Está en la página 1de 82

Universidad de Costa Rica

Sistema de Estudios de Posgrado

Ecología de la tortuga candado Kinosternon scorpioides (Testudines: Kinosternidae) en

el Parque Nacional Palo Verde, Costa Rica.

Tesis sometida a la consideración de la Comisión del Programa de Estudios de Posgrado


en Biología para optar al grado de Magíster Scientiae.

Franklin Enrique Castañeda Menéndez

Ciudad Universitaria "Rodrigo Facio"


Costa Rica

2005
Dedicatoria

Esta Tesis esta dedicada a la Memoria de mi padre, Franklin Devis Castañeda

Contreras, por mostranne la creación y al Creador

A mi esposa Nereyda Estrada Andino por darle sentido a lo que hago

A mi madre Perla Miriam Menéndez por enseñarme el valor de la disciplina

A mis hermanos Carolina y Jaime Castañeda

u
Agradecimientos

• A los amigos: Mayori Grimaldo, Jeffi:ey Ortiz Gamboa, Leandro Castillo,


Robert Pushendorf, Eduardo Toral, la familia Roberts, Randal y Cyntia Chaves,
Lina Paola Pérez, Melvin Cartin.

• A la Dra. Virginia Solís por su amistad, consejo y valioso apoyo.


• Al Servicio Alemán de Intercambio Académico (DAAD) por hacer posible mis
estudios en Costa Rica.

• A mi comité de tesis: José Manuel Mora, GerMdo Avalos y Gerardo Umaña

• Al profesor José Manuel Mora por brindarme su amistad y por mostranne los
rincones de Costa Rica

• Al profesor Federico Bolaños por abrirme puertas a oportwtidades de superarme


como investigador y como científico. Por estar siempre en su oficina y tener
siempre el tiempo y las ganas de colaborar.

• Al profesor Gerardo Avalos por su valioso aporte a este proyecto de Tesis,


especialmente por las largas horas vit.~do y analizando mis datos.

• A mi esposa Nereyda Estrada Andino, por rescatanne de mucho huecos


académicos, por ayudarme con la estadística, redacción, ortografia, por
acompañarme y ayudarme dw·ante las giras de campo en Palo Verde a pesar de
los zancudos y el calor. Por estar a mi lado cuando se pierde a quien se ama.

• A la Sociedad para el Estudio de los Anfibios y Reptiles (SSAR)

• A la Organización para Estudios Tropicales (OET)

• A IdeaWild

• A1 Ministerio de Ambiente y Energia (MINAE) por otorgar el permiSO de


investigación para este trabajo

• A los amigos del Ministerio de Ambiente y Energia que trabajan en el Parque


Nacional Palo Verde

• A los amigos de la Organización para Estudios Tropicales que trabajan en el


Parque Nacional Palo Verde

• A James R. McCranie por revisar el idioma ingles del manuscrito de tesis

• A mi familia Castañeda y Estrada por su apoyo y por estar siempre


esperándonos con los brazos abiertos en la patria

• Al pueblo de Costa Rica

111
"Esta Tesis fue aceptada por la Comisión de) Programa de Estudios de Posgntdo en
Biología de la Universidad de Costa Rica. como requisito parcial para optar alg1ado de
Magíster Scientiae".

Jorge Azofeifa Navas. Ph.D.


Representante del Decano deJ Sistema de Estudios de Posgmdo

José annel Mora Benavides., Ph.D.

~T~ ~

B~~-r)

Rafael Arturo Acuña Mesén, M.Sc.


Representante del Director del Programa de Posgrado en Biología

_;7~- : ::i -_É_


Franklín Enrique Casta:tieda Menéndez
Candidato

lV
Índice

Pagina

Dedicatoria ... .............. ....... .. ............................................................ .. ..ii

Agradecimientos.. ......... ...... ...... ... ... .. .. ......... .... .......... ........................... .iü

Índice ...... ............ ... .................. .. .......... ................. ...... .......... .... ..... .... v

Índice de cuadros ... ... ... ........... ....... .... ................... .. ..............................vi

Indice de Figuras ...... ...... ........... ... .............. ........ ........... .... ... ............... vü

l. Introducción .... .. .................................... ............................. . ............... 1

Movimiento y actividad .................................................... .. .... .. ... ....2

Selección de sitios de anidacióo .......... ...... ..... .... ..... . .............................. .3

Impacto de los incendios ...... ..... ....... .... ... ........ ... ... ... ..... .. ........ ........... 5

Distribución y hábitat. ............................. ............. .......................... 5

\~<:>n.<:>m\a..................................................................................6

Características morfométricas... ..... . .... ..... .......... .... ....... ... ..... . .. . ..... .... 7

Biogeografia ......................................... ... ... ....... ... ..........................8

Reproducción ............ ....... ...... .. ........ . ......... ......... ..... . ........ ...........9

Literatura citada ............ ......... ... ........ ....... ............... ...... ..... ... ....... 10

n. Nest site selection and nesting behavior of the mud turtle Kinostemon scorpioides
(Testudines: Kinostemidae) in Palo Verde National Park, Costa Rica: irop1ications for
tnanagement. ..... ......... ... .. ... ....................... . .................. .. . ..................... 15

lll. Terrestrial movements, activity pattern and habitat use of the mud turtle
Kinostemon scorpioides in Palo Verde National Park, Costa Rica.... ..... .. .... ...... ... 37

N . Tbe impact of fue on the population of mud turtles Kinosternon scorpioides


(Testudínes: Kinostemidae) in a tropical wetland ofnorth.westem Costa Rica ......... .. 55

V
Índice de Cuadros

Introducción

Cuadro 1. Principales características que difieren entre sexos en Kinosternon

scorpioides y la tendencia de estas diferencias en hembras y

machos ... .. ........... .. .... ...... ........ ....................... .. .............. .. ........ .... .. ...... 8

Articulo 1

Table l. The mean and standard error of the mean for eight variables of mi<.Tohabitat

measured at nest sites and at random sites in a Kinostemon scorpioides nest:ing area in

Palo Verde Nat:ional Park (units and sample size shown in parenthesis) ............... ... 35

Articulo 3

Table l. Comparison of size between twtles (Kinosternon scorpioides) that were kílled

by frres and mud turtles that swvived tires........ ... . .. ... . ...................................73

Table 2. Density in several species ofturtles (number ofindividuals 1 hectare) ......... .74

VI
Índice de Figuras

Articulo 1

IIK~"-~~ 't, UU:Cerenccs in three variables of microhabitat measured in arcas wtder cattle

grazing and free of cattle grazing in a nesting area of Kinostemon scorpioides in Palo

Verde National Park. A) understory cover (U = 412, p < 0.0001), B) leaflitter cover (U

= 864, p < 0.0001), C) leaflitter depth (U ""' 895, p <0.0001) .............. . .... ...... .. ... 36

Articulo 2

Figure l. Map ofPalo Verde National Park, northwestern Costa Rica ................... .52

Figure 2. Number ofmud turtles (Kinosternon scorpioides) active on land from

February to May 2004 in Palo Verde National Park........ ... ....... ........... .... .. .......52

Figure 3. Mud turtle (Kinostemon scorpioides) daily activity pattern related to ambient

ternperature in Palo Verde National Park ... ............................. . .. ................... 53

Figure 4. Mud turtle (Kinoslernon scorpioides) daily actívity pattern related to ambient

relative hurnidity in Palo Verde National Park. ... .... . ... ... ....... .. ... ...... ... ...... ...... 53

vu
Figure 5. Frequency of aestivating individuals of Kinosternon scorpioides by distance

category, sex, and age, in Palo Verde National Park. .. .. .... ..... . ... ... . ... .. ___..... . ..... .54

Articulo 3

Figure l. Map ofPalo Verde National Park, northwestem Costa Rica... ...... ... ... ... .. 70

Figure 2. Dead mud turtle (Kinosternon scorpioides) density (mean density and

standard error) in all habitats affected by fu·es at Palo Verde National Park (Apastures,

B. areas dominated by Parkinsonia aculeate, C. secondary gallery forest, D. Large areas

dominated by Typha domingensis, E. areas · of diverse and dry aquatic vegetation

(DAV), and F. smaJJ areas of Typha dcmingensis). .. ___......... ... -·-· -·· ...... ... ...... .. . 71

Figure 3. Comparison of turtles (Kinosternon scorpioides) killed by ftres and mud

turtles that survived fires in two groups: adults and juveniles, in Palo Verde National

Parle .. ... . .. ... ... .. . .. ... ... . ........ .. .... . ... .. ..... .... ... .. ............... .. .. ..... ... .... ...... 71

l"igure 4. Mud turtle (Kinosternon scorpioides) distribution in the wetland in Palo

Verde National Park during the rainy season (A), and during the dry season (B).......... 72

Vlll
L

Ecología de la tortuga candado Kinosternon scorpioides(Kinosternidae) en el Parque

Nacional Palo Verde, Costa Rica.

Introducción

Las poblaciones de tortugas alrededor del mundo están declinando de fonna alannante

(Emst & Barbour 1989, Moll & Moll 2004). Actualmente las tortugas terrestres de Costa

Rica están amenazadas por Ja destrucción del hábitat, los incendios y la caceria, factores

que siguen diezmando las poblaciones de tortugas aún dentro de las áreas protegidas

(Acuña 1990, Mora & Ugalde 1991, Acuña 1994b, Acuña 1998). Desafortunadamente, esta

situación es aún más critica en otros países de América Latina, como en Colombia donde

ya se han extinguido poblaciones locales de .la especie Trachemys scripta (Pritchard &

Trebbau 1984); y en Belice donde el gobierno ha tenido que aprobar leyes estrictas para

evitar la extinción de algunas especies de importancia económica como Dermatemys mawii

(Polisar 1994, 1995). Resulta importante entonces, realizar investigaciones sobre aspectos

ecológicos de los Testudines tropicales tales como el uso y la selección de hábitat, actividad

y movimiento.

En Jos últimos años se ha despertado e) interés en ]a comunidad cientifica por

estudiar el movimiento, la actividad y la selección de hábitat de las tortugas terrestres y

semiacuáticas, y un considerable repertorio de artículos científicos ha sido pubhcado sobre

este tema (Toro 1994, Wilson et al. 1994, Nieuwott 1996, Lue et al. 1999, Carter et al.

2000, Morrow et al. 2001a, Morrow et al. 2001b, Ramsay et al. 2002, Galois et al. 2002,

Rowe 2003); sin embargo, todos estos trabajos fueron realizados con especies de zonas

11:mpladas o fuera del Neotrópico.


2

Movimiento y actividad

Entender los patrones de movimiento de una especie es crucial para explicar su

distribución, abundancia, flujo genético y comportamiento social, factores que son clave en

la conservación (Gibbons et al. 1990). La alteración y fragmentación del hábitat separa

{X)blaciones que antes fueron continuas, reduciendo la variabilidad genética e

incrementando la probabilidad de extinción. La implementación de corredores biológicos

pretende amortiguar este impacto, pero los corredores sólo son útiles si los organismos

pueden desplazarse a través de ellos. Desafortunadamente, para muchas especies de

Testudines, especialmente en el Neotrópico, la dispersión y el movimiento son los aspectos

ecológicos menos conocidos (Gibbons 1990, Shawn et al. 2000, Rowe 2003).

Los movimientos de un individu<? ocurren como respuesta a situaciones

ambientales, características demográficas de la población como sexo, tamaño y estructura

de edades, y a condiciones fisiológicas y reproductivas. Por esta razón, el estudio de los

movimientos dentro y entre poblaciones es crítico para entender una variedad de procesos

ecológicos y de historia natural. Por ejemplo: Identificar los patrones temporales y

espaciales del movimiento de los individuos de una población, es un prerrequisito para

entender las razones ecológicas y evolutivas de la dispersión y migración (Gibbons et al.

1990). El estudio del patrón de movimiento y actividad puede revelar como están

distribuidos los recursos en el ambiente que rodea a una especie y como son utilizados por

la población (Rowe 2003).

En la especie Kinosternon scorpioides este aspecto importante de la ecología se

coooce únicamente a partir de los trabajos de Acuña (1990) y Teska (1976). De acuerdo

COD Acuña (1990) el período de mayor actividad de K. scorpioides en el Parque Nacional

.Palo Vede (PNPV) es en las últimas horas de la tarde. Según Teska (1976), en el PNPV
3

encontró tortugas activas principalmente en horas de la mañana. En tortugas mantenidas en

cautiverio se observó que fueron más activas durante la noche. La población de K.

scorpioides del PNPV muestra un comportamiento de migración local relacionado con la

variación climática anual. Al final izar Ja época de Jluvias muchas tortugas abandonan el

área de lagunas que comienza a secarse y entran en las áreas de bosque deciduo; al finalizar

la estación seca y con el inicio de las lluvias el patrón de migración local se invierte (Acuña

1990, Acuña 1994a). En cuanto a las distancias recorridas por esta especie, Teska (1976)

reporta movimientos terrestres sustanciales (entre 175 y 500 m) y Acuña ( 1998) indica que

las hembras grávidas salen del agua y se desplazan entre 5 y 200 m para poner los huevos.

En una especie de zonas templadas (Kinostemon subrubrum), Bennett (1970) reporta

movimientos cortos, generalmente menores a 2 m. Y, una distancia recorrida promedio de lO

m diarios.

Sdección de sitios de anidacióo

El hábitat es la swna total de los factores del ambiente (alimento, cobertura, agua, sitios de

.Udación) que una especie requiere para poder sobrevivir y reproducirse en un área dada.

-Cada animal sHvestre tiene requisitos específicos de hábitat y para cualquier especie, su

- ..ndancia y posible distribución en un área determinada están limitadas por la calidad,

a-ridad y disponibilidad del hábitat favorable (Lyon 1987). Según Litvaitis et al. (1994)

_ se sabe muy poco sobre los requerimientos de hábitat de los vertebrados.

La selección de sitios de anidación se puede definir como la acción de las hembras

: poner sus huevos en sitios que difieren en características fisicas de sitios elegidos al azar

son 1998). En las especies en las que la selección de sitios pard anidar ocurre, este

~---·"'rtamiento puede estar relacionado con la protección del embrión de depredadores y

es ambientales extremos, también con la selección de un sitio que fuciJite Ja


4

construcción del nido y por Jo tanto minimice la inversión del costo energético de la

hembra. Si los sitios seleccionados ofrecen una mayor probabilidad de supervivencia a sus

ocupantes, entonces la selección natural favorecerá el comportamiento de selección de

sitios de anidación (Wilson 1998).

La cobertura vegetal que rodea un determinado sitio de anidación puede influir

sobre la temperatura de desarrollo del embrión (Janzen 1994). Esto es especialmente

importante en especies en que la proporción sexual es determinada por la temperatura,

como es el caso en muchas especies de tortugas (Vogt & Flores 1992). Se ha probado

también que el ambiente que rodea el nido influye sobre el crecimiento de los embriones

(Packard & Packard 1987). Los organismos que muestran cuidado parentaJ después de la

ovoposición hacen que sea dificil de probar la hipótesis de la ventaja selectiva de selección

de sitios de anidación. Las tortugas no muestran _cuidado parental y esto las convierte en

buenos organismos modelos para probar esta hipótesis. K. scorpioies es una de las especies

de tortugas más pequeñas de Costa Rica y en general de Mesoamérica, Jo cual es relevante

en el contexto de que la cobertura vegetal del sitio seleccionado para anidar es

especialmente importante en las especies de tortugas de pequeño tamaño. Esto se debe a

que estas especies excavan nidos de menor profundidad, y por lo tanto al estar los

embriones más cerca de la superficie se exponen a ser afectados. et\ ma.~~~ ~<\..~ ~~ \.~"'-'

factores ambientales como temperatura y humedad. Las tortugas de mayor tamaño

dependen en menor grado de la estructura de la vegetación que rodea su hábitat, ya que

pueden excavar nidos más profundos protegiendo así los embriones de los extremos

unbientales, y se ha observado que en general Jas especies de tortugas de mayor tamaño

tienden a elegir sitios mas abiertos y las de menor tamaño a elegir sitios con mayor

cobertura vegetal (Wilson 1998).


5

Impacto de Jos incendios

En Costa Rica un hábitat importante para K. scorpioídes lo constituyen las lagunas

estacionales del PNPV (Acuña 1990 & 1994). A mediados de la década de los 90, estas

lagunas estaban cubiertas principalmente por la planta acuática Typha dominguen...is y

parches de la especie arbórea Parkinsonia ac.:uleata. Con el propósito de mejorar el hábitat

de las aves acuáticas migratorias, la administración del parque está desarrolJando un

proyecto para eliminar estas plantas en algunos sectores del humedal, creando áreas

abiertas a través de la remoción mecánica, quemas controladas y pastoreo (MINAE: R-009-

ACA-D-2002). Aunque el manejo activo es favorable para las aves migratorias (Trama

2005), se desconoce el efecto que esta tra.nsformación del hábitat puede tener en la

población de otras especies. Los incendios son frecuentes en el humedal del PNPV. En

mayo de 1989 un incendio mató por lo menos 144 K. scorpioides en tres días, afectando

principalmente juveniles y hembras, por lo que se concluyó que el fuego podía modificar la

estructura poblacional de esta especie (Acuña 1990). Desde 1989 otros incendios han

ocurrido en el parque sin que se haya documentado su impacto. Debido a que las quemas

controladas fonnan parte de las técnicas de manejo en el PNPV, es importante documentar

su impacto sobre la población de K. scorpioides.

Distribución y hábitat

El ámbito de distribución de K. scorpioides en la vertiente del Caribe se extiende desde

Tamaulipas, México a través de Centro América hasta el norte de Argentina y el este de

Brasil (Lee 2000, Stafford & Meyer 2000, Kohler 2001). En la vertiente del Pacífico se

distribuye desde Oaxaca, México hacia el sur a través de Centro América (Savage 2002).

En Costa Rica esta especie se encuentra ampliamente distribuida en las tierras bajas de la
6

vertiente del Pacífico (Acuña 1998, Leenders 2001). También se encuentra en la Meseta

Central y en elevaciones de O a 2,500 m.s.n.m. (Acuña et al. 1983, Savage 2002).

K. scorpioides puede encontrarse en una gran variedad de hábitat<; acuáticos como

quebradas, pantanos, tíos y lagos, y en bosques secos, húmedos y lluviosos de tierras bajas

y en bosques húmedos premontanos (Acuña 1998, Savage 2002). En un estudio sobre los

requerimientos de hábitat de K. scorpioides en el Valle Central de Costa Rica, se detemúnó

que esta especie vive en aguas con temperaturas entre 23 °C y 27 °C, y de una dureza entre

54.78 ppm. (suaves) y 125.00 ppm. (duras). Se encontró en cuerpos de agua con abundante

materia orgánica en descomposición y con valores de pH entre 6 y 7.83 (Acuña et al.

1983). En las lagunas del PNPV las tortugas se sumergen en el fondo lodoso o se refugian

entre las raíces de las abundantes plantas acuáticas (Eieocharis mutata, Typha

dcminguensis, Canna tutea y Tafia geniculala). En la época seca K. scorpioides busca

refugio entre las raíces de los árboles que rodean el área de la laguna (e.g. Enterolobium

cyclocarpum, Cordia alliodora, Calycophylum candidissimum, Anacardium excelsum,

Tamarindus indica), permaneciendo períodos prolongados en estos refugios, desde varios

días hasta semanas (Acuña 1990, Acuña 1994a).

Taxonomía

La especie K. scorpioides pertenece a la familia Kinosternidae (Reptilia: Testudines), que

mmprende un grupo de tortugas terrestres y semiacuáticas exclusivas del nuevo mundo.

Actualmente se reconocen tres géneros y 22 especies en esta familia (Savage 2002). En la

especie K. scorpioides se reconocen seis subespecies, de las cuales se encuentra en Costa

ltica la subespecie Kinosternon scorpioides cruentatum, conocida localmente como tortuga

mudado, tortuga de fango, musuca, o tapaculo (Acuña 1998). En su amplio ámbito de

llistribución esta especie muestra gran variabilidad ontogenética e individual, haciendo


7

dificil la interpretación de la variabilidad geográfica. Sin embargo, se reconocen las

siguientes subdivisiones geográficas principales:

l. Poblaciones Mesoamericanas

1.1 . Poblaciones del norte: incluye México, Guatemala, Belice y el Salvador

1.2. Poblaciones de la cuenca del Río Grande de Chiapas: México y Guatemala

1.3. Poblaciones de Honduras, Nicaragua, Costa Rica y Panamá

2. Poblaciones de América del Sur

2. l. Poblaciones del norte de Sur América

2.2. Poblaciones sureñas: incluye a Bolivia y Argentina.

En cada una de estas regiones geográf1cas se observan diferencias morfométricas

particulares (ej : forma y tamafio del caparazón, tamaño del plastrón, muesca anal presente o

ausente, escudos inguinal y axilar en contacto o no, patrón y color de la cabeza), y las

poblaciones (excepto las de Bolivia y Argentina) están conectadas en zonas de transición

donde se observan diferencias intennedias (Savage 2002).

Características morfométricas

De acuerdo con Savage (2002), esta es una especie pequeña con una longitud estándar del

caparazón de hasta 270 mm. K scorpioides tiene un caparazón de color café o negro,

oblongo y convexo, en el que se observan tres ere~ longitudinales, aunque estas pueden

desaparecer en ejemplares de mayor edad. El escudo nuca) tiene forma rectangular o

cuadrada y es particularmente pequeño. Los escudos vertebrales son angostos y de un

ancho similar, tiene cuatro pares de escudos costales y once marginales. El plastrón puede

ser de color amarillo o anaranjado y tanto el lóbulo anterior como posterior son móviles,

característica que le permite cerrar totalmente su concha sin dejar partes blandas expuestas

(Acuña 1998). Acuña et al. (1993), en un extenso estudio sobre los componentes óseos de
8

esta especie, encontraron que el peso de la concha (sin escamas córneas) fue de 301.61 g y

la conforman un total de 52 piezas. Se encontró también, que los elementos esqueléticos de

alta densidad se distribuyen siguiendo un patrón bilateral (Acuña et al. 1993). En esta

especie se observa dimorfismo sexual. Diferencias significativas entre sexos se reportan en

el ancho curvo del caparazón. el largo del plastrón, en el peso, tamaño, patrón y coloración

de la cabeza y la longitud de la cola (Cuadro 1) (Acufia & Cruz 1993).

Cuadro l. Principales características que difieren entre sexos en K scorpioides y la

tendencia de estas diferencias en hembrdS y machos.

Caracter ística morfométrica 1 Tendencia

Ancho curvo del caparazón Mayor en hembras *

Largo curvo del caparazón Mayor en hembras *


-
Peso Mayor en hembras

Longitud de la cola Mayor en machos

Tamaño de la cabeza Mayor en machos

Pigmentación de la cabeza Más marcado en machos


1
* Se encontró lo contrario en esta investigación.
Biogeografia

Las tortugas de la familia Kinostemidae evolucionaron en el continente Americano, sus

fósiles no se han encontrado fuera del nuevo mundo. Aunque el fósil más antiguo que se

conoce (Xenochelys formosa, relacionado al genero reciente Staurotypus y a la familia

Dennatemydidae) pertenece al Oligoceno y se encontró en Estados Unidos, se cree que este

= tDsil más antiguo es el de Kinosternon jlavescens arizonenese del Plioceno de Arizona.


9

También se conocen los fósiles de K. baurii y K. subrubrum del Pleistoceno de Wisconsin

(Emst & Barbour 1989). Los cladogramas hechos para las tortugas Kinosternidae,

RVelaron que los taxones con el plastrón más pequeño son más primitivos y tos de plastrón

-=ís grande son más derivados (Iverson 1991). De acuerdo con su distribución, Savage

(2002) divide la unidad herpetofaunística de Mesoamérica en 4 grupos principales, el

Fiero Kinosternon está comprendido en el grupo tres: América Tropical Media (centros de
--=i!n y distribución en México y Centro América y con un limite de distnoución norte

..: .:~.:o fuera de América tropical del norte). Adicionalmente, de las 7 regiones geográficas

-- que se divide la herpetofauna de Mesoaméri.ca, Kinosternon se encuentra en: ( 1) tierras

-:.!S y húmedas del este y oeste; (2) tierras bajas del Pacifico; (3) tierras panameñas; y (4)

- =:-~ del Chocó.

roduccióo

;::~ la especie K. scorpioides el apareamiento ocurre en enero y julio. Las hembras ponen

Rvos en dos épocas del año, de febrero a marzo y de agosto a noviembre. Los nidos

~ ama profWldidad de alrededor de 5 cm. La incubación tarda entre 115 y 128 días. Se

.-e los huevos de esta especie son depredados por coyotes (Canis latrans), pero es
,~o...oc:.:..- ~ que otros depredadores también utilicen este recurso (Acuña 1998).. Por ejemplo,

..:e que en el PNPV, especies como Dasypus novemcintus y Loxocemus bicolor

-.,....-.~ o entran en los nidos comunales de Iguana iguana y Ctenosaura similis para

""rn----..:use de sus huevos y neonatos (Mora 1987, Mora 1989, Mora 1991).

~:a.:::-;:~Jmente se sabe que L. bicolor también depreda huevos de tortugas marinas en las

·• -.;;....,. de Guanacaste, Costa Rica (Mora & Robinson 1984).


10

Literatura citada

Acuña, R. 1990. El impacto del fuego y Ja sequía sobre la estructura de la población de

Kinosternon scorpioides (Testudines: Kinosternidae) en Palo Verde, Costa Rica.

Brenesia. 33:85-97.

Acuña, R. 1998. Las tortugas continentales de Costa Rica. Universidad de Costa Rica.

San José. 58 p.

Acuña, R. 1994a. Variación morfométrica y características ecológicas del hábitat de la

tortuga candado Kinosternon scorpioides en Costa Rica (Chelonia,

Kinostemidae). Rev. Brasil. Biol. 54(3):537-547.

Acuña, R. 1994b. Conservación y ecología de las tortugas terrestres. semi.acuáticas y

acuáticas (de agua dulce y marinas) de Costa llica. EUNED. San José, Costa

Rica.

Acuña, R. & B. Cruz. 1993. El dimorfismo sexual de Kinosternon scorpioides

(Testudines: Kinostemidae) en Palo Verde, Costa Rica. Rev. BioL Trop. 41(2):

261 - 265.

Acuña, R. , A. Castaing & F. Flores. 1983. Aspectos ecológicos de la distribución de las

tortugas terrestres y semiacuáticas en el Valle Central de Costa llica. Rev. Biol.

Trop. 31(2): 181-192 .

..k:uña, R. , C. Acuña & M. Tejeira. 1993. Los componentes óseos del caparazón y del

plastrón de dos tortugas costarricenses: Kinostemon scorpioides y K

leucostomwz (Testudines: Kínosternidae). Brenesia. 39-40:81-92.

DC:JIUlelt t, D., G. Whitfietd & J. Franson. 1970. Te.rrestrial activity in aquatic twtles.

Ecology. 51(4): 738.


l1

Carter, S., C. Haas & J. Mitchell. 2000. Movements and activity ofbog turtles (Clemmys

muhlenbergii) in southwestem Virginia. J. Herpetol. 34(1): 75-80.

Ernst, C. & R Barbour. 1989. Turtles ofthe world. Smithsonian. Washington. 313 p.

Galois, P., M. Leveille., L. Bouthillier., C. Daigle & S. Parren. 2002. Movement patterns,

activity and home range of the eastern spiny softshell turtle (Apalone spinifera)

in northemlake Champlain, Quebec, Vermont. J. Herpetol. 36(3): 402-411.

Gibbons, J. 1990. Recommendations for future research on freshwater turtles: what are the

questions? p. 311 -317. In l Gibbons (ed.). Life history and ecology ofthe

slider turtle. Smithsonian Institution. Washington D. C.

Gibbons, J., J. Greene & J. Congdon. 1990. Temporal and spatial movement pattems of

sliders and other turtles, p. 201-215. In J. Gibbons ( ed.). Life history and

ecolog_y ofthe slider turtle. Smithsonian Jnstitution. Washington D.C.

l verson, J. 1991. Phylogenetic hypotheses for the evotution of modern kinostemine turtles.

Herpetological Monographs. 5:1-27.

Janzen, F. 1994. Vegetational cover predicts the sex ratio of hatchling turtles in natural

nests. Ecology. 75(6): 1593-1599.

Kohler, G. 2001. Anfibios y reptiles de N icaragua. Herpeton. Offenbach. 208 p .

Lee, J. 2000. Amphibians and reptiles of the Mayan World. Comell University. New

York. 402 p.

Leenders, T. 2001. Amphibians and reptiles ofCosta Rica. Zona Tropical. Miami. 305

p.
12

Litvaitis, J .• K. T1tus & E. Anderson. 1994. Measuring vertebrate use of terrestrial habitats

and foods, p. 254-274. In T. Bookhout (ed.). Research and rnanagement

techniques for wi ldlife and habitats. Maryland

Lue, K. & T. Chen. 1999. Activity, movement pattems, and borne range of the yellow

margined box turtle (Cuora jlavomarginata) in northern Taiwan. J. Herpetol.

33(4): 590-600.

Lyon, J. 1987. Análisis y evaluación del hábitat, p. 321-344. En: R. Rodríguez (ed).

Manual de técnicas de gestión de vida silvestre. Wildlife Society. Maryland.

Ministerio de Ambiente y Energía (MINAE). 2002. Resolución Legal R-009-ACA-D-

2002. Área de Conservación Arenal.

Moll, D & E. Moll. 2004. The ecology, exploitation, and conservation of river turt1es.

Oxford. New York. 393 p.

Mora, J. 1987. Predation by Loxocemus bicolor on the eggs of Ctenosaura similes and

Iguana iguana. J. Herpetol. 21 (4 ): 334-335.

Mora, J. 1989. Eco-behavioral aspect of two communally nesting iguarúnes and the

structure oftheir shared nesting burrows. Herpetologica. 45(3): 293-298.

Mora, J. 1991 . Loxocemus bicolor, feeding behavior. Herp. Rev. 22(2): 61 .

Mora, J. & A. Ugalde. 1991. A note on the population status and exploitation of

Pseudemys scripta emolli (Reptilia: Emydidae) in northem Costa Rica. Bull.

Chicago Herp. Soc. 26(5): 111.

Mora, J. & D. Robinson. 1984. Predation of sea turtle eggs (Lepidochelys) by the snake

Loxocemus bicolor Cope. Rev. Biol. Trop. 32(1): 161-162.


13

Morrow, J., J. Howard., S. Smith & D. Poppel. 2001a. Home range and movements ofthe

bog turtle (Clemmys muhlenbergii) in Maryland. J. Herpetol. 35(1 ): 68-73.

Morrow, J., J. Howard., S. Smith & D. Poppel. 200 lb. Habitat sele<..iion and habitat use by

the bog turtle (Clemmys muhlenbergü) in Maryland. J. HerpetoJ. 35(4): 545-

552.

Nieuwolt, P. 19%. Movement, activity and microhabitat seJection in the westem box

turtle, Terrapene ornata luteo/a, in New Mexico. Herpetologica. 52(4):487-

495.

Packard G. & M. Packard. 1987. Influence of moisture, temperature, and substrate on

snapping turtles eggs and embryos. EcoJogy. 68(4): 983-993.

Polisar, J. 1994. New Jegislation for the protection and management of Dematemys mawii

in Belize, Central America. Herp. Rev. 25(2): 47 - 49.

Po1isar, J. 1995. River turtle reproductive demography and exploitation patterns in Belize:

implications for management. Vida Silvestre Neotropical. 4(1 ): lO - 19.

Pritchard, P & P. Trebbau. 1984. The turtles of Venezuela. Society for the Study of

Amphibians and Reptiles. New York. 403 p.

Ramsay, S., M. Hofineyr & Q. Joshua. 2002. Activity pattems of the angulate tortoise

(Chersima angulata) on Dassen lsland, South Mrica. J. Herpetol. 36(2): 161-

169.

Rowe, J. 2003. Activity and movements of midland painted turtles (Chrysemys picta

marginata) living in a small marsb system on Beaver Island, Míchigan. J.

Herpetol. 37(2):342-353.

Savage, J. 2002. The amphibians and reptiles of Costa Rica. Chicago. Chicago. 934 p.
14

Stafford, P. & J. Meyer. 2000. Reptiles of Belize. The Natural History Museum of

London. Academic. London. 356 p.

Teska, W. 1976. Terrestrial movements of the mud turtle Kinostemon scorpioides in

Costa Rica. Copeia. (3):579-580.

Tom, J. 1994. Microhabitats and use ofburrows ofBolson tortoise hatchlings, p. 139-146.

In R. Bury & D. Germano (eds. ). Biology of North American tortoises. United

States Department ofthe Interior, Washington D. C.

Trama, F. 2005. Manejo activo y restauración del humedal Palo Verde: cambios en las

cobertura~ de la vegetación y respuesta de las aves acuáticas. Ms.C. Tesis,

Universidad Nacional, Heredia. 154 p.

Vaughan, D., C. McCoy, M. Fallas, J. Chaves, H. Barboza, G. Wong., G. Carbonen, M

Rau & J. Carranza. 1994. Plan de manejo y desarrollo Parque Nacional Palo

Verde y Reserva Biológica Lomas Barbuda!. Universidad Nacional. Heredia.

192 p.

Vogt, R. & O. Flores. 1992. Effects of incubation temperature on sex determination in a

com.rnunity of neotropical freshwater turtles in southern Mexico. Hepetologica.

48(3): 265-270.

Wilson, D . 1998. Nest-site selection: microhabitat variation and its.effect on the survival

of turtle embryos. Ecology. 79(6): 1884-1892.

Wilson, D., H. Mushinsky & E. McCoy. L994. Home range, activity, and use ofburrows

of juvenile gopher tortoises in Central Florida, p. 147-160. in R. Bury & D.

Germano (eds. ). Biology of North American tortoises. United States

Department of the Interior, Washington D.C.


15

Article 1. Nest site selection and nesting behavior of tbe mud turtle KiDosternon

scorpioides (Testudines: Kinosternidae) in Palo Verde National Park, Costa Rica:

implications for management.

Abstract

Habitat selection is the process whereby individuals preferentially use, or occupy, a non-

random set of available habitats. Nest site selection can be defined as the placemeot of eggs

by females at sites that differ from random sites within a delimited arca. We located a total

of 59 nests of Kinosternon scorpioides in Palo Verde National Park (PVNP). We compared

eight microhabitat variables at nest sites against those at random sites. Females placed their

eggs at sites with more understory cover, more leaflitter cover, aod greater leaflitter depth.

Additionally, females selected sites with lower air and soil temperature and lower air

humidity. PVNP is subject to active management actions oriented to control inva<;ive plant

species. We found that habitat quality at nesting areas is being threatened by at Jeast one of

these actions: cattle grazing. This is detrimental for three microhabitat traits that turtles

select for nesting sites: understory cover, leaf litter cover, and leaf litter depth. The

continued degradation of microhabitats on nesting areas could affect the population

structure of this turtle and embryo' s survivorship. Many species of amphibians and reptiles

at PVNP may depend upon the natural integrity of the ecotone between the marsh and fhe

dry forest where K. scorpioides nesting areas are. Thus it is imperative that any disturbance

in this important transition zone is stopped, and that other specific studies regarding several

scales of habitat use and selection are undertaken.

Key words: mud turtle, Kinosternon scorpioides, habitat seJectíon, nesting, Palo Verde,

Costa Rica.
16

Resumen

La selección de hábitat es el proceso en el que los individuos utilizan u ocupan un conjunto

no aleatorios de hábitats disponibles. La selección de sitios de anidación puede deflllirse

como el proceso en el que las hembras de una especie dada ponen los huevos en sitios que

difieren de sitios aleatorios dentro de un área definida. En el Parque Nacional Pato Verde

(PNPV) localizamos un total de 59 nidos de la especie Kinostemon scorpioides,

comparamos ocho variables de microhábitat en los nidos con sitios elegidos al azar. Los

nidos se caracterizaron por tener una mayor cobertura de arbustos, de hojarasca seca y una

mayor profundidad de mantillo. Adicionalmente, las hembras seleccionaron sitios con una

menor temperatura del suelo y menor temperatura y humedad relativa deJ aire en la

superficie. En el PNPV se practican acciones de manejo activo orientadas al control de

plantas invasordS. Encontramos que la calidad ~el hábitat en las áreas de anidación de K.

scorpioides esta siendo alterada por una de estas acciones: el pastoreo. Esto va en

detrimento de tres características del microhábitat que las tortugas seleccionan para anidar:

cobertura de arbustos, cobertura de hojarasca y profundidad del mantillo. La continua

degradación del microhábitat en las áreas de anidación podría afectar la estructura de la

población y la sobrevivencia de los embriones. En el PNPV muchas especies de anfibios y

reptiles probablemente dependen de la integridad ecológica de la zona de ecotono entre el

humedal y el bosque seco donde se encuentran también las áreas de anidación de K.

scorpioides. Es importante que cualquier disturbio en esta zona de transición sea

controlado, y que otros estudios específicos utilizando varias escalas de uso y selección de

hábitat sean realizados.


17

Introduction

The abundance of a particular species, and its distribution in space and time is often

detennined by the availability of habitat. Such habitat must include all necessary

components to ensure the species survival: food, water, refuge, and nest sites (Litvaitis et

al. 1994). Thus, nesting areas are important components of a species habítat and need to be

analyzed in terms of the species requirements for successful reproduction. Within protected

natural areas, accumte management actions could only be achieved if we gain a better

understanding of the habitat quality needed by a particular species to nest.

Habitat selection is the process wP.ereby individuals preferentially use, or occupy, a

non-random set of availab1e habitats (Morris 2003 ). Nest site selection can be defined as

the placement of eggs by females at sites that differ from random sites within a delimited

area. The intentíon of such behavior may be to reduce egg predation and exposure to

environmental extremes (Wilson 1998). Nest site selection has been documented for

several species of turtles (i.e., Chrysemys picta bellii [Janzen 1994a], Ktnostemvn bourii,

Kinosternon subrubrum, Pseudemydura umbrina (Wilson 1998], Chelydra serpe11tina

[Kolbe & Janzen 2002] Carettochelys insculpta [Doody el al. 2003], Emydura macquarii

[Spencer & Thompson 2003]). The importance of the structural characteristics of the

environment surrounding the nest and their influence on survival rate, hatcbling size, and

sex ratios, have been documented by several field and laboratory studies (Packard &

Packard 1987, Packard et al. 1989, Vogt & Flores 1992, Janzen 1994a, Wilson 1998,

Weisrock & Janzen 1999, Hewavisenthi & Parmenter 2001, Kolbe & Janzen 2001, Booth

2002, Hewavisenthi & Parmenter 2002). For instance, the humidity of the incu.bation

environment influences embryo size, and larger hatchlings appear to be more successful

escaping predation and surviving the first year of life (Packard & Packard 1987). Thus,
t8

extreme changes in nest si te humidity could lower the hatchling' s survival rate. AJso, in

species with temperature-dependent sex detemlination (TSD), sex ratios can be skewed to

one particular sex as a result of habitat alteration on nesting areas. For iustance, areas

stripped of vegetation would produce more females because of a higher incubation

temperature. These conditions would affect directly the adult population structure (Vogt &

Flores 1992). Furthermore, species with TSD could serve as good 1ndicators of the

biological impact of global warming, since the annual offspring sex ratio is highly

correlated witb mean air temperature (Janzen 1994b). This implies that a smaU increase in

mean temperature ( < 2 °C) may drastically skew the sex ratio. Hence, habitat alterations

that affect nesting areas could have a severe impact on freshwater turtle populations.

Palo Verde National Park (PVNP) in northwestem Costa Rica, is one of the few

national protected areas that is legally subject to habitat changes caused by human actions

(active management) sucb as cattle grazing, controlled fires, and mechanical crushing of

vegetation. The purpose of these actions in the wetland is to create and maintain areas of

open waters favorable to wintering waterfowl, and to reduce the risk of wild fires (Bumidge

2000, Barboza 2002, MINAE: R-009-ACA-D-2002). While bird populations are mon.itored

to asses their response to changes i.n resource availability (i.e., habitat characteristics), no

information is available on the impact of these management actions on any other vertebrate

species within tbe park.

U sed resources could be compared to available resources to reach valid conclusions

conceming resource selection. When resources are used disproportionateJy to their

availability, their use is said to be selective (Manly et al. 2002). The comparison of the

physícal structure of nesting locations with that of random sites within a delimited area is

especially useful in identifying the characteristics of the habitat that detem1ine nest site
19

selection (WiJson 1998, Kolbe & Janzen 2002). In this paper we compared severa}

microhabitat variables at nest sites against those at random sites in a mud turtle's nesting

area within PVNP. We analyze how the active management of this protected area is

changing the habitat structure of nesting areas and its possible impact on a turtle' s

population. Comments are made on female nesting behavior, clutch size and nesting female

Stze.

Study species

Kinosternon scorpioides is a small species of freshwater turtle widely distributed on the

Pacífic sJope ofCosta Rica. Females ofthis species nest from February to March. and from

August to November. Females dig a hole 2.1 - 4.6 cm deep on the ground and lay between

one and six hard b rittle shelled eggs. Shells are 40x 19 mm in length, and hatchlings are 30 -

40 mm in length. Incubation lasts from 115 to 128 days (Acuña 1998, Savage 2002).

Teroperature-dependent sex determination with pattem II (TSD-II) has been reported for at

least six species in the family Kinostemidae, including K. scorpioides (Ewert & Nelson

1991, Janzen & Kienz 2004). No information is available on the microh.abitat

characteristics ofthe areas where K. scorpioides nests. According to Savage (2002), little is

known about the life history traits of this species.

Materials and methods

Study site

We conducted our research at Palo Verde NationaJ Park (PVNP), located in the

°
northwestem lowlands of Costa Rica, Guanacaste Province ( 1O 0 21 ' N - 85 21' W). The

park has an extension of 19,800 h~ average temperature is 27 ° C, and the annual

precipitation is 2,296 mm. The d.ry season lasts from December to April and the rainy

season from May to November. A great variety of habitats are represented in PVNP:
20

pnmary and secondary decíduous dry forest, riparian forest, savannas and wetlands

(Hartshom 1983). The sampling effort was concentrated in an area of 4 ha of old secondary

dry forest adjacent to the Palo Verde lagoon. This area was located between the facilities of

the Organization for 'Tropical Studies (OTS) and those ofthe Park Ranger station (Fig. 1).

Nesting behavior, nest location aod microhabitat variables

We made observations on female nesting bebavior frorn October 2003 to November 2004.

Body size of nesting females, number of eggs, and size of eggs were recorded. To

detennined temale size, we measured carapace curve length (CCL) in cm to the nearest 0.1

with a flexible measuring tape, and body weight in g with a 1000 g capacity Swiss Pesola.

To determined egg size we measure Jength and width in cm to the nearest 0.1 using a 152

mm metal caliper, and weight in g using a 30 g capacity Swiss Pesola. We located and

indivídually marked K. socorpioides's nests during February 2004. Because these nests are

cryptic and difficult to find, nests were located by either following emerging females to

their nest sites, raking leaf Jitter, or sighting the white egg's shelJs exposed by nest

predators. The study area (4 ha) was divided ínto evenly separated transects. Every lO m

along each transect a random site was tocated (site without nest). To test the nul\ hypothesis

of lack of ditTerences between microhabitat characteristics among nests and random sites,

we compared the values of a total of eight microhabitat variables measured at each nest site

to those measured at random sites. The variables studied were: 1) canopy cover (measured

with an spherical densiomenter modeJ - C); 2) understory cover, for this study aH shrubby

and herbaceous vegetation from O. 1 to 1.5 m above the ground was considered understory

cover; 3) leaf litter cover (variables 2 and 3 were measured by centering a 1 m2 grid over

each site, and visually estimating the percentage ground cover); 4) horizontal uriderstory

cover (this was accomplish by using a 35x20 cm wooden board marked with 10 circles of3
21

cm in diameter, the board was located above the ground ata clistance of 1 m ftom the sites.

The number of circles that the vegetation allowed us to see was counted 11-om a height of20

cm above the ground~ two measures were taken on each site, one to the north and one to the

south. This methodology was modified from Higgins et al. (1994))~ 5) leaf Jitter depth

(measured with a metal ruler); 6) air temperature; 7) air humidity (variables 6 and 7 were

measured with a digital hygro-thermometer model 445582 Extech lnstruments, it was

located 5 mm above the ground in the shade at each nest site and each random site and the

reading was taken after 5 min); 8) soil temperature (measure at 5 cm depth using a

thennometer HI-145-30 Hanna lnstruments, the reading was takenafter 5 min).

We tested for differences in average ambient temperature among days in which nest

and randorn site temperatures were taken using the data from the weather station of OTS.

To eliminate the effect of temperature variat:io~ throughout the day, all temperature

measurements used to compare nest sites and random sites were taken from 1200 to 1400

hours (wannest hours of the day). To better estímate the nest average temperature

throughout the day, nest temperatures were also taken from 0500 to 0700 hours (coolest

hours ofthe day).

Two treatment analyses

The study area (4 ha) is divided by a fence running east - west. The park's administration

uses the fence to keep the cattle in the wetJand and out of the forest. However, the fence is

placed 10 to 15m within the forest, resulting in two treatments within the same nesting area

and habitat: 1) areas under cattle grazing, and 2) areas free of cattle gra.zing. In order to test

for microhabitat differences between treatments, 144 1m2 plots were randomly placed

inside and outside the fence (72 plots per treatment). Three microhabitat variables were

measured on every plot: understory cover, leaf litter cover, and leaf litter depth.
22

Data analyses

To test for differences in microhabitat variables between nest sites and random sites the

data was analyzed using a Multivariate Analysis of Variance (MANO VA). A 2-sample t-

test was used to test for differences among each variable. Differences bet\veen both

treatments (1- area wíth cattle b'Tazing, 2- area free of cattle grazing) were tested with the

Mann-Wh:itney U test. Ambient temperature data from the OTS weather station were

analyzed using a One-Way Analysis of Variance (ANOVA). All percentage data was

arcsine transformed before analyses were done (Zar 1999). The data was analyzed using

JMPin 4.0.4 and Systat 9.0.

Results

Nest location and microhabitat variables

No nests were found in open areas or on the wetland border which was dry an.d covered

with aquatic plants or grass tussoc.ks during the dry season. AH nests were found under the

forest canopy. Nests were found at distances that ranged from 5 to 175 m from the wetland.

During the dry season, most nests were found in the first 5 to 50 m of forest from the

wetland border. However, during the wet season nests were found at greater distances ( up

We located a total of 59 nests and 54 random sites. Females placed their eggs at

si tes that differed from random si tes (MANO VA F 8, 90 '= 11.2, p <0.000 1) in at least seven

ofthe eight variables of microh.abitat under analysis. Significant statistical differences were

found in canopy cover (t 109.95 = -2.86, p < 0.005), understory cover (t w9. 95 = 2.44, p =
0.01), leaflitter cover (t I09, 95 ~ 2.39, p = 0.01), leaflitter depth (t tos, 95 - 3 .68, p < 0.001),

air temperature (t 104 , 95 - -2.62, p = 0.01), air humidity (t to3, 95 = -2.53, p = 0.01 ), and soil

temperature (t 105, 95 = -4.74, P < 0.001). Horizontal understory cover was the only variable
23

that showed no differences between nest sites and random sites (t 107,95, = 0.42, p = 0.67).
Regarding the vegetation structure (ground cover), it appears that females were selecting

sites with more understory cover, more leaf litter cover, and greater [eaf litter depth.

Females nested in sites with less canopy cover (Table 1). Regarding temperature and

humidity, it appears that females selected sites with lower air and soil temperature and

lower air humidity (Table 1). Average soil temperature at nest sites was 28.13 °C from

1200 to 1400 h, and 26.40 °C from 0500 to 0700 h, with an overall average of 27.26 °C.

When we analyzed the temperature readings from the OTS weather station, no significant

differences in average ambient temperature among days of sampling were found (ANOVA

F s. t6s = 0.87, p = 0.50). This indicates that environrnental temperature on the study area

was homogeneous throughout the days in which temperature readings were taken at nest

and random sites. Leaf litter depth was negatively correlated with soil temperature (Pearson

r = -0.35), and air temperature was negatively correlated wíth air humidity (Pearson r = -

0.57).

Two treatment analyses

Only two nests were found in areas under cattle grazing (3.4 % of total nests tound, n = 59).

However, outside the fence (area free of cattle grazing) severa! nests were found 1 - 10m

from the fence (72.9 %). Moreover, seyeral nests were found right under the fence (10.1).

Only 13.5 % of the nests were fowtd at distances greater than lO m from the fence.

Significant differences existed between both treatments in all three variables measured

([Mann-Whitney U test] understory cover: U = 412, p < 0.0001; leaf littercover: U = 864, p

< 0.0001; leaf litter depth: U = 895, p < 0.0001). Arcas outside the fence (free of cattle

grazing) had higher percentage of understory cover, leaf litter cover and greater leaf litter
24

depth. The understory cover was almost nonexistent in areas inside the tence (under cattle

grazing) and both leaf litter variables were lower (Fig. 2).

Nesting behavior

Females were seen nesting during February, October and November. Evidence of nesting

activity was also noted in March (fresh mud accurnulation on female's posterior marginal

plates). Females left the water and walked straight into the forest, once under the forest

canopy they spent from 1 to l .5 h moving erratically in areas of about 15x 15 m, apparently

searching for a suitable si te to nest. During this search, females dig with their front legs into

the leaf litter and place their snouts on the ground (ground-nuz.zling), move a few meters

and repeat these actions. This behavior lasted until the turtle finally began to dig a nest with

its hind legs. The nest digging process lasted from 1 to 1.5 h. Sorne females dug their nest

while their shell was completely covered with lea ves. AII females found nesting (n = 5) had

a fresh accumulation of mud on the posterior marginal plates and occasionally also on the

anal plates after completion of the nesting process. This occurs because of the shell position

during the nest excavation. While excavating the nest, females assume an angular posture

with the anterior portion of the shell elevated and the posterior part angled down so as to

allowed the hind legs to reach and dig deeper into the ground. This results in mud

accumulation on the posterior marginal plates. During the dry season when the forest floor

is completely dry, mud accumulation is possible because females empty their bladder while

in the process ofnesting (pers. obs.). Neither juvenile turtles (n - 29), nor aduJt males (n ~

42) were found with this pattern of mud accurnulation. Thus, we are confident that fema1es

showíng this pattern had nested recently (when mud accurnulation was fresh), or had nested

and gone ínto the forest to aestivate before going back to the water (when mud

accumulation was d.ry).


25

Nesting females were found active from 700 to 1100 h and from 1500 to 2300 h.

Out of a total of 102 females foWld on land during the study period, 14 were found with

mud accumulation on posterior marginals (five of which were found while nesting). This

group of 14 femaJes averaged (x±SE) 18.41 ± 0.28 cm curve carapace length (CCL) and

(x±SE) 670 ± 32.9 g of weight after nesting. Al\ fema\es ranged in size from 16.6 cm to

19.9 cm CCL, and from 460 to 790 g. Nesting females deposit from three to eight eggs per

nest. We found a tendency for a positive relationship between the number of eggs and the

females CCL (r = 0.73, p = 0.06). The egg' s shells have an average length of3.26 ± 0.16
cm (n :::: 35), average width of l. 79 ± 0.06 cm (n = 35), an average weight of 6.34 ± O.74 g

(n = 16). One hatchling was seen actively moving over Jand on May 2004, this individual

measured 2.7 cm standard carapace length, 2.1 cm standard carapace width and weighed

4.25 g. Two egg clutches were collected in November 2004 (clutch of three eggs UCR

17395; clutch of seven eggs UCR 17396).

Discussion

Burke & Gibbons (1995) found that in order to grant protection for 100% ofthe nests and

hibernation sites for three species of fresh water turtles (Kinosternon subrubrum,

Trachemis scripta and Pseudemys jloridana), 275 m of terrestrial habitat needed to be

protected around the wetland. Thus, upland habitats adjacent to wetlands are critical to

nesting and the life-cycle functions of freshwater twtles. At PVNP, K. scorpioides nests

were found in the forest at a maximum distance of 175 m from the wetlands, Acuña (1998)

reported nests up to 200 m from the wetlands in the same area. This data should be

considered when establishing buffer zones boundaries around wetlands elsewhere in Costa

Rica. However, at PVNP many kilometers of all terrestrial habitats surrounding the wetJand

are included within the park limits, and therefore are already under protection. Two of our
26

findings are that 86.4% of the nests were found between 5 and 25 m from the wetland and

that K. scorpioides gravid females are selective in terms of nesting sites. This means that in

order to guarantee the successful reproduction of this species, certain habitat traits must be

met by the areas comprising the fírst few meters of terrestriaJ habitat surrou.nding the

wetland. Thus, it is the quality of habitat available that matters and not only its quantity in

tbis particular case. We found that habitat quality at nesting areas is being threatened by at

least one of the active management actions: cattle grazing. Livestock are changing the

understory structure by destroying aH shrubby vegetation and leaf Jitter over the .first 1O to

15 m of forest from the wetland. This ís detrimental for three mícrohabitat traits that turtles

select for nesting sites: understory cover, leaf litter cover, and leaf litter depth. lndeed,. the

large differences found in vegetation structure between areas under cattle grazing (inside

the fence) and areas free of grazing, explain the almost total lack of nests in the first 1O to

15 m from the wetland and the great abundance of nests just pass the fence, apparently

indícating that gravid females seem to avoid nesting on the cattle grazing area, because th1s

area does not comply with the habitat requirements to nest.

The continued degradation of microhabitats on nesting areas could affect the

population structure of this turtle. For instance, as vegetation cover is removed by the

trampling action of cattle, incubation temperatures at nest sites may increase, resulting in

biased sex ratio. Indeed, we found a slight negative correlation between soil temperature at

nest sites and leaf Jítter depth, meaning that the soil tended to be cooler at greater lea.f Jitter

depth. In the snapping turtle Chelydra serpentine, Juliana el al. (2004) found that three

variables of vegetation cover were correlated with nest temperature: cooler nest were

located under higher vegetation cover. Janzen (1994a) found evidence of correlation

between vegetation cover at nest sites and offspring sex ratio in Chrysemys pie/a.
27

Accordíng to Vogt & Flores (1992) sex ratio differences between populations of fresh

water turtles wíth TSD, may reflect the differences in availabJe nesting habítat Moreover,

Wilson (1998) suggested that small species of turtles would be more dependent on

vegetational cover to protect their embryos fTom environmental extremes (i.e., high

temperature) than larger turtles, because the fonner are not able to dig as deep into the

ground, and therefore lay the eggs nearer the surface. Because nest depth is negatively

correlated to nest tem.perature, small turtles would have to rely on vegetation cover to

provide a cooler incubation temperature for their embryos. K. scorpioides is considered

among the smallest species of turtles in Costa Rica (Savage 2002), and Kinostemon is in

the smallest end of turtles sizes in the world (Wilson 1998). At PVNP aiJ K. scorpioides'

nests found were no deeper than 5 cm in the ground, whereas other species of turtles in

Costa Rica wíth greater carapace length such as Chelydra serpen1ína, Rhinoclemmys

pulcherrima, and Trachemys scripta may dig nests as deep as 13, lO, and 18 cm

respectively (Acuña, 1998).

K. scorpioídes has TSD-II, meanmg that males predominate at intennediate

~~fi'ó'9.~Nl?tn téi~~Ó\lé~~éill'~l' 1tin'íWJ~~dóWffi~ée:Jú?lth cx~é\We~~ SpymfiéeJf 1~?J?ila\~?lth

only one threshold temperature, in TSD-II there are two transítion zones. In K scorpioides

70 to 82 %males were produced at temperatures from 24 to 27°C, 78 % females were

produced at 22.5°C and 100% females were produced at 30°C. In a closely related species

Kinosternon leucostomum, 75 %males were produced at 25°C but 100 % females were

produced at 2'PC (Ewert & Ne1son, 1991). This data suggest that a small difference in

incubation temperature can drastically change sex ratios, like the abrupt drop from 75 %

males at 25°C to O % males at 27°C in K. leucostumum. At P\'1\'P average soiJ temperatures

from 1200 to 1400 h at nest sites were 28.13°C, almost one degree Celsi.us lower than the
28

average temperature at random sites (29.04°C). Nest sites had significantJy more vegetation

covering than random sites. Hence, at K. scorpioides nesting areas in PVNP, as vegetation

cover decreases because of the trampling action of cattle, soil temperatures will most Likely

rise, probably taking incubation temperatures over the upper transition zone, and skewing

sex rations towards females. Although our data are useful to compare between selected si.tes

(nests) and random sites, in order to proper\y docwnent incubation temperature at P\lNP,

much longer sampling periods using temperature data loggers are needed.

In addition to impacting on the population sex ratio, a decrease vegetation covering

may also affect the turtle's population in many other ways. For instance, the embryos

survival rate. Embryo's survivorship in Kinosternon baurii at selected sites (near

vegetation) was significantly higher than that of embryos at nonselected sites (away from

vegetation). This was attributed to the higher dai!Y temperature reached by soiJ at open sítes

devoid of vegetation (Wilson 1998). Hatching success was high on wetter and cooler

substrates, but declined at higher temperatures and on drier media in the species Chelydra

serpentina (Packard & Packard 1987). Changes in microhabitat characteristics such as

understory cover and leaf htter may al so increase predation on eggs, hatchlings and nesting

females. During this study several potential predators were commonly seen on the nesting

areas ( e.g., Procyon lotor [Raccoon], Nasua narica [White Nosed Coati], Didelphis spp.

[Opossum], Ctenosaura similis (Black Iguana], Drymarchon corais [Indigo Snake]).

According to Acuña (1998) all the mammals listed above are predators ofturtles or twtJe's

eggs. At PVNP we made direct observations of these mammals while predating eggs of K.

scorpíoides. D. corais has been reported to predate on turtles (Savage, 2002). C. similes is

known to eat K. socorpioides eggs at PVNP (M. Montes pers. com.). Feinale carapaces

with evident signs of predation were found in the nesting areas and many nests found were
29

predated. A reduction of shrubby vegetation and leaf litter may increase the chances of

predators to detect nesting females, hatcWings, and nests.

The availability of various resources is not generally unifonn in nature, and use may

change as availabílity changes (Manly et al. 2002). At PVNP, the dry deciduous forest

generates a yearly variation on the availability of two important microhabitat traits: leaf

htter cover and leaf litter depth. During the dry season when the trees Joose their leaves,

both leaf litter cover and leaf litter deptb increases, and both decrease when the rai.ns start.

Data on microhabitat at nest sites and random sites were taken only during the dry season.

Thus, we are confident that during the dry season nestíng period, females are selecting

areas with higher Jeaf 1itter cover and greater leaf litter depth. However, in PVNP females

of K. scorpioides nest both during the dry and rainy seasons. During the rainy season nests

and nesting females were found at sites without any leaf litter cover. Hence, nest site

selection in K. scorpioides may be influenced by forest phenology and t.herefore have a

yearly variation pattem. A study of nest site selection during the raíny season wouJd be

necessary to clarify this scenery.

This work points out specific microhabitat traits that K scorpioides' nesting areas

require in order to ensure a successful reproduction, and that these traits are being

negatively affected by one active management action: cattle grazing. However, a simple

recommendation may help resol ve this matter. Since the main idea of having cattle within

the wetland in PVNP is to reduce the amount of invasive aquatic plants in the marsh, there

is no need to have the fence running 10 to 15 m within the dry forest lnstead the fence

should be moved and placed at the edge between the marsh and the forest This way, cattle

would be kept strictly on the wetland and would preclude them from disturbing the

vegetation structure in the nesting areas of turtles that surround the marsh. FinaJiy,
30

Semlistsch & Bodie (2003) summarized data from the Iiterature on the use of terrestrial

habitats by amphibíans and reptiles associated wíth wetlands ( 19 species of frogs, 13

salamanders, 5 snakes and 28 turtles), and assessed the biologically relevant size of core

habitats surrounding wetlands for these groups. They found a range from the aquatic site

edge of 159 to 290 m for amphibians and 127 to 289 m for reptiles. Their data also

indicates the importance of terrestrial habitats for feeding, over wintering and nesting.

Thus, the healthy biological interdependence between aquatic and terrestrial h.abitat is

essential for the persistence of populations. In addition to K. scorpioides, many species of

amphibians and reptiles at PVNP may depend upon the natural integrity of the ecotone

between the marsh and the dry forest. Thus it is imperative that any disturbance in this

fragile, yet vital transition zone, is stopped and that other specific studies regarding several

scales of habitat use and habitat selection are undertaken.

Literature cited

Acuña, R. 1998. Las tortugas continentales de Costa Rica. Universidad de Costa Rica. San

José. 92 p.

Acuña, R. & B. Cruz. 1993. El dimorfismo sexual de Kinosternon scorpioides (Testudines:

Kinostemidae) en Palo Verde, Costa Rica. Rev. Biol. Trop. 41 : 261-265.

Barboza, J. 2002. El pastoreo en la restauración del bosque tropical seco: Parque Nacional

Palo Verde, Guanacaste, Costa Rica. Seminario Nov 13-15 2002. Universidad

Nacional/ INTSEFOR. Heredia.

Booth, D. 2002. Incubation ofrigid-sheUed turtJe eggs: do hydric conditions matter? Comp.

Physiol. 172:627-33.

Burke, V. & J. Gibbons. 1995. Terrestrial buffer zones and wetland conservation: a case

study of freshwater turtles in a Carolina bay. Conservation Biology 9 :1365-l369.


31

Bumidge, W. 2000. Cattle and the management of freshwater neotropical wetlands in Palo

Verde National Park, Guanacaste, Costa Rica. Thesis. M.Sc. University of

Michigan. Ann Arbor. 89 p.

Doody, S., P. West & A. Georges. 2003. Beach selection in nesting Pig-Nosed turtles,

Carettochelys insculpta. Journal ofHerpetology 37: 178-182.

Ewert, M. & C. Nelson. 1991. Sex determination in turtles: diverse patterns and sorne

possible adaptive values. Copeia. 91:50-69.

Hewavisenthi, S. & J. Parmenter. 2001. lnfluence of lncubation environment on the

development of the flatback turtle (Natator depressus). Copeia. 1:668-682.

Hewavisenthi, S. & J. Parmenter. 2002. Incubation envíronment and nest success of the

Flatback turtle (Natator depressu<;) from a natural nesting beach. Copeia 2:302-

312.

Higgins, K., J. 01demeyer., K. Jenldns., G. C1ambey & R. Harlow. 1994. Vegetation

sampling and measurement, p. 567-591. In T. Bookhout (ed.). Research and

management techníques for wildlife and habitats. The Wíldlife Society. Maryland

Hartshorn, G. 1983. Introduction, p. 119-186. In D. Jaozen. Costa Rica natural history.

Chicago. Chicago.

Janzen, F. 1994a. Vegetational cover predicts the sex ratio of batchling turtles in natural

nests. Ecology. 75:1593-1599.

Janzen, F. l994b. Climate change and temperature-dependent sex determination in reptiles.

Proc. Natl. Acad. Scí. 91: 7487-90.

Janzen, F. & G. Krenz. 2004. Phylogenetics: which was first, TSD or GSD? p. 121-130. In

N. Valenzuela & A. Lance (eds. ). Temperature dependent sex determínation in

vertebrates. Smíthsonian Instítution, Washington D.C.


32

Juliana, J., R. Bowden & F. Janzen. 2004. The impact of behavioraJ and physiologicaJ

maternal effects on offspring sex ratio in the common snapping turtles, Chelydra

serpentina. Behav. &ol. Sociobiol. 56:270-278.

Kolbe, J. & F. Jan?...en. 2001. The influence of propagule size and maternal nest site

selection on survival and behavior of neonate turtles. Functional Ecology 15: 772-

781.

Kolbe, J. & F. Janzen. 2002. Irnpact of nest-site selection on nest success and nest

temperature in natural and disturbed babitats. Ecology. 83: 269-281.

Litvaitis, J., K. Titus & E. Anderson. 1994. Measuring vertebrate use ofterrestrial habitats

and foods, p. 254-274. In T. Bookhout (ed.). Research and management techniques

for wildlife and habitats. The Wildlife Society. Maryland.

Manly, B., L. McDonald., D. Thomas, T. McDonald & W. Erickson, 2002. Resource

selection by animals. Kluwer Academic. Dordrecht 221 p.

Ministerio de Ambiente y Energía (MINAE). 2002. Resolución Legal R-009-ACA-D-2002.

Área de Conservación Arenal.

Morris, D. 2003. Toward and ecological synthesis: a case for habitat selection. Oecologia.

136:1-13

Packard, G. & M. Packard. 1987. Influence of moísture, temperature, and substrate on

snapping turtles eggs and embryos. Ecology 68: 983-993.

Packard, G., M. Packard & G. Birchard. 1989. Sexual differentiation and hatching success

by painted turtfes incubating in different thennal and hydric environrnents.

Herpetologica 45: 385-392.

Savage, J. 2002. The amph.ibians and reptiles of Costa Rica. Chicago. Chicago. 934 p.
33

Semlistsch, R & R. Bodie. 2003. Biological criteria for buffer zones around wetlands and

riparian habitats for amphibians and reptiles. Conservation Biology. 17: 1219-1228.

Spencer, R. & M. Thompson. 2003. The significance of predation in nest site selection of

turtles: an experimental consideration of macro- and microhabitat preferences.

Oikos 102: 592-600.

Vogt, R. & O. Flores. 1992. Effects of incubation temperature on sex detennination in a

comrnunity of neotropical freshwater turtles in southern Mexico. Hepetologica. 48:

265-270.

Weisrock, D. & F. Janzen. 1999. Thermal and fitness-related consequences of nest location

in Painted turtles ( Chrysemys picta). Futictional Ecology 13: 94-10 l.

Wilson, D. 1998. Nest-site selection: microhabitat variation and its effect on the survivaJ

of turtle embryos. Ecology 79: 1884-1892.

Zar, J. 1999. Biostatístical Analysis. Prentice Hall. New Jersey. 663 p.


34

,J-
I
'l;
--

rrN

----
1 ,$ ,

Figure l. Map ofPalo Verde National Park, northwestem Costa Rica.


35

Table 1. The mean and standard error of the mean for eight variables of microhabitat

measured at nest sites and at random sites in a Kinosternon scorpioides nesting area in Palo

Verde National Park (units and sample size shown in parenthesis).

Variable Nest site Random site

eanopy cover (%) 90.4 ± 0.89 (n = 56) 93.7 ± 0.70 (n = 54)

Understory cover (%) 42.80 ± 3.24 (n = 56) 31.29 ± 3.41 (n = 54)

Leaf litter cover (%) 90.70 + 1.85 (o = 56) 84.35 + 1.88 (n - 54)

Leaflitter deptb (cm) 3.04 + 0.23 (n = 55) 2.0 + 0.15 (n = 51)

Horizontal understmy cover (%) 27.76 + 2.9 (n = 55) 25.94 + 3.25 (n =53)

Ai r temperature ( o e ) 34.82 + 0.17 (o - 52) 35.55 + 0.21 (n =53)

Air hurnidity (%) 43.29 + 0.55 (n = 52) 45.75 + 0.80 (n = 52)

Soil temperature ( o e ) 28.13 + O 12 (n = 53) 29.04 + 0.14 (n =53)


36

i'
':' 25
:11 A
~ p<O.OOOJ
8 20

f t5
-g 10
::J

grazlng no grazing

B
100
~
~ oo p<O.OOOl
8:. 60

..
~
~
"10

20

grazing no graziog

Figure 2. Differences in three variables of microhabitat measured in areas under cattle

grazing and free of cattle grazing in a nesting area of Kinostemon scorpioides in Palo Verde

National Park. A) understory cover (U = 412, p<O.OOOI), B) leaf litter cover (U = 864,

p<O.OOOl), C) leaflitter depth (U = 895, p<O.OOOI).


37

Article 2: Terrestrial movements, activity pattern aod babitat use of the mud turtle

Kinosternon scorpioides in Palo Verde National Par k, Costa Rica.

Abstract

The mud turtle Kinosternon scorpioides is considered a semi-aquatic species known to

engage in substantial overland movements and terrestrial aestivation. To determine the mud

turtles overland daily activity pattern we searched for active individuaJs walking along

transects. To determine the distance that turtles travel from the wetland to the aestivating

sites, sorne individuals found active overland were fitted with a thread-bobbin track:ing

device which we attacbed to the rear of the carapace. Every turtle found was individually

marked, sexed, and measured. A total of 92 observations of K. scorpioides active on land

were made. Two behavíors were identified: crawling and nesting. K. scorpioides shows a

birnodal activity pattem with maximurn activity during early rnorning hours (0600 to 0800

h) and twilight and early evening hours (1600 to 2000 h). We found a total of 41 turtles

aestivating in the forest Turtles were atan average distance (±SE) of 156.38 m ± 13.68

from the border between the wetland and the deciduous forest, ranging from 20.9 to 304.9

m. Adults and males aestivate at greater distances than juveniles and females respectively.

Turtles found aestivating were in three different types of habitats: leaf litter (85%), natural

rock cavities (10%) and tree buttress (5%). In PVNP K. scorpioides is apparently capable

of aestivating for at least 12 weelcs or 25 % of the year, this takes place during the dry

season. We found several traits of K. scorpioides behavior that may account for the abiJity

of this species to withstand desiccation during the aestivation period.

Key words: mud turtle, Kinosternon scorpioides. habitat use, movement, activity, Palo

Verde, Costa Rica.


38

Resumen

La tortuga candado Kinosternon scorpioides es considerada una especie semi-acuática que

puede tncursionar en hábitats terrestres para pasar el estío. Para determinar el patrón diario

de actividad de esta especie registramos individuos caminando a lo largo de tra.nsectos. Para

determinar la distancia que recorren desde el cuerpo de agua hasta los sitios de estivación

se le instaló un carrete de hilo en el caparazón que nos permitió monitorear sus

movimientos. Cada tortuga encontrada fue marcada, medida y se registro el sexo. Hi.cimos

un total de 92 observaciones de tortugas activas en tierra, se identificaron dos

comportamientos: caminar y anidar. Encontramos que K scorpioides tiene un patrón

bimodal de actividad diaria, con picos de actividad a tempranas horas de la mañana (0600 a

0800 h) y al atardecer ( 1600 a 2000 h). Encontramos 41 tortugas pasando e] estío en el

bosque a una distancia promedio (±SE) de 156.38 m ± 13.68 del borde entre e) hwnedal y

el bosque deciduo, el rango fue de 20.9 a 304.9 m. Los adultos y los machos se encontraron

a distancias superiores que los j uveniles y las hembras respectivamente. Las tortugas

utilizaron tres tipos de habitas para pasar el estío: hojarasca o mantillo (85%), huecos

naturales en la roca (lO%) y gambas (5%). En el PNPV K. scorpioides aparentemente tiene

la capacidad para permanecer en letargo fuera del agua por al menos 12 semanas o el 25%

del año. Encontramos varias características del comportamiento de K scorpioides que le

permiten sobrevivir a la deshidratación durante el periodo de estío.


39

lntroductioo

The knowledge about movement patterns is important to understand distribut:ion,

abundance, gene flow, and behavioral interactions among individuals (Shav.n el al, 2000).

The capacity of dispersion of an organism within and between populations is fundamental

to the ecology of the species, and is important to both intra and inter-specific interactions

(Nieuwolt 1996). Studies of activity and movements of freshwater twt1es, and ectothenns

in general, can potentiaHy reveal how variations in environmental parameters (i.e., weather,

temperature} impact energy acquisition, and therefore an organism's hfe history (Rowe

2003). Terrestrial activity of freshwater turtles has been widely documented, suggesting

that it is an important aspect of their ecology. Terrestrial behavior includes movements

associated with nesting, hatchling emergence from nests, movements to or from hibernating

or aestivating sites, and dispersa! of hatchlings, juveniles or males in search of mates

(Morales-Verdeja & Vogt 1997).

The abundance and distribution of a particular species is often deterrnined by the

availability of habitat. This must offer al1 necessary components (f~ water, refuge,. and

nest sites) to ensure the species existence (Litvaitis et al. 1994). Aquatic or semi-aquatic

turtles that experience seasonal fluctuations of their aquatic habitat may be forced to use

terrestrial habitats. Many organisms under these circwnstances will either ma.ke overland

migrations to fmd other bodies of water, or terrestrially aestivate until aquatic habitats are

replenished (Ligon & Peterson 2002). Either way, the terrestriaJ habitat becomes an

important component for the conservation and ecology of these organisms.

Kinosternon scorpioides is a small species of freshwater mud turtle that ranges in

distribution from Oaxaca and Tamaulipas, Mexico to southern Bolivia and northern Peru,

Brazíl and Argentina (Savage 2002). K scorpioides is widely distributed on the Pacific
40

slope of Costa Rica, and is found in many freshwater habitats such as marshes, permanent

and temporary ponds, streams, rivers, and swamps (Acuña 1998, Savage 2002).

Throughout its distributional range, most accounts make referen.ce to the use of aquatic

habitats of this species (Pritchard & Trebbau 1984, Emst & Barbour 1989, Lee 2000,

Savage 2002, Stafford & Meyer 2002). However, in Belize, K scorpioides is known to

move overland from drying ephemeral habitats into permanent streams as the dry season

progresses (Moll & Moll 2004). Similarly, in Costa Rica K. scorpioides is known to make

substantial overland movements (from 175 to 500 m away from the water), and to leave

drying bodies of water and enter adjacent terrestrial habitats to aestivate (Acuña 1998,

Teska 1976). AJthough it is considered a semi-aquatic species that inhabits mostly

ephemeral aquatic habitats and is known to engage in overland movements and terrestrial

aestivation (Acuña 1998, MoJI & Moll 2004, Savage 2002), no research has been devoted

to study specifically the terrestrial movements an.d terrestrial habitats ofthis species.

Materials and metbods

Palo Verde National Park (PVNP), is located in the lowlands ofthe northern Pacific versant

ofCosta Rica, Guanacaste Province (lO 0 21' N- 85 ° 21' W). The park has an extension of

19,800 ha, the average temperature is 27 °C, and the annual precipitation is 2,296 mm (li'ig.

1). The dry season lasts from December to April, and the rainy season from May to

November. A great variety of habitats are represented in this park: primary and secondary

deciduous dry forest, riparian forest, savannas and wetlands (Hartshom 1983). The wetJand

or lagoon accounts for more than 60 % of the park' s area. It is formed by a mosaic of

aquatic and woody vegetation. During the rainy season, the lagoon's water leveJ can rise up
41

to 1.3 m, but it gradually reduces until it dries out completely at the end ofthe dry season

(ApriJ and May).

During the dry season, as the lagoon water level diminishes, mud turtles are

commonly seen rnoving over land (Acuña 1990). To determine the mud turtles overland

daily activity pattem we searched for active individuals walking along three transects. One

transect was located along a trail east of the MINAE station (about 500 m long), another

one was located along a dirt road between the OTS station and the MINAE station (about

900 m long), and another on.e along a dirt road between the MlNAE station and "Puerto

Chamorro" (about 4000 m long) in the Tempisque River (Fig. 1). AIJ transects were

oriented perpendicular to the border between the wetland and the dry forest. The search. was

done during the dry season from February to May 2004, and fonn 0601 to 2400 h. The day

was divided into nine periods of2 h each (i.e. from 0601 to 0800 h~ from 0801 to 1000 h;

from 2201 to 2400 h). To detennine how the abw1dance of K. scorpioides moving overland

varied throughout the day, we kept records ofthe search effort invested, and the number of

active turtles found for each 2 h period. Thus, the abundance of active turtles was given by

the number of individuals found moving per unit of time invested searching each of the

nine periods. Every turtle found active was individually marked, sexed, and measured The

direction in which every turtle was traveling was also recorded. Measurements taken were:

curve carapace length (CCL), curve carapace width (CCW) in cm to the nearest O.J, and

weight in g to the nearest 0.1 (using eíther a 300 g ora 1000 g capacity Swíss Pesola). The

turtle activity pattern was compa.red to the ambient temperature and burnidity in PVNP.

The OTS station in PVNP has a weather station that automatically takes readings every 30

rnin. This allowed us to construct an average profile of temperature and hwnidity from

0601 to 2400 h from February to May 2004.


42

To determine the distance that turtles travel from the wetland to the aestivating sites,

sorne individuals found active overland were fitted ~ith a thread-bobbin tncking device,

which we attached to the rear of the carapace (Wilson L998). Turtles equipped with thread

bobbins were released within a few hours in the same 1ocation where they were captured.

We attached the free end of the thread-bobbin to the vegetation, which allowed the thread

to unwind as the turtle walked away. Twenty-four h later we followed the thread line until

tínding tb.e turtle. We also sea.rch foc acstivating turtlcs in the Jea( letter usin.g a han.d rake.

Additionally, we searched for aestivating turtJes in tree buttresses and rock bwTOws.

Searches for aestivating turtles were carried out to a maximum distance of 350 m from the

wetland. We detennined the geographic location (using a Gannin Geko 201 GPS) for each

site where a turtle was found aestivating. Each GPS point was plotted on an Ikonos satellite

image ofPVNP; the straight line distance from the border between the wetland and the dry

forest to each GPS point (aestivating síte) was measured in m using ArcNiew 3.2. All

turtles equipped with thread bobbins moved mostly in straight Iines (south to north) from

the wetland to the aestivating sites in the forest. Thus a straight line movement was

assumed for aH turtles found aestivating (turtles found by raking or active search in tree

buttress and rock burrows). Each turtle found aestivating was individually marked, sexed

and measured (see above). The microhabitat or refuge selected by turtfes to aestivate was

described, temperature and humidity readings were taken insíde and outside aestivating

refuges during the hottest time of the day (1 100 to 1400 h). For air temperature and air

humidity we used a digital hygro-thermometer model 445582 Extech Instruments, and for

soil temperature we used a HI-145-30 Hanna fnstruments digital thennometer.

AH turtles were divided in two age classes: juveniles and adults. To determine age

class, we use the following criteria. According to Savage (2002), the smaJlcst mature
43

female known of K. scorpioides is 12.2 cm carapace standard length (CSL). Since we only

have CCL for aestivating turtJes, we measured the CSL and tbe CCL of 103 mud turtles

from PVNP. The mean difference between the measurements was 2.29 cm. Thus, we

considered all individua1s measuring Iess than 14 cm CCL to be juveniles, and a11

measuring 14 cm or more to be adults.

Results

We searched for active mud turtles for a total of 57 days (16 days in February, 15 days in

March, 14 days in April, and 12 days in May). During these days, we actively searched for

turtles for a total of 233 h. We found a total of 76 mud turtles actively moving overland.

When we arrived at PVNP on 9 February, turtles were already leaving the wetland and

entering the forest to aestivate. From February to April all observed turtles were movjng

from the wetland towards the torest (south to north). During February, 12 female turtles

were seen moving from the wetland to the forest to nest, after nesting they went back to the

wetland. During April, only three turtles were found moving actively over land (south to

north), and aH three individuals were found on April 21st during and after a single and

isolated episode of heavy rain. During May only one turtle was tound moving north

(towards the forest), but starting on May 24th all active turtJes were moving back frorn the

forest to the wetland (north to south). March registered the highest overland activity and

April the Jowest (Fig. 2). During the rainy season (between October 10th and 16th 2003, and

between October 30th and November 7th 2004), we found 16 K. scorpioides active overland.

Fourteen of these individuals were females, four of which were nesting temaJes~ one femaJe

was found buried under Ieaf litter in November 2004 with mud accumulation in the

posterior marginals, which is indicative of recent nesting. The two maJes found on Jand

were less than one m away from the water. However, females were far from water {sorne
44

more than 200 m). A total of 92 observations of K. scorpioides active on land were made.

Two behaviors were identified: crawling and nesting. No feeding or social behavior was

seen on land. All turtles reacted by closing their plastron on approach. After five to lO min

of manipulation (while measuring the carapace or mark:ing the individual) most turtles

empty their bladder generating a strong musk odor.

K. scorpioides in PVNP shows a bimodal activity pattem (Fig. 3) with max:imum

activity during early morning hours (0600 to 0800 h) and twilight and early evening hours

(1600 to 2000 h). K. scorpioides daily activity pattem was positively related to relative

humidity (.-2 = 0.46, p = 0.01, n = 12) and negatively related to ambient temperature (r2 =

0.45, p = 0.01 , n- 12) during day light hours (from 0600 to 1800 h). However, during rught

hours (from 1800 to 2400), although ambient temperature continued to drop and humidity

continued to increase, turtle activity diminisbed (Fig 3 & 4).

We found a total of 41 turtles aestivating in the forest. Turtles were atan average

distance (±SE) of 156.38 m ± 13.68 from the border between the wetland and the deciduou.')

forest, ranging from 20.9 to 304.9 m. Adult individuals were at a significant greater

distance tban juveniles (Mann-Whitney O = 274.5, p = 0.009). Adults (n = 28) were at

183.17 m± 15 (x ± SE), and juveniles (n = 13) were ata distance of 98.66 m± 21.26 (x ±

SE ). Males (n = 13) were ata distance of 190.69 m± 92.26 (x ± SD), and females (n = 19)

were ata distance of 163.2 m± 80.5 (x ± SD). However, no significant difterences were

found between maJes and females (Mann-Whitney U = 94.5, p = 0.26). Most juveniles were

found in the frrst 150 m from the wetland, females were more evenly distributed from O to

300 m, and males had a tendency to occupy the farther distance categories (Fig 5). The

greater number ofturtles was found between 50 and 100m from the wetland, and only one

individual was found between 300 and 350m. Aestivating turtfes range in size from 8.5 to
45

21.4 cm CCL, from 7.7 to 18.9 cm CCW, and from 59 to 870 g of weight There is a

positive relation between turtle size and distance from the wetland to aestivating sites

(CCL: .-2 = 0.2, p - 0.003; CCW: e = 0. 18, p = 0.005; weight: .-2 = O.l5, p = 0.009). In
general, smaller turtles were found mostly at a short distance from the wetland, but large

turtles were found at any distance (close or far) from the wetland.

TurtJes found aestivatíng were in three difl'erent types of habitats: leaf litter, rock

cavíties and tree buttress. Most turtles (85 %) aestivate under the leaf li.tter, these

individuals construct a shallow fonn in the ground (about 5 cm deep) that covers the

plastron and the lower hatf of the carapace, the rest of the shell is covered by leaves. Ten

percent of the turtles were found aestivating inside natura] lime stone cavities, and five %

were found in tree buttress. We measured temperature in a total of l3 sites, and humidity

in a total offive sites (inside and outside aestivating refuge). During the hottest time ofthe

day (1100 to 1400 h), the refuge selected by turtles to aestivate maintained lower

temperatures and higher humidity than the envirorunent extemal to the refuge. Average

temperature was 29.59 °C ± 0.5 (x ±SE) and 34.31 °C ± 0.57 (x ±SE) inside and outside

refuges respectively. Average hwnidity was 59.9% ± 6.18 (x ±SE) and. 41.14% ± 3.84 {x

± SE) inside and outside refuges respectively.

Discussion

During February 1974, eight ofnine actively moving K. scorpioides found in PVNP. were

crawling northward or moving away from the water and toward.s the forest, turtles were

seen buried under leaf litter, and most active individuals were found before 900 h and at

1730 h (Teska 1976). Acuña (1990) reports that during the dry season in PVNP many K

scorpioides are seen leaving the wetland and entering the forest to aestivate, when the qtiny

season starts, turtles Me seen mo~'ing back ta the wetJand. This is consistent witb our
46

findings. During our study period (February to May) all individuals found moving overland

were heading to the forest and away from the wetland, it was not until the end ofMay with

the first rains ofthe rainy season, that we started seeing turtles movin.g back from the forest

to the wetland. Under these circumstances, the turtles that abandoned the wetland during

February would have to spend about three months (12 weeks) aestivating in the forest.

Moreover, since the dry season in PVNP begins in December, it is possibJe that sorne

individuals migrated during January. The only source of water in the forest at the study area

(a water hole) was completely dry by the end of March, the water hole was constantJy

monitored and only two K scorpioides were seen in it·during the study period before it was

dry. Thus, the turt]es that migrated to the forest do not have access to water until they return

to the wetland in May or during a casual rain in April. According to Peterson & Stone

(2000) sorne populations of Kinosternon sonoriense are physiologically capable of

surviving many weeks without food and water. When individuals of this species were

experimentally subjected to l l weeks of deprivation, all but one recovered on rehydration.

However, by the end of the ll weeks several blood chemistry parameters had increased

anhomeostaticalJy to very high levels. In a similar work with the species Kinosternon

flavescens, K. hirtipes, and K. sonoriense, Ligon & Peterson (2002) found that severa!

individuaJs had to be rehydrated after only 30 days (about 4 weeks) of deprivation, while

other individuals were able to survive 55 days (about 7.5 weeks) without food and water,

two turtles died during the experiment. Thus, sorne kinosterníds are capabJe of surv1ving

long periods of deprivation, but not without suffering certain level of physiological stress.

In PVNP K. scorpioides is apparently capable of aestivating for at least 12 weeks (84 days ).

However, during this study most turtles abandoned the wetland during March~ reducing the

aestivating period to about 9 weeks (63 days). During ApriJ and May we found many
47

individuals of K. scorpioides buried under mud and under dry aquatic plants in the wetland.

Hence, not all turtles abandon the wetland and enter the forest to aestivate. Acuña (1990)

also reported this dual behavior in the population of K. scorpioides in PVNP. Although, in

PVNP the dry season starts in December, in a normal year the wetland holds water until

March and sorne scattered and small puddles remain even until the first half of April. Thus,

turtles that remain in the wetland would be able to shorten significantly the ínactivity and

deprivation period, probabl:y to only four to five weeks (30 days). Why sorne individuals

start to rnigrate to the forest during February when there is stiH plenty of water in the

lagoon, and what factors determine which individuals stay in the wetland and wbi.c h

migrate to the forest is unknown.

In PVNP, during the dry season apparently K. scorpioídes engages in terrestrial

activity with only two purposes: to aestivate or t~ nest. Feeding and social behavior were

never observed in this species while on land However, terrestriaJ activity was also

registered during the rainy season. During the rainy season no juveniles were seen on land,

most individuals were females (sorne related to nesting activity), and the two maJes seen on

land during the rainy season were very close to the water (less than 1 m). Hence, it is

possible that nesting is the main purpose of terrestrial movements that take place during the

ramy season.

Because ofthe physiological constraints of ectothenny, arnbient parameters such as

temperature and humidity rnay have an important effect on activity pattems of turtJes.

Activity level was found to be negatively related to arnbient temperature in desert box

turtles (Terrapene ornata luteola), resulting in a bimodal activity pattem during the

summer, in which turtles were active in the rnorning and aftemoon and inactive during

rnidday. Terrapene ornata ornata exhibited a similar activity pattem, with rnost activity
48

taking place from 0600-1000 h and from 1600-1800 h, and the midday inactive period was

lengthened during the hottest portion of the summer (Converse & Savidge 2003). Dming

the dry season i.n PVNP, the average daily temperature exhibits a variation of 65 °C,

changing from a low of26.7 °C in early morning and late afternoon hours to 33.2 °C during

midday hours. Additionally, average daily humidity varies from 60 to 71 % i.n late

aftemoon and early moming hours to 44 % during mid.day hours. Thus, the bimodal

activity pattem exhibited by K. scorpioides in PVNP is a convenient mechanism to reduce

the risk of dehyd.ratation and avoid extreme temperatures while migrating over]and during

the dry season.

Ligon & Peterson (2002) concluded tbat the differences in capacity of estivation

among populations of three different species of Kinosternon are primarily li.nked to variable

behavioral responses to dry conditions. Ectothennic species must resort to behavioraJ

thennoregulation. As selection of optimal microclinÚltes is an important compon.ent of this

behavior, temperature and moisture should be driving factors in the selection of

thennoregulatory sites such as shallow excavated pits (known as fonns) used by inactive

individuals ofsome species ofturtles (Converse & Savidge 2003). Peterson & Stone (2000)

found that individuals of K. sonoriense subject to dry conditions used water stored in the

urinary bladder to osmoregulate, and individuals that were surface active became more

dehydrated than those that remained buried. Risk of dehydration may be reduced

behaviorally by mínimizíng activity, reducing operative surface area, and selecting

aestivation sites that maintain relative low temperatures and high relative humidity (Ligon

& Peterson 2002). In PVNP, K. scorpioides may use terrestria) habitats for up to 25 % of

the year (three months). Moreover, the use of terrestrial habitats takes place mostly during a

critica! period for this species (the d.ry season), and this may extend more than normal
49

during d.rought years. Thus, the behavioral responses of an individual of this species in

PVNP may determine its success in surviving to the next rainy season. In PVNP we fotmd

several traits of K. scorpioides behavior that may account for the ability of this species to

withstand desiccation during the aest'ivation period (prompt by an extended dry season and

the reduction or disappearance of their aquatic habitat): 1- the use of aestivating refuges

with lower temperature and higher humidity than the environment~ 2- a bimodal daily

activity pattern that ensures that the turtle overland movement occurs during hours with

optimum conditions of ambient temperature and humidity; 3- juveniles (which are more

prone to dehydration than adu1ts) travel shorter distances than adults from the wetland to

aestivation sites; and 4- tbe storage of water in the urioary bladder.

Literature cited

Acuña, R. 1990. E l impacto del fuego y la sequía sobre la estructura de la población de

Kinosternon scorpioides (Testudines: Kinostemidae) en Palo verde, Guanacaste,

Costa Rica. Brenesia 33:85-97.

Acuña, R. 1998. Las tortugas continentales de Costa Rica Universidad de Costa Rica. San

José. 92 p.

Converse, S. & J. Savídge. 2003. Ambient temperature, activity, and microhabitat use by

omate box turtles (Terrapene omata omata). Joumal of Herpetology 37(4):665-

670.

Ernst, C. & W. Barbour. .1989. Turtles ofthe World. Smithsonian. Washington. 313 p.

Hartshom, G. 1983. Introduction, p. ll9-186.In D. Janzen (ed.). Costa Rica naturaJ history.

Chicago. Chicago.
50

Lee, J. 2000. Amphibians and reptiles ofthe Mayan World. Comen University. New York.

402p.

Ligon, D. & C. Peterson. 2002. Physiological and behavioral variation in est:ivation among

mud turtles (Kinosternon spp). Physiological and Biochemical Zoology

75(3 ):283-293.

Nieuwolt, P. 1996. Movement, activity, and microhabitat selection in the westem box

turtle, Terrapene ornala /uteola, in New Mexico. Hepetologica 52(4):487-495.

Moll, D. & E. Moll. 2004. The ecology, exploitation, and conservation of river turtles.

Oxford. New York. 393 p

Morales-Verdeja, S. & R. VogL 1997. Terrestrial movements in relation to aestivation and

the annual reproductive cycle of Kinostemon leucostomum. Copeia 1997(1):123-

130.

Peterson, C. & P . Stone. 2000. Physiological capacity for estivation of the sonoran rnud

turtle, Kinosternon sonoriense. Copeia 2000(3):684-700.

Pritchard, P & P. Trebbau. 1984. The turtles of Venezuela. Society for the Study of

Amphibians and Reptiles. New York. 403 p.

Rowe, J. 2003. Activity and movements of midland painted turtles (Chrysemys picta

marginata) living in a small marsh system on Beaver lsland, Michigan. Joumal of

Herpetology 37(2):342-353.

Savage, J. 2002. The amphibians and reptiles ofCosta Rica. Chicago. Chicago. 934 p.

Shawn, L., C. Haas & J. MitchelL 2000. Movements and activity of Bog turtles (C1emmys

muhlenbergii) in southwestem Virginia. Joumal ofHerpetology 34(1):75-80.

Stafford, P. & J. Meyer. 2000. Reptiles ofBehze. The Natural History Museum ofLondon.

Academic. London. 356 p.


51

Teska, W. 1976. Terrestrial movements of the mud turtle Kinosternon scorpioides in Costa

Rica. Copeia 1976(3):579-580.

Wilson, D. 1998. Nest-site selection: mícrohabitat variation and its effect on the survival

of turtle embrvos. Ecology 79(6): 1884-1892.


52

n
N

Figure 1. Map ofPalo Verde National Park, northwestem Costa Rica.

O Total
lo Females
1
• Males

5
o -1--·.__--'--~

February Match April May

Figure 2. Nurnber ofmud turtles (Kinosternon scorpioides) active on land from

February to May 2004 in Palo Verde National Park.


53

0.8 1
,~ 0.7 -
•.. - .... - abundance
_.,..._ temperature
.~
e '
::1
el> 0.6

~
el> 0.5
t::1
....
.~ 0.4 #
#

.. ..
! ~ 0.3 •' ,
,t
.. ...
... - ..
1

a
~ 0.2
;:l
'
' \ , -- .rl

~ 0.1

0 -·~--~-----r----~----r---~~--~-----r----+---~ ~
600-800 BOO·fJOO 1>00-1200 1200-1400 1400-'600 '600-~ 'e00-2000 2000-2200 2200-0000
Time ofday

Figure 3. Mud turtle (Kinostemon scorpioides) daily activity pattem related to ambient
temperature in Palo Verde National Park.

O.S l - ... - abundance


0.7 • -+--hwnidit.y%
Z'
·¡:::
~ 0.6 {
e 1
¡ 0.5 l
t: 1
.a
0.4 1
....>-
el>

ü
.$., . ....
u
e
Q) 0.3 .. ...
~
e 02
=
~
0.1 L
T45
1
o - -1 --~~----~----r---~ •o
6:00-8:00 8:00-1):00 1l:Q0-12:00 12:00-14:00 14:00-'6:00 ti:Q0-1!:00 tl:00-20:00 20:00-2200 22roo:oo

Time ofday

Figure 4. Mud turtle (Kinosternon scorpioides) daily activity pattem related to ambient
relative humidity in Palo Verde National Park.
54

100 • Males
90 O Females
80 O Juveniles

--
~
¡¡;....
70
60
~

= 50 ~
~
=
i
""
¡;...
40
30
j
20
10
j
o
0-50 50-100 100-150 150-200 200-250 250-300 300-350
Distance (m)

Figure 5. Frequency of aestivating individuals of Kinosternon scorpioides by distance

category, sex, and age, in Palo Verde Natíonal Park.


55

Article 3. Tbe impact of fire on tbe population of mud turtles .Kínosternon scorpioides

(Testudines: Kinosternidae) in a tropical wetland of northwestern Costa Rica.

Abstract

Turtle populations have been declíning at an alanning rate throughout the worJd in recent

years. The freshwater turtles of Costa Rica are threatened by habitat destruction., tires~ and

over hunting. Wild fires and controlled fires are common in the wetland of Palo Verde

National Park (PVNP). The purpose of controlled fires is to eliminate or reduce the areas

covered by the invasive species of aquatic plant Typha domíngensis. Although extensive

areas of this plant have been bumed in the past years, the impact of fire on the population

of any vertebrate has not been assessed. We gathered infonnation on turtle mortality caused

by tires that occurred from May 2003 to May 2005 in PVNP. After the fires we located

dead turtles by walking strip transects along the recently bumed areas. the turtles found

were measured and age and sex was also noted We found significant differences in dead

Kinosternon scropioides density among the different habitats affected by tires (ANOVA F

s. 1s = 14.41, p < 0.0001), the mean density of dead turtles was highest in small patches ofT.

domingensis and areas of diverse dry aquatic vegetation (29.02 turtles/h~ 23.71 turtleslh~

respectívely). A total of 370 turtles were found through the study period. Results indicate

that fires are killing a greater proportion of adults than j uveníles, and individuals of greater

size. The cattail invasion probably changed many aspects ofthe wetland ecology in PVNP,

including mud turtles aestivating behavior, tire frequency, and fire intensity. We concluded

that this new tire regimen in PVNP !s introducing high mortality rates in age and size

classes accustomed to hígh survival rates. This wiiJ likely diminish the popuJation

reproductíve output and stability.

Key words: mud turtle, Kinosternon scorpioides, fire, mortality, Palo Verde, Costa Rica.
56

Resumen

Las poblaciones de tortugas han declinado de forma alannante en los últimos años. Las

tortugas de agua dulce de Costa Rica están amenazadas por la destrucción del hábitat,

incendios y caceria. Durante la estación seca los incendios son comunes en el humedaJ del

Parque Nacional Palo Verde (PNPV). El propósito de los incendios controlados es eliminar

o reducir las áreas dominadas por Ja planta acuática invasora Typha domingensis. Aunque

extensas áreas han sido quemadas en los últimos años, el impacto del fuego en las

poblaciones de vertebrados no ha sido evaluado. Se obtuvo datos sobre .l a mortaJidad de

tortugas provocada por el fuego en incendios que ocurrieron entre mayo del 2003 y mayo

del 2005 en el PNPV. Se ubicaron y cuantificaron las tortugas muertas a lo largo de

transectos de franja en las áreas recientemente quemadas, se registro el tamaño, sexo y edad

de las tortugas. Encontramos diferencias significativas en la densidad de tortugas muertas

de Kinosternon scorpioídes entre los diferentes hábitats afectados por el fuego (ANOVA F

s. ts = 14.41 , p < 0.0001). La densidad media de tortugas muertas fue mas alta en pequeños

parches de T. domingensis y en áreas de vegetación acuática seca (29.02 tortugas/ha, 23.7 L

tortugas/ha, respectivamente). Encontramos un total de 370 tortugas durante el estudio. Los

resultados indican que los incendios afectan principalmente a la porción adulta de la

población y a individuos de mayor tamaño. La invasión de T. domingensis probablemente

cambio muchos aspectos de la ecología del humedal en el PNPV, incluyendo el

comportamiento de estivación de K scorpioides, y la frecuencia e intensidad de los

incendios. Concluimos que este nuevo régimen de incendios en el PNPV está introduciendo

una alta mortalidad en clases de edad y tamaño que normalmente tienen altas tasas de

sobrevivencia, deteriorando la capacidad reproductiva y la estabilidad de la población.


57

Introduction

Freshwater turtles play ímportant, but incompletely understood roles ín both the ecology of

their respective ecosystems, and in the economy and sociology of human culture (Moll &

Moll 2004). Unfortunately, turtle populations have been declining at an alarming rate

throughout the world in recent years (Emst & Barbour 1989, Moll & Moll 2004). At

present, the freshwater turtles of Costa Rica are threatened outside and inside protected

areas by habitat destruction, fires, and over hunting (Acuña 1990, Mora & Ugalde 1991,

Acuña 1998).

An example of tbe above is Palo Verde National Park (PVNP) in northwestem

Costa Rica. Pires are common in the wetland of PVNP since the mid 80's. Fires kill a

considerable amount of mud turtles in this protected area (Acuña, 1990; Ulises Chavarria

pers. com. ). However, to what extent fires are affecting turtle populations has not been

studied in detail. To understand the ecological impact of tires, we addressed the following

questíons: a) how many turtles do the tires kili?; b) is the mortality caused by fires the same

acros habitats?; e) are tires affecting all portions of the population randomly with respect to

age, size, and sex?

The mud turt1e (Kinosternon scorpioides) is a small species of freshwater turtle

widely distributed on the Pacific slope of Costa Rica. Pernales of this species nest from

February to March, and from August to November. Females diga hole 2.1-4.6 cm deep in

the ground and Jay between one and six hard brittle shelled eggs. Shells are about 40 x 19

mm in length, and hatchlings are 30- 40 mm in length. Incubation takes from 115 to 128

days. This turtle is carnivorous and feeds mostly on macro invertebrates such as insects,

worms, and snails, and vertebrates such as fishes and tadpoles. lt may also behave as a

scavenger (Acuña 1998, Savage 2002). Limbs are developed for bottom crawlin& but
58

always have sorne toe webbing (Emst & Barbour 1989). Temperature-dependent sex

detennination has been reported for at least six species in the family Kinostemidae,

including K. scorpioides (Ewert & Nelson 1991, Janzen & Krenz 2004).

Materials and metbods

We conducted our research at Palo Verde National Park (PVNP), located in tbe lowlands of

the Pacific versant of northwestern Costa Rica, Guanacaste Province ( 1O 0 21' N - 85 21' °
W, Fig. 1). The park has an extension of 19,800 ha, the average temperature is 27 C and °
the annual precipitation ís 2,296 mm. The dry season extends from December to April and

the ra.íny season from May to November. A great variety of habitats are represented in this

park: primary and seco~dary deci.duous dry forest, riparian forest, savannas and wetlands

(Hartshom 1983). The wetland or lagoon accounts for more than 60% ofthe park's area.

The wetland is formed by a mosai.c of vegetation including patches of Parkinsonia ac1deata

(a small woody tree), patches ofthe aquatic plant Typha domingensis (cattail), pastures, and

patches of di verse aquatic vegetation that grows during the rainy season and dies in the dry

season forming a flat and thick mat of dry vegetation over the ground. During the rainy

season the lagoon's water level can rise toa depth of 1.3 m, but it gradually recedes untit it

dries out completely at the end of tbe dry season (April and May). PVNP is one of the few

protected areas legally subjected to active management, such as cattle grazing, controlled

fires, and mechanical crushing ofvegetation. The purpose ofthese actions in the wetland is

to elirninate or reduce the areas covered by an invasive species of the aquatic plant Typha

domingensis. This plant started to expand during the 80's, until it covered most of the

wetland by the m id 90's. When no more areas of open water were left~ the populations of

migratory and resident birds declined (Bumidge 2000, Barboza 2002, MINAE: R-009-

ACA-D-2002). Apparently, the rnanagement actions in PVNP have shown to be beneficia!


59

for bird populations (Trama 2005). However, no information is available on the impact of

these actions on any other species of vertebrate within the park.

In PVNP, the risk of fires is high at the end of the dry season. and these fires are

triggered mainly by illegal deer hunters. However, during the previous three years, fires

were also set by the Park's administration in the wetland to intentionally burn areas of T.

domingensis in an attempt to eliminate this plant. Extensive fire brake lines are made, but

because of the high speed of the wind and the subterranean root system of T. domingen..¡is,

control fires get out of control and often burn extensive areas of all types of vegetation.

During our study the fires affected five different types of habitats within the wetland: L

Areas dominated by P. aculeate; 2. Areas dominated by diverse aquatic vegetation (DA V)

that during the dry season forms a thick mat of dry vegetation over the ground. This habitat

may also include areas where T. domingensis was crushed by mechanical actions during the

rainy season. resulting in a thick mat of dry cattail over the ground during the dry season; 3.

Extensive areas (> 200 ha) dominated by T. domingensis; 4. Small patches (lO- 50 ha) of

T. domingensis (these patches are formed by the mechanical crushing of extensive areas of

T. domingensis as part of the management actions to eliminate this plant, a tractor crushes

extensive areas leaving scattered, small, and isolated patches of cattaíls); and 5. Pastures.

One fire event also affected a secondary gallery forest outside the wetland

To assess the impact of fires on the population of K. scorpioides, we gathered

infonnation on turtle mortality caused by fire from five different fire events at PVNP: 1)

May 2003 at "Sector Palo Verde", 2) April 2004 at "Varillal", 3) May 2004 at "Sector Palo

Verde", 4) April 2005 at "Sitio Azul", and 5) May 2005 at "Sector Palo Verde" (Fíg 1).

After the fires, we quantified the amount of dead turtles by walking along strip transects

throughout the burned areas, we measured the transect Jength in meters using a GPS
60

(Garmin, Geko 201) and selected randorn bearings using a compa.c;s_ Thus, transects were

randomly placed on the bumed areas_ The length and width of tbe transects \vas variable,

depending on the size of the burned area and the kind of habitat A total of 24 transects

were placed on the bumed areas. Transects varied from 290 to 884 m long and from 4 to 10

m wide. All dead turtles found fresh and with evident signs of burns were measured in cm

to the nearest 0.1 (curve carapace length CCL and curve carapace width CCW), and age

and sex was al so recorded. These individuals were considered to represent the portion of

the popu1ation that died due to fues. Using the total area of every transect, we estimated the

dead turt1e density for all five habitats represented in the wetland describe ahove, and also

for the secondary gallery forest During our study period, all fires occurred at the end ofthe

dry season (April and May) and according to tbe park's admioistration, tbe great majority

of ftres since the mid 80's have occurred at this time of the year (Rector Calvo pers. coro)-

At PVNP K. scorpioides shows a migratory behavior, every year during the dry season part

of the population leaves the wetland and enters the forest to aestivate_This portion of the

population survives the tires that affect the wetland because most of the turtles leave the

rnarsh during February and March, and remain in the forest until the end of May when the

rainy season begins (pers.Obs. ). In view of the above, during the dry season of 2004 and

2005, we collected data on densíty, CCL, CCW, age and sex of turtles aestivating in the

forest, and. considered these individuals to represent the portion of the population that

survived the tires. The density of turtles in the forest was estimated using 100 m2 plots that

were located randomly.

All turtles were d.ivided in two age classes: juveniles and. adults. To determine age

class, we used the following crite.ria. Accord.ing to Savage (2002), the smallest mature

female known for K. scorpioides is 12.2 cm carapace standard length (CSL)- Since we only
61

have CCL for dead turtles, we measured the CSL and the CCL of 103 Ji ve mud turtles from

PVNP. The mean difference between the measurements was 2.29 cm. Thus, we considered

all individuals measuring less than 14 cm CCL to be juveniles, and all measuring 14 or

more to be adults.

The objective of this research was to study how the fires are atlecting the abundance

and structure of the population of K. scorpioides in PVNP, and if the turtle's mortality

caused by tires is similar in all habitats. We used a one-way analysis of variance (ANOVA)

to test for differences in density of dead turtles among the different habitats affected by fire.

To detennine if fires are randomly affecting males and females, juveniles and adults we

used a Chi-Square test. Finally, we used a t-test to compare CCL and CCW between the

turtles that were killed and the ones that su.rvived the fires, to determine if fires are

randomly affecting the turtle's population with respect to size. For the data analysis we

used Systat 9.0., and Statistica 6.0.

Results

The mean density of dead turtles was highest in small patches of T. domingensis and areas

of DAV (29.02 turtleslha, 23.71 turtleslha, respectively). Patches of 1: domingensis range

in density from 19.77 to 40.22 turtleslha, and areas ofDAV range in dens1tyfrom 18.0 to

29.24 turtleslha. The density of dead turtles was zero in areas of pastures, and of

intermediate values in areas of P. aculeata and extensive areas of T. domingensis (4.37

turtleslba, and 9.06 turtleslha, respectively). Outside the wetland, dead turtle density was

5.06 turtles!ha in the secondary gallery forest. We found significant differences in dead

turtle density among habitats affected by frres (ANOVA F 5, ts = 14.41, p < 0.0001). Thus,

fires are not affecting equally the populatión of turtles in all habits (Fig. 2). A total of 13

plots (1002 m each) were randomly place in the forest (areas not affected by fires), the mean
62

number of live turtle per hectare in this habitat was 84.61 ± 38.97 (x±SE). The overall mean

number of dead turtles per hectare in the wetland was 14.16 ± 2.86 (x±SE).

A total of 370 turtles were found through the study period_. 120 live turtles and 250

turtles killed by tires. Sex was detennined in a total of315 turtles, 100 live and 215 dead

Sex ratio was 2.33:1 (females to males), and 1.86:1 (females to mates) in li.ve and dead

turtles, respectively. However, resu1ts indicate that the proportion of the population of K

scorpioides affected by fires is independent of sex (X2 := 0.73, p = 0.39). Age was

determined in a total of 367 turtles, 120 live turtles and 24 7 turtles killed by fires. Results

indicate that the proportion of the population of K. scorpioides affected by fires is not

independent of age class (X2 = 13. 77, p < 0.000). The sample of dead tmtles has a greater

proportion of adults in com¡nrison to the sample of live turtles (Fig. 3). Thus, fires did not

randomly affect the population of mud turtles with r~spect to age.

We measured the CCL for a total of 367 mud turtles and the CCW for a total of 370

mud turtles. We found significant differences in CCL between the turtles that died dueto

fires, and the turtles that survived tires, but no significant differences were found in CCW

(Table 1). Turtles killed by fires were of greater size than turtles that survived. Wben males

and females were analyzed separately, we found significant differences in CCL and CCW

between males killed by fires and males that survive (Table 1). Male turtles killed by fires

were of greater size than males that survived. No significant differences were found in size

between females that were killed by fires and femaJes that survived (Ta.ble 1).
63

Discussion

The low density of dead turtles found in extensive areas of T. dominguensis compared to

the high density found in small patcbes or islands of T. domingensis and DAV (Fig. 4) can

be explained by several traits of the behavior of K scorpioides: l. during the rainy season

the small patches of T. domingensis are surrounded by either open water or diverse aquatic

vegetation ( which during the dry season gradually turns into DAV). These conditions

would probably favor the abundance and diversity of fish and macro invertebrates.

Considering the food habits of K scorpioides, turtles are most Jikely to remain in these

areas. 2. Mud turtles are considered "bottom walking" turtles (Moll & Moll 2004). K.

scorpioides in particular is adapted to walk in the bottom of rivers and lagoons (Acuña

1998). Thus, during the rainy season the areas dominated by T. domingensis woutd not be

favorable for this spectes, since it would be extremety dífficult for a "bottom walker" to

move through the thick and dense stems of cattails. On the other hand, the areas of diverse

aquatic vegetation are composed manly of floating plants that leave free most of the

bottom. Thus, areas of open water and of diverse aquatic vegetation would most likely be

favorable for a bottom walker such as K. scorpioides. 3. When the lagoon begins to dry out

during the dry season, turtles are most likely to enter the small patches of cattail to

aestivate, or seek shel~r under the thick mat of dry aquatic vegetation (DA V). The few

individuals found dead in extensive areas of T. dominguensis were always in the border

zone between this and the areas of open water and DAV (Fig. 4). As mentioned above, the

small patches of 1: domingesis are the result of mechanical crushing with a tractor during

the rainy season, later during the peak of the dry season these patches are bumed by the

personnel of the park's administration (MJNAE and OTS). We consider this practice to be

the most detrimental for turtle populations in PVNP. When islands of cattail are created by
64

mechanical crushing, these become favorable aestivating habitat and will be packed with

turtles by the end of the dry season (Fig. 4), thus increasing turtle mortality caused by fire.

Our recommendation is not to burn two kinds of habitats in the wetland: smaU patches of Z:

dominguensis, and areas oíDAV. If patches of cattail are to be created ( which may increase

habitat heterogeneity and offer refuge for many species), these should never be bumed or

eliminated in anyway, especially not during the d.ry season.

The wetland in PVNP was fonnerly an open body of water during the 70's, but it

was completely covered with T. domingensis by the end of the 80's, and the probJem

persists today. Because we did not know the behavior ofthis turtle before this phenomenon

occurred, it is difficult to determine how tires impacted the popuLation ín the past. One

possibilíty is that before the invasion of cattail when the marsh dried out and it was bare-

ground, a greater number of turtles mígrated to th~ forest to estivate, because there was

líttle or no vegetation cover under which they could estivate in the marsh. But during the

80 ·s, as the marsh was gradually covered wüh vegetation, there were an increasing number

of turtles that stayed (aestivated) in the marsh duríng the dry season. Consequently, the

situation in PVNP regarding the turtle mortality caused by fires can also be seen as a

problem of invasive species. Fire as a d.isturbing force is itself influenced by the biota,

particular!y the plant community. Alteration of the vegetation by any factor can influence

the nature of a subsequent fire (Whelam 1995). The cattail invasion probably changed

many aspects ofthe wetLand ecology in PVNP, including mud turtles aestivating behavior,

fire frequency, and fire regimen. Pires in tlte marsh usé to Oé táre, back when the wet1and
had no fuel material (Torres pers. com.; Triguero pers. com.). Also, before cattaiJ invasion,

tires affected habitats with low turtle density, such as areas of P. aculeata or pastures.

Thus, the mortality of K scorpioides caused by fires at present time is an indirect resuft of
65

the invasion of T. domingensis throughout the wetland. According to Gibbons {2000), in

North America, the desert tortoise (Gopherus agassizii) and the gopher tortoise (Gopherus

polyphemus) are threatened by the introduction of non-native ptant spccies that alter among

other things fire frequency.

Before the invasion of cartaj), when the wetland did not offered any aestivating

refuge, the population of K scorpioides most likely concentrated in the dry forest area.c;

surrounding the marsh during the dry sca~n. The d.ry forest is not exempt from tire risk.

However, according to Whelan (1995) a grass fire may be more letha1 to many animals than

a forest fíre. This occurs becau.se dwing a fire event the continuity of the flame front,

frequency of occurrence of patches of unburned vegetation, and evenness in topography

vv'ill determine the abílíty of animals to dodge flames and find refuges. During fires in areas

of T. dominguensis in PVNP, many individuals of K. scorpioide.s· were sccn attempting to

escape the fire front TurtJes come out of their aestivating sites (under the veget-ation and

semi-buried underground) one or two meters ahead of the fire front and rapidly walk away

üom the Oames, but the high density of cattail stems would finally interrupt their flight On

the other hand, most of the areas of dry forest that surround the wetland in PVNP where K.

scorpioides aestivates during the dry season, have relatively opcn understory vegetation

( compared to the areas of T dominguensis in the wetland). Open undestory would probabty

allow a greater nwnber ofmud turtles to escape the fire front. Thus, it is possible that a tire

in the dry forest wou1d cause Jess turtle mortality than a fire in the wetland (in areas

dominated by T. domin~uensis) during the dry season. For instance, dead turtle density in

the secondary gallery forest was low compared to that of areas domidated by T.

dominguensis (Fig. 2).


66

PVNP fire history provided by the Organization for Tropical Studies (OTS) showed

that fTom 1987 to 2005 wild tires and controlled burns occurred in the park every year,

except in 1996 and 1997. From 1987 to 2005 the average area bum per year was 912 ha

mostly in the wetland. In May 2003, when a programmed bum got out of control, over 400

ha were burned in the wetland. Using the average dead turtles density (14.16 turtles/ha), we

estimate that 5,664 turtles were killed. Sorne fires were of great magnitude, in 1992 nearly

3,200 ha were bumed, this would account for an estimated 45,312 turtles killed. Certainly,

this 1s a speculated quantity, since the real mortality caused by a given fire event depends

on the real turtle density in that year and on the tire regimen. However, this gives us a good

idea of how fires can decimate turtle population. There is no data available about relative

abundance or population density of K scorpioides in PVNP for the previous years (no data

is available for neither dead turtle density nor live turtle density). Therefo.re, we are notable

to compare our results. This prevents us from answering the question: is the population of

K. scorpioides declining in PVNP? However, we postulate that the K. scorpioides

population in PVNP is declining for two reasons: l. The increased turtles mortality caused

by fires in the wetland is a recent event (starting in mid 80's) provoked by the invasion of

T dominguensis (which in turn resulted in the accumulation of fuel material and the arise

of a new habitat for turtles to aestivate in the wetland). Thus, high turtle mortality caused

by f'íres in the areas of T. dominguensis is a new factor with which the population did not

have to deal with in previous times. 2. Since we do not have data on turtle density for

PVNP in previous years, the best we can do is to compare our results ·with turtle densities

for other areas and species (Table 2). When doing so, K. scorpioides density in PVNP

seems to be low. In general, density for the species of Kinostemon throughout its

distributional range tend to be high, and density ofPseudemys scripta (1,475 twtles/ha) in
67

Caño Negro, Costa Rica (Mora & Ugalde 1991) is much higher than that of mud turtles in

PVNP (84 turtles/ha). Moreover, according to M.oll & Moll (2004) freshwater turtles are

known for their capacity to attain very high densities and biomass in both lentic and lotic

waters. In view of all of the above, it seems reasonable to speculate that K. scorpioides

density was higher in the past in PVNP.

Fires in PVNP are not on)y affecting the abundance of turtles but, also their

population structure. Fires are impacting mainly adults and individuals of greater size. This

trend needs to be considered. carefully. According to Gibbons (1990), a common life history

trait of turtles is to experience a high (> 90 % in sorne species) mortality rateas juveniles.

Turtles compensate this trend with a very Jow aduJt mortality rate ( 10-15 % in sorne

species) under natural conditions. Additionally, turtles are considered the epítome of long

lived animals and delayed sexual maturity. This means that a hatchling will take many

years to reach sexual maturity, during which time the organism is mainly investing in

growth. A greater size brings as a consequence greater safety from predators (as the shell

gets bigger and harder, therefore more effective); an increase in the reproductive output

(number and size of eggs that females can contain); and an increase in mate acqwsition

capacity formales (Gibbons 1990, Polisar 1995, Moll & Moll 2004). Once an individual

reaches sexual maturity, it wiJI compensate linger by living many reproductive years.

Extended iteroparous longevity compensates for both delayed maturation and hígh egg and

hatchling mortality. Because adult survivorship is high, sporadic years of high egg or

juvenile mortality may not affect population stability (Polisar 1995). Fires in the wetland,

especially in areas of T. domínguensis (a recent development in K. scorpioides history in

PVNP) is introducing high mortality rates in age and size classes accustomed to high

survival rates. This wilJ likely dimínish the population reproductive output and stability.
68

Literature cited

Acuña, R. 1990. El impacto del fuego y la sequía sobre la estructura de la población de

Kinosternon scorpioides (Testudines: Kinosternidae) en Palo verde, Guanacaste,

Costa Rica. Brenesia 33:85-97.

Acuña, R. 1998. Las tortugas continentales de Costa Rica. Universidad de Costa Rica San

José. 92 p.

Barboza, J. 2002. El pastoreo en la restauración del bosque tropical seco: Parque Nacional

Palo Verde, Guanacaste, Costa Rica. Seminario Nov 13-15 2002. Universidad

Nacional 1INISEFOR. Heredia.

Bumidge, W. 2000. Cattle and the managernent of freshwater neotropical wetlands in Palo

Verde National Park, Guanacaste, Costa Rica. Thesis. M.Sc. University of

Michigan. Ann. Arbor. 89 p.

Emst, C. & W. Barbour. 1989. Turtles ofthe World. Smithsonían. Washington. 313 p.

Emst, C., W. Barbour & J. Lovich. 1994. Turtles of the United States and Canada.

Smithsonian. Washington. 578 p.

Ewert, M. & C. Nelson. 1991. Sex determination in turtles: diverse patterns and sorne

possible adaptive values. Copeia. 91 :50-69.

Forero, G., O. Castaño & O. Montenegro. 2003. Abundance population structure and

conservation of Kinosternon scorpioides albogulare in San Amires isJand

(Colombia). Joint meeting of ichthyologists and herpetologists. Brazil.

Gibbons, W., D. Scott., T. Ryan., K. Buhlmann., T. Tuberville., B. Metts., J. Greene., T.

Mills., Y. Leiden., S. Poppy & C. Winne. 2000. The global decline of reptiles,

déjá vu amphibians. BioScience 50:653-666.


69

Hartshom, G. 1983. Introduction, p. 119-186. In Janzen, D. (ed.). Costa Rica natural

history. Chicago. Chicago.

Janzen, F. & G. Krenz. 2004. Phylogenetics: which was first, TSD or GSD? P. 121-l30. In

N. Valenzuela & A Lance (eds.). Temperature dependent sex detennination in

vertebrates. Smithsonian. Washington D.C.

Ministerio de Ambiente y Energía (MINAE). 2002. Resolución Legal R-009-ACA-D-2002.

Áiea de Conservación Arenal.

MoJI, D. & E. Moll. 2004. The ecology, explotation, and conservation of river turtles.

Oxford. New York. 393 p

Mora, J. & A Ugalde. 1991. A note on the population status and exploitation ofPseudemys

scripta emolli (Reptília: Emydidae) in Northern Costa Rica. BuU. Chicago Herp.

Soc. 26: 111.

Morales-Verdeja, S. & R. Vo!:,Tt. 1997. Terrestrial movements in relation to aestivation and

the annual reproductive cycle of Kinosternon Jeucostomum. Copeia 1997: 123-

130.

Polisar, J. 1995. River turtles reproductive demography and exploitation patterns in Belize:

implications fro management. Vida Silvestre Neotropical4: 10-19.

Savage, J. 2002. The amphibians and reptiles of Costa Rica. Chicago. Chicago. 934 p.

Trama, F. 2005. Manejo activo y restauración del humedal Palo Verde: cambios en las

coberturas de la vegetación y respuesta de Jas aves acuáticas. Ms.C. Tesis,

Universidad Nacional, Heredia. 154 p.

Whelan, R. 1995. The ecology offire. Cambridge. Cambridge. 346 p.


70

"...
'\
r:/-l,.,,.
..
· ~-- ,~~
'"'
-.·.
. .......<;\.·'
~
í'- .
"'
/ 1..

-., -- \
·--..
.-----------------------------------------------~ , -~ ~-~v
., ~
..,
'
•... ,J~

-,
~ ' .....'""
'"•

n N

Tcmp1Sque Rn·Ct'
-


Figure 1. Map of Palo Verde National Park. northwestem Costa Rica
71

40

35
'(;'
¿: 30

~
~

'-'
.o
.....
25

20
I I
~q)
15

~
"t:)
tl3
10

II I
-
q)
5
o
o
-5
A 8 e D E F
Habitat type
Figure 2. Dead mud turtle (Kinosternon scorpioides) density (mean density and standard

error) in all habitats affected by fires at Palo Verde NationaJ Park (A pastures, B. areas

dominated by Parkinsonia aculeate, C. secondary gallery forest, D. Large areas dominated

by Typha domingensis, E. areas of diverse and dry aquatic vegetation (DA V), and F. sma11

areas of Typha domingensis).

250

• Líve turtles
200
.!!4 b3 Dead turtles
§
:i 150
...
:;
.e
..
'O
.e
~
100

E
:S 50
z

0 -t--

l Adults Juveniles

Figure 3. Comparison of turtles (Kínosternon scorpioides) killed by fires and mud turtles

that survived fires in two groups: adults andjuveniles, in Palo Verde National Park.
72

----------- --·-----
A
- :-~l~~I~~===~==~~¡@~~~~~~~~~~===
----~------ ---------

-=~ ¡l~ ¡~ ¡ i~¡ ¡J~¡~ t~¡¡¡;¡~~=:-.-


~¡~~¡¡~~¡ ~¡~¡t~~-~~~~~~~~~~¡¡gj¡¡~¡¡m¡;;=
_-_-_
---- -_-_-_ ----
-
-:_ ----------
-----
------ ----------
---------
----------- ----------
----
--- ----- - - -------- - -~--
--------- ----
-_::::::~:::::::::::::::::~:--:::::::::::::::::::::-
-:=:=:=:• :=:=:=:=:=:=:=:=:=:=::
------- =
------------------~----------_-_-_-

--=-=-:-:-:-:-:-:-:-:-:-:-:--:-:-:-:-:-:-~:-:-:-:-:-:-:-:-:-:-:
• ==::=::=:__
----- --=____
:=:=·_
--
--- --- -------------- ------- -------
-_::.-__::-__:-_:-_:-_:-_:-_::-_:-_:-_:ji:::::::::::::::::::::::::::::::::::_:-_:-_::-_:-_:
D - Typha dominguensis ---=--=~=--==:::::::::::::::::::::::::::::::::::::•:=======---
-----------------------
------- ------------- ----------
------------------------ ----
----------
--- --- --
-_:-_:-_::-_:-_:-_::-_:-_:-_:-_::-_-..::::::::::::::::::::::::::::::::::::::::::::::::::
- Open water and diverse ----=~----::-:-:-::-:-:-:-:-:-:-:-•:-:-:-:-:-::-:-:-:-::-:..:-:-:
aquatic vegetation ----~-------------------------------------------------~-----
--------- ---------------- --- --- -
------ - ----~------------- ------ -

~1:1-
~...=

- - --.
.-
--~ -.

- _-:-:-:=~=:
...... ---
4---------
___
:..-...· ----- --
-~.:::::~:- --
·-----
Figure 4. Mud turtle (Kinosternon scorpioides) distribution in the wetland in Palo

Verde National Park during the rainy season (A), and during the dry season (B).
73

Table l. Comparison of size between turtles (Kinosternon scorpioides) that were killed by fires and turtles that survived fires.

All individuals Only males Only females

Curve carapace length (CCL) dead survivors dead survivors dead survivors

n 247 120 75 30 134 70

x± SE (cm) 17.7± 0.19 16.7 ± 0.28 19.76 ± 0.1 18.8 ± 0.26 17.25 ± 0.22 17.62 ± 0.29

t test statistic t == 2.79, df= 365, p = 0.005 t = 2.96, df= 103, p:;:: 0.004 t:::: -1.02, df:;:: 202, p . 0.3

Curve carapace width (CCW) dead survivors dead survivors dead survivors

n 250 120 75 30 137 70

x±SE(cm) 15.14 ± 0.26 15.75±0. 18 16.74±0.14 16.07 ± 0.23 15.94 ± 0.21 16.38 ± 0.3

t test statistic t = -1.9, df= 368, p == 0.05 t = 2.4, df= 103, p = 0.01 t :;:: -1.17, df= 205, p =0.24
74

1'ablc 2. Density in several species ofturtles (nwnber ofindividua1s / hectare)

Spe(ies/family Density Place Source

Kinosternon scorpioides 84.61 dry forest, PVNP-Costa Rica Present study

K.inoslemidae

Kinoslemon scorpioides 14.16 * wetland, PVNP-Costa Rica Present study

Kínosternidae

Kinostemon scorpioides 165.5 San Andrés lslan~ Colombia Forrero et al. 2003

Kinostemidae

Kinostemon leucostomun 449 V eracruz, México Morales-Verdeja &

Kinosternidae Vogt. 1997

Kinostemon subrubmm 208 Oklahoma, US Emst et al. 1994

Kinostemidae

Kínosternon submbrum 39 South Carolina, US Emst et al. 1994

Kinosternidae

Kinosternon jlavescens 27.9 Oklahoma. US Emst el al. 1994

Kino~1Cmidae

Kinostemon sonoriense 787 Arizona, US Emst el al. 1994

Kinosternidae

Kinosternon hirtipes 189.5 Texas, US Emst el al. 1994

Kffiostemidae

Pseudemys scripta 1475 Caño Negro, Costa Rica Mora & Ugalde, 1991

Emydidae

• Density of dead turtJes killed by fires in tbe wetland of PVNP.

74