Morphology, Diet, and Population Structure of the Southern White-lipped Mud Turtle Kinosternon leucostomum postinguinale (Testudines: Kinosternidae) in the Nus River Drainage, Colombia

Abstract

Most existing studies on the White-lipped Mud Turtle, Kinosternon leucostomum, have been based on northern Central American populations, leaving a lack of information on populations from southern Central America and South America. Herein we studied morphology, diet, and population structure of a population of the southern Kinosternon leucostomum postinguinale inhabiting four creeks in Colombia. Observed habitats used were highly variable, ranging from relatively clean waters to streams used for sewage disposal of wastewater from a human settlement. Body size was smaller than that of other populations of southern K. l. postinguinale and also than that of the northern K. l. leucostomum. Sexual dimorphism was evident, with males heavier, longer, and wider than females. Body size was associated with the habitat of origin, with Barrio Nuevo individuals being the largest. The main components of the diet were plant material, insects, snails, and algae. We did not find evidence of sexual differences in the diet, but we found geographic differences in the body size. The population with the largest individuals, from Barrio Nuevo Creek, consumed more snails while those from Totumo Creek, the population with the smallest individuals, consumed more ants and plant material as compared to the other creeks. Additionally, we found a highly male-biased sex ratio, with 2.5 adult males per female, very few juveniles, and no nests, which suggests a dangerous risk of population decline. We suggest continued monitoring of the demography of this population, emphasizing its reproductive biology.

Full text

Journal of Herpetology, Vol. 0, No. 0, 000–000, 0000
Copyright 2016 Society for the Study of Amphibians and Reptiles
Morphology, Diet, and Population Structure of the Southern White-lipped Mud Turtle
Kinosternon leucostomum postinguinale (Testudines: Kinosternidae) in the Nus River
Drainage, Colombia
CLAUDIA P. C EBALLOS,
1,2
DANIEL ZAPATA,
1
CAROLINA ALVARADO,
1
AND EISINHOWER RINC ´
ON
3
1
Grupo Centauro, Escuela de Medicina Veterinaria, Facultad de Ciencias Agrarias, Universidad de Antioquia, Medellı´n, Colombia
3
Escuela Ambiental, Facultad de Ingenieria, Universidad de Antioquia, Puerto Berrı´o, Colombia
ABSTRACT.—Most existing studies on the White-lipped Mud Turtle, Kinosternon leucostomum, have been based on northern Central
American populations, leaving a lack of information on populations from southern Central America and South America. Herein we
studied morphology, diet, and population structure of a population of the southern Kinosternon leucostomum postinguinale inhabiting
four creeks in Colombia. Observed habitats used were highly variable, ranging from relatively clean waters to streams used for sewage
disposal of wastewater from a human settlement. Body size was smaller than that of other populations of southern K. l. postinguinale and
also than that of the northern K. l. leucostomum. Sexual dimorphism was evident, with males heavier, longer, and wider than females.
Body size was associated with the habitat of origin, with Barrio Nuevo individuals being the largest. The main components of the diet
were plant material, insects, snails, and algae. We did not find evidence of sexual differences in the diet, but we found geographic
differences in the body size. The population with the largest individuals, from Barrio Nuevo Creek, consumed more snails while those
from Totumo Creek, the population with the smallest individuals, consumed more ants and plant material as compared to the other
creeks. Additionally, we found a highly male-biased sex ratio, with 2.5 adult males per female, very few juveniles, and no nests, which
suggests a dangerous risk of population decline. We suggest continued monitoring of the demography of this population, emphasizing
its reproductive biology.
Resumen.—La informaci ´
on existente sobre la Tortuga Tapaculo, Kinosternon leucostomum, ha sido generalmente basada en las
poblaciones del norte de Centro Ame´rica, dejando un desconocimiento sobre las poblaciones del sur de Centro Ame´rica y Sur Ame´rica.
En este trabajo nosotros documentamos la morfologı´a, dieta, y estructura de una poblaci ´
on de la subespecie el sur K. l. postinguinale que
habita en 4 quebradas en Colombia. Los ha´bitats observados fueron muy variables, desde quebradas relativamente limpias hasta
quebradas que reciben aguas negras de asentamientos humanos. El taman˜o corporal fue menor a lo reportado por otras poblaciones de la
misma subespecie e incluso de la subespecie del norte, K. l. leucostomum. El dimorfismo sexual fue evidente, con machos ma´s pesados,
largos, y anchos que las hembras. El taman˜ o corporal estuvo asociado al ha´bitat de origen, siendo las tortugas de Barrio Nuevo las ma´s
grandes. La dieta estuvo compuesta de material vegetal, insectos, caracoles, y algas. Nosotros no encontramos diferencias sexuales en la
dieta pero si diferencias geogra´ficas en el taman˜ o. Las tortugas de Barrio Nuevo que fueron las ma´ s grandes consumieron ma´s caracoles,
mientras que las tortugas ma´ s pequen˜ as de El Totumo consumieron ma´s plantas y hormigas en compariaci ´
on con los otras quebradas.
Adicionalmente, encontramos una proporci ´
on sexual sesgada hacia los machos, con 2.5 machos por hembra, muy pocos juveniles y
ning ´
un nido, lo que sugiere un riesgo peligroso de disminuci ´
on poblacional. Recomendamos un monitoreo continuo de la demografı´a de
esta poblaci ´
on y sobre su biologı´a reproductiva en particular.
Within the order Testudines is the Neotropical family
Kinosternidae, comprised of 26 species and 38 subspecies
distributed from Canada to Argentina (van Dijk et al., 2014).
The traditional phylogeny of Kinosternidae has been challenged
by a recently proposed phylogenetic hypothesis (Iverson et al.,
2013) in which some species were reallocated to a new genus
and some subspecies were supported as distinct species. Other
authors have objected to these new phylogenetic relationships
(Spinks et al., 2014). Nevertheless, this debate highlights the
need for further genetic and biologic studies covering the widest
possible distributional range of all entities.
Within this family is the White-lipped Mud Turtle, Kinoster-
non leucostomum (Dume´ril and Bibr ´
on, in Dume´ril and Dume´ril
1851), with its two subspecies: the Northern White-lipped Mud
Tur tle , Kinosternon leucostomum leucostomum,rangingfrom
Me´xico to Nicaragua, and the Southern White-lipped Mud
Turtle , Kinosternon leucostomum postinguinale, ranging from
Nicaragua to Peru (van Dijk et al., 2014). This species is not
currently assessed by the International Union for Conservation
of Nature (IUCN, 2015) but it was listed as Least Concern in
1996 (van Dijk et al., 2014). Turtle specialists, however, have
suggested listing it as Least Concern (van Dijk et al., 2014). In
Colombia, K. leucostomum faces numerous problems including
habitat degradation, human consumption, exploitation for the
crafting of ornaments, and as pets (Giraldo et al., 2012).
In general, K. leucostomum is recognized for its unicarinate
carapace, plastron with two hinges, and head with a yellow to
brown stripe that extends from the eye to the neck (Berry and
Iverson, 2001; Giraldo et al., 2012). Kinosternon leucostomum
exhibits sexual dimorphism (e.g., males are recognized by a
longer and thicker tail and the presence of a patch of horny
scales on the inner surface of the hind limbs; Berry and Iverson,
2001). Their habitat includes slow waters with soft substrates
and abundant vegetation (Ernst et al., 1997) but also creeks with
abundant rocks and fast-flowing water (Acun˜ a-Mese´n, 1993).
Turtles may move 2–3 m away from their aquatic habitat and
between 50–200 m parallel to the shoreline, though they can
move up to 600 m away to nest or aestivate (Morales-Verdeja
and Vogt, 1997). This species is omnivorous with its diet
including snails, grasses, algae, aquatic insects, and seeds
consumed in an opportunistic pattern (depending on the
habitat quality or displacement in the feeding niche by
coexisting turtle species; Vogt and Guzman, 1988; Acun˜a-
Mese´n, 1993). Most of the studies of their ecology have been
2
Corresponding Author. E-mail: claudiaceb@gmail.com
DOI: 10.1670/15-035

conducted on the northern subspecies, with fewer on the
southern subspecies and none in Colombia (Giraldo et al., 2012).
Given the lack of information on the southern subspecies, K. l.
postinguinale, we monitored a population inhabiting the
Corregimiento San Jose´ del Nus, Departmento de Antioquia,
Colombia. During four field trips in 2013–2014 we trapped
turtles in four creeks that drain into the Nus River. Herein we
report on the morphometrics of this subspecies and test for
differences associated with sex and creek of origin. We identify
and quantify the diet at each site from fecal samples and
quantify the population abundance, structure, and the sex
ratios. We also describe habitat quality by identifying and
quantifying the aquatic macroinvertebrates and some physical
parameters of each creek to discuss potential relationships
between habitat, morphology, and diet. In spite of its wide
distribution in Colombia, K. l. postinguinale populations have
not been studied, so we provide new information on the biology
of this southern subspecies. We expect this information will
contribute to the understanding of conservation status for this
species and provide insight for future management plans at the
local and national levels.
MATERIALS AND METHODS
Site Description.—We conducted this study in Corregimiento
San Jose´ del Nus, Municipio San Roque, Departmento de
Antioquia, Colombia. San Jose´ del Nus has an average
temperature of 23.38C; two rainy seasons, April–May and
October–November; and two dry seasons, January–March and
July–September. The main land use is cattle grazing (CORNARE,
2012).
Habitats.—We trapped turtles in four shallow creeks that drain
into the Nus River during four field trips of 4–7 days each: Barrio
Nuevo, Puente Tierra, Totumo, and Vega (Fig. 1). The Nus River
drains into the Magdalena River, a Caribbean drainage. Trips
were done in June, August, and October 2013 and in April 2014.
To characterize the four creeks, we measured four variables at
each site: water velocity (m/s), water depth (cm), transparency
(percentage of visibility of the water depth), and water
temperature (8C) recorded between 0700–0900 h. To determine
water velocity, we measured the time it took for a leaf to drift 5 m
downstream. To measure water depth, we used a pole and a
metric tape in the deepest point, where the traps were then set.
We used a Secchi disk to measure water transparency and a
thermometer submerged 30 cm for 1 min to measure tempera-
ture.
Because populations of the same species may have different
feeding habits, depending on habitat quality (Vogt and Guz-
man, 1988), we also surveyed the aquatic macroinvertebrates of
each creek (indicative of water quality). These organisms have
relatively long life cycles and are widely distributed and
sedentary; therefore, their presence is considered to correspond
with their tolerance of the water quality (Rolda´n-Pe´ rez, 1999;
Aguirre-Ramı´rez and Caicedo-Quintero, 2013). For this purpose
we collected all macroinvertebrates present in the water with a
nylon fish net during 1 h in a 10-m long transect along the
shoreline at each trapping site. We separated all collected
macroinvertebrates from organic material, stored them in 70%
ethanol, and transported them to the lab to be identified to the
lowest taxonomic level possible. We used a taxonomic key of the
most-commonly found macroinvertebrate families in Antioquia,
Colombia (Rolda´ n-Pe´ rez, 1988). We assigned a tolerance score to
each family and placed the sum of all scores into six habitat
quality categories: <15 indicated the water is critically polluted,
16–36 very polluted, 36–60 moderately polluted, 61–100 slightly
polluted, 101–150 clean, and >150 very clean (Rolda´ n-Pe´rez,
1999).
Turtle Sampling.—Because our objective was to trap the largest
number of individuals possible, we set 2, 1, 4, and 3 traps on
Barrio Nuevo, Puente Tierra, Totumo, and Vega creeks, respec-
tively, depending on water depth in each creek. We captured
turtles using hand-made hoop traps (175 cm long ·80 cm in
diameter) from 3-inch fishing-net mesh. We baited the traps with
small pieces of fresh pork viscera (lung, liver, heart, or spleen),
checked them every morning, and replaced the baits daily. All
captured turtles were taken to a nearby fish culture station,
property of the Universidad de Antioquia (682902300
N,
FIG. 1. Map of San Jose´ del Nus, Antioquia, Colombia, indicating sampling sites of Kinosternon leucostomum postinguinale in this study.
0 C. P. CEBALLOS ET AL.

7485902400
W) (Wildlife Conservation Society, Bogota´), for pro-
cessing.
We cleaned, weighed (g), measured, and marked turtles by
notching the carapace (Cagle, 1939). We also photographed
them with a digital camera (Nikon Coolpix P510) set on a tripod
to confirm identification when recaptured. We took linear and
curved measurements of the carapace and plastron along the
medial axis with a Brown & Sharpe 30 cm (12 inch) digital
caliper (Hexagon Manufacturing Intelligence, North Kingstown,
Rhode Island, U.S.A.) to the nearest 0.01 mm and with a metric
tape to the nearest 1 mm, respectively.
Diet.— As a defensive mechanism, turtles usually defecate
when being manipulated; therefore, feces were collected directly
from the cloacal orifice or from the plastic containers in which
they were temporally housed for 1–3 h. Each collected sample
was preserved in 70% ethanol to examine dietary content. All
fecal samples were refrigerated until their arrival at the Escuela
de Veterinaria at the Universidad de Antioquia. Once in the
Parasitology Lab, feces were separated into five categories: plant
material (leaves, stems, seeds), gastropods (snails), insects, algae,
and unidentified items. Individuals of Gastropoda were identi-
fied to the species level by a specialist from the Programa de
Estudio y Control de Enfermedades Tropicales (PECET) at the
Universidad de Antioquia, and insects were identified to the
family level following the taxonomic key of a guide to the
macroinvertebrates of Antioquia (Rolda´n-Pe´rez, 1988). Algae
were identified by a specialist from the herbarium of the
Universidad de Antioquia.
Data Analyses.—Morphological data were normally distributed
except for body weight and linear carapace width of females. To
test for sexual dimorphism and morphological differences
associated with habitat quality, we used an analysis of variance
(ANOVA) for data that were normally distributed and a Kruskal-
Wallis test for data not normally distributed. To test for
differences in the diet associated to the habitat of origin, we
used the overall sex ratio as the null hypothesis. Pairwise
comparisons were tested using a Bonferroni-corrected alpha
(0.05/6 =0.0083). To test for differences in linear carapace length
(LCL) and body weight (BW) distributions of males and females,
we used a Kolmogorov-Smirnov test. All data analyses were
completed with R 3.1.1 software (R Core Team, 2014).
RESULTS
Turtle Habitats.—All four creeks were perennial streams and
their channels were shallow and slow moving (1.8 m deep,
0.3 m/s); however, during the rainy season the water volume
may increase drastically and form temporary waterfalls which
may flood nearby cattle pastures. We estimated physical
parameters (Table 1) and a water contamination index based on
the identified macroinvertebrate families (Appendix 1) for each
creek.
Barrio Nuevo Creek takes its name from a human settlement
along the creek and a railroad line; residents use this creek for
sewage disposal. Accordingly, the biological monitoring work-
ing party (BMWP) scores (Appendix 1) indicated this creek to be
heavily polluted, the worst water quality of all four creeks.
Puente Tierra Creek was the shortest of the four study creeks,
originating in a small gallery forest and draining into Vega
Creek; we set up only one trap because the creek was too
shallow. The BMWP score in this creek suggested a slightly
polluted water quality. Totumo Creek was surrounded by open
areas of cattle pasture and fed a small pond that was densely
covered by a patch of gallery forest. The BMWP index
suggested a slightly polluted to very polluted water quality.
Finally, Vega Creek also was surrounded by cattle pastures and
had the fastest water velocity of all the creeks. The BMWP index
suggested the water quality of this creek was slightly to
moderately polluted.
Population Structure and Sex Ratio.—We recaptured 12 of the
total 80 turtles. Most of the time turtles were captured in traps (n
=79, 85.9%), but several were opportunistically captured by
hand (n=13, 14.1%). Totumo Creek was the site with the highest
number of turtles captured (n=35, with 8 recaptures) followed
by Barrio Nuevo (n=21, with 1 recapture), Vega (n=18, with 1
recapture), and Puente Tierra (n=6, with 2 recaptures). We
opportunistically captured turtles by hand in Barrio Nuevo and
Vega ( n=6 and 7, respectively) but none in the other two creeks.
The capture success (number of turtles captured per trap per day)
TABLE 1. Average physical and water quality characteristics calculated from the number of macroinvertebrate families observed at each creek
where Kinosternon leucostomum was captured (* indicates that the minimum and maximum values are reported when there were 2+sampling sites; n
=number of sampling sites at each creek).
Variable Barrio Nuevo (n=2) Puente Tierra (n=1) Totumo (n=4) Vega (n=3)
Water velocity, m/s (range) 0.09 (0–0.19) 0.06 (0.04–0.07) 0.08 (0–0.25) 0.25 (0.21–0.30)
Water depth, cm (range) 29.4 (8–50) 34.5 (21–48) 100.6 (29–184) 53.5 (24–104)
Transparency, % (range) 23 (8–40) 84 (79–95) 36.3 (13–100) 44 (16–100)
Water temperature, 8C (range) 24.5 23.3 (24–22.5) 23.2 (23–24) 23.2 (21.5– 26)
Number of macroinvertebrate families 1–2* 12 5–10* 6–11*
BMWP index 10–11* 84 32–67* 40–76*
Water quality Heavily
contaminated Slightly
polluted Slightly to
moderately polluted Slightly to
moderately polluted
TABLE 2. Abundance and sex distribution of Kinosternon leucostomum in San Jose´ del Nus, Antioquia, Colombia. (NA =______)
Creek Males Females Total Male-biased structure?
Barrio Nuevo 16 5 21 Yes (v
2
=5.7619, df =1, P=0.0163)
Puente Tierra 4 2 6 NA
Totumo 24 10 35 Yes (v
2
=4.8286, df =1, P=0.0279)
Vega 11 5 18 No (v
2
=2.25, df =1, P=0.1336)
Total 55 22 80 Yes (v
2
=13.254, df =1, P=0.0002)
BIOLOGY OF KINOSTERNON LEUCOSTOMUM POSTINGUINALE 0

was 0.67, 0.58, 028, and 0.46 for Totumo, Barrio Nuevo, Vega, and
Puente Tierra, respectively.
The overall observed sex ratio (28 male : 11 female) was 2.5 :
1, which deviated significantly from an expectation of 1 : 1. This
highly masculinized sex ratio was observed overall and at two
creeks, Barrio Nuevo and Totumo (Table 2). The small number
of turtles captured at Puente Tierra precluded analysis of the sex
ratio at that site.
Turtle Morphology.—Body measurements of adult turtles are
summarized in Table 3. The smallest individual was a neonate
captured by hand on the beach of Vega Creek on June 11, 2013. It
had an LCL =35.3 mm, linear carapace width (LCW) =25.8 mm,
linear plastron length (LPL) =27.5 mm, and it weighed 6.4 g.
We found significant sexual size dimorphism with males
being heavier, longer, and wider than females (Table 3). The LCL
and BW distributions of males and females differed significantly
(P<0.001 for LCL and P<0.001 for BW; Fig. 2). Most males
measured between 120–150 mm of LCL while most females
measured between 110–130 mm.
Given the habitat differences among the four creeks, we
tested for morphological differences related to the creek of
origin. Overall, turtles from Barrio Nuevo were heavier, longer,
and wider than were turtles from Totumo Creek (Table 4). This
was interesting given that aquatic macroinvertebrates found in
Barrio Nuevo Creek categorized it as heavily contaminated, yet
the turtles were the largest. On the other hand, Totumo Creek
was the site with the highest number of captures (see below) but
the turtles were the smallest.
Diet.—Of the 80 turtles captured, we collected a fecal sample
from 40 individuals. From this total, 65% had consumed plant
material (seeds, leaves, roots, or flower parts), 50% consumed
insects, 35% consumed snails, 17.5% consumed algae, and 17.5%
TABLE 3. Sexual dimorphism of Kinosternon leucostomum in San Jose´ del Nus, Antioquia, Colombia. Abbreviations: BW =Body weight (g), LCL =
Linear carapace length (mm), LCW =linear carapace width (mm), LPL =linear plastron length (mm), CCL =curved carapace length (mm), CCW =
curved carapace width (mm). Sexual dimorphism was significantly male biased in all variables (P<0.001).
Variable
Females (n=23) Males (n=54)
Male-biased sexual size dimorphismMean Min Max SD Mean Min Max SD
BW 201.4 100.0 382.6 60.2 285.5 125.0 475.0 83.5 29.5% heavier
LCL 118.2 105.2 136.2 7.6 132.8 105.0 156.0 12.1 11% longer
LCW 75.8 67.6 95.5 7.0 83.4 66.0 117.1 8.6 9.1% wider
LPL 104.9 94.2 119.0 6.1 113.0 92.0 132.3 8.1 7.2% longer
CCL 132.0 118.0 148.0 8.9 154.1 117.0 188.0 14.6 14.4% longer
CCW 115.5 96.0 140.0 10.3 125.9 93.0 152.0 12.7 8.3% longer
FIG. 2. Size class (A) and body weight (B) distributions of a population of Kinosternon leucostomum postinguinale in San Jose´ del Nus, Antioquia,
Colombia.
0 C. P. CEBALLOS ET AL.

consumed unidentified or animal material (e.g., shed skin,
unidentified muscle tissue). Specifically, we identified six insects:
leafcutter ants Atta sp. (n=15 turtles) as well as individuals, or
parts, of Gyrinidae (n=3), Gerridae (n=1), Gryllidae (n=1),
Odonata (n=1), and Hebridae (n=1). Snail (Gastropoda) shells
were barely digested, so we identified them to the species level:
Melanoides tuberculata (n=12), Pomacea palmeri (n=4), Subulina
octona (n=1), Aroapyrgus sp. (n=1), and Marisa cornuarietis (n=
1). The only green alga identified was Spirogyra sp. (n=4).
To account for the observed strong male-biased sex ratio, we
tested for diet differences associated with sex and the habitat of
origin. We did not find sexual differences in the consumption of
plant material (n
males
=22, n
females
=4, v
2
=2.11, P=0.146),
insects (n
males
=14, n
females
=6, v
2
=0.031, P=0.858), snails
(n
males
=10, n
females
=4, v
2
<0.001,P=0.976), or algae (n
males
=3, n
females
=3, v
2
=0.34, P=0.559). To test for geographic
differences, our null hypothesis tested the overall frequency of
the five food types (proportions of each category out of all 74
food items found in all creeks: 0.35 plant material, 0.27 insects,
0.19 snails, 0.095 algae, and 0.095 unidentified items). The
analysis was repeated for each creek. Food type occurrence in
Barrio Nuevo (v
2
=22.58, df =4, P=0.0001), Puente Tierra (v
2
=86.87, df =4, P<0.0001), and Vega (v
2
=58.47, df =4, P<
0.001) differed from the overall percentages, but not those in
Totumo Creek (v
2
=4.85, df =4, P=0.30). In Barrio Nuevo,
snails were consumed almost twice (35%) as much as overall
(Fig. 3). In Puente Tierra, algae were consumed more than three
times (33%) as much as overall, and no snails were found. In
Vega, the main food type was plant material (56%) and no algae
were found.
DISCUSSION
The habitat of K. l. postinguinale in northern Colombia was
highly variable. This turtle can live in heavily contaminated
environments that receive wastewater and garbage from rural
human settlements but also in muddy habitats, cleaner water
with abundant vegetation shading the stream, or in creeks with
rocks and fast-flowing water. Interestingly in this region, this
turtle also persists where waterfalls with a strong current form
during the rainy seasons, and these relatively small turtles
manage not to be washed away, as suggested from our
recaptures (6.7%). A similar habitat with cascades was reported
for this same subspecies in Costa Rica (Acun˜ a-Mese´ n, 1993).
Body size was smaller than in other populations of the same
subspecies and of the species overall. For example, on average
we found an LCL of 132.8 and 118.2 mm for males and females,
but K. l. postinguinale in Costa Rica is larger, measuring 174 and
158 mm, respectively (Acun˜ a-Mese´n, 1993). Similarly, maximum
LCL for males and females was 156 mm and 136 mm,
TABLE 4. Geographic differences in the morphometrics of Kinosternon leucostomum in four creeks in San Jose´ del Nus, Antioquia, Colombia.
Significant geographic differences were found in all variables.
Variable
Group means
Geographic
differences Pairwise comparisons
Barrio Nuevo (BN)
(n=21)
Puente Tierra (PT)
(n=6)
Totumo (T)
(n=34)
Vega (V)
(n=16)
BW 331.0 240.0 221.0 263.0 (P<0.001) BN >T(P<0.001)
LCL 138.3 122.4 124.8 126.3 (P<0.001) BN >T(P=0.0001),
BN >V(P=0.0061)
LCW 87.0 77.3 79.3 79.4 (P=0.0041) BN >T(P=0.0063)
LPL 116.6 106.1 104.8 110.4 (P=0.0193) None
CCL 157.6 140.0 143.1 148.1 (P=0.0071) BN >T(P=0.0064)
CCW 130.2 115.7 118.1 126.7 (P=0.0009) BN >T(P=0.0017)
FIG. 3. Frequency of occurrence of five food types found in the diet of Kinosternon leucostomum postinguinale in four creeks in San Jose´ del Nus,
Antioquia, Colombia.
BIOLOGY OF KINOSTERNON LEUCOSTOMUM POSTINGUINALE 0

respectively, while the maximum LCL reported for the species is
larger (175 mm and 165 mm, respectively; Berry and Iverson,
2001). Body size can vary with genetics, but also with the diet,
which in turn depends on food availability (Vogt and Guzman,
1988; Moll, 1990). Indeed, the larger individuals in our study
came from Barrio Nuevo Creek, the most-heavily contaminated
habitat; however, the main component of the diet at this stream
was snails, which are higher in protein than are insects and
plant material (mean nitrogen contents of Melanoides tuberculata
and freshwater angiosperms =9% and 2.4% dry weight,
respectively; Duarte, 1992; Mehler and Acharya, 2014). In
addition, the smaller and lighter individuals came from Totumo
Creek, where plant material and insects constituted 70% of the
diet.
Our findings agree with other studies that reported K. l.
leucostomum as an omnivorous generalist (i.e., they consume
insects, snails, fish, seeds, fruits, leaves, and even carrion in
Me´xico, northern Belize, and Costa Rica; Moll and Legler, 1971;
Vogt and Guzman, 1988; Moll, 1990; Acun˜a-Mese´ n, 1993). Other
authors have suggested that this species is primarily carnivo-
rous and eats plant material when meat is not available
(Medem, 1962). We did not find sexual differences in diet,
suggesting both sexes are similarly generalists or opportunists.
Finding aquatic snails and algae in their diet suggests they eat in
the water, as suggested by Medem (1962) and Moll and Legler
(1971). We also found terrestrial leafcutter ants in the diet,
however, and given the opportunistic nature of this species and
that it frequently moves on land (Moll and Legler 1971,
Morales-Verdeja and Vogt, 1997), it may feed on land as well
(Acun˜ a-Mese´ n, 1993).
We found a strong male-biased sex ratio (55 : 22) overall as
well as locally at the two sites with the most turtles trapped:
Barrio Nuevo (16 : 5) and Totumo (24 : 10). Factors that may
help explain the observed male-biased sex ratio of this
population include differential mortality, differential migration,
biased primary sex ratios, and even sexual size dimorphism of
the species (Gibbons, 1990; Girondot and Pieau, 1993; Lovich et
al., 2014). Regardless of the reason, the situation is of critical
concern as it suggests a serious risk of population extirpation.
And there is precedent for this concern: A population of
Clemmys insculpta in the United States, monitored for 20 yr,
revealed a reduction in the number of females and an increase in
age associated with human development until the turtle
population disappeared (Garber and Burger, 1995). We urge
further monitoring of the demography and reproduction of this
population, especially regarding the number of eggs and
clutches produced and the specific mechanism of sex determi-
nation.
Acknowledgments.—We thank our undergraduate students
and the staff of the Estaci ´
on Piscı´cola in San Jose´ del Nus,
particularly O. Franco, for logistical support during fieldwork. J.
F. Toro granted us access to the Experimental Station El Nus of
Corpoica. We thank L. E. Vela´ squez from the PECET for
gastropod identification and F. Cardona for algae identification.
This study was sponsored by a research grant from Convoca-
toria de proyectos de investigaci ´
on (CODI), Regionalizaci ´
on
2012 to CC, and the Programa de Sostenibilidad 2014–2015 of
the Universidad de Antioquia. This research was approved by
the Comite´de ´
Etica para la Experimentaci ´
on con Animales of
the Universidad de Antioquia as stated in Act 85 of 2013 and
research permit 135–0079 issued by the local government
authority CORNARE in 2012.
LITERATURE CITED
ACUN
˜A-MESE
´N, R. A. 1993. Las tortugas continentales de Costa Rica.
ICER, Costa Rica.
AGUIRRE-RAMI
´REZ, N. J., AND O. CAICEDO-QUINTERO. 2013. Me´todos de
Campo y Laboratorio para Hidrobiologı´a Sanitaria. Reimpresos,
Universidad de Antioquia, Colombia.
BERRY,J.F.,AND J. B. IVERSON. 2001. Kinosternon leucostomum (Dume´ ril and
Bibr ´
on) White-lipped Mud Turtle. Catalogue of American Amphib-
ians and Reptiles 724:721–728.
CAGLE, F. R. 1939. A system of marking turtles for future identification.
Copeia 1939:170–173.
CORNARE. 2012. Evaluaci ´
on y zonificaci ´
on de riesgos por avenida
torrencial, inundaci ´
on y movimiento en masa y dimensionamiento
de procesos erosivos en el municipio de San Roque. CORNARE,
Colombia.
DUARTE, C. M. 1992. Nutrient concentration of aquatic plants: patterns
across species. Limnology and Oceanography 37:882–889.
DUME
´RIL, A. M. C., AND A. H. A. DUME
´RIL. 1851. Catalogue Me´thodique
de la Collection des Reptiles (Muse´um d’Histoire Naturelle de Paris).
Gide and Baudry, France.
ERNST, C. H., R. G. M. ALTENBURG,AND R. W. BARBOUR. 1997. Turtles of the
World. World Biodiversity Database [Internet]. Available from:
http://wbd.etibioinformatics.nl/bis/turtles.php. Accessed 16 Feb-
ruary 2015.
GARBER, S. D., AND J. BURGER. 1995. A 20-yr study documenting the
relationship between turtle decline and human recreation. Ecological
Applications 5:1151–1162.
GIBBONS, J. W. 1990. Sex ratios and their significance among turtle
populations. Pp. 171–182 in J. W. Gibbons (ed.), Life History and
Ecology of the Slider Turtle. Smithsonian Institution Press, USA.
GIRALDO, A., M. F. GARCES-RESTREPO,AND J. L. CARR. 2012. Kinosternon
leucostomum. Pp. 332–339 in V. P. Pa´ ez-Nieto, M. A. Morales-
Betacourt, C. A. Lasso, O. V. Castan˜ o-Mora, and B. Bock (eds.), V.
Biologı´a y Conservaci ´
on de las Tortugas Continentales de Colombia.
Instituto de Investigaci ´
on de Recursos Biol ´
ogicos Alexander von
Humboldt (IAvH), Colombia.
GIRONDOT, M., AND C. PIEAU. 1993. Effects of sexual differences of age at
maturity and survival on population sex-ratio. Evolutionary Ecology
7:645–650.
IUCN. 2015. The IUCN Red List of Threatened Species. Version 105–3:
an online reference. International Union for Conservation of Nature.
Available from: http://www.iucnredlist.org. Accessed 14 October
2015.
IVERSON, J. B., M. LE,AND C. INGRAM. 2013. Molecular phylogenetics of the
mud and musk turtle family Kinosternidae. Molecular Phylogenetics
and Evolution 69:929–939.
LOVICH, J. E., J. W. GIBBONS,AND M. AGHA. 2014. Does the timing of
attainment of maturity influence sexual size dimorphism and adult
sex ratio in turtles? Biological Journal of the Linnean Society 112:142–
149.
MEDEM, F. 1962. La distribuci ´
on geogra´fica y ecologı´a de los Crocodylia y
Testudinata en el Departamento del Choc ´
o. Revista de la Academia
Colombiana de Ciencias Exactas, Fı´sicas y Naturales 11:279–303.
MEHLER, K., AND K. ACHARYA. 2014. Size distribution and nutrient
excretion of Melanoides tuberculata in a southern Nevada spring
ecosystem. Western North American Naturalist 74:386–395.
MOLL, D. 1990. Population sizes and foraging ecology in a tropical fresh-
water stream turtle community. Journal of Herpetology 24:48–53.
MOLL, E. O., AND J. M. LEGLER. 1971. The life history of a Neotropical
slider turtle, Pseudemys scripta (Schoepff) in Panama. Bulletin of the
Los Angeles County Museum of Natural History Science 11:1–102.
MORALES-VERDEJA, S. A., AND R. C. VOGT. 1997. Terrestrial movements in
relation to aestivation and the annual reproductive cycle of
Kinosternon leucostomum. Copeia 1997:123–130.
RC
ORE TEAM. 2014. R: A language and environment for statistical
computing. R Foundation for Statistical Computing, Vienna, Austria.
Available from: http://www.R-project.org/
ROLDA
´N-PE
´REZ, G. 1988. Guı´a para el Estudio de los Macroinvertebrados
Acua´ticos del Departamento de Antioquia. Fondo para la Protecci ´
on
del Medio Ambiente ‘‘Jose´ Celestino Mutis’’ (FEN Colombia) y
COLCIENCIAS, Colombia.
ROLDA
´N-PE
´REZ, G. 1999. Los macroinvertebrados y su valor como
indicadores de la calidad del agua. Revista de la Academia
Colombiana de Ciencias Exactas, Fı´sicas y Naturales 23:375–387.
0 C. P. CEBALLOS ET AL.

SPINKS, P. Q., R. C. THOMSON,M.GIDIS,AND H. B. SHAFFER. 2014.
Multilocus phylogeny of the New-World mud turtles (Kinosternidae)
supports the traditional classification of the group. Molecular
Phylogenetics and Evolution 76:254–260.
VAN DIJK, P. P., J. B. IVERSON,A.G.J.RHODIN,H.B.SHAFFER,AND R. BOUR.
2014. Turtles of the world, 7th edition: annotated checklist of
taxonomy, synonymy, distribution with maps, and conservation
status. Chelonian Research Monographs 5:329–479.
VOGT, R. C., AND S. G. GUZMAN. 1988. Food partitioning in a Neotropical
freshwater turtle community. Copeia 1988:37–47.
Accepted: 5 January 2016.
APPENDIX 1. Aquatic macroinvertebrate families identified at each of the 10 trapping sites in Barrio Nuevo, Puente Tierra, Totumo, and Vega creeks
and the BMWP scores (Rolda´n-Pe´rez, 1999) assigned to each family to estimate water quality of the creek.
Macroinvertebrate
Barrio Nuevo Puente Tierra Totumo Vega
1 2 3 4 5678 9 10
Aeshnidae - - - - - 6 - - - -
Belostomatidae 5 - 5 - - - 5 - - -
Calopterygidae - - 7 - - - - - 7 7
Coenagrionidae - - - - 7 7 - - - 7
Corixidae - - 7 - - - - - 7 -
Dytiscidae - - - - 9 - - - - -
Elmidae - - - 6 - - - 6 - -
Ephemerellidae - - - - - - - - 9 -
Gelastocoridae - - - - - - - 5 5 5
Gerridae - - 8 - 8 8 8 8 8 -
Gomphidae - - 10 - 10 10 - - - 10
Gyrinidae - - 9 - - - 9 - 1 -
Hebridae - - - 8 - - 8 - - -
Hidrometridae - - - - - 4 - - - -
Hydropsychidae - - 7 - - - - - - -
Libellulidae 6 - 6 6 6 6 - - 6 6
Mesovelidae - - - 5 - - - - 5 -
Naucoridae - - 7 - 7 7 - - - -
Nepidae - - - - - - - - 5 5
Noteridae - - 4 - - 4 - - - -
Notonectidae - - - 7 7 - 7 - 7 -
Planariidae - - - - - - - 7 - -
Pleidae - - - - - - - 8 - -
Ptylodactylidae - 10 - - - - - - - -
Saldidae - - - - - - - - 8 -
Scirtidae - - - - 7 7 - - - -
Staphylinidae - - 6 - - - 6 - - -
Veliidae - - 8 - - 8 8 8 - -
Total BMWP 11 10 84 32 61 67 51 42 68 40
Water quality Heavily polluted Slightly polluted Slightly to
very polluted Slightly to
moderately polluted
BIOLOGY OF KINOSTERNON LEUCOSTOMUM POSTINGUINALE 0