydrous Fe chloride daughter mineral; (4),
(5) and (6), unidentified daughter minerals. (C) Highly
saline brine inclusion (Th --- 400~ in early vein quartz
(pre-cassiterite). (1), Vapour bubble; (2), hydrous Fe
chloride daughter mineral; (3), NaCI or NaKCI daughter
mineral; (4), fluid. Inclusion length is 45/lm. (D) Moder-
ately saline brine inclusion (T h -- 320~ in pre-cassiterite
vein quartz. Maximum dimension is 20/lm. (1), Vapour
bubble; (2), NaCI daughter mineral; (3), fluid. (E) Low-
salinity (liquid-vapour) inclusion in vein quartz intergrown
with cassiterite (T h = 280~ Inclusion length is 40/~m

homogenization temperatures below 500~ and usually

nearer 400~ and salinities of about 40 wt % NaCI
equivalent. Vapour-filled inclusions are also present, which
suggests that boiling was still taking place. This indicates
that vein formation did not begin until temperatures had
dropped substantially from those of the alteration stage
and that the fluid had been partially diluted. These inclu-
sions contain numerous daughter salts at room temperature
and the composition is probably closely similar to that of
the alteration stage fluid, though more dilute. A combin-
ation of optical microscopic, electron microscopic and
X-ray spectrometric methods have been used to identify the
solid daughter salts. The results indicate that the major
heavier elements present are iron, sodium, potassium, cal-
cium, chlorine and sulphur.
By studying the fluid inclusions through the progressive
growth stages of the Chorolque veins from the walls inward,
a decline in temperature and salinity of the hydrothermal
fluid can be traced as vein growth continued. No cassiterite
deposition took place in the early stages of vein growth, al-
most all being deposited in the range of homogenization
temperatures from 300~ to 250~ from a fluid that had
been substantially diluted (less than 26 wt % NaCI equiva-
lent). Fig. 10 shows the progressive changes in temperature
and salinity of the Chorolque hydrothermal system during

123
 Discussion
60O I I Following the evidence provided by the study of fluid inclu-
sions and by the gross pattern of hydrothermal alteration
ul
and mineralization, it is suggested that the most important
w
stages in the evolution of the sub-volcanic systems were as
40 follows. Initiation occurred as the partly crystallized acid
magma consolidated in the volcanic vents. The pattern of
.~ 300 25 ~z hydrothermal activity was at first controlled by the balance

i, i i
between confining lithostatic pressure and the pressure of
"~ 200 the hydrous fluid residuum concentrating in the upper
9 part of the stock. The explosive escape of these fluids pro-
duced the early stage of pervasive hydraulic fracturing,
brecciation and alteration, of which the first generation of
disperse cassiterite mineralization was an integral part. The
Time =
hydrothermal fluid with its contained ore metals 'may have
Fig. 10 Summary diagram of fluid inclusion data for
been largely derived as a residual aqueous phase by the
Chorolque hydrothermal system illustrating very high
crystallization of the magma, but there may also have been
initial temperature and salinity of the fluid during pervasive
a significant contribution from meteoric or formation
quartz-tourmaline alteration and their major decline prior
waters drawn in from the surrounding sedimentary rocks to
to earliest cassiterite deposition
react with the cooling igneous rocks at high temperatures.
At Chorolque the fluid inclusions show that this early
alteration and vein formation. At other deposits that have fluid consisted of a low-density vapour phase and a very
been studied there is much less evidence of an earliest fluid high-density liquid phase, which coexisted at temperatures
of high temperature and extreme salinity. At Llallagua of about 500~ By use of the data of Sourirajan and
fluid inclusion temperatures both in the altered igneous Kennedy 23 for the behaviour of a N a C I - H 2 0 solution,
rocks and in the veins are much more uniform than those at the maximum pressure under which the fluid could have
Chorolque (Fig. 11). The data (uncorrected for pressure) boiled is about 360 bars. This suggests a depth of the order
suggest that the pervasive alteration and early vein growth, of 1 - 4 km, depending on whether hydrostatic or lithostatic
including cassiterite deposition, took place at temperatures
load prevailed. On geological grounds a depth of between
in the region of 400~ to 350~ falling to 300~ and lower
1 and 2 km seems most reasonable.
during deposition of the sulphides. As at Chorolque, boiling
The high-density fluid probably derived some of its con-
took place in the alteration and early vein stages, but the
stituent salts during reactions with the igneous rocks that
fluid salinity remained below 26 wt % NaCI equivalent
produced the intense quartz-tourmaline alteration, and the
throughout the period of vein growth. Kelly and
salinity would also have been increased as a result of boiling.
Turneaure 16 measured salinities ranging from 25 wt %
No ore deposition took place at this stage in the vent.
NaCI down to about 4 wt % NaCI from early vein material
Further out in the fl.anking volcanics and underlying sedi-
from Llallagua. Some more highly saline inclusions have
mentary rocks, however, there was strong sericite-tourma-
been found that suggest that there was a fluid of higher
line alteration of the volcanics and a general deposition of
salinity, at least locally, during the pervasive alteration of
metal sulphides. This produced the distinct geochemical
the stock.
dispersion patterns shown in Fig. 6. This sericitic alteration
In the volcanic complexes, such as Chocaya, Tatasi and
and metal deposition were essentially contemporaneous
Colquechaca, the sulphide-rich ores and more argillic alter-
and, therefore, it is suggested that the sericitic zone is the
ation assemblages are less suitable for inclusion studies. We
region where the high-temperature highly saline fluid mixed
have obtained limited data, however, that indicate that ore
with meteoric waters, and that it was the mixing process
deposition in these very high-level vein systems took place
that produced the physical and chemical changes leading
from fluids at temperatures below 300~ with salinities
to hydrous silicate alteration and sulphide and cassiterite
below
9 26 wt % NaCI equivalent.
deposition.
At other centres, such as Llallagua and San Pablo, the
fluid inclusions indicate that, in general, the brecciation and
60(
4--STOCK

"Pervasive
> Io :,,a II sulphide I
VEINS
Late Stage
alteration were also produced by a boiling fluid of relative-
P I
ly low salinity (less than 26 wt % NaCI) and moderate tem-
Alteration iterite Stage (Fluorite-Siderite)
Stage perature (300-400~ If an early very high-salinity fluid
50(

r
=
I I I =_=
5 did form at these centres (and there is some evidence from
40r -30~ fluid inclusion and alteration mineralogy that this was the
I.- -10 case), it was mainly confined to deeper levels in the vent.
.--
300 5 ~
z At the level of present exposure, mixing of meteoric and
N
.-- magmatic fluids had taken place throughout the permeable
200
oE
I I I igneous rocks. Thus, pervasive sericitic alteration and dis-
seminated mineralization were developed throughout -
~" 1 0 0

I I I rather than the pattern of distinct radial zonation displayed

at Chorolque.
Time > The earliest phase of pervasive alteration and widespread
Fig. 11 Summary diagram of fluid inclusion data for the disseminated mineralization produced the features used by
Llallagua hydrothermal system. Compared with Chorolque, Sillitoe and co-workers 22 to define the tin porphyry type
lower initial temperature and, in particular, salinity are of deposit. Following this, the upper levels of the miner-
accompanied by pervasive sericitic alteration and dissemin- alized volcanic structure became consolidated as magmatic
ated cassiterite mineralization through much of stock activity and hydrothermal fluid generation at that level

124
 declined. A new pattern of fluid movement and ore deposi-
tion was then superimposed by the establishment of dis-
crete vein systems. These veins clearly cut across the earlier
pattern of pervasive alteration and hydrothermal breccia-
tion and do not display radial or concentric distribution
with regard to the vents themselves. The formation of the
vein systems can be ascribed to the reimposition of regional
tectonic stresses following the decline in localized
magmatic/hydrothermal pressures after the hydrothermal
'exhaustion' of the upper levels of the volcanic infrastruc-
ture. The veins evidently tapped a deeper, but genetically
related, magmatic source as hydrothermal fluid pressures
built up and dilatation along defined systems of tectoni-
cally controlled fractures took place. At Potosi, for
example, the vein system continues downward below the
level at which the stock has narrowed down until it forms
only a small dyke (Fig. 3). The activity of the deeper
source regions also declined rapidly - there is rarely much
evidence of resurgent high-temperature activity in the
veins, or renewed hydrothermal brecciation. The vein
fluids became progressively cooled and diluted, probably
because of the influx of meteoric water into the system,
and it appears that convective circulation was not suffici-
ently sustained by the 'heat-engine' effect of the cooling
intrusive to produce large-scale overprinting of the earlier-
formed alteration assemblages by late, lower-temperature
assemblages.
The length of time involved in the development of
these deposits is not known. Geochronological work
currently in progress 11 shows that a time-span of the
order of 1-2 m.y. separated the beginning of igneous
activity from the pervasive alteration that marked the
onset of major hydrothermal activity at several of the
eruptive complexes. At Potosi there is K - A r evidence
that the altered and mineralized Cerro Rico stock was
being eroded a little over 1 m.y. after the time of alter-
ation. The time-span of hydrothermal activity cannot,
however, be precisely defined.
Conclusions
The model of hydrothermal evolution outlined here illus-
trates the wide range of physical and chemical conditions
and the complex interplay of geological processes that are
involved in the formation of ore deposits of this type. In
particular, it can be seen that the shallow depth of em-
placement is an important factor in permitting the rapid
fluctuations in pressure and temperature with time, and
in allowing the large-scale influx of meteoric water into
initially magmatic-hydrothermal systems. These factors
can result in the overlap or telescoping of mineral assem-
blages (which is characteristic of this deposit type): thus,
for instance, cassiterite and silver-bearing sulphosalts can
be accommodated in the same veins at the same level.
Historically, the Bolivian sub-volcanic deposits were regard-
ed as type examples of the 'xenothermal' class, because of
the emplacement of mineral parageneses of inferred high
temperature of formation at geologically shallow depths. 8
Recent studies of magmatic-hydrothermal ores have shown
that the range of conditions that governs the formation of
particular mineral parageneses is too broad for such narrow,
essentially temperature-based, classifications to be useful.
Routhier 20 has, indeed, suggested that the term 'xenother-
mal' be abandoned. The description 'sub-volcanic', how-
ever, successfully embraces the full range of geological
processes contributin9 to the formation of the deposits of
the Bolivian type and accounts for the complex thermal
and mineralogical patterns observed. This study suggests,
for example, that the chemical conditions of the fluid,
especially changes that accompany reaction between the
fluid and the wallrock and the mixing of magmatic with
meteoric water, are more important factors than tempera-
ture in controlling the transport and deposition of
cassiterite.
Acknowledgment
We are indebted to the Natural Environment Research
Council, London, for financial support; to GEOBOL for
providing transport and additional field support; and to
the Management of the Corporaci6n Minera de Bolivia
(COMIBOL) for permission to visit the mines and for
providing accommodation and the assistance of the mine
geological staff. We acknowledge, in particular, the assis-
tance of Ing. Jos~ Torrez, Director of GEOBOL, and the
COM IBOL geologists Salom6n Rivas (Llallagua), Humberto
Mallo (Potosi), and Rene Gonz~lez (Chorolque). The
diagrams and photographs were prepared at the Royal
School of Mines with the assistance of Ella Hin, Helen
Richardson, Mike James and Ken Jones. Technical services
in the preparation of thin and polished sections for fluid
inclusion study were provided in the Division of Mining
Geology, Royal School of Mines, with the assistance of
Barry Foster, Elizabeth Morris and Niel Wilkinson. The
manuscript was critically reviewed by Dr. A. Rankin and
Dr. R. H. Sillitoe.
References
1. Ahlfeld F. Metallogenic epochs and provinces of
Bolivia. Mineral. Deposita, 2, 1967, 291-311.
2. Ahlfeld F. E. Geologla de Bolivia (La Paz: Los Amigos
del Libro, 1972), 190 p.
3. Ahlfeld F. E. and Schneider-Scherbina A. Los
yacimientos minerales y de hidrocarboros de Bolivia. Boln
Dep. Nac. Geol. no. 5 (Especial), 1964, 388 p.
4. Audebaud E. Laubacher G. and Marocco R. Coupe
g~ologique des Andes du Sud du P~rou de I'Oc~an Pacifique
au Bouclier Br~silien. Geol. Rdsch., 65, 1976, 223-64.
5. Avila W. Un modelo de tect6nica de placas para el
origen del Cintur6n Esta~[fero Boliviano. Boln Soc. Geol.
Boliviana no. 22, 1975, 3-23.
6. Best M. G. Migration of hydrous fluids in the upper
mantle and potassium variation in calc-alkalic rocks.
Geology, 3, 1975, 429-32.
7. Brown G. C. and Fyfe W. S. The transition from
metamorphism to melting: status of the granulite and
eclogite facies. In 24th Int. geol. Congr. (Montreal: The
Congress, 1972), Sect. 2, 27-34.
8. Buddington A. F. High-temperature mineral associa-
tions at shallow to moderate depths. Econ. Geol., 30,
1935, 205-22.
9. Clark A. M. and Farrar E. The Bolivian tin province:
notes on the available geochronological data. Econ. Geol.,
68, 1973, 102-6.
10. Cobbing E. J. and Pitcher W. S. Plate tectonics and
the Peruvian Andes. Nature, phys. Sci., 240, 1972, 51-3.
11. Grant J. N. et al. Potassium-argon ages of igneous
rocks and mineralization from part of the Bolivian tin belt.
In preparation.
12. GriggsD. T. The sinking lithosphere and the focal
mechanism of deep earthquakes. In The nature o f the solid
earth Robertson E. C. ed. (New York: McGraw-Hill, 1972),
361-84.
125
13. Harris P. G. Zone refining and the origin of potassic
basalts. Geochim. cosmochim. Acta, 12, 1957, 195-208.
14. JamesD. E. A plate tectonic model for the evolution
of the Central Andes. Bull. geol. Soc. Am., 82, 1971,
3325-46.
15. Keevil N. B. Vapor pressures of aqueous solutions at
high temperatures. J. Am. chem. Soc., 64, 1942, 841-50.
16. Kelly,W. C. and Turneaure F.S. Mineralogy, para-
genesis and geothermometry of the tin and tungsten deposits
of the Eastern Andes, Bolivia. Econ. Geol., 65, 1970,
609-80.
17. Nash J. T. and Theodore T. G. Ore fluids in the
porphyry copper deposit at Copper Canyon, Nevada.
Econ. GeoL, 66, 1971,385-99.
18. Roedder E. Fluid inclusion studies on the porphyry-
type ore deposits at Bingham, Utah, Butte, Montana, and
Climax, Colorado. Econ. GeoL, 66, 1971,98-120.
19. Roedder E. Fluid inclusions as samples of ore fluids.
In Geochemistry of hydrothermal ore deposits Barnes H. L.
ed. (New York: Holt, Rinehart and Winston, 1967), 515-74.
20. Routhier P. Les gisements m~tallif~res: g~ologie et
principes de recherche (Paris: Masson, 1963), 2 vols, 1282 p.
21. Sillitoe R. H. Relation of metal provinces in western
America to subduction of oceanic lithosphere. Bull. geoL
Soc. Am., 83, 1972, 813-7.
22. Sillitoe R. H. Halls C. and Grant J. N. Porphyry tin
deposits in Bolivia. Econ. GeoL, 70, 1975, 913-27.
23. Sourirajan S. and Kennedy G. C. The system H20 -
NaCI at elevated temperatures and pressures. Am. J. Sci.,
260, 1962, 115-41.
24. Stewart J. W. Evernden J. F. and Snelling N. J. Age
determinations from Andean Peru, a reconnaissance survey.
Bull. geoL Soc. Am., 85, 1974, 1107-16.
25. Takenouchi S. and Imai H. Glass and fluid inclusions
in acid igneous rocks from some mining areas in Japan.
Econ. Geol., 70, 1975, 750-69.
26. Toselli A. J. Caracter(sticas del plutonismo
famatiniano en el noroeste Argentina. In Anales de la IV
convenci6n nacional de geologfa ( La Paz: Sociedad
Geol6gica Boliviana, 1975) vol. 2, 369-76.
27. Toks6z M. N. Minear J. W. and Julian B. R. Temper-
ature field and geophysical effects of a downgoing slab.
J. geophys. Res., 76, 1971, 1113-38.
28. Turcotte D. L. and Schubert A. Frictional heating
of the descending lithosphere. J. geophys. Res., 78, 1973,
5876-86.
29. Turneaure F. S. The tin deposits of Llallagua,
Bolivia. Econ. GeoL, 30, 1935, 14-60; 170-90.
30. Turneaure F.S. A comparative study of major ore
deposits of Central Bolivia. Econ. Geol., 55, 1960, 217-54;
574 -606.
31. Turneaure F. S. The Bolivian tin-silver province.
Econ. GeoL, 66, 1971,215-25.
32. Turner J. C. M. The Andes of northwestern
Argentina. GeoL Rdsch., 59, 1970, 1028-63.
33. Wyllie P.J. Experimental petrology and global
tectonics- a preview. Tectonophysics, 17, 1973, 189-209.

126
Occurrence, origin and significance
of mechanically transported
sulphide ores at Buchans,
Newfoundland

J. G. Thurlow
Memorial University of Newfoundland, St. John's,
Newfoundland, Canada

553.068.27:553.435(718)

Summary
The Buchans orebodies are high-grade Kuroko-like massive
sulphide deposits associated with Silurian (?) felsic
volcanism of the northern Appalachians. Since 1928,
17 000 000 tons of ore has been mined, of which approxi-
mately one-half consists of mechanically transported
sulphide breccia accumulations. Most of the transported
ores occur as discrete orebodies composed of angular
(competent) and wispy (plastic) sulphide fragments
intimately mixed with a large variety of angular to rounded
lithic fragments. Abundant evidence exists to suggest that
these ores were fragmented by surface volcanic explosions
and transported as rapidly moving sub-aqueous density
flows. Fragmentation of the in-situ orebodies occurred
intermittently throughout massive sulphide deposition and
thus the zonation of the in-situ deposits is mimicked by
zonation of the transported ores. These ores were pre-
ferentially deposited in palaeotopographic depressions along
the path of density flow and maintain economic grade for
distances in excess of 2 km from their source. Other
deposits appear to have moved as slowly advancing 'mud
glaciers'.
The direction of transport can be determined from
imbrication of ore fragments, direction of elongation of the
ore zones and clast fabric, and from increasing ore grades,
clast size and individual bed thickness in more proximal
areas. Limited data on the distribution of disorganized,
inverse to normally graded and normally graded beds
yield transport directions compatible with those obtained
from the other parameters. On the basis of the above
criteria, as well as stratigraphic considerations and com-
parison of metal ratios in transported and in-situ deposits,
the source of most transported ores can be identified as
one of the two major in-situ massive sulphide deposits.
At least one far-travelled sub-economic prospect has,
however, no known source orebody.
In terms of exploration the Buchans orebodies are
significant in that economic sulphide accumulations may
occur as discretetransported orebodies well beyond the
limits of stockwork mineralization and significant alteration.

127

Geological setting of the Skorovas
orebody within the allochthonous
volcanic stratigraphy of the
Gjersvik Nappe, central Norway*
C. Halls
Division of Mining Geology, Royal School of
Mines, London, England
A. Reinsbakken
Geologisk I nstitutt, Norges "rekniske
H~gskole, Trondheim, Norway
I. Ferriday
Division of Mining Geology, Royal School of
Mines, London, England
A. Haugen
Grong Gruber A/S, Joma, Norway
A. Rankin
Division of Mining Geology, Royal School of
Mines, London, England
553.067: 553.277:553.661.21 (484.3)
Synopsis
The Skorovas orebody is one of the chief stratiform base-
metal deposits within the allochthonous greenstone belt
o f the Central Norwegian Caledonides. It is contained in
the volcanic level of a complex eruptive association of Lower
to Middle Ordovician age defined as the Gjersvik Nappe. The
rocks of this nappe are contained as a depressed segment
of the larger K~li Nappe and defined to the north and
south, respectively, by the B~rgefjeli and Grong-Oiden
basement culminations. The principal components of this
nappe are a plutonic infrastructure o f composite gabbroic
intrusions within which has been emplaced a series o f
dioritic to granodioritic (trondhjem itic) bodies that form
the roots of a consanguineous submarine polygenic volcanic
sequence. The eruptive rocks are overlain unconformably
by a sequence of polymict conglomerates and calcareous
flysch sediments, the composition of which suggests
immediate derivation by erosion from the underlying
igneous complex.
Pre-tectonic segregations, veins and vesicle fillings of
epidote, albite, chlorite, carbonate and quartz related to
primary volcanic flow structures in the lava pile provide
evidence of pervasive inositu sea-floor metamorphism, and
this interpretation is verified by the abundance o f nearly
monomineralic epidote clasts in the derived conglomerates.
The relationship of the eruptive and sedimentary suites
is interpreted in terms of the evolution o f an ensimatic
island arc, o f Lower to Middle Ordovician age, which under-
went uplift and erosion prior to emplacement on the Fenno-
scandian basement during the climactic stages of collision
tectonism o f the Caledonian Orogeny in Silurian times.
The entire igneous and sedimentary assemblage has been
affected by the tectonic stages of allochthonous emplace-
ment, but the gross differences in competence between the
component lithologies has resulted in a particularly hetero-
*UNESCO-lUGS International Geological Correlation Programme,
project no. 60: Correlation of Caledonian stratabound sulphides.
Norwegian--British contribution no. 1.
128
geneous style o f deformation in which folding, componen-
tal sliding, fracturing and penetrative metamorphic refabric-
ation have been governed largely by the geometry of the
most competent lithologies, notably gabbro, diorite and
granodiorite (trondhjemite) intrusives and, within the
extrusive sequence, compact dacitic flows and their spili-
tized aphanitic equivalents (keratophyres). The hetero-
geneous pattern o f deformation is resolved in terms of two
main stages o f folding complicated by componental sliding
movements.
Mineralization occurs at two levels in the eruptive
sequence. The layered gabbros and lensoid metagabbros of
the plutonic infrastructure contain small cumulus bodies of
nickel-, copper- and platinum-bearing p y r r h o t i t e - p y r i t e -
magnetite ore of magmatic derivation. Mineralization of
this type is at present only known in sub-economic quan-
tities.
The Skorovas orebody, in common with other widely
dispersed volcanic exhalites in the Gjersvik Nappe, occurs
within the volcanic sequence at a level marked by episodes
o f explosive dacitic volcanism and associated fumarolic
activity. The Skorovas orebody consists o f approximately
I0 000 000 tons of massive and disseminated predominantly
pyritic ore with an approximate average grade o f 1.3% Zn
and 1.0% Cu, together with trace amounts o f Pb, As and
Ag. The complex lensoid geometry of the orebody is re-
solved in terms of the disjunction of a single stratiform unit
by tight isoclinal folding and componental movements,
probably involving both translation and rotation.
Enrichment o f sphalerite, chalcopyrite and, locally,
galena within the magnetite-pyrite ores at the stratigraphic
top and margins o f the ore lenses is interpreted as a primary
feature. The banded magnetite-pyrite ores are commonly
associated with magnetitic cherts or jaspers and are thus
transitional in aspect to the thin, iron- and silica-rich, base-
metal-depleted, exhalative sedimentary horizons that occur
extensively within the extrusive sequence of the Gjersvik
Nappe. Theseare interpreted as the products o f settling o f
colloidal iron and silica hydrosols following explosive dis-
persal into an oxidizing submarine environment. They are
valuable time-stratigraphic markers and indicators of way-up
in complicated structures and are a potentially valuable tool
in exploration for massive sulphide bodies formed in limited
reducing environments,
The belt of metamorphosed Lower Palaeozoic rocks, chiefly
of Ordovician age, within which the important stratiform
pyritic copper- and zinc-bearing orebodies of the Scandina-
vian Caledonides are located extends over 1500 km from
Rogaland in southwestern Norway to Nord Troms. The
divisions of this complex metallogenic belt have been described
by Vokes 73 and Vokes and Gale,75 and Fig. 1 shows the
relationship of the principal districts to the thrust front of
the Caledonian allochthon. The culminations of the under-
lying Precambrian basement; together with the effects of
erosion, have produced the segmentation of the allochthon
on which the division into separate districts is broadly based.
Structural and stratigraphic correlations along the length of
the belt are made difficult by the structural complexity of
the allochthon, the sparsity of fossil remains and the pene-
trative effects of tectonic deformation and regional meta-
morphism. Sufficiently detailed studies have been made,
however, in the regions of South Trbndelag (Trondheim
district), 49,s~ North Tr~ndelag (Grong-Gjersvik
district) 4~ and the geographically adjacent areas of
Jamtland and Va'sterbotten in Sweden e1~2~3 to show that
the stratiform ores of Skorovas, Joma, Stekenjokk, L~kken
and R6ros lie within the K61i Nappe, which is the upper
structural unit of the Seve-K6li Nappe complex first de-
fined by TSrnebohm. 6s The broad correlation within the
K61i structural level can reasonably be carried into the
Sulitjelma district of Nordland, 39,8~ and in all probability
this correlation can be extended into the ore district of
Nord Troms.
It is clear that the separate districts that comprise the
Ordovician province of stratiform pyritic ores lie at a broadly
comparable structural level in the Caledonian allochthon of
the Scandinavian peninsula, but there are significant differ-
ences in the stratigraphy and metamorphic grade of the host
Fig. 1 Synoptic geological map of Scandinavian
Caledonides showing main districts of stratiform volcano-
genic ores at KSli structural level
rocks from district to district. In general, the Ordovician
host rocks comprise a varied assemblage of supracrustal
volcanic and sedimentary rocks with closely associated
plutonic masses of ultrabasic, basic and acid composition.
The conspicuous quantity of basaltic to andesitic volcanics
in the supracrustal sequences, taken together with their
deformed and metamorphosed condition, ranging in grade
from lower greenschist to almandine amphibolite facies,
has led to the familiar use of the terms greenschist and
greenstone in descriptions of the stratigraphy of various
districts. 61 Goldschmidt 22 early lent authority to this
usage by defining the 'Stamm der grQnen Laven und
I ntrusivgesteine' as an important constituent rock kindred
of the south and central parts of the Caledonian allochthon
at the structural level now under discussion.
It is generally recognized that the stratiform pyritic ore-
bodies have a close genetic relationship to the volcanic rocks
with which they are associated 73 and that this relationship
originated with the formation of tholeiitic and calc-alkaline
eruptives at the margins of the Caledonian orogen in
Ordovician times. 15,16,47,75 The genetic process that re-
lates the ores and host rocks has been masked by the effects
of metamorphic recrystallization and polyphase deform-
ation, which affected both ores 73,74 and host rocks during
the process of allochthonous tectonic emplacement conse-
quent upon collision of the Scandinavian and Laurentian
cratons during Middle Silurian times. 1~ The palaeo-
environmental interpretation of the rock assemblages con-
tained in the structural elements of the KSli nappe is
clearly of the greatest importance in interpreting the genesis
of the associated ores; in a region of the tectonic complexity
displayed by the Caledonian allochthon, however, it is clear
that the primary geological framework must be established
by a study of the field relationships at a level of regional
detail such that the ore deposits can be considered at the
scale of the geological phenomenon responsible for their
formation. If a volcanogenic origin is postulated, an under-
standing of the volcano-stratigraphy and structure in an
area that extends from 1 to 10 km outside the orebody itself
must be sought. This has been the basis on which the
present study of the environment of the Skorovas deposit
was undertaken.
Regional structural and stratigraphic setting
Existing knowledge of the major structural and stratigraphic
units of the Grongfelt originated with the regional geological
mapping undertaken by Statsgeolog Steinar Foslie 12,14 during
the period 1922-27, the details of which were amplified
and interpreted by T. Strand 14 and C. Oftedahl. More recent
regional studies by Zachrisson 81 in the adjacent Swedish
area of J~mtland and V~sterbotten have given an idea of
the succession of structural units within the K~li Nappe
sequence between the Grong and Stekenjokk areas. A com-
pilation from these sources is made in Fig. 2, which shows
the main second-order tectonic divisions that have been
recognized within the K6li level of the Seve-K61i nappe.
Combining the terminologies of Foslie, 12 Oftedahl 41 and
Zachrisson, 81 there are four divisions to be recogr~ized. The
first and uppermost of these is the Gjersvik Nappe, within
which lie the Skorovas (Sk) and Gjersvik (Gj)orebodies.
Below this lies the Leipik Nappe, within which, by extending
the structural interpretation of Zachrisson, the Joma
orebody (Jo) must lie. Below this lies the Gelvernokko
Nappe and, finally, the Lower K~li Nappe unit, within which
are situated the Stekenjokk orebodies (St) (the Stekenjokk
maim and the Levimalm). 82
129
 O-j~m
....
4-J

~2M:~
Q.

~z

E~g

_~ ~..~_
m ~
u~ E

E~-~
~o~

" ~ E~
.~~o ~-
..,~ "~
' - 8l -

o~,.,
~z
v I~ .-- e-
~ E ..~ o
~ v >

9- o ..o +~ .o

~9 ~.~
elzg
~.-~

-~a~
4" "O

"~Z~
~,~ ~
~=~-~
:O O

"~ e~''~"
- .~
E
O
,~_x~

e-

~'~

130
 The broad classification into the second-order tectonic
units shown in Fig. 2 provides a useful basis for descriptions
of the regional geology, but the exact status of the second-
order thrust boundaries is difficult to establish because
these are taken, for the most part, to follow stratigraphic
boundaries. 41~1 For the purpose of the present discussion,
however, the precise location of the second-order structures
and their relative tectonic status is less important than the
plutonic and stratigraphic relationships preserved within the
Gjersvik Nappe itself. In Fig. 2 the upper tectonic contact
with the Helgeland Nappe 23 is clearly defined. The plutonic
and supracrustal stratigraphy is revealed in the passage from
southwest to northwest across the area of the map covering
the Gjersvik Nappe. Without precise knowledge of the
relative ages and finer lithological divisions of the various
units the following sequence is conspicuous. Large masses
of gabbro and granodiorite (trondhjemite) in the southwest
are succeeded spatially to the northeast by the Gjersvik vol-
canic greenstone sequence with the contained orebodies at
Skorovas and Gjersvik. A period of relative quiescence is
indicated by the presence of a marble bed intermittently
preserved at the uppermost level of the volcanic greenstone
sequence. The marble is best preserved in the terrain north
of the Limingen Lake, but a limited thickness is found to
the NNE of Skorovas mine in the terrain to the south of
Tunnsj~en. The volcanics with the overlying marble are
followed by a spectacular polymict conglomerate, the typi-
cal aspect of which is shown in Fig. 12. The final part of
the sequence is made up by the clastic sediments of the
Limingen group, composed by a variety of schistose conglo-
mer;atic, sub-arkosic and phyllitic rocks, the majority of
which are distinctly calcareous.
Oftedahl, 41 in his discussion of the nappe units of the
Grongfelt, defined a thrust boundary of intermediate
significance that separates the polymict conglomerate and
the Limingen sequence of calcareous and conglomeratic
metasediments, so that the Gjersvik Nappe, in its original
definition, does not include the Limingen Group. It seems
reasonable, however, to extend the compass of the Gjersvik
Nappe to include the sediments of this group, which seem
to be laterally related, in part, to the basal polymict conglo-
merate and to have derived most of their clastic components
from the Gjersvik plutonics, greenstones and overlying
limestones.
AGE W TRONDHEIM REGION
L SILURIAN HORG GROUP Hovin sst
( Llandoverian } Lyngestein quartzite cong (Horg Disturbance)
Ashgithan UPPER HOVIN Slate. greywacke
I GROUP Limestone
Caradocian Volla polymict cong. ( Ekne Disturbance )
Shale
Caradocian Rhyolitic tuff
Sandst , shale, l i m e s t
LOWER HOVIN
o GROUP Stokkvola polymict cong (Stokkvola Disturbance ) - -
Jonsvatn greenstone , H~londa andesite
Limest
Slate. s a n d s t
Arenlg Venna potymict cong (Trondheim Disturbance}
Arenig STSREN Tholeiitic eruptive greenstone sequence of
I GROUP basalts with gabbro intrusions and minor
Tremadocian
I trondhjemite and quartz keratophyre.
CAMBRIAN
0RDOVICIAN
I GULA GROUP Schists and quartzites with
CAMBRIAN trondhjemite intrusions
Fig.3 Inferred stratigraphic correlation between Lower Palaeozoic sequences to south and north of Grong Culmination.
Correlation is approximate and based on information from Vogt, 72 Zachrisson, 82 Oftedahl 45 and Roberts. 53 Tectonic
disjunction within the two areas is shown schematically by oblique parallel lines
The rocks of the Gjersvik Nappe have, so far, yielded no
fossil remains to give a basis for precise dating and correla-
tion with stratigraphies in adjacent segments of the Seve-
K61i Nappe. The volcanic and plutonic units of the
Gjersvik eruptive complex do, however, bear certain simi-
larities to the rocks of the St6ren Group 72 in the
Trondheim region. The St~ren Group, locally, overlies
schists of the Gula Group containing Dictyonema flabelli-
forme. 62 The contact between the two groups is, however,
markedly tectonic 16 and, thus, the graptolite fossil evidence
can only be used to suggest a possible maximum age of
Upper Cambrian-Lower Ordovician (Tremadocian) for the
St~ren Group, and it is conceivable that the tholeiitic
eruptive activity recorded in the St~ren sequence 16 could
have been initiated yet earlier in Cambrian time.
It has generally been proposed that the Gjersvik Group
is of equivalent age to the St6ren Group 45 and, by implic-
ation, that the two groups represent similar stages in the
morphological and magmatic evolution of the Caledonian
orogenic margin in central Scandinavia. Stratigraphic and
geochemical evidence suggests, however, that the eruptive
sequence of the Gjersvik Nappe is more evolved in terms of
calc-alkaline character 16,47 a matter that is given further
_
consideration in a later section of this paper. Gale and
Roberts have therefore suggested that the Gjersvik erup-
tives are of younger age than those of the St~ren Group, 16
and a partial correlation, at least, with the andesitic green-
stones of the Lower Hovin Group (Forbordfjell, H610nda
and equivalent greenstones)53,72 seems reasonable. The age
of the youngest Gjersvik eruptives therefore probably lies
within the Arenig-Caradocian range, whereas the graptoli-
tic fauna of the Bogo shale within the Lower Hovin Group,
which overlies the St~ren Group in the Trondheim region, is
interpreted as belonging within the Didyrnograptus
hirundo zone. 57 The St~ren Group thus has a defined
minimum age in the range Arenig to early Llanvirnian.
A further aspect of the stratigraphic correlation between
the Lower and Middle Ordovician sequences in the Trond-
heim and Grong districts concerns the tectonic and strati-
graphic status of various polymict conglomerate horizons
that occur at intervals within the Lower and Upper Hovin
Groups and, notably, that which overlies the Gjersvik
eruptive sequence.
The widespread occurrence of conglomerates (Venna,
GRONG - STEKENJOKK REGION
Basic - acid (quartz keratophyre} voIcanics
LASTERFJALL GROUP and tuffites ; calc p h y l l l t e s , l : m e s t
~ . ~ ~ Vojtla quartzite cong
RSYRVlK GROUP Joma greenstone
s e r p e n t i n e , caic phyl[ites , cong[
LININGEN Calcareous phyl[ite , subarkose ,greywacke
GROUP
- - Havdalsvatn polymict cong (Gjersvik Disturbance)
Locally preserved l i m e s t (marble) I
GJERSVIK TholeiJtic - calc atka[ine eruptive greenstone sequence I
GROUP basatts ,andesites, k e r a t o p h y r e / d a c i t e with major
intrusions of gabbro diorite and trondhjemite
?
Nica schists etc (Guia Group - ? Seve equivalent )
131
Lille Fundsj~ and Steinkjer conglomerates) 53 at the base
of the Lower Hovin Group, overlying the St~ren Group,
led Holtedahl 26 to propose a tectonic event of regional
significance that he termed the Trondheim Disturbance.
Further comparative studies of stratigraphy in the Trond-
heim region led to the recognition of similar polymict
conglomerates at higher stratigraphic levels. Vogt 72
identified an Ekne (Caradocian) Disturbance and also move-
ments in the Lower Silurian which produced the basal quart-
zite conglomerate of the Horg Group (Lyngestein Conglo-
merate), which identified a Horg Disturbance. Further work
by Roberts 53 has suggested additional refinements to the
chronology of uplift and erosion in the Trondheim District
during the mid-Ordovician, a separate event in Mid-Lower Hovin
times being marked at the level of the Stokkvola conglo-
merate. 53 Tectonic evolution in the Trondheim region in
Lower to Middle Ordovician time was evidently punctuated
by episodes of vertical uplift and erosion, the Trondheim
Disturbance being but the first of these. The polymict
conglomerate, which overlies the Gjersvik eruptives at the
base of the Limingen sedimentary series, evidently records a
disturbance of the Trondheim type, which, to avoid confu-
sion, will be named the Gjersvik Disturbance. This distur-
bance is probably most closely related in age to the
Stokkvola event. 53
Fig. 3 shows the inferred general stratigraphic correlation
between the Lower Palaeozoic sequences in the Grong and
Trondheim regions. Zachrisson82 has cited the faunal
evidence in support of a (Lower ?) Silurian age for the
Stekenjokk orebodies, which lie within the lower part of
the sequence of basic to acid volcanic rocks composing the
upper part of the Lasterfjall Group (Fig. 2); this means that
the rocks composing the Gjersvik, Leipik and Gelvernokko
nappes and the upper parts of the Lower K61i Nappe have a
probable age range from Lower Ordovician to Lower
Silurian, matching the age range of the Trondheim Super-
group as defined by Gale and Roberts. 16 The Skorovas
and Gjersvik ore deposits lie within the Gjersvik Group of
volcanic greenstones and must be approximately Lower to
Middle Ordovician in age. It is, however, interesting that in
the Stekenjokk area, accepting the fossil evidence of
Zachrisson, conditions suitable for the formation of strati-
form pyritic ores also existed in Lower-Middle Silurian
times.
Tectonic style within Skorovas area of Gjersvik Nappe
The programme of field mapping in the Skorovas area, with
which the present writers have been actively involved since
1971, was designed to re-examine the major structural and
lithological boundaries within the plutonic to volcanic se-
quence of the Gjersvik Group and to extend, as far as
possible, the geological interpretations of Foslie and
Oftedahl as they affect the Skorovas area. Mapping in the
scale range of 1:2000 to 1:10 000 has also enabled the
first serious attempt to delineate the principal lithologies
within the volcanic sequence, which were uniformly desig-
nated as greenstones by Foslie 12 off the 1:100 000 scale
map of the Trones quadrangle. The Skorovas area, as shown
in Fig. 4, lies close to the eastern boundary of one of the
main plutonic massifs of the Gjersvik Nappe. From Fig. 2
it is clear that the massifs have distinctly tectonic boundaries
of low to intermediate angle (Fig. 6). The plutonic rocks
within these boundaries frequently preserve their original
igneous fabrics, little modified by the penetrative effects of
tectonic deformation. The volcanic rocks and minor intru-
sives outside them, in contrast, generally show intense
penetrative tectonic fabrics. The plutonic massifs all have
132
tectonized envelopes and the intrusion of the complete
range of basic to acid plutonic rocks evidently took place
prior to the main tectonic event, which led to the emplace-
ment of the Gjersvik Nappe within the allochthon and
which was also responsible for the generation of major
isoclinal folds and the early axial plane schistosity that is
generally well developed within rocks of the volcanic
sequence.
Because of gross differences in competence between the
various rock types, notably between the plutonic masses
and the supracrustal volcanic cover, this particularly hetero-
geneous style of deformation characterizes the intermediate
level of the Gjersvik Nappe, the pattern being controlled,
on the largest scale, by the form of the major gabbro, diorite
and granodiorite bodies. Within the volcanic sequence it-
self, high-level doleritic dykes and sills, together with
compact dacitic flows and their spilitized aphanitic
equivalents, exert a more local influence.
In common with adjacent parts of the allochthon, s132
the history of regional deformation can be resolved in terms
of two major stages, the first of which produced the prin-
cipal Caledonian 'grain' of the terrain, creating isoclinal folds
of the style illustrated in Fig. 5, and imposing the early
schistosity mentioned above. It was during this stage that
the main thrust and slide horizons that separate the plutonic
and volcanic levels of the Gjersvik eruptive sequence were
established. The plutonic bodies evidently behaved as
massive tectonic wedges, piercing and, in part, overriding
the superjacent volcanics to create the present pattern.
It should be emphasized that such planes of high tectonic
strain also exist in several lesser orders within the volcanic
sequence. These surfaces, as was noted above, are similarly
formed at lithologicaiboundaries, showing marked con-
trasts in competency, and can partly be explained in terms
of componental movements along the thinned and extended
limbs of isoclinal folds of the early basaltic lavas and pillow
breccias. These rocl~s, under the influence of intense local
strain, suffer a complete penetrative reorganization of
their mineralogy to form chlorite-albite-epidote schists
devoid of any earlier volcanic fabric. In the field the exis-
tence of these surfaces and the flattening produced in the
adjacent units creates a peculiarly lenticulated style of
deformation through which the early isoclinal fold pattern
must be traced. The 'lenticulate style' appears to be a
characteristic feature of highly deformed volcanostratigraphy
and associated plutonics in other regions, notably in the
Mauretanides of WestAfrica (G. Pouit, personal communi-
cation). Minor fold structures of the early generation are not
conspicuously evident within the volcanostratigraphy and
are best observed in the finely stratified tuff bands and
associated cherts and iron-rich chlorite schists of the
exhalit~ facies (Fig, 7(a)). They can also be mapped over
several tens of metres by following coherent chert horizons,
acid tuff bands and dykes, and thence into the larger iso-
clines of the type illustrated in Fig. 5.
The configuration of these larger isoclines, taken to-
gether with the stratigraphic and structural evidence provid-
ed by the mapping of the surface of unconformity separat-
ing the eruptive sequence and the conglomerate series, de-
monstrates, at the present level of erosion, that the volcanic
sequence in the Skorovas district lies inverted within the
lower limb of a major southeast-facing fold, the identity of
which can be broadly equated with the Gjersvik Nappe.
The second stage of deformation, superimposed on the
grain of the early isoclines and schistosity, has created an
open system of broad folds, which have resulted in an
+ irregular pattern of dome and basin structures, the major
axes of which evidently bear a relationship to the contacts
of the plutonic massifs lying to the west and north (Figs. 2
and 4(a)). The formation of the open dome and basin
structures is accompanied by further movements along the
low-angle planes generated during the first stage of deform-
ation. These movements led to the creation of minor folds
and a second-stage crenulation cleavage, which is typically
local and specifically associated with these horizons of
high strain. The scale of the phenomenon is variable and
Fig. 7(b)) shows part of the well-developed belt of second-
stage folding in the volcanic sequence at the southwestern
margin of the Gr~ndalsfjell massif. The vergence of the
axial planes of these and other similar late folds implies that
the principal tectonic stress responsible for this deformation
was imposed from a west to northwest direction.
The deformation history can be interpreted in the follow-
ing way. (1) Creation of the nappe, isoclinal folds and the
Fig. 4 Simplified geological map (a) of Skorovas area with line of section (Fia. 5) indicated (SSV. Store Skorovatn:
GR. Grubefjellet; ONV, (Z)verste Nes~va~net; TV, Tredjevatnet; BI, Bl~hammeren; HV, Havdalsvatnet)and synoptic map
(b) (see page 134) of principal structural trends
early schistose fabric, together with the several orders of
internal thrust horizons, was a consequence of the stresses
imposed during the main stage of emplacement of the
allochthon during Mid-Silurian times. (2) The second gener-
ation of tectonic structures is considered to have been im-
posed upon the first as a consequence of equilibration be-
tween the depressed Scandinavian basement and the imposed
load of the allochthon. The depression of the granitic base-
ment into a field of higher temperature and pressure can
have given rise to plasticity of the basement, enabling local
isostatic adjustments to take place by the initiation of a
system of domes and basins in the basement. The second
fold phase in the Skorovas region is interpreted as a conse-
quence of forces imposed on the volcanic sequence by the
massive plutonic bodies as they slid under the influence of
gravity in an east to northeast direction from the flanks of
a basement dome in the vicinity of the Gronq culmination.
133
 In addition to the fold and low- to intermediate-angle P l u t o n i c m e m b e r s o f Gjersvik eruptive sequence in
thrust structures created during the first two periods of Skorovas area
folding, the topography and geology of the Skorovas area On the 1:100 000 scale map of the Trones quadrangle com-
has been strongly influenced by the formation of a complex piled from the work of Foslie 12 the plutonic rocks of the
system of high-angle faults and fractures. For the most Skorovas area occur in two groups. The first group com-
part these have suffered small displacement of the order of prises the tectonically bounded massifs of Gr~ndalsfjell and
metres, but along the southwest contact of the Gjersvik Nes~piggen, which, though they have strongly tectonized
eruptive complex with the polymict conglomerate oblique envelopes, preserve much of their original igneous fabric in
slip normal faulting has resulted in a vertical displacement
the interior. The second group occurs as an arcuate belt
of the order of 500 m (Figs. 4(a) and 5). The trend of
lying within the volcanic succession to the north, west and
these fractures is predominantly in a NNE to northeast
south of the Skorovas ore deposit (Fig. 4(a)). The plutonic
direction and their formation post-dates the main periods of
rocks of this belt have been subjected to the penetrative
folding in the area. The late fracture patterns in the
deformation that affected the enclosing volcanic rocks and
Skorovas area remain a problem for future investigation.
have responded tectonically as part of the volcanic level
In all probability they can be attributed to the final stages
during deformation.
of Caledonian tectonism, but the influence of later events, The plutonic rocks of the Skorovas area were divided
such as basement reactivation during Mesozoic rifting, by Foslie into two principal compositional groupings, as
cannot be discounted. shown in the map of the Trones quadrangle. 12 Gabbros

/
\ ,~" /
\ ~\\ \ / / /
\ ~ \ 9 / / /
\ \ \ "-..I / / /
/
Y /
.~..~_..~_~ - - ....._~ / / /

td /" \ ...~-- / I-----

I! / /~',,.~....,_-.__. //
II ~-- /////
/
I I "~' "" /IIL/I'~ ~,x / //// 1,/ / /
I I i( // I~~L-~, I / . /~'/"///'." /
I I ; / ~,' VI ]/ / / ./ _ /// / / / ." /

I I ,~ / / | ! _~-"~-"~ /. . ~. / . /-- / _/ / / .,-- ,./

_../ l/
/",
./_
/ l V ~
I ~,-',n/--"
I ~[ / ~ - : /
/
.,~
_/-/
--- /
/
--
/
/
.: -
"

_.i > ~ - " % ~ , ''LJ /

" lII/1
'~. ,
~ .~/'_/ / ./_/

-----~~ i ' // / // //_/ /

/..~'/ ~ ,< // / / ./" ,./
..../'< // ~ / ' " .///, / / /
I</ . < ' ~ - - ~ (.C/.#'~/ /,< 7 7 / / / / 77
/" ~---j~ ._1 . ~/ ./ /
/ /i x~/_./ ..~1/'3<" .." / >"
/ / /",~ ..,..j / x / / ~ . / / /

Skorovas orebody
S
--..-~,,-~-~~ ~ / f2axial trace
/
~"---.L__J j / / 0 1 2 3kin ...~ "A'''c" fl c[eclvclqe
f
/ I, I I I ~..~
orientation
/ ~.. Skorovasklumpen-
// Grendalsfjell
/ tectonic
-/ discontinuity
(b) I
i

134
 r

:=
o
O
v,

o
.E

t-
O

r
{33
O
O

"O
o~
Nk-

E
,~
rJ 3

It'1

0~
14.

135
 Fig. 6 Panoramic view of southeast margin of Gr~ndalsfjell massif seen from point of
vantage on trondhjemite intrusive of Skorovas intrusive arc. Major thrust horizon
separates diorite and gabbro (d) together with hornfelsed envelope (x) from structurally
underlying schistose extrusives (e). A further thrust separates extrusives from
trondhjemite (t) in foreground. Location of photograph (Fig. 7 (b)) shown by crossed
circle at far left of vista

of various facies were distinguished and at the opposite end
of the compositional scale trondhjemite, tectonized granite
and granitic dykes and sills were also shown. There is no
reference on the map to the occurrence of intermediate
dioritic rocks in the immediate area of Skorovas, although
Foslie was undoubtedly aware of their existence because
diorites are mapped as a thin border zone to the north of
the Gr~ndalsfjell massif and to the west of Heimdalshaugen.
The detailed mapping carried out by the present writers
has shown that dioritic rocks of intermediate composition
form an important component in the plutonic sequence and
that a definite relative chronology of intrusion can be
recognized.
It has already been noted that the plutonic sequences in
the Gr~ndalsfjell and NesSpiggen massifs and the plutonic
bodies that compose the arcuate intrusive belt (Fig. 4) are
tectonically separated, and it is convenient to discuss their
plutonic histories separately.
Grdndalsfjell massif
The starkly exposed rocks that compose the Gr~ndalsfjell
massif provide spectacular evidence of their relative ages.
The earliest intrusives are fresh layered olivine gabbros,
which occur as large xenolithic masses or rafts with
maximum dimensions of the order of 70 m x 200 m, con-
tained in a matrix of metamorphosed gabbro and horn-
blende diorite. The cumulus layering of the gabbro bodies
is sub-vertical in attitude with a predominantly east-west
trend. This must be accepted as evidence of significant
post-cumulus displacement.
The composition of the layered gabbro varies from troc-
tolite to hypersthene gabbro and in all facies hypersthene
occurs, either as a reaction rim around olivine or as indepen-
dent ophitic grains. The mineralogy of the gabbro ~sthus
compatible with crystallization from a tholeiitic magma.25'67
The nature of the xenolithic relationship is shown in
Fig. 8(a), and it is clear that the hornblende diorite is a
major component of the Gr~ndalsfjell massif. The
peripheral contacts of the fresh layered gabbro with the
diorite display a distinctive pattern of retrograde alteration,
which partly follows the primary igneous layering and
partly exploits crosscutting joints to produce a distinctive
weathered surface (Fig. 8(b)). The alteration leads to the

136
uralitization and chloritization of the augite and hypers-
thene, the serpentinization of the olivine and saussuritic
degradation of the calcic plagioclase to produce albite,
epidote, clinozoisite and calcite. In the troctolitic facies
of the gabbro the growth of considerable quantities of
chlorite within the plagioclase accompanies this breakdown.
The alteration is ascribed to the contribution of water
from the dioritic magma, which led to aretrograde sub-
solidus hydration in the pre-existing mass of layered
gabbro.
The various facies of altered gabbro may extend for a
considerable distance beyond the boundaries of the fresh
layered rocks, and the distinction between altered gabbro
and hornblende diorite is made in the field on the basis of
the persistence of fluxion banding and layered structure
within the surrounding aureole of hydration. The horn-
blende diorite is characteristically composed of subhedral
dark green grains of hornblende together with saussuritized
plagioclase of intermediate composition and accessory
Fe-Ti oxides. The iron oxides are frequently altered to
sphene and the hornblende is generally partly chloritized.
One of the most striking features of the hornblende
diorite is the occurrence of coarse patches and pegmatoidal
veins, 0.5-3 m wide, consisting of euhedral hornblendes,
commonly up to 10 cm in length, set in a matrix of
andesine feldspar together with accessory amounts of
magnetite and pyrite. The pegmatoid veins show rhythmic
banding parallel to their contacts. This can be interpreted
Fig. 7 Typical dislocated isoclinal style seen in minor
folds of first generation in chert bands to south of
Nes~klumpen (a) (top) and (b)localized post-schistosity
folding and incipient crenulation cleavage of second
generation formed in zone of high strain in schistose
greenstones adjacent to tectonic boundary of G r~ndalsfjell
massif. Location of photograph is shown in caption to
Fig. 6
as a result of episodic deuteric crystallization from hydrous
fluids circulating within the largely consolidated dioritic
body. These rocks can be justifiably described as appinites,
and their presence implies that the level of exposure seen
in the eastern margin of the Gr~ndalsfjell massif corres-
ponds to the upper portion of a differentiated dioritic
body.2S, 78
At the margins of the hornblende diorite, close to the
contact of the plutonic mass with the enclosing greenstones,
a quartz-diorite facies occurs locally.
At least two generations of impersistent basic dykes cut
both the gabbro and the diorite with its appinitic facies. The
dykes are thin, usually less than 20 cm in width, and have a
northeastery trend with steep dips to the northwest. They
are composed of fine-grained hornblende and plagioclase,
together with minor iron oxides, and are locally porphyritic
with plagioclase crystals up to 7 mm long.
The final eruptive event within the Gr~ndalsfjell complex
was the emplacement of a swarm of leucocratic porphyritic
granodiorite dykes, which show a predominantly north-
easterly trend and dip steeply to the northwest. The dykes
are commonly 1 - 2 m thick and can be followed for dis-
tances of 1 - 2 km before they pinch out. Close to the
margins of the plutonic mass, and also within it, these
dykes show well-developed tectonic foliation and, locally,
mylonitic facies, which demonstrates that the northeast-
trending fracture system has been the focus of significant
post-intrusion tectonic strain. The granodiorite dykes are
composed dominantly of sodic plagioclase (roughly of
Fig. 8 Northeast face of Gr~ndalsfjell massif displaying
occurrence of rafts of unaltered layered gabbro (dark)
within dioritic matrix (a) (top) (rafts are of the order of
6 0 - 1 0 0 m x 200 m) and (b) field appearance of hydrated,
uralitized envelope that borders large xenolithic masses of
fresh layered gabbro on Gr~ndalsfjell (Fig. 8(a). Troctolitic
gabbro shows strong differential weathering of pyroxene,
feldspar and olivine, producing pitted surface. Uralitized
assemblage weathers uniformly by comparison
137
 ol igoclase composition), quartz and accessory m icrocline,
biotite, hornblende and sphene. The ferromagnesian
minerals are generally partly chloritized and the feldspars
have been variably altered to fine micaceous aggregates
(sericite or paragonite). Because of the modal composition
of these dykes, which is dominantly oligoclase together
with quartz and with only accessory amounts of potash
feldspar, the rocks may properly be described as trondh-
jemite in the sense of the definition applied by Goldschmidt
in 1916. 22
This summary of the igneous relationships preserved
within the plutonic massif of Gr~ndalsfjell shows clearly
that a considerable volume of dioritic magma was emplaced,
probably at an intermediate to high crustal level, evidently
b y invading a pre-existing mass of layered gabbro, which is
the oldest and presumably the deepest representative of
the plutonic assemblage in the Skorovas area. It may be
added that magmatism must also have been bimodal - that
is to say that the magmas were supplied from two genetically
different sources, the first tholeiitic and the second calc-
alkaline. A range of similar igneous relationships occurs in
the Nes,~piggen massif to the south (Fig. 4).
In addition to the main gabbro-diorite body of the
Gr~ndalsfjell massif delineated by Foslie on the map of the
Trones quadrangle, a significant mass of 'fine-grained gabbro'
is also shown lying directly to the north of Skorovatn. This
forms the imposing topographic feature of Skorovasklumpen
in the basal slope of which lies the extension of the main
thrust surface, which is interpreted as separating the tectoni-
cally 'massive' plutonic level from the highly deformed
volcanic level. This feature is shown on the geological map
of the Skorovas area and in the accompanying structural
synth'esis (Fig. 4). Investigation has shown that Skorovask-
lumpen and the narrow belt of similar character that can
be followed along the eastern margin of the Gr~ndalsfjell
massif are composed predominantly of metamorphosed
basic volcanic rocks, together with interbands of acid
(dacitic-keratophyric) composition and a proportion of
high-level basic intrusive material. The basic rocks of the
belt adjacent to the Gr~ndalsfjell massif are partly incorpor-
ated in a xenolithic screen of considerable complexity. The
original igneous contact of the diorite with the volcanic
country rocks is preserved intact within the main tectonic
boundary (Fig. 4(a)) and can be mapped over a distance of
4 km. Original volcanic structures, notably pillow forms and
vesicles, are preserved within xenolithic masses and testify
to the volcanic origin of the country rocks. Similar textural
evidence of volcanic origin has been found within the basic
sequence that composes Skorovasklumpen.
The reason for the classification of the rocks of
Skorovasklumpen as fine-grained gabbros by Foslie 12 and
other workers lies in their amphibolitic metamorphic grade,
which has produced a mineralogy dominated by hornblende
and intermediate to calcic plagioclase. The presence of
epidote as a constituent mineral throughout a significant
part of the amphibolitic sequence implies that these higher-
grade rocks span the epidote amphibolite facies to enter the
field of amphibolite facies. Since there is no association
with pelitic rocks, a precise description of the prograde
regional metamorphism of the basic rocks of the Skorovas
area depends chiefly upon a determination of the progressive
changes in the composition of the hornblende and plagio-
clase, which must await further detailed work. Broadly,
however, the mineral assemblages accord with the sequences
regarded by Miyashiro 31,32~36 as typical for the regional
metamorphism of mafic rocks at low to intermediate
pressure.
One of the conspicuous features of the mineralogy of the
amphibolite facies rocks of Skorovasklumpen is that
pyrrhotite replaces pyrite as the accessory iron sulphide -
an observation that is readily made in the field. The
amphibolitic lavas locally display distinct pentrative tec-
tonic lineation of the amphiboles, and this lineation can be
observed in the amphibolitized volcanic xenoliths in the
diorite. Amphibolite grade metamorphism evidently took
place under the influence of early tectonic stresses with
which the emplacement of the gabbro-diorite massif was
partly synchronous. The establishment of a precise chrono-
Iogy for these events will depend upon the evidence provid-
ed by future detailed petrographic work. It is probable,
however, that the contact aureole of the Gr~ndalsfjell
massif and the amphibolitic rocks of Skorovasklumpen com-
pose a continuum within the field of low to intermediate
pressure in which regional and contact metamorphism
converge. 34
Rocks of the arcuate intrusive belt
The intrusive arc differs from the plutonic massif of
Gr~ndalsfjell in three distinctive ways: (1) no unmetamor-
phosed gabbroic bodies have been found in which a
plagioclase-pyroxene-olivine assemblage is preserved;
(2) penetrative deformation has produced distinctly tec-
tonic fabrics throughout most of the arc and mineral
assemblages are reduced, for the most part, to those stable
within the greenschist facies; and (3) quartz-rich dioritic to
granodioritic rocks compose a large part of the complex and
the eastern extremity of the arc joins a large granodiorite
mass to the south of Tunnsj~en (see Fig. 4(a)).
Apart from these significant differences, which can pro-
bably be explained in terms of the higher level of emplace-
ment of the arc complex within the volcanic sequence, the
relative chronology of intrusive episodes in the arc is the
same as that observed in the Gr~ndalsfjell massif. The most
basic rocks are the oldest and the successively younger
intrusions become increasingly silicic.
The degree of deformation within the plutonic arc is
often extreme; but, locally, the original geometry of intru-
sion is preserved as shown in Fig. 9. The range of composi-
tions present in the rocks of the arc is very wide and in-
cludes hornblende gabbro, diorite and granodiorite (trondh-
jemite). The definition of the petrographic character of
each generation is complicated by the incorporation of
xenoliths of earlier basic volcanic and plutonic rocks as well
as by extreme deformation, local silicification and reduc-
tion of the primary minerals to greenschist assemblages.
Fig. 9 Trondhjemitic net veining in mafic diorite and
hornblende gabbro on southwest Grubefjell

138
It is sufficient for the purposes of the present discussion to and adjacent competent flow units or intrusives. It is
confirm the presence of gabbro, diorite and trondhjemitic possible, however, in the vicinity of Skorovas, to observe
granodiorite as components of the arc and to suggest that pillowed sequences in which the original geometries are
these are, in part, equivalent to the plutonic complex ob- nearly preserved, as shown in Fig. 10. The dimensions of
served in the Gr~ndalsfjell massif. Prior to the major stages pillows are variable, but diameters within the range 0.5-2 m
of Caledonian deformation leading to the allochthonous are typical. In addition to pillowed basaltic flow units,
emplacement of the Gjersvik Nappe, it is assumed that the there is a significant volume of deformed meta-hyaloclastite
rocks of the intrusive arc and those of Gr~ndalsfjell were pillow breccia associated with the basaltic unit, which
part of the same complex plutonic continuum. structurally overlies the orebody (see Fig. 17). The pillow
breccia lithology is locally transitional to tuffaceous and
Volcanic rocks of Gjersvik eruptive sequence in Skorovas agglomeratic basic pyroclastic facies and can be traced
area and their metamorphic condition within a radius of 3 km around the orebody.
The abundance of amygdales, ranging in size from 2 to
The volcanic rocks of the Gjersvik eruptive complex are of
10 mm and, exceptionally, reaching sizes of 5 cm, indicates
geological and economic interest for they are the host rocks
that the lavas were erupted at relatively shallow depths,
of the Skorovas deposit. The volcanic succession has suffer- probably of the order of 100-500 m. 29~37 The primary
ed extremely from the effects of deformation and low- mineralogy has been completely replaced or pseudomorphed
grade metamorphism under conditions of the greenschist
by assemblages composed of chlorite, albite, epidote,
facies. These modifications, together with the primary
actinolite, calcite and sphene. Stilpnomelane, regarded by
complexity of the volcanostratigraphy, have been obstacles
Miyashiro 36 as atypical of low- to medium-pressure
to the systematic mapping of the greenstones.
It has long been recognized that the Gjersvik greenstones
are composed of a sequence of basic to acid rocks, includ-
ing basalts, andesites and keratophyres of distinctly spilitic
affinity. 21,41 Because of the confinement of systematic
geological studies to the immediate vicinity of the Skorovas
mine itself, previous summaries of the volcanic stratigraphy
have been limited. During the present study an attempt
has been made to document the range of primary volcanic
structures that can be observed at the macroscopic scale
within the acid and basic members of the stratigraphy and
to examine their geometry with respect to metamorphism
and deformation.
It is difficult to assess the relative volumes of basic and
acid rocks within the volcanic sequence, but it can be said
with confidence on the basis of regional mapping that, in
the general area of Skorovas, the dominant volcanic rock
types are basalts and basaltic andesities with lesser amounts
of andesitic and keratophyric rocks. This fact is apparent
from the relative outcrop of acid and basic rocks shown in
Fig. 4(a), although this can only be treated as an approxi-
mate guide. Because of the deformed and dislocated con-
dition of the sequence and the present level of erosion, the
maximum thickness of volcanics is difficult to assess. A
reasonable estimate based on constructed geological sec-
tions, taking into account the effects of tectonic flattening
and extension, can be given as 3 - 4 km.
The sedimentary component within the pile is limited
to very thin, but stratigraphically persistent, iron- and
silica-enriched beds produced as a result of chemical dis-
persion during volcanic activity. Banded calcareous
greenschists, which have been considered by previous
writers to be of sedimentary origin, can be explained as Fig. 10 A, Deformed basaltic pillow lavas observed on
tectonic facies originating from metamorphosed and northern slopes of Grubefjell below orebody. Cuspate
flattened basic flow units. bodies of grey chert that occupy interstices between
The primary mineralogy of all the rocks in the volcanic pillows are conspicuous. In cases of extreme deformation
succession has been degraded to assemblages of the survival of these chert bodies within chloritic schist provides
greenschist facies. Textural evidence shows that the crea- a useful guide to original volcanic structure of rocks.
tion of the greenschist facies assemblages took place during B, Basaltic pillows from flow exposed on southwest shore
two episodes, the first of which was prior to the first stage of Tredjevatnet. Eruption of pillowed basalts followed
of penetrative tectonic deformation. The evidence con- deposition of a dispersed exhalite horizon in vicinity of
firming this metamorphic chronology is best preserved with- Tredjevatnet centre. Layer of ferruginous silica gel, disturb-
in the basic members of the sequence. ed during eruption of the basalts, formed a jasper matrix
for the pillows. Chloritized chilled margin of pillows is
Basaltic and andesitic lavas conspicuous. Significant amounts of pyrite are also found
The state of deformation of the basaltic rocks varies in association with jasper pillow matrix, the pillow lavas
according to their position with relation to the early iso- lying stratigraphically but a few metres from horizon of
clinal folding, the numerous lower-order thrust horizons massive pyrite

139
regional metamorphic assemblages, is a conspicuous com-
ponent of the basaltic andesites in the mine area. This can
probably be explained in terms of the iron enrichment
shown by these rocks (analysis 3, Table 1). Stilpnomelane,
in common with the other greenschist minerals, occurs
dispersed throughout the body of the rock and also as
monomineralic fillings in amygdales and in crosscutting
veinlets. The dominant mineralogy of the amygdales within
the pillowed basalts varies widely. Combinations of two of
the common greenschist mineral species are usual, involving
quartz, epidote, calcite, chlorite, albite and pyrite. Actino-
lite is not usually found in amygdales. Within certain parts
of the Skorovas area the dimensions of the amygdales and
their mineralogy have been useful in discriminating between
individual flow units, although amygdale mineralogy cer-
tainly cannot be applied as a universally reliable criterion of
stratigraphy.
Within the more massive andesitic and basaltic rocks,
original flow textures are preserved by the orientation of
the altered plagioclase microlites. Augite phenocrysts are
pseudomorphed by actinolite and chlorite and the acces-
sory iron-titanium oxides are largely replaced by sphene.
The basalts are not conspicuously porphyritic and igneous
textures are frequently concealed in the meshwork of fine
actinolite, chlorite, epidote-clinozoisite and albite into
which the rocks have been transformed.
The effects of greenschist metamorphism are not only
apparent at the micro scale but are also demonstrated by
the gross redistribution of the rock components, which has
produced massive bands and lenticular knots and spheroidal
bodies, the mineralogy of which is predominantly epidote
with lesser amounts of albite, quartz, etc. These bodies
with dimensions of the order of tens of centimetres are
arranged parallel to the surfaces of the pillow structures or
as discontinuous layers parallel to flow surfaces within
massive basalts and basaltic andesites. The typical form
of these bodies is shown in Fig. 11.
The epidote-rich segregations are evidently pre-tectonic.
During the first period of penetrative deformation the
chloritic mass of the pillowed basalts has tended to develop
a good schistose fabric and the geometry of the pillows, as
a whole, has become flattened to varying degrees. The
epidote layers have behaved as competent bodies and have
deformed by brittle fracturing; in extreme cases the
epidote bodies are preserved as cataclastically reduced
streaks and boudins within the highly flattened pillows.
The textural evidence clearly demonstrates that an
important episode of greenschist metamorphism was re-
sponsible for pervasive alteration and gross reorganization
of the mineralogy of the basic rocks prior to the tectonic
event responsible for the early penetrative schistosity in
the Skorovas region.
Deformation of the volcanic pile also took place under
conditions of the lower greenschist facies and the miner-
alogy established during the primary metamorphic episode
was not changed, but tectonic facies were produced as a
result of further redistribution and segregation of the
various mineral species.
The metamorphic alteration that took place in the
earliest event prior to the deformation of the rocks can be

Fig. 11 A, Pillowed basaltic lavas from northwest of Havdalsvatn showing development of pre-deformational metamor-
phic segregrations of epidote-rich materials (e) parallel to pillow margins. During tectonic flattening epidote layer has re-
sponded by developing a system of brittle fractures. B, Lenticular segregation of epidote (e) of pre-deformation age in
massive andesitic lavas southeast of Store Skorovatn. Conjugate pattern of brittle fractures produced during deformation
of competent lenses is explicitly developed, as in generation of dilatant fractures filled with quartz, chlorite and carbonate

140
Table 1 Whole-rock analyses of Skorovas volcanics. Analyses (1-9) with average values of ocean-floor basalt (10; Cann4)
and island arc tholeiite (11, Pearce and Cann 46) for comparison. 1, Porphyritic quartz keratophyre, Grubefjell; 2,
quartz keratophyre, Grubefjell; 3, andesite with stilpnomelane, Grubefjell; 4, andesite, Grubefjell; 5, andesitic clasts in
agglomerate, Grubefjell; 6, pillowed basalt, Grubefjell; 7, pillowed basalt, Grubefjell; 8, basalt, 6 km southwest of Grubefjell;
9, basalt, northeast I~verste Nes~vatn

% 1 2 3 4 5 6 7 8 9 10 11

SiO2 72.23 70.39 53.07 59.34 56.12 50.15 49.30 48.99 50.13 49.61 52.86
AI20 3 11.82 12.27 14.13 15.40 12.20 13.70 13.81 16.55 14.76 16.01 16.80
TiO2 0.80 0.27 0.77 1.06 0.96 1.54 1.89 1.30 1,24 1.43 0.83
Fe203 2.14 3.37 6.48 3.49 3.31 3o31 + + + + +
FeO 1.28 0.44 6.62 6.01 6.44 7.78 14.70" 13.97" 14.95" + +
MnO 0.03 0.01 0.19 0.23 0.11 0.16 0.21 0.17 0.15 0.18 +
MgO 0.36 0.45 4.40 2.68 4.70 4.70 5.49 5.74 6.00 7.84 6.06
CaO 1.27 0.24 4.66 2.38 4.44 4.89 4.92 5.33 3.50 11.32 10.52
Na20 7.50 8.00 5.21 7.50 6.25 8.81 6.47 6.88 7.30 2.76 2.08
K 20 0.07 0.02 0.51 0.19 0.02 0.52 0.43 0.66 0.55 0.22 0.44
P2Os 0.24 0.03 0.10 0.18 0.12 0.17 0.11 0.06 0.03 0.14 +
Loss on
ignition 1.06 2.24 1.90 2.24 3.57 2.81
Total Fe as
Fe20 3 3.56 3.86 13.83 10.17 10.46 11.95 + + + 12.63 11.45

Total 98.90 99.49 98.04 100.70 98.24 98.54 99.54 99.64 98.24

* Total Fe as FeO.
+ Valuenot obtained by analytical method used.

ascribed to contemporaneous alteration of the volcanic
rocks in situ as a result of the thermally driven circulation
of sea water in the upper layers of the lava pile close to
the site of eruption on the Ordovician sea-floor. Consider-
able evidence has accumulated in recent years to show
that in-situ alteration of the mineralogy of submarine
basalts to produce assemblages of greenschist and lower
amphibolite facies is a phenomenon of wide occurrence
within the upper layers of the sea-floor. 33~s Humphris 27
recognized that the metamorphic assemblages in recent
submarine basalts from the Mid-Atlantic Ridge can be divid-
ed into chlorite-dominated and epidote-dominated types.
It is suspected that this division reflects a process of meta-
morphic segregation similar to that seen in the basalts
of the Gjersvik sequence.
The in-situ hydrothermal alteration processes evidently
involve the convective circulation of large volumes of sea
water relative to the altered rock. Water : rock ratios of the
order o f > 104:1 were calculated by Spooner and Fyfe 59
and the alteration process is believed to extend to a depth of
at least 2 km within the lava pile. 59,6~
The in-situ sea-floor metamorphism of the Gjersvik vol-
canic sequence was evidently an important event and, as well
as causing gross mineralogical changes by chemical redistrib-
ution within the scale of individual flow units, bulk changes
in the chemical composition of the lavas also occurred,
leading to the conspicuously spilitic chemistries shown by
the analysis in Table 1.
The recognition of the pervasive pre-deformation in-situ
sea-floor metamorphism of the Gjersvik basalts also helps
to resolve the controversy that surrounds the tectonic status
of disturbances of the Trondheim type. 11~51~55 The polymict
conglomerate that unconformably overlies the volcanic
sequence was formed prior to deformation and alloch-
thonous transport of the Gjersvik Nappe. This is easily
demonstrated on a local scale by the pervasive schistose
fabric of the matrix and the distinctive stretching of the
competent clasts parallel to the axes of the early isoclinal
folds (Fig. 12(A)). Itcan also be demonstrated on a regional
scale by mapping the level of unconformity through the
isoclinal folds of the first deformation (see Fig. 5).
The conglomerate is composed of boulders directly
derived from the plutonic and volcanic sequence that under-
lies it. Locally, the composition is dominated by marble
clasts with associated pebbles of jasper, and in other places
the clast population is dominated by boulders of phaneritic
granodiorite (trondhjemite), diorite, meta-gabbro and
various of the resistant volcanic rocks. Pebbles of kerato-
phyre are common, but of greatest interest are the pebbles
of the metamorphic epidote assemblage (Fig. 12(B)), which
have evidently been derived by erosion of the metamor-
phosed basalts.
Final and conclusive evidence is thus provided for a
Lower-Middle Ordovician metamorphic event pre-dating
the Gjersvik Disturbance. The metamorphism was produced
by the thermal and hydrothermal effects associated with the
contemporaneous eruptive activity embodied in the l
Gjersvik Nappe. The tectonic movements involved in the
formation of the polymict conglomerate were predomin-
ately vertical as opposed to lateral and must have been re-
lated to an early stage of tectonic evolution within the belt
of Lower-Middle Ordovician eruptives of which the Gjersvik
Complex was a part.
The status of a possible metamorphic event pre-dating the
Trondheim Disturbance has been discussed elsewhere. 11'65
Further investigation will probably reveal the ubiquity of
sea-floor-hydrothermal metamorphic assemblages as clastic
constituents of the polymict conglomerates of the Venna

141
and equivalent horizons. It may be regarded as axiomatic
that such assemblages should be incorporated into the
conglomeratic rocks produced by episodic uplift of the
Ordovician sea-floor and that the history of metamorphism
would be as extended as the history of submarine volcanism.
Magmatic activity in the belt continued after the
erosional event. The evidence for this is provided by
quartz-feldspar porphyry dykes that cut both the eruptive
complex, the unconformity and the overlying conglomer-
ates prior to the first phase of deformation. These dykes
are similar in composition to other granodioritic rocks
within the eruptive complex and are regarded as the latest
product of calc-alkaline magmatism within the Skorovas
area.
Acid to intermediate flows and pyroclastics
There are, within the Skorovas region, a range of acid lavas,
tuffs and agglomerates, which are locally abundant and
form horizons that can be traced laterally over considerable
distances (see Fig. 4(a)). These rocks are of critical interest
because they are closely associated with both the Skorovas
orebody itself and with a variety of iron- and silica-rich
sediments, which, following the conceptual terminology of
Carstens 6,7~s and Oftedahl, 42 are appropriately described
as 'exhalites'.
Fig. 12 A, Typical appearance of polymict conglomerates
as seen to northwest of Havdalsvatnet. Flattened boulder of
trondhjemite (t) displays tectonic fracture pattern character-
istic of its brittle behaviour. Associated boulder of marble
(m) has deformed in a ductile fashion. B, Large pebbles of
pre-deformational epidote-rich metamorphic segregations
derived by erosion from underlying lavas are a common
constituent of greenstone-bearing facies of polymict conglo-
merate. Example photographed close to unconformity on
southern shore of Tredjevatnet
142
Because of the deformation of the volcanic sequence and
the inherent lateral variability of the volcanostratigraphy it
is not possible to describe a unique and widely applicable
type succession. The distribution of the various facies of
acid rocks within specific parts of the Skorovas area suggests
that a minimum of four centres of acid pyroclastic eruption
were active. Their products are preserved, as far as it is
possible to tell, at an approximately similar level in the
volcanic sequence. In the vicinity of the Skorovas orebody
there is stratigraphic evidence of at least two pyroclastic
levels, the lowest of which is exposed in the basal slope of
Skorovasklumpen to the north of Store Skorovatn (this is
shown in Figs. 4(a) and 5).
The orebody itself evidently lies within the vicinity of
one eruptive focus, which will be called the Grubefjell
Centre. The other centres, tentatively distingbished, lie
west and southwest of Tredjevatnet (the Tredjevatnet
Centre), to the east of Overste Nesavatn
o o
(the Nesavatn
O
Centre), and further east in the terrain near Blahammeren
(the Blahammeren
o Centre). The main belts of acid rocks
shown in Fig. 4(a) serve to identify these centres. It is
difficult to judge whether the centres represent independent
volcanic structures or lateral eruptions on the flanks of a
single polygenetic edifice.
The acid volcanic horizons show a range of well-preserved
pyroclastic fabrics to which Oftedahl 41,42 drew specific
attention. Various agglomeratic facies are visible in the
acid horizons in the immediate vicinity of the mine (see
Fig. 14). Distal pyroclastic facies include fine tuff bands
with associated exhalite sediments (Fig. 15(a)). Such
horizons are spread over large areas and are thus valuable
stratigraph ic markers.
Pyroclastic facies can frequently be traced laterally into
compact porphyritic and aphanitic bands of keratophyric
aspect - presumably, flows or highly modified tufts. In the
O
vicinity of the Blahammeren Centre porphyritic flows are
physically continuous with porphyry dykes from which the
eruptions appear to have originated. The dykes, in turn, can
be traced towards the large mass of trondhjemite that occurs
at the eastern end of the northern limb of the intrustive arc.
The disjunction caused by ddformation at the margins of
the intrusive masses and within the volcanic sequence, how-
ever, denies a conclusive statement concerning the connex-
ions between the plutonic and volcanic levels during clima-
tic episodes of acid eruptive activity.
Ti
pprn
9 9 O.EB.
10000-
"
5000-
0 50 1 0 150 200
Zr ppm
Fig. 13 Plot of Ti versus Zr contents for Skorovas basic
extrusives (circles) and basic intrusives (triangles) showing
abundance of low potash (island arc) tholeiites (LKT).
Distinct trend towards field of calc-alkaline basalts (CAB)
and grouping towards ocean-floor basalt (OFB) also shown
 Chemistries of the acid extrusive rocks from the
Skorovas ore level are distinctly soda-rich (see analyses 1 and
2 in Table 1). Petrographically, the rocks display a modal
composition dominated by albite and quartz, occurring both
as phenocrysts and as the constituents in the aphanitic
groundmass, which is a mosaic of albitic plagioclase micro-
iites and quartz. Whatever mafic silicates may have been
present are now represented by dispersed chlorite. Pyrite
is usually present as an accessory. The rocks are properly
described as quartz keratophyres 25,76 and, taking into
consideration the analyses from the basaltic and inter-
mediate rocks shown in Table 1, it is clear that the Skorovas
volcanic rocks are a spilitic suite.
The question is immediately raised as to the relationship
that such a volcanic suite might have to the plutonic rocks
at various structural levels in the immediate vicinity of
Skorovas. The brief account of the plutonic rocks given
above demonstrates the wide variation in the condition of
metamorphism and deformation displayed by these rocks;
there is no suggestion, however, that the compositions are
abnormally sodic and the feldspars, though degraded by
saussuritization, have original compositions in the range
labradorite, in gabbro, to oligoclase, in trondhjemite.
Fig. 14 9 Blocky pyroclastic texture (a) (top) seen in
keratophyric flow unit on Grubefjell about 1200 m west of
Skorovas orebody. Pyroclastic fragments are slightly flatten-
ed add siliceous matrix stands out as a reticular pattern.
Flow is part of major acid horizon with which orebody is
associated. (b) Agglomeratic facies of keratophyric horizon
shown in (a). Locality is in immediate vicinity of ore
horizon above mine entrance on northeast Grubefjell. Acid
fragments are partly silicified and tectonically flattened. A
competent quartz vein with orientation close to principal
stress responsible for flattening during first stage of pene-
trative deformation has responded by buckle folding
Goldschmidt has given analyses of the type trondhjemites
from the Trondheim district and from localities in western
Norway that show total Na20 values in the range 4.3-6.0
wt ~ and K20 values in the range 1-2.5 wt ~ This
gives a typical Na20:K20 ratio for trondhjemite of the
order of 3:1. Partial analyses of three trondhjemitic rocks
from the Skorovas intrusive arc 56 show that the Na20
contents fall in the range 2 - 4 . 5 wt %and K 2 0 values fall
in the range 1-2.5 wt %. Na20:K20 ratios are of the
order 1:1.5-3:1. This range is clearly of the right order for
trondhjemitic to granodioritic rocks with SiO2 contents of
about 70 wt %. The N a 2 0 : K 2 0 ratios of the spilitic
rocks are one to two orders of magnitude greater than those
seen in the regionally associated plutonics (see Table 1).
Fig. 15 Exhalite horizon (a) (top) 2 km east of ~verste
Nes~vatn. Stratigraphic sequence is complex and made up
of graded lapilli tufts overlain by pink to brown coloured
banded cherty sediments incorporating magnetite, hema-
tite, stilpnomelane and iron-rich amphiboles. Purple chert
band shows isoclinal fold style of earliest deformation with
conspicuous refraction of early cleavage. (b) Banded
pyrite-magnetite sediment typical of reduced facies of
iron-rich exhalites (vasskis). Large pyrite porphyrobasts
have suffered cataclasis and dislocation to varying degrees.
o
Specimen from 1.5 km north of Blahammeren. Scale in cm
143
 A comprehensive programme of whole-rock analysis is
being undertaken at the present time to establish the major
differences in chemistry between the plutonic and the
volcanic sequences, but it is clear that the most significant
chemical difference does lie in the conspicuous enrichment
in sodium, which has evidently occurred in the whole range
of the volcanic suite.
The chemical discrepancy displayed by the volcanic and
plutonic suites of the Skorovas area has been the root of a
lengthy controversy concerning the affinities of spilitic
rocks in general. The problem has been discussed by
Wells, 76,77 Sundius 66 and Vallance, 69,7~ others,
and it is clear, after the review of the problem by
Vallance, 69,7~ that the case for post-eruptive metasomatic
alteration of alkali contents by circulating sea water is
strong. Taken in conjunction with the textural evidence
described above, there seems little reason to doubt that the
spilitic character of the Skorovas volcanic sequence is the
result of metasomatism, which accompanied the sea-floor
metamorphism of the volcanic rocks during Lower
Ordovician times. This metasomatic alteration by circulation
of heated sea water changed the chemistry of the rocks,
notably enhancing the Na20 content and concealing the
natural magmatic consanguineity of the volcanic and
plutonic rocks.
Magmatic affinity of Skorovas'eruptives and their
tectonic significance
The relative mobility of the major elements in basic and
acid rocks during metamorphic alteration poses obvious
problems with regard to the determination of the magmatic
affinity of eruptive sequences and the confirmation of
consanguineity within them. Cann, 4 in 1970, recognized
the possibility of using certain elements, notably Y, Zr,
Nb and Ti, which were unaffected by severe secondary
alteration processes, as indicators of the magmatic affinity
of ocean-floor basalts. Pearce and Cann 46 subsequently
extended this concept for use in determining the tectonic
setting of basic volcanic rocks by empirically defining the
ranges of variation of the stable trace elements in suites of
basaltic rocks collected from various defined oceanic and
island arc environments.
Sixty-nine basaltic rocks from various parts of the
Skorovas district have been analysed for stable trace
elements. In Fig. 13 the values for Ti are plotted against
those for Zr with reference to the fields of various basaltic
magma types as defined by Pearce and Cann. 46 In addition,
the Ti/Zr values for eight associated gabbroic to dioritic
rocks from the intrusive arc are superimposed. These rocks
were chosen for their even phaneritic texture and lack of
conspicuous layering. The plot shows that the basaltic
rocks of the Skorovas district concentrate in the field of
island arc tholeiites with a notable trend towards the field
of calc-alkali basalts. It is also possible to recognize a group-
ing of values towards the field of ocean-floor tholeiites.
The coincidence of the analysed values in the plutonic rocks
with the field of island arc tholeiites is regarded as a con-
firmation of consanguineity in the groups of basic plutonic
and volcanic rocks falling in this field.
Study of the trace elements suggests that the eruptive
sequence in the Skorovas area originated in a tectonic
setting in which basaltic rocks typical of an immature
island arc were being generated. 19~s Moreover, a know-
ledge of the field relationships in terms of the chronology
and relative volumes of the eruptive rocks at the plutonic
and volcanic levels confirms this view. Little quantitative
information is available concerning the relative volumes of
144
the various eruptive products in mature calc-alkaline arcs
and in immature tholeiitic arcs. Baker 2 has given some
comparative estimates based on observations of the South
Sandwich Island volcanic sequence, and these are judged
to be in the same order of proportion as those observed in
the Skorovas area, notably basalt >> andesite > dacite and
~hyolite (or their spilitized equivalents). In the case of
mature calc-alkaline arcs the relationship is of a distinctly
different order - andesite ~ basalt. The field evidence,
taken in conjunction with the supporting information from
chemical analysis and petrographic examination, forces the
conclusion that the eruptives of the Skorovas area are, in
fact, the constituents of an immature island arc of Lower
to Middle Ordovician age formed within an ensimatic setting
peripheral to the Laurentian or the Scandinavian craton.
The eruptive sequence, its magmatic evolution terminated,
was emplaced as the structural and stratigraphic core of the
Gjersvik Nappe during the climactic stages of the Caledonian
orogeny in mid-Silurian times. The tectonic decapitation of
the island arc is believed to have originated with the collision
between the Scandinavian craton-arc margin and a
Laurentian counterpart; 10,24 the tectonic transport in-
volved in the process of emplacement is estimated to have
been at least 200-250 km. 16,17,63'64
Skorovas orebody and peripheral exhalative mineralization
The description of the volcanic host rocks given above
confirms the association between the Skorovas orebody and
an eruptive sequence originating in an immature ensimatic
island arc of Lower to Middle Ordovician age. It is appropriate
to consider the morphology and mineralogy of the ore deposit
and the peripheral exhalite mineralization of the Skorovas
region in terms of the exhalative volcanic hydrothermal
origin proposed for it by Oftedahl. 41A2
The orebody is situated within a part of the volcanic se-
quence displaying distinctly calc-alkaline character. Apart
from the keratophyric pyroclastic and flow units, at the
level of which the orebody is located, the sequence includes
a thickness of basaltic andesites and rocks in the range of
sitica contents appropriate to andesite and dacite, now
represented by spilitized equivalents. The precise strati-
graphic location of the orebody with respect to the acid
horizons is difficult to establish owing to the disjunctive
tectonic style, but there is no doubt that the association
PROFILE 73
E -W 680--
/
/
/
/
/
J
J ///!6 E 6~0
J
J
/
/
J -5
Fig. 16 Two sections of east orebody at profiles 59 and
73 east-west situated 140 m apart along morphological
axis of orebody. Progressive development of a first-phase
isoclinal fold is illustrated together with complex digitated
style of isoclinal closures. Open style of second fold phase
shown by undulation of lower contact of ore on profile
59 east-west
 between ore and keratophyric extrusive rocks is intimate
(see Figs. 4(a), 5 and 17).
The Skorovas orebody, at the present state of develop-
ment, is estimated to comprise between 8 000 000 and
9 000 000 tons of massive sulphide ore, including 1 500 000
tons of essentially pyritic ore with minimal base-metal
content. From the initiation of production in 1952 until
1975-76 approximately 4 700 000 tons of ore were milled
to produce pyrite fines with an average grade of 1.2% Cu,
1.8% Zn and 45% S. This concentrate was marketed
primarily for its high sulphur content. Following the de-
cline in the market for sulphur-rich concentrates, a new
beneficiation plant has been constructed for the production
of Cu and Zn concentrates. Present ore reserves are calcu-
lated as approximately 2 000 000 tons with an average
grade of 1.15% Cu and 2.29% Zn. It is a difficult prob-
lem to assess the average grade of the mineralized body as
a whole since this clearly depends upon the geological-econo-
mic criteria chosen to define it. It is, nevertheless, possible
to state that the mineralogy is dominantly pyritic and that
the sulphur content of massive ore is of the average order
of 35 wt % with Zn > Cu >> Pb. Zinc content is of the
order of 2 wt % and Cu ~< 1%.
Structural style of orebody
The morphological complexity of the Skorovas orebody
caused by tectonic disjunction of isoclinally folded lenses
and the extreme tectonic deformation of the wallrock
envelope has been a considerable obstacle to the clear
formation of a genetic model. 2~
The orebody can be described as an en-~chelon array of
closely spaced groups of massive sulphide lenses, the dis-
Fig. 17 Representative section through east and west orebodies at profile 42 east-west showing principal lithological
divisions of host rocks and position of zinc-rich facies along footwall of principal eastern and western lenses. According to
structural interpretation zinc-rich level is stratigraphic top of ore. Complex digitation of ore is well illustrated
tribution of which has created an elongate ore zone with
a length of approximately 600 m lying in a north to NNE
orientation and with a width of the order of 200 m. A
representative cross-section of the orebody is shown in
Fig. 17.
The lenticular bodies have their principal planes orientat-
ed parallel to the axial planes of first-phase isoclinal folds
and the individual lenses are apparently, to a significant
degree, the products of partial disjunction of fold limbs
within that fold system. In detail, as is shown by Fig. 17,
the ore zone shows a longitudinal division into an eastern
and a western orebody. This division may reflect the shape
of the orebody at the site of accumulation prior to deform-
ation. The lateral extremities of the ore lens systems char-
acteristically show multiple digitation and bifurcation and
there are frequently zones of sulphidic impregnation reach-
ing ore grade that lie between the digitations of massive ore.
As Gjelsvik 20 noted, discordance is locally observed be-
tween the contacts of some of the larger massive lenses and
t h e schistosity of the wallrocks. This evidence, together
with the irregular geometry of the orebody as a whole, was
used in support of an epigenetic mode for the formation of
the deposit, although Gjelsvik conceded that early folding
had probably been an influence in creating its present
morphology and that emplacement took place immediately
following the eruption of the volcanic sequence in Lower
Ordovician times.
It is possible to explain the local discordance between
early schistosity and the contacts of the massive lenses in
terms of the contrast in the mechanical behaviour of the
base-metal-poor pyritic lenses and the volcanic wallrocks
during the flattening and isoclinal folding of the first stage
145
of deformation. The disjunction created by componental
movements at the ore contacts during this early phase must
also have been magnified in response to the stresses imposed
during the second period of folding.
The early deformation in the immediate contact zone of
the orebody was sufficient, because of the contrast in com-
petency, to create a schistose tectonic facies composed
predominantly of chlorite, carbonate and, locally, talc.
These components were derived by segregation from the al-
tered basic host rocks - andesite, basaltic, andesite arid
basalt. The schistose tectonic envelope is shown locally in
Fig. 17. The creation of this envelope facilitated the con-
tinuance of componental movements within the vicinity of
the ore contacts during later deformation.
The history of structural deformation within the orebody
can be summarized as follows:
(1) Early isoclinal folding, accompanied by creation of a
schistose envelope with componental movements in
the vicinity of the orebody contacts, led to a
tectonically disjunct style.
(2) Periods of post-schistosity deformation produced
folds of various scales. In the immediate contact
zone small folds of up to several metres in wavelength
occur sporadically in response to local variation in
orebody geometry. The orebody as a whole, however,
was folded on a broad open style, which is typical
of later deformation in the Skorovas region. This is
shown in the isometric projection (Fig. 16).
(3) The final episode of deformation was marked by high-
angle fractures of low displacement with a general
northerly trend.
The early isoclinal structures display axial alignment in a
north to NNE direction with axial planes dipping at approxi-
mately 25 ~ towards the east. This is reflected in the axial
elongation of the orebody. The later open folds, part of
the regional dome and basin system shown in the structural
analysis (Fig. 4(b)), have steeply dipping axial planes and
9an axial trend of approximately NNW orientation con-
cordant with the pattern of the adjacent structural basis,
on the flanks of which the orebody lies.
Mineralogy and stratigraphy within orebody
The bulk composition of the Skorovas orebody reflects a
mineralogy of comparative simplicity. Pyrite, sphalerite,
magnetite and chalcopyrite are the dominant ore mineral
species. Pyrrhotite is conspicuously absent. Galena occurs
in much smaller amounts, and arsenopyrite and tennantite
occur locally as accessory constituents. This mineralogy
accounts for the average range of trace and minor metallic
elements recorded in analyses of the orebody, the following
values being considered as representative averages: Co,
100 ppm; Ni, 20 ppm; As, 300 ppm; Ag, 10 ppm; and
Au, 0.1 ppm. Cadmium is notably enriched in sphalerite-
rich facies of the ore, reaching values of several hundred
ppm, and Mn reaches similar values in the pyritic facies.
Most of the minor chemical variation can be accounted for
by diadochic substitution within the common ore minerals.
Arsenic and silver are notably contributed by arsenopyrite and
tennantite, and grains of native gold have been observed as
inclusions of 5/Jm in size in arsenopyrite from peripheral
parts of the ore. The principal gangue mineralogy of the
ore consists of chlorite, quartz and calcite, together with
lesser amounts of sericite and, locally, stilpnomelane.
The structural and stratigraphic evidence summarized
here and by other authors 2o21,41 has confined the choice
of genetic models for the orebody to the following
146
alternatives: (1) syngenetic deposition of the stratiform
orebodies under submarine conditions as a result of
emission of metal-rich fluids in the vicinity of an acid
eruptive centre or (2) epigenetic emplacement of the ore-
body by replacement of part of the volcanic sequence in
the vicinity of the eruptive centre, this taking place during
post-eruptive hydrothermal activity in early Ordovician
times.
If the first alternative is to be given favour, it would be
desirable to be able to recognize some evidence of
stratigraphy within the orebody. Gjelsvik 2o21 conducted
a systematic analytical study of the major base-metal con-
tents of ore from 43 drill-holes on selected profiles spanning
the length and breadth of the orebody. The results of this
study showed that the contents of zinc and copper varied
antipathetically, zinc showing a tendency towards enrich-
ment in the peripheral zones of the orebody and copper
tending to concentrate in enriched core regions. It was
also noted that the overall content of copper and zinc
showed an increase towards south of the orebody. In the
southern part Gjelsvik noted that zinc, in particular, is
enriched towards the hanging-wall and in the eastern and
western extremities of the ore lenses. In the central zone it
is enriched in the vicinity of the footwall contact (Fig. 17).
In the northern part of the orebody the composition is
essentially pyritic, with minimal base-metal content. The
analytical data prove a systematic variation in base-metal
content both laterally and vertically within the orebody, and
this is confirmed by petrographic studies and field
observation.
In the course of the present study it has been possible to
recognize facies of the ore that are probably of chemical-
sedimentary origin and those which are essentially tectonic.
The pattern described by Gjelsvik 2o21 probably reflects
the influence of both processes. The primary textural
evidence for the operation of sedimentary processes in ore
deposition is given by the graded banding of the pyritic ores
in which rapid changes of modal composition and grain
size occur from band to band. This type of texture is
shown in the banded pyrite, sphalerite magnetite ore of
Fig. 18(C). It is highly unlikely that such banding is of
tectonic origin. Moreover, where tectonism has had a per-
vasive effect on the ore, the textures are of distinctly tectonic
style (see Figs. 18(B) and 18(D). Figs. 18(A) and 18(B)
show that the deformation of the pyritic lenses was marked
by mutual impaction and cataclasis of the constituent
grains. Any gross tectonic flattening or extension of the
lenses must have been accomplished by relative movement
between the individual grains accompanied by cataclastic
degradation. This mechanism has been described as macro-
scopic ductility by Atkinson, 1 who has also shown that
cataclasis is probably the only significant deformation
mechanism available to pyrite, under dry conditions in the
P - T range appropriate to the greenschist facies. It is un-
likely that deformation took place under dry conditions, 4s
but the range of textural evidence strongly suggests that,
within the massive pyrite, cataclasis was the dominant de-
formation mechanism. Atkinson 1 also notes that the
strength of polycrystalline pyrite is strongly and inversely
dependent on porosity. Large volumes of the Skorovas ore-
body are composed by nearly monomineralic close-packed
aggregates of pyrite with low porosity and, when lithified,
these masses must have behaved in a highly competent
manner relative to the adjacent chloritized lavas and pyro-
clastics. Under the influence of the tectonic stresses pre-
vailing during the first period of deformation it seems reason-
able to propose that the style of deformation within the
orebody may have been controlled by the development of
narrow zones of cataclastic flow within which much of the
tectonic strain would have been accommodated. In this
way, the formation of a disjunct lenticular arrangement of
ore lenses could be explained as well as the rarity of well-
preserved isoclinal structures.
Tectonic mineralogical facies of the orebody are un-
doubtedly recognizable in the base-metal-enriched lenses
and extremities on the lateral periphery of the ore. Zinc
values are enhanced by an order of magnitude and lead
values by two orders of magnitude. This is shown by analy-
sis 5 in Table 2. The typical foliated texture of this ore is
shown in Fig. 18(D), which also displays the incipient
development of a crenulation cleavage related to the second
phase of deformation.
Tectonic mechanisms are not, however, the sole explan-
ation of the peripheral enrichment of base-metal values; nor
do they completely explain the separation between maxi-
mum zinc and copper values in the pyritic ores. There
appears to be a definite stratigraphy in which cupriferous
pyritic ores (analyses 1 and 2 in Table 2) are overlain by
zinc-rich ores with laterally developed facies rich in banded
magnetite and carbonate. Analyses for these ore types are
shown as 3 and 4 in Table 2.
It appears also thata distinct primary lateral variation
may also have been present to account for the generally de,
pleted levels of copper and zinc in the northern part of the
orebody. Final evidence of the operation of chemical-
sedimentary processes in the formation of the orebody is
provided by the occurrence of magnetitic and hematitic
chert bands (jasper)in the foot- and hanging-walls of the
orebody stratigraphically overlying the magnetite and zinc-
rich facies.
Evidence of a primary stratigraphy within the orebody
clearly exists despite considerable tectonic modification.
It is also plain that the metal distribution can be interpreted
in terms of a stratigraphic zonation, which resembles that
found in orebodies of undisputed volcanic exhalative
origin in such areas as the Miocene Green Tuff belt of
Japan.3~
The detailed palinspastic reconstruction of the lateral
and vertical facies variation within the complex Skorovas
orebody is the subject of a current study by Reinsbakken
Fig. 18 A, Typical compact pyritic ore with minor
amounts of sphalerite (grey). B, Mutual impaction relation-
ship in pyrite grains from coarser facies of massive ore
showing cataclastic mechanism of deformation. C, Zinc-
rich ore with magnetite typical of upper stratigraphic levels
of orebody showing evidence of probable primary gradation
and sedimentary banding. Pyrite, white; sphalerite and
magnetite, grey; silicate, dark grey. D, Foliated texture of
zinc; and lead-rich peripheral tectonic facies of ore. Incipient
crenulation cleavage is visibly developed with selective con-
centration of galena (white), Gangue matrix (dark grey) is
composed of carbonate and chlorite
147
and will not be discussed further here. It may be said, (3) The sulphide facies are characteristically improver-
however, that the zonal distribution of copper and zinc ished in base metals other than iron and manganese
within the pyritic mass suggests that precipitation of the (see analysis 6, Table 1).
ore minerals could be explained in terms of an evolving These widespread bands can be explained by a mechan-
chloride-complex model such as that used by Sato to explain ism of explosive volcanic dispersal during the climactic daci-
tic eruptions associated with the various volcanic centres.
Table 2 Average metal values for Skorovas ore types and In the course of such a process rapid and complete mixing
sulphide facies of an extensive exhalite of the residual hydrous fraction of the dacitic magma with
oxidizing sea water will have occurred. The base metals will
% 1 2 3 4 5 6 7 have been subjected to infinite dilution in the course of
such a process, leaving oxidized iron and silica hydrosols in
S 46.80 47.20 38.90 42.28 27.50 51.10 suspension. The hydrosols will have suffered greater dis-
Cu 1.09 2.30 0.99 0.79 1.47 0.06 0.20 persion than the pyroclastic fragments and by subsequent
Zn 0.15 0.80 3.90 9.33 44.20 0.02 0.41 settling will have produced a thin stratum of iron- and silica-
Pb 0.03 0.04 0.05 0.04 4.00 0.01 rich sediment that extends well beyond the limits of the
1, Massivepyritic ore (27 samples); 2, copper-rich ore (14 samples); latter. It is for this reason that the extensive exhalite hori-
3, banded magnetite-rich pyrite--sphalerite ore with carbonate (18 zons are so named. They also constitute valuable time-
samples); 4, pyritic zinc-rich ore at stratigraphic top of orebody (13 stratigraphic markers within the intrinsically variable
samples); 5, Zn-Pb-Cu-rich peripheral ore -- probably a tectonic
facies (2 samples;6 massivebase-metal-depletedpyrite or "vassk[s', volcanostratigraphy.
Havdalsvatn (1 sample);7, relatively enriched pyritic ore, Skorovas The sulphide-magnetite mineralogy of the reducate
(30 samples). facies is to be ascribed to post-depositional bacterial reduc-
tion of iron, deposited in the oxidized condition. A typical
zonation within the Kuroko deposits, s5 The applicability facies of this type is shown in Fig. 15(b).
of such a model depends on the existence of conditions such The simple stratigraphy shown in the ideal section (Fig.
that the metal- and sulphur-enriched hydrothermal solutions 20) can be regarded as the product of a single dispersal
are not rapidly and widely dispersed into the dominantly event. Some exhalites, however, give evidence of episodic
oxidizing conditions of the submarine environment. This explosive and fumarolic activity that results in a complex
requirement must be met by topographical barriers in the cyclic stratigraphy in which tuff bands are intercalated
vicinity of the hydrothermal emanations or by density con-
trasts between the emanating brines and sea water. 54 It is
upon the presence or absence of the conditions outlined that Solutions denser than seawa!er, settling in topographic
the distinction between the hydrothermally intensive and TYPE 1 lI ~ n s l t y layering
the hydrothermally extensive exhalite phenomena in the ~...:..-.-....-.:..--....:.-.-.-...-:~...
Skorovas area is based.
~"~- ..~. Solutions denser than seawater mixing causes increase
Peripheral exhalative mineralization "~"~'~- in density and thus mixture sinks beneath undiluted
TYPE 2a ~ ~ ' ~ " ~ . solutions which rise and are continually mixing and
The magnetitic cherts and jasper found at the stratigraphic stoking
top of the Skorovas orebody signify the restoration of
chemically normal oxidizing conditions in the vicinity of ~

the orebody. These ferruginous siliceous horizons represent "7-,,"

. . ~ . / . ..~.'.- .- . \ SoLutionslighter than seawater, mixing causes increase
~ i i : \ in density and settling occurs as in cases 1 and 2a
a continuum between the intensive and extensive facies of
mineralization (see Fig. 19). The relative frequency of the TYPE 2b II ~ ~ ~
association between acid pyroclastic horizons of various
facies and banded magnetite-pyrite and chert in the
~.. ../-
Skorovas area, and within the Grongfelt as a whole, was one \ / Solutions lighter than seawater end remaining so despite
--. \ / _ __ m i x i n g .thus dispersion is accompanied by
of the primary inspirations for the theory of exhalative- "-.. i / infinite dilution
sedimentary ore genesis expounded by Oftedahl in \ \ / / --- __ .. ~
\\ / ~"\ \
1958, 41,42 who carried forward the concepts formulated by \1 / \ \
C. W. Carstens 7~ in his studies of the Leksdal type /LI~ \
IJf~ \ \
of sedimentary sulphide deposit in the Trondheim district. physical settling of iron and silica hydrosols
Oftedahl 42 emphasized the association between acid
TYPE3 or fl
pyroclastic activity and the formation of the iron- and silica- EXPLOSIVE j ~,,~
DISPERSION "~..~'~"~~_ ~
enriched sediments. Understanding of the various exhala-
tive facies has been carried forward in the course of the (a)
present study.
perlpherQ[ escape of solutions
The main characteristics of the extensive peripheral ex-
halites are noted below. l
(1) The exhalite horizons are relatively thin, 0 . 1 - 2 m in ORE TRAP

thickness, are laterally persistent within the volcano-

stratigraphy and can be traced over distances of the
order of several kilometres.
EPISODIC CO~IPLEX EXHALITE

(2) Internal variations of stratigraphy occur in detail. ~) " CLOSE TO HAJOR FOCUS

The sequence is always marked, however, by a change

from a reducate sulphidic or magnetitic banded Fig. 19 Scheme of interaction of hydrothermal brines
stratum to an oxidate ferruginous chert (jasper). with sea water (a) (top) (after Sato 54) and schematic
These changes occur in a vertical sense (Fig. 20) and eruptive and hydrothermal events in Skorovas volcanic
also, generally speaking, in a lateral sense. centre during climactic dacitic episode (b)

148
with iron-enriched chert bands that show a complex miner-
alogy, including stilpnomelane, iron-rich amphiboles and
chlorites, together with a spinel, commonly of magnetite
composition (Fig. 15(a)).
Fig. 20 Ideal section showing products of single event
dispersal in extensive exhalite as observed in vicinity of
O
Blahammeren centre
As well as being valuable time-stratigraphic markers, the
exhalites may be developed as a tool in identifying vent-
proximal and vent-distal environments and have obvious
value as a guide in exploration. An investigation of exhalites
as an exploration tool is currently being carried out in the
Skdrovas area by Ferriday, Halls and Hembre.
Conclusions
It was recognized in the early stages of the present study in
1972 that the Skorovas area provided a unique window on
the eruptive and ore-forming processes that take place within
a Palaeozoic island arc environment. An attempt has been
made in this paper to describe the major eruptive, hydrother-
mal metamorphic and tectonic processes that have acted to
produce the present geology of the Skorovas area in the
context of its position in the Gjersvik Nappe.
Attention has been specifically directed to the hydro-
thermal processes that take place at the volcanic level, but
it is important to record the occurrence of cumulus ores of
magmatic origin within the plutonic complex. At
Lillefjellklumpen, to the north of Skorovas (see Fig. 4(a)),
a small platinum-bearing pyrrhotite-chalcopyrite-pent-
landite lens has been found in association with a minor body
of metagabbro. This occurrence was described by Foslie
and Johnson-H~st in 1932. i3 The present study has shown
that small cumulus bodies of chalcopyrite-pentlandite-
bearing ore occur at a variety of sites in the layered gabbros
of the deeper plutonic level. At the present time these
bodies are of incidental economic interest only. The whole
range of phenomena described can therefore be said to typify
the ore-forming environment within an ensimatic pericratonic
island arc, and only the porphyry style of sub-volcanic miner-
alization appears to be absent. This may, however, reflect
the immature character of the arc.
The study has also placed the Gjersvik, Trondheim and
related disturbances in their proper geological context as
episodes of uplift associated with the stages of evolution of a
pericratonic arc system in Lower to Middle Ordovician times.
Vertical movements of this style can be said to be a character-
istic feature of the evolution of arc systems,71 and Murphy 38
has described fault-bounded back-arc basins of Tertiary age in
Indonesia that contain up to 8-km thickness of clastic sediments,
which were deposited under subaerial to shallow marine condi-
tions. It is, perhaps, adebatable exercise to attempt to correlate
the timing of such movements, which may be intrinsically of
intra-arc origin, with tectonic events of differing style taking
place in other provinces of the Caledonides that could have
been Iocated,'in Lower Ordovician times, on separate geo-
graphically and tectonically isolated margins of the orogenic
system.3
Acknowledgment
The geological investigation reported in this paper has been
carried out as a joint project between Elkem Spigerverket
A/S (Skorovas Gruber), Norges Teknisk Naturvitenskaplig
Forskningsr~d, The Royal School of Mines, Imperial
College, London, and Geologisk Institutt, Norges Tekniske
H~gskole, Trondheim, under the direction of the Steering
Committee of the Grongfelt Exploration Project, to whom
the authorsare grateful for permission to publish. Particu-
lar thanks are due to Dir. Gunnar L6vaas for continued
encouragement and material support and to Ing. ~ivind
Johannsen, Ing. Ole Sivert Hembre, Herr Jan. Skinstad, Herr
Asbj~rn Lund and Skorovas Jakt og Fiskelag for the
provision of advice, hospitality and strategic shelter,
The authors are notably indebted to Professor Frank
Vokes, Dr. David Roberts, Hr Sigbj~rn Kollung, Dr. Hans
Peter Geis, Professor Christoffer Oftedahl, ProfessorJens A.W.
Bugge, Ing. Roar Jensen, Professor Janet Watson, Dr. M. G. m
Audley-Charles, Dr. R. Mason, Atom. Tore Prestvik and Dr.
G. P. L. Walker, who have shared their knowledge with us
to the benefit of this paper.
Use of the analytical facilities and laboratories at
Geologisk Institutt (N.T.H.), Department of Geology,
Bedford College, University of London, and at the Royal
School of Mines, together with the specialist advice and
assistance of Dr. lan Gibson, Dr. Gloria Borley, Hr Ivar
R~mme and Ms. Giselle Marriner, is gratefully acknowledg-
ed.
The following graduate geologists of the Royal School
of Mines, University College London, and Bedford College,
University of London, made significant contributions to
the investigation by independent field mapping and
laboratory studies: Volker Hirsinger, Quentin Palmer,
Steve Flitton, Roger Scott, Peter Walker, Robert Horsley,
Paul McCormick, lan Mayfield, Michael Horder and Roger
White.
Thanks are due to Mrs Arda Halls and Mrs Christine
Norris, who typed the manuscript for this paper.
References
1. Atkinson B. K. Experimental deformation of poly-
crystalline pyrite: effects of temperature, confining pressure,
strain rate, and porosity. Econ. Geol., 70, 1975, 473-87.
2. Baker P. E. Volcanism at destructive plate margins.
J. Earth Sci., 8, 1972, 183-95.
3. Berry W. B. N. Age of Bogo Shale and western Ireland
graptolite faunas and their bearing on dating early Ordovician
deformation and metamorphism in Norway and BHtain.
Norsk geol. Tidsskr., 48, 1968, 217-30.
4. Cann J. R. Rb, Sr, Y, Zr and Nb in some ocean floor
basaltic rocks. Earth Planet. $ci. Lett., 10, 1970, 7-11.
5. Carmichael I. S. E. Turner F. J. and V~.rhoogen J.
Igneous petrology (New York: McGraw-Hill, 1974), 739 p.
149
6. CarstensC.W. Oversigt over Trondhjemsfeltets
bergbygning. Kgl. norske Videnskabers Selsk. Skr. no. 1,
1919, 152 p.
7. CarstensC.W. Der Unterordovicische Vulkanhorizont
in dem Trondhjemgebiet. Norsk geol. Tidsskr., 7, 1922,
185-270.
8. CarstensC.W. Zur Frage der Genesis der norwegischen
Kiesvorkommen. Z. prakt. Geol., 40, 1932, 97-9.
9. CarstensH. Jernmalmene i det vestlige Trondhjems-
felt og forholdet til kisforekomstene. Norsk geol. Tidsskr,
35, 1955, 211-20.
10. Dewey J. F. Evolution of the Appalachian/Caledonian
orogen. Nature, Lond., 222, 1969, 124-9.
11. Dewey J. F. Rickards R. B. and Skevington D. New
light on the age of Dalradian deformation and metamorphism
in western Ireland. Norskgeol. Tidsskr., 50, 1970, 19-44.
12. Foslie S. Geological maps prepared during 1922-1927
and published at 1:100 000 scale by Norges Geologiske
Unders~kelse as follows:
Namsvatnet med en del av Fr~yningsfjell, Oslo 1957
Tunnsj~, Oslo 1958
Sandd~la, Oslo 1958
Trones, Oslo 1958
Nordli, Oslo 1959
13. Foslie S. and Johnson-H~st M. Platina i sulfidisk
nikkelmalm. Norg. geoL Unders, no. 137, 1932, 77 p.
14. Foslie S. and Strand T. Namsvatnet med en del av
Fr~yningsfjell. Norg. geoL Unders. no. 196, 1956, 82 p.
15. Gale G. H. and Roberts D. Palaeogeographical implic-
ations of greenstone petrochemistry in the Southern
Norwegian Caledonides. Nature, phys. ScL, 238, 1972, 60-1,
16. Gale G. H. and Roberts D. Trace element geochemistry
of Norwegian Lower Palaeozoic basic volcanics and its
tectonic implications. Earth Planet. ScL Lett., 22, 1974,
380-90.
17. Gee D. G. and Zachrisson E. Comments on strati-
graphy, faunal provinces and structure of the metamorphic
allochthon, central Scandinavian Caledonides. GeoL Fi~r.
Stockh. F~rh., 96, 1974, 61-6.
18. Gee D. G. A geotraverse through the Scandinavian
Caledonides - (3stersund to Trondheim. Sver. geoL Unders.
Arsb. 69, no. 9, 1975, 66 p.
19. Gill J. B. Geochemistry of Viti Levu, Fiji, and its
evolution as an island arc. Contrib. Mineral PetroL, 27,
1970, 179-203.
20. Gjelsvik T. The Skorovass pyrite deposit, Grong area,
Norway. In Rep. 21st Int. geoL Congr. Part 16: Genetic
problems of ores (Copenhagen: Berlingske Bogtrykkeri, 1960),
54-66.
21. Gjelsvik T. Distribution of major elements in the wall
rocks and the silicate fraction of the Skorovass pyrite deposit,
Grong area, Norway. Econ. GeoL, 63, 1968, 217-31.
22. Goldschmidt V. M. Ubersicht der Eruptivgesteine im
Kaledonischen Gebirge zwischen Stavanger und Trondhjem.
Kristiania Vidensk.-Selsk. Skr., Math.-Naturv. K 1, 2, 1916,
60--112- (Der Opdalit-Trondhjemit-Stamm)
23. Gustavson M. The low-grade rock~ of the Sk~lvaer
area, S. Helgeland, and their relationship to high-grade rocks
of the Helgeland Nappe Complex. Norg. geoL Unders. no.
322, 1975, 13-33.
150
24. Harland W. B. and Gayer R. A. The Arctic
Caledonides and earlier oceans. Geol. Mag., 109, 1972,
289-384.
25. Hatch F. H. Wells A. K. and Wells M. K. Petrology
of the igneous rocks, 13th edn (London: Murby, 1972),
551 p.
26. Holtedahl O. Paleogeography and diastrophism in
the Atlantic-Arctic region during Paleozoic time. Am. J.
Sci., 49, 1920, 1-25.
27. Humphris S. E. The hydrothermal alteration of
oceanic basalts by seawater. Ph.D. thesis, MIT and Woods
Hole Oceanographic Institute, 1976, 247 p.
28. Jake~P. and Gill J. Rare earth elements andthe
island arc tholeiitic series. Earth Planet. Sci. Lett., 9,
1970, 17-28.
29. Jones J. G. Pillow lava and pahoehoe. J. Geol., 76,
1968, 485-8.
30. Lambert I. B. and Sato T. The Kuroko and
associated ore deposits of Japan: a review of their features
and metallogenesis. Econ. GeoL, 69, 1974, 1215-36.
31. Miyashiro A. Regional metamorphism of the
Gosaisyo-Takanuki district in the central Abukuma Plateau.
J. Fac. Sci. Tokyo Univ., Sect. 2, 11, 1957-59, 219-72.
32. Miyashiro A. Metamorphism of mafic rocks, In
Basalts: the Poldevaart treatise on rocks of basaltic com-
position Hess H. H. and Poldevaart A. eds (New York:
Interscience, 1968), vol. 2, 799-834.
33. Miyashiro A. Shido F. and Ewing M. Metamorphism
in the Mid-Atlantic Ridge near 24 ~ and 30~ Phil. Trans.
R.Soc., A268, 1971, 589-603.
34. Miyashiro A. Metamorphism and related magmatism
in plate tectonics. Am. J. ScL, 272, 1972, 629-56.
35. Miyashiro A. Pressure and temperature conditions
and tectonic significance of regional and ocean-floor
metamorphism. Tectonophys., 13, 1972, 141-59.
36. Miyashiro A. Metamorphism and metamorphic belts
(London: Allen and Unwin, 1973), 492 p.
37. Moore J. G. Petrology of deep-sea basalt near Hawaii.
Am. J. ScL, 263, 1965, 40-52.
38. Murphy R.W. Tertiary basins of Southeast Asia.
Proc. S. E. Asia Pet. Explor. Soc., 2, 1975, 1-36.
39. Nicholson R. and Rutland R. W. R. A section across
the Norwegian Caledonides; Bod~ to Sulitjelma. Norg.
geoL Unders. no. 260, 1969, 86 p.
40. Cftedahl C. Om Grongkulminasjonen og Grongfeltets
Skyvedekker. Norg. geoL Unders. no. 195, 1956, 57-64.
41. Oftedahl C. Oversikt over Grongfeltetsskjerp og
maimforekomster. Norg. geoL Unders. no. 202, 1958, 76 p.
42. Oftedahl C. A theory of exhalative-sedimentary ores.
GeoL FSr. Stockh. Forh., 80, 1958, 1-19.
43. Oftedahl C. A manganiferous chert in the Caledonian
greenstone of Trondheim. Norske Vidensk. Selsk. Forh.,
40, 1967, 48-54.
44. Oftedahl C. Greenstone volcanoes in the central
Norwegian Caledonides. GeoL Rdsch., 57, 1968, 920-30.
45. OftedahlC. Norgesgeologi (Trondheim: Tapir, 1974),
169 p.
46. PearceJ. A. and Cann J. R. Tectonic setting of basic
volcanic rocks determined using trace element analysis.
Earth Planet. Sci. Lett., 19, 1973, 290-300.
47. PearceJ. A. and Gale G. H. Identification of ore-
deposition environment from trace-element geochemistry of
associated igneous host rocks. I n Volcanic processes in ore
genesis (London: Institution of Mining and Metallurgy and
Geological Society, 1977), 14-24.
48. Price N.J. Laws of rock behavior in the earth crust.
In Rock mechanics - theory and practice. Somerton W. H.
ed. (New York: AIME, 1970), 3-23. (Proc. 11th Symp. on
Rock Mech.)
49. Roberts D. Structural observations from the Kopper~-
Riksgrense area and discussion of the tectonics of Stj~dalen
and the N. E. Trondheim region. Norg. geol. Unders. no.
245, 1967, 64-122.
50. Roberts D. Springer J. and Wolff F.C. Evolution of
the Caledonides in the northern Trondheim region, Central
Norway: a review. Geol. Mag., 107, 1970, 133-45.
51. Roberts D. Timing of Caledonian evorogenic activity
in the Scandinavian Caledonides. Nature, phys. Sci., 232,
1971, 22-3.
52. Rui I. J. and Bakke I. Stratabound sulphide mineraliz-
ation in the Kj~li area, R~ros district, Norwegian
Caledonides. Norsk geol. Tidsskr., 55, 1975, 51-75.
53. Roberts D. The Stokkvola conglomerate - a revised
stratigraphical position. Norsk geol. Tidsskr., 55, 1975,
361-71.
54. Sato T. Behaviours of ore-forming solutions in sea-
water. Min. Geol. Japan, 22, no. 111, 1972, 31-42.
55. Sato T. A chloride complex model for Kuroko
mineralization. Geochem. J., 7, 1973, 245-70.
56. Scott R. D. The Grubefjell volcanics. Thesis sub-
mitted in partial fulfilment of the requirements for the
degree of B.Sc., Royal School of Mines, London, 1973,
129 p.
57. Skevington D. A note on the age of the Bogo Shale.
Norsk geol. Tidsskr., 43, 1963, 257-60.
58. Skevington D. and Sturt B. A. Faunal evidence
bearing on the age of late Cambrian-early Ordovician
metamorphism in Britain and Norway. Nature, Lond.,
215, 1967, 608-9.
59. Spooner E. T. C. and Fyfe W. S. Sub-sea-floor
metamorphism, heat and mass transfer. Contrib. Mineral.
Petrol., 42, 1973, 287-304.
60. Spooner E. T. C. etal. 018 enriched ophiolitic
metabasic rocks from E. Liguria (Italy), Pindos (Greece),
and Troodos (Cyprus). Contrib. Mineral. Petrol., 47, 1974,
41-62.
61. Strand T. The Pre-Devonian rocks and structures in
the region of Caledonian deformation. In Geology of
Norway Holtedahl O. ed. Norgo geol. Unders. no. 208,
1960, 170-284.
62. Strand T. Cambro-Silurian deposits outside the Oslo
region. In Geology of Norway Holtedahl O. ed. Norg.
geol. Unders. no. 208, 1960, 151-69.
63. Strand T. and Kulling O. Scandinavian Caledonides
(New York: Wiley-lnterscience, 1972), 302 p.
64. Strand T. The Scandinavian Caledonides: a review.
Am. J. Sci., 259, 1961, 161-72.
65. Sturt B. A. Miller J. A. and Fitch F.J. The age of
alkaline rocks from West Finnmark, northern Norway, and
their bearing on the dating of the Caledonian orogeny.
Norsk geol. Tidsskr., 47, 1967, 255-73.
66. Sundius N. On the spilitic rocks. Geol. Mag., 67,
1930, 1-17.
67. Tilley C. E. and Muir I. D. Tholeiite and tholeiitic
series. Geol. Mag., 104, 1967, 337-43.
68. T~rnebohm A. E. Grunddragen af det centrala
Skandinaviens bergbyggnad. Kgl. Sv. Vet. Akad. Handl.,
28, no. 5 1896, 212 p.
69. Vallance T. G. Concerning spilites. Proc. Linn. Soc.
N. S. W., 85, 1960, 8-52.
70. Vallance T. G. Spilites again: some consequences of
the degradation of basalts. Proc. Linn. Soc. N. S. W., 94,
1969, 8-51.
71. van Bemmelen R.W. The geology of Indonesia
(The Hague: Mart. Nijhoff, 1949), 997 p.
72. Vogt T. The geology of part of the H~londa-Horg
district, a type area in the Trondheim region. Norsk geol.
Tidsskr., 25, 1945, 449-528.
73. Vokes F. M. Regional metamorphism of the
Palaeozoic geosynclinal sulphide ore deposits of Norway.
Trans. Instn Min. Metall. (Sect. B: Appl. earth sci.), 77,
1968, B53-9.
74. Yokes F. M. A review of the metamorphism of
sulphide deposits. Earth ScL Rev., 5, 1969, 99-143.
75. Vokes F. M. and Gale G. H. Metallogeny relatable
to global tectonics in southern Scandinavia. In Metallogeny
and plate tectonics Strong D. F. ed. Spec. Pap. geoL Ass.
Can. no. 14, 1976, 413-41.
76. Wells A. K. The nomenclature of the spilitic suite.
Part I: the keratophyric rocks. Geol. Mag., 59, 1922.
346-54.
77. Wells A. K. The nomenclature of the spilitic suite.
Part I1: the problem of the spilites. Geol. Mag., 60, 1923,
62-74.
78. Wells A. K. and Bishop A. C. An appinitic facies
associated with certain granites in Jersey, Channel Islands.
Q. JI geoL Soc., 111, 1955, 143-66.
79. Wilson J.T. Did the Atlantic close and then re-
open? Nature, Lond., 211, 1966, 676-81.
80. Wilson M. R. The geological setting of the Sulitjelma
ore bodies, Central Norwegian Caledonides. Econ. GeoL,
68, 1973, 307-16.
81. Zachrisson E. Caledonian geology of northern
Ja'mtland - southern Vasterbotten: K61i stratigraphy and
main tectonic outlines. Sver. geoL Unders., Ser. C,
no. 644, Arsb. 64, no. 1 1969, 33 p.
82. Zachrisson E. The structural setting of the
Stekenjokk ore bodies, Central Swedish Caledonides.
Econ. GeoL, 66, 1971, 641-52.
83. Zachrisson E. The westerly extension of Seve rocks
within the Seve-K61i Nappe Complex inthe Scandinavian
Caledonides. GeoL F~r. Stockh. F&rh., 95, 1973, 243-51.
151
The Planes--San Antonio pyritic
deposit of Rio Tinto, Spain: its
nature, environment and genesis*
D. Williams
Emeritus Professor of Mining Geology, University
of L o n d o n ; f o r m e r l y Geological Consultant to
Companla Espafiola Minas de Rio T i n t o , Ltda, and
Rio T i n t o Patino, S.A.
R. L. Stanton
Department of Geology, University of New
England, Armidale, New South Wales, Australia
F. Rambaud
Uni6n Explosivos Rio T i n t o , Madrid, Spain

553.661.21 (468.12)

Summary
The San Antonio stratiform cupriferous deposit of Rio
Tinto, discovered in the early 1960s, is now being exploited
as an eastward extension of the bedded pyritite sheet in the
old Planes mine, which is there underlain by a stockwork
feeder pipe.
The Planes pipe consists of sulphur-rich-copper-poor
sulphide ore formed by the almost complete replacement of
a stockwork conduit which traverses Lower Carboniferous
felsic pyroclastic rocks. Overlying the pipe, and extending
far beyond its margin, is the genetically related Planes-San
Antonio cupriferous pyritite, a 'sedimentary-volcanogenic'
deposit precipitated as a chemical sediment derived from
sea-floor hot springs associated with the Lower Carbon-
iferous volcanic island arc of southwest Iberia. Most of the
pyritite bodies of Rio Tinto were deposited directly above or
close to their underlying feeder stockworks. The Planes--
San Antonio pyritite displays well-preserved stratification,
slump structures and interdigitated tuff bands, and was
mainly precipitated comparatively distant from the
nourishing Planes stockwork.
Across-layer base-metal zoning in the Planes-San
Antonio sheet is usually characterized by an increase in
copper, and to a lesser extent in zinc and lead, towards the
base of the deposit. Much of this enrichment may have
been due to late hypogene processes involving the ascent of
juvenile-meteoric solutions which leached base metals from
the underlying volcanic pile and deposited them as
chalcopyrite, sphalerite and galena by partial replacement
of the lower part of the pre-existing pyrite-rich stratiform
sulphide sheet.
A notable feature of the pyritite is the abundance of
colloform and framboidal pyrite. The mean isotopic
composition of the sulphide sulphur of the Planes stock-
work ((~34S ~' + 100/00) is distinctly heavier than that of the
stratiform Planes-San Antonio pyrite (~34 S ~ + 2.7O/0o).
The cupriferous pyritic mass of Planes-San Antonio
exhibits some of the features characteristic of pyritic deposits
formed during the initial stages of arc development ('Cyprus'
or 'ophiolite' type ores). It was actually formed, however, at
a late stage of arc evolution, during a period of relative
quiescence within the waning episodes of submarine
explosive felsic volcanism, and it displays environmental
traits analogous to those of the Japanese 'Kuroko' ores.

*Paper published in Trans. Instn Min. Metal/. (SectB: Appl. earth

sci.), 84, 1975, B73--82.

152
Kuroko deposits: their geology,
geochemistry and origin
Takeo Sato
G e o l o g i c a l S u r v e y o f Japan, T o k y o , Japan ( l a t e l y ,
Department of Geology, Memorial University of
N e w f o u n d l a n d , St J o h n ' s , N e w f o u n d l a n d , Canada)
553.067: 553.277: 553.43/.44(52)
Synopsis
The Kuroko deposits of Japan are predominantly massive,
stratiform sulphide ores with associated underlying stock-
work ores of Miocene age in the Green Tuff volcanic se-
quence~ They formed during a very limited period around
13 m.y. ago in association with the late-stage felsic rocks of
the mafic to re/sic volcanic cycle that began in the late
Oligocene or early Miocene. They occur in narrow zones
parallel to island arc~ The Benioff zone is at about 150-km
depth below the Kuroko zones, which were areas of high
heat flow during the Miocene.
The major constituents of the deposits are pyrite, chal-
copyrite, sphalerite, galena, barite and quartz. They are
zoned vertically. Gypsum and/or anhydrita commonly
occur as separate bodies beside the sulphide-barite bodies.
I t is thought that the ores were precipitated on the Miocene
sea-floor from saline solutions in a weakly acidic, reducing
environment at about 200-250~ as a result o f falling tem-
perature~ The zoning of minerals may have resulted from
changes in density of the ore solution-sea water mixtures
as precipitation proceeded.
Studies of light stable isotopes suggest that the mineral-
izing solutions were predominantly coeval sea water. Lead
isotope data, together with base-metal contents of the host
rocks, however, indicate that leaching by circulating sea
water cannot have been the dominant mechanism. So, mag-
matic water probab/y played an important role in the
mineralization. The following mode/is therefore proposed.
About 20--25 m.y. ago, as the Pacific plate was being sub-
ducted, a large amount of basic to intermediate magma was
generated and differentiated in the lower crustal region.
Magmatic fluids produced at the last stage of differentiation,
about 13 m.y. ago, rose to the sea-floor and, on mixing
with sea water, produced the Kuroko deposits. Subsidence
o f the sedimentary basins before the felsic volcanism could
have been the result of volume decrease during solidification
of underlying magma.
The Kuroko ore deposits of Japan are considered to be a
prototype of heavy metal deposits formed by submarine
hot-spring activity. The intimate association of the ore de-
posits with volcanic rocks has also attracted attention.
The aim here is to review present knowledge of the
geology and geochemistry of the Kuroko deposits and to
propose a genetic model that relates submarine mineraliza-
tion to island arc volcanism. Important publications that
supplement the present paper are those by Lambert and
Sato 1 and Ishihara etal. 2
Geological setting
Kuroko and other base- and precious-metal deposits are
found in the so-called Green Tuff sequences, in which
Miocene submarine volcanics are the main constituent. The
Green Tuff rocks are distributed on the Japan Sea side of
the Japanese islands from Hokkaido to Kyushu, which is
known as the Green Tuff region. The area suffered denu-
dation almost throughout Mesozic and Paleogene times,
and then subsided suddenly in the early Miocene. Many
small sedimentary basins about 1 0 - 2 0 km across, bordered
by nearly vertical faults, were formed in zones of varying
degrees of subsidence parallel to island arc trends. This
tectonism was of horst-and-graben type under tensile
stress. 3
Subsidence was followed by violent submarine volcan-
ism, dominated by andesite in the early stages and dacite
and rhyolite in the later stages. This volcanism (the first
cycle) continued for about 10 m.y. from the earliest to the
middle Miocene. The Kuroko mineralization was associated ~
with the last stages, which coincided with the period of
maximum subsidence. Fig. 1 shows the association of the
ore deposits with 13 m.y. old rhyolitic rocks to be remark-
ably constant throughout the 1200 km from Western
Hokkaido to Sanin.
There followed the second-cycle volcanism, which was
much less intense than the first. 4 Again, there was a mafic
to felsic trend. Compressional stress was operative 3 and
the Green Tuff region was uplifted. Thus, the later-stage
(Plio-Pleistocene) products were mostly subaerial.
Limited distribution of the Kuroko deposits is obvious
not only in time but also in space (Fig. 2). All the major
deposits occur in a narrow zone about 50 km wide and
essentially parallel to the east Japan arc system. 7 The zone
is situated near the boundary of Kuno's e tholeiite and
high-alumina basalt provinces, but the volcanics associated
with the Kuroko deposits are calc-alkaline. A less import-
ant Kuroko zone is associated with the west Japan arc. The
present Benioff zones are about 150 km below the Kuroko
zones.
The characteristics of the geologic environment of the
Kuroko deposits are typified in the Hokuroku district,
which covers an area of about 500 km 2 and includes more
than 100 000 000 tons of ore 9 of the Kuroko and vein
type. In this district late Palaeozoic (Permian?) shale and
chert form the basement of the Green Tuff strata. Miocene
volcanism started with the eruption of andesite in a shallow
marine basin, which soon became filled. After some
erosion, the area subsided rapidly in the early middle Mio-
cene to form a volcano-sedimentary basin. During sub-
sidence, basalts and mudstone were deposited in the west-
ern part of the basin, followed by extensive dacite extru-
sion throughout, the thickness of dacite exceeding 600 m
in the central part of the basin. There followed sporadic
eruption of rhyolitic rock, mainly around the periphery of
the basin. These rhyolites, known as 'white rhyolite', form
characteristic domes. The Kuroko deposits are intimately
associated with them.
After the mineralization there was intermittent inter-
mediate to felsic volcanism with interspersed argillaceous
sedimentation. Basalt flows and pillow lavas locally overlie
the deposits. In the late Miocene the area was uplifted and
denuded. Small caldera-like depressions, filled with felsic
volcanic rocks and siltstone, were formed in the latest Mio-
cene. In these rocks low-grade disseminated lead--zinc
mineralization is known. 9
Horikoshi lo interpreted the Kuroko mineralization as a
volcanic event related to the eruption of the white rhyolite,
and concluded that ore fluids were exhaled during the
waning stages of the explosive rhyolitic volcanism. He
recognized nine lava domes in the Kosaka mining area, of
which four are associated with economic orebodies and two
with sub-economic bodies.
153
 Fig. 1 Stratigraphic positions of Kuroko deposits
Some Kuroko deposits are overlain by basaltic pillow
lavas, as in Ainai and the Fukasawa mines; but, generally,
the deposits are covered by mudstone, which suggests that
volcanism became dormant after the mineralization.
Main features of Kuroko deposits
The Kuroko orebodies are typically sited on the flanks of
the brecciated white rhyolite lava domes, but in some places
Fig. 2 Distribution of Kuroko deposits in Japanese
islands in relation to depths of Benioff zone 6 and volcanic
front
154
they are hosted by tuff breccia. In both cases a unit ore-
body (defined as one continuous geologic body) consists of
massive, stratiform ore and the underlying stockwork ore in
various proportions, the former representing syngenetic
mineralization on the sea-floor and the latter disseminated
and fissure-filling mineralization in the pre-existing rocks.
The major ore-forming minerals are pyrite, chalcopyrite,
sphalerite, galena, tetrahedrite-tennantite, barite and
quartz. Small amounts of auriferous and argentiferous
minerals are present, as well as many minor minerals. 13,14
The orebodies show a vertical zoning of mineral content.
The typical variation of mineral assemblages, in ascending
order, is pyrite-chalcopyrite--quartz (keiko or siliceous
ore) in the stockwork ore, pyrite-chalcopyrite (oko or
yellow ore) in the lower part of the massive ore and sphaler-
ite--galena-tetrahedrite-barite (kuroko or black ore) in
the upper part of the massive ore. Gold and silver are con-
centrated in the black ore zone. Barite shows highest con-
centrations in the uppermost part of the massive ore, where
it forms an almost monomineralic zone, up to several
metres thick, traceable continuously along the contact with
the hanging-wall rocks.
In mudstone or tuff overlying the orebody there often
occur thin beds of hematitic, sometimes manganiferous,
chert (known as tetsusekiei). Where they directly overlie
the orebody, the rocks are highly siliceous, but more
commonly they contain much mud or clay. Laterally, the
ferruginous chert zone rarely continues more than 1 km
from the orebody.
Gypsum-anhydrite masses occur in the orebodies. The
spatial relationship of the gypsum body to the sulphide
ores may be either (1) the gypsum body is beside the sul-
phide orebody, usually partly covered by the black ore,
(2) the gypsum body is between the stockwork and the
massive sulphide ores or (3) the gypsum body is the domi-
nant constituent ofi a unit orebody, sulphide minerals being
concentrated as small lenses in the upper parts.
Mineralizing fluids must have ascended through the
stockwork ore. Fissure-filling veinlets, sometimes up to
1 m wide and concentrated in the central part of the stock-
work ore, may represent ore-fluid conduits. At Kosaka
mine pipe-shaped network zones about 30 m across are
found 200 m below the top of the stockwork ore. 15 The
veinlets continue upwards in the massive ore to the contact
with the hanging-wall rock, where they are overlain by a
thin pyrite bed. In the massive ore the veinlet swarm is
surrounded by highly siliceous ore with textures that sug-
gest a primary gel origin of the silica. It is likely that this
Fig. 3 Schematic cross-section of unit Kuroko orebody
(relation of gypsum body to sulphide body is of type 1)
ore formed on the sea-floor above the feeder vent in a
manner similar to silica sinters in hot springs.
The schematic cross-section of a unit orebody shown in
Fig. 3 illustrates these features.
Physico-chemical environment of ore deposition
The following review of the physical and chemical environ-
ment of Kuroko deposition relates mainly to deposits in the
Hokuroku district.
Temperature
Homogenization temperatures of fluid inclusions in quartz
from the stockwork ore and the fissure-filling veinlets are
mostly in the range 200-270~ 1,16 Inclusions in barite
of the massive ore give temperatures that vary from 100 to
300~ even within a single grain, but generally somewhat
lower temperatures are indicated than those of the stock-
work ore. Stability relations of ore-forming minerals 17
and sulphur isotope partitioning between coexisting
minerals 18 indicate similar temperatures to those estimated
from fluid inclusion studies. Results for various different
orebodies also give similar temperatures. The major part of
the Kuroko deposit formed at temperatures of 200-250~
There appears to have been a temperature drop of 50~
or more from the stockwork ore to the top of the massive
ore.
Pressure
From palaeogeographic studies lo, 19 it is estimated that
the pressure during formation of deposits must have been
well below 100 atm. On the basis of the ecology of micro-
fossils, maximum water depths of 3 0 0 - 5 0 0 m, representing
pressures of 3 0 - 5 0 atm, are estimated at Shakanai mine. 2~
From consideration of the vapour pressure of water in the
N a C I - H 2 0 system it is thought probable that the ore fluid
boiled before it reached the sea-bottom. 21 The vapour
pressure of 5 wt % NaCI solution is 38.1 atm at 250~ 22
If boiling occurs below the sea-bottom, no significant ore
deposition will take place on the sea-floor. This may have
been the case where there are well-developed stockwork
ores but only traces of massive ore, as at Tsuchihata mine. 5
Fugacities of oxygen and sulphur
Estimates of these parameters have been attempted by use
of the stability relations of ore minerals, particularly those
which form solid solutions. Kajiwara and Honma 23 esti-
mated the oxygen fugacity during the formation of black
ore from lead content of barite coexisting with galena. On
the basis of the reaction
(PbS)Gn + 202 = (PbSO4)Brt
where (PbS)Gn and (PbSO4)Brt denote the components of
galena and angelsite in galena and barite, respectively, the
oxygen fugacity for the above reaction can be calculated
thermodynamically by analysis of the two minerals in co-
existence and by assuming values for the equilibrium tem-
perature and the activity coefficients. Kajiwara and Honma,
assuming galena to be pure PbS and a dilute solution model
for the anglesite component in barite, obtained 10 -37 atm
at 200~ and 10 -34 atm at 250~ for the oxygen fugacities
of black ore deposition at Furutobe mine.
Sulphur fugacities are estimated on the basis of sul-
phidation reactions among minerals. The following react-
ions are particularly important for the Kuroko deposits:
5CuFeS2 + $2 = Cu5 FeS4 + 4FeS2
(FeS)sp h + 1/2S 2 = FeS2
2(Ag)EI + 1/2S2 = Ag2S
where (FeS)sph and (Ag)Ei represent the pyrrhotite com-
ponent in sphalerite and the native silver component in
electrum, respectively. Because the FeS contents Of sphaler-
ite from the Kuroko deposits are mostly less than 4 mole % 24
and bornite is not common, the sulphur fugacities must
have been in a relatively narrow range (see Fig. 4). If one
Ean find an appropriate combination of minerals, both
temperature and sulphur fugacity are obtainable by cam-
-5
~,, Siliceous ore / /
..... Yellow ore ~ /
....... Lower black ore J t A
" - - - Upper blOCkore / ~7~o~!
-1C
o~
-15
-21
100 150 200 250 300
TernperQture t ~
Fig. 4 Progressive change of temperature and sulphur
fugacity of Sato's model Kuroko-forming solutions 27
155
 e-
I "0
~'o (3
m~L O

0
I.
r v
c-
O
n
O3

0
m.m 0.r !
O O O O r r 0 0 0
0
c-
.O
~

0% ,-
0 -9, r o.~.
I"~ ,,-' r O o d ~ oo
L_
g
r

l0 0 ~- 0 moo od
" r0

q q
O O r,,.

,.Q

eq
o o
s
=,--
q o
E
o
o868 UO e'J O
I.O ,~"
C0
~
03
~

o
.~

t-
o E
8og 8~
O -"- O r r
I.O r t,D Ole0
I..O I.D I..O e0

E --
o r
o 1,..,.
O
9r "N
c~ 0}

gggg O O3 ,--
,~- 03
LO
r r r
i Z Z Z O0O31.O O3
r ',="-

0 0
o o =8
0 0
0 0 0 '~" LO O0 CXl ,--
(~ 1.0 r O0
t"~ (N ,-- I
~o
0 0 0 0 O O O O O
oN gggg O
,~"
I.O
r
O
f,D
r
O3
O
r

gE 03
r
eO I..O ,~"
r '~"- ~"=
O
I.t')
03
I~
f,D
O3
'r"" f,.O

~') tO

s
r
~m
8
03
8~ 8g
r
i,:

E
o ~ -~
o8
_>
e~
= ~ o s
o o O O ~ .g~ Ng
~ c c
0 0 u
| .

.0. . . .0. . . 0 0
~ ~ ~.~
r
| :::) .=_ S
o o o o %
"~ "O "O "O --' _ 0 ~ E
O O O O ~: E
* '4--

156
bining two sulphidation reactions. Sato 17 observed pyrite,
sphalerite, argentite and electrum in coexistence in black
ore from Kosaka mine and estimated the sulphur fugacity
and the temperature for equilibrium deposition of the four
minerals as 10 "14 atm and 200~ respectively.
pH
The most common method in estimating the pH values of
the Kuroko-forming environment is to use stability relations
of alteration minerals, especially those which contain
potassium. Urabe 14 classified the alteration mineral assem-
blages of the Kuroko deposits into three types: (1) kaolinite
-pyrophyllite-diaspore, (2) sericite-chlorite and (3) seri-
cite--chlorite--calcite, and concluded that the pH becomes
more alkaline from (1) to (3). Most of the Kuroko deposits
in the Hokuroku district are classified as his sericite-
chlorite type. Not uncommonly, however, a small amount
of kaolinite and/or pyrophyllite is associated with
sericite. 2s,26 Thus, the pH of the ore fluid must have been
in the stability field of sericite and close to the boundary
with that of kaolinite and/or pyrophyllite. Although we
have no specific knowledge of the potassium content of the
ore fluid, by assuming a content typical of hydrothermal
solutions 19,27 we can limit the pH to almost neutral or
weakly acidic.
A completely different approach was attempted by
Kajiwara and Honma, 23 who combined the oxygen fugacity
estimate and the interpretation of the sulphur isotope ratios
in pyrite. They concluded that the pH during the formation
of bJack ore was 5.5 + 0.5 at 250~ which is close to the
neutral pH of this temperature.
Model ore-forming system
Our knowledge of the chemistry of the Kuroko-forming
fluid is scanty. Freezing temperatures of fluid inclusions
indicate that the NaCI concentrations vary from nearly
zero to about 2.5 molar, mostly falling in the range 0 . 4 -
1.5 M. Inclusions in sphalerite tend to have somewhat
higher equivalent concentrations than those in quartz, and
salinities decrease outwards from the centres of mineral-
ization. It should be noted, however, that freezing tem-
perature measurements were made on relatively large
crystals, which may be a product of recrystallization or of a
later stage than the main mineralization.
Theoretical physical and chemical models have been
made, and one of these is discussed. The model is based on
many assumptions, however, and can only give an
a~proximation.
Mode/solution
Sato 27 calculated the physico-chemical conditions of the
Kuroko formation by use of the thermodynamic data of
high-temperature aqueous solutions. The premises on
which his calculations are based were that (1) chemical
equilibrium is maintained in the individual partial systems,
except for the contaminated sea-water sulphate, which was
assumed to persist metastably in reducing environments;
(2) chloride complexes are the only important complexes
for transporting ore metals; (3) the initial total chloride
concentration of the ore fluid is 1 - 3 M, the K/Na atomic
ratio being less than 0.5; and (4) the initial copper con-
centration of the fluid is more than 1 ppm. The important
conclusions from these calculations are: (1) model ore
fluids that successfully reproduce the observed m ineral
assemblages have very limited chemical compositions;
(2) mineral zoning is explained by differential precipitation
from a polymetallic solution; (3) variations of physico-
chemical parameters from the bottom to the top of the
orebody are established, temperature changes being the
most important cause of mineral deposition; and (4) mixing
of the ore fluid w i t h sea water is necessary for barite and
anhydrite to be precipitated with sulphides from the same
solution. The chemical compositions of Sato's model solu-
tions and the calculated progressive changes of his model
ore-forming environments are shown in Table 1 and Fig. 4.
The calculated physico-chemical parameters show good
agreement with those estimated from fluid inclusions and
by other methods. This is noteworthy, since the model in-
cludes no assumptions based on measured parameters.
Compared with natural saline geothermal waters, the model
solutions have higher K/Na ratios and lower Ca concentra-
tions. Although no analytical data on the Kuroko-
forming fluid are available, there is a partial analysis of
fluid inclusions from veins of the Osarizawa mine, 34 which
are believed to have a genetic affinity with the Kuroko
deposits. The analysed inclusions contain 7730 ppm Na,
3400 ppm K and 830 ppm Ca. These concentrations are
similar to those in Sato's model.
The K/Na ratios of hydrothermal solutions equilibrated
with the surrounding rocks are known to increase with
temperature, 28,35 and Ca concentrations relative to K and
Na in natural hydrothermal solutions decrease with in-
creasing temperatures. 36 It is therefore possible that the
low Ca, high K/Na solutions represent equilibration with
rock at high temperature, lower-temperature equilibrium
not being established owing to the rapid ascent of solutions
from depth. This interpretation is consistent with the fact
that potassium was conspicuously added to and sodium and
calcium were leached from the stockwork ore and the
alteration zone directly underlying the Kuroko
deposits.37, 38
Behaviour of ore fluid in sea water and mechanism of
mineral zoning
The density difference between ore solution and sea water
is probably the most important factor in controlling the
behaviour of the ore fluid in sea water and, consequently,
the nature of ore deposition. When the ascending solution
has a high salinity and a relatively low temperature, it is
heavier than normal sea water and will collect in depres-
sions on the sea-floor. 4~ Massive ores will consequently
form in depressions and therefore may not necessarily be
underlain by feeder vents or stockwork ore. When a h o t
solution with a salinity not much greater than that of sea
water, such as circulating heated sea water, is discharged
into normal sea water, it will float upwards, mixing with
the surrounding sea water. The precipitates from such a
solution will cover a vast area without any control by sea-
bottom topography. The ferromanganoan sediments on the
East Pacific Rise "41 are one of the best examples of such
deposits.
The Kuroko-forming solution would have had a con-
siderably higher salinity than that of normal sea water, and
the temperature when it reached the sea-bottom would have
been between 200 and 250~ It is calculated that such a
solution would be less dense than normal sea water; but,
when cooled by mixing and/or thermal diffusion, it would
become heavier than sea water. The solution will thus con-
vect above the feeder vent. The resultant deposit will
directly overlie the feeder vent and will show a limited dis-
tribution around it, as is the case with most of the Kuroko
deposits. In the above model, since the site of ore deposi-
tion is essentially controlled by the site of discharge of the
157

JI Main stage
~o-~ ~ ~
Last- stage
j( 1) ~ .-~, ~ (3)
i ~~ % . ' . ~ o l .O~o o
9 Later stage ~" ~ Final product
( 2) .r .o'~/%
.. ~ (~)
~ Pyrite--
cholcopyrite [-~ 5pha[erite--
galena-barite ~
Fig. 5 Model for mechanism of mineral zoning in
Kuroko deposits
ascending solution, ores may be deposited on gravitation-
ally unstable sites. The resultant unconsolidated ores will
often be reworked by submarine slides or turbidity currents.
The common occurrence of graded fragmental ore is one of
the characteristic features of the Kuroko deposits.
By combining density considerations with the model
Kuroko-forming system it is possible to construct a
dynamic model for the mechanism of mineral zoning
(Fig. 5). Pyrite and chalcopyrite, which calculations indi-
cate to be precipitated at higher temperatures than galena
and sphalerite, 27 will be deposited within the feeder vent
and on the sea-floor in the proximity of the feeder vent.
Galena and sphalerite, as well as barite, on the other hand,
will be precipitated in the peripheral parts of the convect-
ing brine mass, where the solution is cooled and sulphate
concentration is increased by mixing with sea water.
If both the temperature and the rate of discharge of ore
fluids are constant throughout the course of mineralization,
lateral mineral zoning of deposits will result. It would be
reasonable, however, to assume that the temperature and/
or the rate of discharge were decreasing with time. In
these circumstances the galena and sphalerite of the later
stages would be deposited closer to the feeder vent than
that of earlier stages: thus, earlier pyrite and chalcopyrite
would be overlain, resulting in vertical zoning.
If the temperature of the discharged solution is suffi-
ciently low (about 160~ for a 3 M NaCI solution), its
density exceeds that of sea water. Such a solution will flow
downhill, depositing minerals at a lower level than the
feeder vent. This mechanism could explain the somewhat
asymmetrical shape of the Kuroko orebodies (see Fig. 3).
Deposition of calcium cannot easily be fitted into the
above model because of the variable position of the gyp-
sum-anhydrite body relative to the sulphide-barite ore.
When Sato's model solution is used, i~ can be demonstrated
that anhydrite will be precipitated at temperatures lower
than 150~ after considerable mixing with sea water. 27
Origin o f ore fluid and metallogenesis
Isotope studies of the Kuroko deposits are pertinent to any
discussion of the origin of the ore fluids. Important points
are: (1) oxygen and hydrogen isotopes of the fluid inclu-
sions from the yellow ore zone indicate that most of the
water is of sea-water origin, with up to 25% contamination
by magmatic and/or meteoric water,~2 (2) sulphur isotope
ratios in sulphide and sulphate minerals are most readily
explained by assuming that the overall sulphur isotope ratio
158
of the ore fluid was the same as that of Miocene sea water
(the sulphide sulphur is 2-7%o heavier than magmatic
sulphur);18, 43 (3) lead isotope ratios of the Kuroko and
the contemporaneous vein-type deposits are almost identical
with the rock leads of Neogene volcanics, showing a uni-
lateral variation across the island arc,~4,4s and (4) the
initial 87Sr/86Sr ratios of the basaltic to rhyolitic Neogene
volcanics of northeast Japan fall in a narrow range between
0.7035 and 0.7048, except for those which occur on the
Japan Sea side of eastern Honshu, where the ratios are
higher. 46
Circulating sea-water versus rnagmatic origin
Anhydrite The circulating sea-water model is proposed mainly by the
students of stable isotopes, 18, 42, 43 who argue th at the
Miocene sea water is the major source of the water and
the sulphur in the Kuroko fluid, the metals having been
leached out from the rocks through which the fluid pene-
trated.
Although this model can adequately explain the ratios of
light stable isotopes in the Kuroko ores, several difficulties
are encountered. First, there is no explanation why the
Kuroko deposits are associated only with the middle
Miocene 'white rhyolite', which is apparently m ore d iffer-
entiated than the other Miocene volcanics. Second, the
lead isotopic data rule out the possibility that the circulat-
ing sea water leached out metals from the basement rocks.
If the convection cell of the circulating sea water is res-
tricted within the Miocene volcanics, it must have been of
a very small scale, because the thickness of the pre-
mineralization Miocene volcanics is usually less than 1 km -
sometimes only 100 to 200 m, as in the Motoyama deposit
of the Kosaka mine, which is one of the largest Kuroko
deposits, with more than 10 000 000 tons of ore. In such
small convection cells very effective leaching must have
taken place to give rise to economic concentrations of ore
metals. It has been demonstrated, however, that ore
metals are not significantly depleted in the Miocene vol-
canics in the mining areas.1
Third, underground development and many drillings
have revealed that the average white rhyolite dome has a
mushroom-like shape about 300 m in diameter and about
200 m in thickness. Although the suggestion that the White
rhyolite domes provided heat energy to drive the con-
vection systems appears to be feasible, a simple calculation
shows that a single lava dome can heat up only a very small
amount of water. The weight of a mushroom-shaped
felsic intrusive body 300 m across and 200 m thick is of
the order of 107 tons or 1013 g. If this body intruded as a
liquid at 750~ the maximum amount of heat that can be
obtained during crystallization and cooling to 250~ is
(80 + 0.3 x 500) x 1013 or 2.3 x 1015 cal, which can heat
up 1013 of water from 20 to 250~ Even if this water
dissolves as high as 100 ppm Cu, the convection system
derived by a single lava dome can transport only 109 g or
103 tons of copper. This is smaller by two orders of mag-
nitude than the copper in an average Kuroko deposit
associated with a single lava dome.
Finally, the circulating sea-water model must provide
reasonable mechanisms to increase salinity and to reduce
sulphate ions to sulphide ions. It should be noted in this
context that no evaporite as a source of salts nor sufficient
organic matter as a reducing agent is present in the mining
areas.
The classical hypothesis that metal-rich fluids are
separated during the final stage of fractionation and solidi-
fication of silicate magmas usually implies that all metals,
water, sulphur and salts in ore fluids are of magmatic
origin.
The lead isotopic data and the intimate association of
the deposits with the white rhyolites are well explained by
this model. The relatively high salinities of the inclusion
fluids are also compatible With the magmatic origin of the
waters on the basis of water-magma interaction experi-
ments 47 and the salinity data of fluid inclusions for which
magmatic derivation is inferred from isotopic studies. 48
The difficulties with the magmatic model are that only a
small fraction (up to 25%) of magmatic water is permissible
to explain the hydrogen and oxygen isotopic data, and that
the sulphur in the sulphide minerals is heavier by 2 - 7 % o
than magmatic sulphur.
Proposed model
Although more studies are undoubtedly needed to clarify
the origin of the Kuroko fluid, it is suggested that mag-
matic waters were of prime importance in forming Kuroko
deposits and the associated vein-type deposits. The isotopic
data for hydrogen, oxygen and sulphur may be explained
by isotopic exchange with sea water during mineralization
and/or d iagenesis.
It cannot be mere coincidence that the Kuroko mineral-
ization is contemporaneous with felsic volcanism during the
last stage of the intense volcanism that commenced at the
beginning of the Miocene. It is suggested that the change
of the movement of the Pacific plate at 2 0 - 2 5 m.y. ago
triggered the generation along the Benioff zone of a large
amount of basic to intermediate magma, which subsequent-
ly rose through the upper mantle. Most of the magma must
have stopped rising and undergone differentiation when it
encountered the light continental crust and density equi-
librium was attained. It is probable that crystal cumulates
increased the crustal thickness under the Green Tuff Region, 5o
where the crust is 5 - 1 0 km thicke~ than in volcanically in-
active regions. 49
From the computer simulation of the palaeo-heat flow
patterns of the Japanese islands by Hasebe and co-
workers 51 it was inferred that there was a high heat-flow
zone above the Benioff zone at about 150-km depth during
the Miocene. The high heat-flow zone may have played an
important role in the formation of Kuroko deposits, be-
cause the steep geothermal gradient may have enabled ore
Fig. 6 Speculative east-west cross-section of northeast
Japan in middle Miocene. Heat-flow pattern is after Hasebe
and co-workers 51 and isostatic anomaly is hypothetical
fluids to reach the sea'bottom at high temperatures. This
may explain why the Kuroko deposits occur in narrow
zones, whereas the vein-type deposits are more widely
scatte red. 7
The ascent of magma under the Green Tuff region must
have resulted in a negative isostatic anomaly and high heat
flow in the region. Fig. 6 illustrates this model. Volume
decrease of solidifying magma or diapiring magma itself 52
could have been a cause of the subsidence of basins prior to
volcanism. It is also possible that the general uplift of the
Green Tuff region in the late Miocene 49 was the result of
isostatic adjustment.
In the final stages of differentiation and solidification a
volatile-rich felsic magma ascended to the surface along the
faults bordering the basins and formed the white rhyolite
and the associated pyroclastics. The ore fluids, which
separated from the residual liquid in the magma chamber,
must have ascended through the same conduits as the white
rhyolite magma, and reached the sea-floor through the frac-
tures formed by the explosive eruption of the white rhyo-
lite. It is not denied, however, that the ultimate origin of
the water in the ore fluids may have been either the coeval
sea water or water derived from subducted oceanic sedi-
ments (see Fig. 6).
In conclusion, it should be re-emphasized that all the
Kuroko deposits in Japan formed in a short time-range of,
at most, 1 000 000 years - or possibly only 0.3 m.y., 53
and that the timing of the Kuroko formation correspond~
to the last stage of a large-scale magmatic cycle that con-
tinued for 10 m.y. Although the extensive hydrothermal
alteration of the Green Tuff rocks 54 suggests that there
was vigorous hot-spring activity during the whole course of
the Green Tuff volcanism, only the middle Miocene activity
gave rise to economic concentrations of base metals. It
seems doubtful whether we can find an analogue of the
Kuroko fluid in present-day hot-spring waters, which are
mostly of 100% meteoric and/or sea-water origin, ss
According to Sugimura etal., 4 volcanic activity in the
Japanese islands began intensely again in the Quaternary,
basic to intermediate rocks becoming predominant. Will
this 'rejuvenated' magmatism produce Kuroko deposits
10 m.y. hence?
Acknowledgment
This paper was prepared while the author was with
Memorial University of Newfoundland. Financial support
from the University and the Institute of Science and Tech-
nology of Japan is gratefully acknowledged. Stimulating
discussions with D. F. Strong, R. W. Henley, J. G. Thurlow,
Akira Sasaki, Ei Horikoshi, Yoshimichi Kajiwara and many
others have contributed much to the contents of this paper,
for which, however, the author is solely responsible. Thanks
are extended to Mrs Noriko Shiina for typing the
manuscript.
The paper is published with the permission of the
Director, Geological Survey of Japan.
References
1. Lambert I. B. and Sato T. The Kuroko and associated
ore deposits of Japan: a review of their features and
metallogenesis. Econ. Geol., 69, 1974, 1215-36.
2. Ishihara S. eta/. eds. Geology of Kuroko deposits
(Tokyo: The Society of Mining Geologists of Japan, 1974),
435p. (Min. Geol., Tokyo, Spec. Issue no.6)
3. Fujji K. Tectonics of the Green Tuff region, northern
Honshu, Japan. Reference 2 , 2 5 1 - 6 0 .
159
4. Sugimura A. et aL Quantitative distribution of late
Cenozic volcanic materials in Japan. Bull. volcan., 26, 1963,
125-40.
5. Sato T. Distribution and geological setting of the
Kuroko deposits. Reference 2, 1-9.
6. Sugimura A. Zonal arrangement of some geophysical
and petrological features in Japan and its environs. J. Univ.
Tokyo Fac. Sci., Sect. I I, 12, 1960, 133-53.
7. Sato T. Metallogenesis of Neogene base metal and
A u - A g deposits in Japan with special reference to Kuroko
deposits. Min. Geol., Tokyo, Spec. Issue no.4, 1972,
185-204. (Japanese text)
8. Kuno H. Lateral variation of basalt magma type across
continental margins and island arcs. Bull. volcan., 29,
1966, 195-222.
9. Sato T. Tanimura S. and Otagaki T. Geology and ore
deposits of the Hokuroko district, Akita Prefecture.
Reference 2, 11-8.
10. Horikoshi E. Volcanic activity related to the formation
of the Kuroko-type deposits in the Kosaka district, Japan.
Mineral. Deposita, 4, 1969, 321-45.
11. Tatsumi T. and Clark L . A . Chemical composition of
acid volcanic rocks genetically related to formation of the
Kuroko deposits. Geol. Soc. Japan J., 78, 1972, 191-201.
12. Ishihara S. Magmatism of the Green Tuff tectonic
belt, Northeast Japan. Reference 2 , 2 3 5 - 4 9 .
13. Shimazaki Y. Ore minerals of the Kuroko-type
deposits. Reference 2, 311-22.
14. Urabe T. Mineralogical aspects of the Kuroko deposits
in Japan and their implications. Mineral. Deposita, 9, 1974,
309-24.
15. Kawabe S. and Masubuchi K. The stockwork siliceous
ore deposits at the Kosaka mine, Akita Prefecture. Min.
Geol., Tokyo, 25, 1975, 275-83. (Japanese text)
16. Tokunaga M. and Honma H. Fluid inclusions in the
minerals from some Kuroko deposits. Reference 2,
385-8.
17. Sato T. An estimate of formative environments of the
Kuroko deposits from mineral paragenesis. Koshogaku-noto
Tokyo, no.7, 1969, 1-6. (Japanese text)
18. Kajiwara Y. Sulfur isotope study of the Kuroko ores
of the Shakanai No. 1 deposits, Akita Prefecture, Japan.
Geochem. J., Tokyo, 4, 1971, 157-81.
19. Sato T. Ore deposit and mechanism of its formation
of Uchinotai western orebody, Kosaka mine, Akita Pre-
fecture. Min. Geol., Tokyo, 18, 1968, 241-56. (Japanese
text)
20. Otagaki T. Development of prospecting at the
Shakanai mine. Min. Geol., Tokyo, 16, 1966, 237-48.
(Japanese text)
21. Ridge J. D. Volcanic exhalations and ore deposition
in the vicinity of the sea floor. Mineral. Deposita, 8, 1973,
332-48.
22. Haas J. L. Jr. The effect of salinity on the maximum
thermal gradient of a hydrothermal system at hydrostatic
pressure. Econ. Geol., 66, 1971,940-6.
23. Kajiwara Y. and Honma H. Lead content of barite
coexisting with galena- an indicator of oxygen fugacity
during ore formation. Min. Geol., Tokyo, 22, 1972,
457-65.
160
24. Urabe T. Iron content of sphalerite coexisting with
pyrite from some Kuroko deposits. Reference 2 , 3 7 7 - 8 4 .
25. Shirozu H. Clay minerals in altered wall rocks of the
Kuroko-type deposits. Reference 2 , 3 0 3 - 1 0 .
26. Matsukuma T. etal. Rare minerals from Kuroko ores
of the Uwamuki deposits of the Kosaka mine, Akita Pre-
fecture, Japan. Reference 2, 349-61.
27. Sato T. A chloride complex model for Kuroko
mineralization. Geochem. J., Tokyo, 7, 1973, 245-70.
28. White D. E. Environments of generation of some base-
metal ore deposits. Econ. Geol., 63, 1968, 301-35.
29. Lebedev L. M. Contemporary deposition of native
lead from hot Cheleken thermal brines. Dokl. Akad. Sci.
SSSR, 175, 1967, 173-6. (Russian text)
30. Brewer P. G. and Spencer D.W. A note on the chemi-
cal composition of the Red Sea brines. Reference 40,
174-9.
31. Ellis A. J. The chemistry of some explored geo-
thermal systems. In Geochemistry of hydrothermal ore
deposits Barnes H. L. ed. (New York, etc.: Holt, Rinehart
and Winston, 1967), 465-514.
32. White D. E. Mercury and base-metal deposits with
associated thermal and mineral waters. In Geochemistry
ofhydrothermal ore deposits Barnes H. L. ed. (New York,
etc: Holt, Rinehart and Winston, 1967), 579-631.
33. Krauskopf K. B. Introduction to geochemistry (New
York, etc.: McGraw-Hill, 1967), 721p.
34. Enjoji M. Studies on fluid inclusions as the media of
the ore formation. Sci. Rep., Tokyo Kyoiku Daigaku,
Sect. C, 11, 1972, 79-126.
35. Ellis A. J. and Mahon W. A. J. Natural hydrothermal
systems and experimental hot water/rock interactions
(Part II). Geochim. cosrnochim. Acta, 31, 1967, 519-38.
36. Fournier R. O. and Truesdell A. H. An empirical
N a - K - C a geothermometer for natural waters. Geochim.
cosmochim. Acta, 37, 1973, 1255-75.
37. lijima A. Clay and zeolitic alteration zones surrounding
Kuroko deposits in the Hokuroku district, northern Akita,
as submarine hydrothermal-diagenetic alteration products.
Reference 2 , 2 6 7 - 8 9 .
38. Date J. and Tanimura S. Dacite and rhyolite associated
with the Kuroko mineralization. Reference 2, 261-5.
39. Sato T. Behaviours of ore forming solutions in sea-
water. Min. Geol., Tokyo, 22, 1972, 31-42.
40. Degens E. T. and Ross D. A. eds. Hot brines and
recent heavy metal deposits in the Red Sea (Berl in, etc.:
Springer-Verlag, 1969), 600p.
41. Bostrom K. and Peterson M. N.A. Precipitates from
hydrothermal exhalations on the East Pacific Rise. Econ.
Geol., 61, 1966, 1258-65.
42. Ohmoto H. and Rye R.O. Hydrogen and oxygen
isotopic compositions of fluid inclusions in the Kuroko
deposits, Japan. Econ. Geol., 69, 1974, 947-53.
43. Kajiwara Y. and Date J. Sulfur isotope study of
Kuroko-type and Kieslager-type strata-bound massive sul-
fide deposits in Japan. Geochem. J., Tokyo, 5, 1971,
133-50.
44. Sato K. and Sasaki A. Lead isotopes of the black ore
("Kuroko") deposits from Japan. Econ. Geol., 68, 1973,
547-52.
45. Sato K. Unilateral isotopic variation of Miocene ore
lead from Japan. Econ. Geol., 70, 1975, 800-5.
46. Shuto K. On the source materials of acid volcanic
rocks in the Cenozoic circum-Pacific belt - Considered
from their strontium isotope ratios. Monogr. Ass. Geol.
Collab. Japan, 18, 1974, 91-105. (Japanese text)
47. Holland H. D. Granites, solutions and base metal
deposits. Econ. Geol., 67, 1972, 281-301.
48. Sato T. Origins and classification of exhalative-
sedimentary mineral deposits. A paper submitted to the
25th Int. Geol. Congr., Sydney, 1976.
49. Sugimura A. and Uyeda S. Island arcs; Japan and its
environs (Amsterdam, etc.: Elsevier, 1973), 247p.
(Developments in geotechnics 3)
50. Kuno H. Origin of andesite and its bearing on the
island arc structure. Bull. volcan., 32, 1968, 141-76.
51. Hasebe K. Fujii N. and Uyeda S. Thermal processes
under island arcs. Tectonophysics, 10, 1970, 335-55.
52. Koide H. and Bhattacharji S. Formation of fractures
around magmatic intrusions and their role on ore local-
ization. Econ. Geol., 70, 1975, 781-99.
53. Ueno H. Duration of the Kuroko mineralization
episode. Nature, Lond., 253, 1975, 428-9.
54. Utada M. Zeolitic zoning of the Neogene pyroclastic
rocks in Japan. Sci. Pap. Tokyo Univ. Coll. Gen. Educ.,
21, 1971, 189-221.
55. Sakai H. and Matsubaya O. Isotopic geochemistry of
the thermal waters of Japan and its bearing on the Kuroko
ore solutions. Econ. Geol., 69, 1974, 974-91.

161
Stable isotope studies on water component associated with the later and (?) lower-
temperature sericitization. There is no isotopic evidence
Bougainville and in Matupi to suggest that sea waters or connate waters formed a
recognizable component of the hydrothermal system at
Harbour, New Britain, Papua any stage. Calculations with possible 8180 depletions of
New Guinea 0.5-1.0~ in whole rocks from the Panguna region
indicate that a meteoric leaching modet for the genesis of
the Panguna copper deposit is theoretically possible, but
J. H. Ford unlikely because of the small water/rock ratios.
D e p a r t m e n t o f Geology and Mineralogy, University The area around Matupi Harbour includes the dormant
volcanoes Rabalanakaia and Tavurvur, set in loosely con-
o f Queensland, Brisbane, Australia
solidated, Quaternary ash deposits. Tavurvur last erupted
D . C. Green
in 1941 and 1942 and has since reverted to a quiet
D e p a r t m e n t of Geology and Mineralogy, University solfataric condition. D/H and 180/160 ratios were
o f Queensland, Brisbane, Australia determined on fumarole condensates, hot springs and
J. R. Hulston meteoric waters. The data are grouped into distinct areas
Institute of Nuclear Sciences, D.S.I.R., L o w e r H u t t , close to the meteoric water line and demonstrate that the
New Zealand thermal springs away from the shoreline are of meteoric
I. H. Crick origin and that sea water only enters the springs at the
Bureau o f Mineral Resources, Geology and shoreline. If allowance is made for the vapour-liquid
Geophysics, Canberra, Australia separation at the temperature of collection (100~ the
S. M. F. Sheppard fumarole results from Tavurvur may be interpreted as
Scottish Universities Research and Reactor Centre, steam separating from a water phase at ~- 100~ The
East Kilbride, Glasgow, Scotland calculated water-phase isotopic compositions suggest that
evaporation and possibly a small amount of mixing with
sea water are involved.
539.155.2:553.216.5:553.43(955) These conclusions are in conflict with those drawn from
anion ratio and trace metal contents inferred by previous
authors to be consistent with an hypothesis of modified
Summary sea-water origin. We contend that the chemistry is a
Two case histories are presented in which light stable reflection of the later history of these acid, mineralized,
isotope ratios (D/H and 180/160) place important con- geothermal waters and does not necessarily give a correct
straints on the origins of ore-forming fluids in the Panguna picture of their ultimate origin. The markedly enhanced
Fe, Mn and Zn values of the Matupi springs are purely a
porphyry copper deposit (upper Pliocene) and on the
function of the leaching potential of geothermal fluids at
nature of the volcanic exhalative environment at Matupi
elevated temperatures and of the chemistry of the porous
Harbour, near Rabaul.
rocks through which they pass.
At Panguna the overlap in ~D values between local
meteoric water and inferred primary magmatic waters poses
problems in the interpretation of secondary biotite and
sericite isotope data. The range of hydrogen and oxygen
isotope ratios of the secondary biotites associated with
the mineralized areas is essentially identical to those from
other magmatic hydrothermal deposits, including porphyry
copper deposits. The sericites from alteration haloes
adjacent to mineralized quartz veins have ~180 values of
5.4 to 8.6O/0o and ~D values of -61 to -77. A combina-
tion of thermodynamic and fluid inclusion techniques
indicates that ~- 500~ is a realistic average temperature for
much of the mineralization-quartz veining-biotitization
process. The isotopic composition of the fluids causing this
mineralization and alteration was ~D ~- - 2 0 to -45O/oo and
~180 ~- 6 to 7o/00 . If a meteoric water component
existed at this stage of alteration, meteoric water/rock
ratios were extremely small (less than 0.05 at %).
Assuming that sericitization took place at temperatures
down to 300~ as suggested by the fluid inclusion data,
the fluids responsible for this process had a composition of
~D = - 3 0 to -50o/00 and (~180 of ~- 3.50/00 .
These fluids probably would be largely of meteoric
hydrothermal origin, produced by 180 shifting of ground-
waters during the formation of the propylitic zone and with
a water/rock ratio of about 0.01 at % for the entire
propylitic zone. Quartz ~ 180 values from veins associated
with the different alteration assemblages increase away
from intrusive centres, consistent with an outwards
temperature decline, a decrease in 8180 content of
associated water and a progressive increase in the meteoric

162

Volcanogenic mineralization at
Avoca, Co. Wicklow, Ireland, and
its regional implications
J. W. Platt
A v o c a Mines, Ltd., Avoca, A r k l o w , Co. W i c k l o w ,
Ireland
553.067:553.43(418.4)
A description is given to the geological setting o f the
stratiform pyritic orebodies at Avoca in/reland and a
volcanogenic origin is proposed on the basis of comparison
with stratiform orebodies located in similar volcanic host
rocks in other parts of the Caledonian orogen and else-
where. The Ordovician volcanic sequence in southeast
Ireland is interpreted as a fossil island arc complex, which
was formed above a contemporary margin o f southeasterly
subduction, which locally defined the boundary of the
evolving Caledonian orogenic belt in southeast/re/and.
The A voca ore deposits are located within a sequence of
ca/c-alkaline lavas, pyroclastics and sediments at a level
marked by the predominance of felsic eruptives. The strati-
graphy of the host rocks is described in some detail and
related to the general geology of the Avoca belt. Massive
pyritic ores and disseminated sulphide ores o f "stringer type"
can "both be recognized at Avoca. The massive ore shows a
compositional gradation suggestive of a primary stratigraphw
a Cu-rich pyritic zone is overlain by a zone of P b - Z n - C u
pyritic ore.
The interpretation of the genesis of the orebody is com-
plicated by its deformed state, which is typical o f other
volcanogenic orebodies within the various Caledonian
provinces. A genetic mode/can be proposed, however, by
using the descriptions of the undeformed "Kuroko" deposits
in Japan as a standard for comparison.
The individual orebodies of Avoca type contain
5 000 0 0 0 - 6 000 000 tons of ore, and these are associated
with extensive zones of low-grade disseminated ore in acid
pyroclastic host rocks. Sequences o f manganiferous and
dolomitized sediments are stratigraphically closely related
to the ores.
The principal geological features of the A voca orebodies
having been defined, it is deduced that there is a strong
possibility that similar orebodies lie hidden in the volcanic
rocks of the southeast Irish Caledonides and that an
aggressive exploration programme could lead to new
discoveries.
-Relative-iy little attention has been given in recent years to
the economic potential of the lower Palaeozoic volcanic
province of southeast Ireland with regard to metalliferous
deposits - (a) because there has been a long and unspectacu-
lar history of production of low-grade copper ores from the
Avoca Mines and their genetic significance has been obscur-
ed by the assumption of their epigenetic origin in all but
the most recent studies is and (b) because the spectacularly
rich and accessible strata-bound lead-zinc orebodies in the
lower Carboniferous rocks of the Central Irish province
have diverted attention from the exploration potential of
the Caledonian volcanic terrains.
The Avoca Mines are located in south Co. Wicklow on
the Avoca River about six miles north of its estuary at the
Port of Arklow (Fig. 1). The mine area is in open rolling
farmland, 500-600 ft above sea-level, the only significant
relief being provided by the Avoca Valley. The mines have a
recorded history of orderly mining since 1752. Mining was
for both copper and sulphur (as pyrite). Within the Avoca
volcanic pile, and Spread over five miles of strike both north-
east and southwest of the mine area, magnetite mining has
been important, and one occurrence of placer gold is known.
The Avoca orebodies were discovered at their exposure
on the banks of the Avoca River, and mining has since
spread for 2 88miles along strike. No significant discovery
of new ore has been made at Avoca for 120 years, although
during the past two decades the anatomy of the system of
orebodies has been followed and exploited well beyond the
limits known in the nineteenth century. Historically, the
exploitation of the zinc- and lead-rich facies of the
Avoca ores has been avoided and the traditional selective
mining by eye of the chalcopyrite-rich facies has meant
that the geological and economic significance of the more
polymetallic ores has not been fully realized. It is therefore
timely that an account of the geological relation-
ships of the Avoca orebodies should be presented on the
basis of modern concepts of ore genesis in volcanic environ-
ments.
Palaeotectonic setting of Avoca region in the Caledonian
province of southeast Ireland
A general description of the geology of the Caledonides of
southeast Ireland has been given by Stillman and co-
workers 15 and Brindley and co-workers? and their work
provides a regional basis for the interpretation of the palaeo-
tectonic setting of the Avoca sulphide orebodies and their
volcanic host rocks (Fig. 4). Three major northeast-south-
west-striking belts of deformed volcanic rocks of Caradocian
age can be recognized, each with its own distinctive character:
(1) the Portrane-Lambay-Donabate-Hill of Allen-Kildare
inlier, which contains a sequence of andesites, basic tufts,
slide and flow breccias, acid porphyries and crystal tufts; (2)
the belt that stretches through Donard, Baltinglass, Rathdangan
and Aughavannagh in southeast Leinster to Borris in Co.
Carlow contains a sequence of rhyolites and ignimbrites,
porphyries, basic volcanics and carbonate-rich sediments,
which have locally been strongly affected by the aureole of
contact metamorphism surrounding the Leinster granite; and
(3) the extensive belt from Tramore in Co. Waterford to
Arklow in Co. Wicklow, in which acid eruptives predomin-
ate, probably forms the core to a synclinal structure - the
Campile syncline. (3a) The volcanic tract to the north of
Arklow is included within this belt and probably occupies a
subsidiary synclinal structure. It is in the centre of this
tract that the Avoca Mines are located.
Because of the presence of the Avoca deposits it is this
latter belt that is of immediate economic interest. Basic to
intermediate lavas mark the onset of volcanic activity in the
Waterford-Arklow-Avoca belt, but they are relatively
limited in occurrence. The chief expression of volcanism in
the Caradocian-Llandeilo sequence takes the form of
massive sheets of felsitic eruptives with intercalated sedi-
ments and pyroclastics. It is considered that deposition,
for the most part, took place under shallow marine
conditions.
Stillman and co-workers 15 classed the volcanic rocks of
this belt as calc-alkaline, probably having formed in an
island arc setting. This interpretation is compatible with
recent models for the plate-tectonic evolution of the
Appalachian-Caledonian orogen (Fig. 2) as described by
Dewey, 4 Fitton and Hughes, 6 Mitchell and Reading, TM
163
Fig. 1 Avoca Mines, Ltd.: location and general geological environment
Strong 16 and Phillips and co-workers. 13 It is generally
accepted that a southeast-inclined zone of subduction along
Fig. 2 Outline framework of Appalachian-Caledonian
orogen. After Dewey 4 and Mitchell and Reading 12
164
the southeast margin of the Proto Atlantic Ocean 19 caused
the island arc volcanism which, during Ordovician time,
extended as a complex semi-continuous belt along the active
southeast margin of the Caledonian-Appalachian orogen
through what is now the Scandinavian peninsula and down
through the English Lake District to southeast Ireland,
Newfoundland and New Brunswick and into the Appala-
chian belt of eastern North America. It is within this belt
that lower Palaeozoic volcanism finds its strongest expres-
sion, and these are also the Caledonian terrains of principal
economic interest.
The tectonic model recently suggested by Phillips and
co-workers 13 for the evolution of the British Caledonides
has direct relevance to the interpretation of the environ-
ment in which the Avoca deposits were formed. Three non-
parallel zones of active subduction during the period of
time from the Ordovician to the Lower Devonian are recog-
nized. One of these was the southeast-inclined margin,
which gave rise to the island arc volcanism in the belt that
stretches from the English Lake District through Co.
Waterford in Ireland. They also proposed that an active
destructive margin lay along the line of the northwest
Caledonian front, subduction taking place slowly to the
ESE, and that there was also a plate boundary trending
EN E-WSW to the south of the Scottish Southern Uplands.
Ordovician volcanism associated with subduction in the
southeast began during Llanvirnian time and intensified dur-
ing the Caradocian with the formation of volcanic sequen-
ces of intermediate type with associated tholeiitic lavas.
The strongest expression of acid volcanism did not occur
until Llandeilian time, and in the British Caledonides the
best development is seen in the Waterford-Arklow area
of southeast Ireland. Shield volcanism of island arc tholeiite
type was first initiated on a basement of Precambrian and
early Palaeozoic continental crust and was followed by
calc-al kal ine eru ptives of i ntermed iate com position, culmin-
ating with a sequence of rhyolitic sheets.
Combining the proposals for the metallogenic 16 and
the petrological/geochemical 6 evolution of the calc-alkaline
arcs formed at the southeast margin of the Proto Atlantic
VotcanogenicCu
massive sutphide deposit
. ,.-VoLcanicistand arc
Oceanic surfclce .~ I ~
100 mites ,
Scare very approximate
Fig. 3 Schematic metaliogenic-petrological model for oceanic subduction zone in plate-
tectonic scheme referred to Avoca environment. After Strong 16 and Fitton and Hughes 6
Ocean, it is suggested that the Avoca region can be fitted
to a generalized model of the type shown in Fig. 3. The
setting of the Avoca deposits can be seen in the context of
the general pattern of metallogenesis associated with the
destructive margins of the Caledonian orogen as outlined by
Strong. 16 Mississippi Valley lead-zinc deposits lie on the
ocean basin side of the palaeo arc, followed in the direction
of the arc by copper pyrite deposits of 'Cyprus type' con-
tained in the remnant oceanic lithosphere. Volcanogenic
copper-lead-zinc deposits with minor A g - A u follow in
the calc-alkaline rocks of the arc proper with the possibility
of S n - W - M o deposits towards the ancient continental
margin. Direct application of this model to the Irish Cale-
donides is suggested and may be of value in guiding explor-
ation for new deposits. The increasing alkalinity of the
eruptives towards the continental side of the palaeo arc is
complementary to the metallogenic polarity and is the
basis for the Fitton and Hughes model 6 in which the se-
quence ranging from tholeiitic basalts in north Leinster to
alkaline 'back arc' volcanism in Wales is recognized.
The sequence of sediments, calc-alkaline lavas and
pyroclastics (and the related diorite to granodiorite in-
trusives) which are host to the Avoca deposits can thus
reasonably be interpreted as having formed as an ensialic
island arc during Ordovician time. It therefore follows
that valid geological comparisons can be made between the
Avoca deposits and other stratiform base-metal deposits
in calc-alkaline volcanic sequences in palaeo island arcs in
other provinces of the Caledonides - as, for instance, in
Sweden, 9 New Brunswick 11 or Newfoundland. 17
Similar models for a volcanic-hydrothermai origin can
also be proposed and comparisons with the undeformed
stratiform polymetallic sulphide ores of the well-known
Miocene Kuroko type in Japan can be made. 1~ In the
following sections a detailed description of the volcano-
stratigraphy and mineralogical variations within and
immediately surrounding the orebodies is given to justify
a volcanogenic model of Kuroko type for the Avoca ores.
Structural relationships and stratigraphy of Avoca
volcanogenic belt
The geological interpretation of the Avoca belt is complic-
ated by the fact that, in common with other provinces in
the orthotectonic regime of the Caledonian oro'gen, the
rocks and their contained ore deposits have suffered multiple
episodes of deformation, and have also been subjected to
low-grade regional metamorphism of greenschist facies. The
tectonic fabric of the rocks has frequently obscured or
obliterated the primary mineralogy and also the diagnostic
~ . ; , " ~ / -
Limestone(cor_ol.)
ree_f ,b,Zn ~ ~ ~ /
<,.
o~ c'~
O^o,.~.
~?~
~. c.,.'\
"o,> %,,
volcanic or pyroclastic textures.
The precise structure of the Avoca belt is not well under-
stood, but Gardiner 7 suggested that the belt lies in a broad-
ly synclinorial structure - the 'Wicklow syncline'. The
current working interpretation of the belt is as a tightly
folded recumbent syncline refolded into an antiformal
structure, the axis of which plunges gently to the northeast.
The outcrop of the folded structure thus displays a horse-
shoe-like form, the limbs trending northeast-southwest and
with closure to the northeast.
Charlesworth 3 estimated the thickness of the volcanic
succession at Avoca to be 4 5 0 0 - 8 0 0 0 ft. Slaty sedimentary
sequences both under- and overlie the volcanics, and some
limestones occur in the footwall of the succession. The vol-
canics are dominantly of acid calc-alkaline facies and the
rhyolites at the base of the sequence are notably associated
with magnetite. The pile consists of rhyolite and rhyolite
breccias overlain by a variety of silicified, sericitized and
chloritized pyroclastic rocks. The pyroclastics are schistose,
but a variety of types are recognized, including crystal and
'quartz-eye' tufts and porphyries, lapilli and lithic tuffs,
coarse agglomerates and flow and slump breccias. Within
the pyroclastic sequence, and occurring most frequently as a
lateral stratigraphic equivalent to ore horizons, are beds of
non-volcanic mudstones and carbonaceous sediments.
Heavily dolomitized sericitic and carbonaceous sediments,
in which manganiferous carbonate is common, belong with-
in this category of lateral facies equivalents to ore and may
be comparable to the 'oolitic dolomite' described in associa-
tion with the Rosebery orebodies, Tasmania. This has been
interpreted as a back-arc basinal facies. 1 Parallels are also
suggested with the Buchans polymetallic sulphidedeposit
of the central mobile belt of Newfoundland, 17 in which
orebodies of a similar type to those found at Avoca have
formed in sediment-filled depressions in the palaeotopo-
graphy of the volcanic pile.
The Avoca massive sulphide orebodies occur stratigra-
phically above the rhyolite-rhyolite breccia sheets, spatially
165
closely associated with the rhyolites and also with dolomitiz- Avoca ore geology - general considerations
ed sediments, intertonguing with these and sometimes being The Avoca Mines are separated by the north-south Avoca
enclosed by them. Chloritization is a common feature of River Valley into two geographical sections known as East
the pyroclastic pile in proximity to the ore. and West Avoca. Modern mining has been almost entirely
centred on West Avoca, and it is from this sector that most
Avoca volcanogenic ore types of the current ideas concerning genesis have been developed
and will be tested. The results of work in West Avoca have
Genera/ direct implications for regional exploration in southeast
Ireland.
Sangster 14 defined two ore types as typifying volcanogenic
ore deposits, distinguished primarily by their relative sulphide
content and their relationship to stratigraphy: (a) massive
sulphide ore, markedly conformable with its surrounding
rocks, and (b) 'stringer' or disseminated ore, which crosscuts
stratigraphy. Both ore types are known and mined at
Avoca, and they compare with the Keiko and Kuroko/Oko
ore types, respectively, from the typical Kuroko sequence. 1~
Massive sulphide ore in volcanogenic deposits of Caledonian
affinity is compositionally zoned and bears a closer relation-
ship to the Kuroko model than the Archaean volcanogenic
ore deposits of which Sangster was writing specifically.

PJlassive sulph ides

Massive sulphide ore at Avoca occurs in two varieties. (1) A
chalcopyrite-pyrite ore with minor galena and sphalerite,
which has a banded, sedimentary appearance: the host rock
is a fine-grained chloritized and possibly slightly carbonaceous
tuff. Layers of sulphide are separated, often widely, by layers
of altered host rock and sulphide content may go as low as
15-20% of the whole-rock total. (To avoid discrimination
at Avoca we require only the banded sulphide characteristics
of the copper-pyrite zone to be recognized for diagnostic
purposes.) (2) A dense, fine-grained pyritic l e a d - z i n c -
copper ore with associated silver-bearing sulphosalts. This
variety commonly contains in excess of 90% total sulphide,
lead and zinc values being at their best towards the hanging-
wall. Sedimentary textures are difficult to recognize in the
hand specimen.
The massive sulphides are associated with lateral and
hanging-wall carbonate mineralization - in particular, nodu-
lar or oolitic dolomite, which relates closely to massive lead-
zinc-copper ore. Both varieties of massive sulphide ore are
considered to occur as integral parts of single composite
massive sulphide orebody or volcanogenic ore systems.
The massive sulphide orebodies have sharp contacts with
country rock, particularly in the hanging-wall. The massive
lead-zinc-copper zone is stratigraphically above its
copper-pyrite counterpart. Maximum ore thickness is
about 70 ft, the largest dimensions of the orebody being
parallel to bedding planes.
Stringer ore
The massive sulphides overlie the disseminated or 'stringer'
type ore. At Avoca total sulphid~ content of stringer ore
is rarely greater than 20%. Stringer olebodies are localized
within the main Avoca chloritized pyroclastic pile close to
or in contact with associated massive sulphide ores, and
are distinctively composed of ramifying and braided
veinlets of quartz with sulphide that cut across stratigraphy
for distances in excess of hundreds of feet.
The upward change from stringer ore into massive
sulphide is abrupt - in contrast to the diffuse, gradational
downward and lateral contacts of the pipelike stringer ore-
body. In its underformed state the stringer orebody is
cylindrical or funnel-shaped, the massive sulphide orebody
capping the widest part of the funnel and fingering laterally
into a carbonate facies.
Fig. 4 Caledonian geology of southeast Ireland. After
Stillman and co-workers 15 and Gardiner 7
One Avoca orebody - the Cronebane orebody - has
been mined recently as an open-pit in East Avoca, and its
dissection has shown it to have a similar anatomy to that
of West Avoca. Both Cronebane and West Avoca are distinc-
tive and separate volcanogenic ore systems. These can both
be considered as type examples of Caledonian volcanogenic
sulphide mineralization.
Features common to these deposits, and with direct
application at Avoca, are: 8 (1) they have a carbonate-rich
gangue, which is often dolomitic and manganiferous,
associated with shaly tuffs and carbonaceous shales; (2) they
are found in an environment of felsic volcanic rocks of calc-
alkaline type, with rare basalts; (3) the explosive volcanism
that generated the volcanic pile took place from domical
centres and may have been locally subaerial; (4) the volcanic
pile is associated with prevalent clastic sedimentation; and
(5) the massive sulphide orebodies may occur in clastic
sediments near volcanic rocks rather than within the volcanic
pile proper: the orebodies have a compositional zoning of
stratigraphy with Pb/Zn decreasing and Cu increasing
downwards.
Orebodies of this style tend to occur in provinces or
clusters, and at least three units, including West Avoca and
Cronebane, are present in the area of the Avoca Mines. The
discovery of significant P b - Z n - C u ore at West Avoca will
have important consequences for the mineral exploration of
southeast Ireland.

166
 Models for volcanogenic sulphide ore deposits need to
be applied with caution in any specific instance. Any two
specific volcanogenic sulphide orebodies from widely separ-
ate areas or, indeed, from the same area may have as many
dissimilar features as they have features ih common. In cate-
gorizing Avoca within the broad 'Kuroko' framework the
The principal point that emerges is that Avoca's tradition-
al label of 'copper mine' is an accident of history, and the
clearly volcanogenic origin of the Avoca ores suggests, by
comparison with polymetallic deposits of similar type, that
investigation of the hitherto ignored potential for the
exploitation of Pb-Zn and Ag will prove rewarding.

Avoca volcanogenic ore model

The volcanogenic ore model for Avoca is presented in Fig. 5.
A stockwork, feeder or stringer pipe crosscutting acid vol-
canic and volcaniclastic rocks is capped by a zoned volcano-
genic massive sulphide orebody. The massive sulphide
consists of an underlying Cu-pyrite zone and an overlying
-
v v v ,v \ O,\*
o.oy
9 /v" v" v V P b - Z n - C u zone. There are sharp contacts between zones.
v v v

9\ / V
9

V V
The massive sulphide body tongues laterally into clastic
V V 9 e\e o/e carbonate-rich basinal rocks, and is itself capped by clastic
Hanging-wall pyroclastics ~ Copper-pyrite zone
(massive sulphide) shale and fine-grained volcanic sediments. The stringer zone
I---7 Hanging-walt tuff (dolomitized) ~ Stringer zone is silicified and chloritized and may be surrounded by a
Footwall pyroclastics
sericitized envelope. The basal rhyolite sheet is brecciated
I
F~ DolomitJc sediments, tuffs ~ (sericitized-)
Graphitic shale ~--I Footwall pyrociastics o 20Oft, with sulphide infilling (generally barren pyrite).
D Lead-zinc zone Approximate Subsequent deformation and strike attenuation after
(massive sulphide) scale
regional metamorphism and the imposition of schistosity
Fig. 5 Schematic volcanogenic ore model for Avoca produce present-day configurations. These can be described
mineralization specifically with reference to the West Avoca and Cronebane
orebodies and their local setting in the overall Avoca mine
model that emerges is skeletal only and will be given flesh area.
as the subtleties of the Avoca geological environment be-
come better understood. To define a model for Avoca is, Avoca mine stratigraphy
nonetheless, valid in that guidelines are thereby determined Stratigraphy has been evaluated in broad terms in East and
for study and development. West Avoca to provide the basis for ore genetic modelling

Rockstown
\ ,
Meet
- ~
West Avoca East Avoca 2 ~

I-2-1 Sericitized pyroctastics

Chtoritized pyroclastics
B~ Cu/Pb--Zn zone (massive sulphide)
i-5-1 Crystal tuff/quartz-feldspar porphyry
J
lID Dolomitized sediments (carbonaceous) Pb--Zn Ass.
Diorite
--'~ Fault 0 1//+ 1t2 3/4 1mile
I I I
Fig. 6 Geological plan of East and West Avoca

167
and to determine a mine sequence for exploration purposes but generally massive and aphanitic. A cherty, potash-
in southeast Ireland (Fig. 6). The rock units are listed below. enriched variety has been recognized, with a magnetic content
that may typify the mine rhyolite series 5 and indicate
Maximum
volcanic vent filling. These magnetic rhyolites give a clear
Rock type thickness, ft geophysical anomaly that reflects the broad synclinal
(7) Diorite intrusive structure of the Avoca belt.
(6) Crystal tuffs 2000+ Where rhyolite is seen in juxtaposition with ore
(Cronebane open-pit) it exhibits brecciated textures with
(5) Sedimentary sequence 300 sulphide infilling and associated hematitic areas, which may
(4) Massive sulphide 70 be fumarolic in origin. The massive irhyolite sheets are over-
lain by a sequence of tuffs, which are variously sericitized
(3) Chloritized pyroclastics 1200
and chloritized. It appears at present that stratiform se-
(mineralized tuff)
quences of chloritized tuffs are strike-continuous - i t is
(2) Sericitized pyroclastics 400-500 not certain that the same applies to sericitized pyroclastics,
(rhyolitic tuffs) which may, in effect, mark only the location of alteration
(1) Massive rhyolite 1600-2000 hal oes around chloritized pipel ike mineral ized 'stringer'
orebodies.
The sequence is stratiform and overturned, distorted by The stringer orebodies stand out in the overall mass of
isoclinal folds and strike attenuation. Regional dip is chloritized tuff by virtue of their enhanced grade. Ore dis-

Fig. 7 Cross-section (500W) West Avoca (looking southwest)

55 ~ SE and schistosity and bedding may be coincident. tribution within the chloritized tuff can be compared to
The rhyolites are sheet-like units, occasionally porphyritic, 'streaky bacon'. No areas are completely barren of sulphide.

168
There would appear to have been a widespread permeation
of sulphide within the chloritized tuffs at Avoca, which
form a very large-scale low-grade orebody.
It is useful to summarize the critical geological features
that could be associated with an orebody of the Avoca type
considered as an exploration target in southeast Ireland.
The massive sulphide orebodies always occur in contact
with or in close proximity to stringer ore zones, and have
an overlying sedimentary sequence, which may enclose the
lead-zinc-copper zone of the composite massive sulphide
orebody. The sedimentary sequence is a unique and
variegated horizon of carbonaceous and sericitic shales and
tuffs, pyritic mudstones and beds of nodular or oolitic
dolomite. Large-scale dolomitic nodules may be due to
boundinage. The entire sequence is manganiferous, with
prolific crosscutting veinlets and permeations of white and
pink carbonate and a dolomitic pervasion of variable inten-
sity. The dolomite, which is probably diagenetic, may occur
as isolated spots or nodules in the rock mass, or at conglo-
merations of nodules, which dominate the composition of
the rock. There is a talc association - possibly hydrother-
mal and related to the massive sulphide occurrence.
The dolomitic sediments are themselves mineralized with
discrete grains of galena and sphalerite, sometimes centred
in dolomite nodules or surrounded by radial haloes of talc.
Metal values are of ore grade over selected sections, but it
is not known yet whether or not ther~ is any strike
continuity.
Fig. 8 Cross-section (31200E), Cronebane open-pit (looking southwest)
In the immediate hanging-wall of the mine geological
sequence the crystal turfs form a wide belt, commonly
interfingering with sediments and tuffaceous mudstones,
and also frequently sericitized with orientated quartz and
feldspar crystals in the groundmass. A variety of this rock
has been termed 'bird's-eye porphyry'.
Volcanogenic sulphides at Avoca
West A voca
The detailed tectonic/structural geology at West Avoca is
not well known, but isoclinal folding and strike attenuation
can adequately explain the deformation of the model
orebody to its existing configuration (Fig. 7).
The West Avoca sulphide orebody has been isoclinally
folded into its allied stringer zone, and has dissociated into
C u - p y r i t e and lead-zinc-copper components. Wedges
of tuff and dolomitized sediments have been thrust into
the copper pyrite zone and the P b - Z n - C u zone has 'slid'
off to the southwest along a graphitized shear and into the
dolomitized sediments. Both C u - p y r i t e and P b - Z n - C u
zones display the same fold style.
The concept is of a C u - p y r i t e orebody, probably
continuous over 4000 ft of folded strike, with fold nose
thickened barren pyritic areas. There is a sub-parallel
P b - Z n - C u orebody, which may itself be dissociated by
thrusting and which is separate from the C u - p y r i t e zone.
Cronebane
By contrast, in the Cronebane orebody, East Avoca, the
separate zones of the massive sulphide orebody are in close
association (Fig. 8). There is isoclinal folding and the
trough of the fold is occupied by the dolomitic-sedimen-
tary sequence. The C u - p y r i t e zone, as at West Avoca,
abuts on both footwall and hanging-wall against the
chloritic tuff or 'stringer' zone. The P b - Z n - C u zone inter-
fingers with the sediments on the south limb of the Crone-
bane isoclinal fold.
Brecciated rhyolite at the east end of Cronebane has
been thrust across the orebody and is intimately associated
with sulphide mineralization. This probably represents
domical rhyolite and it can only be a matter of time before
the vent is recognized.
Comparison with Kuroko model
In model terms it is interesting to compare Avoca with the
classic Kuroko model lo (see Table 1).
Conclusion
The identification of Avoca as a field of volcanogenic
stratiform sulphide orebodies, the formation of which was
169
Table 1 References
1. Braithwaite R. L. The geology and origin of the
Recognized Rosebery ore deposit, Tasmania. Econ. Geol., 69, 1974,
Kuroko unit Mineralogy at A voca 1086-101.
2. Brindley J. C. McArdle P. and Schiener E. J. The
Hanging-wall Volcanic and sedi- Yes
Lower Palaeozoic volcanic rocks of south-east Leinster,
mentary formation
Ireland. Sci. Proc. R. Dublin Soc., 4, no. 30, 1973,431-8.
Ferruginous Hematite, quartz and Yes, sulphide
3. Charlesworth J. K. Historical geology of/re/and
quartz zone minor pyrite (iron facies
(Edinburgh, London: Oliver and Boyd, 1963), p. 79.
formation)
4. Dewey J. F. Evolution of the Appalachian/
Barite zone Massive barite ore No
Caledonian orogen. Nature, Lond., 222, 1969, 124-9.
Kuroko zone ZnS/PbS/barite Yes, P b - Z n - C u
5. Downes K. The geology and geochemistry of the
zone
Ordovician volcanics of County Waterford and south west
Oko zone Cupriferous pyrite Yes, C u - p y zone County Wexford. Ph.D. thesis, University of Dublin, 1974.
Sekkoko zone Anhydrite, gypsum, No, dolomitic 6. Fitton J. G. and Hughes D. J. Volcanism and plate
pyrite sequence may be tectonics in the British Ordovician. Earth Planet. Sci. Lett.,
an equivalent 8, 1970, 223-8.
Keiko zone Copper-bearing Yes, 'stringer' 7. Gardiner P. R. R. Regional fold structures in the
siliceous disseminated zone Lower Palaeozoics of south-east Ireland. Bull. geol. Surv.
stockwork ore /re/and no. 1, 1970, 47-51.
Footwall Silicified pyroclastics Yes 8. Hutchinson R.W. Volcanogenic sulphide deposits
with disseminated and their metallogenic significance. Econ. Geol., 68, 1973,
sulphide veins 1223-46.
9. Juve G. Ore mineralogy and ore types of the Steken-
jokk deposit, Central Scandinavian Caledonides, Sweden.
Sver. geol. Unders., Ser. C, no. 706, 68, no. 13, 1974,
closely related to submarine hydrothermal processes taking 162 p.
place at, or peripheral to, centres of eruption of acid calc-
10. Lambert I. B. and Sato T. The Kuroko and associat-
alkaline laves, holds implications for the development of
ed ore deposits of Japan: a review of their features and
an exploration programme in the remaining volcanic ter-
metallogenesis. Econ. Geol., 69, 1974, 1215-36.
rains in the Caledonides of southeast Ireland.
The target orebodies could be expected to contain 11. Lusk J. Base metal zoning in the Heath Steele B-1
5 000 0 0 0 - 6 000 000 tons of massive sulphide ore associat- orebody, New Brunswick, Canada. Econ. Geol., 64, 1969,
ed with extensive zones of low-grade stringer ore. The ores 509-18.
should be distinctly polymetallic and offer opportunities 12. Mitchell A. H. and Reading H. G. Evolution of
for the exploitation of lead and zinc, and possibly silver island arcs. J. Geol., 79, 1971,253-84.
and gold in small amounts, as well as copper. Stratigraphic-
ally, the orebodies of Avoca style should be intimately re- 13. Phillips W. E. A. Stillman C. J. and Murphy T. A
lated to sequences of manganiferous dolomitized sediments Caledonian plate tectonic model. J/geol. Soc., 132, 1976,
and should also lie in close relationship to bodies of rhyolite- 579-609.
containing disseminated sulphides or magnetite. In common 14. Sangster D. F. Precambrian volcanogenic massive
with orebodies of Kuroko type, it is also likely that individual sulphide deposits in Canada: a review. Pap. geol. Surv. Can.
orebodies might occur in clusters around an eruptive centre, 72-22, 1972,.44 p.
as they appear to do at Avoca.
15. Stillman C. J. Downes K. and Schiener E. J.
It is clear that modern concepts of ore genesis in calc- Caradocian volcanic activity in East and South-east Ireland.
alkaline volcanic terrains can be applied with advantage to
Sci. Proc. R. Dublin Soc., 5, no. 6 1974, 87-98.
the interpretation of the geological relationships of the
Avoca orebodies in a palaeo island arc of Ordovician age. 16. Strong D. F. Plate tectonic setting of Appalachian-
If the Avoca model is used to guide exploration in the other Caledonian mineral deposits as indicated by Newfoundland
areas of Ordovician volcanic rocks in the Caledonides of examples. Trans. Am. Inst. Min. Engrs, 256, 1974, 121-8.
southeast Ireland, it is reasonable to expect that significant 17. Thurlow J. G. Swanson E. A. and Strong D. F.
new discoveries will be made. Geology and lithogeochemistry of the Buchans polymetal-
lic sulphide deposits, Newfoundland. Econ. Geol., 70,
Acknowledgment 1975, 130--44.
The concepts discussed in this paper have been developed 18. Wheatley C. J.W. Economic geology of the Avoca
from continuing dialogue between the author and numerous mineralized belt S. E. Ireland and Parys Mountain,
colleagues active in the mining industry. The author is Anglesey. Ph.D. thesis, University of London, 1971.
most grateful to the Management of Avoca Mines, Ltd., 19. Wilson J. T. Did the Atlantic close and then reopen?
for support and encouragement and for permission to Nature, Lond., 211, 1966, 676-81.
publish this paper. Special thanks due to L. Doyle for
drafting the plans and sections.
This paper is an Irish contribution to the International
Geological Correlation Programme (IGCP) project 'Correlation
of Caledonian stratabound sulphide deposits.'

170
Discussion
Dr. R. A. M. Wilson said that Gass had given an excellent
summary of the curent state of knowledge concerning the
ophiolite suite, not only in the context of global tectonics
but also in petrogenesis, thus providing a composite review of
earth processes at constructive plate margins leading ultimate-
ly to sulphide ore deposition. He was particularly indebted to
the author for bringing him up-to-date on the work that had
been done in Cyprus since he had left in 1958. His reference
to some ophiolites having a 'back-arc' rather than a mid-
oceanic association was particularly interesting. Supporting
evidence for that concept could be found in British Columbia,
where, immediately to the east of the Coast Range batholith,
a sheeted intrusive complex was exposed in the Thompson
River above the Frazer Canyon between the villages of
Lytton and Spences Bridge. The complex was about 8 miles
across strike and consisted of inter-sheeted epidosites, dia-
bases and gabbros, together with altered andesite or basalt
dykes such as occurred in Cyprus. The older greenschist
facies dykes, possibly of Cretaceous age, interdigitated with
gabbro, which graded downwards through interbanded
olivine gabbros into ultramafic rocks. The later andesite or
basalt dykes, which occurred at random intervals throughout
the sheeted complex, were probably of late Tertiary age;
all had a common northwesterly strike and, like the banding
in the gabbros, dipped regularly to the northeast at between
40~ 50 ~ The intrusive complex had the characteristics
of rocks believed to occur at constructive plate margins; as
far as he was aware, that British Columbia example was not
described in the literature but was mainly mapped as schist
and gneiss. 2 He had come across the sheeted dykes by
accident in 1968 when he was examining a group of mineral
claims near Lytton.
Returning to the several papers that referred to the
Cyprus ophiolites, and particularly to the concept of 'over-
printing' of Troodos gabbro and Diabase to greenschist
facies metamorphism by the passage of sea water, he was
reminded of earlier discussions on that subject. Later meta-
morphism or 'overprinting' was thought to account for the
obscure contact between Bishopp's Folded Diabase Series 1
and the 'Pack' (i.e. pack of cards) - a field term used for the
Basal Group dykes. It was in order to discover the nature of
that metamorphism that he had systematically sampled
adjacent groups of dykes in the Sheeted Intrusive Complex
and Granophyric Group; also sampled were the centres and
selvedges of individual dykes and sheets. The results of
petrographic work on those samples were summarized in
the Troodos Memoir. 3 It was found that the Granophyric
Group sheets differed substantially across strike from sheet
to sheet (although, locally, the sheeted aspect was lost
through remobilization), and also that adjacent Diabase
sheets differed in detail in their degree of metamorphism, as
did adjacent Basal Group dykes, some of which were fresh
basalts. Moreover, it was shown that the centres of some
dykes were more metamorphosed than their selvedges. All
those characteristics pointed to certain conclusions concern-
ing the nature of the metamorphism.
(1) In general terms the degree of greenschist facies
metamorphism increased towards the interior of the
Troodos Range, and particularly towards the plutonic rocks,
where the Granophyric Group represented, principally, a
zone of silicification under hydrous conditions round the
gabbro; it was analogous in a way to the zone of silicification
round a porphyry copper.
(2) The flow of hydrothermal solutions was controlled by
the dyke structures and travelled preferentially up their
coarser-grained centres.
(3) The more metamorphosed dykes of an immediately
adjacent pair or group dykes were older than the less altered
dykes.
(4) In the Diabase the youngest dykes were gabbro and
in the Basal Group the youngest dykes were fresh basalts.
(5) The younger dykes, although not everywhere concor-
dant, apparently had a random distribution.
Thus, for the Troodos complex as a whole, he would en-
visage an on-going process of magma generation, repeated
dyke intrusion and metamorphism. The propylitic alteration
associated with ore deposition, as Constantinou pointed out,
occurred after the greenschist facies metamorphism had
taken place.
He emphasized that he neither disputed the evidence
provided by isotope geochemistry to support a hypothesis
of sea-water circulation within the crust nol the widely
accepted concept of brines leaching out and redepositing
metallic sulphides to form orebodies. But he was curious to
know how that water-circulation model was reconciled with
the field and petrographic evidence cited above, which would
appear to require the entry of sea water into the magma
chamber to become available for metamorphism.
Evidence had been provided that a primary basalt magma
contained insufficient water to provide greenschist facies
metamorphism. In Troodos the Diaba.~ sheets in the interior
of the massif were generally non-vesiciJlar, but vesicles start-
ed to appear towards the Basal Group, which was generally
vesicular; the interpretation was (or had been) that the gas
phase was only released as the dykes reached higher levels
in the crust. Presumably, the vapour in those vesicles
originated in the parent magma, and the uralite gabbro was
formed on cooling by reaction with the water in that magma.
If the magma were dry, what was the origin of the vesicles?
M. F. Osmaston said that the plate-convergence-related
obduction hypothesis for the mode of emplacement of
ophiolite slices on continental margins had, in one form or
another, been rather uncritically accepted by many, includ-
ing, it appeared, Professor Gass. In his presentation he had
drawn little attention to the existence of serious difficulties
for the hypothesis in any of its forms. Those might be
briefly referred to as the height and trench problem, the
thermal aureole problem and the age problem. The purpose
of his remarks was to suggest that obduction was not a viable
proposition and that it should be discarded in favour of an
existing alternative, with far-reaching implications for
tectonic analysis.
The essence of the original obduction idea was that slices
of ocean floor were launched on to the continental shelf
in the course of floor subduction beneath that shelf. At
present-day subduction sites the floor approaching the trench
was always more than 2.5 km below the level of the shelf,
and the trench was wider than the widest known ophiolite
slices. Many ophiolite slices were too thin to have been
pushed significantly uphill, so one must postulate not only
a quite exceptional elevation (at least 4 km above normal) of
the basic oceanic crust relative to the level of the shelf but
also that the trench was full to shelf level with sediments.
The latter requirement probably ruled out active subduction.
As to the height difficulty, none of the many proposals
made, including ocean-ridge consumption, events related to
a hypothetical 'flip' of subduction direction and impending
continental collision, had been shown to provide a convinc-
171
 ing chance of a solution, as was attested by the continuing
flow of new suggestions.
The thermal aureole problem consisted in the fact that in
a number of well studied true ophiolite suites the basal
peridotite was in apparently intrusive contact with a slice
of supracrustal material (sometimes exhibiting a pre-existing
regional m e t a n ~ p h i c fabric) upon which it had impressed
an aureole of up to garnet amphibolite grade, and tectonic
transport had taken place with both slices welded together.
Those features were particularly clear in the ophiolites of
western Newfoundland, 4-7 but were known in some degree
from ophiolites and Alpine peridotites the world over.
Williams and Smyth 6 were not sure that those aureoles
could be produced by frictional heating during early sliding
at the contact, and suggested that the obduction hypothesis
might not rule out the mantle part of the slice being still
very hot when it reached pre-existing continental crust.
One of the alternatives proposed by Church, s however, was
that such ophiolites were generated at the earliest moment
in the opening of an ocean basin and were possibly caused
to slide on to the marginal continental crust at that time.
That interesting suggestion - a development from the ideas of
Hess and of Gass himself - deserved close attention.
The age problem had a dual aspect. On the one hand,
most ophiolites were substantially older than the regional
deformation and (where present) blueschist metamorphism
and granite intrusion, which accompanied their final dis-
placement to their present sites. On the other hand, where
erosion had not removed the evidence, it had been found
that the time between the constructive generation of an
ophiolite and its initial emplacement over continental
crust could (and might always) be very short indeed (e.g.
in Oman 8) a point entirely consistent with the form-
ation of hot-body contact aureoles. The long time-interval
before the onset of typical subduction-induced metamor-
phism and magmatism implied, however, that initial em-
placement might well have nothing to do with subduction.
Church's proposal 5 met all those problems, provided
that the embryo ocean ridge substantially overtopped the
level of the adjacent continental crust. In such a situation
thick radiolarian sediments might accumulate over the ridge
due to the heating of the water, which could kill the radio-
laria to give dense showers of siliceous debris. 9 It had been
shown lO that processes at constructive margins were
strongly dependent on the rate of heat input. Recent cal-
culations showed that a thick continental plate being split
at a high oceanic separation rate could generate an accreting
edifice standing well above an adjacent deep shelf before
that shelf had risen rauch in response to the heat input.
Subsequent rise and tilting of the shelf, with the accreted
area welded to it, could then initiate a series of d~collements,
first in the shelf cover and then in the accreted material, to
give the slice stacking order that was now so well established.
It appeared, therefore, that the obduction hypothesis
should now be abandoned. Recognition that the primary
emplacement of ophiolites on to continental edges marked
the initiation of rapid plate separative events would obviously
have enormous implications for tectonic analysis. Also, it
would explain at once (whereas misfit plate collision did
not) the extraordinarily tight grouping of ophiolite ages
along individual former plate margins. A fuller treatment of
that topic would be given elsewhere.
As a tailpiece he wondered if Gass would care to com-
ment on whether, in the foregoing context, the unconfor-
mity between the Upper and the Lower Pillow Lavas in the
Troodos ophiolite could mark the time of its primary em-
placement over a shelf.
172
Professor Gass, in reply, stated that recent work suggest-
ed that that was essentially a metamorphic unconformity
and did not represent a major time-interval; he doubted,
therefore, if the tectonic significance indicated by
Osmaston was justified.
Dr. Franco Maranzana, in commenting, first, on the
paper by Pearce and Gale, said that through the study of the
distribution of Ti, Zr, Nb, Y and rare-earth elements it was
possible to recognize the environment in which ophiolites
were formed. It also appeared that 'true' oceanic ophiolites
did not carry economic mineralization, more commonly
found in the other types of oceanic crusts (marginal basin,
back arc, island arc). Could they, therefore, use that
system as an exploration tool? There should, in fact, be a
direct economic implication related to the distinction of
the various environments, not so much to exclude a potential
prospecting ground but to obtain an order of preference:
when dealing with ophiolitic country, if an area were recog-
nized as a 'true' oceanic environment, a lower exploration
priority should be given to that area.
With regard to the paper by Sheppard, by use of the
ratios D/H and 160/180, the magmatic origin of mineraliz-
ation accompanied by the potassium silicate facies of hydro-
thermal alteration seemed to be demonstrated for porphyry
copper (and Mo) deposits. If potassic alteration were a pre-
requisite for that type of mineralization, was it possible to
conceive a porphyry copper without the above-mentioned
alteration facies? He was referring specifically to the Cerro
Colorado, Panama, porphyry copper deposit, where, after
some 50 000 m of drilling, the deepest hole reaching about
1500 m, no potassic alteration had reportedly been
encountered.
Turning to the paper by Hunt, who had referred to the
absence of potassic alteration at Cerro Colorado, Panama,
either that type of alteration was not recognized, probably
because subsequent feldspar-destructive facies of hydrother-
mal alteration had obliterated the previous pattern of the
potassium silicate phase, or the boreholes had not yet inter-
sected potassic alteration (if so, the size of the orebody
would be truly gigantic) or, as a last, and very improbable,
possibility, that type of alteration was not present.
He suggested that the rather dense distribution of Andean
porphyry coppers might be related to the prolonged active
subduction of the Pacific Plate. A certain lag was considered
necessary for the subduction mechanisms to set up the
magmatogene processes giving rise to hypoabyssal calc-
alkaline stocks and the related mineralizing hydrothermal
alteration, as a very substantial thickness of South American
continental plate needed to be perforated to reach a hypo-
abyssal level. Although no figure for such a lag could be
advanced, that could not simply extend from the inception
of subd~Jction to the known (lower to upper Tertiary)
porphyry coppers, as the existence in Chile of an almost
entirely eroded porphyry-type mineralization, located at
Los Loros (near/a Serena) moved to late Cretaceous, the
oldest known emplacement of porphyry copper in the
Andes (89 m.y). 11 Consequently, it was safe to assume that
the active subduction zone in the Andes triggered porphyry
copper emplacement from about 90 m.y. until the present
time, which accounted for the great abundance of that type
of mineralization.
He wondered, therefore, how important for the form-
ation of porphyry coppers was the duration of active sub-
duction, as well as its geometric parameters, direction of
spreading, strike of the trench, thickness of crust, etc. In
the case of oceanic-oceanic interaction, Coco Plate and
Caribbean Plate, for instance, calc-alkaline magmatism
could be generated at hypoabyssal levels without the need
for the perforation of great thicknesses of continental crust;
thus, much less time was likely to be required for that initial
phase. In southern Costa Rica and western Panama subduc-
tion might have been active from the lower Oligocene to the
Mid-Miocene. A porphyry copper deposit was located in
Panama and indications of that type of mineralization were
found in Costa Rica.
The corollary of all that could be summarized in two
questions. If, hypothetically, in the Andes, subduction were
active for only 10 m.y., would that amount of time have
sufficed for the formation of porphyry coppers? (Probably
not.) If the same time were allocated for another subduction
environment, as that of the Caribbean, would that be
sufficient for the emplacement of similar mineralizations?
(Probably yes.)
With regard to the paper by Gass, in many instances
ophiolitic material, derived from the subducted plate, was
emplaced by 'obduction'. In Costa Rica, the northwestern
part of the country was formed by ophiolite (Nicoya penin-
sula) with rncClangein an assembly strongly suggestive of
obduction. From west to east the main geological-geograph-
ical divisions were: Cocos Plate (subducted), Trench,
Nicoyan ophiolites, rest of Costa Rica, Caribbean Plate
(overthrusting). It would appear, therefore, that the ophio-
lites were derived from the Cocos Plate; on the other hand,
the magnetic pattern of northwestern Costa Rica (according
to aeromagnetic data), distinctly displaying an east-west
grain, was in accord with the magnetic anomalies of the
Caribbean, aligned in the same direction, leading to the
correlation between the Nicoyan ophiolites and the
Caribbean Plate. But, if that were so, another emplacement
rather than obduction should be sought. Changes in the
direction of subduction might have somehow contributed
to that problem, and such changes were possible when two
oceanic plates collided, since there were no fixed rules as to
which of the two was going to be subducted first. Another
explanation could come from a common spreading centre
for both plates prior to subduction. His comment was made
merely to enhance Gass's words about their limited know-
ledge on the emplacement mechanism of ophiolites.
As a concluding remark he would like to focus the
importance of geological investigation in Costa Rica, where
in very short distances they moved from an oceanic environ-
ment [ophiolite and massive sulphides) to a continental
setting (calc-alkaline magmatism and porphyry coppers),
which made the area interesting for mineral exploration as
well as academic work.
In a joint contribution Dr. R. A. Ixer and Dr. J. W.
Gaskarth said that the small copper-lead-zinc deposits of
the southern British Caledonides at Avoca and Parys Moun-
tain had recently been ~einterpreted as having a largely
volcano-sedimentary origin. That suggestion had found much
general acceptance, and a Kuroko-style model had been
favoured, though the detailed evidence for that had not
been well documented.
Platt had presented the evidence from Avoca for that
interpretation, leaving no doubt that, although metamor-
phism, deformation and erosion had modified the deposits,
they were volcanogenic Kuroko style. Recent work at
Parys Mountain by Thanasuthipitak 12 and by the speakers
(unpublished) had shown that deposit to have numerous
similarities to Avoca and the classical 'Kuroko' deposits.
Like Avoca, Parys Mountain had an island arc tectonic
setting, with felsic calc-al kaline volcanism; zoned, bedded
sulphides overlay a stringer zone at the top of the volcanic
pile. A crude depth zoning was distinguished in the ore
mineralogy, which was similar to that of the classical
'Kuroko'. Textural interpretations of the ore minerals
supported a low-temperature volcanogenic origin of that
type.
Exploration along the Ordovician-Silurian boundary in
that area might well lead to the discovery of further deposits.
The model predicted that deposits likely to be found would
be small, rich, lead-zinc bodies, with some larger low-grade
copper deposits. Deformation and erosion had made it
unlikely that very large tonnages of rich ore could be
expected.
Dr. W. J. Phillips, in a written contribution, said that
the formation of ore deposits as the result of large-scale
convective circulation of hydrotherma! solutions derived
from the hydrosphere had been demonstrated convincingly
by the stable isotope studies presented at the meeting,
supported by data on the rare-earth element distributions in
the Troodos sediments. Attention had been drawn to the
relatively narrow, faster-flowing, upward plumes of convec-
tion cells, and flow regimes in several different environments
had been discussed. Little consideration had, however, been
given to the hydraulic effects of convection of interstitial
fluids in porous rocks.
It was important to realize that water with a hydrostatic
pressure (p), might flow down a vertical fracture only if
the minimum principal effective stress (03) were horizontal
and (73 -I- T<p, where Twas the tensile strength of the rock.
Those conditions might occur near the t o p o f a rise on the
sea-floor from which the rocks were sliding. Downward flow
of water into the fracture would continue only as long as the
rocks were sliding, and it would cease when o3 + T>p.
It was more likely that in th~ sustained convective circu-
lations required for one genesis, downward flow of water
occurred over large areas through the pore spaces, so the
horizontal effective stresses were carried by the grain con-
tacts. In the regions of the hotter upward-risingplumes the
pressure of the fluid might have exceeded the combined hori-
zontal principal stress and tensile strength of the rock, so
hydraulic fractures developed extensive dykes and stock-
works. The increased permeability due to hydraulic fractur-
ing would play some part in localizing and accentuating the
upward flow of the hydrothermal solution and so contribute
to the ore formation.
Hunt had suggested that some of the fundamental genetic
problems not yet resolved included the sources of both
magmas and metals, and the possible relations, if any, to
normal igneous processes and subduction zones. The stable
isotope studies indicated that the initial major ore-transport
and alteration processes were magmatic-hydrothermal events.
Although retrograde boiling had been referred to in many
papers, but not at the present meeting, little emphasis had
been placed on the mechanical effects of the process. The
explosive release of magmatic hydrothermal fluids from a
largely consolidated, probably sub-volcanic, shallow igneous
stock explained such characteristic features of porphyry copper
deposits as the extensive crackle brecciation, the occurrence
of porphyry and aplitic intrusions within the central, hotter
zones of the intrusive mass, the extensive alteration and
the zoning. 13 The brecciation not only exposed a very large
surface area of the crystalline phases to reaction with
hydrothermal solutions but the large increase in the permeab-
ility allowed the magma-generated convective system to draw
meteoric waters through the reaction zones in which the
redistribution and concentration of the ore metals might
occur. It seemed likely that the composition of the magma
was not critical but that the significant feature in the forma-
tion of porphyry copper deposits was whether or not retro-
173
grade boiling occurred at a late stage in the consolidation of
the igneous rock. The important prerequisite would be the
generation of a relatively hot, volatile-poor magma, which
might rise to shallow depths and consolidate almost com-
pletely before retrograde boiling occurred. Without retro-
grade boiling a normal igneous rock would form, but with
late-stage retrograde boiling a permeable reactive environ-
ment might be formed within which large-scale ore deposits
could develop by hypogene precesses. The most likely sites
for the development of volatile-poor granitic magma were
the deeper levels of orogenic zones possibly associated with
subduction zones.

References

1. Bishopp D.W. The Troodos Massif, Cyprus. Nature,

Lond., 169, 1952, 489-90.
2. Duffel S. and McTaggart K. C. Ashcroft map area,
British Columbia. Memgeol. Surv. Can. 262, 1952, 122 p.
3. Wilson R. Ao M. and Ingham F.T. The geology of the
Xeros-Troodos area with an account of the mineral
resources. Mere. geol. Surv. Cyprus no. 1, 1959, 184 p.
4. Smith C. H. Bay of Islands Igneous Complex, western
Newfoundland. Mem. geol. Surv. Can. 290, 1958, 132 p.
5. Church W. R. Ophiolite: its definition, origin as oceanic
crust, and mode of emplacement in orogenic belts, with
special reference to the Appalachians. Pubis. Earth Phys.
Brch, Dep. Energy Mines Resources Can., 42, no. 3, 1972,
71-85.
6. Williams H. and Smyth W. R. Metamorphic aureoles be-
neath ophiolite suites and Alpine peridotites: tectonic
implications with west Newfoundland examples. Am. J.
Sci., 273, 1973, 594-621.
7. Williams H. Structural succession, nomenclature and
interpretation of transported rocks in western Newfound-
land. Can. J. Earth Sci., 12, 1975, 1874-94.
8. Glennie K. W. etal. Late Cretaceous nappes in Oman
Mountains and their geologic evolution. Bull. Am. Ask
Petrol. Geol., 57, 1973, 5-27.
9. Kanmera K. Paleozoic and Mesozoic geosynclinal
volcanism in the Japanese islands and associated chert sedi-
"mentation. In Modern and ancient geosynclinal sedimen-
tation Dott R. J. Jr. and Shaver R. H. eds (Tulsa: Society
of Economic Palaeontologists and Mineralogists, 1974),
161-73 (Spec. Pubis $oc. Paleont. Miner., Tulsa,
no. 19).
10. Osmaston M. F. Genesis of ocean ridge median valleys
and continental rift valleys. Tectonophysics, 11, 1971,
387 -405.
11. Sillitoe R. H. Environments of formation of
volcanogenic massive sulfide deposits. Econ. Geol., 68,
1973, 1321-5.
12. Thanasuthipitak T. Relationship of mineralization to
petrology at Parys Mountain, Anglesey. Ph.D. thesis,
University of Aston in Birmingham, 1974. (Abstract
published in Trans. Instn Min. Metall. (Sect. B: Appl. earth
sci.), 84,' 1975, B71.
13. Phillips W. J. Mechanical effects of retrograde boiling
and its probable importance in the formation of porphyry
ore deposits. Tran~ Instn Min. Metal/. (Sect. B: Appl.
earth sci.), 82, 1973, B90-8.

174
Name index Bird J. M. 68, 75, 97 Coleman R.G. 53, 55, 56, 59, 68, 71,
Bischoff J. L. 53, 56, 61, 66, 69, 70 72, 75, 97
Abelson P. H. 56 BishopA. C. 151 Colley H. 41, 84, 85, 95
Adamides N. G. 68 Bishopp D.W. 171, 174 Combarnous M. 70
Ahlfeld F. 102, 113, 116, 125 Bjornsson S. 61, 68, 69 Compston W. 59, 71
Albers J. P. 104, 114 Blake M. C. Jr. 75 Conaghan P.J. 97
Alderton D. 22, 24 Blatner P. 40 Conklin N. 115
Alexiev E. I. 24 Bloxam T.W. 22 Constantinou G. 22, 41, 50, 54, 55,
Allen C. 23, 55 Boato G. 39 64, 65, 66, 67, 68, 77, 171
Allen C. R. 54, 65, 70, 75 Bodvarsson G. 69 CooteA. R. 69
Allen P.M. 23, 95 Bonatti E. 68, 96 Corliss J. B. 41, 71, 96
Amstutz G.C. 71 Bonham H. F. Jr. 114 CornR. M. 115
Andel T. H. van 68 Bories S. 70 Corr R. F. 71
Anderson A. T. Jr. 40 Borley G. 149 Cotton S. 67
Anderson C.A. 41 Bornhardt W. 95 Coutts D. 86
Anderson E.T. 39 Borthwick J. 54 Craig H. 39, 40, 69
Anderson G. M. 25, 39 Bostrom K. 41, 160 Cramer J. 22, 24
Anderson R.N. 70 Bottinga Y. 26, 39, 40 Crick I.H. 162
Andrews A.J. 2, 3 Boussinesq J. 69 Cronan D.S. 79, 80
AngusJ. G. 96 Bowden P. 23, 24
Aoki K. 23 BowesD. R. 19,23 Damon P.E. 101,116
Arculus R.J. 23 Bowman H. R: 113 Dansgaard W. 40
Armstrong R. L. 88, 95 Brace W. F. 62,69 Darcy H. P.G. 62, 69
Arnason B. 40 Braithwaite R.L. 170 Date J. 160
Arn6rsson S. 68 Breemen O. van 23 DavisG. R. 96
Ashley R.P. 114, 116 Brewer P.G. 156, 160 Davis N. F. 25, 26, 39
Atkinson B. K. 146, 149 Brindley J.C. 163, 170 DavisS. N. 69
Atwater T. 70, 115 Bromley A . V . 22, 97 DeerW. A. 55
Audebaud E. 125 Brousse R. 94 Deffeyes K.S. 53, 56
Audley-Charles M.G. 149 Brown G.C. 23, 97, 125 Degens E.T. 70, 160
Aumento F. 44, 56 Brown P.E. 41 Delong S. E. 23
AvilaG. 117 BrowneP. R.L. 56,115 de Montreuil D. L.:
AvilaW. 117, 125 BrunfeltA. O. 22 see Montreuil D. L. de
Aziz K. 70 Bryner L. 114 Dengo G. 102, 116
Buddington A . F . 125 Denholm L.J. 114
Badham J. P.N. 92, 97 Bugge J. A.W. 149 Denholm L.S. 95
Bagby W.C. 116 Bullard E. 81, 94 de Vaumas E.: see Vaumas E. de
Bailey D. K. 21, 24, 94 Burbank W.S. 113, 115 Dewey J.F. 68, 75, 93, 97, 150, 163,
Bailey E.H. 75 Burk C.A. 75 164, 170
Bailey R.A. 115 Burnham C.W. 25, 26, 39, 41 DeWiest R. J.M. 69
Baker P. E. 144, 149 Dickinson W. R. 96
Bakke I. 151 Caltagirone J.-P. 70 Dickson F.W. 53, 56, 61, 66, 69
Baragar W. R.A. 22, 23 Cann J. R. 16, 22, 54, 56, 68, 75, Dietrich V.J. 23
BarkerA. L. 96 141, 144, 149, 151 Dietz R.S. 81, 94
Barnes H. L. 12, 39, 40, 56, 115, 160 Carmichael I. S.E. 149 Dixon C.J. 94, 97
Barnes I. 40 Carr J.M. 56 DoeB. R. 25, 39
Barsukov V. L. 21, 24, 96 CarrM. J. 116 Donaldson I.G. 63, 69, 70
Barton P. B. Jr. 41, 114 Carstens C.W. 142, 148, 150 Dorman L. M. 75
Basham I. R. 97 Carstens H. 150 Dott R. J. Jr. 174
Bass-Becking L. G.M. 71 CarterW. D. 116 Downes K. 170
Baumann L. 97 Casadevall T. 115, 116 Doyle L. 170
Bear J. 70 Chambers A. D. 41 DrakeC. L. 75, 116
Bear L.M. 56, 64, 65, 68 Chapman H.J. 23, 56, 67, 68 Dreier J. E. Jr. 114
Beckinsale R. D. 67 Charlesworth J.K. 165, 170 Drury S.A. 19,23
Bell J. D. 82, 86, 94 Chester R. 80 Dudziak K.H. 70
Bellido B. E. 113 Christiansen R. L. 96, 113, 115 Duedall I.W. 69
Bemmelen R. W. van 151 Chung Y. 69 Duffel S. 174
Bernard A. 87, 94 ChurchW. R. 97, 172, 174 Duke N.A. 22
Bernard A.J. 71 Clabaugh S.E. 113, 116 Dunham Sir Kingsley 97
Bernstein M. 89, 96 Clark A.H. 116 Dymond J. 71
Berry F. A. F. 40 Clark A.M. 125 Dymond J. R. 45, 55
Berry W. B.N. 149 Clark L.A. 23, 56, 160
Best M.G. 125 Clark S.P. 70 Easton A.J. 41, 45, 55
Bethke P. M. 41, 114 Clayton R.N. 40, 42, 45, 55, 56, 69 Eaton G.P. 114
Beveridge T. 3 Clifford T.N. 23 Egin E. 95
Bhattacharji S. 161 Cloupeau M. 70 ElderJ. W. 2 , 3 , 4 , 6 , 10, 12,56,62,
Bickle M.J. 23 Cobbing E.J. 125 63, 67, 69, 70
Bignell R.D. 79, 80 Coleman P.J. 76 Ellis A.J. 2, 3, 40, 56, 156, 160
175
El Shatouri H.M. 114 Glennie K.W. 174 Hildreth R.A. 56, 68
ElstonW. E. 115, 116 Godfrey J. D. 55 Hin E. 125
Engel A. E.J. 56 Goldberg E. D. 3 Hirsinger V. 149
Engel C.G. 56 Goldschmidt V.M. 129, 138, 150 Hitchon B. 40
Enjoji M. 114, 160 Gonz~lez R. 125 Hite R.J. 97
Epstein S. 40, 41,42, 44, 54, 55, 56 Goossens P.J. 112, 116 Hodges F.N. 23
Erb E.E. 115 Gordon L. I. 40 Hoefs J. 69
Etminan H. 41 Govett G. J.S. 22, 41, 50, 55, 64, Hoffman A.W. 41
Evernden J.F. 126 65, 66, 68 Holland H.D. 26, 39, 56, 71,161
Ewing M. 55, 68, 75, 150 Gow A.J. 40 Hollister V. F. 83, 87, 88, 89, 94, 95,
Graff P. R. 22 96
Farrar E. 125 Graham C.M. 54 Holloway J. R. 48, 56
Ferguson J. 71 Grant J. N. 22, 23, 39, 116, 117, Holser W.T. 56
Ferriday I. 128, 149 125, 126 HolstP. H. 70
Fiadeiro M. 69 Gray D.E. 12, 13 Holtedahl O. 132, 150, 151
Fisher F.S. 113 Green D.C. 162 Honma H. 155, 157, 160
Fitch F.J. 151 Green H.W. 23 Honnorez J. 71
Fitton J.G. 163, 165, 170 Green T. H. 92, 97 Horder M. 149
Fleet A.J. 22, 78 Greenbaum D. 59, 68, 73, 75 Horikoshi E. 23, 83, 84, 94, 95, 153,
Fleischer M. 115 Grenne T. 22, 23 159, 160
FlettJ. S. 95 Griggs D.T. 125 Horsley R. 149
Flitton S. 149 Grindley G.W. 64, 69 Horton C. Wo 69
Flower M. F.J. 23 Gr~nvold K. 59, 71 Howie R.A. 55
Floyd P.A. 22 GrossW. H. 106, 113 Hudson J.D. 55, 71
Folinsbee R.E. 115 Grubbs D. K. 39 Hughes D.J. 163, 165, 170
Ford J.H. 41,162 Guilbert J. M. 41, 83, 89, 95 Hulston J.R. 162
Forester R.W. 37, 41,115 Guild P.W. 94 HumphrisS. E. 57, 141,150
Forster H. 96 Gustafson L.B. 41, 115 HuntJ. P. 41,56,98, 115, 172, 173
FoshagW. F. 116 Gustafson W. I. 55 Hutchinson C.S. 75
Foslie S. 129, 130, 132, 134, 138, Gustavson M. 130, 150 Hutchinson R.W. 22, 23, 41, 56,
149, 150 61, 68, 170
Foster B. 125 Haas J. L. Jr. 160 Hynes A. 68, 75
Fournier R.O. 39, 160 Hackett J. P. Jr. 66, 70 HynesA. J. 76
Fowler M. 73 Hadjistavrinou Y. 54
FoxW. 116 Hajash A. 53, 56, 66, 71
Igarashi T. 115
Francheteau J. 65, 70 Hall A. 97
Franck E.U. 70 lijima A. 160
HallW. E. 40, 41
FrangosW. T. 69 Imai H. 114, 126
Halls A. 149
Friedman I. 40, 41 Ingham F.T. 174
HallsC. 22, 23, 97, 116, 117, 126,
Inthuputi B. 97
Fries C. Jr. 116 128, 149
Irving E.M. 113
FrutosJ. J. 111,116 Hamilton D. 41
IrwinW. P. 75
Fujii N. 161 Hamilton D. L. 25, 39
Isacks B. 94
Fujji K. 159 Hamilton W. 97
Ishihara S. 96, 114, 115, 153, 159,
Fyfe W.S. 1,2, 3, 22, 23, 53, 56, Haq B.T. 97
160
58, 67, 69, 125, 141,151 Harakal J. E. 95
Ito K. 55, 70
Hards N.J. 96
Ixer R.A. 95, 173
Gabelman J.W. 96 Harland W.B. 150
Gale G.H. 14, 22, 23, 57, 95, 128, Harris P.G. 126
131,132, 150, 151,172 Hart R.A. 56 Jackson N. 24
Gardiner P. R.R. 165, 166, 170 HartS. R. 59, 71 Jahns R.H. 25, 39
Garlick G. D. 41,45, 55 Hartmann M. 70 Jake~ P. 150
Garside L.J. 114 Harvey R.D. 114 James D. E. 24, 126
Garson M.S. 21,22, 81,82, 85, 89, Hasebe K. 159, 161 James M. 125
90, 94, 95, 96 HatchF. H. 150 Jankovic S. 57
Gaskarth J.W. 95,173 Hatherton T. 96 Javoy M. 26, 39, 45, 55
Gass I.G. 22, 41, 54, 55, 56, 67, 68, Haugen A. 128 Jehl V. 40, 45, 55, 71
72, 75, 76, 97, 171,172, 173 Havens R.G. 56 Jensen B.B. 24
Gayer R.A. 97, 150 Hayes D. E. 23 Jensen R. 149
Gee D.G. 130, 150 Heaton T. H.E. 34, 40, 42, 54, 55, Johannsen 0. 149
Gels H.P. 149 60, 67, 69, 79 Johnson A. E. 56, 61, 69
GeyneA. R. 101,113, 116 Hedge C. E. 59, 69, 71 Johnson H. 56
Gibson I. 149 Helgeson H.C. 12, 56, 63, 65, 69, 70 Johnson-H~st M. 149, 150
Giggenbach W. 40 Hembre O.S. 149 Jones D.L. 75
Gill J. 150 Henley R.W. 1, 3, 58, 67, 70, 159 Jones J.G. 150
Gill J.B. 21, 23, 24, 114, 116, 150 Herrmann A.G. 22 Jones K. 125
Gill K . R . 54 Hess H. H. 39, 81, 94, 150, 172 Jones L.M. 94
Gjelsvik T. 145, 146, 150 Hewett D.F. 115 Julian B.R. 126
Glassley W.E. 71 Heyl A.V. 40 Juve G. 170
176
Kajiwara Y. 155, 157, 159, 160 Lorenz V. 97 Morrison H. L. 69, 70
Kamilli R.J. 115 Lortie R.B. 116 Mottl M.J. 71
Kanehira K. 69, 114 L6vaas G. 149 MudieJ. D. 70
Kanmera K. 174 Lowell J. D. 41, 83, 94, 95 Muehlenbachs K. 40, 42, 45, 55, 69
Kaplan I.R. 56 Lowell R.P. 70, 89, 91 Muffler L. J.P. 56, 95
Karig D.E. 24, 71,94 Lowenstam H.A. 1, 3 Muir I.D. 151
Kato T. 86, 95 LuedkeR. G. 113, 115 MLinnich K.O. 69
Katto Y. 69 Lufkin J.L. 116 Murphy R.W. 149, 150
Kawabe S. 160 Lund A. 149 Murphy T.A. 170
KayR. W. 16,23,68 Lusk J. 170 Murray R.G. 3
Keenan J.H. 70 LyonsJ. I. Jr. 116
KeevilN. B. 11,12,126 Nash J.T. 41,106, 114, 126
Kelly W.C. 122, 124, 126 McAnulty W. N. Jr. 115 Needham H.D. 65, 70
Kennedy G.C. 26, 39, 124, 126 McArdle P. 170 Newall G. 12
Kennett B. L.N. 75 McCormick P. 149 Nicholls I.A. 23
Kennett J.P. 40, 47, 55 McCrea J.M. 55 Nicholson R. 150
Kesler S. E. 83, 94 MacDiarmid R.A. 12 Nieid D.A. 69
Kharaka Y. K. 40 Macdonald R. 21,24 Nielsen H. 56
KiddR. G.W. 67, 68 McDougall I. 114 Nielsen R. L. 39, 56, 70, 96
KiddW. S.F. 22 McDowell F.W. 113 Nishiwaki C. 104, 115
Kilinc I.A. 23, 41 MclverJ. R. 23 Nisterenko G.V. 116
Kirkham D.H. 63, 69, 70 McKeeE. 115 Noble D.C. 113
Kistler R.W. 96 McKee E.H. 104, 114 Noble J.A. 24, 96
Kleinhampl F.J. 104, 114 McKerrowW. S. 97 Nockolds S. R. 39
Klitgord K.D. 70 McTaggart K.C. 174 NolanT. B. 113,114
Knake D. 22 Madden T.M. 69 Noltimier H.C. 70
Knauth L.P. 40, 56 Madel J. 95 Norris Co 149
KnightC. L. 114 Magaritz M. 45, 55
Kobe H.W. 114 MahonW. A.J. 160 Oftedahl C. 129, 130, 131,132, 142,
Koide H. 161 Mallo H. 125 144, 148, 149, 150
Kollung S. 149 Malpas J. 22 Ohmoto H. 25, 36, 39, 41, 55, 56,
KoneEn~ V. 115 Maranzana F. 172 61, 66, 69, 71,115, 160
Kraume E. 95 Marocco R. 125 Oliver J. 94
Krauskopf K.B. 156, 160 Marriner G. 22, 159 O'Neil J. R. 40, 41, 96, 107, 114,
Kristmannsd6ttir H. 56, 68 Martin Vivaldi J.L. 114 115
Kroopnick P. 69 Mason B.H. 12 O'Nions R. K. 59, 71
Krusiewski S.V. 96 Mason R. 149 Orange A.S. 69
KullingO. 151 Masson Smith D. 55, 67, 76 OrcuttJ. A. 75
Kuniyoshi S. 55, 70 Masubuchi K. 160 Osborn E. F. 39
Kuno H. 96, 115, 153, 160, 161 Masuoka T. 69 Osmaston M.F. 171,172, 174
Kuril'chikova G.Y. 96 Matsubaya O. 40, 161 Otagaki T. 95, 160
Kvernvold O. 69 Matsukuma T. 160 OwensD. 113
Matthews D.H. 68 Oxburgh E. R. 97
Lambert I.B. 36, 41, 66, 71, 96, May F. 95 Oyarzun J. 94
104, 106, 114, 150, 153, 159, 170 Mayeda T. K. 40, 55 Oyarz6nM. J. 111,116
Landis G.P. 34, 40, 41 Mayfield I. 149
Langley K.M. 40 Mayor J.N. 113 Palm E. 69
Lanphere M.A. 96 Mehnert H.H. 114 P~;Imason G. 64, 69
Lapwood E.R. 69, 71 MelsonW. G. 68 Palmer Q. 149
Laubacher G. 125 MenziesM. 23, 55 Pan Y.-S. 116
Lawrence J. R. 41 Mercer J.W. 70 Pankhurst R.J. 59, 71, 93, 97
Laznicka P. 41,90, 96 Miller J.A. 151 Park C. F. Jr. 12, 116
Leal G. 114 MinearJ. W. 126 Parmentier E.M. 67
Lebedev L.M. 156, 160 Mitchell A . H . 163, 164, 170 Pearce J.A. 14, 16, 22, 23, 57, 68,
LeeW. H.K. 56 Mitchell A. H.G. 21, 22, 81, 82, 86, 75, 141, 144, 151,172
Leedy W.P. 113 92, 94, 95, 96, 97 Pereira J. 94, 97
Levy E. 113 Miyashiro A. 55, 68, 75, 76, 96, 138, Peterrnan Z. E. 53, 55, 56, 59, 68,
LewisA. D. 22 139, 150 69, 71, 96
LindgrenW. 5, 12, 105, 113 Miyazawa T. 114 Peterson M. N.A. 41, 160
Liou J.G. 47, 55, 70 Moghal M.Y. 97 Petropoulos P.G. 68
Lipman P.W. 96, 102, 107, 113, Monster J. 56 Petruk W. 113
114, 115 MontreuilD. L. de 113 . Phillips W. E.A. 164, 170
Lister B. 94 Moore J.G. 68, 150 Phillips W.J. 39, 173, 174
Lister C. R.B. 51, 54, 56, 57, 64, 70 Moores E.M. 47, 55, 67, 68, 75 Philpotts A. R. 111, 116
Livingston D. E. 89, 96 Morel S.W. 56 Pinckney D.M. 40
Locardi E. 97 Morin P. 94 Pinder G. F. 70
Lombardi G. 40 Morris E. 125 Pineau F. 45, 55
Loncarevic B.D. 57 Morris H.T. 115 Pitcher W.S. 125

177
Plant J. 96 Sakai H. 40, 56, 161 Soler E. 87, 94
Platt J.W. 95, 163 SalasO. R. 113 SomertonW. H. 151
Pocock B.G. 66, 71 Sangster D.F. 23, 166, 170 Sourirajan S. 26, 39, 124, 126
Poldervaart A. 39, 150 Sarre M. 22 Spada A. 97
Potter R. R. 96 Sasaki A. 159, 160 Spencer D.W. 69, 156, 160
Potts M.J. 22 SatoK. 160,161 Spooner E. T.C. 22, 23, 45, 53, 55,
Potts P.J. 22 Sato T. 36, 41, 66, 71, 85, 86, 91, 56, 58, 60, 65,68, 69, 71, 75, 79,
Poty B. 49 94,95,96, 104, 106, 114, 115, 141,151
PouitG. 132 148, 150, 151,153, 155, 156, 157, Springer J. 151
Powell C.M. 97 158, 159, 160, 161,170 Stacey J.S. 25, 39
Prestvik T. 149 Saunders M.J. 49 Stanton R. I_. 71, 95, 152
Price N.J. 151 SavinS. M. 40 Steeves M. 22
Priem H. N.A. 23 Sawkins F.J. 22, 41, 87, 94, 95, 115 Steeves M.A. 23
Prostka H.J. 96, 115 Sch~ifer H. 1, 3 Steiner A. 56
Pryor R. N. 95 Scheidegger A. E. 69 Stern C. R. 96
Pucheldt H. 71 Schiener E.J. 170 Steven T.A. 105, 113, 114
Schilberg L.E. 70 Stevens R. 90
Schilling J.G. 68 Stevens R.K. 97
Schmincke H.-U. 23 Stevenson J.S. 95
Rambaud F. 95, 152 Schmitt H. 99, 107, 113 Stewart F.H. 97
Rambaud Perez J. 95 Schneider A. 56 Stewart J.W. 126
Ramsay W. R.H. 114 Schneider-Scherbina A. 125 Stillman C.J. 163, 166, 170
RankinA. 125, 128 Schoell M. 70 Stohl J. 115
Rast N. 12 Schubert A. 126 StoiberR. E. 111,116
Ratt~J.C. 113, 114 Schuiling R.D. 24 Strand T. 129, 150, 151
ReaW. J. 22, 23 Scott R. 149 Straus J.M. 70
Reading H.G. 163, 164, 170 Scott R.D. 151 Strauss G. K. 95
Rebek R.J. 114 Searle D. L. 22, 56, 61, 68, 97 Streckeisen A. 43, 55
Reed B. L. 96 SekineY. 114 Strong D. F. 22, 23, 80, 95, 96, 151,
Reinhardt B.M. 76 Senechal R.G. 16, 23, 68 159, 164, 165, 170
Reinsbakken A. 128, 147 Shackleton N.J. 40, 47, 55 Stumpfl E.F. 23
Rhoden H.N. 95 Shalaby I.M. 95 Sturt B.A. 151
Rhodes R.C. 115 Sharp G.J. 23 Subbarao K.V. 59, 71
Ribando R.J. 64, 70 Sharp R.P. 40 Suffel G.G. 23
Rice C.M. 41, 97 Shaver R.H. 174 SugimuraA. 116, 159, 160, 161
RiceR. 23 Shawe D. R. 97 Sugisaki R. 40
Richardson H. 125 Sheppard S. M.F. 25, 31, 33, 34, 36, Sundius N. 144, 151
Rickards R.B. 150 39, 40, 41,42, 44, 50, 54, 56, 60, Sundvoll B. 22
Ridge J.D. 115, 160 69, 70, 79, 91,96, 162, 172 Suzuoki T. 40, 44, 55
Riding R. 97 Shibata K. 96 Swanson E.A. 23, 95, 170
Ridler R.H. 23 Shido F. 55, 68, 75, 150 Syers J. K. 43, 55
Riley J.P. 80 Shiina M. 159
Ringwood A. E. 23, 76, 96 Shimazaki u 160 Takagiu 95
Ripley E.M. 116 Shirozu H. 160 Takenouchi S. 114, 126
Rivas S. 125 Shuto K. 161 Takeuchi Y. 56, 95, 114
Roberts D. 22, 23, 131,132, 149, SierraJ. 114 TakeudiY. 116
150, 151 Sigvaldason G.E. 40 Tanimura S. 160
Roberts R.J. 113 Silberman M. L. 41,96, 104, 114, Tarling D.H. 76
Robertson A. H. F. 41, 64, 55, 56, 115, 116 Tatsumi T. 84, 95, 114, 160
71, 78 Sillitoe R.H. 22, 23, 83, 87, 88, 89, Taylor H. P. Jr. 25, 27, 33, 37, 39,
Robertson E.C. 125 91,94,95,96,99, 113, 115, 116, 40, 41,44, 45, 46, 53, 54, 55, 56,
Robinson B.W. 41,114, 95 117, 124, 125, 126, 174 67, 69, 70, 96, 107, 115
Roedder E. 12, 39, 96, 115, 126 Simonian K.O. 22, 68, 75 TaylorP. S. 111,116
Roether W. 69 Simons J.H. 116 Thanasuthipitak T. 174
Rogers F. T. Jr. 69, 70 Skevington D. 150, 151 Theodore T.G. 126
R~rnme I. 149 Skinner B.J. 1,3 Thode H.G. 56
Ross D.A. 70, 160 Skinstad J. 149 Thompson R.N. 23
Ross D. I. 57 Smewing J.D. 22, 23, 54, 55, 56, Thurlow J.G. 23, 85, 94, 95, 127,
Routhier P. 125, 126 67, 68, 75 159, 170
Rowe R.B. 41 Smith A.G. 75 Tilley C.E. 151
Rubenach M.J. 23, 76 Smith B. 56 Titley S. R. 95
Rui I.J. 23, 151 Smith C.H. 174 Toksoz M.N. 126
Ruiz F.C. 116 Smith D. M. Jr. 113 Tokunaga M. 160
Runcorn S.K. 76 Smith P.A. 79, 80 T6masson J. 40, 56, 68
Russell M.J. 97 Smith R.L. 115 Tongiorgi E. 40
Rutland R.W.R. 150 SmythW. R. 172,174 Tooms J.S. 80
Rye R.O. 25, 34, 36, 39, 40, 41, 55, Snelgrove A. K. 94 Tornebohm A. E. 129, 151
56, 66, 69, 71, 114, 115, 160 Snelling N.J. 126 Torrance K. E. 70
178
Torrez J. 125 Wright J. B. 22, 23
Toselli A.J. 126 Wyllie P.J. 96, 126
Tourtelot H.A. 69
Trennery T.O. 66, 71 Yamaoka K. 114
Truesdell A. H. 95, 160 YpmaP. J.M. 116
Turcotte D. L. 70, 126
Turi B. 40 Zachrisson E. 129, 130, 131,132,
Turneaure F.S. 96, 122, 124, 126 150, 151
Turner F.J. 149 Zartman R. E. 40
TurnerJ. C.M. 126 ZeisE. C. 116
Zelenov K.K. 71
Ueda Y. 114 Zussman J. 55
Ueno H. 95, 161
Upadhyay H.D. 22, 96
Urabe T. 157, 160
Urey H.C. 40
Utada M. 161
Uyeda S. 96, 161

Vallance T.G. 144, 151

van Andel T. H.: see Andel T. H. van
van Bemmelen R. W.: see Bemmelen
R. W. van
van Breemen O.: see Breemen O. van
Varne R. 23, 76
Vaumas E. de 76
Veizer J. 59,71
Verhoogen J. 149
Verschure R. H. 22
Villal6n M. 95
Vine F.J. 47, 55, 67, 75
Vitaliano C.J. 114
VogtT. 131,132, 151
Vokes F. M. 22, 95, 128, 149, 151

Walker G. P. L. 149
Walker P. 149
Walker R. L. 94
Walsh J. B. 69
Watanabe T. 84, 95
Watson J. 149
Weber J. E. 69
Weiss R. F. 69
Weissberg B.G. 114, 115
Welland M. J.P. 97
Wells A. K. 144, 150, 151
Wells M.K. 150
Wenner D. B. 40, 54
Wheatley C. J.V. 95, 170
White D. E. 31, 39, 40, 56, 70, 83,
95, 108, 109, 115, 156, 160
WhiteR. 149
Wilkinson N. 125
Williams D. 86, 95, 152
Williams D. L. 70
Williams G.J. 114
Williams H. 22, 172, 174
Wilson G.V. 95
Wilson J.T. 151,170
Wilson M. R. 151
Wilson R. A.M. 55, 171, 174
Winchester J.A. 22
Winter C. K. 22
Wisser E. 100, 101,107, 113, 115
Wodzicki A. 114, 115
Wolfe J.A. 95
Wolff F.C. 151
Wooding R.A. 69, 70

179
Subject index Benioff zones 81, 90, 91, 93, 153
zone-related magmas 91
Acanthite 101, 102 Bentonitic clays 78
Adularia 101, 104, 107, 108 Besshi-type
Agglomerates 142 cupriferous sulphides 81
Aguilarite 101 massive pyrrhotite 87
Alaskites 93 stratiform massive sulphides 86
Albite 128, 132, 137, 139, 140, 143 Bimodal basalt-rhyolite suite 113
Albitic plagioclase microlites and quartz 143 Biotite 19, 27, 28, 31-3, 138, 162
Alestos 58 andesite dykes 104
Alkali Bismuthinite 121
basalts 21 Bolivian granites 20, 21
feldspar 112 Bolivian sub-volcanic deposits 125
granite stocks 90 Bolivian tin-bearing granites 21
magmas 93 from Sn-W greisen 36
Alunite 99, 101, 102, 109 Bolivian tin belt 117
Amino acid 1 Eastern Cordillera of Central Andes 117
Amphibole 19-21, 44, 46-9, 51 tin production 122
Amphibolite 21, 43 Bornite 51,110, 155
facies 42, 43, 48, 58, 73, 138, 141 Boulangerite 90, 102
Amygdales 139, 140 Boulder batholith 33
Andesite-rhyolite volcanics 98 Breccias 104, 110, 119, 120, 163, 165
Andesites 2, 15, 17, 93, 104, 110--2, 139, 144, 146, 153, fillings 112
163, 170 Brecciation 117
Andesitic-dacitic host rocks 107 explosive hydrothermal 117
Andradite 43 igneous 117
Anglesite 155 Broadlands system 64, 66
Anhydrite 31, 36, 52, 153, 157 Buchans orebodies 127
importance in prophyry copper environments 52 high-grade Kuroko-like massive sulphide deposits 127
Anorthosites 43, 81,92 polymetallic sulphide deposit 165
Antimony 90
in tin-tungsten belts in China, Bolivia and Burma 90 Cadmium 146
Apatite 19,21,99, 111 Calc-alkaline 99, 112, 144, 153, 163
Appalachian porphyry-type deposits 89 arcs 165
Appinites 137 eruptives 117, 129
Argentite 101, 104, 157 facies 165
Argillic alteration assemblages 124 lavas 163, 165
Arsenic 146 magmas 91, 92
Arsenophyrite 121, 146 magmatic activity 109, 142, 173
Ash extrusion 26 melts 98
Au-Ag deposits 33, 37 pluton 89
Augite 137 rocks 165
Augite andesite 104 volcanic environments 113, 173
Augite phenocrysts 140 volcanic rocks 110, 112
Auriferous quartz veins 93 volcanic sequences 165
Auriferous sulphides 104 Calcareous
greenschists 139
Barite 85, 101, 102, 153-5,157, 158 sandstones 110
gangue 107 Calcic plagioclase 137, 138
Basalts 15, 34, 44, 47, 52, 139-41, 146, 153 Calcites 43,47, 101,102, 108, 110, 111,137, 139, 140,
calc-alkali 144 146
lavas 132 Calc-sil:cate rocks 93
lavas of the Troodos Ophiolite Complex 77 Caledonian deformation 139
hydrothermal alteration 77 Carbonate 128, 146, 147, 163
massive sulphur deposits 77 Cassiterite 99, 112, 121,125
magma 2, 65, 144 deposition 124
oceanic 48, 144 mineralization 112, 117, 120-2
pillow lavas 16, 60, 154 Chalcedonic vein quartz 47, 50, 101,102
rocks 25, 61, 64, 139 140, 144 Chalcedony 43, 77, 108, 112
strato-volcano 104 Chalcopyrite 43, 50, 61, 65, 85, 101,104, 121,128, 146,
Basic 152-4, 158, 166
dykes 137 facies 163
magmas 4, 72 veins 43
pyroclastic facies 139 Charnockitic rocks 93
transport of Ni, Cr, Co 4 Chemical aspects of sea water 66
tufts 163 basalt interaction 66
volcanic rocks 42, 138, 144, 163 ore deposition 66

180
China clay 36 Dioritic
Cornish high-quality deposits 36 magma 137, 138
isotopically supergeneclays 36 pluton 104
Chloride complexes 157 rocks 136
Chloride-rich (S-poor) fluids 2 Dolerite dykes 66, 67
Chlorine 38 Dolerites 34, 44, 46, 52
Chlorite 28, 43, 46, 47, 77, 128, 132, 137, 139, 140, Dolomite 169
146, 157 Dolomitic nodules 169
Chlorite-epidote 89 Dunites 73
Chloritic alteration 120
Chloritized pyroclastics (mineralized tuff) 168 Eclogite 19, 21
Chondrite 78 Electrum 104, 157
Chorolque hydrothermal system 123, 124 El Salvador porphyry copper deposit 31, 37
Chrome deposits 4 Emplacement processes 74
Chromite 73 Enargite 102, 109
bodies 93 Epidote-clinozoisite 140
Cinnabar 83 Epidotes 43, 46, 47, 49, 128, 132, 137-40
Clays 99, 102 rich segregations 140, 142
Clay-pyrite 89 veins 47, 49
Clinozoisite 137 Epithermal deposits 109, 111
Coastal meteoric waters 31 gold 110
Cobalt 77 gold-silver 33, 34, 90
Cobalt arsenides 81 ore deposition 107, 108
Co-magmatic volcanic rocks 108 sulphide deposition 109
Composite batholith 107 vein deposits 107
Compositions of hydrous minerals 46 Epizonal plutons 26
amphibole, chlorite, epidote 46 Euhedral hornblendes 137
Concentrated residuals in lateritic weathering 2 Exhalite facies 132
Connate water 28, 30, 32
Copper 2, 77, 146-8, 163 Fayalite 112
deposits 2 Feldspar-destructive alteration 33
metallogenic belt 111 Feldspars 4, 21,109, 138, 143
mineralization 89 Felsic
--nickel mineralization 93 eruptives 163
porphyry 4, 89 ignimbrite 110
Copper-pyrite deposits 165, 166 lavas and tufts 109
Cornish granites 20 magma 26
Cretaceous sea-watersource for sulphur 51, 52 stocks 107, 108
Cronebane orebody, East Avoca, Ireland 169 volcanic rocks 99, 153, 166
Crystallizing magma 120 volcanism 127, 159
Crystal tufts 163, 168, 169 Ferromagnesian minerals 138
Cumulus ores 149 Ferromanganese oxides 80
of magmatic origin 149 Fluid dynamic parameters 51
Cupriferous iron sulphides 86 control of sea-water circulation in Troodos crust 51
Cupriferous pyrite ore deposits 58, 61, 86 Fluid inclusions 5
Cupriferous pyrites 34, 61, 62, 67, 86 Fluorine 21
Iberian belt, southern Spain 86 Fluorite 106, 112
Cupriferous pyritic ores 147 Fugacities of oxygen and sulphur 155
Cupriferous pyritic stockworks 86 Fumaroles 67
Cerro Colorado, La Zarza and Avoca 86 Fumarolic deposits 2
Cu-rich pyritic zone, Avoca, Ireland 163
Cyprus deposits 64, 66 Gabbro 17, 43-9, 51,128, 131,132, 137-9, 143, 171
Cyprus ophiolitic rocl~s 67, 171 Gabbroic magma chambers 42, 65
Cyprus sulphide deposits 34, 67 Gabbroic rocks 58
massive 42, 90 Gabbro pegmatite 47
Gabbro plutons 53
Dacite 101,119, 144, 153 Galena 61,65, 85, 101,102, 104, 107, 108, 128, 146,
host rock 101 152-5, 158, 166, 169
porphyry dykes 101 Gangue minerals 28
Deep-sea basalt 65 Garnet 19, 21
Destructive plate boundaries 81 Gemstones 81
Deuterium analysesof fluid inclusions 107 deposits 93
Diabase 171 Geological setting of Skorovas orebody, Gjersvik Nappe,
Diaspore 99, 157 central Norway 128-51
Dickite 101 acid to intermediate flows and pyroclastics 142 ~
Diorite 128, 132, 137-9, 165 basaltic and andesitic lavas 139
intrusive 168 Caledonian tectonism 134
Diorite-type porphyry copper 81, 102 Gelvernokko Nappe 129
181
 Gr~ndalsfjell massif 133, 134, 136-9 High-alumina basalt 153
Helgeland Nappe 131 Hornblende 19, 46, 48, 137, 138
K61i Nappe 129, 131 diorite 136, 137
Leipik Nappe 129 gabbro 138
Lower K61i Nappe 129 peridotites 42
magmatic affinity of Skorovas eruptives 144 Hot spring systems 33
mineralogy and stratigraphy within orebody 146 volcanic rocks 33
Nes&piggen 134, 138 Hydrodynamic model for origin of ophiolitic cupriferous
open dome and basin structures 133 pyrite ore deposits, Cyprus 58-71
peripheral exhalative mineralization 148 Companian-Maestrichtian sea water 60
plutonic chemical aspects of sea water 66
masses of ultrabasic, basic and acid composition 134 cupriferous pyrite ore deposits 58
members of Gjersvik eruptive sequence 134 geochemical evidence for formation of sulphide ore
rocks 134 deposits 61
regional structural and stratigraphic setting 129 geochemical evidence for sea-water interaction 58
Skorovasklumpen 138, 142 hydrothermal convection 61
Skorovas orebody 128, 134, 138, 139, 142, 144-7 hydrothermal ore deposit formation 58
area 134, 140, 142, 146, 148, 149 ophiolitic rocks of Troodos Complex 61
massive and disseminated pyritic ore 128 ophiolitic rocks of Troodos Massif, Cyprus 58
Staten Group 131 sulphur isotopic compositions 61
stratiform base-metal deposits, Central Norwegian Hydrogen and oxygen isotope evidence for sea-water-
Caledonides 128 hydrothermal alteration and ore deposition 4 2 - 5 7
structural style of orebody 145 basic volcanic rocks 42
tectonic style within Skorovas area 132 ophiolite sequences 42
Trondheim region 131,132, 143, 148 epidote veins from gabbro, trondhjemites and S.I.C. 49
volcanic rocks 134, 139, 143 exchange experiments 43
Goethermal geological setting 42
springs 30 isotopic composition of hydrothermal waters 47
Coast Range, California 30 Limni stockwork 50
systems 31 massive sulphide deposits 42, 43
waters 30, 31 metamorphic and hydrothermal processes 42
Gjersvik plutonics 131 stockwork zone 43
Gold associated with island arc deposits 83 sulphur isotopes 51
derived from lavas and volcaniclastic rocks 83 ratios in sulphides 51
Gold-bearing pyrite 87 Troodos
magmatic arcs 87 hydrothermal waters 48
Gold deposits 2, 81, 93, 104 ophiolite complex 42
in outer arcs and inter-arc troughs 90 Upper Cretaceous 43
Gold-silver deposits 33, 37 Hydrogen isotope ratio 25
Gold-silver tellurides 104 Hydrothermal
remobilization from relatively gold-rich sediments 2 alteration 124
Gold telluride deposits 113 breccia 102, 120
Granite plutons 81 circulation of sea water 14
Granites 20, 21,92 convection 61, 64
tungsten-bearing of Japan and Asian mainland 90 fluids 6, 25, 31, 33, 42, 47, 52, 61, 112, 123
Granitic metamorphism 60, 61
batholiths 98 metasomatism 52
dykes 136 ore deposits 2, 4, 5, 33, 38
magmas 29 outlets 80
melts 25 sericites (and clays) 33
plutons 90, 118 systems 4 - 6 , 2 5 , 34, 116, 122, 123
rocks 21 water 28
veins 19 Hydrothermal alteration of basaltic lavas of Troodos
Granodiorites 89, 118, 128, 131,132, 165 Ophiolite Complex 77
batholiths 118 Cyprus massive sulphide deposits 77
dykes 137 Lower Pillow Lavas 77
rocks 138, 142, 143 massive sulphide deposits 77
Greenschist 21,42, 43, 48, 58, 73, 129, 138, 140, 141 basaltophilic elements, S, Fe, Cu and Zn 77
facies 34, 43, 4 6 - 8 , 53, 60, 66, 138, 139, 146, 165, Sheeted Complex 77
171 stockwork zone 77
Greenstones 131,132, 137, 139 titanomagnetite 77
Green Tuff strata 153 Hypersthene 137
Groundtnass quartz 47, 50 Hypogene 110
Gypsum 35, 36, 153 chalcocite 110
Gypsum--anhydrite masses 154 clays and sericites 31
Hypothermal ore deposits 4, 5
Harzburgites 43
Hematite 51,99, 111

182
Identification of ore-deposition environment from trace- chloride-rich (S-poor) fluids 2
element geochemistry of associated igneous host rocks geochemical, isotopic and structural studies 1
14--24 heat transfer by large-scale convective motions 1
geochemical characteristics of tin-bearing magmas 20 high source concentration 2
igneous geochemistry 14 large-scale ore-forming processes 1
identification of tectonic environments 14 prospecting 2
massive sulphide deposits 14, 17 subsurface aspects of ore-forming process 1
magmatism in the oceanic environment 14 volcanism 2
source of tin-rich magmas 21 Iron ore 111
tin deposits 18 Iron oxidation ratio 60
volcanogenic ore deposits 14 Iron oxides 80, 137
Identification of origin of ore-forming solutions by use of Iron-rich chlorite schists 132
stable isotopes 25-41 Iron sulphide 138
equilibrium isotope fractionations 27 Iron--titanium ores 81
formation waters 30 deposits in anorthosites 92
geothermal waters 31 Iron-titanium oxides 140
hot spring systems 33 Isotope geothermometers 27
hydrogen and oxygen isotope fractionations 27 Isotopic compositions 31
hydrogen isotope compositions of fluids 33 of hydrothermal waters 47
hydrothermal fluids 26, 33 of hypgene and supergene clay minerals 36
hydrous saline melts 26 Isotopic exchange with oceanic basalt 51
isotopic composition of hydrothermal waters 32, 33
magmas and hydrothermal fluids 25
Japanese islands 159
magmatic waters 29
metamorphic waters 30 Kuroko deposits 154, 159
meteoric waters 29 Juvenile water 28
N C I - H 2 0 system 26
origin of hydrothermal fluids 30 Kaolinite 27, 28, 30, 36, 77, 157
other epizonal ore deposits 33 Kaolinite-water system 27
porphyry copper-molybdenum deposits 31 Keratophyres 139, 141
porphyry deposits 26 Kuroko deposits: their geology, geochemistry and origin
silicate melt 26 153-61
submarine volcano deposits 34 argillaceous sedimentation 153
types of waters 25, 28 behaviour of ore fluid in sea water and mechanism of
water/rock ratios in hydrothermal systems 37 mineral zoning 157
Igneous geochemistry 14 Benioff zone 159
Igneous geology and the evolution of hydrothermal systems chemical compositions, temperatures and pH values 156
in some sub-volcanic tin deposits of Bolivia 117-26 circulating sea-water versus magmatic origin 158
batholiths 119 distribution of Kuroko deposits in Japanese islands 154
Bolivian tin belt 119 geological setting 153
breccia types, Salvadora stock 121 Green Tuff sequences 153, 159
Cordillera Oriental, Bolvia 117 Miocene submarine volcanics 153
complex vein systems 117 gold and silver 154
hydrothermal alteration 117 Kuroko zones 153
pervasive disseminated mineralization 117 mafic to felsic trend 153
evolution of hydrothermal systems 122 massive, stratiform sulphide ores 153
geology of tin belt 118 model for mechanism of mineral zoning in Kuroko
hydrothermal alteration amd mineralization 120 deposits 158
hydrothermal process at volcanic centres 117 model ore-forming system 157
Palaeozoic sedimentary rocks 119 model solution 157
polymetallic deposits in Altiplano 118 origin of ore fluid and metallogenesis 158
regional geology of Central Andes 118 pH estimations 157
southern Bolivia deposits 118 physico-chemical environment of ore deposition 155
acid eruptive complexes 118 pressure 155
Igneous processes 98 proposed model 159
rocks 2, 4, 14, 29, 117, 124 stockwork ore 154, 155, 157
altered quartz porphyries 117 stratigraphic positions of Kuroko deposits 154
mineralization 117 underlying stockwork ores of Miocene age 153
system 37 vein-type deposits 159
Ignimbrites 104, 163 Kuroko deposits 17, 35, 36, 66, 86, 91,93, 104, 107-10,
Illite 77 148, 153, 159, 163
Ilmenite 43 Kuroko fluid 158, 159
Indian Ocean 35 Kuroko-type zinc-copper-lead sulphide 81
Indonesia granites 20
Introductory remarks on transport problem 1 - 3 Labradorite 143
andesites 2 Laramide porphyry copper deposits 31,91
chemical transport 1 Arizona, New Mexico, Sonora 31

183
Laterites 4 Fiji 104
Lavas 14--6 fluid-inclusion studies 105, 106
Cyprus lavas 16 general characteristics of epithermal deposits 99
Layered gabbros 128 genesis of epithermal deposits 107
Leaching of ore metals 42 geotectonic setting of epithermal deposits 109
Lead 38, 152 Great Britain, Nevada, U.S.A. 103
Lead isotopes 25 Hauraki region, New Zealand 104
Lead-zinc manto-type copper deposits 110
deposits 30, 165 precious-metal and volcanogenic tin belts, Mexico 100
mineralization 153 Spain 105
Leucoxene 43 epithermal gold-bearing veins 105
Limestones 131,165 stable-isotope studies 107
Limni opencast mine 58 typical regions of epithermal deposits 100
Limonite 112 Caenozoic volcanic rocks 100
Lithosphere 75 epithermal mineral deposits 100
Lower Palaeozoic rocks 128 vein and dyke systems, Pachuca district, Mexico 101
Ordovician age 128 volcanogenic magnetite deposits 111, 113
Lower pillow lavas 43, 45, 77, 172 volcanogenic tin deposits 112
Metallogenic belts 111, 117
Magma chambers 38 metamorphic assemblages 141
Magma genesis 21 processes at plate boundaries 81
Magma interaction experiments 159 Metamorphism 43, 46, 48, 53, 60, 142
Magmas 2, 14, 19, 29, 117, 118, 138 Metasedimentary rocks 19
tin-rich 21,22 Metasomatism of basaltic and doleritic ophiolitic rocks 61
Magmatic Metavolcanic rocks 17
arc rocks. 93 Meteroric-hydrothermal fluids and systems 29, 31, 33, 37,
belt 81, 92 38, 91,108, 112
fluids 2, 18, 33, 37, 153 Meteoric waters 28, 29, 31,33, 108, 125
-hydrothermal systems 25, 29, 32, 33, 38, 39, 91,125 Micas 19-21
rocks 89 granite 89
source for sulphur 51 lattices 21
water 28, 42, 48, 108, 153, 159 Mimetite 112
Magmatism 3, 14, 18 Mineral deposits in Peruvian Andes 87
magma types 15 gold-copper 87
distribution 18 iron 87
Magnesium 77 lead-zinc-silver 87
Magnetite 19, 43, 51,111, 112, 137, 146, 147, 165 tin-tungsten--silver-molybdenum 87
magnetite--pyrite ores 128 Mineralization at destructive plate boundaries 81-97
Manganese 77,104 back-arc basins 81, 90
precipitation 80 tin-tungsten-fluorite 90
Manganese oxides 78 banded Kuroko-type ore, Buchans mine, Newfoundland
deposits 109 85
Manto-type copper deposits 99, 110-2 calc-alkaline volcanism 84
stratiform bodies 110 collision belts, southwest England 92
Marcasite 67, 104 distribution of main porphyry deposits in North and
Marianas basin 17 South America 88
Massive gold mineralization
cupriferous pyrite deposits 52, 53 auriferous sulphides and tellurides 87
pyrite 146 quartz-diorite plutons 87
rhyolite 168 quartz veins in meta-andesites and volcaniclastic
sulphide deposit types 14, 15, 18, 34, 35, 42, 51, 54, rocks 87
77, 93, 169, 173 island arc-type plate boundaries 81
Mass-spectrometric measurement 26 calc-alkaline magmatic rocks 81
isotope ratios of 26 ensialic arcs 81
Mean concentrations of trace elements 14 ensimatic arcs 81
Mesozoic rifting 134 Kami dacite stock, Bolivia 89
Metabasalts 14, 16, 21, 60 Kuroko-type zinc-lead-copper stratiform sulphides 83,
sub-oceanic 46 84
Metabasic rocks 1, 4, 14, 28 magmatic arc-type plate boundaries 87
Metadolerites 44, 46, 60 magmatic belts 82
Metagabbros 60, 141 mercury
Metallic mineralization affiliated to subaerial volcanism deposition in vapour-dominated reservoirs 83
99-116 native sulphur-pyrite deposits 83
central Andes 101 porphyry copper deposits 82
Colorado, U.S.A. 102 quartz-chlorite stockwork 85
deuterium analyses of fluid inclusions in vein quartz and quartz stockwork 87
adularia 107 source of metals 90

184
 stockwork mineralization 84 Oolitic dolomite 166
tin and tungsten deposits 89 Opal 112
Miocene tin-bearing Windy Fork granites, Alaska 89 Ophiolites 2, 18, 34, 44, 72, 73, 75, 171-3
Mineralized stockwork material 61 as oceanic lithosphere 73
Miocene 158, 159 complexes 17, 42, 43, 52, 73
C u - P b - Z n - A g - A u Kuroko deposits of Japan 66 in collision settings 93
sea water 158 lavas 73
sediments, tuff and breccia 119 processes 73
volcanism 153, 158 Ophiolitic rocks 64, 66, 67, 93
Mixed igneous-sea water source for sulphur 52 Ophiolitic wedges 93
Model of hydrothermal ore genesis 4 - 1 3 Ore fluids 159
extraction from source 8 magmatic origin 159
flow in surface zone 9 Origin and emplacement of ophiolites 7 2 - 6
formulation of the problem 4 comparison of ophiolite sequences 72
gas column 6, 7 emplacement processes 74
operation of hydrothermal systems and key parameters environment and geochemical charactistics of
4 ophiolites 72
reservoir-pipe model of hydrothermal system 5 Greek ophiolites 75
segregation of material 4 idealized complete ophiolite sequence 72
source constraints 5 ophiolites as oceanic lithosphere 73
transport of soluble matter 10 ophiolites in Alpine-Himalayan chain 74
Modern granite plutons 2 Papua and New Caledonia 75
Modern submarine hydrothermal mineralization, Origin of hydrothermal fluids 30
Santorini and the Red Sea 80 Orpiment 102
hydrothermally introduced metals 80 Orthoclase-biotite 162
manganese precipitation 80 Orthopyroxene 19
metallogenesis on mid-ocean ridges 80 Outer and magmatic arcs 93
metal-rich sulphides 80 Oxygen fugacity 157
submarine hydrothermal activity 80 Oxygen isotope composition of fluid 50
World Mid-Ocean Ridge System 80 isotopes 25
Molybdenum deposits 25, 31,81,83
in cratonic porphyry deposits 83
with porphyry copper 31, 87 Palaeo-Benioff zones 81, 89
Montmorillonites 36, 77, 112 Palaeozoic
Morphotectonic divisions 117 porphyry stock 119
Central Andes 117 sedimentary rocks 119
Motoyama deposit, Kosaka mine 158 volcanism 164
Mount Olympus region 42 Paragonite 138
Mudstone 153, 154 Pb-Zn mineralization 106, 153
Muscovite 19 Pb-Zn ore deposits 30, 165
Mylonitic facies 137 Pegmatites 2, 4
Sn, W 2
Native bismuth 112 Pelagic clays 78
Nepheline-syenites 93 Pelitic rocks 138
Nickel 77 Peralkaline obsidians 21
Nickel-copper mineralization in banded gabbroic and Peridotites 43, 73, 74, 93, 171, 172
ultramafic rocks 93 Permeable water-bearing rocks 26
Nigerian granites 21 Phenocrysts 48, 143
concentrations of Zr, Nb and rare earths 21 Phlogopite 19, 21, 117
Nigerian Younger Granites 21 Pillowed basaltic flow units 139, 140
Pillowed breccias 132, 139
Occurrence, origin and significance of mechanically Pillow lava calcite 54
transported sulphide ores at Buchans, Newfoundland Pillow lavas 42, 43, 46-8, 58, 60, 61,63, 65--7, 153
127 Pillow structures 140
Buchans orebodies 127 Placers 4
high-grade Kuroko-like massive sulphide deposits 127 Plagioclase 19, 43-5, 48, 51, 77, 137, 138
direction of transport 127 anorthite 47
fragmentation of in-situ deposits 127 crystals 137
sulphide breccia accumulations 127 microlites 140
Ocean-floor basalts 5, 144 Plagiogranites 43
Oceanic basaltic host rocks 34 Planes - San Antonio pyritic deposit, Rio Tinto, Spain:
Oceanic dolerites 44, 45 nature, environment and genesis 152
Oceanic plagiogranites from ophiolite complexes 53 bedded pyrite sheet 152
Oligoclase 143 cupriferous pyritite 152
Olivine 19, 21, 138 hypogene processes 152
basalts 43 Lower Carboniferous felsic pyroclastic rocks 152
gabbros 73, 136 stratiform cupriferous deposit, Rio Tinto 152

185
 submarine explosive felsic volcanism 152 -latite 119
sulphur-rich-copper-poor sulphide ore 152 -monzonite 89
underlying volcanic pile 152 phenocrysts 120, 122
Plutonic -sericite-pyrite 89
complex 149 -tourmaline rock 120
infrastructure 128
massifs 132, 133, 137 Radiometric age dating 119
rocks 29, 33, 43, 132, 138, 143, 144, 171 Rare-earth element evidence for genesisof metalliferous
sequence 136 sediments of Troodos, Cyprus 78-9
Plutons 2, 54, 92 chondrites 78
Polybasite 101,104 instrumental neutron activation analysis 78
Polycrystalline pyrite 146 Rare earths 14-7, 20, 21,172
Polymetallic (Cu-Pb--Zn-Ag) ore deposits 109, 170 Ratios for chorites, amphiboles and epidotes in Troodos
Polymetallic sulphide mineralization 122 rocks 50
Porphyries 26, 163, 165 Ratios of pyrites reported from Troodos stockwork
Porphyry 119, 120 disseminations 51
copper-bearing stocks 109 Red Sea brines 31, 34
copper deposits 2, 14, 20, 26, 31,33, 36, 81-3, 87, Reykjanes, Iceland 31
89-91,93,98, 109, 111,112, 122, 162, 171-3 geothermal system 64, 66
copper-molybedenum deposits 25, 31, 33 Rhodochrosite 102
Phanerozoic period 31,33 Rhodonite 100
dykes 142 Rhyodacite 117
mineralization 89, 112 magma 117
rocks 32 Rhyolites 38, 53, 101,104, 112, 119, 153, 158, 163, 165
tin deposits 121 high-silica 112
Post-collision magmatic belts 94 plugs 99, 112
Post-Palaeozoic volcanic sequences 99 Rhyolitic
Potash feldspar 138 breccias 165
Potassic granites 38, 53 rock 112
Potassium 157 sheets 165
Potassium tholeiites 73 volcanism 112
Potential magmatic-hydrothermal solution 26 Rock mechanical properties 1
Precambrian systems 38 fracture density 1
Primary magmaticwater 29 permeability 1
Pumice breccias 104 Ruby silvers 101
Pyrargyrite 102, 104, 108 Rutile 4
Pyrite 18, 43, 50-2, 61,65, 67, 77, 81,101, 102, 104,
108, 110, 121,128, 132, 137,138, 140, 143, 146, Salinities 25, 31, 34, 37, 54, 66, 106, 124, 157, 159
157, 158, 166 Sanidine 112
Pyrite--chalcopyrite-sphalerite vein assemblages 50 Santorini 80
Pyrite--chalcopyrite stability field 51 Saussuritization 143
Pyrite sulphur isotope data 52 Scandinavian Caledonides 128
for massive sulphide ores 32 Sea-water types 28
Pyritic basalt hydrothermal experiments 51
deposits 152 hydrothermal metasomatism of carbonate ooze 35
facies 146 pillow lava interface 53
lead-zinc-copper ore 166 Sericites 33, 104, 120, 138, 146, 157, 162
mineralization 33 Sericite-tourmaline alteration 120, 121
ores 146, 163 Sericitic alteration assemblages 120
Pyritite 152 replacement of biotite and feldspar phenocrysts 120
Pyroclastics 163, 165 Sericitized pyroclastics (rhyolitic tuffs) 168
Serpentines 42
beds 99, 163
fabrics 142 Serpentinization 75
facies 142 Sheeted dyke complex 58, 64, 72
levels 142 Sheeted intrusive complex 42, 45, 77
Pyrolusite 78 Shimogamo, Japan 31
Pyrophyllite 99, 157 Silica 50, 144
Pyroxene 19, 21,43, 44, 48, 77, 138 Silicate alteration assemblages 117, 120
Pyrrhotite 18, 51,128, 138, 146, 155 quartz-tourmaline 117
Pyrrhotite and bornite 51 sericite--clay 117
tourmaline-sericite 117
Silicate magmas 158
Quartz 4, 21, 43, 45, 47, 49, 77, 101,104, 106-8, 110-2, Siliceous ore 86
120, 128, 138, 140, 143, 146, 153-5 disseminated pyrite and chalcopyrite 86
--cassiterite veins 117, 120-2 Silicic tufts 17
diorites 43, 83 Siltstone 153
-feldspar porphyry dykes 142 Silver 99, 146, 155
gangue 122 Silver-bearing sulphosalts 104, 125

186
Silver deposits 33, 81,109 mineralogical facies 147
-lead mineralization 108 processes 98
Silver-rich zones, Bolivia 118 setting of massive sulphide deposits 17
Skorovas orebody 128 stages of allochthonous emplacement 128
Central Norwegian Caledonides 128 Tennantite 102, 104, 110, 146
Skouriotissa 43, 46,47 Tertiary meteoric waters 36
Smectite 43 Tertiary Peruvian epithermal deposits 107
Source rocks 2 Tetrahedrite 104
fracture density 2 Tetrahedrite-tennantite 154
pervasive permeability 2 Thermal spring system 110
Specularite 112 Tholeiites 15, 16, 18-21,153
Sphalerite 43, 50, 61,65, 85, 101,102, 104, 106-9, 146, Tin-bearing
153-5, 157, 158, 166, 169 breccias 112
Sphene 21,43, 138-40 granites 14, 20, 21,92
Spilites 2 SiO2 and Nb data 20
Stable isotope studies on Bougainville and in Matupi pegmatites 21
Harbour, New Britain, Papua New Guinea 162 veins 120
alteration assemblages 162 Tin deposits 14, 18,33, 109, 112, 117, 118
Matupi springs 162 magmatism in the continental environment 18
Fe, Mn and Zn values 162 Tin granites 91
ore-forming fluids 162 Tin-rich
Panguna porphyry copper deposit 162 detrital sediments 21
propylitic zone 162 magmas 22
secondary biotite and sericite isotope data 162 Tin-tungsten deposits 81, 92
thermodynamic and fluid techniques 162 Tin-tungsten-fluorite 90
Stannite 121 Topaz 112
Stilpnomelane 139, 140 Tourmaline 120
Stockwork 45, 47, 49, 54, 77 Tracers 60
chlorites 47 strontium isotopic ratios 60
temperatures 54 Transportation of ore material in hydrothermal fluid
Stratified tuff bands 132 Trondhjemites 38, 42, 43, 45-7, 49, 53, 60
Stratiform massive sulphide orebodies 81 Trondhjemitic granodiorite 139
Stratiform pyritic orebodies 129 Troodos
Strontium 59, 60 basaltic rocks 52
isotope ratios 59, 112 complex 43, 58-65, 77
isotopic contamination 60 Upper Cretaceous 43
Submarine hydrothermal mineralization 80 gabbro 171
Atlantis II Deep 80 hornblendes 47, 48
Santdrini, Cyclades Volcanic Arc, eastern Mediterranian igneous complex 52
80 ophiolite complex 42, 58, 77, 172
Submarine volcanogenic deposits 34, 142 pillow lavas 48
Sub-volcanic sulphide deposits 66
hydrothermal ore systems 122 Tuff breccia 154
polymetallic deposits 122 Tuffaceous mudstone 169
systems 124 Tufts 165
tin deposits of Bolivia 117 Tungsten deposits 33
Sulphate-reducing bacteria 2
Sulphide Ultrabasic intrusions, layered 4
breccia accumulations 127 Ultrabasic rocks 90
deposits 14, 104 igneous 72
facies 148 Ultramafic rocks 92, 171
ores 124 Umbers 67, 78, 79
Sulphosalts 102, 104 copper-enriched 79
Sulphur 163 Upper Cretaceous sea water 61
Sulphur-bearing fluids 50 Upper gabbros 42, 58
Sulphur isotopes 51 Upper Pillow Lavas 43, 45
ratios in sulphides 51
Sulphurization of titanomagnetite 77 Vanadium 81
basaltic lavas and euhedral leucoxene 77 Vein
Supergene clays 36 deposits 122
mineralization 122
Taupo district, New Zealand 4, 9, 12 quartz 49, 50
Taupo systems 9, 10 tin and tungsten 122
Tectonically emplaced ocean-floor deposits 90 Volcanic
Tectonic control of distribution of porphyry orebodies 83 arc magmas 19
environments 15, 16 complexes 124
facies 140 greenstones 132
formation of porphyry copper deposits 98 hydrothermal origin 165
187
Volcanism 2
Volcanogenic
clastic sediments 110
copper-lead-zinc deposits 165
magnetite deposits 109, 112
massive sulphide deposit 14, 35
magnetite 99
tin deposits 99
ore deposits 166
ore systems 166
sulphide ore deposits 166
tin deposits 109, 112, 113
Volcanogenic mineralization at Avoca, Co. Wicklow,
Ireland, and its regional implications 163-70
Avoca ore geology 166
Avoca volcanogenic ore types 166
Caledonian geology of southeast Ireland 166
Caledonian orogenic belt 163, 165
disseminated sulphide ores 163
lead--zinc--copper zone 169
manganiferous and dolomitized sediments 163
massive pyritic ores 163
massive sulphide orebodies 165, 166, 168-70
Ordovician volcanic sequences, southeast Ireland 163-70
Palaeotectonic setting, Avoca region 163
stratiform
polymetallic sulphide ores 165
pyritic orebodies, Avoca, Ireland 163
stringer ore 166, 170
structural relationships and stratigraphy 165
Waterford-Arklow--Avoca belt 163
Volcano-tectonic regimes 110

Wairakei geothermal system 31, 63, 64, 66

Water/rock ratios 37, 50, 60
in hydrothermal systems 37
White mica 112
White rhyolite 158, 159
magma 159
Wood tin 99, 112

Xenoliths 119, 138

Zeolite 46, 48, 60, 73, 112

facies 42, 43, 4 6 - 8 , 54, 58
Zeolite--greenschist facies metamorphism 73
Zeolite-greenschist rocks, Troodos 52
Zinc 99, 146-8, 152
Zinc-rich ores 147
Zircon 19, 21

188