Linking morphology, performance, and habitat utilization: adaptation across biologically relevant 'levels' in tamarins

Abstract

Background
Biological adaptation manifests itself at the interface of different biologically relevant ‘levels’, such as ecology, performance, and morphology. Integrated studies at this interface are scarce due to practical difficulties in study design. We present a multilevel analysis, in which we combine evidence from habitat utilization, leaping performance and limb bone morphology of four species of tamarins to elucidate correlations between these ‘levels’.

Results
We conducted studies of leaping behavior in the field and in a naturalistic park and found significant differences in support use and leaping performance. Leontocebus nigrifrons leaps primarily on vertical, inflexible supports, with vertical body postures, and covers greater leaping distances on average. In contrast, Saguinus midas and S. imperator use vertical and horizontal supports for leaping with a relatively similar frequency. S. mystax is similar to S. midas and S. imperator in the use of supports, but covers greater leaping distances on average, which are nevertheless shorter than those of L. nigrifrons .
We assumed these differences to be reflected in the locomotor morphology, too, and compared various morphological features of the long bones of the limbs. According to our performance and habitat utilization data, we expected the long bone morphology of L. nigrifrons to reflect the largest potential for joint torque generation and stress resistance, because we assume longer leaps on vertical supports to exert larger forces on the bones. For S. mystax , based on our performance data, we expected the potential for torque generation to be intermediate between L. nigrifrons and the other two Saguinus species. Surprisingly, we found S. midas and S. imperator having relatively more robust morphological structures as well as relatively larger muscle in-levers, and thus appearing better adapted to the stresses involved in leaping than the other two.

Conclusion
This study demonstrates the complex ways in which behavioral and morphological ‘levels’ map onto each other, cautioning against oversimplification of ecological profiles when using large interspecific eco-morphological studies to make adaptive evolutionary inferences.

Full text

Berlesetal. BMC Ecology and Evolution (2024) 24:22
https://doi.org/10.1186/s12862-023-02193-z
RESEARCH Open Access
© The Author(s) 2024. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which
permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the
original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or
other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line
to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory
regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this
licence, visit http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creativecom‑
mons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data.
BMC Ecology and Evolution
Linking morphology, performance,
andhabitat utilization: adaptation
acrossbiologically relevant ‘levels’ intamarins
Patricia Berles1* , Jan Wölfer1 , Fabio Alfieri1,2,3,4 , Léo Botton‑Divet1 , Jean‑Pascal Guéry5 and
John A. Nyakatura1
Abstract
Background Biological adaptation manifests itself at the interface of different biologically relevant ‘levels’, such
as ecology, performance, and morphology. Integrated studies at this interface are scarce due to practical difficulties
in study design. We present a multilevel analysis, in which we combine evidence from habitat utilization, leaping per‑
formance and limb bone morphology of four species of tamarins to elucidate correlations between these ‘levels’.
Results We conducted studies of leaping behavior in the field and in a naturalistic park and found significant differ‑
ences in support use and leaping performance. Leontocebus nigrifrons leaps primarily on vertical, inflexible supports,
with vertical body postures, and covers greater leaping distances on average. In contrast, Saguinus midas and S.
imperator use vertical and horizontal supports for leaping with a relatively similar frequency. S. mystax is similar to S.
midas and S. imperator in the use of supports, but covers greater leaping distances on average, which are nevertheless
shorter than those of L. nigrifrons.
We assumed these differences to be reflected in the locomotor morphology, too, and compared various
morphological features of the long bones of the limbs. According to our performance and habitat utilization data,
we expected the long bone morphology of L. nigrifrons to reflect the largest potential for joint torque generation
and stress resistance, because we assume longer leaps on vertical supports to exert larger forces on the bones.
For S. mystax, based on our performance data, we expected the potential for torque generation to be intermediate
between L. nigrifrons and the other two Saguinus species. Surprisingly, we found S. midas and S. imperator having
relatively more robust morphological structures as well as relatively larger muscle in‑levers, and thus appearing better
adapted to the stresses involved in leaping than the other two.
Conclusion This study demonstrates the complex ways in which behavioral and morphological ‘levels’ map
onto each other, cautioning against oversimplification of ecological profiles when using large interspecific eco‑mor‑
phological studies to make adaptive evolutionary inferences.
Keywords Biomechanics, Field study, Integrative biology, Leaping behavior, Limb bones, Locomotion
*Correspondence:
Patricia Berles
patricia.berles@hu‑berlin.de
Full list of author information is available at the end of the article

Page 2 of 20
Berlesetal. BMC Ecology and Evolution (2024) 24:22
Background
Each habitat confronts the animals living in it with func-
tional demands, which over time might ultimately drive
morphological adaptations. To gain insight into such eco-
morphological adaptations, integrated studies that include
analyses of both the organisms’ morphology and in-vivo
performance in controlled laboratory environments and
in the field are indispensable to determine the intricate
interplay between the biologically relevant ‘levels’ of mor-
phology, performance, and (behavioral) ecology [1–4]. e
musculoskeletal apparatus of vertebrates constitutes an
insightful study system for such research avenues because,
first, the locomotor behavior and performance are compa-
rably straightforward to measure in the field and the labo-
ratory, and then this information can be used to derive and
experimentally test biomechanical hypotheses of structure-
function relations on the reasoning of Newtonian mechan-
ics. Many studies thus far have focused on the relation
between structure and function in terms of the mechani-
cal properties of morphological characteristics (e.g., [5,
6]) and related these to coarse categories of the animals’
behavioral ecologies (e.g., [7–9]). Far fewer studies aspired
to investigate the fundamental link between structure on
the one hand and in-depth aspects of the behavioral profile,
the habitat utilization, as well as the animal performance in
the field on the other hand (e.g., [10–15]). is can be likely
related to the fact that such in-depth observations of ani-
mal behavior in the natural habitats are difficult to obtain
without disturbance of the animals and relies on well-habit-
uated study groups necessitating research infrastructure
such as field stations. Moreover, aiming to employ the valu-
able tools of the comparative approach, studying more than
one species, requires dealing with time-consuming inter-
specific data collection. We here present such an in-depth
case study on tamarins, a group of callitrichid primates
with apparently very conservative anatomy [16], focus-
ing on four closely-related species to interrelate variability
in habitat utilization, leaping behavior and anatomy of the
locomotor system.
Tamarins (belonging to Callitrichidae, within platyr-
rhine primates), as all other arboreal mammals, are faced
with specific functional demands for the locomotor sys-
tem due to discontinuous, narrow, and flexible supports,
which these animals need to be able to navigate to bridge
gaps and reach food sources such as fruits, flowers, and
invertebrates (see recent review [17]). For example, the
movement on thin and flexible (i.e., precipitously bend-
ing) terminal branches, where usually the fruit and flowers
are located, represents a challenge for the required bal-
ance [17–21]. Accordingly, support diameter and support
flexibility should be considered as environmental vari-
ables that exert significant selective pressure on the loco-
motor apparatus (e.g., to stabilize the shoulder and elbow
on flexible supports [16]) of such arboreal mammals [22,
23]. Similarly, the ability to leap from one support to the
next has often been interpreted as an adaptation to an
arboreal lifestyle (e.g., [18]). In this context, also lianas,
i.e., thin and flexible, vertical supports, were discussed for
their potentially crucial role in the evolution of locomotor
adaptations in primates [24].
Several tamarin species live sympatrically in the Ama-
zon basin [25]. By forming “mixed-species groups’’,
some tamarins are even syntopic, traveling and foraging
together albeit in different forest layers [26–29]. Because
of species-specific preferences for foraging and traveling
in different microhabitats, differential leaping behavior
has been documented among callitrichid species, too
(e.g., [30–35]). is is also the case for two of the focal
species of our study. e first, Saguinus mystax, travels
primarily in the upper layers of the forest (79% of the
time [26]) and uses mostly horizontal supports thinner
than 10 cm during locomotion [34], whereas the second,
Leontocebus nigrifrons, primarily uses vertical supports
of larger diameter in the lower forest layers (87% of the
time) [26, 34–38]. For both species studied by Berles and
colleagues [26], it was shown that there was a preference
for one leaping type regardless of the available supports in
the different forest layers. L. nigrifrons mainly leaps from
trunk-to-trunk, a leaping style that is observed in strep-
sirrhine primates too, although with some differences
[39] (i.e. tamarins exhibit pauses for clinging and scan-
ning the environment and land forelimbs first, whereas
strepsirrhines show a rapid sequence of leaps, landing
with the hindlimbs first; [32, 40, 41]). On the other hand,
S. mystax primarily performs horizontal leaps between
terminal branches [26]. e third species of this study, S.
midas, lives primarily in the lower to upper forest layer
and moves along medium-sized supports (2-10cm) [42–
45], suggesting similar behavior to S. mystax. In contrast,
S. imperator (the fourth focal species of our study), simi-
lar to L. nigrifrons, moves mostly in the lower forest lay-
ers but, unlike L. nigrifrons, predominantly uses smaller
oblique supports for locomotion [37, 46].
In the upper forest layers, crossing gaps usually
involves leaps out of balancing quadrupedal movement
on terminal branches [35, 41]. Due to the lack of hori-
zontal supports in the lower forest layer, primates must
use trunk-to-trunk leaps from a vertical clinging posi-
tion to cross gaps between supports (Fig.1) [47, 48]. Four
types of leaps can be distinguished in tamarins. e first
type represents long acrobatic downward leaps in hori-
zontal body posture, in which tamarins cross horizontal
distances of more than 5m in the upper canopy. Typi-
cally, such leaps begin and end on thin, flexible, termi-
nal branches [31, 33, 40, 41]. Sometimes several small
twigs are grasped by one hand or foot at the same time.

Page 3 of 20
Berlesetal. BMC Ecology and Evolution (2024) 24:22
A second type of leaps in horizontal body posture is
known as a bounding leap. is short leap involves a
quadrupedal locomotor sequence with longer airborne
phase on oblique and horizontal branches and depends
on powerful limb extension to achieve the similar heights
during take-off and landing [33, 40]. A third type of leap-
ing in tamarins is a stationary leap, in which animals hold
a stable posture on inflexible support of various angles of
Fig. 1 Schematic illustration of data acquisition. A Characterization of habitat utilization. Habitat characteristics and posture were recorded
in the field/park and are shown in green (diameter, D; orientation, O; flexibility, F). B Leaping performance and posture. Horizontal leaping distance
(LD) was recorded in the field/park, time‑related performance measures were extracted from camera recordings and an exemplary trunk‑to‑trunk
leap of L. nigrifrons with a leaping distance of 1.5 m is highlighted in violet (take‑off‑, flight‑, and landing‑phases in ms). C Acquired morphological
variables. Measurements on the bones are highlighted in orange. Anterior view of the left humerus, ulna, radius, femur, and tibia of the specimen S.
mystax AMNH 188,178. The numbers in panel C refer to the labelled data points in Fig. 2C. See text for more information

Page 4 of 20
Berlesetal. BMC Ecology and Evolution (2024) 24:22
inclination before leaping [41]. As a fourth type, trunk-
to-trunk leaps (Fig. S1) are specific in that the animals
always start and land in a fully crouched and static ver-
tical head-up posture on vertical supports [41]. ese
trunk-to-trunk leaps require several mechanical adap-
tations of the limbs. In addition to generating sufficient
impulse during take-off through a powerful extension of
the hindlimbs to bridge the horizontal distance, the body
must be rotated during the flight phase [49] and the fore-
limbs, which are usually extended during the flight, must
be able to withstand the compressive forces during land-
ing [40, 41, 50, 51].
Previous studies on differences in morphology in
tamarins, indicate a strong relationship between spe-
cies’ morphological variation and already observed
ecological variability within mixed-species groups
[52–54]. For example, L. nigrifrons, which prefers trunk-
to-trunk leaps, shows greater individual variation in
morphology than S. mystax, which is related to greater
variability in postural behavior found in field studies
[54]. Also, L. nigrifrons exhibits some osteological fea-
tures in the knee (e.g. a longer patellar groove, shorter
articular facet on the patella and longer femoral con-
dyles) that may be related to trunk-to-trunk leaps, as
suggested by Garber and Davis [54], due to the strong
flexion followed by extension during take-off. e fore-
limbs appear more gracile, e.g., due to a narrowing of
the humeral bi-epiphyseal width, medial epicondylar
width, and anterior and posterior trochlear width [55].
In a recent clade-wide comparative study by Botton-
Divet and Nyakatura [16], the authors analyzed a variety
of limb long bones and found some notable anatomical
differences between the trunk-to-trunk leapers and the
horizontal leapers within the callitrichids. For example,
the hindlimbs of the trunk-to-trunk leapers were shown
to have a proportionally smaller femoral head and a
larger lesser trochanter, whereas the horizontal leapers
have a more expanded trochlea at the humerus, which
could provide greater stability of the elbow [16]. How-
ever, the study by Botton-Divet and Nyakatura [16] does
not take into account the different ecological context,
like support use and leaping performance, faced by the
species under study.
Taking all of this into account, the diversity of habitat
characteristics in these arboreal mammals allows for a
detailed correlation between habitat utilization, perfor-
mance, and morphology. Since the monophyletic taxon
of tamarins is only ~ 14ma old [56] and since tamarins
have very similar body sizes [36], morphological differ-
ences are likely reflective of adaptations to the specific
functional demands resulting from their differing ecol-
ogy [16]. In this study, we first provide information on
behavioral data of leaping of four species of tamarins. We
quantify habitat utilization in terms of support dimen-
sions, orientation, and flexibility. Leaping performance
was characterized in terms of leap distance and dura-
tion of sub-phases of leaps, also accounting for postural
differences. While the leaping performance influences
the magnitude of exerted forces, posture during take-off
and landing determines the predominant direction of
the involved forces. We here define performance from a
strictly biomechanical viewpoint as an observed locomo-
tor trait [15] and thus describe a habitual load caused by
leaping, and not, as often used in the literature, the maxi-
mum performance of an individual, which occurs rather
rarely in the natural habitat [57]. is is based on the
experimental observation that the habitual load causes a
dynamic change of bone structure [58–60]. Since habitual
loads are difficult to quantify in the field, we rely on kine-
matic parameters which reflect the forces involved during
leaping to characterize these loads. As proxies, we use the
leaping distance and the temporal subphases of the leaps.
e duration of the take-off subphase was demonstrated
to be correlated to leaping performance, for example, in
mouse lemurs [61]. By determining the duration of the
take-off subphase we gain insight into the time available
for leg extension, and thus, the impulse generated [62].
Similarly, the landing subphase informs on the duration
of leg flexion. e leaping distance should be correlated
to the peak support reaction forces exerted during leap-
ing [63] while the relation between leaping distance and
the duration of the flight subphase provides an idea of the
jumping velocity. In addition, we measure various inter-
nal and external osteological features of the long bones
from museum collection specimens of the four stud-
ied species. External features include muscle in-levers,
robustness features, and limb proportions (Fig.1C). Dia-
physeal and epiphyseal internal structure were measured
since they have been shown to reflect eco-morphological
adaptation to locomotor biomechanical loadings in pri-
mates (e.g., [64–68]). e aim is to first find patterns in
both habitat utilization and leaping performance that
reflect distinct preferences of the tamarin species.
We have the following predictions for the level of
habitat utilization:
1. S. midas, similar to what was already shown for
S. mystax [26], predominantly uses horizontally
oriented supports with a small diameter, due to the
average height of stay in higher forest layers and the
resulting availability of supports [69].
2. S. imperator, independent of the preferred lower
forest layer [37], predominantly uses horizontal and
oblique supports with small diameter [46].
3. L. nigrifrons is a specialist in leaping on large, vertical
supports [26].

Page 5 of 20
Berlesetal. BMC Ecology and Evolution (2024) 24:22
Based on this, we have the following predictions for
the level of body posture and leaping performance:
4. e Saguinus species predominantly perform leaps
in a horizontal body posture, while L. nigrifrons
is a specialist for leaps in a vertical body posture
independent of the available support type.
5. Vertical leaps cover a greater distance, which is also
reflected in a longer duration of the flight subphase.
6. Greater leaps in general and horizontal leaps in
particular [70] have longer durations of take-off and
landing subphases.
Once species categorization is possible on these two
behavioral ‘levels’, i.e., habitat utilization and leaping
performance, we can hypothesize different functionally
relevant demands on their locomotor morphology and
test these on morphological data.
Based on the available literature, we have the
following predictions for the level of locomotor
morphology:
7. L. nigrifrons shows the strongest bone robustness
as a result of the higher compressive stress on the
forelimbs and hindlimbs during landing on large
inflexible supports and relatively longer hindlimbs
that benefit long leaps like in frogs [63] and galagos
[71] instead of constant quadrupedal locomotion.
8. e Saguinus species have greater stability in the
shoulder and hip joint for good balance on the
flexible supports in addition to lower bone robustness
[16] since horizontal leaps are mostly performed on
flexible supports and the compressive stress during
landing is lower.
9. Species with less variable habitat utilization display a
smaller degree of trabecular anisotropy.
We believe that this study will not only contribute to
a better understanding of the evolution and locomotor
adaptations of tamarins, but also to a more general
understanding of how eco-morphological adaptation
manifests on different biological ‘levels’.
Materials & methods
We collected data on three biologically relevant ‘lev-
els’: (i) habitat utilization, (ii) leaping performance and
posture, and (iii) morphology of the locomotor appa-
ratus (Fig.1). Data for habitat utilization and leaping
performance were jointly collected in the wild and in
a naturalistic park with different primates whereas data
characterizing the locomotor morphology were col-
lected from skeletal material of 12 museum collection
specimens (three per species).
Study sites andstudy groups
For this study, it was important to select groups of
individuals whose natural behavior can be observed
without disturbance by the presence of the observers.
For this reason, we chose the following study sites. First,
a field study was conducted during the dry season from
June to October 2017 in the Amazon lowlands of north-
eastern Peru at the “Estación Biológica Quebrada Blanco”
(EBQB). e mean temperature during this period was
26.9°C and the mean monthly precipitation was 175mm/
m² (Tamshiyacu, Peru; data from https:// www. accuw
eather. com/ de/ pe/ tamsh iyacu). e station is located at
4°21’S and 73°09’W, on the right bank of the Quebrada
Blanco, a small tributary of the Rio Tahuayo that empties
into the Amazon (more detail and a figure showing the
location of the station can be found in Berles etal. [26];
Heymann and Tirado Herrera [72]). e primary forest
at EBQB consists predominantly of a dense canopy at a
height of 25-30m [73]. A mixed species group consisting
of six adults and two juvenile individuals of S. mystax
and three adult individuals of L. nigrifrons was studied.
But we collected data on adult individuals only. e sex
of the animals was neglected in this study. e tamarins
of EBQB are well habituated to the presence of human
observers. e activity data of the mixed-species group
were recorded for a total of 68 days in the field. In total, S.
mystax was observed for 586h and L. nigrifrons for 519h
(mostly parallel to S. mystax).
A second behavioral study was conducted in October
2018 in Romagne, France, in a park hosting different
primate species (“La vallée des singes” hereafter referred
to as park study in contrast to the field study at the
EBQB). e mean temperature during this period was
14°C and the mean monthly precipitation was 49.6mm/
m² (Poitiers, France; data from https:// www. wette rkont
or. de/ de/ wetter/ europa/ extre mwerte- frank reich. asp).
e park consists of artificial islands overgrown with
large trees and bushes with ropes between the trees
to provide additional connections. e tamarins have
the option to reside in temperature-controlled houses
during the nights and winter months. e islands are
separated by wide moats to prevent the monkeys from
escaping. On these islands, visitors can explore the
habitat the monkeys live in, so the tamarins are very used
to observers. Here, the subjects of the study were three
individuals of S. midas on one island and five S. imperator
on another island. e activity data of the tamarins in the
park were recorded over 14 days. S. midas was observed
for a total of 21.5h and S. imperator for a total of 23.5h.
Each data set was collected by the first author. Altogether,
5920 leaps were observed, 2347 leaps for L. nigrifrons,
2730 leaps for S. mystax, 359 leaps for S. imperator and
484 leaps for S. midas.

Page 6 of 20
Berlesetal. BMC Ecology and Evolution (2024) 24:22
Determination ofhabitat characteristics
e support properties of the home range of the study-
site in Peru have been described in Berles etal. [26] and
will be used here. In order to characterize the structure of
the forest of the two islands in “La vallée des singes’’ and
to determine the spectrum of available supports [74], one
Whittaker plot for the island of each species was created
(Table S1). e procedure was done identically at both
study sites.
Behavioral data acquisition
In the field study, the horizontal distance covered dur-
ing leaping, the flexibility, orientation and diameter of
the support as well as the posture of the animal at take-
off and landing were visually estimated and recorded
into a protocol (Fig.1 in green, raw data can be found in
Fig.2). For this, every 30min, the first visible individual
was followed and all behavioral data were noted as long
as it remained visible. For each leap that was observed, all
characteristics were noted [26]. In addition, during the
rest of the time, each visible leap of all individuals, with
all associated parameters, was noted. In the park study,
all support properties with a support height ≤ 2 m were
measured and otherwise estimated as in the field study.
Since visual estimation is essential for our study, we had
all data recorded by only one trained observer, in our
case the first author, as suggested by Bezanson and col-
leagues [75]. e first author in our study trained the
visual estimation of support inclination, tree height, and
distances previously both in Germany, in the Steiger-
wald, a full-leaved foliage forest in central Germany, and
in a one-month pilot study directly in the rainforest [26].
In addition, the leaping behavior of the four species was
recorded with a camcorder (Panasonic VW-ACT190,
50fps) in order to evaluate the durations of sub-phases
of the individual leaps later in the lab (see below and vio-
let boxes in Fig.1, raw data see Fig.2). ese videos were
acquired for all leaps observed in the park study, but only
for a subset of all leaps observed in the field, since it was
logistically impossible to film all the documented leaps at
the same time. e field/park study notes were then cor-
related with the videos using the time code of the videos
and a written protocol. Using the camera recordings, each
leap was divided into a take-off phase, a flight phase, and
a landing phase (violet boxes Fig.1). e take-off phase
starts with the first observed loss of contact of any one of
the four limbs and ends with the observed loss of contact
of the last of the four limbs. is is followed by the actual
flight phase. e flight phase ends with the first contact of
any limb with the new support. e subsequent landing
phase ends with the last of the four limbs contacting the
new support. Using high-fps video recordings we meas-
ured the duration in milliseconds of each phase. e time
resolution was 20 ms.
In summary, variables characterizing the ‘level’ of
habitat exploitation were support orientation, support
diameter, and flexibility during take-offs and landings,
respectively. e habitat variables were of categori-
cal nature. We used three orientation categories (0–20°,
30–60°, 70–90°), four diameter categories (< 2cm, <5cm,
< 10cm, ≥ 10cm), and two flexibility categories (yes, no).
e ‘level’ of leaping performance was characterized by
the horizontal distance of leaps as well as the posture
during take-off and landing, respectively, and the dura-
tion of take-off-, flight-, and landing phases. e perfor-
mance variables were of mixed categorical (posture) and
continuous (distances and durations) nature. Posture (i.e.,
the body position during leaping) included two categories
that were noted regardless of the support used: typical
trunk-to-trunk leaping posture (i.e., monkeys clinging to
the supports in a vertical position) and typical horizon-
tal leaping posture (i.e., monkeys were in a pronograde,
quadrupedal body position) (Fig.1).
Bone data acquisition
Bones were obtained from the Field Museum of Chicago
(FMNH) and the American Museum of Natural History
in New York (AMNH). e specimens from the FMNH
Fig. 2 Data split by species. A Habitat utilization, (B) Leaping performance, (C) Morphology (standardized variables, compare to Fig. 1C). Humerus:
(1) surface area of the scapular articulation, (2) surface area of the radial and ulnar articulation, (3) in‑lever of the M. deltoideus, (4) in‑lever of the M.
subscapularis, (5) in‑lever of the M. supraspinatus and M. infraspinatus, (6) cross‑sectional‑area (CSA) at 50% length, (7) anteroposterior second
moment of area (SMAap) at 50%, (8) mediolateral second moment of area (SMAml) at 50% length, (9) trabecular degree of anisotropy (DA), (10)
trabecular bone volume fraction (BV.TV); Radius: (11) Surface area of the humeral articulation, (12) Surface area of the carpal articulation; Ulna: (13)
surface area of the humeral articulation, (14) surface area of the radial articulation, (15) in‑lever of the M. triceps brachii; Femur: (16) surface area
of the pelvic articulation, (17) Cross‑section of the femoral neck, (18) patellar height index (patellar width projected onto surface/patellar width),
(19) in‑lever of the M. gluteus medius, (20) in‑lever of the M. gluteus superficialis, (21) in‑lever of the M. iliopsoas, (22) CSA at 50% length, (23)
SMAap at 50% length, (24) SMAml at 50% length, (25) DA, (26) BV.TV; Tibia: (27) surface area of the femoral articulation, (28) surface area of the talar
articulation. Percentages of utilized categories of the studied support characteristics and posture can be found in numerical form in supporting
information Table S2
(See figure on next page.)

Page 7 of 20
Berlesetal. BMC Ecology and Evolution (2024) 24:22
Fig. 2 (See legend on previous page.)

Page 8 of 20
Berlesetal. BMC Ecology and Evolution (2024) 24:22
were CT-scanned at the PaleoCT Scanner Facility of the
University of Chicago using a GE phoenix V|tome|x.
Specimens from the AMNH were CT-scanned at the
Shared Materials Instrumentation Facility (SMIF) of
Duke University. e resolution of the scans varied from
15.5 to 18μm. As all bones of a single specimen were
scanned as a batch, single bones were then cropped using
Fiji [76] and CT scans were segmented using Amira 6.0.0
(ermo Fisher Scientific). All meshes are available on
demand from Morphosource (see specimen details in
Table S3). Importantly, the epiphyses of two humeri and
one femur were not completely fused, meaning that the
specimens were likely subadult (humeri: S. imperator
FMNH 98,035 and S. midas FMNH 93,236 femur: S.
midas FMNH 93,236).
Quantication ofthebone internal structure
Diaphyseal and epiphyseal internal structure (cross-
sectional properties [CSP] and trabecular architectural
properties, respectively) were quantified in the humeri
and femora following part of the procedure in Alfieri
etal. [77] (Fig.1 in orange, for raw data see Fig.2). e
two bones were oriented in anatomical standard position
following Ruff [66] using VG Studio Max 3.3 (Volume
Graphics, Heidelberg, Germany). Oriented bones,
exported as image TIFF stacks, were then imported
to Fiji for internal structure quantification. Regarding
diaphyseal properties, we quantified the cross-sectional
area as well as the anteroposterior and mediolateral
second moment of area at 50% bone length. Regarding
epiphyseal properties, we included the trabecular degree
of anisotropy (DA) and the trabecular bone volume
fraction (BV.TV). Further details on CSP and trabecular
properties and the procedures employed to quantify
them are included in Supporting information note 1.
Quantication ofexternal bone morphology
e segmented CT scans were exported as a surface mesh
from Amira. Measurements of the outer morphology of
the humerus, ulna, radius, femur, and tibia were obtained
on these meshes in Geomagic Wrap (3D Systems 2017).
We measured the effective length of the humerus, radius,
femur, and tibia to calculate the intermembral index (IMI:
(humeral length + radial length) / (femoral length + tibial
length)). A low IMI indicates relatively longer hindlimbs
and hence, a specialization for vertical leaping [78]. We
further measured robustness variables that inform on the
potential to resist stresses as well as the length of muscle
in-levers that inform on the potential to generate and
absorb joint torques. A detailed description and depiction
of the measurements are provided in the supporting
information (Figs. S2, S3, S4, S5, S6) and a brief overview
of all measurements is displayed in Fig.1. All internal and
external morphological variables were continuous and
we size-corrected them prior to further analysis using the
centroid sizes of the humerus and femur (see Supporting
information note 2 for details).
Missing data imputation andsampling bias correction
All subsequent analyses were conducted in R Version
4.2.2 [79]. Measurement values were missing for the
performance variables due to take-off and landing events
being out of sight during recording. In total, 8.5% of the
recorded leaps had at least one missing subphase. e
take-off was missing in 1.7% of the recordings, the flight
subphase in 7.2% and the landing subphase in 6.7%.
Also, data were missing in the trabecular variables due
to unfused epiphyses in two humeri and one femur. e
missing data were imputed for both data sets separately
using the R package missMDA [80]. See Supporting
information note 3 for more details. We imputed the
data after sampling bias correction (see below) because
the frequency of missing data was linked to the support
use and height of the leaps. Specifically, leaps high up in
the canopy had missing data and removing them before
bias correction would have increased the existing bias
towards including disproportionately more vertical leaps
from the lower forest layers. Hence, we decided to rather
rely on imputed data for these leaps instead of increasing
the disbalance of the recorded leaps even more.
Sampling bias correction fortheperformance data
As most of the video footage used to capture durations
could only be recorded in the lower layer of the forest,
the ratio of horizontal to vertical leaps was likely biased
toward vertical leaps. is was particularly likely for the
two species that were filmed in the wild where trees were
higher (Table S1). us, the data set had to be adjusted
to correct for this bias. is procedure resulted in a
selection of 1092 of the 1271 observations (Table S4C).
See Supporting information note 4 for more details. A
summary of the number of observations per species and
per ‘level’ after each step (data collection, missing data
imputation, and sample bias correction) can be found in
Table S5.
Cluster analysis anddimensionality reduction
e R packages FactoMineR [81] and clValid [82] were
used for data ordination, cluster analysis, and cluster vali-
dation/assessment, respectively. As mentioned above, we
first used cluster analysis on the ‘levels’ of habitat utili-
zation and leaping performance separately to explore
whether the four studied species can be categorized into
groups that reflect a dichotomy between trunk-to-trunk-
leaping and horizontal leaping. Each cluster analysis was

Page 9 of 20
Berlesetal. BMC Ecology and Evolution (2024) 24:22
preceded by an ordination analysis, but all ordination
dimensions were used for clustering. e ordination was
necessary for the datasets concerning the habitat utiliza-
tion and leaping performance to transform the categori-
cal variables into pseudo-continuous ordination scores
from which distance metrics could be computed. We
also used the first two ordination dimensions to visual-
ize the mapping of clusters between the different ‘levels’,
justifying also the ordination of morphological variables.
For habitat utilization, which consisted of categorical
variables only, we used multiple correspondence analy-
sis (MCA; MCA function from FactoMineR), which is
equivalent to principal component analysis (PCA) for
quantitative data. For simplicity, we refer below to prin-
cipal components (PCs) to the dimensions of all dimen-
sionality reduction techniques. In MCA, variables are
split into their categories which are in turn transformed
into continuous variables, giving larger weight in the gen-
eration of PCs to categories shared by fewer individuals.
On the adjusted data set of the performance variables, we
performed a factor analysis of mixed data (FAMD; FAMD
function from FactoMineR). is analysis includes both
categorical and continuous variables. On the morpho-
logical variables, we applied a PCA (PCA function from
FactoMineR).
Hierarchical clustering was conducted each time on
the PCs using the HCPC function from FactoMineR. e
algorithm operates by assigning each observation to a
cluster and merging them to higher ‘level’ clusters step by
step by combining the two clusters at a time that share
the closest Euclidean distance in the multidimensional
space spanned by the PCs. e WARD method was
used to avoid chain effects that can complicate cluster
generation in the presence of coalescing clusters. e
clusters were then selected by cutting the cluster tree and
a k-means consolidation was used post hoc to consolidate
clustering. We decided upon the most meaningful
clustering primarily by using inertia gain (i.e., how much
variance can be additionally explained by adding another
cluster) as provided by the HCPC function. Different
cluster determination criteria exist and, hence, the
resulting interpretations might depend on the favored
criterion. Although we decided to determine the number
of clusters via the criterion of inertia gain, we additionally
assessed the reliability of different numbers of clusters
using three more indices. ese are called connectivity
[83], Dunn index [84], and silhouette width [85]. e
measure of connectivity indicates how well similar
observations are clustered together as determined by the
‘k-nearest neighbors’ method. It can range from zero to
infinity and should be minimized for good clustering.
e Dunn index is a ratio of the smallest distance
between the clusters to the largest intra-cluster distance.
It falls between zero and positive infinity and the larger
the value, the better the cluster discrimination. e
silhouette width is the average of the silhouette values
among observations. e silhouette value measures the
degree of confidence that an observation is assigned to
a particular cluster. e value falls between − 1 and + 1
with poorly clustered observations being close to -1 and
well-clustered observations being close to + 1.
We cut the tree into two to six clusters for these
analyses, each time computing the measures of clustering
quality using the package clValid [82]. Two to six
clusters were chosen since two clusters were expected
regarding the “trunk-to-trunk leaper vs. horizontal
leaper” dichotomy, three clusters seemed plausible
with an additional intermediate or more generalist
locomotor behavior, four clusters could represent the
four species, and five to six clusters would indicate that
no interpretable clustering could be achieved. Testing a
larger number of clusters did not appear insightful to us.
After selecting a clustering for each level, we used the
five data points closest to the respective cluster centroid
to characterize habitat utilization and performance
clusters. To ensure an accessible overview of the
characteristics of the morphological clusters, we created
bar plots of differences between the mean values of
clusters instead. We calculated the relative frequency
of each species falling into the respective cluster and
used a χ²-test to evaluate whether there is a significant
association between the clusters and the species within
each ‘level’. We further generated graphs with the first
two principal components of each ‘level’ and visualized
the loadings of the original variables onto these to
illustrate and explore clustering trends. However, care
must be taken in interpretation in case the loadings of
the original variables onto these 2D subspaces are low,
which would indicate a poor representation of these
variables in these two dimensions. e loading between
a continuous variable and a PC was defined as their
Pearson’s correlation coefficient and the loading between
a categorical variable and a PC was defined as the R²
obtained from an ANOVA analysis.
Inferential statistics
Cluster analysis and dimensionality reduction provide
insight into the major trends between species within
each biological level that we studied. However, we
also plotted the raw data of each level to find patterns
which might not be captured by these analyses. For
this purpose, we used inferential statistics to compare
differences between the four species in each trait of the
levels of habitat utilization and leaping performance

Page 10 of 20
Berlesetal. BMC Ecology and Evolution (2024) 24:22
and posture. We refrained from conducting inferential
statistics with the morphological traits because of
the small intraspecific sample sizes (see Supporting
information note 5 for more details).
Specic challenges oflinking dierent ’levels’
For the habitat characterization it would be optimal to
have a detailed overview about all available supports
within the home range of the monkeys. Unfortunately,
it was not possible to determine exact proportions of
available supports foreach leap. Nevertheless, based on
the similar trends observed in other studies, we believe
that our estimates and measurements reflect a good
pattern of support use during leaping by the four species.
In theory, it would be ideal to have more than four spe-
cies from which the three levels are extracted. With four
species, we only have four degrees of freedom at maxi-
mum for a statistical analysis of adaptation. Additionally,
it would be optimal to study at least one species in both
field and park, to provide a valuable means for validating
and comparing methods and observations. Unfortunately,
this was not possible, due to the rather rare occurrence of
the species in naturalistic parks.
Also, the individuals studied in the field are not the
same from which we obtained the morphological data,
limiting us to the comparison of species’ means. In the
optimal case, data from all the ‘levels’ here analyzed
would be obtained from the same individual. However,
no animals were sacrificed for this study. ese limita-
tions prevented us from conducting any reliable inferen-
tial statistical analysis across the levels. us, we decided
to keep this study within an exploratory framework using
mostly descriptive statistical methods.
Results
Habitat utilization
e inertia gain criterion suggested two clusters of
support use for the leaps of all four species (Fig. 3A,
Fig. S7). On the contrary, the measures used for cluster
validation suggest the selection of six instead of two
clusters (Table S6), indicating that the two clusters were
poorly separated. However, only the connectivity was
strongly improved (187.73 for two clusters and 29.3
for six clusters), whereas the Dunn index (0.22 for two
clusters and 0.28 for six clusters) and the silhouette width
(0.21 for two clusters and 0.26 for six clusters) improved
marginally. We thus decided to use the two clusters
from the inertia gain criterion for further interpretation,
because they reflect a weak, but meaningful trend in the
support use variability across species.
According to the five most typical leaps, cluster 1 was
characterized by vertical (70–90°), thick (> 10 cm), and
inflexible supports, for both take-off and landing, which
can be considered trunk-like supports. Cluster 2 on
the contrary was characterized by horizontal (0–20°),
thin (< 5 cm), and flexible supports (Table S7), which
can be considered canopy-like supports. We found a
significant association between clusters and species
(χ²(df = 3) = 357.77, p < 0.001) that is majorly driven by
L. nigrifrons (Fig.3A). In particular, 81% of the observed
leaps of L. nigrifrons fell into cluster 1, whereas S.
mystax, S. midas and S. imperator were very similar
with 54–60% of leaps falling into cluster 1. Accordingly,
L. nigrifrons can be considered rather specialized in the
exploitation of trunk-like supports, whereas the three
Saguinus species can be characterized as more generalist
(Fig.3A, D). e three Saguinus species, despite being
all generalist, differed significantly in most of the
support characteristics (Table S8), which is not captured
by the cluster analysis. In particular, S. midas used
flexible supports and supports with 0–20° orientation
more frequently than the other two species (Fig. 2). S.
imperator, on the other hand, used the support < 2 cm
more and those ≥ 10cm less frequently compared to the
other two Saguinus species. S. mystax, at last, stood out
in using flexible supports more frequently than the other
two species (Fig.2).
e described cluster trends are well-captured by
the first two principal components, despite them only
explaining 34% of the variance (Fig. 3A, Fig. S8). PC1
separates the two clusters and all variables load strongest
on PC1 compared to PC2 (support diameters are repre-
sented the best and support orientations the worst). Also,
the centroid of L. nigrifrons is positioned on the far left of
cluster 1 and the centroids of other three species closely
together near the intersection of both clusters. is in
(See figure on next page.)
Fig. 3 Principal component graphs of biologically relevant ‘levels’. The planes spanned by the first two principal components (PCs) from the habitat
utilization (A), leaping performance and posture (B) and morphology (C) datasets are shown with grey symbols representing the mean values
of the species and black symbols representing the mean values of the categorical variables. The grey lines connect the species means to highlight
cross‑level differences. Variable loadings of the corresponding PC analysis (D, E, F) are found right to the PC graph. Loadings of continuous variables
represent Pearson’s correlation coefficients and can range from − 1 to + 1. The closer a point falls to the circle’s margin the better is its variable’s
representation in this 2D subspace (falling on the margin indicates total representation). Loadings of categorical variables (indicated by asterisks)
represent R² values from an ANOVA and can range from 0 to + 1. The point labels in panel F correspond to the numbered morphological variables
in Fig. 1C (the point for IMI referring to the effective length measurements used to compute the index)

Page 11 of 20
Berlesetal. BMC Ecology and Evolution (2024) 24:22
Fig. 3 (See legend on previous page.)

Page 12 of 20
Berlesetal. BMC Ecology and Evolution (2024) 24:22
addition to the cluster results described above render the
graphs reliable to illustrate the clustering trend of leaps
and its association with the species. However, the poor
low-dimensional representation of support use vari-
ability, the continuous scatter of data points in the space
spanned by the two first PCs and the poor clustering are
all indicators of a low correlation among habitat utiliza-
tion variables (i.e., the leaps were characterized by a large
diversity of combinations of support characteristics).
Leaping performance andposture
e inertia gain criterion of the cluster analysis again
suggested two clusters also for leaping performance and
posture (Fig.3B, Fig. S7). However, the three measures
used for cluster validation suggest the selection of
different numbers of clusters (Table S6). e connectivity
supported the choice of two clusters, whereas the Dunn
index suggested five clusters and the silhouette width
three clusters instead of two. Nevertheless, both latter
indices were only slightly improved by increasing the
number of clusters, with the Dunn index increasing
from 0.03 to 0.05 and the silhouette width increasing
from 0.367 to 0.372. Hence, we chose the two clusters
supported by the inertia gain criterion and the
connectivity index for further interpretation.
Cluster 1 is characterized by a vertical body posture
during take-off and landing (though one of the five data
points closest to the centroid was a horizontal-to-hori-
zontal leap; Table S7). It is further characterized by rela-
tively longer leaping distances and flight durations and by
short take-off and landing durations. Cluster 2 is repre-
sented by a horizontal body posture during both take-off
and landing, as well as a shorter horizontal leap distance
and a shorter flight time on the one hand and longer
take-off and landing times on the other hand (Table S7).
We found a significant association between the clusters
and the species (χ²(df = 3) = 425.88, p < 0.001) with large
differences among all species. In particular, when look-
ing at the percentage of leaps of each species falling into
the first cluster in descending order, it was 86% for L.
nigrifrons, 66% for S. mystax, 20% for S. imperator, and
3% for S. midas. us, S. midas and S. imperator can be
considered specialized in short leaps with horizontal
postures, L. nigrifrons specialized in long leaps with ver-
tical take-off and landing postures, with S. mystax fall-
ing in between with a non-specialized leaping behavior.
However, the cluster results do not reflect the fact that S.
midas leaps significantly more frequently in a horizontal
posture than the other two Saguinus species, that did not
differ significantly in this regard (Fig.2, Table S9).
e described cluster trends are mostly well-captured
by the first two principal components, which explain
64% of the variance (Fig.3B, Fig. S8). PC1 separates the
two clusters, and body posture, flight duration and hori-
zontal distance load stronger on PC1 compared to PC2.
However, take-off and landing durations load stronger
on PC2 and do not appear to contribute to the cluster-
ing. Using these variables for cluster discrimination as
suggested by the five most typical leaps described above
might thus be misleading. PC2 shows that L. nigrifrons
and especially S. mystax display longer take-off and
landing durations than the other two species (Fig.3E).
e distribution of species means with the close proxim-
ity of the centroids of S. midas and S. imperator reflect
the three clusters supported by the silhouette width
index above. All of this suggests that the categorization
of the performance depends on the performance vari-
ables under consideration.
Locomotor morphology
Based on the results in terms of habitat utilization and
leaping performance, we amended our predictions for S.
mystax. S. mystax is similar to S. midas and S. imperator
in support use, but exhibits larger leap distances. Based
on this, we expect S. mystax to have relatively more
robust bones to withstand higher compressive stress than
S. midas and S. imperator, but at the same time relatively
less robust bones than L. nigrifrons. e detailed predic-
tions for each of the internal and external bone morpho-
logical variables (Fig.1C) can be found in Table S10.
According to the inertia gain criterion, the cluster
analysis revealed two clusters of similar morphologies
(Fig.3C, Fig. S7). ey are also supported by the con-
nectivity (9.34) as well as the silhouette width (0.29). e
Dunn index (0.99) favors six clusters, but only margin-
ally improves compared to its value when choosing two
clusters (0.6) (Table S6). A reason for this could be the
small sample size. Consequently, we chose two clusters
for interpretation.
Cluster 2 is characterized by larger values than cluster 1
in 24 out of 29 variables, although the degree of difference
varies between about 0.25 to 2 standard deviations (Fig.
S9). Cluster 1 is only characterized by larger values
concerning patellar groove height, the length of the
iliopsoas in-lever, IMI, as well as DA and BV.TV in the
trabeculae of the femoral head. We found no significant
association between the clusters and the species
(χ²(df = 3) = 4.89, p = 0.18), which might be attributed
to the small sample size. All three specimens of each
L. nigrifrons and S. mystax as well as two S. imperator
specimens and one S. midas specimen were assigned to
the first cluster, the remaining three specimens falling
into the second cluster. us, L. nigrifrons and S. mystax
have less robust morphologies with reduced potential
for joint torque generation whereas S. imperator and S.
midas are more diverse in their morphology compared to

Page 13 of 20
Berlesetal. BMC Ecology and Evolution (2024) 24:22
L. nigrifrons and S. mystax with a trend toward increased
robustness and heightened potential for joint torque
generation.
e described cluster trends are well-captured by the
first two principal components, which explain 55% of
the variance (Fig. 3C, Fig. S8). PC1 separates the two
clusters and 16 of the 29 variables have a loading above
0.5 on PC1 (Fig. 3F). L. nigrifrons and S. mystax are
positioned in the first cluster which is associated with
low PC1 scores, whereas S. imperator falls clearly into
the second cluster associated with high PC1 scores. S.
midas falls in between the two clusters. Six variables
have a loading above 0.5 on PC2, but since none of the
cluster validation indices suggested a third cluster that
is separated along PC2 from the first two clusters, they
might not be informative to explain the morphological
variation between species.
Discussion
e fundamental phenomenon of adaptation may only be
understood by an analysis of the intricate relationship of
the different biologically relevant ‘levels’ constituting it:
from morphology to function, behavior, and environment
[1–4]. Integrated studies of prey capture and
locomotion in predators show that various performance
characteristics, such as acceleration and deceleration,
may depend on predator ecology [86–89], while factors
such as habitat structure and locomotor performance
[90–92], are reflected in morphological adaptations [12,
13]. Our study contributes to the understanding of the
relationship between morphology, habitat utilization,
and performance in the field. We collected data on
habitat use and leaping behavior of the free-living
mixed-species group L. nigrifrons and S. mystax, and of
S. midas and S. imperator living in a naturalistic park,
showing trends in support preference, body posture
during leaping, and leap distance. Our results agree with
the previously described great locomotor plasticity in
tamarins in relation to their small body size spectrum
[40], highlighting the importance of an in-depth analysis
of support use and performance. Further, many of the
morphological variables studied here contributed to a
clustering of the studied species, but this clustering and
its underlying morphological trends contradicted our
predictions that were derived from the habitat utilization
and performance analyses. ese findings suggest a
complex relationship between these three ‘levels’ of
habitat use, performance, and morphology. We discuss
the results of each ‘level’ in detail to highlight novel
insights and discrepancies with other studies as well as
potential connections between the three studied ‘levels’.
Major dierences inhabitat utilization
betweenLeontocebus andSaguinus
Support inclination, flexibility, and diameter are of great
importance for arboreal primates, as they can vary within
a single movement sequence, which constantly needs to
be adjusted accordingly [93, 94]. We could show on the
basis of the cluster results that the three Saguinus spe-
cies can be described as generalists with a flexible choice
of supports for leaping. L. nigrifrons, on the other hand,
can be described as a trunk specialist in frequently using
vertical (70–90°), thick (> 10cm) and inflexible supports
both during take-off and landing, as expected. is dis-
tinction between Saguinus as generalists and L. nigri-
frons as a trunk specialist could indicate a phylogenetic
signal in the habitat utilization pattern (Fig.4). However,
our results suggest that all studied tamarin species are in
principle capable of coping with a variety of support use
characteristics, as already described in the literature [26,
40, 41, 93]. Also, our results point out that the Saguinus
species differ in how frequently they use certain sup-
port characteristics. e fact that S. imperator showed
a stronger utilization of thin supports (also shown by
Karantanis [93] and Buchanan-Smith [46]), but not par-
ticularly of more flexible and horizontally oriented ones
compared to the other two, reflects the fact the S. imper-
ator moves mostly on branches in the lower forest layers
Fig. 4 Cladogram of the four studied tamarin species with cluster characterizations for each of the three analyzed biologically relevant ‘levels’. The
tree topology follows Botton‑Divet & Nyakatura [16], but branch length information was omitted for simplicity

Page 14 of 20
Berlesetal. BMC Ecology and Evolution (2024) 24:22
[37, 46]. S. midas, on the other hand, shows the most
frequent utilization of horizontal supports as well as of
inflexible supports. e latter fact contradicts previous
studies that found that S. midas mostly leaps between
terminal, flexible branches in the mid-forest layer [43,
44]. Perhaps this is related to the specific group of indi-
viduals under study or differences in habitat structure
and indicates a larger intraspecific variability. S. mystax,
on the other hand, stood out in using flexible supports
most frequently. is is surprising, since this species also
used vertical supports more frequently than the other
two Saguinus species. Perhaps this reflects a more dedi-
cated use of trunks from younger trees.
e only interspecific comparative studies besides
ours [34] investigated the support uses of S. mystax and
L. nigrifrons. e author could not find any differences
in support sizes used for locomotion in these two spe-
cies. Although L. nigrifrons used vertical trunks more
often than S. mystax, both species rarely used supports
larger than 10cm in diameter [34]. In contrast, in our
study, L. nigrifrons primarily used supports thicker than
10cm in diameter for leaping (Fig.2). is is surprising
since we studied the primates at the same study site as
Norconk [34] and might be related to intraspecific vari-
ability. Another explanation could be differences in the
composition of the trees between subareas of the EBQB
study site. However, mirroring our results, Norconk [34]
found a clear difference in the support inclination used,
with L. nigrifrons spending 50% of the observed time on
vertical supports, while S. mystax most frequently used
oblique and horizontal supports. In a study by Smith
[38], L. nigrifrons used larger supports and vertical sup-
ports in more than 70% of the leaps, while S. mystax
preferred thinner supports and used horizontal sup-
ports for leaping in more than 80%. In yet another study
by Garber and Pruetz [33] at a site at the Rio Blanco,
it was shown that S. mystax, used horizontal branches
for locomotion twice as often as vertical ones, but small,
medium and large supports were used relatively equally.
In summary, the support uses of the four species here
studied is mostly consistent with information in the lit-
erature, while differences between study groups might
be related to intraspecific variability or differences in
habitat structure. However, for example, in a compara-
tive study of S. mystax at two different locations, it
was shown that despite significant differences in forest
structure, the overall pattern and frequency of move-
ment and postural behavior hardly differed [33]. Also,
in a mixed species group of L. nigrifrons and S. mystax,
both species preferred their predominant support use
regardless of the availability of supports in the different
forest layers [26].
Leaping performance partly corresponds tohabitat
utilization
We hypothesized that, based on the climbing height
preferences reported in the literature, S. midas and S.
mystax predominantly perform leaps in a horizontal
body posture during take-off and landing. e distances
between the branch-type supports are likely relatively
short. is is why we also expected these species to leap
short distances with short flight phases. On the other
hand, we expected L. nigrifrons to start and land leaps in
a vertical body posture and to leap the largest horizontal
distance with the longest flight phase because of tree
trunks being usually positioned further from each other
than horizontal branches. We expected a more generalist
leaping performance from S. imperator due to its
preference for inclined supports with smaller diameters
in the lower forest layers.
e predictions for L. nigrifrons were met, but the three
Saguinus species showed intricate differences in their
leaping performance that partly opposed our predictions
in the case of S. imperator and S. mystax. According to
our analysis, S. midas and S. imperator can be considered
specialists for short horizontal leaps independent of the
choice of support. is pattern is particularly evident
for S. midas because it adopted a horizontal body pos-
ture during take-off and landing most frequently among
the study species. Nevertheless, S. midas is known to
be capable of long leaps when using vertical trunks and
might leap as far as 7.6m [95]. S. mystax, on the other
hand, performed longer leaps than the other two Sagui-
nus species in our study despite using a horizontal body
posture as frequently as S. imperator. Perhaps, this can be
explained by the use of specific support diameters, since
S. mystax and L. nigrifrons both used the largest support
diameters the most often and also leapt longer distances
than the other two tamarin species. Nevertheless, L. nig-
rifrons covered the largest mean distance in our species
sample with distances greater than 1m making up 65% of
the leaps (see Fig. S10). A similar observation was done in
a study by Garber and colleagues [32], in which L. nigri-
frons covered horizontal distances between 1 and 2m in
51.5% of leaps, while only 9.5% were below 0.5m and dis-
tances greater than 3m accounted for only 2% of leaps.
However, Garber and Leigh [41] found that L. nigrifrons
covered a horizontal distance of less than 1m during ver-
tical leaps in 83% of the observed cases. Perhaps this is
again related to the specific group of individuals under
study as well as the respective habitat structure.
e habitat characteristics and performance cluster
indices did not agree upon a single number of clusters,
neither within each of these levels nor between both
levels. is means that a straightforward relationship
between both levels cannot be inferred. Instead, it

Page 15 of 20
Berlesetal. BMC Ecology and Evolution (2024) 24:22
appears that specific features of support use and leaping
performance are linked to each other. For example, the
fact that we found S. mystax to use flexible supports most
frequently among the four species studied is consistent
with it displaying the longest take-off and landing
times. In contrast, the similarly long push-off durations
during take-off in L. nigrifrons may be related to this
species covering larger distances. Long take-off times
are beneficial for generating large support reaction force
impulses during push-off. is impulse may result in
larger covered distances or may be lost for propulsion due
to the bending of the flexible support [48]. In addition,
it was shown for gibbons that this shift of the center of
mass (CoM) further towards the landing support by
increasing hip joint excursion during the take-off reduces
the leaping distance [70]. is allows to compensate
for the deflection of the support with minimal work for
the CoM [70]. It would be interesting to study whether
S. mystax and L. nigrifrons, when they travel together
through the forest, choose supports with different
flexibility properties while using supports with the same
diameters. Furthermore, according to our results, the
frequency of each posture adopted during take-off and
landing follows the frequency of the support orientation
(Fig.2). In particular, vertical support orientations and
vertical body postures are used with the same frequency
within each species, and the same holds for horizontal
supports and the horizontal body posture. e fact that
this correspondence was observed in all four species
suggests that all of them can adjust their body posture
according to the support orientation faced at each leap.
Morphology does notcorrelate tohabitat utilization
andleaping performance
Based on the results of the habitat utilization and the
leaping performance of the studied tamarins, we expected
the locomotor morphology of L. nigrifrons to reflect a
larger potential to generate joint torques and resist forces
associated with long distance leaps from and to inflex-
ible supports compared to the Saguinus species. Only the
lever-arm of the iliopsoas muscle was expected to reflect
a larger potential for joint torque generation in Saguinus,
because it would benefit stability while climbing on nar-
row supports. e IMI was expected to be smaller in L.
nigrifrons, because relatively longer hind limbs compared
to the other species are expected to contribute to the
long-distance leaps performed by this species.
Although we found many morphological features
to discriminate the studied species, none of these dif-
ferences were according to our predictions. However,
while we expected overall habitat utilization and loco-
motor performance to drive morphological adaptations,
it might be that specific aspects of both could relate to
the morphological similarity between L. nigrifrons and
S. mystax. As discussed above, both these species were
the ones that most frequently used the largest support
diameters, and, more obviously, the ones that covered
the larger leaping distances. Moreover, all the L. nig-
rifrons and S. mystax specimens were associated with
cluster 1, which was characterized by larger values of
DA and BV.TV on average. is is largely in agreement
with expected functional adaptations of these two tra-
becular parameters. Concerning BV.TV, this trait is
expected to positively relate to the magnitude of stresses
acting on a joint (i.e., more bone volume in response to
increased biomechanical stimulations, see Kivell [59]
for a review). Hence, higher BV.TV in L. nigrifrons and
S. mystax reflects a larger potential to resist biomechani-
cal stresses experienced during long-distance leaps. A
higher DA was previously detected in the femoral head
of leaping strepsirrhines and related to their stereotypi-
cally oriented vertical clinging and leaping [96]. Perhaps,
a more preferentially oriented direction of the trabeculae
could also reflect the necessity to resist the large support
reaction forces during take-off and landing. However, for
all other studied traits it appears paradoxical that a lower
potential for the generation of joint torqueand resistance
against forces is associated with larger leaping distances
from comparably wider supports. Rather, leaping of tam-
arins from and to trunk-like supports should require the
generation of large propulsive forces through a powerful
extension of the hindlimbs during take-off and the with-
standing of considerable compressive and bending forces
acting on the forelimbs during landing [40, 97]. In case
of S. mystax, it might be explained by its comparably fre-
quent use of flexible supports. It is known that the peak
take-off force is reduced by the flexibility of the support,
which means that the animals have to generate higher
forces during take-off from flexible supports than from
rigid supports [48].
While we found L. nigrifrons and S. mystax to display
a similar locomotor morphology with regard to skeletal
measurements, Garber [40] found that L. nigrifrons has
relatively longer arms, legs, hands and feet than S. mys-
tax. is limb elongation was explained in connection
with the preferred trunk-to-trunk leaping [40], since, as
already mentioned, this type of leaping differs consider-
ably from horizontal leaping due to a stationary, verti-
cal holding position during take-off, which requires the
animals to overcome substantial inertia. In addition, the
body must be turned almost 180 degrees in flight. Fig-
ure S1illustrates typical postures exhibited by the dif-
ferent species during trunk-to-trunk leaping during
our study. Interestingly, only L. nigrifrons performs a
strong rotation of the body axis during trunk-to-trunk
leaps. e other three species also leap from a vertical

Page 16 of 20
Berlesetal. BMC Ecology and Evolution (2024) 24:22
holding position but are already facing the landing sup-
port with the flight phase being similar to that of a hori-
zontal leap, where the body does not have to be rotated
as much. Since the required muscle power is generated
by the hindlimbs, long and/or well-muscled hindlimbs
are advantageous [50]. Due to the notable dominance of
the quadriceps femoris muscle in specialized strepsir-
rhine primates, the force from the knee and thigh can
be transferred to adjacent joints and segments [98]. In
our study, however, the gluteus medius muscle, based on
its in-lever, appears to be suited for more powerful hip
extension in S. midas and S. imperator than in L. nigri-
frons. It is also difficult to relate the higher potential to
generate joint torques and resist forces in S. imperator
and S. midas to any of the habitat utilization and per-
formance parameters, because both species never stood
out in having similarly large values or frequencies in any
of them. Rather, both covered shorter distances (Fig.2)
than S. mystax and L. nigrifrons, so it can be assumed
that they did not have to exert higher forces than the spe-
cies that covered longer distances. However, our mor-
phological data imply that both these species experience
substantial loads or, at least, can cope with such loads
during take-off and landing. Perhaps, in comparison, the
larger IMI in L. nigrifrons allow for an extended decel-
eration distance due to relatively elongated forelimbs,
which is crucial to slow the body down when landing
on the trunks [51]. S. midas and S. imperator may have
compensated for this lack of elongation with relatively
larger articular surfaces and more powerful muscles for
shoulder stabilization [99]. Although these two species
rarely landed on vertical rigid supports in our study, it
should not be ignored that a rather rarely used ability
may nevertheless be an ecologically significant activity
and may have a major impact on musculoskeletal struc-
tures [100, 101]. Another explanation for the relatively
more robust limb features in S. midas and S. imperator
could be a possible increased risk of falls on thinner and
more compliant branches which could result in a selec-
tion pressure towards more robust bones. It has been
demonstrated that rare loading events, with high poten-
tial costs of failure, such as falls, will favor higher bone
safety factors in comparison to habitual loads [see 102
for a review]. e reason for the falls could be various,
such as breakage of the take-off or landing supports, or
misjudgment of the distance and flexibility of the land-
ing supports [103]. e falls we observed were always
falls from horizontal or slightly oblique flexible supports.
However, this only affected juveniles of S. mystax, which
were not considered in this study. e frequency of free
falls of S. mystax while foraging was three times higher
than that of L. nigrifrons in a study by Peres [104]. And
in a study by Price [103], S. oedipus was able to grasp
lower branches and prevent falls to the ground only 30%
of the time. Even from heights up to 10m, tamarins were
not observed to injure themselves by landing always on
their legs [103]. is fact may suggest that S. midas and
S. imperator have more robust bone features due to pos-
sible falls from compliant supports.
Conclusion: weak integration of‘levels’
Our study revealed that the three considered ‘levels’ of
habitat utilization, leaping performance and locomotor
morphology are surprisingly weakly integrated in our
study system of tamarins. It also appears that specific
variables of support use correspond to specific variables
of leaping performance, while morphological differences
cannot be linked meaningfully to any of these two ‘lev-
els’ except for a few isolated variables. e intraspecific
variability and the poor clustering on the two behavio-
ral levels should caution against the assignment of spe-
cies to coarse ecological categories (vertical leaper vs.
horizontal leaper, or more general ones, e.g. fossorial vs.
arboreal vs. cursorial etc.) for drawing ecomorphologi-
cal inferences, as it is often done in large-scale interspe-
cific studies, at any taxonomic level. Hypotheses on the
form-function relationship depend on the degree of our
previous understanding of the ecology of the animals,
but this might be incomplete, offering much room for
mis- and overinterpretation. Variations in the ecology
of the animals and slight changes in behavior that were
not considered in the respective study may influence and
bias the inferences [2]. In our case, it may well be that
the similar morphology of S. mystax and L. nigrifrons is
determined by shared, but not leaping-related, behav-
ioral characteristics [2] that were not considered in our
study such as clinging to trunks during exudate feeding
[105]. Also, sometimes similar morphologies may result
from or respond to different selective pressures (i.e. ‘one-
to-many mapping’ of form onto function [9, 106]; e.g.
similar ulnar morphology may result from both digging
and climbing in xenarthrans [107]). Additional behavio-
ral data as well as studying other tamarin species might
elucidate which aspects of support use and performance
mainly drive the evolution of the locomotor apparatus.
Our in-depth exploratory analysis, despite its practical
limitations (e.g. two species being observed in a natural-
ist park instead in the wild; small interspecific sample
size etc.), serves as a starting point for the generation of
novel hypotheses.

Page 17 of 20
Berlesetal. BMC Ecology and Evolution (2024) 24:22
Supplementary Information
The online version contains supplementary material available at https:// doi.
org/ 10. 1186/ s12862‑ 023‑ 02193‑z [108].
Additional le1: Figure S1. Different types of trunk‑to‑trunk leaps in
the studied tamarin species. Figure S2. Measurements of the humerus
obtained in Geomagic. Cranial (A) and craniodorsal (B). Figure S3. Meas‑
urements of the ulna obtained in Geomagic. Figure S4. Measurements of
the radius obtained in Geomagic. Figure S5. Measurements of the femur
obtained in Geomagic. Figure S6. Measurements of the tibia obtained
in Geomagic. Figure S7. Hierarchical trees of clustering methods. Figure
S8. Scree plots for the dimensionality reduction analyses. Figure S9.
Characterization of morphology clusters. The difference between Cluster
2 (C2) and Cluster 1 (C1) is illustrated on the scale of standard deviations
for each variable. Standardization was also done to facilitate comparison
among variables like it was done for principal component analysis. Figure
S10. Boxplot for leaping distance of L. nigrifrons. Supporting informa-
tion note 1. Additional information on quantification of internal bone
structure. Supporting information note 2. Additional information on
Body size correction of morphological data. Supporting information
note 3. Additional information on Missing data imputation. Supporting
information note 4. Additional information on sampling bias correc‑
tion of performance data. Supporting information note 5. Additional
information on inferential statistics.
Additional le2: TableS1. Characteristics of trees within the seven
Whittaker plots with the total number (N) and the percentage value of
trees and categories for the score of height of stay of the monkeys during
the focal animal scan method in Peru and “La vallée des singes”. TableS2.
Percentages of utilized categories of the studied support characteristics
and posture. TableS3. Specimens used for morphological analysis.
TableS4. Effect of sample bias correction on sample sizes. TableS5.
Numbers of observations. TableS6. Cluster validation indices. TableS7.
Cluster characteristics, i.e., five most representative observations for each
cluster within the levels of habitat utilization and leaping performance
and posture. TableS8. Inferential statistics regarding the categorical
behavioral variables. TableS9.Inferential statistics regarding the
continuous behavioral variables.
Acknowledgements
We thank field assistant M. H. Arirama for the support in the field, and E. W.
Heymann for the possibility of conducting our study at Estación Biológica
Quebrada Blanco. We are grateful to the “Vallée des Singes”‑zoological park,
Mr. E. le Grelle, and the “Centre pour la Conservation des Primates” for granting
us access to the monkeys and to all the zoological staff that helped during
the study. We are grateful to the collection managers and staff for allowing
access and loan of the specimens: R.S. Voss, N. Duncan, R.D. Macphee, N.B.
Simmons, S. Ketelsen, E. Hoeger and M. Surovy from the American Museum of
Natural History in New York; A.W. Ferguson, L. Smith, J. Phelps and L. Heaney
from the Field Museum of Chicago. We thank all those who helped during
the scanning process: D. Boyer and J. Gladman from Duke University; Z.‑X.
Luo and A.I. Neander from the University of Chicago. We acknowledge critical
comments from anonymous reviewers as well as G. Boulinguez‑Ambroise for
their numerous and helpful comments.
Authors’ contributions
AUTHOR CONTRIBUTIONS: PB and JAN conceived of the study. PB, JAN, JPG
and JW designed the methodology. PB, FA, LBD, and JW collected the data. PB
and JW analyzed the data. PB drafted the manuscript. All authors revised the
manuscript and gave final approval for publication.
Funding
Open Access funding enabled and organized by Projekt DEAL. This study
was funded by the German Research Foundation DFG (NY 63/2 − 1). FA is
financially supported by the project grant TMPFP3_217022 from the Swiss
National Science Foundation (https://www.snf.ch; SNSF Swiss Postdoctoral
Fellowships, SPF).
Availability of data and materials
All data and R code necessary for the reproduction of the results are available
on Figshare (https:// doi. org/ 10. 6084/ m9. figsh are. 24937 131).
Declarations
Ethics approval and consent to participate
Field work was carried out under authorization no. AUT‑IFS‑2017‑062 from
the Servicio Nacional Forestal y de Fauna Silvestre (SERFOR) of the Peruvian
Ministry of Agriculture.
Consent for publication
Not applicable.
Competing interests
The authors declare no competing interests.
Author details
1 AG Vergleichende Zoologie, Institut für Biologie, Humboldt‑Universität
zu Berlin, Philippstr. 12/13, 10115 Berlin, Germany. 2 Present Address:
Institute of Ecology and Evolution, University of Bern, Bern 3012,
Switzerland. 3 Department of Earth Sciences, University of Cambridge,
Cambridge, UK. 4 Museum für Naturkunde, Leibniz‑Institut für Evolutions‑
und Biodiversitätsforschung, Berlin, Germany. 5 La Vallée des Singes,
Romagne 86700, France.
Received: 24 July 2023 Accepted: 19 December 2023
References
1. Bock WJ, Von Wahlert G. Adaptation and the form‑function complex.
Evolution. 1965;269–99. https:// doi. org/ 10. 2307/ 24064 39.
2. Feilich KL, López‑Fernández H. When does form reflect function?
Acknowledging and supporting ecomorphological assumptions. Integr
Comp Biol. 2019;59(2):358–70. https:// doi. org/ 10. 1093/ icb/ icz070.
3. Garber PA. Within‑and between‑site variability in moustached tamarin
(Saguinus mystax) positional behavior during food procurement.
Primate Locomotion. Boston, MA: Springer; 1998. pp. 61–78. https:// doi.
org/ 10. 1007/ 978‑1‑ 4899‑ 0092‑0_4.
4. Reilly SM, Wainwright PC. Conclusion: ecological morphology and the
power of integration. In: Wainwright PC, Reilly SM, editors. Ecological
Morphology: Integrative Organismal Biology. Chicago: University of
Chicago Press; 1994. p. 339–54.
5. Alfaro ME, Bolnick DI, Wainwright PC. Evolutionary consequences of
many‑to‑one mapping of jaw morphology to mechanics in labrid
fishes. Am Nat. 2005;165(6):E140–54. https:// doi. org/ 10. 1086/ 429564.
6. Smith JM, Savage RJ. Some locomotory adaptations in mammals. Zool
J Linn Soc. 1956;42(288):603–22. https:// doi. org/ 10. 1111/j. 1096‑ 3642.
1956. tb022 20.x.
7. Marcé‑Nogué J, Püschel TA, K aiser TM. A biomechanical approach
to understand the ecomorphological relationship between primate
mandibles and diet. Sci Rep. 2017;7(1):1–12. https:// doi. org/ 10. 1038/
s41598‑ 017‑ 08161‑0.
8. Wölfer J, Amson E, Arnold P, Botton‑Divet L, Fabre AC, van Heteren AH,
Nyakatura JA. Femoral morphology of sciuromorph rodents in light of
scaling and locomotor ecology. J Anat. 2019;234(6):731–47. https:// doi.
org/ 10. 1111/ joa. 12980.
9. Zelditch ML, Ye J, Mitchell JS, Swiderski DL. Rare ecomorphological
convergence on a complex adaptive landscape: body size and diet
mediate evolution of jaw shape in squirrels (Sciuridae). Evolution.
2017;71(3):633–49. https:// doi. org/ 10. 1111/ evo. 13168.
10. Fleagle JG. Locomotor behavior and muscular anatomy of sympatric
Malaysian leaf‑monkeys (Presbytis obscura and Presbytis melalophos).
Am J Phys Anthropol. 1977a;46:297–307. https:// doi. org/ 10. 1002/ ajpa.
13304 60211.

Page 18 of 20
Berlesetal. BMC Ecology and Evolution (2024) 24:22
11. Fleagle JG. Locomotor behavior and skeletal anatomy of sympatric
Malaysian leaf‑monkeys (Presbytis obscura and Presbytis melalophos).
Yearbook of Physical Anthropology. 1977b;20:440–53.
12. Rodman PS. Skeletal differentiation of Macaca fascicularis and Macaca
nemestrina in relation to arboreal and terrestrial quadrupedalism. Am J
Phys Anthropol. 1979;51(1):51–62. https:// doi. org/ 10. 1002/ ajpa. 13305
10107.
13. Ward SC, Sussman RW. Correlates between locomotor anatomy and
behavior in two sympatric species of Lemur. Am J Phys Anthropol.
1979;50:575–90.
14. Clifton KB, Motta PJ. (1998). Feeding morphology, diet, and ecomorpho‑
logical relationships among five Caribbean labrids (Teleostei, Labridae).
Copeia, 953–966. https:// doi. org/ 10. 2307/ 14473 42.
15. Irschick DJ, Losos JB. Do lizards avoid habitats in which performance is
submaximal? The relationship between sprinting capabilities and struc‑
tural habitat use in Caribbean anoles. Am Nat. 1999;154(3):293–305.
https:// doi. org/ 10. 1086/ 303239.
16. Botton‑Divet L, Nyakatura JA. Vertical clinging and leaping induced
evolutionary rate shifts in postcranial evolution of tamarins and mar‑
mosets (Primates, Callitrichidae). BMC Ecol Evol. 2021;21(1):132. https://
doi. org/ 10. 1186/ s12862‑ 021‑ 01848‑z.
17. Nyakatura JA. Early primate evolution: insights into the functional
significance of grasping from motion analyses of extant mammals. Biol
J Linn Soc. 2019;127(3):611–31. https:// doi. org/ 10. 1093/ bioli nnean/
blz057.
18. Jenkins FA. Tree shrew locomotion and the origins of primate arboreal‑
ism. Primate Locomotion. New York; London: Academic Press; 1974. pp.
85–115.
19. Schapker NM, Chadwell BA, Young JW. Robust locomotor perfor‑
mance of squirrel monkeys (Saimiri boliviensis) in response to simu‑
lated changes in support diameter and compliance. J Experimental
Zool Part A: Ecol Integr Physiol. 2022;337(5):417–33. https:// doi. org/
10. 1002/ jez. 2574.
20. Silcox MT, López‑Torres S. Major questions in the study of primate
origins. Annu Rev Earth Planet Sci. 2017;45(1):113–37. https:// doi. org/
10. 1146/ annur ev‑ earth‑ 063016‑ 015637.
21. Young JW, Shapiro LJ. Developments in development: what have we
learned from primate locomotor ontogeny? Am J Phys Anthropol.
2018;165:37–71. https:// doi. org/ 10. 1002/ ajpa. 23388.
22. Boinski S. The positional behavior and substrate use of squirrel mon‑
keys: ecological implications. J Hum Evol. 1989;18(7):659–77. https://
doi. org/ 10. 1016/ 0047‑ 2484(89) 90099‑7.
23. Young JW, Stricklen BM, Chadwell BA. Effects of support diameter and
compliance on common marmoset (Callithrix jacchus) gait kinematics. J
Exp Biol. 2016;219(17):2659–72. https:// doi. org/ 10. 1242/ jeb. 140939.
24. Toussaint S, Llamosi A, Morino L, Youlatos D. The central role of small
vertical substrates for the origin of grasping in early primates. Curr Biol.
2020;30(9):1600–13. https:// doi. org/ 10. 1016/j. cub. 2020. 02. 012.
25. Rylands AB, Heymann EW, Alfaro L, Buckner J, Roos JC, Matauschek
C, Mittermeier C, R. A. Taxonomic review of the new world tamarins
(primates: Callitrichidae). Zool J Linn Soc. 2016;177(4):1003–28. https://
doi. org/ 10. 1111/ zoj. 12386.
26. Berles P, Heymann EW, Golcher F, Nyakatura JA. Leaping and differential
habitat use in sympatric tamarins in Amazonian Peru. J Mammal.
2021;103(1):146–58. https:// doi. org/ 10. 1093/ jmamm al/ gyab1 21.
27. Heymann EW, Buchanan‑Smith HM. The behavioural ecology of mixed‑
species troops of callitrichine primates. Biol Rev. 2000;75(2):169–90.
https:// doi. org/ 10. 1111/j. 1469‑ 185X. 1999. tb000 44.x.
28. Nadjafzadeh MN, Heymann EW. Prey foraging of red titi monkeys, Cal-
licebus cupreus, in comparison to sympatric tamarins, Saguinus mystax
and Saguinus fuscicollis. Am J Phys Anthropol. 2008;135(1):56–63.
https:// doi. org/ 10. 1002/ ajpa. 20704.
29. Porter LM. Forest use and activity patterns of Callimico goeldii in com‑
parison to two sympatric tamarins, Saguinus fuscicollis and Saguinus
labiatus. Am J Phys Anthropology: Official Publication Am Association
Phys Anthropologists. 2004;124(2):139–53. https:// doi. org/ 10. 1002/
ajpa. 10343.
30. Buchanan‑Smith H. Field observations of Goeldi’s monkey, Callimico
goeldii, in northern Bolivia. Folia Primatological. 1991;57(2):102–5.
31. Garber PA. Locomotor behavior and feeding ecology of the Panama‑
nian tamarin (Saguinus oedipus geoffroyi, callitrichidae, primates). Int J
Primatol. 1980;1(2). https:// doi. org/ 10. 1159/ 00015 6570.
32. Garber PA, McKenney AC, Mallott EK. The Ecology of Trunk‑To‑Trunk
Leaping in Saguinus fuscicollis: Implications for Understanding Locomo‑
tor Diversity in Callitrichines. Neotrop Primates. 2012;19:(1). https:// doi.
org/ 10. 1896/ 044. 019. 0101.
33. Garber PA, Pruetz JD. Positional behavior in moustached tamarin mon‑
keys: effects of habitat on locomotor variability and locomotor stability.
J Hum Evol. 1995;28:411–26. https:// doi. org/ 10. 1006/ jhev. 1995. 1032.
34. Norconk MA. Mechanisms promoting stability in mixed S. mystax and S.
Fuscicollis troops. Am J Phys Anthropol. 1990;135:56–63. https:// doi. org/
10. 1002/ ajp. 13502 10208.
35. Nyakatura JA, Heymann EW. Effects of support size and orientation
on symmetric gaits in free‑ranging tamarins of amazonian Peru:
implications for the functional significance of primate gait sequence
patterns. J Hum Evol. 2010;58(3):242–51. https:// doi. org/ 10. 1016/j.
jhevol. 2009. 11. 010.
36. Garber PA. Vertical clinging, small body size, and the evolution of
feeding adaptations in the Callitrichinae. Am J Phys Anthropol.
1992;88(4):469–82. https:// doi. org/ 10. 1002/ ajpa. 13308 80404.
37. Izawa K. Foods and feeding behavior of monkeys in the upper Amazon
basin. Primates. 1975;16(3):295–316. https:// doi. org/ 10. 1007/ BF023
81557.
38. Smith AC. Comparative Ecology of Saddle‑Back (Saguinus fuscicollis)
and moustached (Saguinus mystax) tamarins. Ph. D. thesis, Univ. of Read-
ing; 1997.
39. Ankel‑Simons F, editor. Primate anatomy. An introduction. 3rd ed.
Amsterdam, Boston: Elsevier Academic Press; 2007.
40. Garber PA. A comparative study of positional behavior in three species
of tamarin monkeys. Primates. 1991;32(2):219–30. https:// doi. org/ 10.
1007/ BF023 81179.
41. Garber PA, Leigh SR. Patterns of positional behavior in mixed‑species
troops of Callimico goeldii, Saguinus labiatus, and Saguinus fuscicollis in
northwestern Brazil. Am J Primatol. 2001;54:17–31. https:// doi. org/ 10.
1002/ ajp. 1009.
42. Emmons L, Feer F. Neotropical rainforest mammals. A field guide.
Chicago: University of Chicago Press; 1990.
43. Fleagle JG, editor. Primate adaptation and evolution. Second Edition:
Academic Press; 1999.
44. Fleagle JG, Mittermeier RA. Locomotor behavior, body size, and
comparative ecology of seven Surinam monkeys. Am J Phys Anthropol.
1980;52:301–14. https:// doi. org/ 10. 1002/ ajpa. 13305 20302.
45. Mittermeier RA, van Roosmalen MGM. (1981): Preliminary Observa‑
tions on Habitat Utilization and Diet in Eight Surinam Monkeys. In: Folia
Primatologica 36, 1981, S. 1–39. https:// doi. org/ 10. 1159/ 00015 6007.
46. Buchanan‑Smith HM. Tamarin polyspecific associations: forest utiliza‑
tion and stability of mixed‑species groups. Primates. 1999;40(1):233–47.
https:// doi. org/ 10. 1007/ BF025 57713.
47. Delciellos AC, Vinícius Vieira M. Jumping ability in the arboreal
locomotion of didelphid marsupials. Mastozoología Neotropical.
2009;16(2):299–307.
48. Demes B, Jungers WL, Gross TS, Fleagle JG. Kinetics of leaping primates:
influence of substrate orientation and compliance. Am J Phys Anthro‑
pol. 1995;96:419–29. https:// doi. org/ 10. 1002/ ajpa. 13309 60407.
49. Dunbar DC. Aerial maneuvers of leaping Lemurs: the physics of whole‑
body rotations while airborne. Am J Primatol. 1988;16(4):291–303.
https:// doi. org/ 10. 1002/ ajp. 13501 60402.
50. Emerson SB. Jumping and leaping. In: Hildebrand M, Bramble DM, Liem
KF, Wake DB, editors. Functional vertebrate morphology. Cambridge:
Belknap Press; 1985. pp. 58–72.
51. Peters A, Preuschoft H. External biomechanics of leaping in Tarsius and
its morphological and kinematic consequences. Biology of Tarsiers.
1984;96:109.
52. Bicca‑Marques JC. Hand specialization, sympatry, and mixed‑species
associations in callitrichines. J Hum Evol. 1999;36(4):349–78. https:// doi.
org/ 10. 1006/ jhev. 1998. 0272.
53. Garber PA. Seasonal patterns of diet and ranging in two species of tam‑
arin monkeys: stability versus variability. Int J Primatol. 1993;14(1):145–
66. https:// doi. org/ 10. 1007/ BF021 96509.

Page 19 of 20
Berlesetal. BMC Ecology and Evolution (2024) 24:22
54. Garber PA, Davis LC. Intraspecific variability in anatomy and positional
behavior in two tamarin species. Am J Phys Anthropol. 1995;22:110–1.
55. Davis LC. Functional morphology of the forelimb and long bones in the
Calltrichidae (Platyrrhini, Primates). Ph. D. thesis, Carbondale: Southern
Illinois Univerity; 2002.
56. Buckner JC, Alfaro JWL, Rylands AB, Alfaro ME. Biogeography of
the marmosets and tamarins (Callitrichidae). Mol Phylogenet Evol.
2015;82:413–25. https:// doi. org/ 10. 1016/j. ympev. 2014. 04. 031.
57. Druelle F, Aerts P, Ngawolo JCB, Narat V. Impressive arboreal gap‑
crossing behaviors in wild bonobos, Pan paniscus. Int J Primatol.
2020;41:129–40. https:// doi. org/ 10. 1007/ s10764‑ 020‑ 00140‑z.
58. Barak MM, Lieberman DE, Hublin JJ. A Wolff in sheep’s clothing:
trabecular bone adaptation in response to changes in joint loading
orientation. Bone. 2011;49(6):1141–51. https:// doi. org/ 10. 1016/j. bone.
2011. 08. 020.
59. Kivell TLA. Review of trabecular bone functional adaptation: what have
we learned from trabecular analyses in extant hominoids and what can
we apply to fossils? J Anat. 2016;228:569–94. https:// doi. org/ 10. 1111/
joa. 12446.
60. Hirasaki E, Kumakura H, Matano S. Biomechanical analysis of vertical
climbing in the spider monkey and the Japanese macaque. Am J Phys
Anthropology: Official Publication Am Association Phys Anthropologists.
2000;113(4):455–72. https:// doi. org/ 10. 1002/ 1096‑ 8644.
61. Legreneur P, Monteil KM, Pellé E, Montuelle S, Bels V. Submaximal leap‑
ing in the grey mouse lemur. Zoology. 2011;114(4):247–54. https:// doi.
org/ 10. 1016/j. zool. 2011. 03. 004.
62. Channon AJ, Usherwood JR, Crompton RH, Günther MM, Vereecke EE.
The extraordinary athletic performance of leaping gibbons. Biol Lett.
2012;8(1):46–9. https:// doi. org/ 10. 1098/ rsbl. 2011. 0574.
63. Nauwelaerts S, Aerts P. Take‑off and landing forces in jumping frogs. J
Exp Biol. 2006;209(1):66–77. https:// doi. org/ 10. 1242/ jeb. 01969.
64. Demes B, Jungers WL. Long bone cross‑sectional dimensions, locomo‑
tor adaptations and body size in prosimian primates. J Hum Evol.
1993;25(1):57–74. https:// doi. org/ 10. 1006/ jhev. 1993. 1038.
65. Marchi D, Ruff CB, Capobianco A, Rafferty KL, Habib MB, Patel BA.
The locomotion of Babakotia radofilai inferred from epiphyseal
and diaphyseal morphology of the humerus and femur. J Morphol.
2016;277(9):1199–218. https:// doi. org/ 10. 1002/ jmor. 20569.
66. Ruff CB. Long bone articular and diaphyseal structure in Old World
monkeys and apes. I: locomotor effects. Am J Phys Anthropology: Offi‑
cial Publication Am Association Phys Anthropologists. 2002;119(4):305–
42. https:// doi. org/ 10. 1002/ ajpa. 10117.
67. Ryan TM, Shaw CN. Unique suites of trabecular bone features char‑
acterize locomotor behavior in human and non‑human anthropoid
primates. PLoS ONE. 2012;7(7):e41037. https:// doi. org/ 10. 1371/ journ al.
pone. 00410 37.
68. Ryan TM, Walker A. Trabecular bone structure in the humeral and femoral
heads of anthropoid primates. Anat Record: Adv Integr Anat Evolutionary
Biology. 2010;293(4):719–29. https:// doi. org/ 10. 1002/ ar. 21139.
69. Terborgh J. Five New World primates: a study in comparative ecology.
Princeton University Press; 1983.
70. Channon AJ, Günther MM, Crompton RH, D’Août K, Preuschoft H,
Vereecke EE. The effect of substrate compliance on the biomechanics
of gibbon leaps. J Exp Biol. 2011;214(4):687–96. https:// doi. org/ 10. 1242/
jeb. 046797.
71. Aerts P. Vertical jumping in Galago senegalensis: the quest for an obli‑
gate mechanical power amplifier. Philosophical Trans Royal Soc Lond
Ser B: Biol Sci. 1998;353(1375):1607–20. https:// doi. org/ 10. 1098/ rstb.
1998. 0313.
72. Heymann EW, Tirado Herrera ER. Estación Biológica Quebrada Blanco:
Un Sitio Poco conocido para investigación en biodiversidad y ecología
en la Amazonía peruana. Revista peruana de biología. 2021;28(3).
https:// doi. org/ 10. 15381/ rpb. v28i3. 20266.
73. Encarnación F. Introduccion a la flora y vegetacion de la rimatol
peruana: terminologia y descripcion de las formaciones vegetales de la
llanura amazonica. Candollea. 1985;577(10):237–52.
74. Dagosto M, Yamashita N. Effect of habitat structure on positional
behavior and support use in three species of lemur. Primates.
1998;39(4):459–72. https:// doi. org/ 10. 1007/ BF025 57569.
75. Bezanson M, Watts SM, Jobin MJ. Tree truthing: how accurate are
substrate estimates in primate field studies? Am J Phys Anthropol.
2012;147(4):671–7. https:// doi. org/ 10. 1002/ ajpa. 22037.
76. Schneider CA, Rasband WS, Eliceiri KW. NIH Image to ImageJ: 25 years
of image analysis. Nat Methods. 2012;9(7):671–5. https:// doi. org/ 10.
1038/ nmeth. 2089.
77. Alfieri F, Botton‑Divet L, Nyakatura JA, Amson E. Integrative Approach
uncovers new patterns of ecomorphological convergence in slow
arboreal xenarthrans. J Mammalian Evol. 2022;29(2):283–312. https://
doi. org/ 10. 1007/ s10914‑ 021‑ 09590‑5.
78. Turnquist JE, Minugh‑Purvis N. Functional morphology. In: Bennett BT,
C. R. Abee and, Henrickson R, editors. Nonhuman Primates in Biomedical
Research. 2nd ed. Boston: Academic Press; 2012. pp. 87–129.
79. Team RC. R: a language and environment for statistical computing.
2012 ed. Vienna, Austria: R Foundation for Statistical Computing; 2023.
80. Josse J, Husson F. missMDA: a package for handling missing values in
multivariate data analysis. J Stat Softw. 2016;70:1–31. https:// doi. org/ 10.
18637/ jss. v070. i01.
81. Lê S, Josse J, Husson F. FactoMineR: an R package for multivariate analy‑
sis. J Stat Softw. 2008;25:1–18. https:// doi. org/ 10. 18637/ jss. v025. i01.
82. Brock G, Pihur V, Datta S, Datta S. clValid: an R Package for Cluster Valida‑
tion. J Stat Softw. 2008;25(4):1–22. https:// www. jstat soft. org/ v25/ i04/.
83. Handl J, Knowles J. (2005). Exploiting the trade‑off—the benefits of
multiple objectives in data clustering. In International conference on
evolutionary multi-criterion optimization (pp. 547–560). Springer, Berlin,
Heidelberg. https:// doi. org/ 10. 1007/ 978‑3‑ 540‑ 31880‑4_ 38.
84. Dunn JC. Well‑separated clusters and optimal fuzzy partitions. J Cyber‑
netics. 1974;4(1):95–104. https:// doi. org/ 10. 1080/ 01969 72740 85460 59.
85. Rousseeuw PJ. Silhouettes: a graphical aid to the interpretation and vali‑
dation of cluster analysis. J Comput Appl Math. 1987;20:53–65. https://
doi. org/ 10. 1016/ 0377‑ 0427(87) 90125‑7.
86. McElroy EJ, Reilly SM. The relationship between limb morphology,
kinematics, and force during running: the evolution of locomotor
dynamics in lizards. Biol J Linn Soc. 2009;97(3):634–51. https:// doi. org/
10. 1016/ 0377‑ 0427(87) 90125‑7.
87. Higham TE. The integration of locomotion and prey capture in ver‑
tebrates: morphology, behavior, and performance. Integr Comp Biol.
2007;47(1):82–95. https:// doi. org/ 10. 1093/ icb/ icm021.
88. Holzman R, Collar DC, Mehta RS, Wainwright PC. An integrative mod‑
eling approach to elucidate suction‑feeding performance. J Exp Biol.
2012;215(1):1–13.
89. Kane EA, Higham TE. Complex systems are more than the sum of their
parts: using integration to understand performance, biomechanics, and
diversity. Integr Comp Biol. 2015;55(1):146–65. https:// doi. org/ 10. 1093/
icb/ icv033.
90. Priyadarshana T, Asaeda T, Manatunge J. Foraging behaviour of planktiv‑
orous fish in artificial vegetation: the effects on swimming and feeding.
Hydrobiologia. 2001;442:231–9. https:// doi. org/ 10. 1023/A: 10175 78524
578.
91. Robinson SK, Holmes RT. Foraging behavior of forest birds: the rela‑
tionships among search tactics, diet, and habitat structure. Ecology.
1982;63(6):1918–31. https:// doi. org/ 10. 2307/ 19401 30.
92. Whelan CJ. Foliage structure influences foraging of insectivorous forest
birds: an experimental study. Ecology. 2001;82(1):219–31. https:// doi.
org/ 10. 1890/ 0012‑ 9658.
93. Karantanis NE. Comparative positional behaviour in three captive
callitrichid species: Leontopithecus chrysomelas, Saguinus imperator and
Cebuella pygmaea. Doctoral dissertation, MSc Thesis, University College
of London; 2010.
94. McNamara A, Dunham NT, Shapiro LJ, Young JW. The effects of natural
substrate discontinuities on the quadrupedal gait kinematics of free‑
ranging Saimiri sciureus. Am J Primatol. 2019;81:e23055.
95. Thorington RW. Observations of the tamarin Saguinus midas. Folia
Primatological. 1968;9(2):95–8. https:// doi. org/ 10. 1002/ ajp. 23055.
96. Ryan TM, Ketcham RA. The three‑dimensional structure of trabecu‑
lar bone in the femoral head of Strepsirrhine primates. J Hum Evol.
2002;43(1):1–26. https:// doi. org/ 10. 1006/ jhev. 2002. 0552.
97. Garber PA, Blomquist GE, Anzenberger G. Kinematic analysis of trunk‑
to‑trunk leaping in Callimico goeldii. Int J Primatol. 2005;26(1):223–40.
https:// doi. org/ 10. 1007/ s10764‑ 005‑ 0732‑z.

Page 20 of 20
Berlesetal. BMC Ecology and Evolution (2024) 24:22
98. Demes B, Fleagle JG, Lemelin P. Myological correlates of prosimian leap‑
ing. J Hum Evol. 1998;34(4):385–99. https:// doi. org/ 10. 1006/ jhev. 1997.
0203.
99. Eigen L, Nyakatura JA. Architectural properties of the musculoskeletal
system in the shoulder of two callitrichid primate species derived from
virtual dissection. Primates. 2021;62(5):827–43. https:// doi. org/ 10. 1007/
s10329‑ 021‑ 00917‑7.
100. Isler K. 3D‑kinematics of vertical climbing in hominoids. American Jour‑
nal of Physical Anthropology. The Official Publication of the American
Association of Physical Anthropologists. 2005;126(1):66–81. https:// doi.
org/ 10. 1002/ ajpa. 10419.
101. Ruff CB, Harper CM, Goldstein DM, Daegling DJ, McGraw WS. Long
bone structural proportions and locomotor behavior in Cercopitheci‑
dae. J Hum Evol. 2019;132:47–60. https:// doi. org/ 10. 1016/j. jhevol. 2019.
04. 003.
102. Willie BM, Zimmermann EA, Vitienes I, Main RP, Komarova SV. Bone adap‑
tation: Safety factors and load predictability in shaping skeletal form.
Bone. 2020;131:115114. https:// doi. org/ 10. 1016/j. bone. 2019. 115114.
103. Price EC. Adaptation of captive‑bred cotton‐top tamarins (Saguinus
rimata) to a natural environment. Zoo Biol. 1992;11(2):107–20. https://
doi. org/ 10. 1002/ zoo. 14301 10206.
104. Peres CA. Prey‑capture benefits in a mixed‑species group of Amazo‑
nian tamarins, Saguinus fuscicollis and S. mystax. Behav Ecol Sociobiol.
1992;31(5):339–347. https:// doi. org/ 10. 1007/ BF001 77774.
105. Heymann EW, Smith AC. When to feed on gums: temporal patterns of
gummivory in wild tamarins, Saguinus mystax and Saguinus fuscicollis
(Callitrichinae). Zoo Biology: Published in Affiliation with the American
Zoo and Aquarium Association. 1999;18(6):459–71. https:// doi. org/ 10.
1002/ (SICI) 1098‑ 2361.
106. Liem KF. Adaptive significance of intraspecific and interspecific
differences in the feeding repetoires of cichlid fishes. Am J Zool.
1980;20:295–314.
107. White JL. Indicators of locomotor habits in xenarthrans: evidence for
locomotor heterogeneity among fossil sloths. J Vertebr Paleontol.
1993;13(2):230–42. https:// doi. org/ 10. 1080/ 02724 634. 1993. 10011 502.
108. Berles P, Wölfer J, Alfieri F, Botton‑Divet L, Guéry J‑P, Nyakatura JA.
R_Data_Berles et al. 2024 [Internet]. figshare; 2024 [cited 2024Feb8].
https:// doi. org/ 10. 6084/ m9. figsh are. 24937 131. v1.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in
published maps and institutional affiliations.