ArticlePDF Available

Abstract and Figures

Tissue vibrations in the larynx produce most sounds that comprise vocal communication in mammals. Larynx morphology is thus predicted to be a key target for selection, particularly in species with highly developed vocal communication systems. Here, we present a novel database of digitally modeled scanned larynges from 55 different mammalian species, representing a wide range of body sizes in the primate and carnivoran orders. Using phylogenetic comparative methods, we demonstrate that the primate larynx has evolved more rapidly than the carnivoran larynx, resulting in a pattern of larger size and increased deviation from expected allometry with body size. These results imply fundamental differences between primates and carnivorans in the balance of selective forces that constrain larynx size and highlight an evolutionary flexibility in primates that may help explain why we have developed complex and diverse uses of the vocal organ for communication.
Content may be subject to copyright.
Rapid evolution of the primate larynx?
Daniel L. BowlingID
*, Jacob C. DunnID
, Jeroen B. Smaers
, Maxime GarciaID
Asha Sato
, Georg HantkeID
, Stephan Handschuh
, Sabine Dengg
, Max KerneyID
Andrew C. Kitchener
, Michaela GumpenbergerID
, W. Tecumseh FitchID
1Department of Psychiatry and Behavioral Sciences, Stanford University, Stanford, California, United States
of America, 2Department of Behavioral & Cognitive Biology, University of Vienna, Vienna, Austria,
3Behavioural Ecology Research Group, Anglia Ruskin University, Cambridge, United Kingdom, 4Biological
Anthropology, Department of Archaeology, University of Cambridge, Cambridge, United Kingdom,
5Department of Anthropology, Stony Brook University, Stony Brook, New York, United States of America,
6Division of Anthropology, American Museum of Natural History, New York City, New York, United States of
America, 7Animal Behaviour, Department of Evolutionary Biology and Environmental Science, University of
Zurich, Zurich, Switzerland, 8Center for Language Evolution, University of Edinburgh, Edinburgh, United
Kingdom, 9Department of Natural Sciences, National Museums Scotland, Edinburgh, United Kingdom,
10 VetCore Facility for Research, University of Veterinary Medicine Vienna, Vienna, Austria, 11 Klinische
Abteilung fu¨r Bildgebende Diagnostik, University of Veterinary Medicine Vienna, Vienna, Austria
These authors contributed equally to this work.
* (DLB); (WTF)
Tissue vibrations in the larynx produce most sounds that comprise vocal communication in
mammals. Larynx morphology is thus predicted to be a key target for selection, particularly
in species with highly developed vocal communication systems. Here, we present a novel
database of digitally modeled scanned larynges from 55 different mammalian species, rep-
resenting a wide range of body sizes in the primate and carnivoran orders. Using phyloge-
netic comparative methods, we demonstrate that the primate larynx has evolved more
rapidly than the carnivoran larynx, resulting in a pattern of larger size and increased devia-
tion from expected allometry with body size. These results imply fundamental differences
between primates and carnivorans in the balance of selective forces that constrain larynx
size and highlight an evolutionary flexibility in primates that may help explain why we have
developed complex and diverse uses of the vocal organ for communication.
Recent years have witnessed a renaissance in research into the evolutionary and mechanistic
basis of mammalian vocal communication. Two factors underlie this progress. The first is an
extension of fundamental theoretical principles, initially developed for human speech, to a
much broader range of vertebrate taxa, rendering it clear that source-filter theory and myoelas-
tic-aerodynamic theory apply to a wide range of terrestrial species [14]. The second is the
comparative application of these theories to large quantities of interspecific data, using phylo-
genetically controlled methods designed to address fundamental questions about trait evolu-
tion, including, e.g., the roles of sexual selection, “honest” indicators of caller characteristics,
ecological factors, and morphological and neural specializations for vocal signals [510].
PLOS Biology | August 11, 2020 1 / 21
Citation: Bowling DL, Dunn JC, Smaers JB, Garcia
M, Sato A, Hantke G, et al. (2020) Rapid evolution
of the primate larynx? PLoS Biol 18(8): e3000764.
Academic Editor: Simon W. Townsend, University
Received: February 14, 2020
Accepted: July 8, 2020
Published: August 11, 2020
Peer Review History: PLOS recognizes the
benefits of transparency in the peer review
process; therefore, we enable the publication of
all of the content of peer review and author
responses alongside final, published articles. The
editorial history of this article is available here:
Copyright: ©2020 Bowling et al. This is an open
access article distributed under the terms of the
Creative Commons Attribution License, which
permits unrestricted use, distribution, and
reproduction in any medium, provided the original
author and source are credited.
Data Availability Statement: All of the data used in
this paper are recorded in the supporting
information file S1 Data.xlsx.
Funding: This work was funded in part by Austrian
Science Fund (FWF) DK Grant “Cognition &
Communication” (#W1262-B29). DLB was
Much of this recent progress has focused on primates, in which acoustic investigations have
been intense and a broad range of species has been studied [5,6,9,11]. At the same time, analy-
ses of similar scope, but focused on brain size, have taught us that principles that apply gener-
ally within a clade like primates do not necessarily apply to other clades like carnivorans or
ungulates [12]. Carnivorans in particular are of interest for comparative analyses focused on
vocal communication because they are comparable to primates in terms of the range of habi-
tats they occupy (e.g., terrestrial to arboreal), the social systems they exhibit (e.g., solitary to
gregarious), and the body sizes they display (10
x to 10
x kg) [13,14]. Vocalizations in some
carnivoran species have been studied in detail [15,16], and some comparative work has been
done [17,18], but large-scale interspecific analyses applied across the carnivoran order have
not been initiated.
Here, we compare a key mechanistic determinant of vocalization across primates and carni-
vorans. Our sample includes 55 species, ranging in size over three orders of magnitude in both
clades: from pygmy marmoset (Cebuella pygmaea, mean body mass approximately 110 g) to
Western gorilla (Gorilla, approximately 120 kg) in primates (n= 26); and from dwarf mon-
goose (Helogale parvula, approximately 280 g) to tiger (Panthera tigris, approximately 180 kg)
in carnivorans (n= 29). We focus on the larynx (i.e., the main organ of vocal production),
combining three-dimensional computer models built from X-ray computed tomography (CT)
scans with detailed digital measurements, using a protocol designed to characterize gross fea-
tures of laryngeal morphology consistently across species. Phylogenetically controlled compar-
isons of these data with specimen-specific body lengths reveal marked differences in the
evolutionary trajectories of overall larynx size in the different clades. Parallel comparisons with
acoustic vocalization data provide initial evidence that these differences are relevant to vocal
Each larynx was characterized by a set of 10 measurements (Table 1 and Fig 1). Across species,
the most variable measurements were associated with the ventral extent of the larynx, followed
by our proxy for vocal fold length, and the distance between dorsal cricoid and ventral thyroid
cartilages in the midsagittal plane. By contrast, the least-variable measurements were associ-
ated with the width of the thyroid cartilage in the coronal plane, followed by the diameter of
Table 1. Laryngeal measurements sorted from most to least variable by CVs calculated on raw values.
Measurement name Landmark numbers used CV Standardized loading on PC1 “larynx size” Correlation with “larynx size”
Ventral cricoid height 1 to 2 0.858 0.885 0.867
Larynx height 1 to 14 0.837 0.980 0.979
Ventral thyroid height 7 to 8 0.825 0.855 0.834
Vocal fold length [(9 to 12) + (9 to 13)]/2 0.717 0.960 0.944
Crico-thyroid distance 4 to 9 0.669 0.990 0.964
Dorsal cricoid height 3 to 4 0.648 0.940 0.887
Apical cricoid depth 2 to 4 0.587 0.990 0.973
Basal cricoid depth 1 to 3 0.585 0.975 0.943
Basal cricoid width 5 to 6 0.571 0.975 0.903
Dorsal thyroid width 10 to 11 0.559 0.975 0.920
Correlations are Pearson rvalues calculated between larynx size and the log-transformed measurements (all significant at P<0.0001). Landmark numbers correspond
to those in Fig 1.
Abbreviations: CV, coefficient of variation; PC1, first principal component
Rapid evolution of the primate larynx?
PLOS Biology | August 11, 2020 2 / 21
supported by a Lise Meitner grant from the FWF
(#M1773-B24). JCD was supported by a grant
from the Royal Society (RSG\R1\180340) and the
Rhinology and Laryngology Research Fund.
National Museums Scotland thanks the Negaunee
Foundation for their generous support of a
preparator who dissected the larynges use in the
study. The funders had no role in study design,
data collection and analysis, decision to publish, or
preparation of the manuscript.
Competing interests: The authors have declared
that no competing interests exist.
Abbreviations: CT, computed tomography; CV,
coefficient of variation; F0, fundamental frequency;
OU, Ornstein-Uhlenbeck; pANCOVA, phylogenetic
ANCOVA; PC1, first principal component; pGLS,
phylogenetic generalized least squares; RA, relative
age at death; SSD, sexual size dimorphism; SS,
specimen sex.
the cricoid cartilage in the horizontal plane. For allometric comparisons, the dimensionality of
the laryngeal measurement data was reduced by applying principal component analysis. The
first principal component accounted for 91% of the variation in measurements across species
and received high loadings from all 10 measurements (mean = 0.95, SD = 0.05). We refer to
this first principal component as “larynx size” hereafter.
Fig 2 shows larynx size plotted against log10 body length for all 55 species. Allometric scal-
ing was clear in both orders. Comparison of the primate and carnivoran regression lines, using
phylogenetic ANCOVA (pANCOVA [19,20]), revealed similar slopes (β
= 4.71 versus β
= 4.73, F
= 2.162, P= 0.148) but significantly different intercepts (α
=8.04 versus α
=9.09, F
= 8.177, P<0.001). This indicates an evolutionary grade shift between primates
and carnivorans, with primates exhibiting larger larynges for the same body length across our
study sample. We estimated the magnitude of this difference by calculating the mean differ-
ence between our laryngeal measurements for eight primate-carnivoran species pairs with sim-
ilar body length (specimens differed by <1 cm on average; see Fig 2). The results suggest that,
for similar body lengths, the primate larynx is approximately 1.38×larger than the carnivoran
larynx on average (SD = 0.38, range = 0.92–1.98). Importantly, we also found that residual
Fig 1. Laryngeal landmarks. Three-dimensional digital larynx models from our data set showing a primate larynx (top row) and a carnivoran larynx
(bottom row), each from three perspectives. The cricoid cartilage is shown in yellow, the thyroid cartilage in blue, the arytenoid cartilages in purple, and the
hyoid bone in red (gray epiglottal and tracheal cartilages were not assessed). Midsagittal perspectives (right panels) show the left halves of each larynx, with
the ventral aspects pointing right. The 14 anatomical landmarks placed on each larynx model (see Methods) are depicted as black circles. Note that
landmark 12 is not depicted here, but located on the ventral tip of the right arytenoid vocal process in a location corresponding to landmark 13, but on the
right side. Table 1 describes the 10 measurements derived from these landmarks.
Rapid evolution of the primate larynx?
PLOS Biology | August 11, 2020 3 / 21
variation of the primate allometry was significantly higher than residual variation of the carni-
voran allometry (i.e., carnivorans exhibited stronger allometric integration; factor = 2.17x, P =
0.03 obtained using a permutation analysis of primate-to-carnivoran rate ratios; see Methods),
suggesting that the relationship of larynx size to body length is more flexible in primates.
To ensure that these results are not confounded by systematic differences between primates
and carnivorans in the sex and age of our individual larynx specimens, or differences in the
prevalence of species’ sexual dimorphism between clades, we conducted additional phyloge-
netic generalized least squares (pGLS) regressions predicting larynx size as a function of body
Fig 2. Body length versus larynx size. Base 10 logarithm of body length plotted against the PC1 of the larynx
measurements for our sample of 26 primate (blue) and 29 carnivoran (red) specimens. Squares depict atelids and
diamonds depict papionines, clades that stand out for having exceptionally large or small larynges among primates,
respectively (see text). Specimens belonging to one of the eight pairs used to estimate the magnitude of the grade shift
between primates and carnivorans are outlined in black. Regression lines display a significant grade shift between
primates and carnivorans, as quantified by pANCOVA (see text). See Fig 3 for full species names corresponding to the
abbreviated names shown here. The data used to create this figure are located in S1 Data, sheet B, columns B and C.
pANCOVA, phylogenetic ANCOVA; PC1, first principal component.
Rapid evolution of the primate larynx?
PLOS Biology | August 11, 2020 4 / 21
length and the following covariates, each computed across as many of our specimens as possi-
ble: specimen sex (SS; M or F, N= 52), relative age at death (RA; specimen age at death/species’
maximum life span, N= 50), and species’ sexual size dimorphism (SSD; log10 average male
mass/log10 average female mass; N= 55). Each model included one covariate as well as its
interaction with body length. No significant effects of the covariates or their interactions with
body length were observed (SS: t
= 1.597, P= 0.1188; SS
body length: t
P= 0.277; RA: t
= 0.078, P= 0.938; RA
body length: t
=0.108, P= 0.914; SSD: t
0.069, P= 0.945; SSD
body length: t
= 0.039, P= 0.969), indicating that the differences
between primates and carnivorans shown in Fig 2 are robust to differences in the age, sex, and
sexual dimorphism of the specimens and species we sampled. We also conducted a pGLS
regression predicting larynx size as a function of body length with SS
SSD as a covariate, to
assess a possible effect of SS dependent on species’ sexual dimorphism. The effect of the
SSD covariate was not significant (t
= 1.460, P= 0.153).
To further explore allometric patterns that may help explain the relative flexibility of the
primate larynx to body-size relationship, we used multiregime Ornstein-Uhlenbeck (OU) evo-
lutionary modeling to estimate whether additional grade shifts have occurred. These analyses
reveal two additional grade shifts in primates (in Atelidae and Papionini) but no additional
grade shifts in carnivorans. These results demonstrate that the allometric pattern between lar-
ynx size and body length that best fits our data includes a total of three grade shifts, with the
additional shifts occurring in primates. Fig 3A displays the three grade shifts, their estimated
phylogenetic locations, and the residual larynx size associated with each specimen. The signal-
to-noise ratio (pηϕ) associated with this model was 7.29, demonstrating high effect size and
thus high statistical power. The significance of the model fit to the data was confirmed by
= 8.1769, P<0.001). The first grade shift was toward larger residual larynx
size in primates, estimated to have occurred at the root of their divergence from carnivorans.
The second grade shift was toward even larger larynges in atelid primates (represented here by
Alouatta sara,Alouatta caraya, and Ateles fusciceps), estimated at the root of their divergence
from cebids. The third grade shift was toward smaller larynges in papionine primates (repre-
sented here by Macaca sylvanus,Macaca fuscata,Macaca silenus,Papio hamadryas,Mandrillus
sphinx, and Mandrillus leucophaeus) at the root of their divergence from cercopithecines. Fig
3B illustrates the difference in larynx size, indicated by the first grade shift in situ between a
similarly sized primate and carnivoran specimen.
Intuitive insight into the evolutionary diversification of the larynx in our sample can be
gained by working backwards from the residual larynx sizes observed at the branch tips of our
phylogenic tree to estimates of residual larynx size for its internal nodes. Accordingly, the
ancestral phenogram in Fig 4A shows the evolutionary diversification of residual larynx size
derived using a multiple-variance Brownian motion model that allows the rate of evolution
along different lineages in a phylogeny to vary [21,22]. This depiction of evolutionary trait
space shows that in order to achieve the variation observed in residual larynx size, the larynges
of multiple primate lineages underwent rapid diversification. Fig 4B visualizes this difference
in evolutionary rates using a standard Brownian motion Markov chain Monte Carlo proce-
dure to estimate distributions of evolutionary rate that match our residual larynx size results
and the associated phylogeny [23,24]. The resulting primate and carnivoran rate distributions
demonstrate a significant difference in the amount of trait variance accumulated per unit time
(rate ratio = 2.17x, P= 0.03).
To test whether these results are confounded by using body length as our proxy of body size
—e.g., perhaps carnivorans are simply longer-bodied than primates—we repeated the analyses
described above using body mass, rather than body length, as our proxy for body size. The
results were very similar: primate larynges remained larger and more variable than carnivoran
Rapid evolution of the primate larynx?
PLOS Biology | August 11, 2020 5 / 21
larynges, covariate effects and their interactions with body mass were not significant, and the
average rate of primate laryngeal evolution was significantly greater than the rate of carnivoran
laryngeal evolution. The only substantive difference was that in the mass analyses, only one
grade shift beyond that between primates and carnivorans was observed, toward larger laryn-
ges in Alouatta at the root of their divergence from Ateles. Together, the mass analyses indicate
that our findings are not an artifact of using body length as a proxy for body size. See S1 Text
for full reporting of the body mass results.
Fig 3. Ornstein-Uhlenbeck model results. (A) Phylogenetic tree and residuals from a pGLS regression of larynx size to log body-length. Carnivorans (red) exhibited
smaller larynges than expected based on body size, whereas primates (blue) exhibited larger larynges. Among primates, atelids exhibited exceptionally large larynges
(upper set of dashed lines), and papionines exhibited exceptionally small larynges (lower dashed lines). Arrows indicate where grade shifts in mean larynx size are
estimated to have arisen; percentages indicate support for these estimations from a bootstrap analysis (see Methods). (B) Computer larynx models derived from CT
scans depicted in situ for two species with comparable body lengths (71.4 cm for the red fox and 68.5 cm for the siamang), showing the larger relative size of the
primate larynx. The data used to create this figure are located in S1 Data, sheet B, columns B and C. CT, computed tomography; pGLS, phylogenetic generalized least
Rapid evolution of the primate larynx?
PLOS Biology | August 11, 2020 6 / 21
Lastly, we compared larynx size with an acoustic feature measured directly from species-
typical vocalizations across our sample to assess the potential link between our morphological
results and vocal communication (Fig 4C). Fundamental frequency (F0) was selected as the
most relevant acoustic feature here because of its important role in vocal communication and
its close association with vocal fold length [1,25], which is strongly correlated with larynx size
Fig 4. Evolutionary rate analyses. (A) Ancestral phenogram depicting divergence times plotted against residual larynx size from a pGLS regression of larynx size to log
body length. Primate lineages (blue) are characterized by larger and more variablevalues than those of carnivoran lineages (red), aswell as faster rates of change.(B)
Probability density plots indicating the distributions of evolutionary rates required to produce the variance in residual larynx size observed in primates (blue) and
carnivorans (red). The Pvalue is the result of a permutation analysis of primate-to-carnivoran rate ratios (see Methods) and affirms the hypothesis that the two
distributions are significantly different. (C) A comparison of larynx size and log mean-F0 in species-typical vocalizations (data from [9]). See S1 Fig for a labeled version of
C. The data used to create Fig 4A and 4B are located in S1 Data, sheet B, columns B and C. The data used to create Fig 4C are located in S1 Data, sheet B, columns C and
K. F0, fundamental frequency; PC1, first principal component; pGLS, phylogenetic generalized least squares.
Rapid evolution of the primate larynx?
PLOS Biology | August 11, 2020 7 / 21
in our sample (r= 0.94; Table 1). Mean F0 data were obtained for 53 of our 55 species from
[9]; Lemur catta and Otocyon megalotis were excluded because of insufficient data (see Meth-
ods). PGLS regressions between larynx size and mean F0 showed a significant negative rela-
tionship for primates (t
=7.135, P<0.001) and for carnivorans (t
P<0.001). The slope of the primate regression was marginally steeper than that of the carni-
voran regression (β
=0.476 versus β
=0.296, pANCOVA F
= 4.020, P= 0.05), sug-
gesting that a given change in larynx size is associated with a relatively larger change in mean
F0 among primates. These results provide preliminary evidence that the phylogenetic patterns
we observe in vocal morphology are related to the acoustics of vocal communication. The data
used in all analyses described above are included in S1 Data.
The results described above highlight a clear pattern: relative to carnivoran larynges, primate
larynges are significantly larger with respect to body size, more variable in size, and have
evolved faster. Furthermore, this pattern is related to acoustic variation in the mean F0 of spe-
cies-typical vocalizations, suggesting its relevance to vocal communication.
With respect to evolution, an important first point is that the phylogenetic variation in rela-
tive larynx size that we observe does not necessarily reflect variation in selective forces acting
directly on larynx size. Logically, our results are consistent with variation in selection on larynx
size, but also body size, or some combination of both. A parallel example comes from studies
of relative brain size, where interspecific differences are commonly interpreted as reflecting
selection for structural/functional enhancements of the brain, but large-scale comparisons
akin to those performed here suggest a primary role for selection on body size [26]. Here, how-
ever, the evidence supports evolutionary change in larynx size rather than body size as the pri-
mary driver of the differences we observe: independent analyses of evolutionary rate applied to
body length and “crico-thyroid distance”—a proxy for larynx size (see Table 1) measured in
the same units as body length (mm) and thus suitable for direct comparison—indicate that lar-
ynx size has changed significantly faster than body size over time (by 2.15×for primates, 1.67×
for carnivorans, and 1.79×overall; P<0.01 for all tests). This makes it unlikely that the differ-
ences in relative larynx size that we observe have been driven only by evolutionary changes in
body size. Notably, a reversal of this general pattern was found for papionine primates, such
that the rate of body-length evolution outpaced that of crico-thyroid distance by 1.19×(sample
size [n= 6] too small to appropriately test significance). Papionines also stand out here for hav-
ing some of the smallest larynges relative to body length that we observed among primates,
suggesting that understanding the constraints on laryngeal evolution in this clade may be par-
ticularly informative.
As demonstrated by the greater size variance and faster evolutionary rates observed in pri-
mates, larynx size is less tightly coupled to body size in this order than it is in carnivorans. This
pattern of weaker allometric integration suggests that the primate larynx has been relatively
more likely to respond to fluctuations in evolutionary pressure. That is, alterations in the bal-
ance of selective forces that maintain relative larynx size can be expected to have driven greater
diversity in primates than carnivorans, which, as indicated by our comparisons with F0, would
be expected to have had consequences for vocalization. In the following paragraphs we con-
sider the hypothetical selective forces that act on larynx size and discuss whether or not differ-
ences in these forces between primates and carnivorans can plausibly account for our results.
Major adaptive hypotheses can be assigned to each major function of the larynx: protecting
the airway during feeding, regulating the supply of air to the lungs, and vocal communication.
A further relevant distinction is whether the proposed evolutionary trajectories are based on
Rapid evolution of the primate larynx?
PLOS Biology | August 11, 2020 8 / 21
directional selective pressure or a relaxation of selective pressure(s). Either can increase pheno-
typic variability, and multiple pressures can interact to determine phylogenetic trends.
A first possibility is that differences in relative larynx size between primates and carnivorans
reflect the role of the larynx in protecting the respiratory system during feeding [27]. Although
the trachea is mainly occluded by the epiglottis (not examined here) during swallowing, the
larynx protects against aspiration of food or liquid via reflex closure and coughing when the
epiglottis is bypassed [28]. Accordingly, increased variation in primate larynx size may be par-
tially explained by a relaxation of selective forces related to diet and associated feeding behav-
iors relative to carnivorans. For example, it may be that the threat of choking on large pieces of
minimally-chewed meat leads to selection pressure for smaller larynges in carnivorans, and
that this pressure is somewhat relaxed in primates, which eat more plant material and spend
more time chewing as a group [29]. It should be noted, however, that even though relaxed
selection against choking provides a logical explanation for the overall differences in larynx
size we observe between primates and carnivorans, it is not clear how this could explain the
greater variance in larynx size observed among primates. Although carnivoran diets are cer-
tainly more based on animal matter than primate diets on average, there is considerable varia-
tion within both clades, from obligate herbivory among carnivorans (giant pandas) to obligate
carnivory among primates (tarsiers).
A second possibility is that differences in relative larynx size between primates and carni-
vorans reflect the role of the larynx in respiration, e.g., in regulating the amount of oxygen that
can enter the lungs and intrapulmonary pressure [30]. For example, differences in locomotor
behavior between primates and carnivorans may place different demands on oxygen metabo-
lism [31] and the capacity to stiffen the thorax by increasing intrapulmonary pressure [28].
Both factors are closely related to substrate use, which although variable in both orders, is
more often terrestrial for carnivorans and arboreal for primates [32,33]. Notable exceptions
among the carnivorans studied here include Potos flavus,Nasua nasua, and Arctictis binturong,
all of which are relatively arboreal [34]. If the larger larynges of primates observed here were
hypothesized to reflect selection (directional or relaxed) related to respiratory function, no
support is found among these comparatively arboreal carnivorans, which all possess the rela-
tively small larynges determined to be typical of carnivorans here. At the same time, some sup-
port is found among the papionine primates, which are relatively terrestrial compared to other
primates and fit the pattern of having relatively small larynges. Even so, it is not clear how a
relatively small/large larynx would benefit the muscular and locomotor requirements of a ter-
restrial/arboreal lifestyle, respectively. A related alternative hypothesis is that increased oxygen
supply supported by the larger primate larynx serves brain function.
A third possibility Is that differences in relative larynx size between primates and carnivor-
ans reflect the role of the larynx in vocal communication. Here, multiple hypotheses have been
proposed to explain interspecific differences, and we consider the relevance of our data to each
in turn. The acoustic adaption hypothesis suggests that vocalizations are optimized to maxi-
mize their transmission in different habitats [35,36]. The majority of primate species are found
in tropical forest environments, whereas carnivorans occupy a greater diversity of habitats.
Forest environments diminish the value of visual signals, while providing relatively constant
physical conditions for acoustic signal transmission [35]. Together with the fact that primates
are relatively microsmatic, this difference in habitat may increase selective pressure to produce
vocalizations with frequency content suited to propagation in a forest habitat, i.e., the lower
frequencies made possible by larger larynges [35]. This hypothesis appears compatible with the
larger relative size of the primate larynx, but not its increased size variation. If variation in lar-
ynx size is contingent upon variation in habitat, carnivorans would be predicted to exhibit
greater larynx size variation, not less, as observed here.
Rapid evolution of the primate larynx?
PLOS Biology | August 11, 2020 9 / 21
The sexual selection hypothesis proposes that larger larynges and the lower-frequency
vocalizations they afford have been selected for acoustic size exaggeration in deterring oppo-
nents during intrasexual competition and/or attracting mates during intersexual choice—as
appears to be the case in howler monkeys [7]. More broadly, sexual dimorphism in larynx size
(and voice F0) appears to be strongly influenced by mating system in anthropoid primates
(including humans), with polygyny favoring increased dimorphism [37]. This suggests that
sexual selection may be important to understanding the kind of systematic differences in lar-
ynx size we observe here. This would be particularly true if the average strength of sexual selec-
tion differed between our samples of primates and carnivorans and/or if these samples were
discrepant in terms of the sexes of specimens included (they were not; see Methods). However,
using SSD as a proxy for the strength of sexual selection, we failed to find a significant differ-
ence between our primates and carnivorans. Additionally, when added as a covariate, SSD did
not account for significant variation in the allometries between larynx and body size. A final
point here is that although there is no carnivoran analogue for the enlarged larynges of howler
monkeys (that we are aware of), the next most-supported OU model showed a shift toward rel-
atively larger larynges in the roaring cats (Panthera) compared to the purring cats. Intrigu-
ingly, pantherines also possess enlarged vocal pads and an extremely elastic stylohyoid
ligament [38,39], which may serve to further exaggerate body size in vocalizations (often per-
formed in the darkness of night). Thus, although sexual selection and acoustic size exaggera-
tion may play roles in accounting for some of the within-order variation we observed, it
appears unlikely to account for overall differences between primates and carnivorans.
Finally, the social complexity hypothesis proposes that the larger and more varied larynges
of primates reflect selection for enhanced vocal communication, as might be required to maxi-
mize the benefits and minimize the costs of social living. Primates and carnivorans exhibit fun-
damental differences in social behavior [40], with primates being more likely to form larger,
more stable groups characterized by strong bonds and more close-contact time. In our sample,
median group size was 13.8 for primates and 2.6 for carnivorans (Mann-Whitney U= 121,
P<0.0001; see Methods). Primate group size also tends to be more variable [41]. In our sam-
ple, median absolute deviation in group size was 51.9 for primates and 32.6 for carnivorans.
Our results may be consistent with this hypothesis insofar as the more “complex” social lives
of primates can be expected to result in pressure for larger larynges, which could potentially
provide a more robust framework matched to the increased mechanical forces expected from
more frequent use [42]. However, preliminary comparisons of larynx size and social group
size data failed to show any significant correlation between these variables in our sample, for
primates (Spearman r= 0.11, P= 0.6094), carnivorans (r= 0.02, P= 0.9374), or all species con-
sidered together (r= 0.15, P= 0.2762). As a stark example, consider that howler monkeys have
some of the largest larynges relative to body size of any known species and yet live in relatively
small social groups compared to many other primates [6]; conversely, papionines have rela-
tively small larynges but live in relatively large social groups. These data suggest that social
complexity, at least as far as it can be approximated by simple measures of social group size
(see [43] for discussion), is unlikely to explain the clade-wise differences in larynx size that we
A central obstacle in choosing among these hypotheses is that our data only allow us to
assess primates and carnivorans with respect to one another. As a result, we cannot be certain
whether the primate larynx is relatively large and variable or the carnivoran larynx relatively
small and constrained. One reason to expect the former is that only primates appear to include
additional grade shifts, as evidenced, e.g., by laryngeal hypertrophy in howler monkeys. But
settling this issue will ultimately require more context from additional comparisons, e.g., with
Artiodactyla, Rodentia, and/or Chiroptera, in which hammer-headed bats provide another
Rapid evolution of the primate larynx?
PLOS Biology | August 11, 2020 10 / 21
known example of laryngeal hypertrophy [44]. Only such comparisons can tell us how the var-
iability observed here ranks among other mammalian orders. In addition to taking several
years to source and analyze often rare specimens, such comparisons will be complicated by
considerable differences in life history and body size, largely circumvented here by selecting
clades with significant overlap in both domains.
Future studies will also benefit from the inclusion of multiple specimens per species, as
error from within species variation in larynx size was not controlled here [45]. However, with
regard to the magnitude of this error and its potential to bias our results, we note that our sam-
ples cover over three orders of magnitude in body length. This means that interspecific varia-
tion was much larger than the intraspecific variation, a fact that can be expected to mitigate the
influence of intraspecific error on our main results. Importantly, we also controlled for the age
and sex of our individual specimens, as well as sexual dimorphism at the species level, buffer-
ing against error variance arising from associated individual and intraspecific differences. A
final important limitation to consider here is our focus on cartilage and bone only. This was a
consequence of using CT with unfixed and unstained specimens, which remain viable for
functional phonation studies. Although the cartilaginous structure of the larynx clearly corre-
lates with its acoustic capabilities (see Fig 4C), it is ultimately the soft tissues of the larynx that
are responsible for vocal production. Accordingly, studying evolutionary patterns of soft tissue
variation will be important, particularly if this work continues to explore the vocal conse-
quences of evolved laryngeal morphology.
In conclusion, the work described here demonstrates clear differences in larynx size and
larynx-size variation between primates and carnivorans, reflecting fundamental differences in
underlying rates of evolutionary change. These differences are consistent with multiple selec-
tive hypotheses, including directional selection resulting from cladistic differences in ecology
and social structure that impact vocal communication, as well as relaxed selection on the feed-
ing and respiratory functions of the larynx in primates relative to carnivorans. Our results
open an exciting new avenue of study focused on laryngeal variation among further mamma-
lian clades, which will provide the context required to determine how particular the differences
we observe here are to the evolution of the primate larynx. If the relative flexibility of the pri-
mate larynx is robust to future analyses with more clades, it would indicate an increased capac-
ity to explore trait space in our lineage, which may in turn explain why primates have
developed such diverse and complex uses of the vocal organ.
Ethics statement
This research was conducted in accordance with European Union Directive 2010/63/EU.
Approval by a specific IACUC committee was not required. Cadavers were donated to
National Museums Scotland by the zoos of origin, and our use of their laryngeal specimens
here was approved by the museum on a case-by-case basis.
The specimens used in this study lived and died in zoos throughout northern Europe. Death
was followed by a postmortem examination performed by local veterinary staff at the zoo of
origin after death, and the cadavers (excluding the digestive system) were frozen at 20˚C. The
cadavers were then shipped on ice to National Museums Scotland for processing and preserva-
tion. Larynges, from the tongue to below the larynx, were excised during specimen preserva-
tion and refrozen for shipment on ice to Vienna, Austria. Upon arrival in Vienna, specimens
were unpacked, thawed, and cleaned with saline before being mounted and refrozen in
Rapid evolution of the primate larynx?
PLOS Biology | August 11, 2020 11 / 21
preparation for X-ray CT scanning. Any specimens judged to be in poor condition at this stage
(e.g., because of desiccation, decomposition, or insult) were excluded from further consider-
ation. Selected specimens were mounted on polystyrene foam plates with ventral aspects facing
upwards, and toothpicks were inserted into the foam on either side to prevent lateral rolling.
During mounting, care was taken to avoid distortion and to approximate in vivo posture to
the degree possible under the circumstances. The selection of which specific specimens/species
to include in our analysis was based on balancing six factors: (1) availability; (2) including a
wide range of body sizes in both orders; (3) maximizing phylogenetic variation; (4) maximiz-
ing tissue quality; (5) balancing sex across specimens; and (6) body length representativeness.
If multiple specimens were available for a species, the one closest to estimated species’ mean
was selected (see S2 Fig for a comparison of specimen body length and estimates of species-
typical ranges).
CT scanning
The majority of the selected larynx specimens (n= 49) were large enough to be scanned using
a clinical CT scanner (Siemens Emotion 16, Munich, DE). Plate-mounted specimens were
placed directly on the scanner gurney, rolled into the scanner bore, and scanned frozen.
Depending on specimen size, source voltage was either 110 or 130 kV and beam intensity was
either 80 or 90 mA (70 or 130 mAs). Each reconstructed slice (maximum of 1,215) measured
512 ×512 pixels. Depending on sample size, resolution of reconstructed volumes was between
238 and 369 μm
in the XY direction with slice thickness either 200 or 500 μm. Scan time was
approximately 1 minute per specimen. Owing to their size, the six smallest specimens
(Cebuella pygmaea,Saguinus oedipus,Leontopithecus chrysomelas,Saguinus bicolor,Helogale
parvula, and Suricata suricatta, see S2 Fig) were scanned with an Xradia microXCT-400 scan-
ner (Carl Zeiss X-ray Microscopy, Pleasanton, CA) equipped with the 0.4×lens. Because of the
longer scan times associated with this technique (approximately 3 hours per specimen), these
specimens were thawed, remounted upright inside sealed 50-ml Falcon tubes packed with
plastic drinking straws for support and filled to 5 ml with phosphate-buffered saline to prevent
dehydration (no contact with specimen). During this remounting process, care was taken to
avoid distortion and approximate in vivo posture to the degree possible. Immediately after
remounting, these specimens were placed in a tube rack that held them upright in the scanner
bore and scanned with source voltage = 40 keV and beam intensity = 200 μA. Projections were
recorded with 1-second exposure time (camera binning = 4) and an angular increment of
0.25˚. Reconstructed slices measured 512 ×512 pixels. Depending on sample size, isotropic
voxel resolution of reconstructed volumes varied between 29 and 72 μm
Larynx models and measurement
Computer models of each larynx were created from the CT data to facilitate detailed measure-
ments of the laryngeal cartilages in a nondestructive and consistent manner across species.
Models were constructed, using Amira 3D data visualization software (version 5.6.0; FEI,
Hillsboro, OR), and a detailed step-by-step protocol divided into four parts. The first part
described model construction: (1) importing the CT data into Amira; (2) gradually adjusting
the lower limit of the X-ray absorption window to visualize only smooth surfaces of the laryn-
geal cartilages and hyoid bone; and (3) using this lower limit to threshold Amira’s “isosurface”
function, which creates three-dimensional surface models of visualized data. The second part
described the establishment of a framework for landmark placement: (4) defining a midsagittal
plane by manually rotating and translating an Amira “ObliqueSlice” object to divide the cri-
coid cartilage in half along its dorsal-ventral axis; and (5) defining a midcoronal plane by
Rapid evolution of the primate larynx?
PLOS Biology | August 11, 2020 12 / 21
extracting the parameters of the midsagittal plane (using Amira’s “getPlane” function), enter-
ing them into a custom Matlab script (version 2016a; MathWorks, Nantik, MA) that generated
new Amira parameters for a perpendicular coronal plane, using these new parameters to
define a second ObliqueSlice (using Amira’s “setPlane” function), and manually translating
this second ObliqueSlice to divide the cricoid cartilage in half along its medial-lateral axis. The
third part described the placement of anatomically defined landmarks using Amira’s “Land-
mark editor” (see Fig 1): (6) at the intersection of the cricoid cartilage and the midsagittal
plane, the cricoid’s ventral basal extent (#1), ventral apical extent (#2), dorsal basal extent (#3),
and dorsal apical extent (#4); (7) at the intersection of the cricoid and the midcoronal plane,
the cricoid’s basal right and left lateral extents (#5 and 6); (8) at the intersection of the thyroid
cartilage and the midsagittal plane, the thyroid’s ventral basal extent (#7), ventral apical extent
(#8), and the midpoint between these (#9); (9) on the thyroid cartilage at the level of landmark
#4, the thyroid’s dorsal right and left lateral extents (#10 and 11); (10) on the arytenoid carti-
lages at the ventral extent of the right and left vocal processes (#12 and 13); and (11) at the
intersection of the hyoid bone with the midsagittal plane, the hyoid’s apical extent (#14). The
fourth part of the protocol described measurement: (12) exporting the coordinates of land-
marks #1 to #14 as a text file; and (13) using a second custom Matlab script to read this text file
and calculate the 10 Euclidean distances between specific pairs of points used to characterize
laryngeal morphology. These measurements are named and defined in Table 1. Importantly,
the anatomical locations of the 14 landmarks were not determined until after an initial inspec-
tion of all 55 of the digital larynx models included here (i.e., steps 1 to 5 were completed for all
specimens before steps 6 to 13 were initiated). This allowed us to select anatomical locations
where landmarks could be reliably placed across all specimens.
Specimen body length was defined as the distance between the ischium of the pelvis and the
tip of the snout in carnivorans, or the top of the skull in primates. Body lengths were obtained
from the same individual from which the larynx was excised for all but Nasalis larvatus and
Herpailurus yagouaroundi. Average body lengths from [46,47] were substituted for these speci-
mens. Body length was preferred over body mass, because the latter was only available for a
small subset of our specimens, owing to postmortem procedures at the zoos of origin, and
because we considered length to be less affected by zoo diet than mass (analyses based on aver-
age body mass are provided in S1 Text nonetheless). Laryngeal and body length measures were
log-transformed (base-10), because these variables change on a proportional scale, whereas
our linear models quantify variation on a linear scale.
Principal component analysis
After log-transformation, the laryngeal measurements (see S1 Data) were subjected to princi-
pal component analysis in R (version 3.2.3; The R Foundation, Vienna, Austria), using the
“psych” package function “principal” [48] to determine the factors underlying laryngeal varia-
tion across species. Initial analyses, configured to extract 10 components (without rotation),
revealed only a single component with an eigenvalue >1 for both orders. Following Kaiser’s
criterion [49], we thus reconfigured the principal component analyses to extract a single com-
ponent (rotation = varimax). The results indicated that this single component—which we refer
to as “larynx size”—accounted for the vast majority of variation in laryngeal measurements,
receiving high loadings from all 10 measurements (see Table 1).
Phylogenetic comparative methods
A measure of relative larynx size for use in the statistical analyses was derived by calculating
residuals from a pGLS regression of larynx size to body length across all species. A lambda
Rapid evolution of the primate larynx?
PLOS Biology | August 11, 2020 13 / 21
model showed that phylogenetic signal was high for this measure (λ= 0.956), which thus val-
idly represents larynx size after controlling for body size and phylogenetic nonindependence
due to shared ancestry. PGLS regressions were also used to model the effects of covariates on
the relationship between larynx size and body length. The trees used for all phylogenetic analy-
ses were obtained from [50] for primates and [51] for carnivorans.
To estimate the phylogenetic location of evolutionary grade shifts in mean relative larynx
size in our sample, we used multiregime OU evolutionary modeling. This approach estimates
where shifts in trait values occur along the branches of a phylogeny directly from the data. OU
models differ from more widely used Brownian-motion models in that OU models incorpo-
rate additional parameters to account for changes in mean trait value (θ) and the strength at
which the population moves from one mean to another (α). Several such OU methods have
been proposed using different parameter maximization algorithms. Here, we checked estima-
tions between the two most recent methods, “l1ou” and “phylogenetic expectation-maximiza-
tion” [52,53], and confirmed that both yielded the same results for the model presented in Fig
3. The effect size associated with this OU model was quantified using the signal-to-noise ratio
(pηϕ) and by running a bootstrap analysis. The signal-to-noise ratio demonstrates high power
when >1 [54]. Bootstrap analysis was run using the “l1ou” method [53] and further confirmed
strong support for the estimated grade shifts.
To test whether the estimated OU model provided significant results, we used pANCOVA
[19,20] as a least-squares approach. PANCOVA provides a confirmatory test, indicating
whether the clade-wise differences in mean relative larynx size, estimated by the OU model,
are statistically significant by comparing the intercepts of the associated larynx size to body
length allometries. We also used pANCOVA to test for differences in fit between alternative
OU models, extending it to ask, e.g., whether a 2-grade or a 3-grade OU model provided a bet-
ter fit to the data. The results of the OU model presented in Fig 3 were thus validated using
pANCOVA. Although OU modeling approaches provide a powerful way to estimate patterns
of evolutionary diversification, they necessarily entail more uncertainty than confirmatory
least-squares hypothesis-testing approaches. Whereas OU modeling approaches assume more
statistical parameters in order to estimate changes along individual branches of a phylogeny,
pANCOVA is focused on observed tips only and does not infer changes along individual
branches of a phylogeny, providing greater statistical power for confirmatory testing. It is
clear, however, that shifts in trait value among clades observed at tips are expected to be the
result of evolutionary changes that arose deeper in the phylogeny. The results of pANCOVA
using only observed tips are thus expected to align with the estimations based on OU modeling
To estimate trait values for the internal nodes of our phylogeny, we used an ancestral esti-
mation approach. Although not appropriate for hypothesis testing, such approaches provide
an intuitive visual depiction of the evolutionary diversification of a trait. Traditionally, a stan-
dard Brownian-motion model has been used for these purposes. Here, however, we use a mul-
tiple-variance Brownian-motion approach, because this better accommodates possible
changes in the rate of evolution that occur along different lineages in a phylogeny [21,22].
Considering that relative larynx size was found to entail both shifts in mean and rate among
clades, a multiple-variance Brownian-motion model is more appropriate here than a standard
Brownian-motion model in which rate is fixed. The results of these analyses are displayed in
Fig 4A.
To test for differences in rate of change in relative larynx size in primates versus carnivor-
ans, as well as rate of change in larynx size versus body size, we tested whether the ratio of the
Brownian-motion rate in primates relative to the rate in carnivorans is significantly higher
than unity, using a permutation analysis [24,55,56]. If this rate ratio is indeed higher than
Rapid evolution of the primate larynx?
PLOS Biology | August 11, 2020 14 / 21
unity, the rate in primates can be considered to be significantly higher than in carnivorans.
The Brownian-motion rate parameter is the most appropriate measure to perform this test,
because the Brownian motion model is a pure rate model (i.e., it quantifies rate, not direction).
The interpretation of rate of evolution (quantified as σ
in the standard Brownian-motion
model) corresponds to the amount of variance accumulated per unit time. Clades with a higher
rate accumulate more trait variance per unit time, resulting in more observable trait variance
and weaker allometric integration. To visualize the rate difference between primates and carni-
vorans in residual larynx size, we depict the posterior distributions of Markov chain Monte
Carlo estimates of the respective Brownian-motion rate parameters (results of the rate analysis
are displayed in Fig 4B).
Specimen age, sex, sexual dimorphism, and species social group size
The potentially confounding effects of specimen age at death, SS, and species sexual
dimorphism were addressed by modeling these factors as covariates in pGLS regressions
predicting larynx size as a function of body length (and mass; see S1 Text). A separate
regression was considered for each covariate and its interaction with body length. Descrip-
tions of how the data for each covariate were sourced follow below. All data are provided
in S1 Data.
Previous research in humans and nonhuman mammals has reported age effects on laryn-
geal geometry [57,58], indicating that the ages of the animals whose larynges were studied here
may be a determining factor in their morphology. This could confound our results if there
were a systematic age bias between our primates and carnivoran samples. To assess this possi-
bility, we obtained age at death data for 50 of our specimens from National Museums Scot-
land’s research records (missing: Vulpes lagopus,Lynx canadensis,Potos flavus,Puma concolor,
and Vulpes vulpes). To account for differences in longevity across species (e.g., a chimpanzee
might live 50 years in captivity, whereas a bush dog might only live 15), we normalized these
data by dividing each specimen’s age at death by maximum recorded longevity for its species
(as recorded in [59]) to calculate “relative age.” Maximum longevity was preferred over mean
for normalization, because the latter was often distorted by exceptionally early deaths in [59],
especially when sample size was small. A further way in which the age at death of our speci-
mens could have confounded our results is through interaction with sexual maturity. For
example, if the primate specimens studied here tended to be sexually mature at death and the
carnivorans not, this could explain their systematically larger larynges. To assess this possibil-
ity, estimates of age at sexual maturity were obtained for as many of our study species as possi-
ble from a published compilation of zoo records (N= 32) [60] and compared with our data for
age-at-death. For all but one of these 32 cases, the specimen we selected to represent its species
came from an animal that was sexually mature at death (exception: Vulpes corsac; age at
death = 6 months, estimated age at sexual maturity = 10 months). Although we could not
determine age at death and sexual maturity for the remaining 23 of our specimens, taking
these data as representative suggests that the majority are likely to have been sexually mature at
The sex of our specimens represents another potential confounding factor for our results.
For example, if males tend to have larger larynges than females, and more males were included
in our sample of primates than of carnivorans, then the differences we observed between these
clades could be due to SS rather than a true difference between primates and carnivorans.
However, sex data for our specimens from National Museums Scotland’s research records
indicate that our sample was closely matched, including 15 females, 10 males, and 1 unknown
(Chlorocebus pygerythrus) for primates and 16 females, 11 males, and 2 unknowns
Rapid evolution of the primate larynx?
PLOS Biology | August 11, 2020 15 / 21
(Herpailurus yagouaroundi and Tremarctos ornatus) for carnivorans. SS (male or female) was
nonetheless included as a covariate in a pGLS model predicting larynx size as a function of
body length as a further point of caution. Another way in which sex could have impacted our
results is through potential differences in sexual dimorphism between primates and carnivor-
ans—for example, if we happened to select male specimens from highly dimorphic species for
primates but females from highly dimorphic species for carnivorans. In this way, SS could
have impacted our results even though male/female ratios were approximately the same in our
sample of primates and carnivorans. To examine this possibility, we calculated an index of
SSD for each species (log10 average male mass/log10 average female mass). Comparison of the
SSD index between primates and carnivorans did not indicate a significant difference in SSD
strength between orders. Nevertheless, we also included SSD as a covariate in a pGLS model
predicting larynx size as a function of body length. Finally, we tested whether there was an
effect of SS that was dependent on SSD by including the interaction between SSD and SS as a
covariate in a pGLS model predicting larynx size as a function of body length.
Data describing social group size were obtained for all 55 of the species considered here.
Social-group-size data for primates were mostly obtained from a single source [46], which typ-
ically reported minima and maxima for each species, allowing us to calculate an average repre-
sented by the midpoint of these ranges. For Gorilla gorilla and Pan troglodytes, mean group
size was explicitly stated in [46] and used here instead of the midpoint. Social-group-size data
for Ateles fusciceps,Mandrillus sphinx, and Papio hamadryas were not included in [46] and
were instead obtained from [61] (A.fusciceps), [62] (M.sphinx), and [63] (P.hamadryas).
Social-group-size data for carnivorans were mostly obtained from a single source [47]. When
minima and maxima were offered, the average was calculated as the midpoint unless mean
group size was explicitly stated, in which case it was used instead (as for primates). Species
explicitly described as “solitary” were assigned a mean group size of 1; species for which the
“basic social unit” was described as a breeding pair were assigned a mean group size of 2. A
mother and her litter were not considered a social group for this analysis, although doing so
did not affect the significance of the calculated difference between our primates versus carni-
vorans. Data for Zalophus californianus were not included in [47] and were instead obtained
from Table 2 in [64].
Vocalization data
Vocalization data for all 55 of the species examined here were obtained from [9]. In that study,
a mean F0 value for each species was derived from a set of six vocalizations systematically
selected to cover the range of spectral variability present in a larger set of vocalizations. This
procedure ensured that the vocalization data chosen to represent each species were derived in
the same manner across species, promoting valid interspecific assessments. The resulting
“F06” value was only available for 80% (n= 44) of the species examined here. However, using
the same data set and reducing the number of vocalizations required to calculate mean values
from six to three, a comparable “F03” value was derived for 96% (n= 53) of our species. For
species with both F03 and F06 values, these variables were highly correlated (Spearman
r= 0.97 for primates [n= 20] and 0.98 for carnivorans [n= 24], Ps <0.0001), justifying the use
of F03 values here. For the two species without F03 values (Lemur catta and Otocyon megalo-
tis), the data in [9] included less than three vocalizations with measurable F0 values. These spe-
cies were thus excluded from analyses involving F0. It should also be noted that, although
several previous studies have preferred using acoustic measures other than mean F0 in their
comparisons with body size (e.g., minimum F0, maximum F0, or F0 range; [65,66]), attempt-
ing such comparisons across all 53 species included in the present acoustic analysis invariably
Rapid evolution of the primate larynx?
PLOS Biology | August 11, 2020 16 / 21
resulted in weaker relationships with larynx size than that observed for mean F0: the R
for linear models of larynx-size versus log10-transformed F0 mean, maximum, minimum, and
range were 0.583, 0.513, 0.506, and 0.361 respectively. Finally, we note that the acoustic allome-
tries reported using F0 mean in [9] are among the strongest in the literature (in the sense of
highest R
; cf. [6567]), providing further support for the systematic approach taken to deriv-
ing F0 mean in [9] and, by extension, here.
Supporting information
S1 Text. Full description of the body mass results.
S1 Data. (A) Laryngeal measurements. All data reported in millimeters. (B) Body length, lar-
ynx size, covariates, and fundamental frequency. Dashes indicate missing data. Maximum life
span data compiled from [59]. Average male and female mass data are compiled from
[46,47,6878]. Relative age = specimen age/maximum life span. Species sexual
dimorphism = log10(average male mass)/log10(average female mass). (C) Social group size
S1 Fig. A reproduction of Fig 4C with abbreviated species names labeling the individual
data points. See Fig 3 for full species names. The data used to create this figure are located in
S1 Data, sheet B, columns C and K.
S2 Fig. The body length and sex of each selected primate specimen (n= 26; top panel) and
each selected carnivoran specimen (n= 29; bottom panel). Each circle represents one speci-
men and is labeled with species name (blue = male, red = female, gray = sex unavailable). Hori-
zontal bars represent body length ranges and vertical bars represent body length means for
each species as reported in [46,47]. These ranges and means are given for males (blue) and
females (red) separately, when reported separately in [46,47] or in gray when reported
together. The four smallest primate specimens and two smallest carnivoran specimens were
scanned using micro-CT. The data used to create this figure are located in S1 Data, sheet B,
columns B, and G. CT, computed tomography.
The authors are grateful to Nadja Kavcik-Graumann for drawings in Fig 3B and to Thomas
O’Mahoney for comments on an earlier draft. National Museums Scotland thanks the Negau-
nee Foundation for its generous support of a curatorial preparator, who dissected out the
larynges used in this study.
Author Contributions
Conceptualization: Daniel L. Bowling, W. Tecumseh Fitch.
Data curation: Daniel L. Bowling, Asha Sato, Georg Hantke, Andrew C. Kitchener.
Formal analysis: Daniel L. Bowling, Jacob C. Dunn, Jeroen B. Smaers.
Funding acquisition: Andrew C. Kitchener, Michaela Gumpenberger, W. Tecumseh Fitch.
Rapid evolution of the primate larynx?
PLOS Biology | August 11, 2020 17 / 21
Investigation: Daniel L. Bowling, Jacob C. Dunn, Maxime Garcia, Asha Sato, Georg Hantke,
Stephan Handschuh, Sabine Dengg, Max Kerney, Andrew C. Kitchener, Michaela Gum-
penberger, W. Tecumseh Fitch.
Methodology: Daniel L. Bowling, Jacob C. Dunn, Jeroen B. Smaers, Maxime Garcia.
Project administration: Daniel L. Bowling.
Resources: Andrew C. Kitchener, Michaela Gumpenberger, W. Tecumseh Fitch.
Supervision: Daniel L. Bowling, Jacob C. Dunn, W. Tecumseh Fitch.
Visualization: Daniel L. Bowling, Jeroen B. Smaers.
Writing – original draft: Daniel L. Bowling, Jacob C. Dunn, Jeroen B. Smaers, W. Tecumseh
Writing – review & editing: Daniel L. Bowling, Jacob C. Dunn, Jeroen B. Smaers, Maxime
Garcia, Asha Sato, Georg Hantke, Stephan Handschuh, Andrew C. Kitchener, Michaela
Gumpenberger, W. Tecumseh Fitch.
1. Titze. The myoelastic aerodynamic theory of phonation. Denver, CO: National Center for Voice &
Speech; 2006.
2. Taylor AM, Reby D. The contribution of source-filter theory to mammal vocal communication research.
J Zool. 2010; 280: 221–236.
3. Elemans CPH, Rasmussen JH, Herbst CT, Du¨ring DN, Zollinger SA, Brumm H, et al. Universal mecha-
nisms of sound production and control in birds and mammals. Nat Commun. 2015; 6: 8978. https://doi.
org/10.1038/ncomms9978 PMID: 26612008
4. Herbst CT, Stoeger AS, Frey R, Lohscheller J, Titze IR, Gumpenberger M, et al. How low can you go?
Physical production mechanism of elephant infrasonic vocalizations. Science. 2012; 337: 595–9. PMID: 22859490
5. McComb K, Semple S. Coevolution of vocal communication and sociality in primates. Biol Lett. 2005; 1:
381–385. PMID: 17148212
6. Dunn JC, Halenar LB, Davies TG, Cristobal-Azkarate J, Reby D, Sykes D, et al. Evolutionary trade-off
between vocal tract and testes dimensions in howler monkeys. Curr Biol. 2015; 25: 2839–2844. https:// PMID: 26592343
7. Charlton BD, Reby D. The evolution of acoustic size exaggeration in terrestrial mammals. Nat Commun.
2016; 7: 12739. PMID: 27598835
8. Dunn JC, Smaers JB. Neural correlates of vocal repertoire in primates. Front Neurosci. 2018; 12. PMID: 30140202
9. Bowling DL, Garcia M, Dunn JC, Ruprecht R, Stewart A, Frommolt K-H, et al. Body size and vocaliza-
tion in primates and carnivores. Sci Rep. 2017; 7. PMID: 28117380
10. Chen Z, Wiens JJ. The origins of acoustic communication in vertebrates. Nat Commun. 2020; 11: 369. PMID: 31953401
11. Garcia M, Herbst CT, Bowling DL, Dunn JC, Fitch WT. Acoustic allometry revisited: morphological
determinants of fundamental frequency in primate vocal production. Sci Rep. 2017; 7.
10.1038/s41598-017-11000-x PMID: 28874852
12. Shultz S, Dunbar RIM. The evolution of the social brain: anthropoid primates contrast with other verte-
brates. Proc R Soc B Biol Sci. 2007; 274: 2429–2436. PMID:
13. Thompson PR. A cross-species analysis of carnivore, primate, and hominid behaviour. J Hum Evol.
1975; 4: 113–124.
14. Barash DP. Neighbor recognition in two “solitary” carnivores: the raccoon (Procyon lotor) and the red
fox (Vulpes fulva). Science. 1974; 185: 794–796. PMID:
Rapid evolution of the primate larynx?
PLOS Biology | August 11, 2020 18 / 21
15. McComb K, Pusey A, Packer C, Grinnell J. Female lions can identify potentially infanticidal males from
their roars. Proc R Soc London B Biol Sci. 1993; 252: 59–64.
PMID: 8389047
16. Townsend, Allen C, Manser M. A simple test of vocal individual recognition in wild meerkats. Biol Lett.
2012; 8: 179–182. PMID: 21992821
17. Pruitt CH, Burghardt GM. Communication in terrestrial carnivores: Mustelidae, Procyonidae and Ursi-
dae. In: Sebeok TA, editor. How animals communicate. Bloomington, IN: Indiana University Press;
18. Peters G. On the structure of friendly close range vocalizations in terrestrial carnivores (Mammalia: Car-
nivora: Fissipedia). Zeitschrift fu¨r Sa
¨ugetierkd. 1984; 49: 157–182.
19. Smaers JB, Rohlf FJ. Testing species’ deviation from allometric predictions using the phylogenetic
regression. Evolution (N Y). 2016; 70: 1145–1149. PMID: 27060983
20. Smaers JB, Mongle CS. Evomap: R Package for the evolutionary mapping of continuous traits. San
Francisco, CA: Github; 2018. Available from:
21. Smaers JB, Mongle CS, Kandler A. A multiple variance Brownian motion framework for estimating vari-
able rates and inferring ancestral states. Biol J Linn Soc. 2016; 118: 78–94.
22. Smaers JB, Mongle CS. On the accuracy and theoretical underpinnings of the multiple variance Brown-
ian motion approach for estimating variable rates and inferring ancestral states. Biol J Linn Soc. 2017;
121: 229–238.
23. Revell LJ. phytools: an R package for phylogenetic comparative biology (and other things). Methods
Ecol Evol. 2012; 3: 217–223.
24. Adams DC. Quantifying and comparing phylogenetic evolutionary rates for shape and other high-
dimensional phenotypic data. Syst Biol. 2014; 63: 166–177.
PMID: 24335426
25. Titze I, Riede T, Mau T. Predicting achievable fundamental frequency ranges in vocalization across
species. PLoS Comput Biol. 2016; 12: 1–13. PMID:
26. Smaers JB, Dechmann DKN, Goswami A, Soligo C, Safi K. Comparative analyses of evolutionary rates
reveal different pathways to encephalization in bats, carnivorans, and primates. Proc Natl Acad Sci.
2012; 109: 18006–18011. PMID: 23071335
27. Harrison DFN. The anatomy and physiology of the mammalian larynx. New York: Cambridge Univer-
sity Press; 1995.
28. KJ A. Joseph H. Ogura Memorial Lecture. The vertebrate larynx: adaptations and aberrations. Laryngo-
scope. 1993; 103: 1197–1201. PMID: 8412462
29. Hiiemae KM, Crompton W. Mastication, food transport, and swallowing. Functional vertebrate morphol-
ogy. Cambridge, MA: Belknap Press; 1985.
30. D. Jr. B. Respiratory Functions of the Larynx. Physiol Rev. 1989; 69: 33–56.
physrev.1989.69.1.33 PMID: 2643125
31. Bramble DM, Carrier DR. Running and breathing in mammals. Science. 1983; 219: 251–256. https:// PMID: 6849136
32. Taylor ME. Locomotor adaptations by carnivores. In: Gittleman JL, editor. Carnivore behavior, ecology,
and evolution. Boston, MA: Springer US; 1989. pp. 382–409.
33. Larson SG. Nonhuman primate locomotion. Am J Phys Anthropol. 2018; 165: 705–725.
10.1002/ajpa.23368 PMID: 29574838
34. Taverne M, Herbin AFM, Herrel A, Peigne
´S, Lacroux C, Lowie A, et al. Convergence in the functional
properties of forelimb muscles in carnivorans: adaptations to an arboreal lifestyle? Biol J Linn Soc.
2018; 125: 250–263.
35. Ey E, Fischer J. The “acoustic adaptation hypothesis”—a review of the evidence from birds, anurans
and mammals. J Bioacoustics. 2009; 19: 21–48.
36. Morton ES. Ecological sources of selection on avian sounds. Am Nat. 1975; 109: 17–34.
37. Puts DA, Hill AK, Bailey DH, Walker RS, Rendall D, Wheatley JR, et al. Sexual selection on male vocal
fundamental frequency in humans and other anthropoids. Proc R Soc B Biol Sci. 2016; 283: 0–7. PMID: 27122553
38. Owen R. On the anatomy of the cheetah, Felis jubata, Schreb. Trans Zool Soc. 1834; 1: 129–136.
39. Pocock RI. On the hyoidean apparatus of the lion (F.leo) and related species of Felidae. Ann Mag Nat
Hist. 1916; 8: 222–229.
Rapid evolution of the primate larynx?
PLOS Biology | August 11, 2020 19 / 21
40. Pe
´a FJ, Shultz S, Dunbar RIM. Evidence for coevolution of sociality and relative brainsize in
three orders of mammals. Evolution (N Y). 2007; 61: 2811–2821.
2007.00229.x PMID: 17908248
41. Wrangham RW, Gittleman JL, Chapman CA. Constraints on group size in primates and carnivores:
population density and day-range as assays of exploitation competition. Behav Ecol Sociobiol. 1993;
32: 199–209.
42. Grodzinsky AJ, Levenston ME, Jin M, Frank EH. Cartilage tissue remodeling in response to mechanical
forces. Annu Rev Biomed Eng. 2000; 2: 691–713.
PMID: 11701528
43. Bergman TJ, Beehner JC. Measuring social complexity. Anim Behav. 2015; 103: 203–209. https://doi.
44. Schneider R. Der Larynx des ma
¨nnlichen Hypsignathus monstrosus Allen, 1861 (Pteropodidae, Mega-
chiroptera, Mammalia). Zeitschrift fu¨r Wissenschaftliche Zool. 1964; 175: 1–53.
45. Eckel HE, Sittel C, Zorowka P, Jerke A. Dimensions of the laryngeal framework in adults. Surg Radiol
Anat. 1994; 16: 31–36. PMID: 8047966
46. Mittermeier RA, Rylands A, Wilson DE, editors. Handbook of the mammals of the world, volume 3: pri-
mates. Barcelona: Lynx Edicions, Conservation International, IUCN; 2013.
47. Wilson DE, Mittermeier RA, editors. Handbook of the mammals of the world, volume 1: carnivores. Bar-
celona: Lynx Edicions, Conservation International, IUCN; 2009.
48. Revelle W. psych: procedures for personality and psychological research. package version 1.6.12 [soft-
ware]. 2016. Available from:
49. Kaiser HF. The application of electronic computers to factor analysis. Educ Psychol Meas. 1960; 20:
50. Arnold C, Matthews LJ, Nunn CL. The 10kTrees website: a new online resource for primate phylogeny.
Evol Anthropol. 2010; 19: 114–118.
51. Nyakatura K, Bininda-emonds ORP. Updating the evolutionary history of Carnivora (Mammalia): a new
species-level supertree complete with divergence time estimates. BMC Biol. 2012; 10: 1–31. https://doi.
52. Bastide P, Mariadassou M, S. R. Detection of adaptive shifts on phylogenies using shifted stochastic
processes on a tree. J R Stat Soc. 2017; 79: 1067–1093. Available:
53. Khabbazian M, Kriebel R, Rohe K, Ane
´C. Fast and accurate detection of evolutionary shifts in Orn-
stein–Uhlenbeck models. Methods Ecol Evol. 2016; 7: 811–824.
54. Cressler CE, Butler MA, King AA. Detecting adaptive evolution in phylogenetic comparative analysis
using the Ornstein-Uhlenbeck model. Syst Biol. 2015; 64: 953–968.
syv043 PMID: 26115662
55. Adams DC, Ota
´rola-Castillo E. Geomorph: an R package for the collection and analysis of geometric
morphometric shape data. Methods Ecol Evol. 2013; 4: 393–399.
56. Denton JSS, Adams DC. A new phylogenetic test for comparing multiple high-dimensional evolutionary
rates suggests interplay of evolutionary rates and modularity in lanternfishes (Myctophiformes; Mycto-
phidae). Evolution (N Y). 2015; 69: 2425–2440. PMID: 26278586
57. Mueller PB, Sweeney RJ, Baribeau LJ. Senescence of the voice: morphology of excised male larynges.
Folia Phoniat. 1985; 37: 134–138. PMID: 4054768
58. Frey R, Riede T. The anatomy of vocal divergence in North American Elk and European red deer. J
Morphol. 2013; 274: 307–319. PMID: 23225193
59. Weigl R. Longevity of mammals in captivity; from the living collections of the world. Frankfurt: E.
Schweizerbart’sche Verlagsbuchhandlung; 2005.
60. Hayssen V, van Tienhoven A, van Tienhoven A. Asdell’s patterns of mammalian reproduction, a com-
pendium of species-specific data. Ithaca, NY: Cornell University Press; 1993.
61. Schwitzer C, Mittermeier RA, Rylands AB, Taylor LA, Chiozza F, Williamson EA, et al., editors. Pri-
mates in peril: the world’s 25 most endangered primates. Arlington: IUCN; 2014.
62. Abernethy KA, White LJT, Wickings EJ. Hordes of mandrills (Mandrillus sphinx): extreme group size
and seasonal male presence. J Zool. 2002; 258: 131–137.
63. Swedell L. Ranging behavior, group size and behavioral flexibility in Ethiopian hamadryas baboons
(Papio hamadryas hamadryas). Folia Primatol. 2002; 73: 95–103.
PMID: 12207056
Rapid evolution of the primate larynx?
PLOS Biology | August 11, 2020 20 / 21
64. Garcı
´a-Aguilar MC, Aurioles-Gamboa D. Breeding season of the California sea lion (Zalophus califor-
nianus) in the Gulf of California, Mexico. Aquat Mamm. 2005; 29: 67–76.
65. Fletcher NH. A simple frequency-scaling rule for animal communication. J Acoust Soc Am. 2004; 115:
2334–2338. PMID: 15139646
66. Martin K, Tucker MA, Rogers TL. Does size matter? Examining the drivers of mammalian vocalizations.
Evolution (N Y). 2017; 71: 249–260. PMID: 27882540
67. Hauser MD. The evolution of nonhuman primate vocalizations: effects of phylogeny, body weight, and
social context. Am Nat. 1993; 142: 528–42. PMID: 19425989
68. Isler K, Kirk EC, Miller JMA, Albrecht GA, Gelvin BR, Martin RD. Endocranial volumes of primate spe-
cies: scaling analyses using a comprehensive and reliable data set. J. Hum. Evol. 2008; 55: 967–978. PMID: 18817943
69. Kamilar JM, Cooper N. Phylogenetic signal in primate behavior, ecology and life history. Phil. Trans.
Royal Soc. B. 2013; 368: 20120341. PMID: 23569289
70. Borgwardt N, Culik BM. Asian small-clawed otters (Amblonyx cinerea): resting and swimming metabolic
rates. J. Comp. Physiol. B. 1999; 169: 100–106. PMID:
71. Lambert JE, Fellner V, McKenney E, Hartstone-Rose A. Binturong (Arctictis binturong) and kinkajou
(Potos flavus) digestive strategy: implications for interpreting frugivory in carnivora and primates. PLoS
ONE. 2014; 9: e105415. PMID: 25157614
72. Bidau CJ, Martinez PA. Sexual size dimorphism and Rensch’s rule in Canidae. Biol. J. Linn. Soc. 2016;
119: 816–830.
73. de Oliveira TG. Herpailurus yagouaroundi. Mamm. Species. 1998; 578: 1–6.
74. Peters G, Peters MK. Long-distance call evolution in the Felidae: effects of body weight, habitat, and
phylogeny. Biol. J. Linn. Soc. 2010; 101: 487–500.
75. Gompper ME, Decker DM. Nasua nasua. Mamm. Species. 1998; 580: 1–9.
76. Grassman LI Jr., Tewes ME, Silvy NJ, Kreetiyutanont K. Ecology of three sympatric felids in a mixed
evergreen forest in north-central Thailand. J. Mammal. 2005; 1: 29–38.
77. Kays R. Social polyandry and promiscuous mating in a praimte-like carnivore: the kinkajou (Potos fla-
vus). In: Reichard UH, Boesch C, editors. Monogamy: mating strategies and partnerships in birds,
humans and other mammals. Cambridge, UK: Cambridge University Press; 2003. pp. 125–137.
78. Johansson O
¨, Malmsten J, Mishra C, Lkhagvajav P, McCarthy T. Reversible immobilization of free-
randing snow leopards (Panthera uncia) with a combination of medetomidine and tiletamine-zolaze-
pam. J. Wildl. Dis. 2013; 49: 338–346. PMID: 23568909
Rapid evolution of the primate larynx?
PLOS Biology | August 11, 2020 21 / 21
... Vocal fold length was approximated based on the mineralization patterns of the 3D reconstructions. We placed landmarks on the dorsal border of each arytenoid cartilage, and in the area of vocal fold attachment on the thyroid cartilage; we then measured the distance between the arytenoid and thyroid landmarks, resulting in two vocal fold measurements per individual (right and left; Bowling et al., 2020). We recorded the average value as "mean vocal fold length" for each individual to be used for all further statistical analysis. ...
... Many similar modifications for loud low-frequency vocalizations have been documented in a range of ungulates (Frey et al., 2007;Frey et al., 2011;Frey & Hofmann, 2000;Frey & Riede, 2003). Nonhuman primates show a suite of specific vocal modifications, including many types and sizes of laryngeal air sac (Fitch, 2016;Schön Ybarra, 1995), and large-scale comparisons of laryngeal morphology in primates and carnivores suggest that the primate larynx has evolved more rapidly than that of carnivores (Bowling et al., 2020). ...
... This finding is clearly relevant to researchers using USVs to gain insight into the emotional state and stress response, as well as to broader questions in the evolution of communication (cf. Bowling et al., 2020;Dunn et al., 2015;Fitch, 2016). ...
Full-text available
Although laryngeal morphology often reflects adaptations for vocalization, the structural consequences of selection for particular aspects of vocal behavior remain poorly understood. In this study, we investigated the effects of increased ultrasonic calling in pups on the adult larynx morphology in selectively bred rat lines. Laryngeal morphology was assessed using multiple techniques: mineralized cartilage volumes were compared in 3D‐models derived from microCT scans, internal structure was compared using clearing and staining procedures combined with microscopy, cellular structure was compared using histology and microscopy, and element composition was assessed with scanning energy dispersive X‐ray spectroscopy. Our results show that adult rats from lines bred to produce ultrasonic calls at higher rates as pups have shorter vocal folds and a more mineralized thyroid cartilage compared to rats bred to produce ultrasonic calls at lower rates. The change in vocal fold length appears to account for differences in low‐frequency calls in these two rat lines. We suggest that the observed increases in mineralization of the thyroid cartilage in the high‐ultrasound lineage provide increased reinforcement of the laryngeal structure during ultrasonic call production. Our findings therefore demonstrate an effect of selection for vocal behavior on laryngeal morphology, with acoustic consequences.
... Unlike motor systems controlling locomotion and respiration (Ramirez and Baertsch, 2018;Grillner and El Manira, 2020), in-depth inquiries of cellular and network properties of brainstem neurons influencing acoustic signaling remain relatively unexplored, despite advances in characterizing the musculoskeletal periphery (Mead et al., 2017;Kwong-Brown et al., 2019;Riede et al., 2019;Bowling et al., 2020). A neurobehavioral challenge often facing soniferous species is fine temporal control of rapid modulations of acoustic waveforms. ...
Full-text available
Precise neuronal firing is especially important for behaviors highly dependent on the correct sequencing and timing of muscle activity patterns, such as acoustic signaling. Acoustic signaling is an important communication modality for vertebrates, including many teleost fishes. Toadfishes are well known to exhibit high temporal fidelity in synchronous motoneuron firing within a hindbrain network directly determining the temporal structure of natural calls. Here, we investigated how these motoneurons maintain synchronous activation. We show that pronounced temporal precision in population-level motoneuronal firing depends on gap junction-mediated, glycinergic inhibition that generates a period of reduced probability of motoneuron activation. Super-resolution microscopy confirms glycinergic release sites formed by a subset of adjacent premotoneurons contacting motoneuron somata and dendrites. In aggregate, the evidence supports the hypothesis that gap junction-mediated, glycinergic inhibition provides a timing mechanism for achieving synchrony and temporal precision in the millisecond range for rapid modulation of acoustic waveforms.
... The meaning of "pre-exposure" is ill-defined in consonance research. Although "before exposure to music" is often implied, there is no clear line that differentiates what we call music from other forms of tone-based auditory-vocal communication, and the auditory system receives harmonic stimulation from mother's voice as soon as it comes online, which it has presumably done for hundreds of millions of years(Bowling et al., 2020; Chen & Wiens, 2020;Chiandetti & Vallortigara, 2011;Negus, 1949). ...
Evidence supporting a link between harmonicity and the attractiveness of simultaneous tone combinations has emerged from an experiment designed to mitigate effects of musical enculturation. I examine the analysis undertaken to produce this evidence and clarify its relation to an account of tonal aesthetics based on the biology of auditory-vocal communication.
Full-text available
The principle of acoustic allometry—the larger the animal, the lower its calls' fundamental frequency—is generally observed across terrestrial mammals. Moreover, according to the Acoustic Adaptation Hypothesis, open habitats favor the propagation of high-frequency calls compared to habitats with complex vegetational structures. We carried out playback experiments in which the calls of the Guizhou snub-nosed monkey (Rhinopithecus brelichi) were used as stimuli in sound attenuation and degradation experiments to test the hypothesis that propagation of Guizhou snub-nosed monkey calls is favored above vs through the forest floor vegetation. We found that low-pitched Guizhou snub-nosed monkey vocalizations suffered less attenuation than its high-pitched calls. Guizhou snub-nosed monkeys were observed emitting high-pitched calls from 1.5 to 5.0 m above the ground. The use of high-pitched calls from these heights coupled with the concomitant behavior of moving about above the understory may provide a signal for receivers which maximizes potential transmission and efficacy. Our results support the Acoustic Adaptation Hypothesis and suggest that by uncoupling its vocal output from its size, this monkey can produce a high-pitched call with a broad spectral bandwidth, thereby increasing both its saliency and the frequency range over which the animal may more effectively communicate in its natural habitat.
Small carnivorans are generally poorly represented in zoos, probably because they are small, mostly nocturnal and solitary hunters. However, there is limited knowledge about the ecology and behaviour of a large number of these and many species are threatened with extinction or their conservation status is poorly known or even unknown. Although rare in zoos, there are good opportunities for zoos and museums to cooperate to ensure that when animals die, they are subjected to careful post-mortem examinations, and preserved for research into anatomy and functional morphology. In turn museum collections are rich stores of specimens that assist zoos in identification, and provide veterinarians with access to animal remains and their anatomy to facilitate treatment and surgery. New techniques, such as computerized tomography (CT) and magnetic resonance imaging (MRI) scanning, plus the establishment of biobanks allow new ways of capturing vital information about small carnivorans for a wide range of research, including taxonomy and systematics, archaeology and palaeontology, anatomy, pathology and conservation science. Most importantly museum collections of zoo specimens allow the impacts of captivity, such as diet and activity levels, to be investigated. However, the development of this research resource relies on closer cooperation between zoos and museums. As wild and captive populations are increasingly managed under the International Union for Conservation of Nature’s Conservation Planning Specialist Group One Plan approach, it will also be increasingly important for zoos and museums to work together to benefit the conservation of threatened small carnivorans.
Full-text available
Acoustic communication is crucial to humans and many other tetrapods, including birds, frogs, crocodilians, and mammals. However, large-scale patterns in its evolution are largely unstudied. Here, we address several fundamental questions about the origins of acoustic communication in terrestrial vertebrates (tetrapods), using phylogenetic methods. We show that origins of acoustic communication are significantly associated with nocturnal activity. We find that acoustic communication does not increase diversification rates, a surprising result given the many speciation-focused studies of frog calls and bird songs. We also demonstrate that the presence of acoustic communication is strongly conserved over time. Finally, we find that acoustic communication evolved independently in most major tetrapod groups, often with remarkably ancient origins (~100–200 million years ago). Overall, we show that the role of ecology in shaping signal evolution applies to surprisingly deep timescales, whereas the role of signal evolution in diversification may not. Acoustic communication is widespread, but not universal, across terrestrial vertebrates. Here, the authors show that acoustic communication evolved anciently, but independently, in most tetrapod groups and that these origins were associated with nocturnal activity.
Full-text available
The arboreal environment is complex and irregular and imposes significant constraints on the locomotor ability of vertebrates, resulting in morphological adaptations. Carnivorans are of particular interest because they radiated several times independently into the arboreal milieu, leading to convergence in long bone anatomy. We investigated whether the forelimb muscles show differences in arboreal vs. terrestrial species and tested for convergence in muscle architecture. To do so, we studied the forelimb muscles in 17 species of carnivorans with different locomotor modes (arboreal and terrestrial). Our results showed that the forelimb musculature evolved in a convergent manner in arboreal species, resulting in more forceful and heavier wrist rotators and elbow flexors, even when accounting for phylogeny. This suggests that selective forces imposed by the environment drive the evolution of the muscular system towards a convergent phenotype that provides a functional advantage for locomotion in the complex three-dimensional arboreal habitat.
Full-text available
Understanding the nature of the relationship between vocal complexity and brain architecture across non-human primates may help elucidate some of the key elements underlying the evolution of human speech. Here, we report a positive correlation between vocal repertoire size and the relative size of cortical association areas (governing voluntary control over behavioural output) in non-human primates. We further demonstrate that a hominid grade shift in the relative volume of cortical association areas coincides with a similar grade shift in the hypoglossal nucleus (which is associated with the cranial nerve that innervates the muscles of the tongue). Our results support a qualitative continuity in the neural correlates of vocal repertoire, but a quantitative discontinuity in the extent to which the neural system supporting speech is innervated by cortical association areas in great apes and humans.
Full-text available
A fundamental issue in the evolution of communication is the degree to which signals convey accurate (" honest ") information about the signaler. In bioacoustics, the assumption that fundamental frequency (fo) should correlate with the body size of the caller is widespread, but this belief has been challenged by various studies, possibly because larynx size and body size can vary independently. In the present comparative study, we conducted excised larynx experiments to investigate this hypothesis rigorously and explore the determinants of f o. Using specimens from eleven primate species, we carried out an inter-specific investigation, examining correlations between the minimum fo produced by the sound source, body size and vocal fold length (VFL). We found that, across species, VFL predicted minimum f o much better than body size, clearly demonstrating the potential for decoupling between larynx size and body size in primates. These findings shed new light on the diversity of primate vocalizations and vocal morphology, highlighting the importance of vocal physiology in understanding the evolution of mammal vocal communication.
Full-text available
A fundamental assumption in bioacoustics is that large animals tend to produce vocalizations with lower frequencies than small animals. This inverse relationship between body size and vocalization frequencies is widely considered to be foundational in animal communication, with prominent theories arguing that it played a critical role in the evolution of vocal communication, in both production and perception. A major shortcoming of these theories is that they lack a solid empirical foundation: rigorous comparisons between body size and vocalization frequencies remain scarce, particularly among mammals. We address this issue here in a study of body size and vocalization frequencies conducted across 91 mammalian species, covering most of the size range in the orders Primates (n = 50; ~0.11–120 Kg) and Carnivora (n = 41; ~0.14–250 Kg). We employed a novel procedure designed to capture spectral variability and standardize frequency measurement of vocalization data across species. The results unequivocally demonstrate strong inverse relationships between body size and vocalization frequencies in primates and carnivores, filling a long-standing gap in mammalian bioacoustics and providing an empirical foundation for theories on the adaptive function of call frequency in animal communication.
Full-text available
Previous studies of the vocalisation frequencies of mammals have suggested that it is either body mass or environment that drives these frequencies. Using 193 species across the globe from the terrestrial and aquatic environments and a model selection approach, we identified that the best supported model for minimum and maximum frequencies for vocalisation included both body mass and environment. The minimum frequencies of vocalisations of species from all environments retained the influence of body mass. For maximum frequency however, aquatic species are released from such a trend with body mass having little constraint on frequencies. Surprisingly, phylogeny did not have a strong impact on the evolution of the maximum frequency of mammal vocalisations, largely due to the pinniped species divergence of frequency from their carnivoran relatives. We demonstrate that the divergence of signal frequencies in mammals has arisen from the need to adapt to their environment. This article is protected by copyright. All rights reserved
In a recent contribution, Griffin & Yapuncich (2016) (‘G&Y’) report on purported inaccuracies and lack of theoretical underpinnings of our recently proposed ancestral estimation procedure. Our method is designed to provide a good overall fit with the data when different branches may be subject to different rates of change (multiple variance Brownian motion, or ‘mvBM’). Here we demonstrate that G&Y’s theoretical concerns stem from a misinformed account of basic statistical concepts and procedures, a misinterpretation of the primary literature and a circular adherence to a restrictive model of evolution (standard Brownian motion, or ‘standard BM’) whose usefulness has long been considered inappropriate for modelling branch-specific evolutionary patterns. We further apply a series of simple tests that falsify G&Y’s claims on every account. Finally, we show that including a range of sample sizes (from 4 through 100) to G&Y’s own suggested simulation procedure further substantiates what they purport to falsify: the validity of mvBM when modelling potential deviations from standard BM in trait evolution.