Access to this full-text is provided by PLOS.
Content available from PLOS One
This content is subject to copyright.
RESEARCH ARTICLE
Ontogenetic shape trajectory of
Trichomycterus areolatus varies in response to
water velocity environment
Peter C. SearleID
1
*, Margaret Mercer
1
, Evelyn Habit
2
, Mark C. Belk
1
1Department of Biology, Brigham Young University, Provo, Utah, United States of America, 2Facultad de
Ciencias Ambientales y Centro EULA, Departmento de Sistemas Acua
´ticos, Universidad de Concepcio
´n,
Concepcio
´n, Chile
*petersearle94@gmail.com
Abstract
Body and head shape among fishes both vary between environments influenced by water
velocity and across ontogeny. Although the shape changes associated with variation in
average water velocity and ontogeny are well documented, few studies have tested for the
interaction between these two variables (i.e., does ontogenetic shape variation differ
between velocity environments). We use geometric morphometrics to characterize shape
differences in Trichomycterus areolatus, a freshwater catfish found in high and low-velocity
environments in Chile. We identify a significant interaction between velocity environment
and body size (i.e., ontogeny). Ontogenetic patterns of shape change are consistent with
other studies, but velocity environment differentially affects the ontogenetic trajectory of
shape development in T.areolatus. Shape change over ontogeny appears more con-
strained in high-velocity environments compared to low-velocity environments.
Introduction
Morphometric traits in animals and plants are shaped by the selective pressures of their envi-
ronments resulting in interspecific variation in shape [1–4]. Shape responds sensitively to envi-
ronmental pressures such as predation, resource use, competition, temperature, water velocity,
and water availability [5–10]. For example, head and body size changed in two Australian
snakes (Pseudechis porphyriacus and Dendrelaphis punctulatus) following introduction of a
toxic cane toad (Bufo marinus), and head shape changed among species in the Neotropical
cichlid genus Geophagus depending on prey availability [11,12]. Organisms vary in form (i.e.,
shape) among species largely in response to variation in habitat use or trophic niche and corre-
sponding environmental variation among habitats [2,12].
Environmental variation can also lead to intraspecific morphometric differentiation [13–
17]. Examples include differences in head shape in wall lizards (Podarcis bocagei) between sax-
icolous and ground-dwelling habitats; wing morphology in speckled wood butterflies (Parage
aegeria) along a latitudinal gradient; tail, head and ear length in West African Dwarf goats
(Capra aegagrus hircus) among agro-ecological zones and swimming performance and shape
PLOS ONE
PLOS ONE | https://doi.org/10.1371/journal.pone.0252780 June 11, 2021 1 / 17
a1111111111
a1111111111
a1111111111
a1111111111
a1111111111
OPEN ACCESS
Citation: Searle PC, Mercer M, Habit E, Belk MC
(2021) Ontogenetic shape trajectory of
Trichomycterus areolatus varies in response to
water velocity environment. PLoS ONE 16(6):
e0252780. https://doi.org/10.1371/journal.
pone.0252780
Editor: Peter Eklo¨v, Uppsala Universitet, SWEDEN
Received: October 20, 2020
Accepted: May 24, 2021
Published: June 11, 2021
Peer Review History: PLOS recognizes the
benefits of transparency in the peer review
process; therefore, we enable the publication of
all of the content of peer review and author
responses alongside final, published articles. The
editorial history of this article is available here:
https://doi.org/10.1371/journal.pone.0252780
Copyright: ©2021 Searle et al. This is an open
access article distributed under the terms of the
Creative Commons Attribution License, which
permits unrestricted use, distribution, and
reproduction in any medium, provided the original
author and source are credited.
Data Availability Statement: The data set used for
analysis has been uploaded to Dryad. The public
DOI is doi:10.5061/dryad.rn8pk0p8f and URL is
https://doi.org/10.5061/dryad.rn8pk0p8f.
in Creek Chub (Semotilus atromaculatus) due to urbanization [18–21]. Such intraspecific phe-
notypic variation can be the result of genetic [22] or plastic variation [23–25].
Stream zonation is an outgrowth of the river continuum concept, which suggests that as
streams move from their headwaters to their mouths, they differ in many environmental vari-
ables [26]. Across a stream continuum (i.e., across zones) there are multiple environmental
variables that vary including: substrate size, stream width, depth, flow volume, gradient, and
water velocity, as well as biotic variables such as prey type and availability [27]. Each of these
variables may contribute to shape differences observed between fish from different stream
zones. Of these variables, water velocity has been demonstrated as a strong predictor of intra-
specific variation in shape in fishes [28]. Typically, high-velocity environments favor a nar-
rower body shape, whereas low-velocity environments favor a more robust body shape [28].
This general response has been well documented in many fishes including centrarchids [29],
characids [30], cichlids [31], cyprinids [32] and salmonids [33]. Although velocity is typically
considered the primary selective agent in these studies, it is often difficult to separate water
velocity from the other, usually covarying, variables that differ between stream zones.
In addition to variation in shape in response to the environment, an organism’s shape
changes over the course of development [34,35]. Since D’Arcy Wentworth Thompson’s influ-
ential book, On Growth and Form [36], ontogenetic shape change has been well documented
among invertebrates: crustaceans [37] and insects [38]; as well as vertebrates: reptiles [39],
birds [40] and mammals [41]. Most fishes hatch or are born several orders of magnitude
smaller in size compared to adults. Because of this dramatic change in body size as they grow,
fish occupy multiple niches over the course of their ontogenetic development, and they can
experience different selective pressures within each sequential niche [42]. Often associated
with these ontogenetic niche shifts are similarly dramatic ontogenetic shape changes [43–45].
Generally, juveniles have larger eyes and heads relative to body size; whereas, adults have
smaller eyes and heads relative to body size [46,47]. Studies from numerous fish taxa have doc-
umented similar shape changes [43–45,47,48]. Although many studies have addressed the
change in form over developmental stages in organisms with reference to allometry [46,49],
few studies have addressed how differing selective environments interact with ontogenetic pat-
terns to produce phenotypic variation within species [50].
The potential interaction between environment and ontogenetic shape trajectories in fishes
has been quantified in relatively few studies. In the sexually dimorphic livebearer, Brachyraphis
rhabdophora, males exhibited parallel ontogenetic shape trajectories such that shape differ-
ences between juveniles and adults were maintained between predation environments. How-
ever, females exhibited convergent ontogenetic shape trajectories such that shape differences
were more pronounced in juvenile females compared to adult females between predation envi-
ronments [51]. Similarly, three-spined sticklebacks (Gasterosteus aculeatus) exhibited differen-
tial ontogenetic shape trajectories between two marine (with different habitat structure) and
one freshwater habitat resulting in morphologically distinct adults [52] (Also see [53–58]).
Taken together, these studies provide evidence that the interaction between environmental
variation and ontogeny could result in differential shape trajectories within species. However,
few studies have specifically quantified the interaction between water velocity environment
across the ontogenetic shape trajectory in fishes.
In this paper we characterize shape variation in Trichomycterus areolatus across ontogeny
in high-velocity and low-velocity stream environments. Trichomycterus areolatus is a freshwa-
ter catfish found throughout Central Chile. It is a small (up to 140 mm standard length), gener-
alist, benthic carnivore that feeds mostly on insect larvae and crustaceans [59–61].
Trichomycterus areolatus is a good candidate for studying the potential interactive effect of
water velocity environment on ontogenetic shape trajectories because they are one of the most
PLOS ONE
Shape variation in Trichomycterus areolatus
PLOS ONE | https://doi.org/10.1371/journal.pone.0252780 June 11, 2021 2 / 17
Funding: Funding for this work was supported by
grants to MCB by the Roger and Victoria Sant
Foundation and the Department of Biology,
Brigham Young University, as well as EH by
projects 032904 3/R Direccio
´n de Investigacio
´n,
Universidad del Bı
´o-Bı
´o and 204.310.041-1.0
Direccio
´n de Investigacio
´n, Universidad de
Concepcio
´n. The funders had no role in study
design, data collection and analysis, decision to
publish, or preparation of the manuscript.
Competing interests: The authors have declared
that no competing interests exist.
broadly distributed species in Chilean rivers and streams, and they occur in all water velocity
environments (i.e., stream zones) [26]. Furthermore, all size classes of T.areolatus are found in
all stream zones, they appear to occupy the same stream section throughout their entire lives,
and there is no evidence of migration among stream sections [27]. To investigate the potential
interaction between environments characterized by differences in water velocity and ontogeny
(i.e. do ontogenetic shape trajectories differ between high and low-velocity environments), we
used landmark-based geometric morphometrics [62] to quantify body and head shape across
ontogenetic trajectories of T.areolatus in contrasting velocity environments. We document an
interaction between velocity environment and ontogenetic shape variation in T.areolatus.
Materials & methods
Study site and collection
The Andalie
´n River (Concepcio
´n province, Chile, 36.740673 S, 73.016947 W) is a small coastal
drainage (48 km) that lacks geographic or reproductive barriers to gene flow (i.e., no waterfalls
or other flow barriers) along its entire length. In addition, the river exhibits strong zonation
patterns in habitat structure and environmental variables. The river can be divided into three
distinct zones: the rhithron (higher gradient stream characterized by high current velocity,
large substrate, and small stream size), the transition (with intermediate characteristics), and
the potamon (lower gradient river characterized by low current velocity, small substrate and
large stream size) [27]. We recognize that these stream zones differ in several ways (Table 1,
modified from [26]), but these variables covary with stream velocity in predictable ways.
Water velocity is most different between the transition and potamon zones, thus we group the
rithron and transition zones as a high-velocity environment, and the potamon zone as a low-
velocity environment. As in many studies where a continuous environment is divided into dis-
crete segments for comparison [15,63,64], there are several other variables that potentially dif-
fer between high and low-velocity environments and that could create selective differences
between environments. For this reason, we refer to these as water velocity environments and
we acknowledge that other stream variables may contribute to observed variation in shape in
T.areolatus.
We collected T.areolatus from the Andalie
´n River and two of its tributary streams
(Nongue
´n and Queule Rivers) in 2003 and 2004 using electrofishing backpack equipment. All
collections were done under the auspices of Direccio
´n de Investigacio
´n, Universidad de Con-
cepcio
´n and the Undersecretariat of Fisheries (Collections were made before IACUC protocols
were required for field studies in Chile, but we followed the Guidelines for Use of Fishes in
Table 1. Environmental characteristics of the rithron, transition, and potamon zones in the Andalie
´n River, Chile during the wet season.
Parameter Unit Rithron Transition Potamon
Mean Depth Cm 26.5 ±5.4 28.3 ±11.1 25.5 ±7.3
Maximum Depth Cm 50.0 ±6.3 61.8 ±27.7 55.2 ±17.4
Flow m
3
/s 0.7 ±0.6 3.3 ±2.6 3.6 ±2.7
Avg. Velocity m/s
1
0.83 ±0.06 1.0 ±0.24 0.59 ±0.3
Boulders % 76.7 ±2.5 16.6 ±25.8 0 ±0
Width M 5.8 ±2.2 11.8 ±6.3 19.8 ±15.8
Order 3.3 ±0.5 3.6 ±0.5 4.3 ±0.2
Gradient ˚ 0.4 ±0.09 0.2 ±0.07 0.14 ±0.03
Modified from [26].
https://doi.org/10.1371/journal.pone.0252780.t001
PLOS ONE
Shape variation in Trichomycterus areolatus
PLOS ONE | https://doi.org/10.1371/journal.pone.0252780 June 11, 2021 3 / 17
Field Research provided by the American Society of Ichthyologists and Herpetologists, the
American Fisheries Society and the American Institute of Fisheries Research Biologists [65]).
Specimens were originally collected for a study comparing distribution patterns of fish assem-
blages in the Andalie
´n River drainage [27]. According to guidelines in [65], we retained speci-
mens based on population density at each location. The majority of fish were released, but a
representative sample of the entire size range of the specimens was randomly retained as
museum vouchers. It is these specimens retained as vouchers that form the basis of this study.
Because the density of T.areolatus in the rithron zone was low, fewer specimens were retained
(n = 20). Because these specimens from the rithron did not represent the full range of body
sizes available in the other two zones, we combined samples from the rithron zone and the
transition zone. In addition, water velocities in the rithron and transition zones were higher
and more similar to each other than velocities in the potamon zone (Table 1). To determine if
the inclusion of specimens collected in the rithron changed the statistical outcome or interpre-
tation of the results, we ran the overall shape analysis with and without the samples from the
rithron. Neither the statistical significance, nor the estimated least squares means of shape dif-
fered between the two analyses, so we report results from analysis of the full data set.
Trichomycterus areolatus is sexually monomorphic [66], so we did not distinguish between
males and females in our sampling and analysis. We euthanized specimens with an overdose
of BZ-20 (20% ethyl p-aminobenzoate) and assigned ID numbers. We preserved specimens in
ethyl alcohol and measured and photographed them at Brigham Young University. We used a
Canon EOS Digital Rebel XT DSLR camera with a Canon EF-S 18–55 mm f3.5–5.6 lens. Some
specimens were damaged or preserved in unnatural positions such that they could not be
included in the analysis. Mean sizes and size distributions were similar between high and low-
velocity environments. Specimens used for the analysis of lateral body shape from high-veloc-
ity environments (n = 183) had a mean standard length of 55.2 mm (SD = 19.5 mm;
range = 22.4–116.3 mm); and specimens from low-velocity environments (n = 103) had a
mean standard length of 58.3 mm (SD = 23.5 mm; range = 23.5–115.1 mm). Specimens used
for the analysis of head shape from high-velocity environments (n = 182) had a mean standard
length of 58 mm (SD = 22 mm; range = 22.4–134.9 mm); and specimens from low-velocity
environments (n = 113) had a mean standard length of 58.6 (SD = 24.3 mm; range = 23.5–
138.8 mm). We photographed the right lateral view of the body, and the dorsal view of the
head of each fish for morphometric analysis and included a mm-scale ruler in the photograph
for scaling. Preserved specimens are deposited in the Monte L. Bean Life Science Museum at
Brigham Young University in Provo, Utah, USA.
Geometric morphometrics
We used landmark-based geometric morphometrics to quantify body and head shape in T.
areolatus [67]. Prior to landmarking, we meticulously reviewed digital images to confirm that
each specimen was not rotated dorsally or ventrally, was straight on the dorso-ventral and
anteroposterior axis, had a closed mouth, and the operculum was in a relaxed, closed position.
Where possible we re-photographed specimens, but if the error could not be corrected, we
removed the specimen from the analysis. Most specimens were included in both the body and
head shape analysis, but a few were used in only one or the other analysis. We digitized land-
marks and semi-landmarks using the program tpsDig [68]. One researcher landmarked all
specimens in random order without reference to collection location (i.e., velocity environ-
ment) or body size. Two other researchers independently inspected landmarked images to
confirm homologous and consistent placement of landmarks. For analysis of lateral body
shape, we used 13 landmarks defined as follows: 1) rostral-most point, 2) center of eye, 3)
PLOS ONE
Shape variation in Trichomycterus areolatus
PLOS ONE | https://doi.org/10.1371/journal.pone.0252780 June 11, 2021 4 / 17
posterior tip of operculum, 4) juncture of the ventral margin of the operculum with the ventral
outline of the body, 5) vertical projection of the x-value of landmark 4 on the dorsal outline, 6)
posterior tip of spine on preoperculum, 7) midpoint between landmarks 5 and 9 on the dorsal
outline, 8) anterior origin of pelvic fin, 9) anterior origin of dorsal fin, 10) anterior origin of
anal fin, 11) midpoint between landmarks 4 and 8 on the ventral outline, 12) dorsal projection
from anterior extent of caudal fin, and 13) ventral projection from anterior extent of caudal
fin. Landmarks 7 and 11 are sliding semilandmarks (Fig 1).
The dorsal view of the head represents a symmetrical structure, so we landmarked only the
right half for analysis [69]. We used 7 landmarks defined as follows: 1) anterior extent of pre-
maxilla at midline of head, 2) midline of head projected from point 7, 3) lateral edge of nostril,
4) medial margin of eye, 5) lateral margin of eye, 6) lateral extent of the head projected from
point 5, and 7) posterior tip of opercular spine (Fig 2).
We generated shape variables from our landmark data using tpsRelW [70]. tpsRelW uses a
generalized Procrustes analysis to remove non-shape variation (i.e., position, orientation, and
scale) [71,72], and generates shape variables in the form of partial warps and uniform compo-
nents (i.e., W, the weight matrix). The program then calculates a series of relative warps that
result from a principal components analysis of the weight matrix [62,67,73]. We used these rel-
ative warps as our shape variables for analysis. In addition, we used centroid size (a multivari-
ate measure of size) derived from this analysis to represent the stage of ontogenetic
development. We excluded relative warps that individually accounted for <1% of shape varia-
tion to avoid inflating the real degrees of freedom associated with analysis of shape variation
[17,74]. Consequently, we used the first 12 of 22 relative warps for the body (accounting for
96.15% of total shape variation) and the first 9 of 10 relative warps for the head (accounting for
99.91% of total shape variation).
Statistical analysis
To determine the effects of velocity environment and ontogeny on shape variation in T.areola-
tus, we used a multivariate linear mixed model [16,17,74]. We analyzed body and head shape
separately, using relative warps as the response variable. We used location (rithron/transition
versus potamon) as a discrete predictor in the model to represent either high or low-velocity
Fig 1. Body landmarks. Landmarks used in the analysis of shape of the lateral body view of T.areolatus.
https://doi.org/10.1371/journal.pone.0252780.g001
PLOS ONE
Shape variation in Trichomycterus areolatus
PLOS ONE | https://doi.org/10.1371/journal.pone.0252780 June 11, 2021 5 / 17
environments, and we used centroid size as a covariate predictor in the model to represent
ontogenetic stage of development. We included the interaction between velocity environment
and centroid size (ontogeny) to test for different patterns of shape variation across ontogeny
between velocity environments. The multivariate linear mixed model required that the matrix
of shape variables (relative warps) be vectorized such that each individual specimen was repre-
sented by 12 rows (body) or 9 rows (head) of data, and each row corresponded to one relative
warp [74]. Thus, the response variable is transposed from a row in a matrix to a vector column
of responses. This vectorization creates the need for an index variable that preserves the iden-
tity of specific relative warps. This index variable is used as a predictor variable in the model
much like time is used as a predictor in a typical repeated measures design [15,74–76]. Relative
warps are principal components, and as such they are orthogonal to each other, they have a
mean of zero, and their ordination is arbitrary relative to each other. For example, individuals
that would have a positive score on one component would not necessarily score positively on
any other component. For this reason, the index variable, by preserving the identity of each
relative warp, is necessary to test our hypotheses of shape differences among groups. Thus, it is
the two-way interactions between velocity environment and the index variable, and centroid
size and the index variable, and the three-way interaction between velocity environment, cen-
troid size, and the index variable that allow us to test the hypothesis of significant shape differ-
ences between velocity environments across ontogeny on all relative warps simultaneously
Fig 2. Head landmarks. Landmarks used in the analysis of shape of the head of T.areolatus.
https://doi.org/10.1371/journal.pone.0252780.g002
PLOS ONE
Shape variation in Trichomycterus areolatus
PLOS ONE | https://doi.org/10.1371/journal.pone.0252780 June 11, 2021 6 / 17
[51,74,75,77]. We included individual ID number as a random effect in our analysis because
each specimen had multiple shape variables (i.e., relative warps) as response variables. We esti-
mated degrees of freedom using the Kenward–Roger method [78] and conducted analyses
using Proc MIXED in SAS (SAS version 9.4, SAS Institute Inc., Cary NC, USA). To determine
which relative warps differed across ontogeny and between velocity environments, we plotted
least squares means and associated 95% confidence intervals for small and large individuals
separately in both velocity environments.
To quantify the magnitude and direction of shape change between velocity environments
over ontogeny we used a phenotypic change vector analysis (PCVA) [79–81]. Because PCVA
requires two categorical variables to form the two vectors that are compared, we used size at
maturity (51 mm SL) [66] to divide the sample between juveniles and adults, and then divided
the adults into small adults and large adults by dividing the size range of adults in half (adults
ranged in size from 51mm to 139 mm, dividing line between small adults and large adults = 103
mm SL) [66]. We used juveniles as the small size class and large adults as the large size class to
characterize the PCVA across the entire size range. The PCVA tests for differences in the mag-
nitude and/or direction of ontogenetic shape change between velocity environments across all
relative warps used in the analysis. We tested for significant differences in magnitude and
direction of shape change between small and large individuals in different velocity environ-
ments using ASReml-R version 4 [82] using modified R [83] scripts from [74,84].
To visualize shape variation between velocity environments at small and large sizes we cal-
culated divergence vectors as described in [85,86]. These divergence vectors characterize dif-
ferences in shape across all relative warps between velocity environments for small and large
size classes separately. We calculated the divergence vector as the sum of the products of the
first eigenvector (from a principal components analysis of the least squares means for each rel-
ative warp in the two velocity environments) times the associated relative warp scores for each
individual. We then used these divergence scores for each individual as a regressor on shape in
tpsRegr [87] to generate a thin-plate spline visualization of the extremes of shape variation
between velocity environments. These thin-plate spline plots of shape deformation represent
shape divergence across all relative warps between velocity environments. Data used for all
analyses are available in the Dryad Digital Repository (doi:10.5061/dryad.rn8pk0p8f).
Results
Body shape of T.areolatus (right lateral view) differed significantly by centroid size (over ontog-
eny), and between velocity environments, and centroid size and velocity environment exhibited
a significant interaction (see the two-way and three-way interaction with the index variable in
Table 2). Ontogenetic shape change in the body was mainly associated with relative warp 1,
with minor contributions from relative warps 4, 5, and 6. Shape differences associated with
velocity environment were mainly evident on relative warp 1, with minor contributions from
relative warps 5, 6, 8, and 10 (Fig 3). The magnitude of shape change of the body (p = 0.0024),
but not the direction (p = 0.3889) differed significantly between velocity environments. Fish
from the high-velocity environment had a significantly lower magnitude of shape change over
ontogeny compared to fish from the low-velocity environment (D
High
= 0.0339, D
Low
= 0.0494,).
Correspondingly, the change in shape over ontogeny in the high-velocity environment mainly
involves a slight reduction in the proportional length of the head relative to the body and a slight
deepening of the body in the midsection (Fig 4, right side). In contrast, the change in shape over
ontogeny in the low-velocity environment includes a substantial reduction in the proportional
length of the head relative to the body and a substantial narrowing of the body in the midsection
(Fig 4, left side). In addition, shape differs between high-velocity and low-velocity environments
PLOS ONE
Shape variation in Trichomycterus areolatus
PLOS ONE | https://doi.org/10.1371/journal.pone.0252780 June 11, 2021 7 / 17
at both ends of the ontogenetic spectrum (Fig 3). Smaller T.areolatus from the low-velocity
environment had proportionally longer heads and deeper bodies and heads (dorsoventrally)
compared to specimens from the high-velocity environment (Fig 4, upper); whereas, larger
T.areolatus from the low-velocity environment had proportionally shorter heads compared to
specimens from the high-velocity environment (Fig 4, lower).
Table 2. Multivariate analysis of covariance effects for body shape and head shape.
Source Degrees of Freedom F-Value p-Value
Body Shape
Velocity environment 1,2244 1.50 0.2201
Centroid size (CS) 1, 2244 46.72 <0.0001
Index 11, 1334 29.25 <0.0001
Velocity environment�Index 11, 1334 3.88 <0.0001
Centroid size�Index 11, 1334 33.20 <0.0001
Velocity environment�CS�Index 12, 1325 3.12 0.0002
Head Shape
Velocity environment 1, 1723 0.02 0.8882
Centroid size (CS) 1, 1723 19.25 <0.0001
Index 8, 1089 24.42 <0.0001
Velocity environment�Index 8, 1089 10.73 <0.0001
Centroid size�Index 8, 1089 24.26 <0.0001
Velocity environment�CS�Index 9, 1075 7.14 <0.0001
Centroid size represents ontogenetic variation.
https://doi.org/10.1371/journal.pone.0252780.t002
Fig 3. Relative Warp least squares means for body. Least squares means (error bars represent 95% confidence
intervals of the mean) for each of 12 relative warps from the multivariate linear mixed model of body shape variation of
T.areolatus. High-velocity environment is represented by closed circle symbols and solid connecting lines, and low-
velocity environment is represented by open circle symbols and dashed connecting lines. Small and large ontogenetic
stages are represented by centroid sizes of 35.00 and 150.00, respectively.
https://doi.org/10.1371/journal.pone.0252780.g003
PLOS ONE
Shape variation in Trichomycterus areolatus
PLOS ONE | https://doi.org/10.1371/journal.pone.0252780 June 11, 2021 8 / 17
Head shape of T.areolatus (dorsal view) differed significantly over ontogeny (centroid
size), between velocity environments, and ontogeny and velocity environment exhibited a sig-
nificant interaction (see the two-way and three-way interaction with the index variable in
Table 2). Ontogenetic shape change in the head was associated with relative warps 1, 2, 3, 5,
and 6. Shape differences in the head associated with velocity environment were evident on rel-
ative warps 1–7 (Fig 5). The direction of shape change of the head (p = 0.0024), but not the
magnitude (p = 0.0886) differed significantly between velocity environments. The multivariate
direction of the ontogenetic trajectory of head shape differed by 47.63˚ between velocity envi-
ronments. This difference in direction of the ontogenetic trajectory of head shape between
velocity environments represents a crossing norm of reaction in that relative warp scores for
the large and small ends of the ontogenetic trajectory alternate in sign or magnitude between
high and low-velocity environments (warps 1–6; Fig 5). Visually, the change in shape over
ontogeny in the low-velocity environment involves a substantial reduction in head width and
an especially marked reduction in relative eye size (Fig 6, left side). In contrast, the change in
shape over ontogeny in the high-velocity environment involves an increase in head width and
a slight elongation of the head anterior of the eye (Fig 6, right side). Head shape differs between
high-velocity and low-velocity environments at both ends of the ontogenetic spectrum (Fig 5).
Smaller T.areolatus from the high-velocity environment had narrower heads, and smaller eyes
compared to small specimens from the low-velocity environment (Fig 6, upper); whereas,
larger T.areolatus from the high-velocity environment had broader and longer heads, com-
pared to specimens from the low-velocity environment (Fig 6, lower).
Discussion
Velocity environments differentially shape the ontogenetic trajectory of shape in T.areolatus.
Two patterns emerge from our data that inform how these two velocity environments differen-
tially affect ontogenetic shape trajectories. First, in both head and body, differences in shape
exist between environments at the smallest ontogenetic sizes. This pattern is similar to that
observed in B.rhabdophora between high and low-predation environments—even the smallest
individuals were divergent in shape between environments [51]. Development of shape is
highly plastic in fishes [88], and induced effects from hatchery environments are especially
well documented [89]. In cases where fish shape is induced by environment, shape typically
does not differ at the earliest stages of ontogenetic development (before environmental effects
Fig 4. Visualization of variation in body shape. Thin plate spline deformations of divergence in body shape between high-velocity and low-velocity
environments for large ontogenetic stage and small ontogenetic stage. Vertical arrows represent the direction of ontogeneticshape change from small to large.
Horizontal arrows represent divergence between velocity environments.
https://doi.org/10.1371/journal.pone.0252780.g004
PLOS ONE
Shape variation in Trichomycterus areolatus
PLOS ONE | https://doi.org/10.1371/journal.pone.0252780 June 11, 2021 9 / 17
are manifest), but rather diverges as a consequence of time spent in the divergent environ-
ments [88]. This is not consistent with our data. However, it could be that important shape
divergence between environments occurs at the larval stage or the larval to juvenile transition
[89] and is thus already manifest by the smallest stage at which we sampled.
The second important pattern is that the magnitude of overall shape change in both body
and head is substantially lower in fish from the high-velocity environment compared to fish
from the low-velocity environment. This suggests that shape at all stages may be more con-
strained in the high-velocity environment. Higher water velocities, along with other potentially
selective effects, produce more convergent forms relative to environments with lower water
velocities [90]. Where interactive effects have been identified in determining shape of fishes,
one environmental effect is often more constraining than the other, thus creating conditions
for significant interactions [17]. For example, Utah chub (Gila atraria) exhibit less variation in
shape in response to diet in high predation environments, compared to low predation environ-
ments, suggesting that predation constrains fish shape more than diet [17]. Similarly, female B.
rhabdophora from high predation and low predation environments converge in shape as
adults, apparently in response to the demands of pregnancy on shape. Thus, pregnancy exerts
a stronger effect on shape than does predation in females of this species [51]. In T.areolatus,
high-velocity environments appear to constrain shape to a greater extent than low-velocity
environments.
Multiple variables differ between velocity environments (i.e., stream zones [27]), and it is
not clear from our data which of these variables are most important in determining shape vari-
ation over ontogeny. However, most studies that report differences in shape of stream fishes
based on differences in environments focus on variation in stream velocity or surrogate
Fig 5. Relative warp least squares means for head. Least squares means (error bars represent 95% confidence
intervals of the mean) for each of 9 relative warps from the multivariate linear mixed model of head shape variation of
T.areolatus. The high-velocity environment is represented by closed circle symbols and solid connecting lines, and the
low-velocity environment is represented by open circle symbols and dashed connecting lines. Small and large
ontogenetic stages are represented by centroid sizes of 5.00 and 20.00, respectively.
https://doi.org/10.1371/journal.pone.0252780.g005
PLOS ONE
Shape variation in Trichomycterus areolatus
PLOS ONE | https://doi.org/10.1371/journal.pone.0252780 June 11, 2021 10 / 17
variables such as stream gradient or flow rate [28–32]. Effects of water velocity are well docu-
mented [29–33]. Generally, narrower body and head shapes are favored in high-velocity envi-
ronments which require steady swimming, whereas robust body and head shapes are favored
in low-velocity environments which require unsteady swimming [28,91]. Our results were
consistent with this pattern across the ontogenetic stages we sampled. T.areolatus sampled
from high-velocity environments had narrower bodies and heads than T.areolatus sampled
from low-velocity environments, and the general shape of body and head changed relatively
little in the high-velocity environments. Overall, water velocity seems to be a likely constraint
on variation in shape in the high-velocity environment, but water velocity appears to be rela-
tively less of a constraint in its effect on shape in the low-velocity environment.
Generally, juvenile fish have larger heads (relative to body size) and narrower bodies,
whereas adults have smaller heads (relative to body size) and more robust bodies [92–95].
Although proportional size of the head in T.areolatus changes over ontogeny consistent with
the pattern observed in other studies, body shape does not become more robust with increas-
ing size. This pattern of narrow body shape in adults may be associated with the benthic nature
of T.areolatus, and changes in habitat use over ontogeny. Body shape of T.areolatus is conver-
gent with body shape of other taxa that inhabit the benthic environment in high-velocity sys-
tems [90]. It may be that species that inhabit the hyporheic environment in high-velocity
environments are constrained in shape such that more robust body forms may be selected
against because of the small size of openings available in the substrate. A comparison of onto-
genetic shape trajectories among these convergent species may lend support to this idea [90].
The presence of shape variation between velocity environments within the Andalie
´n River
is surprising, given the relatively small size of the river (48 km) and the apparent lack of
Fig 6. Visualization of variation in head shape. Thin plate spline deformations of divergence in head shape between high-velocity and low-velocity environments
for large ontogenetic stage and small ontogenetic stage. Vertical arrows represent the direction of ontogenetic shape change from small to large. Horizontal arrows
represent divergence between velocity environments.
https://doi.org/10.1371/journal.pone.0252780.g006
PLOS ONE
Shape variation in Trichomycterus areolatus
PLOS ONE | https://doi.org/10.1371/journal.pone.0252780 June 11, 2021 11 / 17
geographic or reproductive barriers to gene flow (i.e., no waterfalls or other flow barriers that
would preclude movement between velocity environments). With no evidence of genetic isola-
tion between velocity environments, the observed morphometric variation is unlikely to be
based on underlying genetic differences. It is more likely an example of phenotypic plasticity
in response to environmental conditions [24,25], but much of the induced difference must
occur during the larval to juvenile transition at smaller sizes than we sampled. A common-gar-
den experiment covering all sizes (larvae to adult) would be needed to directly address this
question [96].
Acknowledgments
We thank the undergraduate and graduate students at Brigham Young University and the
Universidad de Concepcio
´n that helped collect, measure, and photograph the fish. We thank
Jillian Campbell for landmarking the photos and preparing the data for analysis.
Author Contributions
Conceptualization: Evelyn Habit, Mark C. Belk.
Data curation: Peter C. Searle, Mark C. Belk.
Formal analysis: Peter C. Searle, Mark C. Belk.
Funding acquisition: Evelyn Habit, Mark C. Belk.
Investigation: Evelyn Habit, Mark C. Belk.
Methodology: Evelyn Habit, Mark C. Belk.
Project administration: Evelyn Habit, Mark C. Belk.
Resources: Evelyn Habit.
Supervision: Mark C. Belk.
Validation: Mark C. Belk.
Visualization: Peter C. Searle, Margaret Mercer, Mark C. Belk.
Writing – original draft: Peter C. Searle, Margaret Mercer.
Writing – review & editing: Peter C. Searle, Margaret Mercer, Evelyn Habit, Mark C. Belk.
References
1. Casatti L, Castro RMC. Testing the ecomorphological hypothesis in a headwater riffles fish assemblage
of the Rio Sao Francisco, Southeastern Brazil. Neotrop Ichthyol. 2006; 4(2):203–14. https://doi.org/10.
1590/s1679-62252006000200006 WOS:000239040600006.
2. Bock WJ. Concepts and methods in ecomorphology. J Biosci. 1994; 19(4):403–13. https://doi.org/10.
1007/bf02703177 WOS:A1994PW58500006.
3. Bhat A. Ecomorphological correlates in tropical stream fishes of southern India. Environmental Biology
of Fishes. 2005; 73(2):211–25. https://doi.org/10.1007/s10641-005-0561-0
4. Motta P, Kotrschal K. Correlative, experimental, and comparative evolutionary approaches in ecomor-
phology. Netherlands Journal of Zoology. 1992; 42(2–3):400–15.
5. Cattau CE, Fletcher RJ, Kimball RT, Miller CW, Kitchens WM. Rapid morphological change of a top
predator with the invasion of a novel prey. Nature Ecology & Evolution. 2018; 2(1):108–15. https://doi.
org/10.1038/s41559-017-0378-1 WOS:000426514800018. PMID: 29180705
6. Marchand LJ, Tarayre M, Dorey T, Rantier Y, Hennion F. Morphological variability ofcushion plant Lyal-
lia kerguelensis (Caryophyllales) in relation to environmental conditions and geography in the Kergue-
len Islands: implications for cushion necrosis and climate change. Polar Biol. 2021; 44(1):17–30. https://
doi.org/10.1007/s00300-020-02768-2
PLOS ONE
Shape variation in Trichomycterus areolatus
PLOS ONE | https://doi.org/10.1371/journal.pone.0252780 June 11, 2021 12 / 17
7. Pergams ORW, Lawler JJ. Recent and Widespread Rapid Morphological Change in Rodents. PLOS
ONE. 2009; 4(7):e6452. https://doi.org/10.1371/journal.pone.0006452 PMID: 19649284
8. Chivers DP, Zhao XX, Brown GE, Marchant TA, Ferrari MCO. Predator-induced changes in morphology
of a prey fish: The effects of food level and temporal frequency of predation risk. Evol Ecol. 2008; 22
(4):561–74. https://doi.org/10.1007/s10682-007-9182-8 WOS:000256963300005.
9. Hugueny B, Pouilly M. Morphological correlates of diet in an assemblage of West African freshwater
fishes. Journal of Fish Biology. 1999; 54(6):1310–25. https://doi.org/10.1111/j.1095-8649.1999.
tb02057.x
10. Webb PW. Body Form, Locomotion and Foraging in Aquatic Vertebrates. American Zoologist. 1984; 24
(1):107–20. https://doi.org/10.1093/icb/24.1.107
11. Phillips BL, Shine R. Adapting to an invasive species: Toxic cane toads induce morphological change in
Australian snakes. Proceedings of the National Academy of Sciences of the United States of America.
2004; 101(49):17150–5. https://doi.org/10.1073/pnas.0406440101 PMID: 15569943
12. Blasina G, Molina J, Cazorla AL, de Astarloa JD. Relationship between ecomorphology and trophic seg-
regation in four closely related sympatric fish species (Teleostei, Sciaenidae). C R Biol. 2016; 339(11–
12):498–506. https://doi.org/10.1016/j.crvi.2016.07.003 WOS:000387441300007. PMID: 27562196
13. Pakkasmaa S, Piironen J. Water velocity shapes juvenile salmonids. Evol Ecol. 2000; 14(8):721–30.
https://doi.org/10.1023/a:1011691810801 WOS:000170783200004.
14. Hendry AP, Taylor EB, McPhail JD. Adaptive divergence and the balance between selection and gene
flow: Lake and stream stickleback in the misty system. Evolution. 2002; 56(6):1199–216.
WOS:000176820900011. https://doi.org/10.1111/j.0014-3820.2002.tb01432.x PMID: 12144020
15. Meyers PJ, Belk MC. Shape variation in a benthic stream fish across flow regimes. Hydrobiologia.
2014; 738(1):147–54. https://doi.org/10.1007/s10750-014-1926-1 WOS:000340479400011.
16. Mercer M, Searle PC, Cifuentes R, Habit E, Belk MC. Morphometric Response of Galaxias maculatus
(Jenyns) to Lake Colonization in Chile. Diversity. 2020; 12(6):13. https://doi.org/10.3390/d12060219
WOS:000551690100001.
17. Williams TJ, Johnson JB, Belk MC. Interaction between predation environment and diet constrains
body shape in Utah chub, Gila atraria (Cypriniformes: Cyprinidae). Biol J Linnean Soc. 2017; 122
(1):147–56. https://doi.org/10.1093/biolinnean/blx050 WOS:000409213700011.
18. Taylor-Cox ED, Macgregor CJ, Corthine A, Hill JK, Hodgson JA, Saccheri IJ. Wing morphological
responses to latitude and colonisation in a range expanding butterfly. PeerJ. 2020; 8:29. https://doi.org/
10.7717/peerj.10352 WOS:000590404900004. PMID: 33240660
19. Kaliontzopoulou A, Carretero MA, Llorente GA. Intraspecific ecomorphological variation: linear and geo-
metric morphometrics reveal habitat-related patterns within Podarcis bocagei wall lizards. J Evol Biol.
2010; 23(6):1234–44. https://doi.org/10.1111/j.1420-9101.2010.01984.x WOS:000277710100011.
PMID: 20406345
20. Ofori SA, Hagan JK, Kyei F. Morphometric characterization and differentiation of West African Dwarf
goat populations in Ghana. Tropical Animal Health and Production. 2021; 53(1):69. https://doi.org/10.
1007/s11250-020-02498-x PMID: 33398459
21. Kern EMA, Langerhans RB. Urbanization Alters Swimming Performance of a Stream Fish. Frontiers in
Ecology and Evolution. 2019; 6(229). https://doi.org/10.3389/fevo.2018.00229
22. Stearns S. The Evolution of Life Histories: New York: Oxford University Press; 1992.
23. Nonaka E, Svanback R, Thibert-Plante X, Englund G, Brannstrom A. Mechanisms by which phenotypic
plasticity affects adaptive divergence and ecological speciation. Am Nat. 2015; 186(5):E126–E43.
https://doi.org/10.1086/683231 WOS:000363928900003. PMID: 26655782
24. Fitzpatrick BM. Underappreciated consequences of phenotypic plasticity for ecological speciation. Inter-
national Journal of Ecology 2012; 2012:1–12.
25. West-Eberhard MJ. Phenotypic plasticity and the origins of diversity. Annu Rev Ecol Syst. 1989;
20:249–78. https://doi.org/10.1146/annurev.es.20.110189.001341 WOS:A1989CA83100011.
26. Vannote RL, Minshall GW, Cummins KW, Sedell JR, Cushing CE. The river continuum concept. Can J
Fish Aquat Sci. 1980; 37(1):130–7.
27. Habit E, Belk M, Victoriano P, Jaque E. Spatio-temporal distribution patterns and conservation of fish
assemblages in a Chilean coastal river. Biodiversity and Conservation. 2007; 16(11):3179–91. https://
doi.org/10.1007/s10531-007-9171-9 WOS:000248910700011.
28. Langerhans RB. Predictability of phenotypic differentiation across flow regimes in fishes. Integrative
and Comparative Biology. 2008; 48(6):750–68. https://doi.org/10.1093/icb/icn092
WOS:000262760700006. PMID: 21669830
PLOS ONE
Shape variation in Trichomycterus areolatus
PLOS ONE | https://doi.org/10.1371/journal.pone.0252780 June 11, 2021 13 / 17
29. Brinsmead J, Fox MG. Morphological variation between lake- and stream-dwelling rock bassand pump-
kinseed populations. Journal of Fish Biology. 2002; 61(6):1619–38. https://doi.org/10.1006/jfbi.2002.
2179 WOS:000181220000023.
30. Sidlauskas B, Chernoff B, Machado-Allison A. Geographic and environmental variation in Bryconops
sp. cf. melanurus (Ostariophysi: Characidae) from the Brazilian Pantanal. Ichthyological Research.
2006; 53(1):24–33. https://doi.org/10.1007/s10228-005-0310-6
31. Langerhans RB, Layman CA, Langerhans AK, Dewitt TJ. Habitat-associated morphological divergence
in two Neotropical fish species. Biol J Linnean Soc. 2003; 80(4):689–98. https://doi.org/10.1111/j.1095-
8312.2003.00266.x WOS:000187247200010.
32. Langerhans RB, Chapman LJ, Dewitt TJ. Complex phenotype-environment associations revealed in an
East African cyprinid. J Evol Biol. 2007; 20(3):1171–81. https://doi.org/10.1111/j.1420-9101.2007.
01282.x WOS:000245989600033. PMID: 17465926
33. McLaughlin RL, Grant JWA. Morphological and behavioural differences among recently-emerged brook
charr, Salvelinus fontinalis, foraging in slow- vs. fast-running water. Environmental Biology of Fishes.
1994; 39(3):289–300. https://doi.org/10.1007/bf00005130 WOS:A1994NA68400007.
34. Strelin MM, Cosacov A, Chalcoff VR, Maubecin CC, Se
´rsic AN, Benitez-Vieyra SM. The role of ontoge-
netic allometry and nonallometric flower shape variation in species-level adaptive diversification–Cal-
ceolaria polyrhiza (Calceolariaceae) as a case study. Evol Dev. 2020;n/a(n/a):e12363. https://doi.org/
10.1111/ede.12363 PMID: 33372721
35. Lo
´pez-Romero FA, Klimpfinger C, Tanaka S, Kriwet J. Growth trajectories of prenatal embryos of the
deep-sea shark Chlamydoselachus anguineus (Chondrichthyes). Journal of Fish Biology. 2020; 97
(1):212–24. https://doi.org/10.1111/jfb.14352 PMID: 32307702
36. Thompson DW. On growth and form: Cambridge Univ. Press; 1942. 116 pp. p.
37. Ocampo EH, Spivak ED, Baeza JA, Luppi TA. Ontogenetic changes in the external anatomy of the par-
asitic castrator crab Calyptraeotheres garthi: implications for the timing of host colonization and sexual
behaviour. Biol J Linnean Soc. 2017; 120(1):54–74. WOS:000400943400004.
38. Gottardo M, Vallotto D, Beutel RG. Giant stick insects reveal unique ontogenetic changes in biological
attachment devices. Arthropod Structure & Development. 2015; 44(2):195–9. https://doi.org/10.1016/j.
asd.2015.01.001 PMID: 25601633
39. Monteiro LR, Cavalcanti MJ, Sommer HJS III. Comparative ontogenetic shape changes in the skull of
Caiman species (Crocodylia, Alligatoridae). Journal of Morphology. 1997; 231(1):53–62. https://doi.org/
10.1002/(SICI)1097-4687(199701)231:1<53::AID-JMOR5>3.0.CO;2-P. PMID: 29852629
40. de la Hera I, Pulido F, Visser ME. Longitudinal data reveal ontogenetic changes in the wing morphology
of a long-distance migratory bird. Ibis. 2014; 156(1):209–14. https://doi.org/10.1111/ibi.12112.
41. Hernandez G, Garcia S, Vilela JF, de la Sancha NU. Ontogenetic variation of an omnivorous generalist
rodent: the case of the montane akodont (Akodon montensis). J Mammal. 2017; 98(6):1741–52. https://
doi.org/10.1093/jmammal/gyx135 WOS:000417241500023.
42. Claessen D, Dieckmann U. Ontogenetic niche shifts and evolutionary branching in size-structured pop-
ulations. Evolutionary Ecology Research. 2002; 4:189–217.
43. Lima-Junior SE, Goitein R. Ontogenetic diet shifts of a Neotropical catfish, Pimelodus maculatus (Siluri-
formes, Pimelodidae): An ecomorphological approach. Environmental Biology of Fishes. 2003; 68
(1):73–9. https://doi.org/10.1023/a:1026079011647 WOS:000185646800009.
44. Rodriguez-Mendoza R, Munoz M, Saborido-Rey F. Ontogenetic allometry of the bluemouth, Helicole-
nus dactylopterus dactylopterus (Teleostei: Scorpaenidae), in the Northeast Atlantic and Mediterranean
based on geometric morphometrics. Hydrobiologia. 2011; 670(1):5–22. https://doi.org/10.1007/s10750-
011-0675-7 WOS:000291222700002.
45. Baliga VB, Mehta RS. Ontogenetic allometry in shape and flexibility underlies life history patterns of lab-
rid cleaning behavior. Integrative and Comparative Biology. 2016; 56(3):416–27. https://doi.org/10.
1093/icb/icw028 WOS:000381708600005. PMID: 27252204
46. Gould SJ. Allometry and size in ontogeny and phylogeny. Biol Rev. 1966; 41(4):587–638. https://doi.
org/10.1111/j.1469-185x.1966.tb01624.x PMID: 5342162
47. Ahnelt H, Sauberer M, Ramler D, Koch L, Pogoreutz C. Negative allometric growth during ontogeny in
the large pelagic filter-feeding basking shark. Zoomorphology. 2020; 139(1):71–83. https://doi.org/10.
1007/s00435-019-00464-2 WOS:000514996200006.
48. Merigoux S, Ponton D. Body shape, diet and ontogenetic diet shifts in young fish of the Sinnamary
River, French Guiana, South America. Journal of Fish Biology. 1998; 52(3):556–69. https://doi.org/10.
1111/j.1095-8649.1998.tb02017.x WOS:000072711500009.
49. Shea BT. Ontogenetic scaling of skeletal proportions in the talapoin monkey. J HumEvol. 1992; 23
(3):283–307. https://doi.org/10.1016/S0047-2484(05)80004-1.
PLOS ONE
Shape variation in Trichomycterus areolatus
PLOS ONE | https://doi.org/10.1371/journal.pone.0252780 June 11, 2021 14 / 17
50. Gru¨nbaum T, Cloutier R, Mabee PM, Le Franc¸ois NR. Early developmental plasticity and integrative
responses in arctic charr (Salvelinus alpinus): effects of water velocity on body size and shape. Journal
of Experimental Zoology Part B: Molecular and Developmental Evolution. 2007; 308B(4):396–408.
https://doi.org/10.1002/jez.b.21163 PMID: 17358017
51. Hassell EMA, Meyers PJ, Billman EJ, Rasmussen JE, Belk MC. Ontogeny and sex alter the effect of
predation on body shape in a livebearing fish: Sexual dimorphism, parallelism, and costs of reproduc-
tion. Ecology and Evolution. 2012; 2(7):1738–46. https://doi.org/10.1002/ece3.278 PMID: 22957177
52. Pistore AE, Barry TN, Vanderzwan SL, Schutz H, Rogers SM, Jamniczky HA. The role of genetic and
environmental background driving ontogenetic trajectories of skeletal variation in the threespine stickle-
back (Gasterosteus aculeatus L.). Evolutionary Ecology Research. 2019; 20(1):27–50.
53. Parsons KJ, Sheets HD, Skulason S, Ferguson MM. Phenotypic plasticity, heterochrony and ontoge-
netic repatterning during juvenile development of divergent Arctic charr (Salvelinus alpinus). J Evol Biol.
2011; 24(8):1640–52. https://doi.org/10.1111/j.1420-9101.2011.02301.x WOS:000292699200002.
PMID: 21599773
54. Svanba
¨ck R, Eklo
¨v P. Effects of habitat and food resources on morphology and ontogenetic growth tra-
jectories in perch. Oecologia. 2002; 131(1):61–70. https://doi.org/10.1007/s00442-001-0861-9 PMID:
28547511
55. Landaeta MF, Bernal-Duran V, Castillo MI, Diaz-Astudillo M, Fernandez-General B, Nunez-Acuna P.
Nearshore environmental conditions influence larval growth and shape changes for a temperate rocky
reef fish. Hydrobiologia. 2019; 839(1):159–76. https://doi.org/10.1007/s10750-019-04004-3
WOS:000476499300011.
56. Arnqvist G, Johansson F. Ontogenetic reaction norms of predator-induced defensive morphology in
dragonfly larvae. Ecology. 1998; 79(6):1847–58. https://doi.org/10.1890/0012-9658(1998)079[1847:
ornopi]2.0.co;2 WOS:000075752700001.
57. Hjelm J, Svanba
¨ck R, Bystro
¨m P, Persson L, Wahlstro
¨m E. Diet-dependent body morphology and onto-
genetic reaction norms in Eurasian perch. Oikos. 2001; 95(2):311–23.
58. Dunn NR, O’Brien LK, Closs GP. Phenotypically induced intraspecific variation in the morphological
development of wetland and stream Galaxias gollumoides McDowall and Chadderton. Diversity. 2020;
12(6):19. https://doi.org/10.3390/d12060220 WOS:000551688400001.
59. Chara JD, Baird DJ, Telfer TC, Rubio EA. Feeding ecology and habitat preferences of the catfish genus
Trichomycterus in low-order streams of the Colombian Andes. Journal of Fish Biology. 2006; 68
(4):1026–40. https://doi.org/10.1111/j.0022-1112.2006.00984.x WOS:000236718200004.
60. Scott S, Pardo R, Vila I. Trophic niche overlap between two Chilean endemic species of Trichomycterus
(Teleostei: Siluriformes). Revista Chilena De Historia Natural. 2007; 80(4):431–7.
WOS:000252317900004.
61. Habit E, Belk MC, Tuckfield RC, Parra O. Response of the fish community to human-induced changes
in the Biobio River in Chile. Freshwater Biology. 2006; 51(1):1–11. https://doi.org/10.1111/j.1365-2427.
2005.01461.x WOS:000233834400001.
62. Rohlf JF, Marcus LF. A revolution in morphometrics. Trends in Ecology & Evolution. 1993; 8(4):129–32.
https://doi.org/10.1016/0169-5347(93)90024-J
63. Kern EMA, Langerhans RB. Urbanization drives contemporary evolution in stream fish. Global Change
Biology. 2018; 24(8):3791–803. https://doi.org/10.1111/gcb.14115 PMID: 29700897
64. Johnson JB, Belk MC. Predation environment predicts divergent life-history phenotypes among popula-
tions of the livebearing fish Brachyrhaphis rhabdophora. Oecologia. 2001; 126(1):142–9. https://doi.
org/10.1007/s004420000504 WOS:000166243500016. PMID: 28547433
65. Nickum JG. Guidelines for Use of Fishes in Field Research. Fisheries. 1988; 13(2):1–62. https://doi.org/
10.1577/1548-8446-13-2
66. Manriquez A, Huaquin L, Arellano M, Arratia G. Reproductive aspects of Trichomycerus areolatus
Valenciennes, 1846 (Pisces, Teleostei, Siluriformies) in Angostura River, Chile. Stud Neotrop Fauna
Environ. 1988; 23(2):89–102. https://doi.org/10.1080/01650528809360749 WOS:A1988Q403800004.
67. Bookstein FL. Morphometric tools for landmark data: Geometry and biology: Cambridge University
Press; 1991.
68. Rohlf FJ. tpsDig: Digitize landmarks & outlines from image files, scanner, or video. Stony Brook, NY:
Department of Ecology and Evolution, State University of New York at Stony Brook; 2003.
69. Rohlf FJ, Loy A, Corti M. Morphometric Analysis of Old World Talpidae (Mammalia, Insectivora) Using
Partial-Warp Scores. Syst Biol. 1996; 45(3):344–62. https://doi.org/10.1093/sysbio/45.3.344
70. Rohlf FJ. tpsRelw32: Relative warps analysis. Stony Brook, NY: Department of Ecology and Evolution,
State University of New York at Stony Brook; 2003.
PLOS ONE
Shape variation in Trichomycterus areolatus
PLOS ONE | https://doi.org/10.1371/journal.pone.0252780 June 11, 2021 15 / 17
71. Rohlf FJ, Slice D. Extension of the procrustes method for the optimal superimposition of landmarks.
Systematic Zoology. 1990; 39(1):40–59. https://doi.org/10.2307/2992207 WOS:A1990DP85800004.
72. Rohlf FJ. The tps series of software. Hystrix, the Italian Journal of Mammalogy. 2015; 26(1):9–12.
https://doi.org/10.4404/hystrix-26.1–11264
73. Bookstein FL. Principal warps—Thin-plate splines and the decomposition of deformations. IEEE Trans
Pattern Anal Mach Intell. 1989; 11(6):567–85. https://doi.org/10.1109/34.24792 WOS:
A1989U674900003.
74. Roth-Monzo
´n AJ, Belk MC, Zuniga-Vega JJ, Johnson JB. Beyond pairwise interactions: Multispecies
character displacement in Mexican freshwater fish communities. Am Nat. 2020; 195(6):983–96. https://
doi.org/10.1086/708513 WOS:000537567400006. PMID: 32469659
75. Ingley SJ, Billman EJ, Belk MC, Johnson JB. Morphological divergence driven by predation environ-
ment within and between species of Brachyrhaphis fishes. PLoS ONE. 2014; 9(2):11. https://doi.org/10.
1371/journal.pone.0090274 WOS:000332389000147. PMID: 24587309
76. Heins DC, Baker JA. Stream-flow environment predicts divergent life history phenotypes among popu-
lations of the Blacktail Shiner Cyprinella venusta: Temporal stability of a large-scale pattern. Ecology of
Freshwater Fish. 2018; 27(1):453–9. https://doi.org/10.1111/eff.12360 WOS:000417191100040.
77. Wesner JS, Billman EJ, Meier A, Belk MC. Morphological convergence during pregnancy among preda-
tor and nonpredator populations of the livebearing fish Brachyrhaphis rhabdophora (Teleostei: Poecilii-
dae). Biol J Linnean Soc. 2011; 104(2):386–92. https://doi.org/10.1111/j.1095-8312.2011.01715.x
78. Kenward MG, Roger JH. Small sample inference for fixed effects from restricted maximum likelihood.
Biometrics. 1997; 53(3):983–97. https://doi.org/10.2307/2533558. PMID: 9333350
79. Adams DC, Collyer ML. Analysis of character divergence along environmental gradients and other
covariates. Evolution. 2007; 61(3):510–5. https://doi.org/10.1111/j.1558-5646.2007.00063.x PMID:
17348916
80. Collyer ML, Adams DC. Analysis of Two-State Multivariate Phenotypic Change in Ecological Studies.
Ecology. 2007; 88(3):683–92. https://doi.org/10.1890/06-0727 PMID: 17503596
81. Adams DC. Character displacement via aggressive interference in Appalachian Salamanders. Ecology.
2004; 85(10):2664–70. https://doi.org/10.1890/04-0648.
82. Butler DG, Cullis BR, Gilmour AR, Gogel BJ, R. T. ASReml-R Reference Manual Version 4. Hemel
Hempstead, United Kingdom: VSN International Ltd; 2018.
83. R Development Core Team. R: A language and environment for statistical computing. Vienna, Austria:
R Foundation for Statistical Computing; 2020.
84. Billman EJ, Rasmussen JE, Creighton JC, Johnson JB, Belk MC. A multivariate approach to the analy-
sis of within lifetime variation in life history. Methods Ecol Evol. 2014; 5(8):797–805. https://doi.org/10.
1111/2041-210X.12211.
85. Langerhans RB. Trade-off between steady and unsteady swimming underlies predator-driven diver-
gence in Gambusia affinis. J Evol Biol. 2009; 22(5):1057–75. https://doi.org/10.1111/j.1420-9101.2009.
01716.x PMID: 21462405
86. Langerhans RB, Makowicz AM. Shared and unique features of morphological differentiation between
predator regimes in Gambusia caymanensis. J Evol Biol. 2009; 22(11):2231–42. https://doi.org/10.
1111/j.1420-9101.2009.01839.x PMID: 20069725
87. Rohlf FJ. tpsRegr: Multivariate regression of shape onto one or more independent variables. Stony
Brook, NY: Department of Ecology and Evolution, State University of New York at Stony Brook; 2003.
88. Wimberger PH. Plasticity of fish body shape. The effects of diet, development, family and age in two
species of Geophagus (Pisces: Cichlidae). Biol J Linnean Soc. 1992; 45(3):197–218. https://doi.org/10.
1111/j.1095-8312.1992.tb00640.x
89. Belk MC, Benson LJ, Rasmussen J, Peck SL. Hatchery-induced morphological variation in an endan-
gered fish: a challenge for hatchery-based recovery efforts. Can J Fish Aquat Sci. 2008; 65(3):401–8.
https://doi.org/10.1139/f07-176
90. Helfman G, Collette BB, Facey DE, Bowen BW. Special habitats and special adaptations. The diversity
of fishes: Biology, evolution, and ecology: John Wiley & Sons; 2009.
91. Langerhans RB, Reznick DN. Ecology and evolution of swimming performance in fishes: Predicting
evolution with biomechanics. In: Domenici P, Kapoor BG, editors. Fish Locomotion: An Eco Ethological
Perspective. Enfield: Science Publisher; 2009. p. 200–48.
92. Bravi R, Ruffini M, Scalici M. Morphological variation in riverine cyprinids: a geometric morphometric
contribution. Ital J Zoolog. 2013; 80(4):536–46. https://doi.org/10.1080/11250003.2013.829129
WOS:000327469300010.
PLOS ONE
Shape variation in Trichomycterus areolatus
PLOS ONE | https://doi.org/10.1371/journal.pone.0252780 June 11, 2021 16 / 17
93. Bernal-Duran V, Landaeta MF. Feeding variations and shape changes of a temperate reef clingfish dur-
ing its early ontogeny. Sci Mar. 2017; 81(2):205–15. https://doi.org/10.3989/scimar.04555.09A
WOS:000403554400007.
94. Villares GA, Goitein R. Morphological aspects in the ontogeny of Salminus hilarii Valenciennes, 1850
(Ostaryophysi: Characidae). Braz J Biol. 2016; 76(4):905–11. https://doi.org/10.1590/1519-6984.04315
WOS:000387656700013. PMID: 27143050
95. Reis RE, Zelditch ML, Fink WL. Ontogenetic allometry of body shape in the Neotropical catfish Cal-
lichthys (Teleostei: Siluriformes). Copeia. 1998;(1):177–82. https://doi.org/10.2307/1447715
WOS:000071900500020.
96. Rader RB, Belk MC, Shiozawa DK, Crandall KA. Empirical tests for ecological exchangeability. Animal
Conservation. 2005; 8(3):239–47. Epub 2005/08/26. https://doi.org/10.1017/S1367943005002271
PLOS ONE
Shape variation in Trichomycterus areolatus
PLOS ONE | https://doi.org/10.1371/journal.pone.0252780 June 11, 2021 17 / 17
Available via license: CC BY 4.0
Content may be subject to copyright.
