Estimating body mass in New World "monkeys" (Platyrrhini, Primates), with a consideration of the Miocene platyrrhine, Chilecebus carrascoensis

2008·1.8 MB·English

by  Sears, Karen E.

A tamerican museum
Novitates
PUBLISHED BY THE AMERICAN MUSEUM OF NATURAL HISTORY
CENTRAL PARK WEST AT 79TH STREET, NEW YORK, NY 10024
Number 3617, 29 pp., 1 figure, 6 tables June 16, 2008
Estimating body mass in New World “monkeys”
(Platyrrhini, Primates), with a consideration of the
Miocene platyrrhine, Chilecebus carrascoensis
KAREN E. SEARS,1 JOHN A. FINARELLI,2’3 JOHN J. FLYNN,4 AND
ANDRE R. WYSS5
ABSTRACT
Well-constrained estimates of adult body mass for species of fossil platyrrhines (New World
“monkeys”) are essential for resolving numerous paleobiological questions. However, no
consensus exists as to which craniodental measures best correlate with body mass among extant
taxa in this clade. In this analysis, we analyze 80 craniodental variables and generate predictive
equations applicable to fossil taxa, including the early platyrrhine Chilecebus carrascoensis.
We find mandibular length to be the best craniodental predictor of body mass. There is no
significant difference in predictive value between osteological and dental measures. Variables
associated with the mandible and lower dentition do significantly outperform the cranium and
upper dentition. Additionally, we demonstrate that modern platyrrhines differ, morphometrically,
from early fossil forms. Chilecebus possesses unusual cranial proportions in several key features, as
well as proportionally narrow upper incisors and wide upper cheek teeth. These variables yield
widely divergent body mass estimates for Chilecebus, implying that the correlations observed in a
crown group cannot be assumed a priori for early diverging fossils. Variables allometrically
consistent with those in extant forms yield a body mass estimate of slightly less than 600 grams for
Chilecebus, nearly a factor of two smaller than prior preliminary estimates.
Scaled to body mass, the brain of Chilecebus is markedly smaller than those of modern
anthropoids, despite its lowered body mass estimate advocated here. This finding, in conjunction
with a similar pattern exhibited by fossil catarrhines, suggests that increased encephalization arose
independently in the two extant subgroups of anthropoids (platyrrhines and catarrhines).
1 Department of Animal Biology, University of Illinois at Urbana-Champaign, Urbana, IL 61801.
2 Department of Geological Sciences, University of Michigan, Ann Arbor, MI 48109.
3 University of Michigan Museum of Paleontology, Ann Arbor, MI 48109.
4 Division of Paleontology, American Museum of Natural History, New York, NY 10024.
5 Department of Geological Sciences, University of California- Santa Barbara, Santa Barbara, CA 93106.
Copyright © American Museum of Natural History 2008 ISSN 0003-0082
2 AMERICAN MUSEUM NOVITATES NO. 3617
INTRODUCTION mass than do cranial or dental variables.
These elements are rare as fossils, however,
Motivated by the discovery of the well-
and when found as isolated elements can be
preserved cranium of the early, fossil platyr-
difficult to identify to low taxonomic levels
rhine (New World “monkey”) Chilecebus
(Damuth and MacFadden, 1990). The most
carrascoensis (Flynn et al., 1995), we explore common primate fossils by far are isolated
methods for estimating body mass in fossil
teeth or jaws, elements that typically correlate
platyrrhine species from cranial morphomet¬
with body mass more poorly (Gingerich et al.,
ries. Adult body mass is highly correlated with
1982; but also see Dagosto and Terranova,
a diverse suite of ecological, macroevolution¬
1992).
ary, physiological, and life history variables
Using morphometric proxies to estimate
among living mammals (Gittleman and
body mass in fossil taxa requires the poten¬
Harvey, 1982; Schmidt-Nielsen, 1984; Gittle¬
tially problematic assumption that the allo-
man, 1986; Damuth, 1987; McNab and Eisen-
metric relationship between a given variable
berg, 1989; Gittleman, 1991, 1993; Gillooly et
and body mass observed among living forms
al., 2005). It follows that improved estimators
also applies to fossil taxa. As an example,
of body mass for fossil taxa are essential to
estimates of body mass from different mor¬
understanding the biology of extinct species
phometric variables for the early catarrhine
(Gingerich and Smith, 1984; Martin, 1984;
Aegyptopithecus yield widely varying results
Damuth and MacFadden, 1990; Eisenberg,
(Radinsky, 1977; Martin, 1990; Dagosto and
1990; Jungers, 1990). However, it remains
Terranova, 1992; Simons, 1993). Thus, fossil
unclear which morphometric variables corre¬
taxa might be characterized by fundamentally
late best with body mass among living mam¬
different allometries of the skull and dentition;
mals, let alone among fossil taxa. Previous
the more unusual these morphometries are,
attempts to estimate body mass from various
the more uncertain body size estimates (and
skeletal and dental proxies in anthropoids
derived indices such as encephalization quo¬
have emphasized the Catarrhini (Old World
tients) become. In practice, such deviations are
“monkeys” and hominoids) (e.g., Delson et
difficult to identify because they require a
al., 2000), rather than platyrrhines (Hills and
robust and well-resolved phylogeny (of living
Wood, 1984; Conroy, 1987; Martin, 1990).
and fossil forms) and well-preserved and
Since the most reliable morphometric corre¬
relatively complete fossil material. To mini¬
lates of body mass are identified in the context
mize these potential problems with a single
of comparisons to phylogenetically closely
related taxa (Eaglen, 1984; Conroy, 1987; estimator, body mass estimates from several
Damuth and MacFadden, 1990; Dagosto proxies often are averaged (e.g., Radinsky,
and Terranova, 1992), we investigate these 1971; Martin, 1990). This approach makes the
associations in platyrrhine primates. optimistic assumption that inaccuracies among
There has been considerable research into variables offset one another, such that variables
various cranial, postcranial, and dental vari¬ yielding overestimates counterbalance those
ables as body mass estimators for extinct underestimating mass.
primates (e.g., Gingerich, 1974; Gingerich et To elucidate the pattern of increased en-
al., 1982; Gingerich and Smith, 1984; Bouvier, cephalization (brain volume scaled to body
1986; Gingerich, 1990; Martin, 1990; Ruff, mass) among platyrrhines, we apply a suite of
1990; Dagosto and Terranova, 1992; Spocter craniodental body mass estimators to the
and Manger, 2007). While a particular vari¬ 20.1 Ma fossil Chilecebus carrascoensis from
able’s reliability as an estimator of body mass the Abanico Formation of central Chile
for a fossil taxon depends on the strength of (Flynn et al., 1995). Represented by the best-
the correlation with body mass among extant preserved and -dated early primate skull from
taxa, those variables that tend to perform South America, Chilecebus provides an excel¬
better in a statistical sense are not typically the lent test case for application of these methods.
most useful to paleobiologists. Limb bone The completeness and exceptional preserva¬
measures generally correlate better with body tion allow many of the proxy variables to be
2008 SEARS ET AL.: ESTIMATING BODY MASS IN NEW WORLD ‘MONKEYS’ 3
applied. Moreover, the antiquity of this fossil TABLE l
allows methods of estimating body mass in Body mass for measured extant platyrrhine species
extinct taxa to be judged against the backdrop Masses given in grams, from Smith et al. 2003.
of potentially significant evolutionary change
in allometry. Genus Species Mass [g]
Alouatta caraya 5862.5
METHODS Aotus vociferans 873.0
Ateles geoffroyi 5284.9
Variables, Specimens, and Body Mass Data
Brachyteles arachnoides 13499.9
Cacajao melanocephalus 3800.0
We assembled a suite of 80 morphometric
Cacajao calvus 5796.0
variables from the literature (Gingerich and Callicebus moloch 854.7
Schoeninger, 1979; Radinsky, 1981a, 1981b, Callimico goeldii 480.0
1982; Gingerich et al., 1982; Anthony and Callithrix pygmaea 125.0
Kay, 1993; Hartwig, 1993; Kobayashi, 1995). Callithrix jacchus 292.0
Appendix 1 lists measurement variables and Cebus apella 2500.0
their descriptions. Of these potential body mass Chiropotes satanas 3000.0
Lagothrix lagotricha 6300.0
proxies, 32 were measurements of osteological
Leontopithecus rosalia 535.5
features on the cranium and mandible, while 48
Pithecia pithecia 1375.5
were measurements of the dentition. All teeth
Saguinus mystax 618.0
except the third molars (which do not occur in Saimiri sciureus 743.2
callitrichines) were measured. Specimens of
extant taxa examined were from the collections
of the Mammalogy Division, Zoology Depart¬ variables using ordinary (Model 1) least-squares
ment of the Field Museum of Natural History, regression (Sokal and Rohlf, 1995). Because
and the Mammalogy Collections, Division of similarity in morphometry may arise from close
Vertebrate Zoology of the American Museum phylogenetic relationship (Felsenstein, 1985),
of Natural History. In total, 157 specimens we performed phylogenetically corrected re¬
representing 17 platyrrhine species were mea¬ gressions (Garland and Ives, 2000), which were
sured. Specimen numbers are given in appen¬ executed in the PDAP (Midford et al., 2003)
dix 2. Extant taxa were chosen to ensure a module for the Mesquite software package
sample containing at least one species from each (Maddison et al., 2002, 2004). Phylogenetic
of the 15 platyrrhine genera and the full range of relationships among extant taxa were obtained
body sizes (from 13.5 kg for Br achy teles ara- from molecular phylogenies of the Platyrrhini
chnoides to 125 grams for Callithrix pygmaea). (Canavez et al., 1999; von Dornum and Ruvolo,
Body mass data were obtained from the Masses 1999) (figure 1). Slopes and intercepts describ¬
of Mammals Database (v. 3.03: Smith et al., ing the relationship between the morphologic
2003). Average body masses used in the predictor and body mass were calculated for
regression analysis are listed by species in each variable.
table 1. In addition, we measured preserved To assess a variable’s efficacy in predicting
morphometric variables for the fossil taxon body mass, we quantified the relative perfor¬
Chilecebus carrascoensis (SGOPV 3213, Museo mance among estimators using the log-likeli¬
Nacional de Historia Nacional, Santiago, hood fit of each regression model to the data:
Chile). Mean measurement values for extant
taxa are given by species in appendix 3.
Allometric Regression Analyses (Burnham and Anderson, 2002).
Morphometric variables and body masses for In the equation, n is the sample size, and ESS
the living platyrrhines were log-transformed is the error sum of squares of the regression;
(base 10) to determine allometric relationships therefore, the regression likelihood is propor¬
between the variables and body mass. Species tional to its ability to minimize residual
average masses were fit to species-average log- variance. Model log-likelihoods were rescaled
4 AMERICAN MUSEUM NOVITATES NO. 3617
Brachyteles arachnoides
Lagothrix tagotricha
Ateles geoffroyi
Alouatta caraya
Gacajao calvus
Gacajao mefanocephalus
Chiropotes satanas
Pithecia pithecia
Cailicebus moloch
Gallithrixjacchus
Gallithrix pygmaea
Callimico goeldii
Leontopithecus rosafia
Saguinas mystax
Aotus voctferans
Gebus apella
Saimiri sciumus
Fig. 1. Cladogram from molecular phylogenies (Canavez et al., 1999; von Dornum and Ruvolo, 1999) of
platyrrhine primates used in the phylogenetically corrected regressions of body mass on morphometric
variables. This cladogram is a synthetic topology of these two phylogenetic analyses. The topologies for the
two analyses were congruent for overlapping taxa, with two exceptions. First the Callicebus/Cacjaol
Chiropotes/Pitehca clade was basal to all other Platyrrhini in von Dornum and Ruvolo (1999), but was allied
with the Ateles/Bachy teles/Lagothrix/Allouata clade in Canavez et al. (1999). Second, Aotus was allied with
Saimiri and Cebus in Canvez et al. (1999), but left in an unresolved polytomy in von Dornum and Ruvolo
(1999). Therefore both nodes are conservatively left in polytomies here.
such that the maximum value across all hood difference of 2 (Edwards, 1992; Royall,
variables was 0, with poorer predictors having 1997; Wagner, 2000a, 2000b) as the cutoff for
progressively more negative log-likelihoods identifying one variable as a significantly
(Edwards, 1992). We employed a log-likeli¬ better predictor of body mass than another.
2008 SEARS ET AL.: ESTIMATING BODY MASS IN NEW WORLD ‘MONKEYS’ 5
Body Mass Estimation in Chilecebus calculated for the extant catarrhine and
platyrrhine taxa considered in Martin (1990).
The body mass of the fossil platyrrhine The encephalization quotient (EQ) is the ratio
Chilecebus carrascoensis was estimated from of observed brain volume to expected brain
46 measurements preserved in SGOPV 3213,
volume for a given body mass (Jerison, 1970;
the holotype cranium, using the regression
Radinsky, 1971). The metric of interest,
equations generated above from living taxa.
however, is not relative volume but deviation
We calculated a weighted average of these
of observed volume from the allometric
estimates using the proportional likelihoods for
regression of brain volume on body mass,
each regression model over the set of all models
i.e., the natural logarithm of the EQ (logEQ)
(Burnham and Anderson, 2002). Model aver¬
(Marino et al., 2004; Finarelli and Flynn,
aging is appropriate when there is no clearly
2007; Finarelli, 2008). Positive logEQs imply
identifiable “correct” model. This method has
larger than expected brain volumes for a given
been employed to a limited extent in previous
body mass, and negative values the opposite.
work (e.g., Radinsky, 1971; Martin, 1990),
To assess the relative brain volume for extant
where the mass estimates from several proxies
platyrrhines and the extinct taxon Chilecebus,
were averaged to reduce errors associated with
we used the allometry relating body mass to
any single predictor. However, simply calculat¬
brain volume for strepsirrhine primates (as an
ing the mean of the individual variable
outgroup to anthropoids) given in Martin
estimates implicitly assumes that all of the
(1990) as a frame of reference.
variables perform equally well in predicting
body mass. Since regression model likelihoods
RESULTS AND DISCUSSION
quantify an individual variable’s relative pre¬
dictive ability, a more appropriate averaging Body Mass Estimators
technique takes advantage of this information.
Using proportional likelihoods, poorly per¬ We performed phylogenetically corrected
forming models influence the final body mass regression of body mass on each of the 80
estimate less, whereas better fitting models morphometric variables for the extant taxa.
more strongly influence the final estimate. Slopes and intercepts for the predictive equa¬
tions and the corresponding log-likelihoods
for the phylogenetically corrected regressions
Estimating Relative Brain Size in Chilecebus
are given for each variable in table 2; scatter-
X-ray computed tomographic scans of the plots for the regressions are given in appen¬
Chilecebus carrascoensis cranium were per¬ dix 4. The relative ability of the morphometric
formed on the medical scanner at Children’s variables surveyed in this analysis to accurate¬
Hospital of San Diego (CT HiSpeed Adv ly predict body mass, as measured by the log-
SYS#HSA1; 140 Kv, 170 mA) to estimate likelihood fits for the regressions, is highly
endocranial volume (Flynn et al., 1995). The uneven. The maximum likelihood observed
skull and surrounding matrix were embedded in among the variables is for mandibular length,
modeling clay to better mimic the density of the indicating that this variable is the single best
human body, and scanned at 1.0 mm resolution. predictor of body mass among the 80 variables
The series of individual axial and transverse scan surveyed. Only one variable (pterygoid-zygo¬
slices were composited into 3-D images using matic length) had a log-likelihood fit within
Cemax VIP software® (Freemont, CA). The 2LnL units of mandibular length, and there¬
resulting brain volume measure was used in fore, there is no significant difference in the
conjunction with body mass estimates to deter¬ ability of those two variables to predict body
mine the relative brain volume in Chilecebus. mass among extant platyrrhine species.
To determine whether increased encephali- The two optimal predictors are both cra-
zation observed in extant anthropoids was niomandibular, osteological measurements,
inherited from their last common ancestor, or not dental variables. Teeth are frequently
whether New and Old World anthropoid employed as body mass predictors for fossil
primates evolved this trait independently, mammals (Gingerich, 1974; Legendre, 1986;
log-encephalization quotients (logEQs) were Legendre and Roth, 1988) because of their
6 AMERICAN MUSEUM NOVITATES NO. 3617
TABLE 2
Regressions of body mass on morphometric variables for extant taxa
Slopes and intercepts of predictive equations for phylogenetically corrected regressions for each measurement
variable (appendix 1) in living platyrrhines sampled (table 1). Also given are R2, error sum of squares (ESS)
and the log-likelihood (LnL), rescaled such that the maximum likelihood observed among models (mandibular
length) is 0. Measurement regressions are ordered in decreasing LnL. Teeth are upper (U) or lower (L) lengths,
widths (anterior or posterior for molars), and areas.
Measurement Slope Intercept R2 ESS LnL
Mandibular length 2.689 -1.349 0.636 0.117 0.000
Pterygoid-zygomatic length 2.268 0.210 0.592 0.135 -1.136
Bizygomatic width 2.828 -1.616 0.626 0.166 -2.783
Moment arm of the masseter 1.672 0.916 0.510 0.175 -3.225
LM1L 2.205 1.822 0.449 0.195 -4.077
Skull length 3.123 -2.720 0.537 0.196 -4.105
Temporal fossa length 3.202 -2.233 0.374 0.198 -4.193
LM1A 1.136 1.856 0.431 0.202 -4.371
Moment arm of the temporalis 1.760 1.434 0.434 0.213 -4.786
Mandibular height 2.016 1.010 0.516 0.215 -4.841
LM2A 1.056 1.942 0.375 0.232 -5.470
Foramen magnum width 3.339 -0.253 0.387 0.232 -5.474
LP3L 1.767 2.419 0.374 0.234 -5.535
LMlWPos 2.187 1.955 0.350 0.239 -5.716
Bulla width 2.939 0.302 0.327 0.241 -5.781
Occipital width 2.488 -0.525 0.315 0.242 -5.803
LMlWAnt 2.269 1.968 0.353 0.244 -5.864
Condyle Width 2.614 2.008 0.447 0.244 -5.879
Cheek tooth row, upper 1.982 0.653 0.273 0.244 -5.880
Cheek tooth row, lower 2.054 0.471 0.359 0.245 -5.885
Facial height (length) 2.299 -0.075 0.532 0.248 -6.003
Foramen magnum area 1.483 0.199 0.277 0.249 -6.015
LP3A 0.893 2.340 0.323 0.249 -6.018
Braincase width 3.086 -1.814 0.551 0.250 -6.070
LP4L 1.935 2.265 0.325 0.251 -6.083
LP4A 0.980 2.192 0.303 0.251 -6.099
LP2L 1.760 2.395 0.506 0.252 -6.118
LM2WPos 1.952 2.109 0.356 0.252 -6.134
LM2WAnt 2.259 1.952 0.419 0.253 -6.169
Temporal chord 3.246 -1.778 0.183 0.255 -6.229
UP3L 1.866 2.326 0.306 0.261 -6.396
UP4L 2.028 2.220 0.313 0.262 -6.438
LM2L 1.957 1.960 0.496 0.263 -6.452
Braincase length 2.792 -1.345 0.288 0.270 -6.678
LP2A 0.844 2.355 0.456 0.272 -6.724
UP3A 0.932 2.138 0.275 0.273 -6.764
UP2W 1.752 2.117 0.384 0.274 -6.783
UP4A 1.015 2.009 0.267 0.277 -6.893
Palate length 2.025 0.388 0.394 0.281 -6.985
UP2A 0.877 2.254 0.422 0.287 -7.170
LP3W 1.704 2.316 0.412 0.296 -7.408
LP4W 1.830 2.205 0.209 0.297 -7.450
Foramen magnum height 2.476 0.754 0.264 0.303 -7.598
UM1A 1.080 1.747 0.399 0.309 -7.743
UP3W 1.782 2.006 0.383 0.309 -7.758
UM1L 2.078 1.875 0.393 0.310 -7.793
UM2WAnt 2.212 1.674 0.398 0.311 -7.813
UP4W 1.964 1.847 0.199 0.311 -7.813
UM2WPos 2.046 1.799 0.397 0.312 -7.826
2008 SEARS ET AL.: ESTIMATING BODY MASS IN NEW WORLD ‘MONKEYS’ 7
TABLE 2
0Continued)
Measurement Slope Intercept R2 ESS LnL
UM2A 0.955 1.965 0.395 0.313 -7.858
Palate width 1.830 1.034 0.400 0.315 -7.908
UMlWAnt 2.157 1.676 0.378 0.318 -7.998
Masseteric fossa length 1.778 0.544 0.291 0.320 -8.027
Maxillary incisor size 2.476 0.525 0.286 0.322 -8.086
LP2W 1.527 2.373 0.350 0.323 -8.102
UP2L 1.683 2.427 0.353 0.323 -8.118
UM2L 1.677 2.186 0.353 0.331 -8.315
UC1L 1.497 2.293 0.459 0.333 -8.355
UMlWPos 2.130 1.695 0.326 0.344 -8.624
LC1L 1.403 2.527 0.328 0.349 -8.728
UC1W 1.310 2.351 0.296 0.366 -9.116
UI2L 1.900 2.341 0.240 0.381 -9.432
Postorbital constriction 3.056 -1.526 0.234 0.395 -9.720
UI2W 1.584 2.394 0.293 0.418 -10.176
Bulla length 2.383 0.061 0.181 0.425 -10.308
Condyle length 2.493 0.950 0.094 0.430 -10.398
Frontal chord 2.163 -0.202 0.446 0.431 -10.409
LC1W 1.178 2.436 0.188 0.434 -10.469
Vertical face height 1.808 0.683 0.357 0.440 -10.585
Occipital chord 2.033 0.687 0.111 0.456 -10.877
LI2W 1.569 2.340 0.111 0.480 -11.284
UI1L 1.660 2.295 0.059 0.488 -11.407
Parietal chord 2.852 -1.035 0.089 0.499 -11.591
LI1L 1.670 2.692 0.230 0.500 -11.601
LI1W 1.427 2.489 0.141 0.515 -11.847
UI1W 1.417 2.391 0.243 0.535 -12.150
LI2L 1.511 2.663 0.025 0.540 -12.216
Skull vault height 1.723 0.573 0.006 0.558 -12.479
Orbit width 1.791 0.885 0.003 0.610 -13.192
Basal length of pterygoid fossa 0.928 2.093 0.092 0.646 -13.655
higher preservation potential. However, den¬ dental variable by at least 2 LnL units. While
tal measurements are generally considered to osteological features include the best predic¬
be less reliable estimators than osteological tors of body mass, they also include the three
variables (including measurements of the worst predictors: vault height, orbital width,
postcranium) (Gingerich et al., 1982; Dagosto and pterygoid length. Further, a Mann-
and Terranova, 1992), potentially reflecting Whitney test (Sokal and Rohlf, 1995) recovers
specialized adaptations to unusual diets/food no significant difference in the median rank
items or unusual food-processing methods log-likelihood scores across osteological and
(Gingerich and Smith, 1984). To test this as¬ dental partitions (U = 656, p > 0.05). Thus,
sumption, we partitioned the data into osteo¬ while median log-likelihood for the osteolog¬
logical (both cranium and mandible), and ical variables is slightly higher than that of
dental (both upper and lower) splits. dental variables (-6.45 versus -7.60), the
Eight of the 10 best predictors in this variance of log-likelihoods for osteological
analysis are osteological (table 2): mandibular variables also is higher. Thus, while the best
length, pterygoid-zygomatic length, bizygo¬ predictors are osteological, among platyrrhine
matic width, moment arm of the masseter, primates osteological characters do not uni¬
skull length, temporal fossa length, moment versally outperform dental ones.
arm of the temporalis, and mandibular height. Among dental variables, the best predictor
Among these, both mandibular length and of body mass is lower Ml length (table 2), a
pterygoid-zygomatic length exceed the best commonly used proxy for body mass in
AMERICAN MUSEUM NOVITATES NO. 3617
mammals (Van Valkenburgh, 1990; Janis et population sample for a species. It is therefore
al., 1998a, 1998b). Lower Ml area, another difficult to rigorously evaluate sexual dimor¬
commonly used proxy (Gingerich, 1974; phism or other forms of intraspecific variation.
Legendre, 1986; Legendre and Roth, 1988; However, large degrees of within-species vari¬
Finarelli and Flynn, 2006), was the next best ation can negatively impact the precision of
estimator and was not significantly different prediction allometries.
than lower Ml length (LnL difference = More troublesome would be if there were a
0.298). As noted above, the distribution of tendency for the range of variation to increase
log-likelihoods for dental variables was less or decrease as a function of body mass (e.g.,
dispersed than the distribution for osteological Rensch’s rule: Rensch, 1950). To assess this
variables. Eight dental measures (as opposed potential impact of intraspecific variation on
to two osteological) were within 2 LnL units model estimates for extant Platyrrhini, we
of the optimal dental score. However, among calculated the difference between log-trans¬
these, only upper cheek tooth row length was formed maximum and minimum values for the
from the upper dentition. A Wilcoxon signed- 17 species, yielding an observed proportional
rank test performed on 24 upper/lower pairs range (table 3). This range was not calculated
of dental measurements demonstrates that for Brachyteles due to its low sample size. The
variables from the lower dentition significant¬ observed intraspecific proportional ranges
ly outperform their upper dentition counter¬ average approximately 14% of the total
parts as predictors of body mass (W = 160, p observed variation across all taxa. In addition,
< 0.05). there is no association between the percentage
Thus, regression model likelihoods demon¬ of total range that the proportional range
strate: 1) that the single best predictor is represents, and the major variable partitions.
mandibular length, and 2) that the lower Mann-Whitney tests (Sokal and Rohlf, 1995)
dentition significantly outperforms the upper comparing osteological/dental splits (U = 641,
dentition in predicting body mass. Thus, it p > 0.05) and mandibular/cranial splits (U =
appears that predictors associated with the 818.5, p > 0.05) were not significant. We also
mandible may outperform those of the crani¬ regressed observed proportional range against
um among extant platyrrhines. To test this, the species average body mass for each taxon.
log-likelihoods were separated into mandibu¬ For each measurement variable, a t-test for
lar and cranial partitions, rather than osteo¬ the slope of the regression (Sokal and Rohlf,
logical (including the mandible) and dental 1995) was not significant; regression slopes
(including both upper and lower) partitions. ranged from -0.073 to 0.066 with a median
Median mandibular log-likelihoods were high¬ value of 0.013. Therefore, proportional ranges
er than cranial log-likelihoods (-6.12 versus do not vary as a function of species body size.
— 7.81), a significant difference (Mann-
Whitney test, U = 519, p < 0.05).
Body Mass Estimates for Chilecebus
Intraspecific variability must be a concern
when evaluating evolutionary allometries One of the main objectives in deriving
through species mean values. Platyrrhines, morphometric proxies for body mass among
and primates in general, are sexually dimorphic extant taxa is to apply the predictive equations
(Martin et al., 1994), a feature that leads to a to fossil taxa. Here we apply the body mass
range of variance that is discounted when proxies derived above to the fossil platyrrhine
averaging over species. When estimating body Chilecebus carrascoensis (Flynn et al., 1995).
mass for fossil taxa allometries must use species The only known specimen of this taxon (a skull,
averages. Sex identification for fossil mammals SGOPV 3213) lacks the mandible, and ptery¬
generally is not possible, except in rare goid-zygomatic length (the second best predic¬
instances (e.g., Coombs, 1975; Fleagle et al., tor variable) could not be measured. There are,
1980; Krishtalka et al., 1990; Van Valkenburgh however, 46 measurements that can be evaluat¬
and Sacco, 2002). This is due to the fact that ed on this nearly complete skull. Of these, 24
fossil taxa are often known from a limited variables are osteological and 22 dental. Body
number of specimens and rarely represent a mass estimates derived from these variables
2008
SEARS ET AL.: ESTIMATING BODY MASS IN NEW WORLD ‘MONKEYS’
co O
'o> ^ S s -inH'-*0l-0'o\f-NH-inH<oNoiOnt-NH'©d-ovo>o©>©oo<rN''-©Hcr<oi^-HO'©OT©t ^^t<-mnii^ni'nooi^ooO>'-(HNOOO<oNoro'0o
Oh ^ G OOOCN-hO-h-h-hO-h-hOOOOO-h 00000-h<N-h-h0-h
ooooooooooooooooooooooooooooooo
s
‘o0o1i (li- Oeh?I .5a«3 ^(Nohohr'oo^hO^'O^-bHNOhOOooHrH^momhHomtmo/iiHootObo'oH^'OohoOoK/KioOiO'OH^c^offhiortor^t
OCOh OOOmO-H-H-H'-HO'-HOOOOOOOOOOO-HOOOO-HOOO
•rC-HO C3W Ico 3 ooooooooooooooooooooooooooooooo
£G
tP5h .h ^ >i 0o\i0c\'o«oOoOnN~Oi'o\<oNm3'if^NrC-r<-lrH-—<NlOlOi~pitiNoiO<ON^^itn-o^o^oH—^H Chl<^VoO<roohiColHo—inlrOoOh^mHp<'©o©h!^'-
©©©-h-h©^h^h-h©©-h©©©^h©©-h©©^h*-h©-h-h—<<N^h©©
N g ooooooooooooooooooooooooooooooo
*
© *£ (Nco<nt^in(N|oiocot^ioco(NHtoo<30Htoo<NiorooiHto-Hi>^o-Hr'Oiio
a ^pt -©^ opo -pH p'o pm ^p< pN 'po <po )p< p^< oN p^ tc©o oco ©m p< Npr -po op^ p|- ro-i ^©'o —o; m—;o po ^—t;< —N; oo ro^
ooooooooooooooooooooooooooooooo
3 ^3
^■'OrtHtmir)M(fiH(soorfimoiOrtHt<OHt
bi>0i ~<a41 Cr^0o0o0^^toOo'Oo0'd0«-Noo0lT'OtO^o<cNNO '0-rO<0'b^t'bOi^t<OOl(N<0^^t(Nf'OHt^O
00^h^h^h0<N0-h<N-h-h OOOOOOOOO-h-hOO-h^h-^-hOO
OXhIC •^'a ooooooooooooooooooooooooooooooo
Cg Oh nhmNHOt<Hooiofoiooocoofo'OoOo\oMo<hnM<o'tomo\i(fN)N'Otr(tioNohH(s<oooiOnMooio\iofin^ocHo<HNiOc<-HoOiiio(rS'
prtfSHpprtppHpHpppPpHHppHppOHHNpPP
Ph ooooooooooooooooooooooooooooooo
U
p ^ •§ -S
P§0 .v3. O o<>nr<oOoof-OH-THtri<nio^oi-o'oor<tsior-\('^t|<cNi^>oni>Tnfi(^N'-H<'oitnr<HiTt|-oH(iS^oi^orH^^<Hoo<r'0i^(Soo^o<oo<oNoO(O'cHi-
! i * 0-h0<N-h-h<N<N-h00-h000-h0-h«N000-h-h0-h00-h-h0
►» o0 p>3~*Hi '■ « u ooooooooooooooooooooooooooooooo
«
"ft g <^t'od(-S'cHi-rt'('scio-o<rn(N|O'oTi^j<-oNoO'0''cni-mo-'Hci-r'no\ii^'0'00|'«nohcocro<io'>\o<^oi-o«rooooooi<r'sOhl(So)r’-tH(^fln-rooi
WD OO-h-h-h-hCN-hO-h-h-hOOOOOOOOO-h-hOO-h-h-h-hOO
So Jc3o O01 OC3 =Soh .3G u ooooooooooooooooooooooooooooooo
^ s at bo-^i ?
P-1 S3 ■ -1 x oHotM^'d--H-HO'(-NH’<HNM'oM^oiino<>i)'-0H^i|^noHcoti>onTo)<-iNn<OOOTcto'-oHo{inNoosoi>o-ocioo(iNo(ON>-Hc<aii<MN^oooiro'
d -s U G p p —; <N —; p —: p —; p —; <N p p p p p p p —; © p p p p —; <N —; p p p
<Pr Q< >So-» .„ -vr 3 r5 ooooooooooooooooooooooooooooooo
H ci op Si§ B*
r^<NC0'd-0><N0S<N<N0>c<l'>00>l^'c}-<Nr0i<N^<0i00l^'O00<N^'-Hr0ir'<N'-H
o a 2^ i.3 0p0 O— ^t—-' Crjtl —l~ ipt^ Orj- H—0 Np 0—0 (^p< N—l pn 'OpI ^pI ^p0 0p'O —O O<N0 —0i np'c l—-C —^ 0p0 <pN -pH p< Np< Np I^pC p0
_ ai & $ ooooooooooooooooooooooooooooooo
§ a
•~2 a3_ g .y CgO ■^■oo^HOtNOO\fNt~-vo^H<NtN-HrNvooioo\'Oc<or~r~vo-Hioir-m-Hr~-vo
^ 2 2 a ^ lOiOMrOO^O^'OOllOiO'O'OVO'O'lOOiOl^VOOlOOCTliOVOOOCNrOO'O
C3 Cu OO-h<NOO-hc0^hO-h-hOOOOOOOOOOOOO-hO-h-h<N-h
ooooooooooooooooooooooooooooooo
I
0&c6J1sD 4O031 ._0OT1' ~OW01- )og -OLHOl "O-Hd- oOOoO o-'^Hol- OhOmlOtofOOO-O-HhOOOOl-OOhHMOtlOhChCOhNOO'OOO'tOOOhinOrMOOObO^lCCOOOOlnOfOOrOOhh'HOOil-oihOir-OKh'ONOMMO
Oh ooooooooooooooooooooooooooooooo
u
-c <2 ° SaP
inooo><NOO’-H'Orocon^Tt'cl-in'0'0'c}-
©(N<OI^^tO\CT\fOi-H'>OOOHtOlO
!—H « ot>a ^0t0-io-)ho<mNooo-oh<0N-rh---Hhio©)<-Nh<0N->hr0)'c0i-0in0'c0i- O0-lhO0f©Ob-hC<ON-H©M©'O-hh^tHNO'^-rth'O©i0n
S|| 3 u ooooooooooooooooooooooooooooooo
h-I1 ^ Co h3J ohoOvio^^h^-MinfONN<NCOO\omoi-nH«Nrt^hHlOin'mO'>OcifOf'h\cl-O-o^i^O-iin'Ovol~OHtOSOCVOOiO©'b©h©r^ri-h'^'-O'^-b'©b
© © P © © P CO <S P P P © © © © © tN P © © P © © © © fSj P P © ©
S§| a < ©©©©©©©©ooooooooooooooooooooooo
a
2bO PI <Oro-HrtHtio(NHtroooo'Oio'0^(N(S-H(Oioh(NOOoohroo<OMrt
fOWHHCfiooo^HtcsoocoooM^iomooiocoinMmHtfSMHtf'OO'cf^io
►3^ p — p p — p — p p © — — p p p p p p p ri p — p p © p p p p © ©
O ©©©©©©©©©©©©©©©©<©©©©©©©©©©©©©©©
c3 oj ”7o3
O) n a.
00\OOOlH(fiOlOO<0'tOlOO'CHhoO<OHHfrilo^O\^C'0\<OHtiCOOOO
OOmt'^COrtiOfNrflh'OhOlOlHism- OOnOOl-HMhrOOOOirOiiOOl
f«°A -H-Hrt(Srt(SMf(l(Nrt-H-H-H-H<N(N(N<S<S<S|0(N(N<S-H|(1(S^(NrtO
< ©©©©©©©<©©©©©©©©©©©©©©©©©©©©©©©©
^ "I I
O
.2 o J2 «,
aao .“>3 ^U <> S3 V p 2 H G J
4H3S<3L-H> <r-O0S51h & 00 O U Upc fOUofhi ^OoShh OPothnh Ph P—h PJ3Q 3^PQ U^0§1 Hi& oGPOh ^Og hdS>i§ Go rO—h 3 © ^e£c <au <su O 5
10 AMERICAN MUSEUM NOYITATES NO. 3617
u
_sci 060 091 084 132 o-rH- v©<No mOvo -Omh <Ovon -m-Hh OOOO\-Sh-'OOO(NOmOt^O00O0\-0h0OO-\hO-\hO-\hOO(NO'nO't OO-h-hOOOO
m 0.0.0.0.ooooooooooooooooooooooooooooooooo
S

g_myst 0.0610.0900.0690.198oo oo oo oo oo5 oo omO^(mo-Shhmo-Oh©o-\OHho-oihcooOmmoO^oOoo-ph\ohoOo(oSO<ONo-h\mOoO©\oO^oOo-ohi\^hoO©(N-oth''O<o-h©\oOO<\NoOO<Oo\N(S-o©h0Oo©C- oh3_ oOO omt— oOt" oOVO
S

Pi_pith 0.119 0.139 0.074 0.1280-oH'>ot<^NmoOmOo-OHoO-'oH>O^oti-tH'o-©©om<N©o<i-oHN©©o©©o<©©o<TNto<©No<-HNo'C-oHTl^\o-OHoO-<HoNCOoNOmo<N-oH'3©-o©0'o-0H©oO-OHo©0o^0©©o<<No©©to©^t<^o©^

a
e_ros 0.0900.1530.0680.162mpom -©oH>iop^- mHopm© op<m ©o©m <o©a© 'io©<p ©’op-Ht '~o©<- <No©© ©opom oo©m© oC^oNi o-o©©© o©©m oop©o ao<pNN omp© -oCHiN^ oO'p;3O -o©©'3 o©©-o <oo©©t ^o©ppm o©©o© oo©m© omm© o©CiN~ o<'pmN oo©t^o o©>©-o oH©o
L

o
g 6953rtrt(SOOrt(s|inOO\^'OHiOO\Hin»OrtHON(N^^«rt(SliOOOO(N^OO^
Laja 0.12 0.08 0.08 0.34<po N<oNC o—N■_ -o©H ©o© l'o—'-■_ Ho©0 Oo© 0o—O ©o© <o©© o©© ©o©© o©m oCpN to^mo0_ Oop< o(NN omp ©o©< o©© o0©O o©0O op© o©^m o©© opm o0©O opT of©r< opN omCN

a
_sat 278094105110««p \MOp OHp \^p^\ ^Op) l^pO0 O0p O0p0 Op(O» p'Hr tpiOn H'pC ('pnS h^pO' t^p\0,0 p0\l 0pO\ N<iNf 'lOC^N tOpNH \pifV lp0) i\pn0 «0pM ^p'VO pOh p«W 0p'\t hipn( O»©H \pOr Ntp0 '00p
Ch 0. 0. 0. 0.ooooooooooooooooooooooooooooooooo

Ce_apel 0.085 0.113 0.047 0.111m-oinH©ooOoio^OHmoO7mtoOm©Ooi©nOoimnOo'moOoo<-^NoH(SiOo'-M~Oo^TO'o©fOomooO©ion-THootoO(oo^-m\oHH'©oOTmf-omHooOoom\oOo'\©oO^m'Oo©(O(oN^OmoOOio(^f-oioHoo oo^OOO oVOO oOcn om-©H

g
y 6710cmN'3'©-©m©'©mmi^'©mti'~'C-'N©^©tm'-H^iC--NH©voC©Ntm~-©mrp-m’-H©<<©©mi^m<©m©mto'-oiooio^m'3'-©mic^N'©ocoN(©'-H<©m-oHo
p 3268
_ 0001
Cx 0. 0. 0. 0.ooooooooooooooooooooooooooooooooo
c
ac 81558625^^tor^oooo'OOoor^i^roir^<NO<Ni^oo'Oi^oo'ooot^i^ONi^o-Hr<ioooo
Cxj 0.0 0.0 0.0 0.1-oho^hoOoOoOoOo-hoOoOoOoOOoOoOo-oh-oh-oho-ho-ho-hoOoOoOoOoOoOOoOo-oh-ohoOOo-oh
Cl_goel 0.064 0.127 0.093 0.161m-c©nH <©pO© O©C©N© ©cmCn N©©cnC o©Npo© O-pHO© OcOn©' O>i-<h©© ©p^©| p-—C©© N—-©OH Op©m -r©H©^ |’^©—-aH '©\p'i©~ >pa'iN©o cm©n© ^>hj©-- mHr©j- CHo©NO Oo>Oo ©tp©^i ©O'—>C >o©Np© ©©—©©m ©©o( ^Op©<O ©©pOm m©ptm '^©©mp ©©mp ii©'Op^

olo 3007m(^mvomvoooo<o©pfoNoa©'iim^(v^oTmtmr~mt©^pmoosoO'>\voom'©t'''f©sm©'p-Ho\'3©-mm(m^ic^^'3p-pcNtNi'-^H'©ro~or©~o©\mm
Ci_m 0.05 0.07 0.09 0.28-©h>-©h©©>©©>©©>©©©>©©©©>©<©N>©-h>©-h>-©h>©©>©©->h©->h©<>N©><©N>-©h>-©h>-©h>©©©>©->h©-h>©-h>-©h>-©h>-©h>©©-©h>C©<>©-1>

a
el 9591Oh,1'COiHinr'i(S|n®^(fi^r'Oorfio\hoo^)rtVOhcn(Nf,i'-i|fl(S mm©
m 8456-H©-H©-H©©©©©©©©©©-H©-H©-H-H©©©©©©-H-H CN © ©
a_ 0.00.00.00.0©©©©©OO©©©©©©©©©©©©©©©©©©©©©©©©© ©© ©-H
C

v 0\Mm(NO«OOh'OOt'Of'mO\ ^t-^tr^mm©-HOompoooo©©©t^Tt
Ca_cal 0.094 0.062 0.094 0.169©^©hp©^oh©oOt©N-hm©©p©©m©©pv©©o©©oo©©oo©©^-©Hh©^©^-H©h©t©^ ©©CN ©i1—r1 i©m©vo©pv©o©i©ri ©p©v ©©o i©r©i t©©^> ©©^ ©ppO©Ot©~ ©-©t~ ©©- ©
At_geof 0.133 0.089 0.080 0.104a©—© o>pm<i© ^CsN^© htCmN© oCi©Nor ^p\©©o ©pom ©p\©' c©—mmI ©pm>' t©p©m -p©H^ ©p^'t <«>pN©i fp'-©l Hop© ©oC©N«© p©iin ^>p©i ©»N©(© (smS© |mW0© (©W>s r^©C-No< ©p\N' ©Opm© hrp^©' H-CN©H rCr©Nt' o<p©oN ©po'- h©Hpm ©©p

voci 62687549©©<©Nr^mm©inmoro^omom-Hr-mp--HH^i^©©<rN^©mr©-fp^ot^o©oo-H©rm^m©Tt^t©-Hr-'-oHo©m©<N<N©o©o©©mmm©©©
o_ 0.00.00.00.1©©©©©©©©©©©©©©©pp©h ©p ©© ©p< ©© ©© ©p ©© ©ph ©p ©ph ©© ©© ©© ©© ©© ©© ©© ©p ©© ©© ©© ©© ©© ©p
A

a
Al_car 0.105 0.105 0.015 0.184-mt©NH <cMmn ©rpH^© ©©O© mp\©m^ ©o©© om-©Hr tmm©m om©^o p©©m o©rmio ©-c«Hn o©—©\ <<s©oN© —o©© \<©(©SN c©On< Np©( oNC©oN (i©CrNN^ <©CONN ©pMm ©p'-mH ©cnrO< tprO©t ©(p©N ©p©o< mm©N P©—o© '©O©m ©rpTt t©M©p ©m^cn

W -H <N -H <N -H <N <N
Species Pal LMl Pal MineUI1W ^JSh4 S<hhN4J hpJ,