On the Súrya Siddhánta, and the Hindú Method of Calculating Eclipses

Author(s): William Spottiswoode

Source: The Journal of the Royal Asiatic Society of Great Britain and Ireland, Vol. 20
(1863), pp. 345-370
Published by: Cambridge University Press
Stable URL: http://www.jstor.org/stable/25581255
Accessed: 23-06-2018 01:58 UTC

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide
range of content in a trusted digital archive. We use information technology and tools to increase productivity and
facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
http://about.jstor.org/terms

Cambridge University Press is collaborating with JSTOR to digitize, preserve and extend
access to The Journal of the Royal Asiatic Society of Great Britain and Ireland

This content downloaded from 97.86.236.53 on Sat, 23 Jun 2018 01:58:15 UTC
All use subject to http://about.jstor.org/terms
 345

Aht. XIV.?On the Surya Siddhdnla, and the Hindu Method of

Calculating Eclipses. By William Sfottiswoode, M.A.,
F.R.S., &c.
[Read Idth January, 1803.]

Seimc years ago it was suggested to mc to undertake an edition

and translation of the Surya Siddhanta; but a long and careful
stuely of the work convinced me that it would be impossible satis
factorily to accomplish the object without more assistance than
was to be found in Europe. The MSS. were few and insufiicient
in accuracy; the lexicon was deficient iu most of the technical
terms; anel the einly works that threw any light upon the subject
were tlmse* of Colebroeike, invaluable as far as they went, anel the
Abbe Gucriu's Astreineunic Iiidieuiic. The other writers who hael
touched upon the subject merely reproduced what was to be found
in Colebrookc.
Mr. 1 La-IPs edition in the* Ribliotheca ludica, aud Mr. Rurgess'
elaborate translation and notes, published by the American Oriental
Society,1 neiw se*em to leave little wanting upon the subject. Rut
feir those who wish to comprehend the nature, and estimate the
real value, of the Hindu astronomical methods, without entering
moro deeply than necessary into the complexities of either text or
commentary, it may still be useful to present the processes iu as
compendious a form as possible. I have therefore attempted to
translate into modern mathematical language anel formula* the
rules of the work in question.2 The text, like all Sanskrit works,
gives no account whatever of its rules or metheiels; and as the
explanations of the commentators, being of comparatively recent
date, have little or no interest for the history of the science, it
appeared to me foreign to the present purpose to retain them.
Under these circumstances I have contented myself iu the case of
exact formula?,, with.occasionally adding a few of the steps necessary
for verification ; anel iu the case of approximate, with pointing out
what assumptions arc involved iu the rules which they express.
It should be added that, the assumptions so indicated are neit
1 Journal of the American Oriental Society, vol. vi, pp. 145?498.
2 These ruleB arc contained in Chaps. I?VI. To the remaining chapters the
process is not applicable.

This content downloaded from 97.86.236.53 on Sat, 23 Jun 2018 01:58:15 UTC
All use subject to http://about.jstor.org/terms
 346 ON THE SURYA SIDDHANTA.

necessarily hlontical with, although in some sense equivalent to,

tlmse* originally made by the author of the treatise.
I'Yeun sue h observations sh they wero able; tei make*, the) Himhis
deduced values fen* the mean motions of the sun, meion, and planets,
supposed to revolve about the earth, ami of their apsides and node*s.
Hy means eif these value's they calculated buck to remote ejpeiedis,
when, according fo I heir ehtfu, I here wouhl have1 been a general
conjunctieiii eif parts or eif the entire) system.
The following is an em time e)f the process eif calculating a lunar
eclipse. First find the number of days elapsed from the commence
ment of the age, or period, to the mean midnight next before the
full inoeni for which an eclipse was te) he calculated. The original
determination of the mean motions hael eif course given the current
year of the period. This being done, an easy arithmetical process
gave the mean longitude of the sun, the moon, anel the moon's
apsis.
The process of correction, whereby the true hingitudes were
thence deduced, is curious and peculiar. It had been neiticed that
the apsieles, or points of slowest movement, and the jtositions of con
juuetion with the sun had proper motions. These were attributed
to iidluenccs residing in the apsides and conjunctions respectively ;
and corrections due to each were accordingly devised. The un
disturbed orbit was considered a circle with the earth (E) iu the
centre; and iijxm it the centre of a smaller circle e>r epicycle moved
with a uniform angular velocity, equal, but npjtositc iu direction, tei
that of the undisturbed planet; so that M being the centre, anel m
any given point on the epicycle, Mm always remained parallel to
itself. If then at the apsis, or conjunction (according as the correc
tion of ono or of the other was being calculated), M m was in a
straight line with B m, the true position of the planet was conceived
to be at the point where E m cut the umlisturbed orbit. The
radius moreover of the epicycle was variable; anel its magnitudes
at the eidel and even ijuadrants being elctcrmimul sei as te) satisfy
observation, its intermediate variation was considered proportional
to the sine of the mean aneunaly.
The true longitudes and elaily motions of the sun anel moon
having been found, the interval between mean midnight anel the
end of the half month, or moment of opjiosition in longitude, en*
middle of the eclipse is then determined. Rut since the Hindu
time is reckoned from true sunrise to true sunrise, it is next required
to determine the interval between mean midnight anel true sunrise.
This is effected by means of (1) the cepiation of time, found by a

This content downloaded from 97.86.236.53 on Sat, 23 Jun 2018 01:58:15 UTC
All use subject to http://about.jstor.org/terms
 ON THE SURYA SIDDHANTA. 347
simple; but rather rough method ; (2) the precession, of which more
below ; anel (']) the ascensional difference}.
The* diameters of the sun, moon, and shadow, are found on the
principle that their true* are to their mean diameters, as their true
are to their mean motions.
Lastly are determined the moon's latitude at the middle of tho
eclipse; the amount eif greatest obscuration; the duration of the
eclipse; of total obscuration (if it be total); and the times e?f first
and hist contact of immersion and emergence ; by methods which do
not require particular notice apart from the details themselves.

SURYA SIDDHANTA.

Chatteii f.

On the Mean Motions oe the Rjlanets.

The divisions of time are as follow (vv. 11, 12) :?

10 hmg syllables = 1 respiration (= 4 seconds),
G respirations = 1 vinaelf,
GO vimidf = 1 naeli,
60 naelfs = 1 day.
Those of the circle arc the same as ours (v. 28).
The civil day is reckoned front sunrise to sunrise, and feir astro
logical purposes a mouth consists of 30 such days, anel a year of 12
such months.
The astronomical day is the interval from midnight tei midnight.
The lunar month is the interval from one new, or full, moon to
the next. It is elivideel into 30 lunar days, whiem of course elo not
correspond with civil days. The lunar month is named after the
solar month in which it commences. When two lunar months
begin in the same solar month, the former is called au intercalary
month, anel both bear the same name. The civil day is named
after the lunar elay in which it ends; when two lunar elays end in
the same civil day, the feirmcr is " omitted," and both bear the
same name (see also vv. 34-40).
The solar year is sidereal, and the zodiac is elivideel into 12
signs, to each of which corresponds a solar month (vv. 12, 13),

This content downloaded from 97.86.236.53 on Sat, 23 Jun 2018 01:58:15 UTC
All use subject to http://about.jstor.org/terms
348 ON THE SURYA SIDDHANTA.

The following is the composition of the "Great Age," an

imaginary period (vv. 15-17):?
Solar Years. Solar Years.
Dawn. 144,000
KritaYuga .. .. 1,440,000
Twilight .. .. 144,000
Total .. 1,728,000
Dawn. 108,000
TretaYuga .. .. 1,080,000
Twilight .. ., 108,000
Total .. 1,200,000
Dawn. 72,000
Dvapara Yuga .. 720,000
Twilight .. .. 72,000
Total .. 864,000
Dawn. 3G,000
Kali Yuga .. .. 3G0,000
Twilight .. .. 3G,000
Total .. 432,000
Total of Great Age 4,3

Furthermore, the Kal pa (^T^I) is thu

Solar Years. Solar Year
Dawn. 1,728,000
71 Great Ages .. 300,720,000
1 Twilight .. .. 1,728,000
1 Manvantara .. 308,448,000
14 Manvantaras .. .. 4,31
1 Kalpa . 4,320,000,000
The Kalpa is a day of Brahma. His night is of the same
and his age consists of 100 years, each of 300 such d
nights. The total duration is 311,040,000,000,000 sol
(vv. 20, 21).
The following is a computation of the time from the commence
ment of the Kalpa to the end of the present Treta Yuga
(vv. 21-23):?

This content downloaded from 97.86.236.53 on Sat, 23 Jun 2018 01:58:15 UTC
All use subject to http://about.jstor.org/terms
 ON THE 8URYA SIDDHANTA. 349
Solar Years.
Dawn of current Kalpa. 1,728,000
6 Manvantaras .1,850,G88,000
27 Great Ages . 11(5,640,000
TretaYuga. 1,728,000
1,970,784,000
But from the clapseel portion eif the \
present Kalpa there must be deducted
i, ,.time
the i ..occupied
/ <tA in> creation
17,064,000
(v. 24, ( 7
see also vv. 45-47) .. .. .. /
1,953,720,000

In their daily westward motion the planets lag behind the fixeel
stars each by the same absolute mean distance, viz., 11,858,717
yojaiuis (fJ^t*T0? ?ml their angular motion is inversely as the
radius of the orbit. The initial point of the sphere is the end o
the constellation Revati (vv. 25-27).
[The principal star of Rcvatf is saiel to be 10' W. of the .abov
mentioned point, and is supposed to be ? l'isciutn.]
The numbers of revolutions of the planets, &c, are as follow
(vv. 29-34, 41-44):?
In a Great Age. In a Kalpa.
Revolutions of the Planets. Apsides. Nodes.
Sun . 4,320,000 .. 387
Mercury. 17,937,060 .. 368 488
Venus. 7,022,376 .. 535 903
Mars . 2,296,832 .. 201 214
Jupiter. 3(14,220 .. 930 174
Saturn. 146,568 .. 39 662
Moon:?
Sidereal rev. 37,753,336
Apsis. 488,203
Node. 232,338
From the foregoing data the followi
(vv. 34-40):?
In a Great Age.
Sielereal elays .1,582,217,828
Deduct solar revolutions. 4,320,000
Civil days. .. 1,577,917,828
vol. xx. 2 B

This content downloaded from 97.86.236.53 on Sat, 23 Jun 2018 01:58:15 UTC
All use subject to http://about.jstor.org/terms
350 ON THE SURYA S1DDI1ANTA.

Sidereal solar years . 4,320,000

12

Solar months . 51,840,000

Melon's sidereal revolutions .. .. 57,
Deduct solar revolutions .. .. .. 4,
Synodical revolutions (lunar months) .. 53
Deduct solar months .. .. .. 51,840,000
Intercalary months .. .. .. 1,593,330
Lunar months x 30 = luuar days .. 1,G03,000,080
Deduct civil days.1,577,917,828
Omitted lunar days. 25,082,252
In order to find the number of civil days that have elapsed
since the creation, or any other given epoch, to a given date,
proceed first as in vv. 23, 24. Then (vv. 48-51) let?
Y =: No. of years to end of last Krita Yuga,
y = ? since Y,
m = complete solar months since y,
d = lunar days elapsed of current month.

Then the required number of lunar days

= 30(12 (V+,) + .?)0 + 5JS?) + "

And the corresponding number of civil duys

/A0,v,\ , n 2220389 w 25082252 >.
= { [12 fi + V) + '?;-72000- + rf}0 - 1G03000080)
/Ao/Vj xL -\ 2226389 ,1394479457

Suppose the planets were arranged in the order of their supposed

distance from the Earth, viz., Saturn, Jupiter, Mars, Sim, Venus,
Mercury, Moon. The first hour of the first day of the first month
of the first year was assigned to the Sun ; and so on for all the

This content downloaded from 97.86.236.53 on Sat, 23 Jun 2018 01:58:15 UTC
All use subject to http://about.jstor.org/terms
 ON THE SUJtYA SIDDHANTA. 351

other hours in tho order given above. The succession for the
days, months, and years will consequently fall as follows:?
Hours. Days. Months. Years.
1 1=7x0+1 1=7x0+1 1=7x0+1
2 25 = 7 x 3 + 4 31 = 7 x 4 + 3 36
3 49 = 7x7 61=7x8 + 5 721 = 7x103
4 73 = 7 x 10 + 3 91 = 7 x 13 1081 = 7 x 15
5 97 = 7 x 13 + 6 121 = 7 x 17 + 2 1441 = 7 x
6 121 = 7x17 + 2 151 = 7x21 + 4 1801 = 7
7 145 = 7 x 20 + 5 181 = 7 x 25 + 6 2161 = 7 x
Hence the following scheme of arrangement of planet
months, and years:?
Hours. Days. Months. Years.
1 Sun Sun Sun Sun
2 Venus Moon Mars Mercury
3 Mercury Mars Jupiter Saturn
4 Moon Mercury Saturn Mars
5 Saturn Jupiter Moon Venus
6 Jupiter Venus Mercury Moon
7 Mars Saturn Venus Jupiter
Hence, to find the planet of the day we have only to find the
remainder of I) (or the number of days)-4-7, and the planet opposite
tho corresponding placo in the days column will be the planet
required. Now the positions of the planets in the day colum
being of the form D, those in the month column are of the form
21) + 1, and those in the year column of the form 3 D + 1. Hence
q implying quotient, and r remainder, the formula; for finding the
planet corresponding to a given?
f?\
Day will be \j)r

Month ? (2\Mjq * * )
^ 7 yr,
Year ? (\MQSg + lJ
The mean place of the planets at any given time (No. of days
elapsed = D) is given by the formula jr, where?
R = No. of revolutions in an age (vv. 53-4),
C = No. of civil days in an age.
2 B2

This content downloaded from 97.86.236.53 on Sat, 23 Jun 2018 01:58:15 UTC
All use subject to http://about.jstor.org/terms
35ii on this HiiitYA kiihuiAnta.
To find the current year of the cycle of Jupiter (00 years); if
.1 bo the number of past revolutions and signs of Jupiter?

0?m,.tye?r = Q0)/.
The above method gives the mean places of the planets, &c,
for the prime meridian (through Ujjayini); wo now proceed to find
them for any other meridian (vv. 59, 00)?

Let n = radius of Earth = 1000 yojanas,

I = latitude.

Then /> *Ao = circumference of equator,

r ^10 cos. I = circumference of [larallel, whose latitude is /.

Let t, t' be the calculated anel observed lines eif immersion anel
emersion of a total lunar eclipse, then the correction for longitude
(and latitude) expressed in yojanas (vv. 03-05)

{t~t')r Vio -=
= i-f-- cos. I ,) .,.
(tr^t COS.,80I VlcT
00 v ' 3
And if n be the planet's m
for the meridian of the p

DR n(t~t')r *.'Tb
0 ~ GOr VlO cos. I 0-00
And, if /" = time before eir after midnight expressed in nadi's,
then the planet's mean position at that time will be expressed
by (v. 67)?
DR nT_
0 - 00
The orbits are however inclined to the elliptic as fo
(vv. 68-70):?
Moon .. .. .. .. .. 4 30
Mars .1 30
Mercury .. .. . . . ..20
Jupiter .. .. .. ... ..10
Venus .2 0
Saturn . ..2 0

This content downloaded from 97.86.236.53 on Sat, 23 Jun 2018 01:58:15 UTC
All use subject to http://about.jstor.org/terms
 ON THE HliltVA HIDDIIANTA. 353

chatteii 11.

On the True Motions op the Plakets.

The planets arc advanced or retarded, or diverted in latitude, in

various degrees, from their mean positions, by agencies situated in
their nrmides, nodes, &c. (vv. 1-14). To determine their true
positions, a Table of Sines is necessary. The intervals of arc feu*
which the sines are calculated are 225' ; then, if s, s', s", be the
sines of 225', 2x225', 3x225', we have the following rule fen
calculation :?
8 = 225'

b =a + *- ;
n , S S
*" =
S S S + 8

And tho Table given as the result

from the 7th to the end some m
is (vv. 15-27) :?

Arc. Sine. Arc. Sine. Arc. Sine.

o * I > lo f| I I O f I f
3 45 225 33 45 1910 63 45 3084
7 30 449 37 30 2093 67 30 3177
11 15 671 41 15 2267 71 15 3256
15 0 890 45 0 2431 75 0 3321
18 45 1105 48 45 2585 78 45 3372
22 30 1315 52 30 2728 82 30 3409
26 15 1520 56 15 2859 86 15 3431
30 0 1719 60 0 2978 90 0 3138

It must be remembered that the sine is a line,

consequently that sin. 90? = radius.

This content downloaded from 97.86.236.53 on Sat, 23 Jun 2018 01:58:15 UTC
All use subject to http://about.jstor.org/terms
354 ON THE SURYA SIDDIIANTA.
The sine of any arc not an exact multiple of 225' is given by
the formula (vv. 31-33)?

Bin. (? 225' + 0) = ^(Bi"-(?->-l)225;-Bi?.?225')j

ZZo

Similarly the arc might be found from the sine.

If S be the Sun, N its node, and N SII a spherical triangle
right angled at II, the Sun's declination D (= S II) is given by the
equation (v. 28) :?
R sin. D = sin. S N sin. S N II.

The corrections of the mean longitudes are made by means of

epicycles, the magnitudes of which vary in different parts of the
orbit. The dimensions are cxpresscel in arcs of the orbits to which
they belong, as follow (vv. 34-38) :?

Circumference of Epicycle

Planet. Of Apsis Of Conjunction

At even At odd At even At odd
Quadrant. Quadrant. Quadrant. Quadrant.
o o o o

Sun .. .. 14 13 40
Moon .. 32 31 40
Mercury .. 30 28 0 133 132
Venus .. 12 11 0 262 260
Mars .. 75 72 0 235 232
Jupiter ..33 32 0 70 72
Saturn .. 49 48 0 39 40

Let e0, cp be the magnitudes of the epicycle at the even and

quadrauts respectively, expressed as above, Se the correction
applied to e0 for any point whose mean anomaly is nt; let
radius of circular orbit; then?
fo _ sin. nt
ex ? e0 ~" R '
and the circumference of the epicycle e at that point is?
sin. nt, .
<" = o, ? , (ei ? e0).

This content downloaded from 97.86.236.53 on Sat, 23 Jun 2018 01:58:15 UTC
All use subject to http://about.jstor.org/terms
 ON THE SURYA 8IDD11ANTA. 355

The following is the calculation of the correction for the apsis

(v. 39). Let E be the Earth, and API* the undisturbed circular
orbit; A a the radius of the _./>
epicycle at A, V p the radius
of the epicycle at P; draw
n m, Q q, ?r to E p. The
movement is represented by
supposing the epicycle to
move with its centre on the
circle A P R, without revolv
ing about its centre. Conse t?_?1/ I_a] Y
quently, the radius A a, or
its equivalent P n, is parallel
to E A; in other words, / p P n = i P E A = 0. The point Q,
in which E n cuts the circle APR, is the true position of the
planet; and P Q is the correction sought.
Since the circumferences of circles are as their radii,?

300? _ sin. Q
6 sin. p Pn'
And when, as in the caso of the epicycles of the apsis, e is small,
we have approximately?
m n = q Q = arc P Q.
Hence the correction for the apsis?
5 0 = _JL. Bin. 0.
300?
Rut if, as in the case of the conjunction, the epicycle is not small,
we have (vv. 40-42)?
Al' = US, E ro1 + wn? = E?\
n ?m CJ q
Hence, if 8t 0 be tho correction for the conjunction,?
^ in m
sin. St 0 = VE
Q m*
7 =-f^wR.
?r

The correction for the apsis is the only one required for the Sun
and Moon. For the other planets, calculate (1) the correction of
conjunction, and apply half of it to the mean place; thence (2) calcu
late that of the apsis, and apply half of it to the place already
corrected ; thence (3) calculate that of the apsis afresh, anel apply
it to the original mean place of the planet; and lastly, thence (4)

This content downloaded from 97.86.236.53 on Sat, 23 Jun 2018 01:58:15 UTC
All use subject to http://about.jstor.org/terms
356 ON THE SURYA SIDDHANTA.

calculate that of the conjunction, and apply it to the last place

(vv. 43-45).
The part of the equation of time, depending upon the difference
between the Sun's mean and true places, is given iu minutes by
the formula (v. 46)?
O's equation x n
36007
To calculate the correction 8 n eif the mean daily motion n of a
planet due to the influence of the apsis ; let v be the mean motion
of the apsis, then?
6* sin.
/i (n0? =v)V-1-.-;
B tabular sin. 0
225'
Also, if * be the arc of the epicycle (e), corresponding to 0 in tho
orbit,?
360? _ 3 sin. 0 _ 3 sin. c9
e "~ & sin. ? ~~ B n '
Whence? . c (n=?- p)
o n $ tab. sin 0
i-1-.
360 225

To calculate the corresp

correction 8t n, due to the
of conjunction (vv. 50,
the following figure?
Let E be the Earth,
M' M the correct
motion of the centr
the epicycle at M',
m' m points in the epi
corresponding to in
Make E n == E m;
Join M in, and M m' s
Then in the Ab Mint, z M tm = rt I, . . M
inc'iri, Z 7/1 o'm' = rt /. = m iri o,
.*. the A8 {ll'<s similar.

And since the epicycle is small compared wi

approximately t s = o' m'.
. o iri _ t m _ t s
m m' M in in iri'
, t s _ in m _ MM'
7~n~i M~m ~~ "EAf

This content downloaded from 97.86.236.53 on Sat, 23 Jun 2018 01:58:15 UTC
All use subject to http://about.jstor.org/terms
 ON THIS SURYA SIDDHANTA. 357
The text, however, substitutes Em for EM in the above expression
without explanation; so that?
?, n = M M'1<jA-Il
m

When the commutation in the* final process of 43-45 is between

the follervving limits, the motion becomes retrograde (vv. 52-55):?
Mercury .. 144? to 215?
Venus .. .. 163? ? 197?
Mars .. .. 164? ? 426?
Jupiter.. .. 130" ? 230?
Saturn .. .. 115? ? 245?
To find the latitude of a planet (vv. 56-58). Subtract from the
mean [dace of the planet, corrected for the apsis only, that or its
node ; then, if?
V = distance of planet from its node,
L = extreme, or greatest latitude,
V = latitude at mean distance R,
I = ? true ? ;*;
We havo for the Moon?
,_ 7l_ Ij sin. V
?_ I-& ;
And for the other planets?
r V , , L sin. V
r- = ?, whence I =
R / r

This is to be added to or subtracted from th

the difference between arcs measured em se
and eeliptic.
The day anel night of a planet, or the
twice over tho same meridian, differs from
by a quantity thus found. Each sign cont
number of respirations occupied by the
which the planet is, across the meridian (det
the quantity required (v. 59)?
np
1800'.
The radius of the heiur circle (v. 60)
~ R ros. I) = R ? R vers. D.

This content downloaded from 97.86.236.53 on Sat, 23 Jun 2018 01:58:15 UTC
All use subject to http://about.jstor.org/terms
 358 ON THE SURYA SIDDIIANTA.

To find the length of the day of a planet, or tho time during

which it is above the horizon (vv. 01-03). In tho following figure?
Let N P Z S represent the meridian
of an observer at 0,
P P' the N. and S. poles,

/ \\ -A
E E' the points of the meridian
cut by the equator.

Si_J_|J )>t^?_N E1) the declination of the planet.

Draw D R A parallel to E 0,
E IL 1/ to N S,
P A G the arc of a great circle
through P and A.
Then the A s, A R 0, 0 II E, are
similar.
EII_OR_^
- ' II0 "" R A a'
And? CB = DF = sin. D.
Hence- A R = S eilh D*.
Rut the arc, of which A R is the sine, is the same part
diurnal circle that the arc, of which 0 0 is the Bine, is o
equator.
. A R _ GO
* ED ~ 0 E'
which determines G 0. And the arc, of which G C is the sine, is the
measure in time of the difference between a quadrant and tho arc
of a diurnal circle intercepted between the horizon and meridian.
The ecliptic is divided into 27 lunar mansions, each of which
consequently contains 800'. Hence, in order to find in what man
sion a planet is at any given time, let 0 = its longitude.

No. of complete mansions traversed = (?^)

\o00 J{I'

Portion traversed by current mansion = ( ? ~,)

No. of days elapsed = -(? )

J l n\8007'
A lunar day is ^ of a lunar month, or of the interval in which
the moon gains 300? in longitude on the Sun. It is therefore
measured by -s/^? = 12? = 720'. Hence wo may find the number
of lunar months and days elapsed by proceeding as above with a
divisor 720'.

This content downloaded from 97.86.236.53 on Sat, 23 Jun 2018 01:58:15 UTC
All use subject to http://about.jstor.org/terms
 ON THE SURYA S1DDHANTA. 359

Tho yoga (^n"Ti:) (;/) is tho perioel during which ilio lemg
itudcs eif the Sun (0) anel eif the Moon (0,) together amount to tho
Bpacc of a lunar mansion (v. 65).
. *. J?i,?
800 = No. of yogas passed + portion of current yoga
= qy + z suppose.
Then if n, n? be the daily motions?
-_ =- nidi's elapsed of current yoga.
n, + n
Similarly for the lunar days (rf,)?
?i~0 .
~lW = ** + *'
_ = nadi's elapsed of curren
nt ? n
Each lunar day is divided into two halves (qi^*!!!), which
have particular names and portions assigned to them. But they
appear to have no practical use.

Chaptek III.

On Direction, Place, and Time.

To construct the dial (vv. 1-7).

Describe a circle ; at its centre erect a gnomon 12 digits high.
Mark the two points where in the forenoon and afternoon the extre
mity of the shadow meets the circle. _ N_
From each point as a centre, with
the distauce between them as a
radius, describe two circles; the
line joining their points of section
will be a N and S line. Ry similar
processes draw E and W, NE and
SW, NW and SE lines. Draw a
circumscribing square, and mark off
the sides passing through the E
and W points in digits. Let e be s
the extremity of the shadow at
noon, when the Sun is on the equinox ; t
E W.

This content downloaded from 97.86.236.53 on Sat, 23 Jun 2018 01:58:15 UTC
All use subject to http://about.jstor.org/terms
360 ON THE SURYA SIDD1IANTA.
Then (v. 8) if g = height of gnomon,
/* = length of hy[iothcnuse,
s = ? ce[iiinoctial shadow,
h* = g2 + s\
z
In a yuga the sid
lags behind em the zo
lutieins. And the pos
initial point of the
period is consequentl
following proporti
When d = No. of d
</,== ? ina
^ COO rev.
d x

But it is an oscillatory movem

of sign like a sine; only the
90? x 30 _ 2?o
10
The annual precession conseque
_. 365 x 360? x 36
4320000 x 365 10
In the accompanying figure le
C b represent the gnomon
be ? equinoctial sh
b d ? any other ,, = s',
C e ? equinoctial hypothenuse = h,
C d ? any other ? = h',
C E ? radius = R,
d e ? measure of amplitude = an
Z ? zenith,
S ? south point.
Then if the Sun be upon the equator?
sin. / = B E = i^,
h

cos. /=CB h=J^.

If it be not on the equator?
sin. Z=BE,
sin. A. = 0 B,
and .90? ~ L = Z + D.

This content downloaded from 97.86.236.53 on Sat, 23 Jun 2018 01:58:15 UTC
All use subject to http://about.jstor.org/terms
 ON TUB SURYA SIDDIIANTA. 3G1

Similarly from the above equations we may find the shadow or

hypothenuse for a given latitude or declination.
The true and mean longitudes may be found, when the latitude
and declination arc known, by an inversion of the processes of
ii 28-30.
If a represent the Sun's amplitude, then CA = sin. a; and from
the similar As 01 b e, OF A, we have?
// sin. D
sin. a =
9
Also from the similar As Ode, D C A, we have?
h Bin. D h'
ax = - ....;
(j R,
and s = s ? #!.
Returning to the figure?
Z 0 will be the projection of the prime vertical,
V ? ? point where the Sun passes it.
Then V C = sin. A.
And generally sin. A = ?r^.
Now since the As V 0 F, E 0 II arc similar,
A RA
sin. sin.
= D?:
riii. e

Hence if h" be the value of the hypothenuse when the su

on the prime vertical?
sin. D _ g
sin. e h"'
Also since?
sin. / ___ s
cos.I g
j" ? s cos* I
' ' l ~~ muTTF
Again, since the As II E 0, F C A arc similar?
sin. a =-.
R sin. D
cos. /
,.ri . oox ft1 sin. a
Whence (v. 28) a,
I* = ?

To find, for a given declination and latitude, the sine of th

Sun's altitude at the moment when it crosses the SE or the SV
vertical circle.
Sup]lose that the plane of the above figure is turned about C
ns an a?cis through 45?, ro that Z 1) E S represents an arc of SE, or

This content downloaded from 97.86.236.53 on Sat, 23 Jun 2018 01:58:15 UTC
All use subject to http://about.jstor.org/terms
362 ON THE SURYA SIDDHANTA.

of a SW circle. Then b e = V2 s, D E' = v7? sin. a. The alti

tude is then given by the formula?

Bin A = + 2^Bin'fl 4- /(Ba - 2 sin.'q) g* 4 ?? g* sin.? a

- ?* + 2 *a V <7a + 2 s* <y + 2 **)* *
To verify this, form the quadratic equation in sin. A?
(*/* + 2s,)sin.,A ? 4 se7 sin. a sin. A + 2g2 sin.-V* ? RV = 0,
or
2 (s sin. A ? g sin. a)' ? g2 (R* ? sin.1 A) = 0,

or, referring to the figure?

(be.VK-Gb.V E'y -Ci'OK^O.
But
C K
DK+ D
C E'
o = *i, or&e.DK-C*.DE' = CK
which renders tho equation identical.
If the declination be south, D E' falls in
hence the double sign.
To find the sine of the Sun's altitude
distance from the meridian, the declina
known (vv. 34-36).
In the same figure as before, let 0 be
Sun's place at any time, and POQ that of
P A G that of another.
Then
CO = sine of ascensional difference = sin. a.
E Q = versine of hour angle = versin II.
0 R = sin. A.
R, = radius of diurnal circlo.
Then

sin. A = (R + sin. a - versin II) |? ^?J.

_ G Q . F D . E II
~ 0Ea
But?
E II _ OR , DJ*" _ G_Q . 6Q.FD.EH _ 0
TTE ~ 0 A 0 E "~ A 0 ' ' 0 Ea
In a given latitude to find the Sun's declination, and thence it

This content downloaded from 97.86.236.53 on Sat, 23 Jun 2018 01:58:15 UTC
All use subject to http://about.jstor.org/terms
 ON THE SURYA SIDDIIANTA. 363

true and mean longitude from the shadow at any hour (vv. 40-61).
It was shown above that?
h _ E 0 _ E II
a CA F 0

h
To describe on the dial the path of the extremit
for any day, set off three liases in the forenoon, n
(,y-co-ordinates); calculate the distances E. and W
and draw a circle through the three points. Th
path required (vv. 41, 42).
To determine the time occupied by each sign
for a point on the equator. Let R? R2, R3 be t
1, 2, 3 signs respectively. Then if tx, t2, tB, be th
and s,", 52", s8" the number of seconds in 30?, 60?
then?

Bin. tx = R? 7' *'", Bin, t, = R?

1*,|Bi"-
lij lv3s>", Bin. /. = R' ?" *"

For the two sets of quantifies (s,", s3", 5,"), {tl9 t2, h), being
measured, one on the ecliptic, and the other on the equator, form
respectively the hypothenuscs and bases of three rt. I. d As; and
by ordinary spherical trigonometry?

sin. s" = cos. s,'' t, sin. tt = fJ sin. L.

Ra
For the next three signs the expressions will be the same as
these, only in an inverse order; and so on for the other six.
For a point neit on the equator we must add (or subtract) the
ascensional difference.
Given the Sun's longitude and the local time, to find the points
of the ecliptic on the horizon and on the meridian (vv. 4G-49).
The preceding method gives the ascensional equivalents for the
various signs; and for the portions of signs, in which either the Sun
h S " v"
or the horizon is, we have ?2- = -1?
O tq t%
A similar process will give the local time at which any given
point on the ecliptic will be upon the horizon (vv. 50, 51).

This content downloaded from 97.86.236.53 on Sat, 23 Jun 2018 01:58:15 UTC
All use subject to http://about.jstor.org/terms
 364 ON THE SURYA S1DDHANTA.

Chapter IV.

On Eclipses.

The diameter of the Sun = 6,500 yojanas (v. 1).

? Moon = 480 ?
Let, as before, n, nv represent the mean, and n
daily motions of the Sun and Moon respectively;
n'
Corrected diameter of the Sun = 6500 -?
n

? Moon = 430 !ti.

?i
In order to find the apparent diameters in minutes of arc, the
corrected diameter of the Sun is projected on a circle at the Moon's
mean distance, by multiplying the expression by the ratio of the
Sun's revolutions in an age to those of the Moon, or by that of tho
Moon's to the Suivs orbit. At the distance in question 1' = 15
yojanas (vv. 2-3).
Corrected diameter of the Earth = 1600 'ii-.
nx
To find the diameter of the Earth's shadow upon tho Moon's
mean orbit. Project the difference of the Sun's anel the Earth's
corrected diameters on the Moon's orbit, and subtract the result
from the Earth's corrected diameter (vv. 4, 5). The formula is?

1G00 <7i,
- ( 6500
VVin
- HJ00
nt<\J??.
J 6500
Calculate the longitudes of the Sun aud Moon at midn
preceding or following the opposition or conjunction; t
eclipse be probable, calculate the interval to the instant of
or conjunction, by the methods of Chapter 11 (vv. 6-9).
If ; ? 7*, be the radii of the eclipseel and eclipsing bodie
the latitude of the former, the amount of the obscuration
by the formula (vv. 6-11)?
rx + r - /,.
The times of duration of the eclipse, and of total obscuration
will be expressed by (vv. 12, 13)?

60 being the number of nadia in a day.

This content downloaded from 97.86.236.53 on Sat, 23 Jun 2018 01:58:15 UTC
All use subject to http://about.jstor.org/terms
 ON THE SUItYA SIDDIIANTA. 3C5

This method assumes that the latitude remains unchanged

during the eclipse; but if greater accuracy is required, with the
above formula as a first approximation recalculate both longitude
and latitude of the Moon; aud repeat the process as often as
desired (vv. 14, 15).
The instant of true opposition or conjunction is considered as
the middle of the eclipse (vv. 16, 17).
If from the formula of vv. 12, 13, corrected by vv. 14, 15, we
subtract any interval of time {t), and reconvert the remainder into
arc, we may regard the result as the perpendicular, and the latitude
as the base of a right-angled triangle; the hypothenuse will then
represent the amount of obscuration at the time f. In the case of
a solar eclipse, a correction for parallax during its continuance must
be made. This is explained by Chapter V, vv. 14-17 (vv. 18-21).
Cemvcrsely we may rccpiirc to know when the obscuration will
attain to a given amount. The method, being similar to those
given above, need not be given in detail (vv. 22, 23).
In projecting an eclipse (a process which is explained in
Chapter VI), the eclipseel body is represented in the centre of the
figure with a N. and S. line, and an E. and W. line drawn through
it as co-orelinates, or lines of reference. The N. and S. line repre
sents a great circle drawn through the N. and S. points of the
horizon; the E. and W. line a small circle parallel to the prime
vertical. The position of the ecliptic is fixed by calculating, first,
the angle (t>) between the E. and W. line and the circle of diurnal
motion; and secondly, the angle {iv) between the latter circle and
the ecliptic.
For the first process let P be the pole of the equator, M
the eclipsed body, N the north point of the horizon. Then
P N = I, MPN = 180? -hour angle
= 180? -II.
If M were on the horizon anel at the E. or W. point, then P M and
N M would be quadrants, anel PMN= v. It is, however, assumed
that the same relations would remain approximately unchanged for
other positions of M; hence for the triangle P M N, we have?

sin. M P N = sin. P M N of si|i y _. sin. II sin. /

sin. M N sin. PN ' R
Secondly, it seems supposed that the diurnal circle and the equator
meet at 90? from M; hence moving M to a point M', 90? forward
vol. xx. 2 C

This content downloaded from 97.86.236.53 on Sat, 23 Jun 2018 01:58:15 UTC
All use subject to http://about.jstor.org/terms
366 ON THIS MUJIYA HIDDHANTA.

on the ecliptic, the declination of M' will measure the angle between
the evliptie and the eliurnal cirede.
The* sine of the dellectiou (v + w) se) found is laid off on a straight
line on the scale of radiiiB = 49 digits; i.e., 3438 ? 49 = 70, or
70' = 1 eligit.
To take; account of the apparent increase of heavenly bodies
near the horizon ; it is assumed that 3' at the horizon are equivalent
to 4' at the zenith. Hence it is calculated (v. 26)?
i day _ altde. in time
1' excess over 3'
Whence the rule?
n . i . c vof..digits
Equivalent , r "-It- i|1 time
in minutes + =-.---=-^-?.
of arc 3i days
i day

Chapter V.

On Parallax in a Solar Eclipse.

When the Sun is on the meridian, it is considered that there

is no parallax in longitude. When the latitude is eepial to the
eleclination, there Jb none in latitude (v. 1).
The first step iowarels finding the parallax at the moment of
conjunctieiu, is te) determine the sine of amplitude of the point of
the ecliptic on the eastern horizon (vv. 2, 3). For this purpose, let
D, = the greatest eleclination ; then, adopting the notation hitherto
used, we have by ii, 28?
R sin. 1) = sin. 0 in. sD?
And by iii, 22, 23?
R sin. 1)
sin. a = -;
cos. /
Whence?
?... ? = Bi"-gBi"i?..
cos. /

To find the sines of the Z 1) and altitude of the point of tho

ecliptic having the greatest altitude (vv. 4-6),?

This content downloaded from 97.86.236.53 on Sat, 23 Jun 2018 01:58:15 UTC
All use subject to http://about.jstor.org/terms
 ON TIIK SURYA SIDDllANTA. 367
Let Z' = meridian Z D,
C| = Z I) eif point iu question,
A, = Altitude ?
Then if, iu tho spherical triangle Z R L, Z be the zenith, Z L a
meridian, Z L JL to the ecliptic B L, and M n
the arc of a great circle from M to the pede of /\\
the ecliptic, / \\
ZL = Z', / \
Z B = Z., B \^/_Wjv
Z L B = 90? - B Z L, approximately, \y
m ii = parallax iu longitude, m
And?
sin, z, _ sin. Z LB__ sin.ZLB sin. (90?-a) _ VR'-
sin. z7~* BinTZ irii"~~ R ~ "R """ R
And? sin.2Z, + sin.'A, = R;
Whence? _
0:0 7 _ / . , , sin.'z' sill.'a
sin. Z, - ^J flln.?z - -__
sin. Ai = ^RJ - sin. 2Zr
But (v. 7) we may approximately take Z, = z'.
To find the parallax in longitude (vv. 7-9). The Moon's greatest
horizontal parallax = 4 midfs. Hence the proposition?
sin. Z m R
~M in "~ 4*
But? sin. Z m . siu.Bm __ sin. A,
M m ' ii m R
. sin. Bin sin. Ax ___ R
;i?/i R 4

But 5 = sin. 30? ; hence?

? ,, . , ... sin. B m
Parallax in longitude = -.?-?r
b Bin. -30? -+ sin. A,
[The term used in the text seems rather
of B m.]
The formula must be useel to correct the time of conjunction
previemsly found; the parallax must then he calculated afresh, anel
the pre)ccss repeated.
202

This content downloaded from 97.86.236.53 on Sat, 23 Jun 2018 01:58:15 UTC
All use subject to http://about.jstor.org/terms
368 ON THE SURYA S1DDIIANTA.

To determine the parallax in latitude (v. 10), we have only to

substitute, from tho formula)?
n
Sun's greatest parallax = t;
Moon's ? = 315,

the quantity Hl ~~n for 4 in the equation -1~??

15 M = - ;in
whence
4

rnrallax in latitude
15 R
= &^?) 1?!.

This formula may be simplified for calculation by the following

considerations (v. 11):?
nt ? n = 731' 27"
15 R = 51570',
and?
51t57Q'?27"
731' ? 70J *
= R-49
nearly.
J
Hence, approximately?
v n . i in
Parallax 4-llatitude
i Sin.Z1
= ??49Sin.Zl
l, R
or = --?!.
70

With the value of the par

to be corrected. The para
beginning, middle, and en
calculated; the quantiti
Pi ? Px, Pi ?
are to be added as correc
determined.

Chapter VI.

On Planetary Conjunctions.

To find whon two planets will have the samo longitude (vv. 8-0).
Let 0', 0\, be their longitudes,
ii, n? their daily motions.

This content downloaded from 97.86.236.53 on Sat, 23 Jun 2018 01:58:15 UTC
All use subject to http://about.jstor.org/terms
 ON THE SURYA SIDDIIANTA. 3G9

Then they arc distant from tho point where they will have the
same longitude, respectively,?

fiLziA)-,
n + 7i, nand iLzM?.
? n,
To find the moment of conjunction, i.e., when they will bo on th
same secondary to the ecliptic (vv. 7-12).
Lot V, S, bo the two planets having the n
same longitude,
N the* north point eif the horizon,p \\
P, P', the poles eif the cepiator and
ecliptic,
P S, P V, great circles from P, cutting
the ecliptic in s and v,
N S, N V, great circles freim P, cutting
the ecliptic in s' and v',
then the two planets are removed from con s

junction by the distance v s. To determine

this, find?
M v + M s = v v ? M v' + s s'
Let ss0 be the value of s s when S is on
is an angle which = 0? when the pole
= 90? when it is at the zenith; and is on t
vary with the elevation of the pole ; in o
= the latitude (L) of the observer. And S
near the ecliptic, s S s is regarded as a pl
angle s' s S = 90?. Hence?
ss0 __ sin. L _ s
I cos. L g
Again, for any other position of S, we have the propo

s s' : s s'0 = mcrid. dist. in time : ? day,

or?
Is t
ss = ? ??.
9 i^y
In the same way, in the A MS s, the angle at s' is considered
as = 90?, and consequently?
M S _ R
M s siu. M S s'

This content downloaded from 97.86.236.53 on Sat, 23 Jun 2018 01:58:15 UTC
All use subject to http://about.jstor.org/terms
370 ON TUB SURYA S1DD11ANTA.

Rut since M is suppeise*d always near the ecliptic, M S s = P' S P

= P' M P, nearly; and if D' be the declination of a body 90? in
advance of M,?
MS _ R,
M s sin. 1)''
Rut
Sine of greatest declination = sin. 24? = 1397' = 5
whence it is concluded that sin. D' = 58 x D':
also radius = 3438' = 58 x GO'nearly;
whence? *,- Ms
, / =IV
?.
00
In the same way v v, and M v' may be found; .and
pletely determined.

This content downloaded from 97.86.236.53 on Sat, 23 Jun 2018 01:58:15 UTC
All use subject to http://about.jstor.org/terms