close

Вход

Забыли?

вход по аккаунту

?

Dental sexual dimorphisms in some extant hominoids and ramapithecines from China A quantitative approach.

код для вставкиСкачать
American Journal of Primatology 9:305-326 (1985)
Dental Sexual Dimorphisms in Some Extant Hominoids and
Ramapithecines From China: A Quantitative Approach
SUSAN s.LIEBERMAN’, BRUCE R. GELVIN~,AND
CHARLES E. OXNARD3
‘Department ofAnatomy and Cell Biology, School o f Medicine, University o f Southern
Californiq ‘Department o f Anthropology, California State University, Northridge,
“Departments ofAiiatomy and Cell Biology, School of Medicine, and of Biological Sciences,
College of Letters, Arts and Sciences, University o f Southern California
Studies of sexual dimorphism in the dental dimensions of some extant and
fossil hominoids have been carried out by means of univariate statistical
methods [Oxnard et al., 19851. The study reported here extended these
studies with multivariate statistical methods (canonical variates analyses).
The extant genera studied were Gorilla, Pan, Pongo, and Homo. The fossil
teeth interpolated were those of the recently discovered ramapithecines
from Yunnan Province, China. For both extant and fossil species, the lengths
and breadths of all maxillary and mandibular teeth were used except for
the third molar, which was excluded because of its absence in so many
human subjects. The nature of sexual dimorphism in the dentition of extant
apes and humans was assessed, and the positions of the fossil teeth within
the multivariate results for the extant forms was examined. Among the
apes, the greatest sexual dimorphism was seen in Gorilla; the least was
seen in Pan. Three different patterns of sexual dimorphism were apparent
among the three ape species. The maxillary and mandibular patterns were
different in Gorilla and Pan but more similar in Pongo. The African apes
showed greater differences between variances for each sex and of each jaw;
these features may have evolved most recently. The conventional notion
that sexual dimorphism is mainly due to size and size-related shape effects
along a single continuum or axis was rejected. The interpolation of the fossil
data placed Siuapithecus close to each of the more dimorphic apes, especially
Pongo, but also showed that it had higher order differences from the extant
forms studied. Ramapithecus was most similar to Homo. These results have
implications both for the role of sexual dimorphism in the evolution of
higher primates and for the phylogenetic relationships among them.
Key words: dental dimensions, multivariate statistics, canonical variates analysis, sexual dimorphisms, humans, chimpanzees, gorillas, orangutans, ramapithecines, Gorilla, Pan, Pongo, Homo
Received June 10, 1985; revision accepted August 27, 1985.
Address reprint requests to Professor Charles E. Oxnard, 820 Oak Knoll Circle, Pasadena, CA 91106.
0 1985 Alan R. Liss. Inc.
306 I Lieberman, Gelvin, and Oxnard
INTRODUCTION
Structural sexual dimorphism in primates, with respect to both body size and
body proportions, has been studied extensively [Campbell, 1972; Montagu, 1974;
Friedman et al, 1974; Wood, 19761. It is known that females are half the size of
males in baboons, orangutans, and gorillas, are similar in size to males in squirrel
monkeys, colobs, and gibbons, and are larger in size than males in some species of
spider monkeys, lemurs, and marmosets [Schultz, 1969; Napier and Napier, 1967;
Leutenegger & Kelly, 19771. Differences in shape are believed, mainly on the basis
of univariate statistical analyses, to be associated with the above size differences,
being greatest in those species such a s orangutans and howler monkeys that show
the greatest size dimorphism, and smallest in those species such as marmosets and
spider monkeys that show the least size dimorphism. Wood [1975, 19761, examining
several variables together and employing methods such as Penrose size and shape
statistics found that differences in shape between the sexes in primates were attributable primarily to size. However, Geist [1974] and Leutenegger and Kelly [ 19771
found that sexual dimorphism in size was not correlated closely with sexual dimorphism in other dimorphic characters. For example, sexual dimorphism in overall
body size did not correlate well with canine dimorphism, and when there was a
relationship between these two dimorphic features, it was not the same across
several genera and species. Thus, different structural dimorphisms may be both
species specific and character specific. This is consistent with the concept that there
is no single gene or group that codes for the entire range of sexually dimorphic
characters and that the heritability of each quantitative trait (such as body size,
tooth length and breadth, and body weight) is separate and often largely independent. Many authors recognize the roles of multiple factors in the evolution of sexual
dimorphism, such as intensity of both sexual and fecundity selection, mating patterns, and behavior [Gautier-Hion, 1975; Clutton-Brock et al, 1977; Leutenegger &
Kelly, 1977; Leutenegger, 1982; Leutenegger & Cheverud, 19821. Nevertheless,
these investigators still speak of greater or lesser degrees of sexual dimorphism,
thus clinging to the notion that sexual dimorphism is primarily a unidimensional
trait that can be quantified along a single, simple continuum.
More recently, overall body proportions of different primate species were studied
by means of both univariate and multivariate statistics [Oxnard, 1983a,b, 19841. As
expected, the reality was far more complex than a simple unidimensional spectrum
would indicate. Instead, up to seven qualitatively different patterns of sexual dimorphism were found among the twenty primate species included in those studies.
These patterns became evident only when the multivariate statistical approach was
used, because univariate statistics focus, by design, on only one variable and cannot
detect patterns that exist when several variables together with their interactions
are studied. Though the role of size-dependent sexual dimorphism was found to be
important in primates, much of the above picture was independent of size; for
example, sexual dimorphism was almost as great proportionately, in bushbabies a s
in orangutans, two genera a t opposite ends of the scale of difference in sexual
dimorphism of size.
Subsequent studies, with univariate statistical methods, analyzed sexual dimorphism in the dentition of extant and fossil hominoids [Oxnard et al, 1985;
Blumenberg, personal communication]. Oxnard et al, [1985] used the same base of
extant data as that reported here. In comparing the four extant large hominoid
genera, Homo and Pan both had small canine dimorphism, and Gorilla and Pongo
had large canine dimorphism. In both cases, these differences were between mean
values for the sexes, so that males, on the average, had larger canines. However,
when the dispersions for each sex were compared, Gorilla and Pan had significantly
308 / Lieberman, Gelvin, and Oxnard
TABLE I. Numbers of SDecimens Per Sex and Per Genus
Genus
Maxillary teeth
Gorilla
Pan
Pongo
Homo: sexes unknown
Total
Mandibular teeth
Gorilla
Pan
Pongo
Homo: sexes unknown
Total
Females
Males
42
42
38
42
44
41
Total
84
86
79
59
308
39
44
43
40
39
38
79
86
81
56
302
This confirmed that the two groups originally identified by Wu and colleagues truly
existed, and the authors returned to this concept [Wu et al, 19831.
The 955 teeth suitable for study in this way (adult, unbroken, undistorted, and
not too badly worn-data from badly worn teeth were specifically excluded) represented every tooth position in both maxilla and mandible, with 17-49 teeth per tooth
locus. About half of the teeth were found in situ in fragmentary to almost complete
jaws; the other half were recovered as isolated specimens. Lengths and breadths of
these teeth have been analyzed univariately both as an overall group and as
subgroups [Wu & Oxnard, 1983a,b], and in comparison with data from extant higher
primates [Oxnard et al, 19851. Of course, these dimensions have been analyzed as
samples from the two species with sexes undefined. Sex cannot be determined for
individual tooth dimensions in most primates except, perhaps, for dimensions of
canines in especially heavily dimorphic species. It is the fact that each species shows
bimodal univariate distributions possessing all the properties of the bimodal distributions that result from sexual dimorphisms in extant humans and apes [Oxnard et
al, 19851 that implies that the fossil distributions are due to sexual dimorphism.
Methods
Four separate multivariate analyses were performed on the data. For the extant
primates, mandibular and maxillary data were analyzed separately by means of
unweighted canonical variates analyses. For each jaw, the seven groups used were
Gorilla females, Gorilla males, Pan females, Pan males, Pongo females, Pongo
males, and Homo, sex unknown. In each case, therefore, six canonical axes were
generated.
The distributions for the two fossil groups, Ramapithecus and Sivapithecus,
were already known from prior study to be bimodal in many of their measurements.
Where this was the case, the central values for the lower and upper peaks were
taken as best representing the means for “females“ and “males” respectively. Where
the distribution for a given dimension was unimodal, the values for females and
males were taken as equal even though, given complete information about sex for
each specimen, there would be further significant separations between the sexes for
at least some of these variables. These values for females and males were then
interpolated into the canonical variates analyses previously done on the extant
species, employing the previously generated eigenvectors. All analyses were carried
out with the IBM 3081 computer a t the University of Southern California Computing Center, using SAS, the Statistical Analysis System. All graphs were generated
using the SASGRAPH procedure of SAS.
Morphometric Studies of Hominoid Teeth I 307
different dispersions between sexes and Pongo and Homo did not. Thus four different
patterns of sexual dimorphism were found in the canines alone for these four genera.
For non-canine teeth, the patterns were even more complicated. Gorilla and Pongo,
with large overall size differences between the sexes, had large differences between
sexes in mean dental dimensions a t most loci. Pan and Homo, with small overall
size differences between the sexes, had smaller differences in mean dental dimensions. Gorilla and Pan had significant dispersion differences between sexes in most
non-canine teeth as they did in canines. In contrast, in non-canine teeth in Pongo
and Homo, there were no differences between sexes in the dispersions for each
measurement. Thus, sexual dimorphism is related to more than just size. Several
different sexually dimorphic patterns were apparent even in univariate studies.
Blumenberg (personal communication) used coefficients of variation to study
the dentition of extant great apes and fossil dryopithecines and ramapithecines.
Differences between the sexes were measured by their effects (together with those
of other factors) on the variance of the total sample. The dispersion for each sex and
sex ratio differences were not considered. Nevertheless results of this univariate
study strongly support the hypothesis that sexual dimorphism involves several
different patterns, is related to more than size, and is multidimensional.
The goal of the study reported here was to expand the studies of the above data
by employing multivariate statistical analyses. Several questions concerning the
data were addressed: (1)What is the multivariate nature of sexual dimorphism in
the dentition of the large extant Hominoidea? (2) What patterns are apparent in
multivariate analyses that might not appear in univariate studies? (3) Are such
patterns of sexual dimorphism, if they exist, the same in all species? (Further
questions about evolutionary implications stem from the interpolation of the data
on fossils into the multivariate statistical results for extant forms.) (4)Are the
patterns of sexual dimorphism in the dentition of the fossils like those in any of the
extant species, or are they unique? (5) What insights into the evolution of sexual
dimorphism do patterns of sexual dimorphism provide? (6)What are the implications
for primate and human phylogeny?
MATERIALS AND METHODS
Materials
The data on extant primates were from 307 individuals representing four genera: Pan, Pongo, Gorilla, and Homo (Table I). The data were culled from the studies
of Gregory and Hellman [1926], Pilbeam [1969], Wolpoff [1971a], Frayer [ 19731, and
Mahler [1973]. The sexes of apes were known from field or museum records; information on sex was not available for humans. The data consisted of measures of
length and breadth from each tooth (except the third molar) from both the mandible
and the maxilla of adult individuals with established permanent dentition. A univariate description of these data has already been reported [Oxnard et al, 19851.
The fossil data consisted of length and breadth measurements from the newly
discovered Siuapithecus and Ramapithecus teeth from Shihuiba, Lufeng County,
Yunnan Province, China [Wu et al, 1981, 1982, 19831. These fossils were discovered
in late Miocene deposits and dated to approximately eight million years ago [Wu &
Olsen, 19851. The fossils were assigned to these genera on the basis of morphological
examination [Wu et al, 1981, 19821. At this point, influenced by studies of nonChinese ramapithecines [eg, Andrews & Cronin, 19821, Wu et a1 [1983] agreed that
the fossils might comprise only the two sexes of a single orangutan-like species,
Sivapithecus. Subsequently, however, Wu and Oxnard [ 1983a,b] demonstrated an
absolute separation between the canines of these two groups that was far larger
than that found between canines of even the most highly dimorphic living primates.
Morphometric Studies of Hominoid Teeth I 309
TABLE 11. Canonical Variates Analysis of Maxillary Teeth
Variable
Axis 1
Axis 2
I1L
I1B
I2L
12B
CL
CB
P3L
P3B
P4L
P4B
M1L
M1B
M2L
M2B
0.48
0.19
-0.49
-0.16
1.76
0.45
0.34
0.96
0.53
0.22
-0.47
0.02
0.26
-0.28
-0.47
-2.30
0.30
0.42
0.25
1.35
-0.59
0.78
-0.45
-0.54
0.97
-0.31
0.87
-0.15
Standardized canonical coefficients
Axis 3
Axis 4
Axis 5
-0.30
0.56
-0.13
-0.61
-1.16
-1.56
0.53
0.87
0.39
0.25
0.12
0.56
0.86
-0.20
-0.55
0.48
0.48
-1.66
-0.63
2.33
-0.26
-1.90
0.75
0.54
0.30
0.74
-1.18
0.59
0.73
-0.52
-0.35
0.33
-0.07
-0.30
0.42
1.07
-0.91
-1.56
-0.56
1.93
-1.36
1.10
Axis 6
0.53
0.32
0.79
-0.18
-0.21
-0.11
-0.90
-1.35
-0.14
0.27
1.06
0.29
-0.18
0.53
RESULTS AND DISCUSSION
Extant Species
Maxilla. Table I1 lists the standardized canonical coefficients (eigenvectors) for
the maxillary analysis. The 14 variables consisted of the lengths and breadths (L,
B) of the incisors (I1, 11
' , canine (C'), premolars (P3, P4),and the first two molars
(MI, M'). Only the first four axes were statistically significant, comprising more
than 95% of the information within the data. The first canonical axis was accounted
for mainly by canine length, with some involvement of third premolar breadth. The
second axis primarily contained information contributed by the breadth of the first
incisor and by canine breadth. Molars contributed in only lesser degrees to the
separations between the groups in the second and higher axes. With the exception
of canine length, all important variables in the first four axes were breadths. This
was consistent with the concept that most sexual dimorphism can be attributed to
breadth dimensions rather than lengths, whether in regard to bodily dimensions
[Oxnard, 1983~1
or to dimensions of teeth [Wu & Oxnard, 1983a,b]. This was noted
much earlier for the teeth of humans [Garn et al, 19671.
Figure 1presents the scores on the first four canonical axes for all seven of the
extant groups in the maxillary analysis. For the first two axes, the greatest
separation between males and females, representing the greatest degree of sexual
dimorphism, was in Gorilla. The greatest female/male overlap in canonical space,
representing the smallest sexual dimorphism, was in Pan. It should be noted,
however, that the entire distribution of the genus Homo was considerably smaller
than the combined spread for females and males for Pan, indicating that, though
we did not know the sexes for the sample of Homo, dimorphism was even smaller
in Homo than in Pan. This fits well, of course, with the generally accepted views
about these genera. The position of Homo was nearest to female Pan, though
completely separate from them.
The first axis separated the sexes for all species, and the second axis separated
Gorilla, Pan, and Homo from Pongo. The sexes of Pongo, in contrast with those of
Gorilla and Pan, did not differ in the second axis (made up of the breadths of the
first incisors and the canine). Thus, it is immediately clear that the nature of sexual
dimorphism in Pongo differs from that of the other apes. There is also information
310 I Lieberman, Gelvin, and Oxnard
ClVP
' 1
Fig. 1. Plots of canonical axes 1 versus 2, and 3 versus 4 for maxillary studies of extant species. H,
Homo; FTF and PT.M, Pan female and male, respectively; PG.F and PG.M, Pongo female and male,
respectively; GG.F and GG.M, Gorilla female and male respectively. The scales are in standard deviation
units. Clear separations are achieved between the sexes of each ape.
TABLE 111. Canonical Variates Analysis of Mandibular Teeth
Standardized canonical coefficients
Axis 3
Axis 4
Axis 5
Variable
Axis 1
Axis 2
I1L
I1B
12L
12B
CL
CB
P3L
P3B
P4L
P4B
M1L
MlB
M2L
M2B
0.13
0.07
0.31
-0.20
3.40
- 1.04
2.18
-0.06
0.34
-0.54
0.26
-1.29
-0.18
0.72
-0.56
-1.46
-0.28
-0.34
-0.45
0.46
0.89
0.46
-0.77
0.88
1.17
-0.00
1.41
-0.32
-0.16
-1.32
0.23
0.16
1.71
2.41
- 1.83
0.23
-0.55
-0.43
0.34
0.32
-0.48
-0.57
0.87
0.74
-0.95
-0.50
-2.00
1.76
0.11
0.59
1.42
-0.41
- 1.15
1.09
-1.43
0.42
-0.03
0.97
0.88
-0.30
-3.76
1.93
1.34
0.24
-1.43
0.28
-0.76
-0.78
1.57
0.18
Axis 6
0.58
0.07
-0.25
1.64
0.06
0.12
-3.15
-0.22
0.11
0.32
0.98
-0.24
0.33
0.38
that, though more difficult to interpret, shows high-dimensional differences in
sexual dimorphisms between the different genera. This information is addressed in
subsequent sections.
Mandible. Table I11 lists the standardized canonical coefficients for the
mandibular analysis. In this case, the first five axes were significant, and variables
from a different suite of tooth positions than that of the maxilla accounted for the
separation of the groups. The first axis primarily represented canine length and
Morphometric Studies of Hominoid Teeth I 311
0
-
1
-
,
-10
5
0
CAN
5
2
4
CAY,
Fig. 2. Plots of canonical axes 1 versus 2, and 3 versus 4 for mandibular studies of extant species. H,
Homo; FTF and W.M, Pun female and male, respectively; PG.F and PG.M, Pongo female and male,
respectively; GG.F and GG.M, Gorilla female and male, respectively. The scales are in standard deviation
units. Clear separations are achieved between the sexes of each ape; in addition, the overall degree of
separation is greater for the mandibular analyses than for the maxillary (Fig. 1).
the length of the third premolar, together with some contributions from canine
breadth and the breadth of the first molar. The lengths of the first and second
molars and the breadth of the first incisor were the predominant variables in the
second axis. It can be seen that the lengths of the teeth were as important as the
breadths in separating the groups; this contrasts with the situation in the upper
jaw. It is possible that this was related to the effects of interstitial wear [eg, Wolpoff,
1971b], a factor that could not be allowed for in these data.
Figure 2 shows the scores for each group on the first four canonical axes for the
mandibular analysis. The within-group variability was much lower than that for
the maxilla. Homo was much more distinct from the other species in the mandibular
analysis than in the maxillary analysis, which was due mainly to the first axis. As
judged by the shape of its plot, Homo displayed a very different pattern of
dimorphism, mainly in the second axis, than that of apes whose patterns of
dimorphism were mainly in the first axis. For the first two axes, the greatest
separations between males and females, representing the largest dimorphisms,
were, again, in Gorilla; there was no overlap a t all between Gorilla and any of the
other species.
The third and fourth axes also contributed to the pattern of sexual dimorphism.
The third axis separated Gorilla females and males from Pan females and males
almost completely. It required, however, the combination of the third and fourth
axes to separate the sexes of Pongo almost completely, suggesting a totally different
pattern of sexual dimorphism in Pongo than in the other great apes.
High-dimensional plots. For both the maxilla and the mandible, information
in higher axes (though real and statistically significant) is often difficult to interpret,
having graphical and conceptual limitations that preclude ready visualization in
more than two or three dimensions simultaneously. Therefore, the data are
312 / Lieberman, Gelvin, and Oxnard
GORILL.1: LLYILL1RY DATA
GORILLA: 1 J A N D I B U W R D.AT.4
‘‘I
I , :
0;
POPU’GO: X i X D I B U L 4 R DATA
Fig. 3. High-dimensional plots for each extant species for both maxillary and mandibular analyses. In
each case the curve placed highest on the plot is the curve for the maie. The patterns of sexual dimorphism
a r e different for each ape. Humans a r e closest to Pun. The differences are greater for mandibular than for
maxillary analyses. A, Gorilla; B, Pongo; C, Pan and Homo.
graphically represented here with high-dimensional plots (Figs. 3-5, 8-11). The
sine-cosine function [Andrews, 1972, 19731 that generates the high-dimensional
plot for group j follows: f,(t)=xl,/JZ+xgsin(t)+x3Jcos(t)+x4,sin(2t)+xS,cos(Zt)+....,
where xlJ is the canonical mean for group j on axis 1,etc.
These plots have been used in a similar way in canonical variates analyses of
overall bodily proportions in primate sexual dimorphism [Oxnard, 1983~1.The area
between the two plots for two sex subgroups is the squared Mahalanobis generalized
distance (D2) between the centroids of each sex subgroup [Andrews, 1972, 1973;
Morphometric Studies of Hominoid Teeth / 313
P h S AND HOMO: 1USILLARY DATA
t
PAN AND HOMO: hIANDIBULAR DATA
2j
\,{
'0'
0:
8'
4
6'
-81
-9-
..o:
c
- 2
~, .>
.,
c
1
2
,
4
5
-5
-4
-3
-2
-1
0
I
I
I
+
5
Figure 3. Continued.
Oxnard, 1983~1.When males and females are compared, a species with large sexual
dimorphism (as judged by a large squared generalized distance between the sexes)
will have a large area between the high-dimensional plots for each sex. A species
with small sexual dimorphism will have a small area between the plots for each
sex.
It follows that if two sets of two sexes, or any two sets of two groups, have a
similar generalized distance between them, then the areas between the two highdimensional plots will be the same. But if these similar generalized distances are
made up of different combinations of canonical variates values that are due to
different patterns of sexual dimorphism, then this will be reflected in the different
shapes of the two pairs of functions and in the different loci a t which they cross. It
is possible to examine differences in patterns of sexual dimorphism with this
technique.
Oxnard [ 1 9 8 3 ~found,
)
despite different overall curvatures for different species,
similar patterns of curvatures between the sexes in those species in which similar
patterns of sexual dimorphism existed. Oxnard [ 1983~1also found, however, quite
different patterns of curvatures between the sexes in some species that conventionally had been thought to display a similar and very large degree of sexual
dimorphism.
In this way, the high-dimensional display yields more information about the
canonical variates analyses than do the summary figures of the generalized distance
values. Therefore, this method of display is particularly applicable in this study for
assessing the level and degree of complexity of sexual dimorphism in groups about
which we know very little, such as fossils. Work by one of us (S.S.L.)is in progress
to quantify the pattern similarities between pairs of such plots.
Figure 3 shows the sine-cosine plots of the canonical variates means for the
seven extant groups for both maxillary and mandibular analyses. For comparison,
the Mahalanobis distances (D) between the groups are presented in Table IV. As
seen from Figure 3A, male and female Gorilla had more similar patterns in their
lower than in their upper jaws, although the absolute generalized distance between
the sexes for each jaw was not so distinct (6.34 for the maxilla, 6.83 for the
314 I Lieberman, Gelvin, and Oxnard
TABLE IV. Mahalanobis Generalized Distances (D) Between Groups for.Maxillary and
Mandibular Canonical Variates Analyses*
Grow
HS-U
Maxillary distances
HS-U
GG-F
7.41
GG-M
11.06
PT-F
4.34
PT-M
5.75
PG-F
6.71
PG-M
8.98
Mandibular distances
HS-U
GG-F
9.86
GG-M
14.15
PT-F
8.81
PT-M
10.19
PG-F
10.21
PG-M
11.77
GG-F
GG-M
PT-F
PT-M
PG-F
6.34
7.25
7.12
5.59
6.28
10.40
8.17
9.09
6.44
3.34
5.58
7.62
6.07
6.18
4.20
6.83
8.65
9.56
6.70
7.37
11.54
10.04
10.35
7.47
3.89
4.65
6.35
6.77
5.76
4.91
PG-M
*Groups: HS, Homo; GG, Gorilla; PT,Pan; PG, Pongo. F, females; M, males; U, sexes unknown
mandible). This disimilarity between the mandibular and maxillary patterns in
Gorilla may reflect the unique pattern of canine dimorphism found in Gorilla a s
seen from univariate analyses [ Oxnard et al, 19851.
The high-dimensional plots for Pongo (Fig. 3B) show a completely different
pattern of dimorphism than that found in Gorilla, as well as even more marked
differences in pattern between the maxilla and the mandible. Thus, even though
the Mahalanobis distances were less than those in Gorilla (4.2 and 4.9, respectively),
the contrast in pattern between the maxilla and the mandible is greater.
There was less sexual dimorphism in Pan than in the other apes, as seen from
Figure 3. However, notwithstanding this smaller difference between the sexes that
was shown by the Mahalanobis generalized distances, once again there was a quite
different high-dimensional pattern for each jaw, representing a quite different
pattern of separation in the substructure of the canonical variates for each jaw.
The maxillary and mandibular plots for Homo (sexes combined: Fig. 3C) were
quite similar to each other. Indeed, smaller between jaw differences were also
evident in Pongo. Thus, it is possible that increased between-jaw difference (and
possibly increased between-sex variability as well) is the more recently evolved
state; lower between-jaw and between-sex differences in Pongo and Homo ithe low
between-sex differences for Homo are known from the univariate data [ Oxnard, et
al, 1985) may be the ancestral (and probably generalized) state.
The hypothesis that sexual dimorphism in the great apes can be represented
along a single dimension such as size or by a single number such as the coefficient
of variation is rejected. Canonical variates analysis shows that there are four
statistically significant axes for the maxilla and five for the mandible. Three
different patterns of sexual dimorphism emerge from each of the mandibular and
maxillary studies. Some of the differences between the sexes are greatest in axes
other than the first, which might be expected to be the axis carrying the greatest
component of overall size. These pattern differences are strong evidence in support
of a hypothesis of multiple causality of sexual dimorphism, even in so limited an
anatomical region as the dentition.
Morphometric Studies of Hominoid Teeth / 315
No single hypothesis or mechanism is likely to be the source of such a
multiplicity of pattern. Indeed, the nature of sexual dimorphism in each genus may
be associated with specific genetic effects [for review, see Biggerstaff, 19791;
developmental gradients along the tooth row (first suggested as applying to the
dentition by Butler [ 19391);functional differences related to diet (dietary differences
do exist between the sexes as well as between the species in many primates
including some of the Hominoidea); ecological differences (clearly present between
sexes as well as between species, particularly in the case of Pongo); the effects of
interstitial wear leg, Wolpoff, 1971b], including its differential effects in relation to
lengths of teeth a s compared with breadths, and differential sexual selection. It is
likely that a complex interaction between some or all of the above mechanisms
accounts for the different patterns of sexual dimorphism seen in the extant apes.
Further research into this problem is critical.
RESULTS AND DISCUSSION
Fossils
The univariate results for the ramapithecine fossils from Lufeng indicated that
a t least two species or species groups existed eight million years ago in China [Wu
& Oxnard, 1983a,b; Oxnard, 1983b, 19841. The smaller fossil, Ramapithecus, exhibited a small degree of dental sexual dimorphism, less than that in any extant apes.
The larger form, Siuapithecus, showed a somewhat greater degree of dental sexual
dimorphism, more akin to that in Pongo. However, the univariate studies also
showed that the locations along the tooth row of the most marked sexually dimorphic
dimensions were different from those in any extant form and, in particular, did not
involve especially marked canine dimorphism in either fossil species [Oxnard et al,
19851, though enlarged canines with some dimorphism were evident in both upper
and lower jaws in Siuapithecus.
The univariate studies also indicated that approximately equal numbers of teeth
from presumed males and females have been recovered separately for many individual tooth loci for Ramapithecus. Of course, one must be most cautious in suggesting
hypotheses about social structure based upon this finding alone. However, when
equal sex ratios are associated with reduced sexual dimorphism together with
reduced canine sexual dimorphism and reduced canines, as is the case in Ramapithecus, strong morphological sexual selection can probably be discounted. Either
pair bonding or social groups of equal numbers is suggested [Wu & Oxnard, 1983a,b;
Oxnard, 1983b, 1984; Oxnard et al, 1985; Blumenberg, personal communication;].
Siuapithecus, on the other hand, was represented by teeth from presumed males
and females in ratios ranging between two to one and four to one. This was associated with considerably more marked dental sexual dimorphism overall (though not
specifically canine sexual dimorphism) and increased canine size in both jaws compared to the other teeth. Thus, this species may have been polygynous and under
increased sexual selection pressure [Wu & Oxnard, 1983a,b; Oxnard, 198313 1984;
Oxnard et al, 1985; Blumenberg, personal communication]. Unequal numbers of
adult males and females are to be expected in a polygynous situation. Polygyny is
well known in different forms among many monkeys and all the great apes. In each
case that has been examined (some monkeys in detail, most apes only anecdotally),
the sex ratio in adult life is heavily biased toward the female, though the sex ratio
a t birth is generally believed to be even. This results from death rates perinatally,
in infancy, in juveniles, and in adolescents that are biased toward the males (see
Drickamer [1974], Butynski [1982], and Meikle et al, [1984] for detailed studies of
some polygynous monkeys; and Hrdy [1981] for more anecdotal materials on the
polygynous apes).
316 I Lieberman, Gelvin, and Oxnard
CA!NOXIC~ YARLirnS MEAVS: huxILL&RY
C1*,
I .,Of
b.?5.
2
0.50.
6
H
2 c
0.11-
0.00-
-0.11-
* B
D S
-3.10.
-0.?l
C
6
Fig. 4 . Canonical variate means for each extant group and the interpolated positions of the putative
mean for each sex of the fossil groups in the maxillary analysis. H, Homo; 1 and 2, females and males of
Gorilla; 3 and 4, females and males of Pan; 5 and 6, females and males of Pongo; A and B, putative
females and males of Ramapithecus; C and D, putative females and males of Siuapithecus. Putative
females, especially, and males of Ramapithecus (A and B) fall closest to Horm (H); axes 3 and 4 make it
clear that any apparent similarity with Pun (3 and 4) in the first plot is spurious. Both putative sexes of
Siuupithecus (C and D) are closest to the large dimorphic apes, especially Pongo (5 and 6).
To the degree that dental cross-sectional areas provide information about the
areas that take up stresses of mastication in the body of each tooth, Ramapithecus
may have been somewhat more omnivorous (areas are similar to those in humans),
and Sivapithecus may have been somewhat more herbivorous (areas are similar to
those in apes) [Wu & Oxnard, 1983b; Oxnard, 198313, 19841. Thus, based on both
univariate statistical analyses of dentition and functional implications of the data,
Ramapithecus appears to be closer to a “Homo grade” and Sivapithecus to a “Pongo
grade” (but not, of course, to these particular genera). These terms are used here in
an heuristic sense only, rather than as a means of generating a cladogram. These
univariate results are, then, the background for the multivariate studies of ramapithecines presented here.
Canonical variates analyses. Figures 4 and 5 are plots of canonical variates
means for the seven groups described earlier for maxillary and mandibular analyses, respectively. The fossils are, additionally, interpolated into these diagrams.
For the first two axes in the maxillary analysis, the mean position of the
Ramapithecus putative females (A) was closest to Homo, and that of Ramapithecus
putative males (B) was closest to female Pan. The mean position of putative female
Sivapithecus (C) was closest to male Pongo, and that of putative male Sivapithecus
was between male and female Gorilla. The direction from female to male Siuapithecus was similar to that of females to males of each extant ape. But the direction
from female to male Ramapithecus was the reverse of that of extant apes. Perhaps
this will also be the case when data for humans of known sex are interpolated.
Morphometric Studies of Hominoid Teeth / 317
CANONICAL I7.4RL4TES IIEANS: MNDIBCL4R
CAVONICAL J'ARLiTES W L W S : MANDIBULAR
Fig. 5. Canonical variate means for each extant group and the interpolated positions of the putative
means for each sex of the fossil groups in the mandibular analysis. H, Homo; 1 and 2, females and males
of Gorilla; 3 and 4, females and males of Pun; 5 and 6, females and males of Pongo; A and B, putative
females and males of Runupithecus; C and D, putative females and males of Siuapithecus. Putative
females, especially, and males of Ramapithecus (A and B) fall closest to Homo (HI; axes 3 and 4 make it
clear that any apparent similarity with Pun (3 and 4) in the first plot is spurious. Both putative sexes of
Siuupithecus (C and D) are closest to the large dimorphic apes especially Pungo (5 and 6). These results
are considerably clearer than in the case of the maxillary analysis (Fig. 4).
For the first two axes in the mandibular analysis, the means for both putative
sexes of Ramapithecus were little closer to Homo than to Pan (but basically midway
between them). The mean for putative male Siuapithecus was very close to male
Pongo, whereas it was closer to Gorilla in the maxillary analysis. In this analysis,
the direction between the means of Siuapithecus putative males and females was
the reverse of that of extant apes. For both maxilla and mandible, there was greater
overall sexual dimorphism in Siuapithecus than i n Ramapithecus.
High-dimensional plots. As in the case of the extant genera, high-dimensional
plots were generated i n order to include information in higher axes. The importance
of this procedure is twofold. First, as with the extant species, i t allows us to view
multidimensional information in a way denied us by the more usual two- or threedimensional methods. However, because the fossils are interpolated into the analyses that use eigenvectors to which variances and covariances had not contributed,
it is not all unlikely that biologically important and statistically significant information may be parlayed into these axes for the fossil groups. This is so even though
the highest axes may contain little or no significant information for the extant
forms. This matter has been explored by Oxnard [1972, 19751.
Figure 6 shows the high-dimensional plots of the maxillary and mandibular
analyses for the fossils alone. Total sexual dimorphism (represented by the area
between the plots and degree of difference between the plots) in Ramapithecus was
larger in the maxilla than in the mandible, unlike that in extant apes. In contrast,
sexual dimorphism in Siuapithecus was greater in the mandible than in the maxilla, as in all extant great apes. But the basic difference may be a reflection both of
318 I Lieberman, Gelvin, and Oxnard
FOSSILS: XflILLARY DATA
FOSSILS: MANDIBULAR DATA
Fig. 6 . High-dimensional plots for each putative sex of each fossil in the maxillary and mandibular
analyses. The curves for Siuapithecus are placed above those for Ranapitheeus. In each case the curve
placed highest on the plot is the curve for the male. The patterns of each genus share certain highdimensional similarities, but the pattern of putative sexual dimorphism in each is different.
sexual dimorphism at different loci than in extant apes and reduced canine dimorphism relative to extant apes.
The separation between the two fossils was very clear, with no overlap in the
maxillary analyses and only very small overlap in the mandibular analyses. This
separation was greater by far than any sexual dimorphism in any extant ape,
further refuting the recent claims that the two fossils are really males and females
of the same species. Other anatomical evidence [Wu & Oxnard, 1983a,b; Etler,
1983; Wu & Olsen, 19851 further supports the taxonomic distinction between the
two.
Figures 7-9 combine the high-dimensional plots of the fossils (Fig. 6 )with those
of the extant species (Fig. 3). Figure 7 compares Sivapithecus with Gorilla and
Figure 8 compares Sivapithecus with Pongo. Sivapithecus was closer to Pongo than
to Gorilla, particularly in regard to the mandible.
Figure 9 compares the plots for Ramapithecus with those for Pan and Homo.
Ramapithecus was closer to Homo Sor both jaws, but somewhat less close to Homo
for the mandible than for the maxilla.
Previous univariate analyses [Oxnard et al, 19851 indicated that Siuapithecus
was like the extant apes in having unequal numbers of males and females and
fairly large sexual dimorphism of means at many tooth positions. The univariate
results also showed that Siuapithecus was specifically like Pongo in having equal
dispersions for each dimension in each sex. It is of interest that zero dispersion
differences between the sexes were also characteristic of Homo, although this might
have been due simply to the small dimorphism in all features that generally
characterize Homo. The multivariate statistical results reported here (from canonical variates analyses and high dimensional plots) confirm that Siuapithecus shows
marked sexual dimorphism and is similar overall to Pongo. The results also indicate however, that Siuapithecus differs from any extant ape in having major sexual
dimorphism in the posterior teeth and small specifically canine dimorphism.
Morphometric Studies of Hominoid Teeth / 319
GORILL4 AND SIVAPITIIECCS: MAXILLiRY
GORILW AYD SIVAPITHECUS: MPuUDlBULm
’2:
I l l
I 0’
Fig. 7. High-dimensional comparisons for Gorilla and Siuupithecus. Again, in each case, the curve for
each pair that is placed highest on the plot is the curve for the male. Although there are, of course,
differences of the fossil from Gorilla, the fossil overlaps that genus considerably so that it is not easy to
see the differences between them.
PONGO AND SIVAPITHECUS: MAXILLARY
PONGO AND SNAPITHECUS: MANDIBULAR
-1:
-”
Fig. 8. High-dimensional comparisons for Pongo and Siuupithecus. Once again, the curve for each pair
that is placed highest in the plot is the curve for the male. Although there are differences between the
fossil and Pongo, the fossil overlaps that genus even more than was the case for Gorilla (Fig. 7).
Both univariate and multivariate results indicate that Ramapithecus resembles no extant ape. Like Homo, Ramapithecus has small sexual dimorphisms a t
fewer tooth positions, equal dispersions for each sex, a 1:l female to male ratio, and
canines not much larger than incisors. It differs from humans in having slightly
less dental sexual dimorphism overall and in having that dimorphism present a t
different tooth positions.
320 I Lieberman, Gelvin, and Oxnard
PA??, HOMO, RXX4PITHECUS: MAXILL4Rk’
PAN, HOMO, RAhUPITHECUS: MANDIBULAR
Fig. 9. High-dimensional comparisons for the sexes of Pan, for Homo, and for the interpolated means for
each putative sex Ramapithecus. Although there is a vague overall similarity among all three genera,
Ramapithecus has a generally horizontal plot that is similar to the horizontal plot for Homo. It is quite
different from the markedly bipeaked plot for Pan. The differences of Ramapithecus from even Homo is
evident in the high-dimensional elements represented by the small high frequency waves in the plot for
Ramapithecus.
It appears, therefore, that sexual dimorphism among hominoids may have
existed in a number of states. One state consisted of large mean dimorphism, large
dispersion dimorphism, large differential between numbers of females and males,
large differences between the two jaws, and large ratios of back teeth areas to front
teeth areas. All of these features were found in extenso in Gorilla and, except for
somewhat smaller mean dimorphism, also in Pan. They were not found in any
other species examined here. That places the African great apes together, and given
the dates for the fossils, suggests that these features are derived more recently in
the African apes.
A second state consists of small (or zero) dispersion differences together with
large differences for each of the other characteristics enumerated above. This is the
case for both Pongo and Siuapithecus and suggests that these may be conservative
characteristics that have been around in a particular radiation for a t least 8 million
years. It is unlikely to represent a single lineage because the chances that Siuapithecus is actually ancestral t o Pongo must be vanishingly small.
A third state consists of small differences for every character enumerated above.
It is evident in both Homo and Ramapithecus. This could mean that these features
have been characteristic of a radiation of a t least some hominoids, again for about 8
million years. And again, this combination is unlikely to represent a single lineage
because the chances that Ramapithecus is actually ancestral to Homo must also be
vanishingly small.
One way of examining these groups of characteristics is as follows. The ancestral
state might be represented by large mean sexual dimorphism, as in Pongo, Gorilla,
and Siuapithecus. The reduced sexual dimorphism of Pan, Homo, and Ramapithecus
may represent the more recently evolved state. But the reverse of the above is also
possible, with small sexual dimorphism being the ancestral condition. Both of these
possibilities contain totally unlikely combinations of genera.
Morphometric Studies of Hominoid Teeth / 321
A second way of displaying these characteristics is as follows. The ancestral
state might be represented by large dispersion sexual dimorphism, as in Gorilla and
Pan. The reduced dispersion sexual dimorphism in Homo, Pongo, Ramapithecus,
and Siuapithecus may be the more recently evolved state. The reverse condition is
presumably possible, and again, both of these possibilities contain unlikely combinations of genera.
However, there is also the question of sex ratio. Is 2:l the original ratio, with
1:l derived @ace the finding in the 8-million-year-old Ramapithecus)? Is 1:l the
original ratio (pace the finding of 2:l in the 8-million-year-oldSiuapithecus)?
The whole question is complex and cannot be easily accommodated by any of
the conventional ways of looking a t hominoid fossils without assuming multiple
appearances of similar complex combinations of traits a t different times. It is likely
to be the case that when univariate and multivariate studies like these are performed on species such as Homo erectus, H. habilis, and the various australopithecines, interrelationships will be easier to discern. Preliminary data already suggest
that H. habilis has small dimorphism and equal bimodal distributions for those of
its dental dimensions that can be examined in this way. This is not a n inappropriate
finding for such a species. Preliminary data also suggest, however, that these
features of contemporaneous australopithecines are different [Oxnard, 1985a,b].
It does appear likely that small sexual dimorphisms of means, dispersions, and
interjaw differences, together with equal ratios between the sexes, are the generalized situation or template, from which increased selection pressures can mold increased sexual dimorphisms of various kinds and increased female to male ratios.
Other recent work also suggests that sexual dimorphism is complex and that greater
sexual dimorphisms are of more recent origin in hominoids [Blumenberg, personal
communication].
The possible evolutionary links between the Chinese ramapithecines and modern forms must be examined. Based on tooth morphology and coefficients of variation, several authors believe that ramapithecines are ancestral to hominids [de
Bonis, 1983; Kay, 1982a,b; Kay & Simons, 19831. On the other hand, ape status for
both ramapithecine groups is argued by others, based not only upon anatomical
information but also upon ecological and biogeographical interpretations [Bernor,
1983; Andrews & Cronin, 1982; Ward & Kimbel, 1983; Ward & Pilbeam, 19831. The
above multivariate statistical analyses indicate that Siuapithecus is clearly more
like Pongo, while Ramapithecus is consistently more like Homo. Neither is like the
African apes. It is possible that Sivapithecus, with its more ape-like features and
marked sexual dimorphism, could have been close to ancestral ape lineages, most
likely orangutans or their progenitors. Ramapithecus, with its more humanlike
features, could be related to very early prehuman lines; however, that assimption
appears to be negated by presumed later links of prehumans with African apes
based on two decades of biomolecular work.
Thus, most of the earlier work utilizing the concept of the biomolecular clock
[reviewed in Goodman et al, 19831 has placed the common humanAfrican ape
ancestor at up to 5 million years ago. If this date is even marginally correct, and if
Ramapithecus is ancestral to humans, Ramapithecus would also have to be ancestral
to the African apes. The series of structural reversals in the complex of sexual
dimorphisms that would be necessary for such a phylogeny make it unlikely.
Newer biomolecular clock results [Goodman et al, 1983; Cronin, 19831 dated the
human-African ape split at 7.6 to 8 million years ago. Even more recently, DNADNA hybridization studies [Sibley and Ahlquist, 19841 placed it even further back
in time (upper limit 10 million years ago). Gingerich’s 119851 most recent estimations
allowed the upper limit to be as early as 13 million years ago. Considering all the
322 I Lieberman, Gelvin, and Oxnard
assumptions and weaknesses of the biomolecular clock concept leg, see Lewin, 19851,
particularly those that underestimate evolutionary time (failure to include the
effects of regulatory genes, failure to account for the role of selection, failure to
account for varying rates of point mutations, failure to account for chromosomal
mutations, failure to allow for biomolecular redundancy are only a few examples), it
is more likely that the actual date approaches or may be even earlier than the
estimates referenced above.
Since Ramapithecus and Sivapithecus from China (not, it should be noted, the
meager remains of much earlier ramapithecines from the rest of the world) were
sympatric approximately 8 million years ago, a date for the human-African ape split
at 10 million years or even earlier is consistent with the findings of this study: that
Ramapithecus is closely similar to hominids but not to the African pongids. It is
possible that the cladogenesis that led to Ramapithecus was preceded by the humanAfrican ape divergence.
It is irrelevant to ask whether or not this particular Ramapithecus is ancestral
to humans; indeed, it is irrelevant to ask whether any specific fossil of this approximate age i s ancestral to an extant form. The statistical chances involved in the
accidents of fossilization almost always preclude any specific form from being such
a n ancestor. Modern evolutionary ideas, such as punctuated events, suggest that
little or no fossil record of actual cladogenesis is available. This lack of a fossil record
of any “transition” is consistent with the “island“ theory of allopatric speciation
and punctuated equilibrium theory, and should be viewed as neither artifact nor
anomaly [Eldredge & Gould, 1972; Gould & Eldredge, 1977; Minkoff, 19831.
In the past, Pongo has been thought to be highly specialized as compared with
the African apes, but similarities between Pongo and Sivapithecus (noted here)
require that this idea be reexamined. Anatomical similarities between Siuapithecus
and Pongo have been found [Wu & Olsen, 19851. Similarities have also been suggested by others [eg, Ward & Pilbeam, 19831 on the basis of the examination of much
older Siuapithecus from other geographic localities. The large canine dimorphism of
extant African apes was probably not the general hominoid situation. Rather, large
dimorphisms in mean size of canines for each sex, especially when combined with
large dispersion differences for canine dimensions of each sex, are probably recently
evolved features that are not typical of either earlier hominid or hominoid ancestors.
Indeed, some orangutan characters (e.g., zero dispersion differences between the
sexes) may be shared with humans because they are more general for hominoids.
This is consistent with anatomical views of the orangutan as a “living fossil”
[Walker as reported by Lewin, 19831.
Since a “living fossil” is a species that has undergone little or no anagenesis
over a period of evolutionary time, further studies of the orangutan should have
important implications for hominid evolution. Many characters found in the orangutan can be assumed to be more similar to early hominids than are characters found
in other extant apes, such as Gorilla and Pan, genera that may have undergone both
more anagenesis and more cladogenesis. Other work, employing different characters, supports the idea that African apes are the most derived hominids [Ward &
Kimbel, 1983; Templeton, 1983; Etler, 19831. Templeton [ 19831, employing restriction endonucleases, found that horninids lay between Pongo and the African apes.
Living hominoids, with the exception of Pongo, are considered to be atypical relicts
[Corruccini & Ciochon, 19831.
Finally, it is important to reevaluate the role of structural dimorphism in
human evolution. It is apparent from this and prior studies from this laboratory
that structural sexual dimorphism is not a single primate pattern with a possibly
rigid genetic basis and little evolutionary plasticity. If structural differences have
Morphometric Studies of Hominoid Teeth / 323
differentiated as recently as the time of the human-African ape split (only 10 million
years or more ago), then sexual dimorphism has changed in parallel within each
hominoid lineage, with concomitant changes in within-sex variability. That proportion of sexual dimorphism that is heritable is therefore much newer and more
modifiable than has previously been assumed. Other recent work supports the model
of sexual dimorphism as highly modifiable [Blumenberg, personal communication).
Sexual dimorphism can no longer be thought of as a single dimension or continuum,
slowly changing over long periods of time in a consistent direction, but rather as a
complex phenomenon resulting from multiple factors and multiple selection pressures in relatively short time periods. The genetic basis of any nonstructural sexual
dimorphism (such as behavior) is probably even smaller, younger, more plastic, and
more modifiable than has been previously supposed.
CONCLUSIONS
1. The multivariate statistical investigation reported here confirmed the findings from univariate study of extant ape and human teeth [Oxnard et al, 19851 that
sexual dimorphism in hominoid teeth is not a simple unidimensional phenomenon
related to size differences but is complex and involves a different pattern in each
species. Thus, the sexes of Pan, Pongo, and Gorilla (clearly identifiable as bimodal
distributions in univariate data) were markedly separated from each other multivariately in different ways. Of course, all were markedly different from Homo. These
distance and pattern differences are further strong evidence in support of the hypothesis of multiple causality of sexual dimorphism, even in so limited an anatomical region as the dentition.
2. The investigation also showed the “ramapithecine” teeth from Lufeng in
China formed two statistically completely separate groups as judged by multivariate
statistical analysis. This, too, confirmed the findings from prior univariate statistical
studies [Wu & Oxnard, 1983a and b].
3. One of the “ramapithecine” groups consisted of the cluster of specimens
identified as Siuapithecus. These lay close to the extant apes, especially Pongo.
4. The second “ramapithecine” group consisted of the cluster of specimens
identified as Ramapithecus. These lay close to extant Homo and away from extant
apes.
5. The investigation further showed that the bimodal distributions that existed
for each of these subgroups in tooth dimensions examined univariately also existed
in a multivariate state in a form that indicated that they likely represented sexual
dimorphism.
6. The multivariate dimorphism in Siuapithecus resembled that which exists
in the living apes, especially Pongo.
7. The pattern of multivariate dimorphism in Ramapithecus differed markedly
from that in the apes but was almost as small as that existing in Homo.
8. High-dimensional analyses, which displayed all canonical axes together and
thus contained additional information that was not pertinent to the living species,
further confirmed the complete difference between Sivapithecus and Ramapithecus,
thus demonstrating that these two groups of specimens could not be the two sexes of
a single “ramapithecine” species.
9. High-dimensional analyses also showed, notwithstanding the overall resemblances of Sivapithecus to the living apes, especially Pongo, that there were features
of Sivapithecus that rendered it different from even Pongo. These differences were
in higher canonical directions in which the living forms showed little or no significant separations but in which Sivapithecus exhibited significant separations.
324 I Lieberman, Gelvin, and Oxnard
10. High-dimensional analyses showed in addition notwithstanding the overall
resemblances of Ramapithecus to living Homo, that there were features of Ramapithecus that rendered it different from Homo. These differences also were in higher
canonical directions in which the living forms showed little or no significant separations but in which Ramapithecus showed significant separations.
11. These findings have implications for higher primate phylogeny because
they demonstrate the existence eight million years ago of (a) a complex pattern of
features in which one of these groups, Siuapithecus, has marked similarities with
the extant genus Pongo, and (b) another complex pattern which is quite dissimilar
to that of any ape but which has many similarities with modern Homo. The standard
molecular clock assumption (that the link of humans with African apes was 6 million
years ago) is incompatible with the last conclusion. But there are sufficient discrepancies among the many different molecular clock estimates (in 1985 alone they
ranged from just under 2 million years to over 12 million years) that this need not
be the case.
12. Equivalent univariate and multivariate studies are needed of those fossils
(fossil Homo and australopithecines) intermediate in time between the extant species and these eight million year old “ramapithecines.” (These particular “ramapithecines” are, themselves, different from “ramapithecines” [sivapithecines] known
from earlier dates elsewhere in the world).
13. These results have, finally, considerable implications for the study of the
evolution of sexual differences in higher primates. The existence of many complex
patterns of difference between the sexes rather than simple differences along a
single unidimensional axis of sexual separation in each of these living and fossil
species implies that sexual dimorphisms are markedly more plastic than we had
previously thought. They must have changed a great deal in different ways in
different evolutionary lines within only the last few million years.
ACKNOWLEDGMENTS
Our thanks are due to Professor F.P. Lisowski of the Department of Anatomy,
University of Tasmania for much helpful discussion of the research and this manuscript, and to Professor Wu Rukang of the Institute of Vertebrate Paleontology and
Paleoanthropology, Academia Sinica, Beijing, for the provision of data from the
Lufeng fossils. The investigations are supported by University Research Grants, a
Faculty Research and Innovation Fund Grant, and N.I.H. Biomedical Research
Support Grants RR 05351 and 05356 to C.E.O. from the University of Southern
California.
REFERENCES
Andrews, D.F. Plots of high-dimensional data.
BIOMETRICS 28:125-136,1972.
Andrews, D.F. Graphical techniques for highdimensional data. pp 37-59 in DISCRIMINANT ANALYSIS AND APPLICATION. T.
Cacoullos, ed. New York, Academic Press,
1973.
Andrews, P.; Cronin, J.E. The relationship of
Siuapithecus and Ramapithecus and the evolution of the orang-utan. NATURE
297~541-546,1982.
Bernor, R.L. Geochronolom and zoogeographic relationships of Miocene Homkoidea. pp 21-66 in INTERPRETATIONS O F
APE AND HUMAN ANCESTRY. R.L. Ciochon, R.S. Corruccini, eds. New York,
Plenum, 1983.
Biggerstaff, R.H. The biology of dental genetics. YEARBOOK OF PHYSICAL ANTHROPOLOGY 22~215-227,1979.
Butler, P.M. Studies of the mammalian dentition. Differentiation of the postcanine dentition. PROCEEDINGS OF THE ZOOLOGICAL SOCIETY, LONDON 109:l-36,
1939.
Butynski, T.M. Harem-male replacement and
infanticide in the blue monkey (Cercopithecus mitis stuhlrnunni) in the Kibale forest,
Uganda. AMERICAN JOURNAL OF PRIMATOLOGY 3:l-22, 1982.
Camubell. 8. SEXUAL SELECTION AND
T H k DESCENT OF MAN. Chicago, Aldine,
1972.
Clutton-Brock, T.H.; Harvey, P.H.; Rudder,
B. Sexual dimorphism, socioeconomic sex
ratio, and body weight in primates. NATURE 269:797-800,1977.
Corruccini, R.S.; Ciochon, R.L. Overview of
ape and human ancestry: Phyletic relationships of Miocene and later Hominoidea. pp
Morphometric Studies of Hominoid Teeth / 325
3-20 in INTERPRETATIONS OF APE
AND HUMAN ANCESTRY. R.L. Ciochon,
R.S. Corruccini, eds. New York, Plenum,
1983.
Cronin, J.E. Apes, humans and the molecular clock. pp. 115-149 in INTERPRETATIONS OF APE AND HUMAN ANCESTRY. R.L. Ciochon, R.S. Corruccini, eds.
New York, Plenum, 1983.
de Bonis, L. Phyletic relationships of Miocene hominoids and higher primate classification. pp 625-649 in INTERPRETATIONS OF APE AND HUMAN ANCESTRY. R.L. Ciochon, R.S. Corruccini, eds.
New York, Plenum, 1983.
Drickamer, L. A ten year summary of reproductive data for free-ranging Macaca mulatta. FOLIA PRIMATOLOGICA (Basel)
21%-80, 1974.
Eldredge, N.; Gould, S.J. Punctuated equilibria: a n alternative to phyletic gradualism.
pp 82-115 in MODELS IN PALEOBIOLOGY. T.J.M. Schopf, ed. San Francisco,
Freeman, 1972.
Etler, D. The Lufeng fossil ape site and the
place of the Chinese sivapithecines in hominoid phylogeny. M.Sc. dissertation, University of Wisconsin-Milwaukee, 1983.
Frayer, D.W. Gigantopithecus and its relationship to Australopithecus. AMERICAN
JOURNAL OF PHYSICAL ANTHROPOLOGY 39:413-429,1973.
Friedman, R.C.; Richart, R.M.; Vande Wiele,
R.L. SEX DIFFERENCES IN BEHAVIOR.
London, Wiley, 1974.
Garn, S.M.; Lewis, A.B.; Swindler, D.R.; Kerewsky, R. Genetic control of sexual dimorphism in tooth size. JOURNAL OF
DENTAL RESEARCH 46:963-972,1967.
Gautier-Hion, A. Dimorphisme sexuel et organisation sociale chez les cercopithecine forestiers africains. MAMMALIA 39:365-374,
1975.
Geist, V. On the relationship of social evolution and ecology in ungulates. AMERICAN
ZOOLOGIST 14:205-220, 1974.
Gingerich, P.D. Nonlinear molecular clocks
and the divergence of apes and humans.
PROCEEDINGS, TAUNG DIAMOND JUBILEE SYMPOSIUM, Johannesberg, 20,
1985.
Goodman, M.; Baba, M.L.; Darga, L.L. The
bearing of molecular data on the cladogenesis and times of divergences of hominoid
lineages. pp 67-86 in INTERPRETATIONS
OF APE AND HUMAN ANCESTRY. R.L.
Ciochon, R.S. Corruccini, eds. New York,
Plenum, 1983.
Gould, S.J.; Eldredge, N. Punctuated equilibria: The tempo and mode of evolution reconsidered. PALEOBIOLOGY 3:115-151,1977.
Gregory, W.K.; HeIlman, M. The dentition of
Dryopithecus and the origins of man.
AMERICAN MUSEUM OF NATURAL
HISTORY ANTHROPOLOGICAL
PAPERS 2%-123,1926.
Hrdy, S.B. THE WOMAN THAT NEVER
EVOLVED. Cambridge, Harvard, 1981.
Kay, R.F. Siuapithecus sirnonesi, a new species of Miocene hominoid with comments on
the phylogenetic status of the Ramapithecinae. INTERNATIONAL JOURNAL OF
PRIMATOLOGY 3:113-174,1982a.
Kay, R.F. Sexual dimorphism in Ramapithecinae. PROCEEDINGS OF THE NATIONAL ACADEMY OF SCIENCES OF
THE USA 79:209-212,1982b.
Kay, R.F.; Simons, E.L. A reassessment of
the relationship between later Miocene and
subsequent Hominoidea. pp 577-624 in
INTERPRETATIONS OF APE AND HUMAN ANCESTRY. R.L. Ciochon, R.S. Corruccini, eds. New York, Plenum, 1983.
Leutenegger, W. Scaling of sexual dimorphism of body weight and canine size in primates. FOLIA PRIMATOLOGICA 37:163176,1982.
Leutenegger, W.; Cheverud, J. Correlates of
sexual dimorphism in primates: EcoloLrical
and size variables. &INTERNATIONAL
JOURNAL OF PRIMATOLOGY 37:387402, 1982.
Leutenneger, W.; Kelly, J.T. Relationships of
sexual dimorphism in canine size and body
size to social, behavioral, and ecological correlates in anthropoid primates. PRIMATES
18:177-186, 1977.
Lewin, R. Is the orangutan a living fossil?
SCIENCE 222122-123, 1983.
Lewin, R. Molecular clocks scrutinized. SCIENCE 228571,1985.
Mahler, P.E. Metric variation in the pongid
dentition. PhD Dissertation, University of
Michigan, Ann Arbor, 1973.
Meikle, D.B.; Tilford, B.L.; Vessey, S.H. Dominance rank, secondary sex ratio, and reproduction of offspring in polygynous primates.
THE AMERICAN NATURALIST 124:173188,1984.
Minkoff, E.C. EVOLUTIONARY BIOLOGY.
New York, Addison-Wesley, 1983.
Montagu, A. THE NATURAL SUPERIORITY OF WOMEN. New York, Colliers, 1974.
Napier, J.R.; Napier, P.R. A HANDBOOK
OF THE LIVING PRIMATES. London. Academic Press, 1967.
Oxnard, C.E. Some African fossil foot bones:
A note on the interpolation of fossils into a
matrix of extant species. AMERICAN
JOURNAL OF PHYSICAL ANTHROPOLOGY 37:3-12. 1972.
Oxnard, C.E. UNIQUENESS AND DIVERSITY IN HUMAN EVOLUTION: MORPHOMETRIC STUDIES OF AUSTRALOPITHECINES. Chicago, University Press,
1975.
Oxnard, C.E. Anatomical, biomolecular and
morphometric views of the living primates.
326 / Lieberman, Gelvin, and Oxnard
pp. 113-142 in PROGRESS IN ANATOMY.
Harrison, R.J., and Navaratnam, V. Cambridge, University Press, 1983a.
Oxnard, C.E. THE ORDER OF MAN: A
BIOMATHEMATICAL ANATOMY OF
THE PRIMATES. Hong Kong, University
Press, 1983b, New Haven, Yale, 1984.
Oxnard, C.E. Sexual dimorphisms in the
overall proportions of primates. AMERICAN JOURNAL OF PRIMATOLOGY 4:l22,1983~.
Oxnard, C.E. Hominids and hominoids, lineages and radiations. In PAST, PRESENT
AND FUTURE OF HOMINID EVOLUTION. P. V. Tobias, ed. New York, Liss,
1985a.
Oxnard, C.E. Humans, apes and Chinese fossils: New implications for human evolution.
HONG KONG UNIVERSITY PRESS OCCASIONAL PAPERS’ SERIES 4:l-46,
1985b.
Oxnard, C.E.; Lieberman, S.S.; Gelvin, B.R.
Sexual dimorphisms in dental dimensions
of higher primates. AMERICAN JOURNAL OF PRJMATOLOGY. 8:127-152,1985.
Pilbeam, D. Tertiary Pongidae of East Africa:
Evolutionary relationships and taxonomy.
PEABODY MUSEUM BULLETIN, YALE
UNIVERSITY, NO 31,1969.
Schultz, A.H. THE LIFE OF PRIMATES.
New York, Universe Books, 1969.
Sibley, C.G.; Ahlquist, J.E. The phylogeny of
the hominoids as indicated by DNA-DNA
hybridization. JOURNAL OF MOLECULAR EVOLUTION. 20:2-22,1984.
Templeton, A.R. Phylogenetic inference from
restriction endonuclease cleavage site maps
with particular reference to the evolution of
humans and apes. EVOLUTION 37221224,1983.
Ward, S.C.; Kimbel, W.H. Subnasal alveolar
morphology and the systematic position of
Siuapithecus. AMERICAN JOURNAL OF
PHYSICAL ANTHROPOLOGY 61:157-171,
1983.
Ward, S.C.; Pilbeam, D.R. Maxillofacial morphology of Miocene hominids from Africa
and Indo-Pakistan. pp 211-238 in INTERPRETATIONS OF APE AND HUMAN ANCESTRY. R.L. Ciochon. R.S. Corruccini. eds.
New York, Plenum, 1983.
Wolpoff, M.H. Metric trends in hominid dental evolution. CASE WESTERN RESERVE
UNIVERSITY STUDIES IN ANTHROPOLOGY 2:1-244,1971a.
Wolpoff, M.H. InterstitiaI wear. AMERICAN
JOURNAL OF PHYSICAL ANTHROPOLOGY 34:205-225,1971b.
Wood, B.A. An analysis of sexual dimorphism in primates. Ph.D. Dissertation, University of London, 1975.
Wood, B.A. The nature and basis of sexual
dimorphism in the primate skeleton. JOURNAL OF ZOOLOGY, LONDON 180-15:34,
1976.
Wu, R.; Oxnard, C.E. Ramapithecines from
China: Evidence from tooth dimensions.
NATURE 306:258-260,1983a.
Wu, R.; Oxnard, C.R. Ramapithecus and Siuupithecus from China: Some implications
for higher primate evolution. AMERICAN
JOURNAL OF PRIMATOLOGY 5:303-344,
1983b.
Wu, R.; Olsen, J. PALAEOANTHROPOLOGY AND PALAEOLITHIC ARCHAEOLOGY IN THE PEOPLE’S REPUBLIC OF
CHINA, London, Academic Press, 1985.
Wu, R.; Xu, Q.; Lu, Q. Morphological features
of Ramapithecus and Siuapithecus and their
phylogenetic relationships. ACTA ANTHROPOLOGICA SINICA 2:l-14,1983.
Wu, R.; Han, D.; Xu, Q.; Lu, Q.; Pan, Y.;
Zhang, H.; Zheng, L.; Xiao, M. Ramapithe
cus skulls found first time in the world.
KEXUE TONGBAO 26:1018-1021,1981.
Wu, R.; Han, D.; Xu, Q.; Qui, G.; Lu, Q.; Pan,
Y.; Chen, W. More Ramapithecus skulls
from the Lufeng, Yunnan-Report on the
excavation at the site in 1981. ACTA ANTHROPOLOGICA SINICA 2:101-108,1982.
~~~
Документ
Категория
Без категории
Просмотров
5
Размер файла
1 527 Кб
Теги
china, ramapithecines, approach, hominoid, dental, sexual, dimorphisms, extant, quantitative
1/--страниц
Пожаловаться на содержимое документа