Osteological correlates of tail prehensility in carnivorans

Abstract

Among mammalian morphological specializations to arboreality, prehensile tails are the least studied. In carnivorans, two phylogenetically and ecologically distant representatives possess truly prehensile tails: the kinkajou Potos flavus, a neotropical procyonid and the binturong Arctictis binturong, a viverrid from south-east Asia. This paper examines osteological characters associated with tail prehensility by comparing carnivorans with and without prehensile tails. The prehensile-tailed taxa are characterized by: (a) a relatively longer proximal caudal region in length and number of vertebrae; (b) more robust distal caudal vertebrae, which possess expanded transverse processes. Comparable features have been also reported for prehensile-tailed primates, indicating evolutionary convergence. These features can be functionally associated with enhanced flexion–extension of the proximal part of the tail and increased strength and flexing capacity at the distal end of the tail. These tail movements are briefly described in free-ranging prehensile-tailed carnivorans.

Full text

J. Zool., Lond. (2003) 259,423–430 C
2003 The Zoological Society of London Printed in the United Kingdom DOI:10.1017/S0952836903003431
Osteological correlates of tail prehensility in carnivorans
Dionisios Youlatos*
Aristotle University of Thessaloniki, School of Biology, Department of Zoology, GR-54124, Thessaloniki, Greece
(Accepted 14 August 2002)
Abstract
Among mammalian morphological specializations to arboreality, prehensile tails are the least studied. In
carnivorans, two phylogenetically and ecologically distant representatives possess truly prehensile tails: the kinkajou
Poto s flavus,aneotropical procyonid and the binturong Arctictis binturong,aviverridfrom south-east Asia. This
paper examines osteological characters associated with tail prehensility by comparing carnivorans with and without
prehensile tails. The prehensile-tailed taxa are characterized by: (a) a relatively longer proximal caudal region in
length and number of vertebrae; (b) more robust distal caudal vertebrae, which possess expanded transverse
processes. Comparable features have been also reported for prehensile-tailed primates, indicating evolutionary
convergence. These features can be functionally associated with enhanced flexion–extension of the proximal part
of the tail and increased strength and flexing capacity at the distal end of the tail. These tail movements are briefly
described in free-ranging prehensile-tailed carnivorans.
Key words:Po t o s flavus,Arctictis binturong,Carnivora,prehensile tail, vertebrae
INTRODUCTION
The specific mechanical properties, limited dimensions
and random disposition of branches compel arboreal
dwellers to constant problems of balance and safety from
falling. To cope with these major problems, arboreal
mammals have evolved anatomical and behavioural
specializations that involve relatively shorter limbs, sharp
claws, suspensory habits, and prehensile appendages
including tails (Cartmill, 1985).
Among these specializations, the prehensile tail is
perhaps the least well understood. However, it is encoun-
tered in most didelphid and phalangerid marsupials,
anteaters, pangolins, some rodents, and a few primates
and carnivorans. Emmons & Gentry (1983: 513) defined
aprehensile tail as the ‘... one which can support alone
the weight of the suspended body’ to distinguish from the
semi-prehensile tail ‘... which can support a significant
part, but not all, of body weight’. Among mammals,
tail prehensility has been principally studied in the
prehensile-tailed neotropical monkeys involving Cebus
and the monophyletic Atelinae (Alouatta,Lagothrix,
Ateles,Brachyteles). In these monkeys, behavioural
studies have associated the use of prehensile tails with the
negotiation of large body size on small branches (Grand,
1972), the use of palm fronds and the relatively fragile
*E-mail: dyoul@bio.auth.gr
branches in neotropical forests (Emmons & Gentry, 1983),
and the accommodation of suspensory foraging for fruit
and flowers on the extremities of tree crowns (Grand,
1972; Cant, 1986; Garber & Rehg, 1999; Youlatos,
1999). These behaviours are further reflected in some
morphological specializations exhibited by convergence
in Cebus and the Atelinae: a developed area in the
cortex, an expanded sacroiliac joint and long proximal
caudal region, a powerful caudal flexor musculature and
well-developed mm. intertransversarii that span a few
vertebrae, and short, robust distal vertebrae with expanded
transverse processes (Dor, 1937; Ankel, 1972; Grand,
1977; Falk, 1980; German, 1982; Rosenberger, 1983;
Lemelin, 1995).
Some of these morphological specializations have also
been reported for the prehensile-tailed representatives of
other mammalian orders implying a comparable use of the
prehensile tail (Dor, 1937; Grand, 1977). More precisely,
two prehensile-tailed carnivorans, the kinkajou Potos
flavus (Shreber, 1774) (Procyonidae) and the binturong
Arctictis binturong (Raffles, 1821) (Viverridae) seem to
possess a high number of proximal caudal vertebrae, short
and robust distal caudal vertebrae, long flexor and extensor
tendons that span a few vertebrae and well developed mm.
intertransversarii (Dor, 1937). Kinkajous are medium-
sized solitary frugivores that exploit the high canopy
layers where they travel by climbing and clambering,
while binturongs are large omnivores that exploit all
forest layers travelling alone or in small groups and

424 D. YOULATOS
Tab l e 1 . Morphological and behavioural features of the studied viverrids and procyonids listed by body weight in each family. Habits: N,
nocturnal; C, crepuscular; D, diurnal. Substrate use: A, arboreal; T, terrestrial. Foods: M, animal matter; F, fruit-seeds; V, leaves/grasses;
E, nectar (prehensile-tailed taxa in bold). Superscript numbers, source (see footnote)
Geographic Habits6Substrate Foods1,4,6,8,9Mass6Head–Body Tail length6TL/HBL Tail use 1,5−7
range6use 1−7(kg) length6TL (mm)
HBL (mm)
Viver ridae
Arctictis SE Asia N A M, F, L 9–14 610–965 560–890 0.92 Prehensile
Cryptoprocta Madagascar N-C T, A M 7–12 610–800 610–800 1.00 Non-prehensile
Par adoxurus S, SE Asia N A M, F 1.5–4.5 432–710 406–660 0.93 Non-prehensile
Genetta Africa N A, T M 1–3420–580 390–530 0.91 Non-prehensile
Nandinia CAfrica N A F, M 1.7–2.1 440–580 460–620 1.07 Non-prehensile
Procyonidae
Procyon N, C, S America N-D T, A M, F 2–12 415–600 200–405 0.67 Non-prehensile
Nasua C, S America D A, T F, M 3–6410–670 320–690 1.03 Semi-prehensile (?)
Pot os C, S America N A F, N, M 1.4–4.6 405–760 392–570 0.75 Prehensile
Bassariscus CAmerica N T, A M, F 0.82–1.33 305–420 310–441 1.05 Non-prehensile
1Kaufmann, 1962; 2Taylor, 1970; 3Trapp, 1972; 4Charles-Dominique, 1978; 5Laborde, 1986; 6Nowak, 1991; 7McClearn, 1992; 8Julien-
Laferri`
ere, 1999; 9Kays, 1999.
(?) Kaufman proposed semi-prehensile tail but McClearn disagrees.
cautiously among tree branches (Table 1). Moreover, both
species share a nocturnal, mainly arboreal way of life.
Limited field studies report that the tail is grasped in tail-
only and tail-assisted suspensory feeding postures and is
coiled around branches during deliberate clambering and
head-first descent (Dor, 1937; Nowak, 1991; McClearn,
1992). Among other carnivorans, coatis Nasua nasua (L.,
1766) (Procyonidae) have been observed to use their
tails (classified as semi-prehensile by Kaufmann, 1962)
frequently for stabilization during postures or head-first
descent, but mainly to keep them off any arboreal supports
(McClearn, 1992). In addition, fossas Cryptoprocta ferox
Bennett, 1833 (Viverridae) sometimes coil the extremity
of the tail around branches in head-first descents but
primarily use it as counterbalance in locomotion and
postures (Albignac, 1970; Laborde, 1986). For these
carnivorans, there are no reports of dexterous tail use or tail
hanging postural and locomotor patterns and no studies
reporting any osteological and muscular similarities or
differences with their prehensile-tailed relatives.
Based on descriptions of differences in tail use, this
study aims to investigate any osteological features of
the caudal vertebrae associated with tail prehensility
in carnivorans. As discussed above, similar features
have been already examined in the caudal vertebrae of
the prehensile-tailed primates (Ankel, 1972; German,
1982). These features have successfully distinguished
prehensile-tailed from non-prehensile-tailed primates and
their validity can be tested in carnivorans. The sacroiliac
joint (SIJ) interconnects the medial surface of the ilium to
the sacrum and is usually expanded in prehensile-tailed
taxa. The tail is divided into a proximal and a distal
caudal part, based on the morphology of the proximal and
distal vertebrae respectively. The proximal vertebrae bear
ventral and neural arches, a pair of transverse processes
and articulate with zygapophyses, like the lumbar
vertebrae (Fig. 1a). This enables enhanced sagittal
flexibility in the region that would be further related to
(a) (b) (c)
31
2
Fig. 1. Dorsal view of the form of caudal vertebrae of a typical car-
nivore (Canis): (a), proximal; (b) transitional; (c) distal. Measure-
ments: (1) vertebral length; (2) proximal width; (3) mid-width.
the relative length of the region, the number of vertebrae
involved as well as individual length of these units (Slijper,
1946; Ward, 1993; Shapiro, 1995). The distal vertebrae are
long and rounded in cross-section and articulate through
intervertebral discs only (Fig. 1c). Their morphology and
overall robusticity would indicate the presence or absence
of high and frequent extrinsic (gravity, reaction) and
intrinsic (muscles, tendons, ligaments) stresses (German,
1982; Wainwright et al., 1982; Lemelin, 1995). At the
junction between the two regions of the tail, lies the transit-
ional vertebra that bears a proximal articulation with
zygapophyses and a distal articulation with a vertebral
disc only (Fig. 1b). Caudad to the transitional vertebra,
vertebral length increases gradually until the longest
vertebra is reached. The relative location of these
vertebrae is more caudal in the prehensile-tailed taxa
(Ankel, 1972). These landmarks divide the tail into a
proximal region, limited between the first caudal and the
transitional vertebra, a transitional region, limited between
the transitional and the longest distal vertebra, and a distal
region, lying caudad to the longest vertebra to the tip of
the tail. Prehensile-tailed primates bear a longer proximal
caudal region.

Carnivoran prehensile tails 425
Based on these observations in primates, the aim of this
study is to test whether similar osteological differences
occur in the caudal vertebrae of the prehensile-tailed and
non-prehensile-tailed taxa of related carnivoran families.
If the two distantly related prehensile-tailed carnivorans
present common osteological features in the caudal
vertebrae this would most likely suggest comparable
adaptations related to the prehensile action of the tail.
The lack of such features in the other phylogenetically
related arboreal and scansorial carnivorans that do not use
their tail in any prehensile way would further argue in
favour to the convergent adaptations to tail prehensility.
Unfortunately, the lack of systematic behavioural data
of tail use for both prehensile-tailed taxa cannot
provide firm correlations between vertebral morphology
and tail function. On the other hand, this study could
provide further insight for the study of osteological
correlates to tail prehensility in other mammalian taxa,
such as arboreal marsupials and rodents.
MATERIALS AND METHODS
The axial skeleton of museum specimens of the viverrids
and procyonids listed in Table 1 were examined. All
the studied material is housed in the collections of the
Laboratoire d’Anatomie Compar´
ee and the Laboratoire
des Mammif`
eres et Oiseaux of the Mus´
eum National
d’Histoire Naturelle in Paris, France. The genera exam-
ined were selected for their main arboreal-scansorial way
of life (Table 1) and their close phylogenetic relationship
with the 2 prehensile-tailed carnivorans (Bininda-
Emonds, Gittleman & Purvis, 1999). A preliminary survey
of tail skeletons of terrestrial viverrids showed no
qualitative and quantitative differences with their arboreal-
scansorial relatives.
The first step was to identify the landmark (proximal,
transitional, longest and distal) vertebrae and define the
proximal, transitional and distal caudal regions (Table 2).
This was not an easy task considering that many tail
skeletons were not articulated and many vertebrae were
missing. Thus, fewer tail skeletons were used for collec-
ting quantitative data than for identifying the landmark
vertebrae (see differences in sample sizes in Tables 2–4).
Once the landmark vertebrae were identified in complete
tail skeletons, the relative number of vertebrae included
in each region was expressed as a percentage of the total
number of caudal vertebrae for each skeleton. The percent-
ages for each specimen were used to calculate means for
each taxon (Table 3). For reasons of comparability with
published studies (German, 1982), similar measurements
were used on the caudal vertebrae that have successfully
traced quantitative differences between prehensile-tailed
and non-prehensile-tailed primates (Fig. 1c):
(1) length:the craniocaudal dimension;
(2) proximal width:the mediolateral dimension of the
proximal end including vertebral body and processes;
(3) mid-width:the mediolateral dimension at mid-
length restricted to vertebral body (this measurement
requires precision on the proximal vertebrae).
Tab l e 2 . Skeletal features in the tails of the studied viverrids and
procyonids listed by body weight (the prehensile-tailed taxa are
in bold). SIJ, number of sacral vertebrae involved in the sacroiliac
joint; TRANS, position of the transitional vertebra; LONG, position
of the longest vertebra; TOTAL, number of all caudal vertebrae
SIJ TRANS LONG TOTAL n
Viver ridae
Arctictis 28–10 11–13 32 4
Cryptoprocta 1–2 5 10–11 31 2
Par adoxurus 1–2 5–6 10–11 28 4
Genetta 1–2 5 9 26 3
Nandinia 1–2 5–6 9–10 28–31 5
Procyonidae
Procyon 25 8 15–17 5
Nasua 25–6 9–10 22–24 9
Pot os 26–7 10–11 26–28 5
Bassariscus 15 8 20 1
Tab l e 3 . Mean percentages of proximal, transitional and distal
caudal regions based on the number of the respective vertebrae
on total number of tail vertebrae. In each family the taxa are listed
by body weight; the prehensile-tailed taxa in bold
Proximal (%) Transitional (%) Distal (%) n
Viver ridae
Arctictis 28.1 9.4 62.5 2
Cryptoprocta 16.1 16.1 67.8 1
Par adoxurus 17.8 17.8 64.4 2
Genetta 19.2 15.3 65.5 2
Nandinia 16.1 16.1 67.8 3
Procyonidae
Procyon 30.1 18.1 51.8 3
Nasua 22.8 20.1 57.1 3
Pot os 23.9 14.6 61.5 4
Bassariscus 25.0 15.0 60.0 1
Ve r tebral length was used to calculate the lengths of the
3caudal regions:
(a) total tail length:the sum of lengths of all caudal
vertebrae;
(b) proximal region length:the sum of lengths of
the proximal vertebrae between the first and the
transitional included;
(c) transitional region length:the sum of lengths of the
vertebrae between the transitional and the longest
included;
(d) distal region length:the sum of lengths of the distal
vertebrae caudad to the longest vertebra.
Thus, each region was also expressed as a percentage of
total tail length (Table 4).
In addition, the mean vertebral length was calculated
for each caudal region as the ratio of the length of each
region (L=sum of lengths of vertebrae) to the number
of vertebrae involved (n). Mean vertebral length was
subsequently divided by the cube root of body weight
from the literature. Body weight is proportional to the
third power of linear dimensions and its cube root, which
makes the units comparable to linear measures, and is

426 D. YOULATOS
Tab l e 4 . Mean percentages of proximal, transitional and distal
caudal regions based on the sum of the lengths of the respective
vertebrae on total tail length (sum of the lengths of all tail vertebrae).
In each family the taxa are listed by body weight; the prehensile-
tailed taxa in bold
Proximal (%) Transitional (%) Distal (%) n
Viver ridae
Arctictis 30.2 12.8 57.0 2
Cryptoprocta 13.4 18.0 68.6 1
Par adoxurus 14.0 25.1 60.9 2
Genetta 13.1 20.7 66.2 2
Nandinia 11.9 22.4 65.7 3
Procyonidae
Procyon 21.7 24.8 53.5 3
Nasua 12.4 29.7 57.9 3
Pot os 21.9 18.4 59.7 4
Bassariscus 13.9 23.0 63.1 1
used to facilitate comparisons between taxa that vary in
size (Shapiro, 1995). Mann–Whitney U-tests were used
for statistical comparisons of the relative lengths of the
different caudal regions and mean vertebral lengths (Zar,
1996).
The 3 vertebral measurements were used to calculate 2
indices:
(1) robusticity index =([mid-width]/[vertebral length])∗
100; this reflects the robustness of each vertebra,
which is closely related to the magnitude of the applied
forces on the bone. A more robust vertebra is expected
to withstand higher exerted forces than a more gracile
one (German, 1982);
(2) relative expansion of the transverse process index =
([proximal width]/[vertebral length])∗100; this meas-
ures the expansion of the attachments of the main
rotator and flexors of the tail. Higher values indicate
more expanded muscular attachments and accordingly
more powerful muscles (German, 1982; Lemelin,
1995).
The values of both indices for each vertebra of the
specimens of a species were tested for intraspecific
variability using the Wilcoxon signed ranks test (Zar,
1996). All individuals of a single species showed the same
patterns of variability for both indices and differences
were statistically insignificant. Therefore, for subsequent
interspecific comparisons, the mean values of each index
were used for each vertebra for each species. Thus,
the Wilcoxon signed ranks test was then used for
statistical comparison of the indices across species (Zar,
1996).
RESULTS
Using the data from the literature (Table 1), the relative tail
length (external tail length/head-body length) is not cor-
related with body weight (r=0.381, d.f. =9, NS) for the
carnivoran species examined. This means that prehensile-
tailed carnivorans do not possess the longest tails among
0
Mean vertebral length
2
4
6
8
10
12
14
16
Potos
Procyonids
Arctictis
Viverrids
Proximal Transitional Distal All
6.38
7.69 9.61
7.42
13.54
13.77
10.85
10.85
10.85
14.08
9.33 10.29
6.88
10.52
9.61
9.95
7.93 10.41
Fig. 2. Mean length of individual vertebrae divided by the cube
root of body weight for each caudal region and total tail length.
the arboreal and scansorial viverrids and procyonids
examined (Table 1). However, the viverrids examined
possess a higher number of caudal vertebrae than
procyonids, without necessarily possessing absolutely and
relatively longer tails (Tables 1 & 2).
The prehensile-tailed genera Arctictis and Potos bear
ahigher number of caudal vertebrae compared to other
representatives of the respective families. This difference
is more pronounced in procyonids (Table 2). In both
viverrids and procyonids, there are two sacral vertebrae
involved in the SIJ and there are no differences between
prehensile and non-prehensile taxa (Table 2). In Arctictis
and Potos,the transitional and longest vertebrae are
located more caudally than in the other taxa (Table 2).
Differences are more evident in viverrids than in
procyonids (Table 2). Among the latter, Nasua show
an intermediate position between Potos and the non-
prehensile-tailed forms (Table 2).
Potos and Arctictis possess longer proximal vertebrae
than their non-prehensile-tailed relatives, but differences
are not significant (Fig. 2, Table 5). On the other
hand, they possess shorter distal vertebrae than the
non-prehensile-tailed forms, but only Arctictis shows
significant differences (Fig. 2, Table 5).
Viverrids and procyonids do not present any significant
differences in relative vertebral number and length of
the proximal caudal region (Tables 3–5). By contrast,
viverrids possess a significantly shorter transitional region
in both relative vertebral number and length and a
significantly longer distal region in relative vertebral
number and length than procyonids (Tables 3–5).
Compared to other viverrids, the prehensile-tailed
Arctictis bear a significantly longer proximal region
in both relative vertebral number and length, and
significantly shorter transitional and distal caudal regions
(Tables 3–5).
Among procyonids, Potos bear the longest proximal
region and a significantly shorter transitional region in
relative length (Tables 4 & 5). By contrast, Nasua possess
the significantly shortest proximal region in relative
vertebral number and length within the family (Tables
3–5). Both taxa, differ significantly in relative proximal
and transitional caudal region length (Tables 4 & 5). Lastly,

Carnivoran prehensile tails 427
Tab l e 5 . Mann–Whitney U-tests between the prehensile-tailed and
non prehensile-tailed carnivorans for proportion of each caudal
region based on the number of vertebrae (CRNV), proportion of
each caudal region based on the sum of the lengths of the vertebrae
involved in each region (CRRL), mean length of the individual
vertebrae of each caudal region divided by cube root of body weight
(MLV)
Proximal caudal region CRNV CRRL MLV
Arctictis vs viverrids <0.05 <0.05 NS
Pot o s v sNasua NS <0.05 NS
Pot o svsprocyonids (-Nasua)<0.05 NS NS
Nasua vs procyonids (-Pot o s )<0.05 <0.05 NS
Viver rids vsprocyonids NS NS NS
Transitional caudal region
Arctictis vs viverrids <0.05 <0.05 NS
Pot o s v sNasua NS <0.05 NS
Pot o svsprocyonids (-Nasua)NS NS NS
Nasua vs procyonids (-Pot o s )NS NS NS
Viver rids vsprocyonids <0.01 <0.05 NS
Distal caudal region
Arctictis vs viverrids <0.05 <0.05 <0.05
Pot o s v sNasua <0.05 NS NS
Pot o svsprocyonids (-Nasua)<0.05 NS NS
Nasua vs procyonids (-Pot o s )NS NS NS
Viver rids vsprocyonids <0.001 <0.05 NS
Potos possess the significantly longest distal region in
relative vertebral number, but rank behind Bassariscus in
relative length (Tables 4 & 5).
Ve r tebral robusticity expressed as the ratio of mid-
width to length reveals significant differences between
prehensile-tailed and non-prehensile-tailed genera in both
families (Fig. 3a,b, Table 6). In general, the prehensile-
tailed taxa possess more robust vertebrae than the
non-prehensile-tailed forms (Table 6). The two prehensile-
tailed species do not differ significantly (Table 6).
Differences between the prehensile-tailed and non-
prehensile-tailed species are more pronounced caudad to
the transitional vertebra and become particularly evident
in the last distal vertebrae (Fig. 3a,b). In prehensile-tailed
forms, these vertebrae are highly robust maintaining a
constant mid-width on decreasing length.
Tab l e 6 . Wilcoxon signed ranks tests for robusticity index and relative expansion of transverse processes index between the prehensile-
tailed and non prehensile-tailed carnivorans. NS, not significant
Robusticity Relative expansion
ZSignificance ZSignificance
Pot o s v sNasua −4.28 <0.001 −3.04 <0.05
Pot o s v sProcyon −2.24 <0.05 −1.76 NS
Pot o s v sBassariscus −3.66 <0.001 −3.21 <0.001
Pot o s v sArctictis 1.05 NS −1.22 NS
Arctictis vs Nandinia 3.81 <0.001 4.58 <0.001
Arctictis vs Paradoxurus 3.70 <0.001 4.47 <0.001
Arctictis vs Cryptoprocta 3.16 <0.01 4.36 <0.001
Arctictis vs Genetta 3.64 <0.001 4.60 <0.001
0
Potos Nasua Procyon Bassariscus
20
40
60
80
100
120
Arctictis Paradoxurus Nandinia Genetta Cryptoprocta
1
Robusticity index
3579111315 17 19 21 23 25 27 29
0
20
100
10
30
40
50
60
70
80
90
(a)
(b)
1357911131517192123 25 27 293133
Caudal vertebrae
Fig. 3. Plot of robusticity index ([mid-width]/[length]∗100) on
number of caudal vertebrae: (a) procyonids; (b) viverrids.
The relative expansion of the proximal transverse
processes expressed as the ratio of the proximal width
to vertebral length also reveals differences between
prehensile-tailed and non-prehensile-tailed species
(Fig. 4a,b, Table 6). The prehensile-tailed taxa possess
more developed proximal transverse processes and the
differences increase in the distal vertebrae (Fig. 4a,b).
In prehensile-tailed forms, these vertebrae possess
well-developed processes while they are weak in the
non-prehensile-tailed taxa. Differences between the two
prehensile-tailed taxa are not significant (Table 6).

428 D. YOULATOS
0
Potos NasuaProcyon
20
40
60
80
100
120
1
Relative expansion index
35 79111315 17 19 21 23 25 27 29
(a)
Arctictis Paradoxurus Nandinia
0
20
40
60
80
100
120
1357911131517192123252729313335
(b)
Caudal vertebrae
Fig. 4. Plot of relative expansion of transverse process index
([proximal width]/[length]∗100) on number of caudal vertebrae:
(a) procyonids; (b) viverrids.
DISCUSSION
This study of the osteological characters in the caudal
region of carnivorans revealed certain features that group
the prehensile-tailed genera and separate them from their
non-prehensile-tailed relatives. The prehensile tails of
Potos and Arctictis are shown to be characterized by
the more distal location of the transitional and the
longest vertebrae, the longer relative length of the proxi-
mal caudal region, the shorter relative length of the
transitional region, and the more robust distal caudal
vertebrae bearing expanded transverse proximal pro-
cesses. Moreover, these features are absent from non-
prehensile-tailed carnivorans, providing further support to
assumptions of comparable tail use in the two prehensile-
tailed taxa. The way these morphological features can
be functionally associated with certain tail movements,
which have been reported in the overall tail use patterns of
prehensile-tailed carnivorans (Dor, 1937; Nowak, 1991;
McClearn, 1992) is discussed below.
In the proximal part of the caudal region, both
prehensile-tailed carnivorans possess more caudally
located transitional and longest vertebrae, a feature also
characteristic of prehensile-tailed primates (Dor, 1937;
Ankel, 1972; contra German, 1982). The more caudal
location of the transitional vertebra enlarges the proximal
caudal region. As proximal vertebrae are characterized by
lumbar-like intervertebral articulations, this is translated
into a greater number of this type of articulations
in the proximal caudal region. The lumbar type of
articulation amplifies intervertebral mobility enhancing
sagittal (flexion–extension) movements (Slijper, 1946).
Moreover, the increased number of articulating vertebrae
and their greater individual length, that contribute to an
increased relative length of that region, seem to further
enhance the flexibility of that region (Wainwright et al.,
1982; Ward, 1993; Shapiro, 1995). This morphology
contrasts the mechanically more stable relatively short
vertebral regions, which consist of few and relatively
short vertebrae. Assuming greater vertebral mobility
of the proximal part of the tail in prehensile-tailed
carnivorans, the equal development of the mass of caudal
flexors and caudal extensors in Arctictis may imply
equally powerful flexion and extension movements in
that part of the tail (Dor, 1937; Ankel, 1972; Lemelin,
1995). Increased flexion and extension of the proximal
part of the tail have been reported for both Arctictis
and Potos during slow climbing patterns or head-first
descents where the tail engages in controlled movements
in search of grasping points on the surrounding branches
(Nowak, 1991; McClearn, 1992). By contrast, the tail
of non-prehensile-tailed carnivorans seems to follow
passively body displacement and infrequently moves
back and forth to accommodate balance during climbing
(Kaufmann, 1962; Albignac, 1970; Taylor, 1970; Laborde,
1986; Nowak, 1991; McClearn, 1992; Rozhnov et al.,
1992). These behavioural differences in proximal tail
use between prehensile-tailed and non-prehensile-tailed
carnivorans may be related to the reported differences in
the osteological features of the proximal caudal region.
Despite these differences, the indices of robusticity
and relative expansion of the transverse processes of
the proximal vertebrae present no differences between
prehensile-tailed and non-prehensile-tailed carnivorans.
As both indices are associated with the action of gravity
and reaction forces as well as the forces exerted by the
muscles, tendons and ligaments in each vertebra, this may
imply similar loadings in the proximal caudal vertebrae
of both groups (German, 1982; Lemelin, 1995). Similar
findings are reported for prehensile-tailed primates
where the proximal caudal region failed to distinguish
between the two groups (German, 1982; Lemelin, 1995).
This was related to the notable dexterity of the proximal
tail of many non-prehensile-tailed primates (German,
1982; Lemelin, 1995; Meldrum, 1998). Apparently, this
wasfound in non-prehensile-tailed carnivorans, but more
field studies of tail use are needed to determine this.
The vertebral features of the distal caudal region
differentiate successfully between prehensile-tailed and
non-prehensile-tailed carnivorans (Table 5). The differ-
ences are mainly based on the higher values of the
indices of robusticity and relative expansion of transverse
processes and less on the shorter individual length of
distal vertebrae. On the other hand, there are practically
no differences in the relative number of vertebrae and
length of the distal caudal region. Potos and Arctictis
possess shorter and more robust vertebrae than their
non-prehensile-tailed relatives (Figs 2 & 3). Absolute

Carnivoran prehensile tails 429
mid-width of the vertebrae tends to be constant caudally
as the vertebrae shorten in length resulting in relatively
higher values and a peak in robusticity in the last distal
vertebrae (Fig. 3). The relative increase of width of the last
caudal vertebrae most likely indicates the action of higher
compressive and bending forces at these bony elements
(Wainwright et al., 1982). In addition, the reduction in the
length of the vertebrae seems to reduce the displacement
of one end of the region relative to the other, resulting
in the decrease of these bending moments about the
intervertebral discs (Ward, 1993). Potos and Arctictis
frequently use tail-assisted suspensory feeding postures
and anchor their tails while walking and clambering on
inclined and small branches (Nowak, 1991; McClearn,
1992; Julien-Laferri`
ere, 1993; pers. obs.). These grasping
patterns are likely to impose greater and more frequent
stresses along every distal caudal vertebra because of
the continuous action of gravity and the forces exerted
by the contraction of the well-developed caudal flexors,
which exceed the caudal extensors in mass (Dor, 1937;
Grand, 1977; German, 1982; Lemelin, 1995). In this
way, prehensile-tailed carnivorans seem to withstand the
action of these forces along the distal region of the tail.
Ve r tebral robusticity is also significantly higher in
the distal vertebrae of prehensile-tailed primates that
commonly engage in tail-only and tail-assisted suspensory
postural and locomotor patterns and impose their tails
in higher and more frequent stresses (German, 1982;
Lemelin, 1995; Meldrum, 1998).
The other feature that distinguishes well between
prehensile-tailed and non-prehensile-tailed carnivorans
is the relative expansion of the transverse processes
of mainly the distal caudal vertebrae (Fig. 4). Both
Potos and Arctictis possess significantly more expanded
transverse processes than the non-prehensile-tailed taxa.
Anatomical studies have shown that two muscles take
their multiple origins on the transverse processes: the
mm. intertransversarii, that contribute to the axial rotation
of the distal tail in respect with the more proximal part,
and the m. flexor caudae longus, that contributes to ventral
flexion of the tail (Dor, 1937; German, 1982; Lemelin,
1995). More expanded origins are usually associated
with bulkier muscles, as they provide larger surface for
muscular bundles (Hildebrand, 1995). Both muscles are
well developed and their tendons cross a few vertebrae in
Arctictis;mm. intertransversarii span two vertebrae and
the long caudal flexor spans three to four vertebrae
(Dor, 1937). Lemelin (1995), in his detailed work on
the prehensile tails of ateline monkeys, suggests that
if a tendon crosses fewer intervertebral joints, it is
likely to provide a greater degree of curvature for a
given cranial displacement. This results in an increased
flexion or rotatory ability of the tail, a function probably
also occurring in prehensile-tailed carnivorans. These
tail movements are frequent in tail-assisted suspensory
feeding or in repositioning of the body and forelimbs
during postural and locomotor activities (McClearn, 1992;
Julien-Laferri`
ere, 1993). Similar anatomical findings in
the prehensile-tailed primates have been also associated
with the frequent suspensory locomotor and postural
habits of ateline monkeys (Grand, 1977; German, 1982;
Lemelin, 1995).
This study of the caudal vertebrae of carnivorans, has
shown that the two distantly related prehensile-tailed
carnivorans, Arctictis and Potos,exhibit similar osteo-
logical features of the caudal vertebrae, which are lacking
from the other related carnivorans that do not use their
tails in a prehensile way. These features are functionally
associated with enhanced sagittal movements of the
proximal part of the tail, as well as increased flexing and
rotating abilities withstanding great and frequent stresses
in the distal part. Similar tail movements frequently
occur when the tail is used in grasping and suspensory
activities, but this is difficult to verify because of the
lack of behavioural data on the carnivorans studied. On
the other hand, the presence of similar morphological
features in the caudal vertebrae of the relatively well-
studied prehensile-tailed primates provides some support
but does not necessarily suggest similar tail use and
neuromotor control patterns (Lauder, 1998). This stresses
the need for more field data on the locomotor and postural
behaviour of scansorial and arboreal carnivorans, and
provides insight for the search of similar morphological
features in the caudal vertebrae of other scansorial and
arboreal mammals that are known to use their tails in a
prehensile manner.
Acknowledgements
Part of this research was funded by PARSYST. I thank Dr
M. Tranier, Mr R. Manceau, and J. Cuisin for access to the
collections of the Laboratoire des Mammif´
eres et Oiseaux,
Mus´
eum National d’Histoire Naturelle (MNHN). I am
also greatly indebted to Dr D. Robineau for access to
the collections of the Laboratoire d’Anatomie Compar´
ee
(MNHN). I am greatly indebted to Drs J.-P. Gasc,
F. Jouffroy and P. Lemelin and an unknown reviewer for
constructive remarks on the paper.
REFERENCES
Albignac, R. (1970). Notes ´
ethologiques sur quelques carnivores
malgaches: le Cryptoprocta ferox (Bennett). Te r r e Vie 24:395–
402.
Ankel, F. A. (1972). Vertebral morphology of fossil and
extant primates. In The functional and evolutionary biology
of primates:223–240. Tuttle, R. (Ed.). Chicago: Aldine-
Atherton.
Bininda-Emonds, O. R. P., Gittleman, J. L. & Purvis, A. (1999).
Building large trees by combining phylogenetic information: a
complete phylogeny of the extant Carnivora (Mammalia). Biol.
Rev. 74:143–175.
Cant, J. G. H. (1986). Locomotion and feeding postures of spider and
howling monkeys: field study and evolutionary interpretation.
Folia primatol. 46:1–14.
Cartmill, M. (1985). Climbing. In Functional vertebrate
morphology:73–88. Hildebrand, M., Bramble, D. M., Liem, K.
F.&Wake, D. B. (Eds). Cambridge: Belknap Press.
Charles-Dominique, P. (1978). Ecologie et vie sociale de
Nandinia binotata (Carnivores, Viverrid´
es): comparaison avec

430 D. YOULATOS
les prosimiens sympatriques du Gabon. Terre V i e 32:477–
528.
Dor, M. (1937). La morphologie de la queue des mammif`
eres. Paris:
Pierre Andr´
e.
Emmons, L. & Gentry, A. (1983). Tropical forest structure and the
distribution of gliding and prehensile-tailed vertebrates. Am. Nat.
121:513–524.
Falk, D. (1980). Comparative study of the endocranial casts of
New and Old World monkeys. In Evolutionary biology of New
Wo r l d m onkeys and continental drift: 275–292. Ciochon, R. L.
&Chiarelli, A. B. (Eds). New York: Plenum Press.
Garber, P. A. & Rehg, J. A. (1999). The ecological role of the
prehensile tail in white-faced capuchins (Cebus capucinus).Am.
J. Phys. Anthropol.110:325–339.
German, R. Z. (1982). The functional morphology of caudal
vertebrae in New World monkeys. Am. J. Phys. Anthropol.58:
453–459.
Grand, T. I. (1972). A mechanical interpretation of terminal branch
feeding. J. Ma mmal.53:198–201.
Grand, T. I. (1977). Body weight: Its relation to tissue composition,
segment distribution, and motor function. I. Interspecific
comparisons. Am. J. Phys. Anthropol.47:211–240.
Hildebrand, M. (1995). Analysis of vertebrate structure.NewYork:
Wiley.
Julien-Laferri`
ere, D. (1993). Radio-tracking observations on
ranging and foraging patterns by kinkajous (Pot o s flavus)in
French Guiana. J. t rop. Ecol.9:19–32.
Julien-Laferri`
ere, D. (1999). Foraging strategies and food
partitioning in the neotropical furgivorous mammals Caluromys
philander and Potos flavus. J. Zool. (Lond.) 247:71–80.
Kaufmann, J. H. (1962). Ecology and social behaviour of the coati,
Nasua narica,onBarroColoradoIsland, Panama. Univ. Calif.
Publ. Zool.60:95–222.
Kays, R. W. (1999). Food preferences of kinkajous (Potos flavus): a
frugivorous carnivore. J. Mammal.80:589–599.
Laborde, C. (1986). Description de la locomotion arboricole de
Cryproprocta ferox (Carnivore, viverrid´
emalgache). Mammalia
50:369–378.
Lauder, G. V. (1998). On the inference of function from structure.
In Functional morphology in vertebrate paleontology:1–18.
Thomason, J. J. (Ed.). Cambridge: Cambridge University
Press.
Lemelin, P. (1995). Comparative and functional myology of the
prehensile tail in New World Monkeys. J. Morphol.224:351–
368.
McClearn, D. (1992). Locomotion, posture, and feeding behavior
of kinkajous, coatis, and raccoons. J. Ma mmal.73:245–261.
Meldrum, D. J. (1998). Tail-assisted hind limb suspension as a
transitional behavior in the evolutionof the platyr rhine prehensile
tail. In Primate locomotion: recent advances: 145–156. Strasser,
E., Fleagle, J. G., Rosenberger, A. L. & McHenry, H. (Eds). New
Yo r k : P l e num Press.
Nowak, R. M. (1991). Wal ker’s mammals of the World.Baltimore:
Johns Hopkins Press.
Rosenberger, A. L. (1983). Tale of tails: parallelism and prehensility.
Am. J. Phys. Anthropol.60:103–107.
Rozhnov, V. V., Kuznetsov, G. V., Neklyudova, T. I., Fam Chong
An’ & Chan Van Dyk (1992). Ecological and ethological
observations of arboreal species of true civets in Vietnam. In
Zoological Investigations in Vietnam: 132–148. Sokolov, V. E.
(Ed.). Moscow: Nauka. (In Russian.)
Shapiro, L. (1995). Functional morphology of indrid lumbar
vertebrae. Am. J. Phys. Anthropol.98:323–342.
Slijper, E. (1946). Comparative biologic-anatomical investigations
on the vertebral column and spinal musculature of mammals.
Ve r h . K . N e d .Akad. Wet.42:1–128.
Taylor, M. E. (1970). Locomotion in some east African viverrids.
J. Ma mmal.51:42–51.
Trapp, G. R. (1972). Some anatomical and behavioral adaptations
of ringtails Bassariscus astutus. J. Mammal. 53:549–557.
Wai n w r ight, S. A., Biggs, W. D., Currey, J. D. & Gosline, J. M.
(1982). Mechanical design in organisms. Princeton: Princeton
University Press.
Ward, C. (1993). Torso morphology and locomotion in Proconsul
nyanzae. Am. J. Phys. Anthropol. 92:291–328.
Youlatos, D. (1999). Tail use in capuchin monkeys. Neotrop. Prim.
7:16–20.
Zar, J. H. (1996). Biostatistical analysis.London: Prentice-Hall.