Академический Документы
Профессиональный Документы
Культура Документы
With 17 figures
186..222
DAVID KENFACK*
Center for Tropical Forest Science, Arnold Arboretum, Harvard University, 22 Divinity Avenue,
Cambridge, Massachusetts, 02138, USA
Received 28 January 2010; revised 7 October 2010; accepted for publication 22 October 2010
The taxonomy of the amphi-Atlantic tree genus Carapa (Meliaceae) has long been controversial. Of the three
species currently recognized in the genus, two are known to present substantial morphological variation that has
been used in the past to distinguish several taxa, most of which are currently placed in synonymy. Here, a
combination of field observations, univariate analyses of leaf, floral and seed characters and principal coordinate
analyses of floral characters in the context of a molecular phylogenetic analysis was used to investigate the
patterns of variation and delimit morphological species anew in the genus. These results support the recognition
of 27 species in Carapa, of which 16 are previously described and 11 are new. In general, phylogenetically related
species occurred in the same geographical area, but were morphologically distinct. 2011 The Linnean Society
of London, Botanical Journal of the Linnean Society, 2011, 165, 186221.
INTRODUCTION
Species form the basis of analyses of biogeography,
ecology and conservation biology (Sites & Marshall,
2004), but many remain undescribed. With the
increasing disturbance and destruction of natural ecosystems, there is an urgent need to document and
catalogue new species before they become extinct. For
centuries, systematic biologists, particularly monographers, have relied on morphological characters for
diagnosing and delimiting species. However, speciation is not always accompanied by clear morphological
differentiation. Morphology in some instances is difficult to use in delineating species, especially in
so-called cryptic species, i.e. morphologically similar
populations that are probably, or at least to some
extent, reproductively isolated. Cryptic species are
often found in poorly studied species complexes
or typically widespread morphologically variable
species. Cryptic species are now being uncovered with
the aid of molecular techniques (e.g. Dick, Abdul*Corresponding author. E-mail: kenfackd@si.edu
186
2011 The Linnean Society of London, Botanical Journal of the Linnean Society, 2011, 165, 186221
MORPHOMETRICS OF CARAPA
187
morphological (Pennington & Styles, 1975), anatomical (Kribs, 1930) and molecular data (Muellner et al.,
2003) place Carapa in subfamily Swietenioideae.
Within this subfamily, the septifragal capsule and
unwinged seeds with a woody testa are unique distinctive features of this genus. In the past, questions
have been raised about the generic boundary between
Carapa and the morphologically similar mangrove
genus Xylocarpus Koen. In fact, taxa currently placed
in the two genera were treated under Carapa by
Lamarck (1785) and de Candolle (1878), and it was
only in 1896 that Harms reinstated Xylocarpus
(Harms, 1896). Since then, while there has been
agreement concerning the generic boundaries of
Carapa, there has been substantial disagreement
over the number of species recognized within the
genus and their circumscription.
In his monograph of Meliaceae, de Candolle (1878)
recognized six species of Carapa, of which two were
later transferred to Xylocarpus. By 1917, seven additional species had been described in the genus from
the African rainforest. Harms (1940) in the second
edition of Die natrlichen Pflanzenfamilien, recognized 11 species of Carapa, and Staner (1941), in the
treatment of Meliaceae of the Belgian Congo (DR
Congo), indicated 14 species of which five were found
in tropical America and nine in tropical Africa. The
most recent and comprehensive treatment of the
genus was that of Noamesi (1958). In his revision of
2011 The Linnean Society of London, Botanical Journal of the Linnean Society, 2011, 165, 186221
188
D. KENFACK
MOLECULAR
DATA
DNA was extracted from leaf tissue (silica gelpreserved or herbarium material) for one representative of all but three of the morphospecies (Table 1)
and representatives of 15 other genera of Meliaceae.
The amplification of the entire internal transcribed
spacer (ITS) region of the ribosomal DNA (ITS1-5.8SITS2) was carried out by polymerase chain reaction
(PCR), using the primers ITS4 (White et al., 1990)
and LEU1 (Baldwin, 1993). After sequencing in both
forward and reverse directions, phylogenetic analyses
were performed using maximum likelihood (ML) and
all but one of the remaining 13 genera currently
recognized in Swietenioideae and three of Melioideae
as outgroups. For the ML analysis, the GTR+I+G
nucleotide substitution model was chosen using the
Akaike information criterion (AIC) in Modeltest
version 3.06 (Posada & Crandall, 1998, 2001) as the
best fit for our data set. Heuristic searches were
carried out in PAUP* with tree bisection
reconnection (TBR) branch swapping and Multrees
option in effect. Clade support was estimated using a
likelihood bootstrap (LB) of 100 replicates with TBR
branch swapping and Multrees option in effect. The
best tree obtained from this analysis was used as a
guide for partitioning morphological data in the morphometric analyses.
MORPHOLOGICAL
DATA ANALYSIS
2011 The Linnean Society of London, Botanical Journal of the Linnean Society, 2011, 165, 186221
MORPHOMETRICS OF CARAPA
189
Figure 2. Variation in leaf size, number of leaflets and leaflet shape in nine morphospecies of Carapa. Numbers
correspond to the codes of the morphospecies in the multivariate analysis.
2011 The Linnean Society of London, Botanical Journal of the Linnean Society, 2011, 165, 186221
Figure 3. Variation in flower colour of 12 morphospecies of Carapa. Numbers correspond to the codes of the morphospecies in the multivariate analysis.
[Photographs: G and L, Pierre-Michel Forget; remaining, David Kenfack].
190
D. KENFACK
2011 The Linnean Society of London, Botanical Journal of the Linnean Society, 2011, 165, 186221
Figure 4. Variation in fruit shape, surface ornamentation and colour in Carapa: Numbers corresponds to the codes of the morphospecies in the multivariate
analysis. [Photographs: J, Charles Doumenge; M, Sainge Moses; O, Mathieu Gueye; P and Q, Pierre-Michel Forget; remaining, David Kenfack].
MORPHOMETRICS OF CARAPA
2011 The Linnean Society of London, Botanical Journal of the Linnean Society, 2011, 165, 186221
191
192
D. KENFACK
Table 1. List of quantitative and qualitative characters assessed for the study of morphological variation in Carapa
Leaf characters
Floral characters
Qualitative data
Leaflet texture
Leaflet acumen (presence absence)
Petiole and rachis indumentum (RACIND)
Leaflet lamina indumentum
Inflorescence length
Indumentum of the inflorescence
Indumentum of the pedicel (PEDIND)*
Indumentum of sepals (CALIND)*
Colour of sepals
Indumentum of petals
Colour of petals
Colour of the nectary
Ornamentation of the surface of the fruit
Colour of the mature fruit
Ornamentation of the seed coat
material. Carapa flowers are imperfect, with functional organs of one kind (stamens or ovary) and
developed, but non-functional organs of the other
(Pennington & Styles, 1975). Furthermore, the two
flower types are easily distinguishable. Morphometric
comparisons can be made only between flowers of the
same type, and carpellate flowers represented only 37
and 17%, respectively, of the tetramerous and pentamerous flowers dissected. Because of this paucity of
carpellate flowers, measurements from staminate
flowers were used primarily in the analyses presented
2011 The Linnean Society of London, Botanical Journal of the Linnean Society, 2011, 165, 186221
MORPHOMETRICS OF CARAPA
193
Figure 5. Variation in seed coat colour, ornamentation and hilum size in 16 morphospecies of Carapa. Numbers refer to
the denomination of the morphospecies in the multivariate analysis.
used in the morphometric analysis for reasons enumerated below. Each specimen was scored for as
many of the 72 characters in Table 1 as possible and
assigned to a morphospecies (see above) numbered 1
to 27.
2011 The Linnean Society of London, Botanical Journal of the Linnean Society, 2011, 165, 186221
194
D. KENFACK
GEOGRAPHICAL
DISTRIBUTION
TAXONOMIC
DECISIONS
The primary goal of reassessing patterns of morphological variation in Carapa was to look for morphological discontinuities among groups of specimens
(morphospecies) that could allow one to distinguish
them from other such groups. By and large, the
decisions were based on the PCoA of staminate
2011 The Linnean Society of London, Botanical Journal of the Linnean Society, 2011, 165, 186221
MORPHOMETRICS OF CARAPA
flowers. However, in some instances, results from the
univariate analysis of leaf and seed characters and
other qualitative characters, such as flower colour,
fruit shape and seed coat ornamentation, were used
to supplement the results of the multivariate analysis. In making decisions as to whether these groups of
specimens represented species, as far as possible a
combination of morphological, molecular and spatial
data was used to distinguish separate species. Groups
of specimens that (1) showed strong morphological
discontinuities (e.g. non-overlapping number of leaflets, sessile vs. pedicellate flowers, different number
of ovules per locule, etc.), (2) had overlapping ranges
and/or occur in sympatry and/or (3) belonged to different well-supported clades of the ITS phylogenetic
tree were considered as belonging to separate species.
RESULTS
GLOBAL
MORPHOLOGICAL VARIATION
195
MORPHOLOGICAL VARIATION
AMERICAN SPECIMENS
IN
2011 The Linnean Society of London, Botanical Journal of the Linnean Society, 2011, 165, 186221
196
D. KENFACK
Table 2. Number of significant differences (sig. diff.) resulting from multiple comparisons (using Tukeys post hoc
analysis) of means of leaf, floral and seed quantitative characters among 27 morphospecies of Carapa from Africa and
America
All specimens
American specimens
Variables
d.f.
Number of
sig. diff.
Per cent
Leaf
BLAW
BLL
BLL/TLL
BLPET
BLV
BLW
GNUM
LL
LNUM
MLAW
MLL
MLPET
MLV
MLW
PETL
PETL/LL
RACD
RACL
TLAW
TLL
TLL/TLW
TLPET
TLV
TLW
24
24
24
24
24
24
24
24
24
24
24
24
24
24
24
24
24
24
24
24
24
24
24
24
95
97
43
72
103
85
48
99
127
91
88
67
88
100
95
118
41
103
94
71
38
37
84
88
29
30
13
22
32
26
15
31
39
28
27
21
27
31
29
36
13
32
29
22
12
11
26
27
Floral
ANTL
ANTW
CALD
NECD
NECH
OVAL
OVAW
PEDL
PETAL
PETAW
PISL
SEPL
STATD
STATL
STATLL
STID
STYL
25
25
25
25
25
25
25
25
25
25
25
25
25
25
25
25
25
116
100
165
109
98
40
57
98
161
140
104
149
88
158
171
124
68
Seed
HILL
HILL/SEEL
HILL/SEEW
HILW
HILW/HILW
SEEH
SEEL
SEEW
18
18
18
18
18
18
18
18
128
131
129
105
87
82
105
100
African specimens
Number of
sig. diff.
Per cent
d.f.
Number of
sig. diff.
Per cent
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
24
21
2
14
14
19
4
17
19
18
23
15
19
25
11
2
14
17
23
19
17
15
22
22
53
47
4
31
31
42
9
38
42
40
51
33
42
56
24
4
31
38
51
42
38
33
49
49
15
15
15
15
15
15
15
15
15
15
15
15
15
15
15
15
15
15
15
15
15
15
15
15
19
24
28
21
22
14
28
39
52
26
23
19
14
27
38
64
13
39
27
26
2
5
15
34
16
20
23
18
18
12
23
33
43
22
19
16
12
23
32
53
11
33
23
22
2
4
13
28
39
33
55
36
33
13
19
33
54
47
35
50
29
53
57
41
23
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
24
17
28
19
20
11
20
39
29
27
20
28
24
33
32
15
12
53
38
62
42
44
24
44
87
64
60
44
62
53
73
71
33
27
15
15
15
15
15
15
15
15
15
15
15
15
15
15
15
15
15
36
40
58
49
33
16
11
13
53
50
37
51
21
48
27
56
21
34
38
55
47
31
15
10
12
50
48
35
49
20
46
26
53
20
75
77
75
61
51
48
61
58
8
8
8
8
8
8
8
8
31
28
28
21
15
24
32
29
86
78
78
58
42
67
89
81
9
9
9
9
9
9
9
9
34
39
40
32
22
21
15
19
76
87
89
71
49
47
33
42
d.f.
2011 The Linnean Society of London, Botanical Journal of the Linnean Society, 2011, 165, 186221
MORPHOMETRICS OF CARAPA
Table 3. Loadings of the three first axes of the principal
coordinate analysis (PCoA) of all 292 Carapa specimens
from throughout its range, eigenvalues, percentage of variance and cumulative percentage of variance explained by
the first three axes
Variables
PCoA1
PCoA2
PCoA3
MERO
PEDIND
CALIND
OVUN
CAUL
PEDL
NECH
OVAL
STID
STYL
PISL
OVAW
NECD
ANTW
ANTL
STATD
STATLL
STATL
PETW
PETL
SEPL
CALD
Eigenvalue
Percentage of variance
Cumulative percentage
of variance
0.7093
0.6959
0.5584
0.3752
0.3457
0.1919
0.1596
0.1180
0.0781
-0.0949
-0.0980
-0.1074
-0.1414
-0.2120
-0.2649
-0.2765
-0.3009
-0.3138
-0.3171
-0.3218
-0.3631
-0.4202
10.2893
28.9474
28.9474
-0.0584
-0.2815
-0.2482
0.0651
-0.3339
0.2762
0.4921
0.3299
0.3137
0.1613
0.3299
0.0144
0.2190
-0.3123
-0.0364
-0.3452
-0.7813
0.0624
0.0755
0.1569
-0.0336
-0.0655
8.8386
24.8661
53.8135
-0.2825
-0.1053
-0.1003
0.7218
0.0744
0.0319
-0.0390
0.1858
-0.3609
0.2120
0.2378
-0.2365
-0.2893
-0.2757
-0.0574
-0.0146
0.2850
0.2744
-0.0254
0.1203
-0.1494
-0.2074
5.1058
14.3645
68.1781
197
2011 The Linnean Society of London, Botanical Journal of the Linnean Society, 2011, 165, 186221
198
D. KENFACK
Figure 7. Scatter plots of the two first axes of the principal coordinate analysis of specimens of Carapa from Africa and
America. A, all specimens with overlay of the four major clades from the internal transcribed spacer (ITS) topology.
B, with overlay of the 27 morphospecies.
2011 The Linnean Society of London, Botanical Journal of the Linnean Society, 2011, 165, 186221
MORPHOMETRICS OF CARAPA
Table 4. Loadings of the three first axes of the principal
coordinate analysis (PCoA) of all 131 specimens of Carapa
from the New World Tropics, eigenvalues, percentage of
variance and cumulative percentage of variance explained
by the first three axes
Variables
PCoA1
PCoA2
PCoA3
MERO
STID
CALIND
PEDIND
CAUL
OVAL
NECH
OVUN
PEDL
ANTW
OVAW
NECD
STATD
PISL
PETAW
SEPL
ANTL
STYL
CALD
STATLL
PETAL
STATL
Eigenvalue
Percentage of variance
Cumulative percentage
of variance
0.9342
0.8584
0.7548
0.7423
0.6694
0.4623
0.2049
0.1486
0.0414
0.0169
-0.0496
-0.0841
-0.2747
-0.3593
-0.3665
-0.4513
-0.4785
-0.4963
-0.4973
-0.5231
-0.5826
-0.6698
4.8557
38.8527
38.8527
-0.3260
0.5964
-0.8177
-0.7465
-0.1830
0.6356
0.5750
0.6929
0.3398
0.0466
0.0222
0.2381
0.1656
0.1999
0.1566
-0.4978
-0.2701
-0.3269
-0.4528
-0.1909
0.0849
0.0580
2.22586
17.8101
56.6628
-0.1541
-0.2548
0.0956
0.1096
0.2457
-0.3761
0.0153
0.2729
0.2043
0.7291
-0.4137
-0.1375
-0.0497
-0.1808
0.1423
-0.4977
0.4169
0.0317
-0.5030
0.2519
0.0321
0.0201
1.461783
11.6964
68.3592
MORPHOLOGICAL
VARIATION IN
AFRICAN
SPECIMENS
199
2011 The Linnean Society of London, Botanical Journal of the Linnean Society, 2011, 165, 186221
200
D. KENFACK
0.30
Principal Coodinate 2
0.20
0.10
0.00
0.10
0.20
0.30
Clade I
Clade II
not sequenced
0.40
0.00
0.10
0.20
0.30
0.40
0.20
0.30
0.40
Principal Coodinate 1
0.25
0.20
0.15
Principal Coodinate 3
0.10
0.05
0.00
0.05
0.10
0.15
0.20
Clade I
Clade II
not sequenced
0.25
0.30
0.50 0.40 0.30 0.20 0.10
0.00
0.10
Principal Coodinate 1
Figure 8. Scatter plots of the principal coordinate analysis of specimens of Carapa from America. AB, axes 1 and 2 and
1 and 3, with overlay of the two major clades I and II. CD, axes 1 and 2 and 1 and 3 with overlay of the 11 morphospecies.
The first three axes represent 68.8% of the total variation. See Table 4 for the loadings of the different characters used.
2011 The Linnean Society of London, Botanical Journal of the Linnean Society, 2011, 165, 186221
MORPHOMETRICS OF CARAPA
0.30
Principal Coodinate 2
0.20
0.10
0.00
0.10
0.20
0.30
1
5
9
2
6
10
3
7
11
4
8
0.40
0.50 0.40 0.30 0.20 0.10
0.00
0.10
0.20
0.30
0.40
0.20
0.30
0.40
Principal Coodinate 1
0.25
0.20
0.15
Principal Coodinate 3
0.10
0.05
0.00
0.05
0.10
0.15
0.20
1
5
9
0.25
2
6
10
3
7
11
4
8
0.30
0.50 0.40 0.30 0.20 0.10
0.00
0.10
Principal Coodinate 1
Figure 8. Continued
2011 The Linnean Society of London, Botanical Journal of the Linnean Society, 2011, 165, 186221
201
202
D. KENFACK
PCoA1
PCoA2
PCoA3
STID
STATD
MERO
CALD
PEDL
SEPL
OVAL
ANTL
PEDIND
STATLL
OVAW
PETL
NECH
PETW
NECD
PISL
STYL
STATL
ANTW
OVUN
Eigenvalue
Percentage of variance
Cumulative percentage
of variance
0.4076
0.3083
0.2966
0.2893
0.2808
0.2532
0.0566
0.0462
-0.0566
-0.0685
-0.0697
-0.0750
-0.0857
-0.1206
-0.1622
-0.1712
-0.2370
-0.2655
-0.3095
-0.3170
1.6840
35.5336
35.5336
-0.1481
-0.0829
-0.1096
0.1902
0.0832
0.2067
-0.4510
0.3039
0.0078
0.3988
-0.4764
0.1927
0.0241
0.1679
-0.1678
-0.2733
-0.0645
0.0778
0.0088
0.1118
0.9423
19.8836
55.4172
-0.0849
-0.1438
0.0792
0.0304
0.3254
-0.2820
0.0713
-0.0507
0.1221
-0.3225
0.1107
0.1230
0.0540
0.4762
0.0570
-0.3015
-0.4359
-0.0298
-0.1118
0.3136
0.5001
10.5525
65.9698
DISCUSSION AND
TAXONOMIC INFERENCES
The ITS sequence data and statistical analyses of
leaf, floral and seed characters showed correlated
molecular and morphological variation in Carapa. In
2011 The Linnean Society of London, Botanical Journal of the Linnean Society, 2011, 165, 186221
MORPHOMETRICS OF CARAPA
203
Figure 9. Scatter plots of axes 1 and 2 of the principal coordinate analysis of specimens of Carapa from the Cis-Andes
with overlay of the five morphospecies of the area. See Table 5 for the loadings of the different characters used.
2011 The Linnean Society of London, Botanical Journal of the Linnean Society, 2011, 165, 186221
204
D. KENFACK
PCoA1
PCoA2
PCoA3
OVUN
NECH
PEDL
STATD
STATLL
OVAL
STID
PISL
STATL
PETW
PETL
ANTW
NECD
ANTL
OVAW
STYL
CALD
SEPL
PEDIND
CALIND
Eigenvalue
Percentage of variance
Cumulative percentage
of variance
0.2132
0.2017
0.1514
0.1275
0.1249
0.1227
0.1226
0.1134
0.1121
0.1050
0.1004
0.0679
0.0310
0.0015
-0.0566
-0.0577
-0.1242
-0.1455
-0.5997
-0.6116
3.97535
62.5358
62.5358
0.1964
0.2778
0.1784
-0.0410
-0.1051
0.2003
0.3916
-0.1489
-0.2052
-0.1239
-0.1500
0.2887
0.0446
-0.1016
-0.1114
-0.2642
-0.3841
-0.3618
0.2067
0.2125
0.7504
11.8048
74.3405
-0.1066
-0.0600
0.0549
-0.0527
0.0673
0.4622
0.1033
-0.0238
-0.0191
-0.0532
-0.0275
-0.4007
0.2081
-0.4309
0.4345
-0.3618
0.1708
0.0783
-0.0066
-0.0367
0.3946
6.2081
80.5486
2011 The Linnean Society of London, Botanical Journal of the Linnean Society, 2011, 165, 186221
MORPHOMETRICS OF CARAPA
205
Figure 10. Scatter plot of axes 1 and 2 of the principal coordinate analysis of specimens of Carapa from the Trans-Andes
region with overlay of the six morphospecies of the clade II. See Table 6 for the loadings of the different characters used.
7) were placed in clade I. These morphological, biogeographical and phylogenetic differences suggest
that these morphospecies (2, 4, 5 and 6) can be
considered as distinct species from C. guianensis s.s.
Of the five morphospecies of clade II, morphospecies 2, restricted to Panama, has the narrowest distribution range (Fig. 15). It can be distinguished from
the remaining morphospecies of clade II, based on
their coriaceous leaflets with a rounded to emarginated apex and their distinctly pedicellate flowers with
four-ovulate locules. In the field, individuals belonging to morphospecies 2 are readily distinguished from
those of morphospecies 4 that occur in the same area.
The first is a tall tree branched near its top, the
branches spreading upwards, whereas the second is a
small tree of marshy areas branching low and often
with the branches arching downwards. Specimens of
morphospecies 2 do not match any existing type and
are to be described as a new species.
Morphospecies 4 is the second most widespread of
those in the New World, occurring from Nicaragua
to Panama and along the Pacific coast of Colombia
and Ecuador (Fig. 15). It occurs mostly at low altitudes (< 300 m) and favours marshy conditions. Its
range overlaps with those of morphospecies 1, 5, 6,
and 8. In the herbarium, specimens of morphospecies 4 are readily distinguished on the basis of their
numerous prominent secondary veins, the presence
of a rusty indumentum on the lower surface of
2011 The Linnean Society of London, Botanical Journal of the Linnean Society, 2011, 165, 186221
206
D. KENFACK
0.30
Principal Coodinate 2
0.20
0.10
0.00
0.10
0.20
0.30
Clade I
Clade II
not sequenced
0.40
0.10
0.20
0.30
0.40
0.50
0.30
0.40
0.50
Principal Coodinate 1
0.30
Principal Coodinate 3
0.20
0.10
0.00
0.10
0.20
0.30
Clade I
Clade II
not sequenced
0.40
0.40 0.30 0.20 0.10
0.00
0.10
0.20
Principal Coodinate 1
Figure 11. Scatter plots of the principal coordinate analysis of Carapa specimens from Africa. AB, all specimens with
overlay of the two major clades III and IV. CD, all specimens with overlay of the 16 morphospecies. See Table 7 for the
loadings of the different characters used.
2011 The Linnean Society of London, Botanical Journal of the Linnean Society, 2011, 165, 186221
MORPHOMETRICS OF CARAPA
0.40
0.30
12
16
20
24
13
17
21
25
14
18
22
27
15
19
23
27
Principal Coodinate 2
0.20
0.10
0.00
0.10
0.20
0.30
0.40
0.40 0.30 0.20 0.10
0.00
0.10
0.20
0.30
0.40
0.50
Principal Coodinate 1
0.30
Principal Coodinate 3
0.20
0.10
0.00
0.10
0.20
12
15
18
21
24
27
0.30
0.40
0.40 0.30 0.20 0.10
0.00
0.10
0.20
0.30
13
16
19
22
25
14
17
20
23
26
0.40
Principal Coodinate 1
2011 The Linnean Society of London, Botanical Journal of the Linnean Society, 2011, 165, 186221
0.50
207
208
D. KENFACK
PCoA1
PCoA2
PCoA3
ANTW
STID
OVUN
OVAW
NECH
ANTL
NECD
STATD
OVAL
STATLL
PEDL
PEDIND
PETW
CALIND
PETL
CALD
SEPL
STATL
STYL
PISL
Eigenvalue
Percentage of variance
Cumulative percentage
of variance
0.3340
0.3171
0.2463
0.2358
0.1977
0.1886
0.1285
0.1184
0.0467
0.0396
0.0373
0.0360
-0.0715
-0.0745
-0.2162
-0.2351
-0.2671
-0.3419
-0.3524
-0.3673
6.19
33.4732
33.4732
-0.0442
-0.1278
0.4234
-0.1179
0.1652
-0.0389
-0.2830
-0.2187
0.2798
-0.0601
0.1902
0.1835
-0.3147
0.1654
-0.2038
-0.3954
-0.1445
0.0227
0.2569
0.2618
4.1249
22.306
55.7793
-0.2695
0.0970
-0.0402
0.2830
-0.1962
-0.3473
0.3155
0.5042
-0.1689
-0.0911
0.2119
0.0930
-0.2917
0.0145
-0.2056
-0.0294
-0.2302
0.0787
0.0722
0.2002
2.92
15.79
71.5692
PCoA1
PCoA2
PCoA3
STATLL
CALD
SEPL
ANTL
STATD
ANTW
OVAW
NECD
PETW
PEDL
STID
PETL
MERO
STATL
STYL
NECH
PISL
OVAL
PEDIND
OVUN
Eigenvalue
Percentage of variance
Cumulative percentage
of variance
0.2964
0.2763
0.2557
0.2216
0.2029
0.1921
0.1725
0.1211
0.0886
0.0414
-0.0026
-0.0267
-0.0983
-0.0984
-0.1155
-0.1846
-0.1928
-0.2787
-0.3179
-0.5531
4.3945
56.6603
56.6603
0.3120
-0.1201
0.0184
-0.0780
0.3048
-0.2339
0.1019
-0.1376
-0.4507
0.3302
-0.1664
-0.3346
0.1644
0.0484
0.2980
-0.2406
0.1373
-0.1202
0.2056
-0.0389
1.377
17.7547
74.415
0.3164
-0.0070
-0.0680
-0.1139
0.0768
0.0890
-0.0244
-0.0702
0.0689
0.1287
0.1579
-0.1015
0.2778
-0.3096
-0.4443
0.1053
-0.4804
-0.1696
0.2396
0.3285
0.7228
9.3188
83.7337
(Styles & White, 1991). Clade III included two wellsupported subclades and six morphospecies, all from
the Central African forest block (Fig. 16).
Morphospecies 16, 19, 23 and 25 formed a wellsupported clade and also formed discrete clusters in
the PCoA morphospace. With the exception of specimens assigned to morphospecies 19, an entity
endemic to the Eastern Africa montane forest (in
Burundi, Rwanda and Uganda), the ranges of these
morphospecies largely overlap.
Morphospecies 16 has a narrow distribution in
lowland humid forest of southern Cameroon, an area
where morphospecies 23 and 25 also occur. It comprises understory treelets growing to only 3 m tall,
whereas morphospecies 23 and 25 are trees to 20 m
tall. The leaves of morphospecies 16 are only three- or
four-jugate, the leaflets being distinctly cuspidate at
the apex, whereas those of the two other morphospecies have up to 22 pairs of leaflets, with rounded to
acuminate apices. Moreover, specimens of morphospecies 16 have bi-ovulate locules, whereas morphospe-
2011 The Linnean Society of London, Botanical Journal of the Linnean Society, 2011, 165, 186221
MORPHOMETRICS OF CARAPA
209
Figure 12. Scatter plots of the principal coordinate analysis of 52 specimens of Carapa from the central African forest
block belonging to clade III of the internal transcribed spacer (ITS) topology. A, axes 1 and 2. B, axes 1 and 3. See Table 8
for character loadings.
2011 The Linnean Society of London, Botanical Journal of the Linnean Society, 2011, 165, 186221
210
D. KENFACK
Figure 13. Scatter plots of the principal coordinate analysis of specimens of Carapa from Africa belonging to the clade
IV of the internal transcribed spacer (ITS) topology. A, all specimens with overlay of the 10 morphospecies. B, subset of
specimens from West Africa. C, subset of specimens from central Africa. D, subset of a composite group of five
morphospecies from Central Africa. See Table 9 for character loadings.
2011 The Linnean Society of London, Botanical Journal of the Linnean Society, 2011, 165, 186221
MORPHOMETRICS OF CARAPA
2011 The Linnean Society of London, Botanical Journal of the Linnean Society, 2011, 165, 186221
211
212
D. KENFACK
Variables
PCoA1
PCoA2
PCoA3
Variables
PCoA1
PCoA2
PCoA3
PEDL
SEPL
CALD
PETL
PETW
STATD
STID
STATL
STATLL
ANTL
ANTW
NECH
NECD
STYL
OVAL
PISL
OVAW
OVUN
PEDIND
Eigenvalue
Percentage of variance
Cumulative percentage
of variance
0.60467
-0.76507
0.52578
1.88102
0.58930
0.26560
-0.61538
1.44072
-1.04728
-0.96648
-1.21717
-0.78089
0.25741
-0.55421
-0.83612
0.16540
-0.90895
1.45100
0.51065
2.61446
30.91000
30.91000
0.25132
0.02516
0.16590
0.08867
0.07424
0.18351
-0.02243
0.08128
0.01937
-0.01362
-0.03399
-0.06634
0.09845
-0.03253
-0.07738
-0.03270
0.02852
-0.42149
-0.31593
2.03721
24.08530
54.99530
0.44919
-0.02055
-0.08148
-0.16028
-0.22492
-0.00473
-0.01292
-0.00846
-0.02863
-0.03127
-0.02765
-0.01977
-0.06804
0.02968
-0.02046
0.05027
-0.02125
-0.02957
0.23085
1.16258
13.74480
68.74020
PEDL
STATLL
STATD
OVAW
STID
NECH
NECD
STATL
PETL
ANTL
CALD
SEPL
OVAL
PEDIND
PETW
ANTW
LIND
PISL
STYL
Eigenvalue
Percentage of variance
Cumulative percentage
of variance
0.3718
0.3543
0.3383
0.2639
0.2140
0.1877
0.0545
-0.0170
-0.0323
-0.0413
-0.0605
-0.0607
-0.0944
-0.1490
-0.1566
-0.1648
-0.1806
-0.4066
-0.4207
1.3525
35.57
35.57
0.3474
0.3363
-0.0013
-0.2390
-0.0312
-0.2576
-0.3307
-0.1485
-0.1778
0.1049
0.1566
0.1822
-0.2829
0.3537
-0.2797
-0.1704
0.1796
0.0026
0.2557
1.0553
27.76
63.33
0.0301
0.1996
-0.1792
-0.0697
-0.0528
0.2833
0.0836
-0.0260
0.0231
-0.3263
-0.4258
-0.4025
0.2047
0.5022
-0.1070
0.0428
0.2680
-0.0032
-0.0448
0.322
8.47
71.80
2011 The Linnean Society of London, Botanical Journal of the Linnean Society, 2011, 165, 186221
MORPHOMETRICS OF CARAPA
Table 11. Loadings of the three first axes of the principal
coordinate analysis (PCoA) of seven specimens groups of
Carapa of clade IV from Central Africa, eigenvalues, percentage of variance and cumulative percentage of variance
explained by the first three axes
Variables
PCoA1
PCoA2
PCoA3
OVUN
OVAL
PEDIND
PISL
STYL
NECH
STATL
MERO
PEDL
PETL
STID
NECD
STATD
OVAW
PETW
ANTL
SEPL
CALD
ANTW
STATLL
Eigenvalue
Percentage of variance
Cumulative percentage
of variance
0.4821
0.3117
0.2906
0.2596
0.1749
0.1621
0.1601
0.0535
0.0240
-0.0139
-0.0253
-0.1247
-0.1403
-0.1513
-0.1684
-0.2100
-0.2198
-0.2544
-0.2758
-0.3347
1.7837
45.8165
45.8165
0.0885
-0.1920
0.2069
-0.0094
0.1697
-0.2626
-0.0524
0.2761
0.1118
-0.3483
-0.1532
-0.0857
0.4918
0.1432
-0.3627
-0.1359
-0.0004
-0.1151
-0.1292
0.3588
0.5764
14.8054
60.6219
0.0464
-0.4102
-0.1283
-0.0140
-0.0482
0.2612
0.0683
-0.2487
0.2831
-0.0591
0.4151
0.3229
0.3597
-0.0781
-0.1156
-0.2514
-0.0814
-0.0184
0.0072
-0.3105
0.4435
11.3923
72.0143
logical differences, even in sympatry, they are considered here to belong to two separate species. The
comparison of their morphological characteristics to
the protologues of validly published species and type
specimens seen, showed that material of morphospecies 23 more or less matches C. parviflora Harms and
specimens of morphospecies 25 match C. macrantha
Harms, both described from central Africa. The name
C. parviflora was applied to specimens of morphospecies 25 because they have small flowers and despite
one of their main characteristic features, the farinose
indumentum of the inflorescence branches and
pedicel, not mentioned in the protologue of C. parviflora. The name C. macrantha was assigned to specimens of morphospecies 23 which has 1022 pairs of
leaflets, despite the protologue of the C. macrantha
mentioning only four to 11 pairs.
The second subclade of the African clade III comprised morphospecies 15 and 27, both restricted to the
rainforest of Cameroon, Equatorial Guinea and
Gabon (Fig. 16). Specimens assigned to the former all
came from submontane cloud forest (8001200 m
elevation). They are tall trees to 35 m, their leaves
213
Molecular
clade
1
2
3
4
5
American
American
American
American
American
I
II
I
II
II
American II
7
8
N/A
American II
9
10
N/A
American I
11
12
13
14
15
16
17
18
American I
African IV
African IV
African IV
African III
African III
African IV
N/A
19
20
21
22
23
24
25
26
African
African
African
African
African
African
African
African
27
African III
III
IV
IV
IV
III
IV
III
IV
2011 The Linnean Society of London, Botanical Journal of the Linnean Society, 2011, 165, 186221
214
D. KENFACK
Figure 14. Distribution of the five morphospecies of Carapa from the Cis-Andes region of the New World, four of which
formed clade I of the internal transcribed spacer (ITS) topology.
based on their habitat specialization and the morphological differences. In the transitional zone between
submontane and montane forest, specimens of morphospecies 15 matched the protologue of C. angustifolia Harms, one of the five species described from
Cameroon (Harms, 1917), whereas those of morphospecies 27 had a novel combination of characters
that indicate that they belong to a new species.
The African clade IV is the most species rich of the
four clades of the ITS tree, including nine morphospecies that span almost the entire distribution range of
the genus in Africa (Fig. 17). Specimens of morphospecies 18 from So Tome were included in the
PCoA of this clade as they originate from the same
area. The nine morphospecies formed three wellsupported subclades.
Subclade I included specimens from the two morphospecies 20 and 22. The former is restricted to
montane forest in Central Africa, where it often
occurs in sympatry with individuals of morphospecies
15 in the transitional zone between submontane and
montane forest, whereas the latter occurs in coastal
forest from sea level to only 200 m elevation. The two
members of subclade I also present differences in
2011 The Linnean Society of London, Botanical Journal of the Linnean Society, 2011, 165, 186221
MORPHOMETRICS OF CARAPA
215
Figure 15. Distribution of the six morphospecies of Carapa from the Trans-Andes region of the New World, five of which
formed clade II of the internal transcribed spacer (ITS) topology.
2011 The Linnean Society of London, Botanical Journal of the Linnean Society, 2011, 165, 186221
216
D. KENFACK
Figure 16. Distribution of the six morphospecies of Carapa from Central Africa, which formed clade III of the internal
transcribed spacer (ITS) topology.
2011 The Linnean Society of London, Botanical Journal of the Linnean Society, 2011, 165, 186221
MORPHOMETRICS OF CARAPA
217
Figure 17. Distribution of the nine morphospecies of Carapa from Central and West Africa, which formed clade IV of the
internal transcribed spacer (ITS) topology.
By and large, the patterns of morphological variation revealed in this study are consistent with the
delimitation of many species recognized in the past,
but placed in synonymy by recent workers (Table 12).
Indeed, 15 of the 27 morphological entities recognized
here can be assigned to previously described species,
based on comparisons with their protologues and/or
the type specimens (Table 12), whereas the 12 morphospecies remaining exhibit novel combinations of
characters that suggest that they represent undescribed taxa.
IMPORTANT
DISCRIMINATING CHARACTERS
2011 The Linnean Society of London, Botanical Journal of the Linnean Society, 2011, 165, 186221
218
D. KENFACK
ACKNOWLEDGEMENTS
I am grateful to the Kellogg Lab for laboratory assistance; Pierre-Michel Forget, Sainge Moses, Rosa Ortiz
and Fidy Ratovozon, for collection; David Neil,
Gretchen Walters, Kayombo, Ludovic Ngok Banak,
Mathieu Gueye, Renato Valencia, Richard Condit and
Yves Issembe for their help during fieldwork and for
helping with exportation permits; Pete Lowry, Peter
Stevens and Pierre Michel Forget for their useful
comments on the original manuscript. Financial
support for this study was provided by the National
Geographic Society, the Center for Tropical Forest
Sciences, the Whitney R. Harris World Ecology
Center, the Missouri Botanical Garden and Idea Wild.
REFERENCES
Adjanohoun E, Ahyi MRA. 1985. Contribution aux tudes
ethnobotaniques et floristiques au Mali. Paris: Agence de
Cooperation Culturelle et Technique, (A.C.C.T.).
Adjanohoun E, Ake Assi L. 1979. Contribution au recensement des plantes mdicinales de Cte dIvoire. Abidjan:
Centre National de Floristique. Universit dAbidjan.
Baldwin BG. 1993. Molecular phylogenetics of Calycadenia
(Compositae) based on ITS sequences of nuclear ribosomal
DNA: chromosomal and morphological evolution reexamined. American Journal of Botany 80: 222238.
Bouquet A. 1969. Fticheurs et mdecines traditionnelles du
Congo (Brazzaville). Paris: Mmoire O.R.S.T.O.M No 6.
de Candolle C. 1878. Meliaceae. In De Candolle A & De
Candolle C, eds. Monographiae phanegamarum. Paris:
A Masson. 1: 389758.
Chae SS, Warde WD. 2006. Effect of using principal coordinate analysis and principal components on retrieval of
clusters. Computational Statistics and Data Analysis 50:
14071417.
Cresswell JE, Hagen C, Woolnough JM. 2001. Attributes
of individual flowers of Brassica napus L. are affected by
defoliation but not by intra-specific competition. Annals of
Botany 88: 111117.
Dick CW, Abdul-Salim K, Bermingham E. 2003. Molecular
systematic analysis reveals cryptic tertiary diversification of
a widespread tropical rain forest tree. The American Naturalist 162: 691703.
Dorken ME, Barrett SCH. 2004. Phenotypic plasticity of
vegetative and reproductive traits in monoecious and dioecious populations of Sagittaria latifolia (Alismataceae): a
clonal aquatic plant. Journal of Ecology 92: 3244.
Forget PM, Poncy O, Thomas RS, Hammond DS,
Kenfack D. 2009. A cryptic new species of Carapa Aublet
(Meliaceae) from Central Guyana. Brittonia 64: 366374.
Gentry AH, Dodson CH. 1988. New species and a new
2011 The Linnean Society of London, Botanical Journal of the Linnean Society, 2011, 165, 186221
MORPHOMETRICS OF CARAPA
combination for plants from trans-Andean South America.
Annals of the Missouri Botanical Garden 75: 1434
1436.
Gower JC. 1966. Some distance properties of latent root and
vector methods used in multivariate analysis. Biometrika
53: 325338.
Gueye M, Kenfack D, Forget PM. 2010. Importance socioculturelle, potentialits conomiques et thrapeutiques du
Carapa (Meliaceae) au Sngal. In: Van der Burgt X, van
der Maesen J, Onana J-M, eds. Systematics and conservation of African plants, Proceedings of the 18th AETFAT
Congress, Yaounde, Cameroon: 359367.
Harms H. 1896. Meliaceae. In: Engler A, Prantl K, eds. Die
natrlichen Pflanzenfamilien 3. Leipzig: W. Engelmann,
258308.
Harms H. 1917. Meliaceae africanae. Notizblatt des Knigl.
botanischen Gartens und Museums zu Berlin 7: 223
232.
Harms H. 1940. Meliaceae. In: Engler A, Prantl K, eds. Die
natrlichen Pflanzenfamilien. ed. 2, 19b 1. Leipzig: W.
Engelmann, 1172.
Hebert PDN, Penton EH, Burns JM, Janzen DH, Hallwachs W. 2004. Ten species in one: DNA barcoding
reveals cryptic species in the Neotropical skipper butterfly
Astraptes fulgerator. Proceedings of the National Academy
of Science of the United States of America 101: 14812
14817.
Henderson A. 2006. Traditional morphometrics in plant systematic and its role in palm systematics. Botanical Journal
of the Linnean Society 151: 103111.
Holmgren PK, Holmgren NH, Barnett LC. 1990. Index
herbariorum. Part I: the herbaria of the world, 8th edn. New
York: New York Botanical Garden.
Hutchinson J, Dalziel JM. 1958. Flora of West Tropical
Africa (2nd edition, revised by R. W. J. Keay), Vol. 1(2).
London: Crown Agents for Overseas Governments and
Administrations.
Keay RWJ. 1989. Trees of Nigeria. Oxford: Clarendon
Press.
Keay RWJ, Onoche CFA, Stanfield DP. 1964. Nigerian
trees. Ibadan: Federal Department of Forest Research.
Kenfack D, Prez A. 2011. Two new species of Carapa
(Meliaceae) from Western Ecuador. Systematic Botany. In
press.
Kenfack D, Thomas DW, Chuyong GB, Condit R. 2007.
Rarity and abundance in a diverse African forest. Biodiversity Conservation 16: 20452074.
Kerharo J, Adam JG. 1974. La pharmacope sngalaise
traditionnelle. Plantes mdicinales et toxiques. Paris: Editions Vigot Frres.
Kribs DA. 1930. Comparative anatomy of the woods of the
Meliaceae. American Journal of Botany 17: 724738.
Lamarck JBAP. 1785. Encyclopdie mthodique 1: 621.
Legendre P, Legendre L. 1998. Numerical ecology, 2nd edn.
Amsterdam: Elsevier.
Martnez-Ortega MM, Luis Delgado L, Dirk C, Albach
DC, Juana A, Elena-Rossell JA, Rico E. 2004. Species
boundaries and phylogeographic patterns in cryptic taxa
219
2011 The Linnean Society of London, Botanical Journal of the Linnean Society, 2011, 165, 186221
220
D. KENFACK
APPENDIX
LIST
MORPHOMETRIC ANALYSIS
2011 The Linnean Society of London, Botanical Journal of the Linnean Society, 2011, 165, 186221
MORPHOMETRICS OF CARAPA
1958; GUINEA. A. Chevalier 14865. A. Chevalier
14869; A. Chevalier 20474; Cordonier 243; F. Malaisse
2541. LIBERA. A.G. Voorhoeve 1288; D.H. Linder 905;
D.H. Linder 1156; J.G. Adam 28702; J.T. Baldwin Jr
10284; J.W.A. Jansen 1360. SENEGAL. A. Chevalier
3164; Anonymous 3166; D. Kenfack, M. Gueye & M.
Sadio 2076; D. Kenfack, M. Gueye & M. Sadio 2077;
D. Kenfack, M. Gueye & M. Sadio 2080; D. Kenfack,
M. Gueye & M. Sadio 2085; Fotius K811; Kaoussou
Sambou, 1; Kaoussou Sambou, T. Sarr sn; Madsen
J.E. 3123. SIERRA LEONE. C.E. Lane-Pook 349; D.
Cledhill 301; G.F. Scott Elliot 4153; M. Heudelot 749;
M. Heudelot 749.
Morphospecies 13: BENIN. C. Barter 3248. GHANA.
A.A. Enti 613; C. Vigne 2525; Hutchinson 146; J.
Deaw 380; J.J. Chipp 262; J.K. Morton 25327; N.H.
Johnson 146; IVORY COAST. A. Chevalier 16233;
IVORY COAST. A. Chevalier B. 22279; IVORY COAST.
J.J.F.E. de Wilde 3120; IVORY COAST. Martineau 303;
NIGERIA. J.M. Dalziel 342.
Morphospecies 14: GUINEA. A. Chevalier 408; A.
Chevalier 461; A. Chevalier 18192; H. Pobguin sn.
GUINEA-BISSAU. E. Santo 1310. IVORY COAST. H.
Pobguin 264; L. Gautier, R. Spichiger & H. Tr
2857. MALI. M. R. Dubois 38. SENEGAL. D. Kenfack
2070; D. Kenfack 2074; D. Kenfack 2084; D. Kenfack,
M. Gueye & M. Sadio 2071; D. Kenfack, M. Gueye &
M. Sadio 2072; D. Kenfack, M. Gueye & M. Sadio
2079; D. Kenfack, M. Gueye & M. Sadio 2081; D.
Kenfack, M. Gueye & M. Sadio 2083; Goudiaby A.
1273; Jacques-George A. 17514.
Morphospecies 15: CAMEROON. Kenfack 2007;
Sainge M 318.
Morphospecies 16: CAMEROON. G.L. Bates 535; P.
Tchouto 2963; R. Letouzey SRFK 1289.
Morphospecies 17: CAMEROON. A.J.M. Leeuwenberg
5279; R. Letouzey 14950; A. Binuyo 45466; A.S. Jhoneuill 220; D. Kenfack 737; D. Kenfack 1024; D.
Kenfack 1169; D. Kenfack 1170; D. Kenfack 1364; D.
Kenfack 1656; D. Kenfack 2118; D.W. Thomas 2352;
D.W. Thomas 8256; D.W. Thomas & J. Nemba 5932; G.
Zenker 145; G. Zenker 3713; G. Zenker sn; G. Zenker
sn; J.J. Bos 4374; M. Akogo 174; Mainoud 430; T.D.
Maitland 431. EQUATORIAL GUINEA. M.F. de Carvalho
2252. GABON. C. Barter 153. NIGERIA. P.A. Talbot
1462. NIGERIA. P.W. Richards 3013; G. Mann 1767.
221
2011 The Linnean Society of London, Botanical Journal of the Linnean Society, 2011, 165, 186221