Reproductive Ecology of Sceloporus utiformis (Sauria: … · 2008-06-09 · Reproductive Ecology of Sceloporus utiformis (Sauria: Phrynosomatidae) from a Tropical Dry Forest of Me´xico

Journal of Herpetology, Vol. 37, No. 1, pp. 1–10, 2003Copyright 2003 Society for the Study of Amphibians and Reptiles

Reproductive Ecology of Sceloporus utiformis(Sauria: Phrynosomatidae) from a Tropical Dry Forest of Mexico

AURELIO RAMIREZ-BAUTISTA1,2,3 AND GUADALUPE GUTIERREZ-MAYEN4

1Laboratorio de Ecologıa, Unidad de Biologıa, Tecnologıa y Prototipos (UBIPRO), FES-Iztacala, Universidad NacionalAutonoma de Mexico. Av. de los Barrios s/n, Los Reyes Iztacala, Tlalnepantla, Edo. de Mex. C.P. 54090, A.P. 314.

Mexico; E-mail: [email protected] de Herpetologıa. Escuela de Biologıa. Benemerita Universidad Autonoma de Puebla. Edificio 76.

Ciudad Universitaria. Av. San Claudio y Blvd. Valsequillo s/n. Col. San Manuel. C.P. 72570. Puebla, Pue. Mexico;E-mail: [email protected]

ABSTRACT.—The reproductive cycle and cycles in fat body and liver mass are described for male andfemale Sceloporus utiformis from Chamela, Jalisco, Mexico, during 1988 and 1990. These cycles varied amongmonths and between sexes. Males reached sexual maturity at 45 mm snout–vent length (SVL) at an age ofseven months. Females reached sexual maturity at 56 mm at an age of nine months. Testes of males beganto increase in volume in early May, and maximal testicular activity occurred from June to October. Testissize began to decrease in November and was small in December. Reproductive activity of females occurredbetween July and March. Gonads of females began to increase in volume in mid-June, and maximal eggproduction was from August to November. Maximal testicular growth was positively correlated with in-creasing temperature and precipitation, but not with photoperiod, whereas in females follicular growth waspositively correlated with photoperiod but not with temperature and precipitation. Egg production occurredprimarily during the wet season and early dry season. Females started to lay their eggs in late July. Mean(6 SE) clutch size of oviductal eggs (6.94 6 0.28) and vitellogenic follicles (7.3 6 0.3) were not significantlydifferent. Clutch size was significantly correlated with female SVL (r 5 0.26, P , 0.05) and varied betweenyears and months, as in other lizard species, presumably reflecting the effects of annual variation in resourceavailability on females.

RESUMEN.—El ciclo reproductivo y ciclos del cuerpo graso y masa del hıgado son descritos para las hem-bras y los machos de Sceloporus utiformis durante 1988 y 1990 en Chamela, Jalisco, Mexico. Estos ciclosvariaron entre meses y entre sexos. Los machos alcanzaron la madurez sexual a una longitud hocico-cloaca(LHC) de 45 mm a una edad de 7 meses. Las hembras alcanzaron la madurez sexual a los 56 mm a unaedad de 9 meses. Los testıculos de los machos comenzaron a incrementar en volumen a principios de mayo,con la maxima actividad testicular de junio a octubre. El tamano de los testıculos empezo a decrecer ennoviembre, y en diciembre fue pequeno. La actividad reproductiva de las hembras ocurrio entre julio ymarzo. Las gonadas de las hembras comenzo a incrementar en volumen a mediados de junio, la maximaproduccion de huevos fue de agosto a noviembre. El maximo crecimiento testicular estuvo correlacionadopositivamente con el incremento de la temperatura y precipitacion, pero no con el fotoperiodo, mientras queen las hembras, el crecimiento folicular estuvo positivamente correlacionado con el fotoperiodo pero no conla temperatura y precipitacion. La produccion de huevos ocurrio durante la estacion de lluvias y a principiosde la estacion de secas. Las hembras comenzaron a poner sus huevos a fines de julio. La media (6 SE) deltamano de la puesta de los huevos en el oviducto (6.94 6 0.28) y de folıculos vitelogenicos (7.3 6 0.3) nofueron estadısticamente diferentes. El tamano de la puesta estuvo significativamente correlacionado con laLHC de la hembra (r 5 0.26, P , 0.05) y vario entre anos y meses como en otras especies que habitan enla region.

Since the pioneering working of Tinkle, agreat diversity of reproductive patterns in liz-ards has been uncovered (Tinkle, 1969; Tinkle etal., 1970). Tinkle’s works stimulated numerousstudies on lizard reproduction and life-history

2 Corresponding Author.3 Present address: Centro de Investigaciones Bıolo-

gicas Universidad Autonoma del Estado de HidalgoA. P. 1–69 Plaza Juarez, C. P. 42001 Pachuca, Hidalgo,Mexico

variation such as reproductive period, clutchsize, egg size, clutch frequency, fecundity, age atmaturity, and survivorship (Ballinger, 1979;Dunham, 1981; Vitt, 1986; Lemos-Espinal andBallinger, 1995; Ramırez-Bautista et al., 1995,1996). These studies have shown that variationin life history is explained by environment(proximate environmental effects), ecology (for-aging mode and habitat use), and phylogeny(Vitt and Congdon, 1978; Dunham and Miles,1985; Vitt, 1986, 1990, 1992). Length of the re-

2 A. RAMIREZ-BAUTISTA AND G. GUTIERREZ-MAYEN

productive period varies among species: for ex-ample, although some tropical lizards reproducecontinuously throughout the year (Vitt, 1986;Benabib, 1994), others have seasonal reproduc-tive schedules (Ramırez-Bautista, 1995; Ramı-rez-Bautista et al., 1995, 1996). In those tropicalspecies with continuous reproduction, there cannevertheless be annual variation in egg produc-tion associated with the rainy season and highfood availability (Guyer, 1988 a,b; Rose, 1982)

Reproductive activity in tropical lizards liv-ing in seasonal (wet-dry) environments is cycli-cal (Ramırez-Bautista, 1995; Ramırez-Bautista etal., 1995, 1996; Ramırez-Bautista and Vitt, 1997;Lemos-Espinal et al., 1999). In most lizard spe-cies from the tropical dry forests of westernMexico, males begin to establish territories forcourtship, mating, and copulation at the begin-ning of the rainy season (Ramırez-Bautista,1995; Ramırez-Bautista and Vitt, 1997, 1998).Egg development starts during the rainy season,and hatching occurs at the end of the rainy sea-son (Ramırez-Bautista, 1995; Ramırez-Bautistaand Vitt, 1997). Ramırez-Bautista and Vitt (1997)and Ramırez-Bautista et al. (2000) showed thatreproductive patterns of lizards are affected byrainfall. However, it also has been found thatreproduction is influenced by temperature(Marion, 1982), food availability (Ballinger,1977), photoperiod (Marion, 1982), and phylog-eny (Miles and Dunham, 1992). Several studiesindicate that life-history variation among highertaxa (Dunham and Miles, 1985; Miles and Dun-ham, 1992) is often greater than variationamong populations of the same species (Grantand Dunham, 1990; Smith and Ballinger, 1994;Ramırez-Bautista et al., 1995) or between yearswithin populations (Ballinger, 1977; Dunham,1978; Smith and Ballinger, 1994; Smith et al.,1995).

Similar to other species from other latitudes,variability in reproductive characteristics of liz-ards from the tropical dry forest of westernMexico is associated with variation in resourceavailability, temperature, and precipitation (Ra-mırez-Bautista, 1994; Ramırez-Bautista, 1995;Ramırez-Bautista and Vitt, 1997, 1998; Lemos-Espinal et al., 1999). To date, there are only an-ecdotal notes concerning the general biology ofthe lizard, Sceloporus utiformis (Ramırez-Bautis-ta, 1994; A. Ramırez-Bautista and Z. Uribe, un-publ. data). Here we present data on a popula-tion of S. utiformis to address the following gen-eral questions: (1) Are sexually mature femalesand males the same size? (2) What is the repro-ductive cycle of females and males? (3) Doesclutch size vary with female size? (4) Is there acorrelation between peak reproductive activityand environmental factors (temperature, precip-itation, and photoperiod)?

MATERIALS AND METHODS

Field Site.—Reproductive data were obtainedon female and male S. utiformis from Chamela,Jalisco, near the Estacion de Biologıa ‘‘Chamela’’(EBCH, 5 km from the Pacıfic coast at approxi-mately 198309N, 1058039W, at elevations varyingfrom 10 to 584 m) in Jalisco, Mexico. Mean an-nual temperature is 24.98C, and vegetation istropical dry forest. Rain falls in late spring andsummer and annual average rainfall is 748 6119 mm (SE; Bullock, 1986). Data on photope-riod were taken from the Astronomical Alma-nac of the World (1984). Mean annual temper-ature and precipitation over a 10-yr period andduring the study were recorded at the EstacionMeteoreologica of the region and have been re-ported elsewhere (Ramırez-Bautista, 1995; Ra-mırez-Bautista and Vitt, 1997).

Reproductive Analysis.—A total of 268 (94 adultfemales and 113 males, and 61 young) individ-uals were sampled from June 1988 to September1990 (28 months). Because samples were oftensmall for individual months and varied consid-erably among years, data were pooled to de-scribe the annual reproductive cycle of femalesand males. All specimens were humanely killedand fixed (10% formalin) in the laboratorywhere gonadal analysis was performed. The fol-lowing linear measurements were taken on nec-ropsied lizards: snout–vent length (SVL) to thenearest 1.0 mm, length and width of left testis,length and width of oviductal eggs, and of vi-tellogenic and nonvitellogenic follicles (all to thenearest 0.1 mm). Volume of the gonads of fe-male (oviductal eggs, nonvitellogenic, and vitel-logenic follicles) and male (testes) was estimatedwith the formula for a prolate spheroid (Selby,1965):

V 5 4/3 p (1/2L) (1/2W)2

where W is width and L is length. Testes ofmales, and livers and fat bodies of both sexeswere weighed (to the nearest 0.0001 g). Thesmallest females (56 and 57 mm) that containedenlarged vitellogenic follicles ($ 5.6 mm3) oroviductal eggs (165.8 mm3) were used as an es-timate of minimum SVL at maturity. All femaleswith a SVL , 56 mm were considered nonre-productive with nonvitellogenic follicles (4.8mm3). Males were considered sexually mature(SVL 5 45 mm) if they had enlarged testes ($15.6 mm3) and epididymides typically associat-ed with sperm production (Goldberg and Lowe,1966). For all seasonal analyses, data were re-stricted to sexually mature lizards. Because or-gan volume (gonads) and mass (liver and fatbodies) may vary with SVL, we first calculatedthe regression of log10-transformed organ vol-ume and/or mass data with log10 of SVL. For

3REPRODUCTION OF SCELOPORUS UTIFORMIS

FIG. 1. Size distribution of sexually mature malesand females of Sceloporus utiformis from Chamela(June 1988 to September 1990).

those regressions that were significant (indicat-ing a body size effect), we calculated residualsfrom the relationship of organ mass and volumeto SVL (all variables log10-transformed) to pro-duce SVL adjusted variables. We used these re-siduals to describe the organ and/or reproduc-tive cycles. This technique maintains variationbased on extrinsic factors (i.e., season) whileminimizing the confounding effect of individualvariation in SVL. For regressions that were notsignificant (i.e., no body size effect), we used theactual volume (gonads) and organ mass (liverand fat bodies) to describe the reproductiveand/or organ mass cycles. We performed AN-OVAs on organ mass and volume with monthas the factor to determine whether significantamong-month variation existed, including onlythose months for which N $ 3.

The number of nonvitellogenic and vitello-genic follicles and/or oviductal eggs was re-corded for females. Clutch size was determinedby counting eggs in the oviduct or vitellogenicfollicles of adult females during the reproduc-tive season. Reproductive potential, as usedhere, is the average number of eggs producedin one year (reproductive season) by one female.Incubation period was estimated as the intervalbetween the date on which individual femaleshad their first oviductal eggs of the season (July25) and the date on which hatchlings first ap-peared on the study site (October 7). Snout–ventlength of neonates was taken from hatchlings (N5 85) marked for demographic study of thisspecies. Seasonal distribution of body sizes(adults, juveniles, and hatchlings) was estab-lished from marked specimens.

Morphological comparisons (e.g., sexual di-morphism) were restricted to the upper 50% ofthe sample of sexually mature females (64–73mm SVL) and males (64–84 mm SVL) to reducebias resulting from sampling error. Means arepresented 6 SE unless otherwise indicated.Standard parametric statistical tests were usedwhen possible; otherwise, appropriate nonpara-metric tests were substituted. Statistical analy-ses were performed with the Macintosh versionof Statview 4.01 (Abacus Concepts, Inc., Berke-ley, CA, 1992). Specimens are deposited at theColeccion Nacional de Anfibios y Reptiles delInstituto de Biologıa, Universidad Nacional Au-tonoma de Mexico, Mexico City.

RESULTS

Body Size and Sexual Maturity.—Reproductive(N 5 207) lizards varied in size from 45–84 mmSVL. Sexually mature males ranged in size from45–84 mm SVL (mean 6 SE; 61 6 0.9, N 5 113;Fig. 1) and reached maturity in their first repro-ductive season after hatching at an age of sevenmonths (based on unpublished capture-recap-

ture data and growth rate). Sexually mature fe-males ranged in size from 56–73 mm SVL (63.86 0.5, N 5 94; Fig. 1) and reached sexual ma-turity at an age of nine months. Based on com-parison of the largest 50% of sexually maturemales and females, males were slightly larger(69 6 0.6 mm, 64–84) than females (67 6 0.4mm, 64–73; t 5 22.28, df 5 99, P , 0.05).

Male Reproductive Cycle.—One hundred thir-teen males were sexually mature based on theirSVL. There were significant linear relationshipsbetween SVL of sexually mature males and tes-tes volume (r2 5 0.62, F1,111 5 182.1, P , 0.001),fat body mass (r2 50.69, F1,111 5 11.84, P , 0.001)and liver mass (r2 5 0.63, F1,111 5 185.7, P ,0.001). Consequently, the relationships betweenseason and males testes volume, fat body massand liver mass are best represented by plots ofregression residuals (Fig. 2). An ANOVA on re-siduals of the regression revealed significantamong-month variation in testes volume (F11,101

5 11.06, P , 0.001), fat body (F11,101 5 6.4, P ,0.001), and liver mass (F11,101 5 3.9, P , 0.001;Fig. 2). There were significant linear relation-ships between testes volume and fat body mass(r2 5 0.29, F1,111 5 10.44, P , 0.005) and betweentestes volume and liver mass (r2 5 0.62, F1,111 5183.7, P , 0.005). Peak testicular volume oc-curred from May through June and November.However, one male from March and another onefrom April 1990 were in reproductive condition.Testes reached their maximum size during Junethrough October. A significant decline in gonad-al size occurred in November, and testes weresmall in December (Fig. 2). The period of max-imal testicular growth of S. utiformis was posi-tively correlated with increasing temperature (r5 0.90, P , 0.0001) and precipitation (r 5 0.84,P , 0.001) but not with photoperiod (r 5 0.29,P . 0.05).

Female Reproductive Cycle.—The annual repro-ductive cycle of females was based on 94 indi-


FIG. 2. Male testis, fat body, and liver cycles of Sce-loporus utiformis. Data are mean (6 1 SE) residualsfrom a regression of log10-testis volume (mm3), fatbody mass (g), and liver mass (g) against log10 SVL.

FIG. 3. Female gonad, liver mass, and fat body cy-cles of Sceloporus utiformis. Data are mean (6 1 SE)residuals from a regression of log10-gonad volume(mm3), liver mass (g) against log10 SVL, and fat bodymass (g).

viduals from 1988, 1989, and 1990. There weresignificant linear relationships with female SVLfor gonadal volume (r2 5 0.31, F1,92 5 9.86, P ,0.05) and liver mass (r2 5 0.55, F1,92 5 40.5, P ,0.001) but not for fat body mass (r2 5 0.03, F1,92

5 0.082, P . 0.05). Consequently, gonadal vol-ume and liver mass cycles of females are bestrepresented by plots of regression residuals,

whereas the fat body cycle is best representedby log-transformed fat body mass (Fig. 3). AnANOVA on residuals of the regression revealedsignificant among-month variation in gonadvolume (F11,82 5 3.18, P , 0.05) and liver mass(F11,82 5 2.5, P , 0.05: Fig. 3). There was a pos-itive significant linear relationship between go-nadal volume and liver mass (r2 5 0.197, F1,.92 522.50, P , 0.001 but not between gonadal vol-ume and fat body mass (r2 5 0.001, F1,92 5 0.097,P . 0.05). Average female gonadal volume in-creased from mid-June to late-July when femalesbegan to ovulate and decreased from January to


FIG. 4. Relationship between female log10 SVL andclutch size for Sceloporus utiformis.

FIG. 5. Seasonal distribution of body sizes for Sce-loporus utiformis.

March of the following year. Reproductive activ-ity in 1988 continued until January and March1989 because three (75%) and two (66.3%) fe-males were found with oviductal eggs, respec-tively. A similar pattern occurred in 1989 whenthree females (100%) were found with vitello-genic follicles in February 1990.

Females containing vitellogenic ovarian folli-cles were observed between early summer (July)and late winter (March) during 1988–1989 and1989–1990 with peak reproductive activity oc-curring from August to November. All classesof eggs (oviductal eggs or vitellogenic follicles)were found in 25 females (83.3%) in 1988, 33females (80.5%) in 1989, and 14 females (61%)in 1990.

Vitellogenesis and follicular growth of femaleS. utiformis were related to photoperiod (r 50.56, P , 0.05) but not to temperature (r 5 0.15,P . 0.05), or precipitation (r 5 0.29, P . 0.05).

Clutch Size.—Mean clutch size of oviductaleggs (6.94 6 0.28, range 3–10, N 5 31) and vi-tellogenic follicles (7.3 6 0.3, range 2–11, N 541) were not significantly different (Mann-Whit-ney U-test, z 52 1.12, P 5 0.2651). Mean clutchsize of oviductal eggs and vitellogenic follicles(combined) was 7.1 6 0.21 (2–11, N 5 72).Clutch size was significantly correlated with fe-male SVL (r 5 0.26, P , 0.05, N 5 72; Fig. 4).Two-factor ANOVAs on the residuals from theclutch size versus SVL regression revealed sig-nificant effects of years (F2,69 5 2.39, P , 0.05)and months (F8,63 5 9.72, P , 0.005). Meanclutch size was similar in 1988 and 1990, but itwas higher (7.6 6 0.27, N 5 39) than in 1989(6.6 6 0.36, N 5 33, F2,69 5 3.2, P , 0.05). Clutchsize ranges varied among years: 3–10 eggs in1988, 2–11 eggs in 1989, and 4–9 eggs in 1990.Clutch size of females from January (5.33, N 53) and March (2.0, N 5 2) from 1989 was lowerthan for females that reproduced during the wet

season. However, three females from February1990 had a mean clutch size of 8.0 vitellogenicfollicles. Using egg volume data and restrictingthe dataset to those months (January and Julyto December) for which data were sufficient,two-factor ANOVAs showed no effect of years(F2,28 5 0.08, P . 0.05) or months (F6,24 5 1.21,P . 0.05) on egg volume.

Females started to lay their eggs on 25 July in1988 and 1990. The first hatchlings were ob-served on 7 October (Fig. 5). These data suggestthat incubation period of eggs is approximately75 days. Most hatchlings appeared in the fieldbetween October 1989 and January 1990. MeanSVL of neonates from October was 27.8 6 1.07(26–32, N 5 5), and average of the offspringduring the hatching period (October to January)was 29.1 6 0.21 mm (range 5 25–32 mm, N 585).

DISCUSSION

Sexual Dimorphism.—Males of S. utiformisfrom Chamela reached sexual maturity at asmaller size yet attained a larger maximumbody size than did females. Many species of liz-ards are sexually dimorphic in body size (Smithet al., 1997; Stamps, 1983; Ramırez-Bautista etal., 1996; Ramırez-Bautista and Vitt, 1997). Inlizard species there are a variety of possible ex-planations for sexual size dimorphism (Stamps,1983). Male body size could be related to sexualselection and defense of territory. Reproductiveactivity of both sexes began at the same time,and sexual dimorphism in body size of malescould be a result of sexual selection by females.Larger males can have an advantage over small-er males in acquiring mates (Ruby, 1981; Ra-mırez-Bautista and Vitt, 1997). Sexual dimor-phism may represent an evolutionarily conser-vative trait maintained in species with a com-mon ancestor (Stamps, 1983; Anderson and Vitt,


FIG. 6. Seasonal changes in the frequencies of var-ious reproductive states for female Sceloporus utiformis.Sample sizes appear above bars.

1990). In male S. utiformis, a combination ofthese factors could be involved.

Male Reproductive Cycle.—Reproductive activ-ity in male S. utiformis was cyclical with an ac-tivity peak during June through October. Testessize increased as soon as temperature and pre-cipitation increased. Temperature, precipitation,or a combination of these factors appears to in-fluence reproduction in male lizards (Licht andGorman, 1970; Gorman and Licht, 1974; Marion,1982; Ramırez-Bautista and Vitt, 1997, 1998).Male reproductive activity in 1989 began inmid-May and continued through March 1990.On 24 February 1990, a marked male was ob-served in copulation in the field. These resultsshow that the reproductive period of males in1989 extended into the dry season of 1990. Thisreproductive pattern is different from that ofseveral sympatric species at this site (Ramırez-Bautista and Vitt, 1997, 1998). For example, inAnolis nebulosus (Ramırez-Bautista and Vitt,1997), Urosaurus bicarinatus (Ramırez-Bautistaand Vitt, 1998), and Cnemidophorus communis(Ramırez-Bautista and Pardo-De la Rosa, 2002)reproduction occurs during the wet season(June to October). Reproductive activity of S.utiformis begins at the end of the dry season(May) and ends early in the next dry seasonwith maximum reproduction occurring duringthe wet season, similar to another sympatricspecies, Cnemidophorus lineatissimus (Ramırez-Bautista et al., 2000). In this respect, reproduc-tive activity of S. utiformis resembles that of oth-er species inhabiting seasonal environments inwhich male reproduction might extend beyondthe wet season, such as Urosaurus ornatus and U.graciosus (Parker, 1973; Vitt and Ohmart, 1975;Van Loben Sels and Vitt, 1984) and Ameiva amei-va (Vitt, 1982). The positive relationship betweentestes development and fat body and betweentestes volume and liver mass suggests a low en-ergetic cost to male reproductive activities. Thispattern differs from other species with cyclicalreproduction occurring in the region, such as A.nebulosus (Ramırez-Bautista, 1995; Ramırez-Bau-tista and Vitt, 1997) and U. bicarinatus (Ramırez-Bautista and Vitt, 1998) in which testes devel-opment is not related to fat body mass, sug-gesting a high energetic cost. Sceloporus utiformismales have a long period of reproductive activ-ity similar to the tropical dry forest lizard, C.lineatissimus (Ramırez-Bautista et al., 2000) anddifferent from other species of the same regionsuch as A. nebulosus and U. bicarinatus (Ramırez-Bautista and Vitt, 1997, 1998).

Female Reproductive Cycle.—The duration of thefemale reproductive period was similar to thatof males. Follicles began to increase in size dur-ing May. Most egg production occurred fromJuly to October (coinciding with the wet season).

However, oviductal eggs were found as late asMarch of the next year, indicating that femalescan produce eggs over a period of nine months(Fig. 6). The timing of reproductive activity isdifferent from other species of the region suchas A. nebulosus and U. bicarinatus that have ashort (July to October) period of egg productionwith the peak occurring during the rainy season(Ramırez-Bautista, 1995; Ramırez-Bautista andVitt, 1997, 1998). In S. utiformis, fat body masswas low during vitellogenic follicle developmentand egg production (primarily August to De-cember) suggesting a high cost of egg produc-tion. However, the relationship between follicu-lar volume and liver mass suggests that livermass is mobilized as energy to support vitello-genesis and egg production, as has been ob-served in other lizard species (Ramırez-Bautista,1995; Ramırez-Bautista and Vitt, 1997, 1998).

The onset of vitellogenesis in females beganas photoperiod increased and, although the cor-relations with rainfall and temperature were notsignificant, a combination of the three factors islikely to play an important role in initiating re-production (Marion, 1982; Licht, 1984; Ramırez-Bautista and Vitt, 1997, 1998). Although photo-period may initiate reproduction, rainfall, andtemperature and, more specifically, the timingof rainfall, may be the ultimate cues for repro-duction through its effect on egg and offspringsurvival. An ultimate effect of rainfall on repro-duction is suggested because vitellogenesis andmost egg production occurred during the peakof the rainy season. Females had vitellogenic fol-licles and eggs in March 1989 and 1990, whichsuggests some plasticity of females to respondto the environmental conditions of the region.

The reproductive pattern of S. utiformis fe-males is different from that of other syntopicspecies (Ramırez-Bautista and Vitt, 1997, 1998)and from some other tropical dry forest speciesof the genus Sceloporus (Table 1). For example, S.horridus (Fitch, 1970), S. siniferus (Davis and Dix-on, 1961) and S. spinosus (Fitch, 1978) have shortreproductive periods (spring to summer). Incontrast, S. gadovae (Lemos-Espinal et al., 1999)


TA

BL

E1.

Rep

rodu

ctiv

ech

arac

teri

stic

s(m

ean

6SE

)of

fem

ales

ofva

riou

sov

ipar

ous

spec

ies

ofth

ege

nus

Scel

opor

usfr

omtr

opic

ald

ryfo

rest

sof

Mex

ico.

M5

mu

ltip

lecl

utch

esp

erye

ar;S

5si

ngle

clut

chp

erye

ar.

Spec

ies

Bod

ysi

ze(S

VL

,mm

)R

ange

Clu

tch

size

Ran

geC

lutc

hfr

eque

ncy

Rep

rod

ucti

vese

ason

Sou

rce

S.ut

iform

isS.

gado

vae

S.ho

rrid

usS.

mel

anor

hinu

sS.

pyro

ceph

alus

S.si

nife

rus

S.sp

inos

us

63.8

60.

555

.06

0.2

—87

.26

1.6

53.7

60.

4352

.36

0.52

87.2

61.

6

56–7

347

–76

—80

–98

47–6

248

–61

77–9

6

7.1

60.

213.

66

0.2

12.0 7.7

60.

335.

86

0.31

5.0

12.7

2–11

1–5

8–15

5–9

4–9

4–6

8–16

S M M S M M M

sum

mer

–win

ter

spri

ng–w

inte

rsu

mm

ersp

ring

–su

mm

ersp

ring

–su

mm

ersp

ring

–su

mm

ersp

ring

–su

mm

er

Thi

sst

udy

Lem

os-E

spin

alet

al.,

1999

Fitc

h,19

70R

amır

ez-B

auti

sta,

unp

ubl

.dat

aR

amır

ez-B

auti

sta,

unp

ubl

.dat

aD

avis

and

Dix

on,1

961

Fitc

h,19

78

and C. lineatissimus (Ramırez-Bautista et al.,2000) have a longer period of reproductive ac-tivity similar to that of S. utiformis.

Gravid females collected in January, February,and March 1989 and 1990 had greater fat bodyand liver mass. These data suggest that foodwas abundant and that females fed well enoughto produce eggs early in the dry season. DuringJuly 1988, 1989, and 1990, most females had vi-tellogenic follicles and oviductal eggs. This in-dicates the importance of rainfall in the repro-ductive activity of females. A climatic and re-productive activity comparison suggests thatthe interaction might be very complex as isshown in other species of the same region (Ra-mırez-Bautista and Vitt, 1997, 1998). If testicularvolume and follicular development increased(May to June) when photoperiod and tempera-ture increased, then a combination of thesecould be the proximate factors for the onset ofreproduction. An increase in temperature andprecipitation early in the reproductive seasoncould facilitate follicular development throughtheir effect on egg production and offspring sur-vival (Andrews and Sexton, 1981; Ramırez-Bau-tista and Vitt, 1997).

Clutch Size.—Mean clutch size of S. utiformiswas 7.1 eggs. Clutch size was correlated withfemale SVL and is among the largest reportedfor the genus Sceloporus from tropical arid zones(Table 1). Among other Sceloporus species withlarge clutch sizes are S. spinosus (12.7; Fitch,1978), S. melanorhinus (7.7; Ramırez-Bautista, un-publ. data), and S. pyrocephalus (5.8; Ramırez-Bautista, unpubl. data). Body size may explainin part the large clutch size in S. utiformis. Largefemales produced more eggs than small fe-males. Females reach maximum SVL at the be-ginning of reproduction and most energy isthen allocated to egg production (Tinkle andBallinger, 1972; Tinkle and Dunham, 1986).Large females could have allocated more energyto clutch size than to growth. A high energeticcost of egg production is suggested by the neg-ative relationship between clutch size and fatbody and liver mass. This pattern is similar tothat in other species (Smith et al., 1995; Ramı-rez-Bautista and Vitt, 1997). Fat body and livermass increased after the completion of the re-productive period. However, some females werefound with oviductal eggs and vitellogenic fol-licles early in the dry season, together with highlevels of fat and liver mass. These data suggestthat food availability for this species is sufficientfor some females to reproduce in the dry season(Ramırez-Bautista, 1995). In addition, success inaccumulating fat reserves during the dry seasonallows females to produce eggs late in the re-productive season. The snout–vent length at sex-ual maturity of females that reproduced later


during the reproductive season was 63.8 mm.These data suggest that these females repro-duced later because sexual maturity wasreached at the end of reproductive period.

Year-to-year variation in reproductive char-acteristics of S. utiformis could be the effect ofyear-to-year variation in precipitation. Precipi-tation was above average in both 1988 (884.3mm) and 1989 (937.1 mm), whereas it was be-low average in 1990 (583.5 mm; Ramırez-Bautis-ta and Vitt, 1997, 1998). Perhaps in response tothe above-average precipitation in the precedingyears, egg production was prolonged towardMarch in 1989 and 1990. In addition, both themean and the range of clutch size varied signif-icantly among years, which could be caused bydifferences in food abundance among years. Thegreatest range in clutch size occurred in yearswith the higher precipitation. Although there isalways food available in the environment (Ra-mırez-Bautista, 1995), higher levels of precipi-tation presumably increase the quantity andquality of food resources. Precipitation increasesarthropod abundance and consequently favorsfemale reproduction (Ramırez-Bautista, unpubl.data). Females that are well fed gather high lev-els of fat reserves. Several authors have reporteda positive correlation between precipitationand/or insect abundance and clutch size andfrequency in lizards (Parker and Pianka, 1975;Vinegar, 1975; Ballinger, 1977, 1979; Dunham,1982; Ramırez-Bautista et al., 1995; Ramırez-Bautista and Vitt, 1997). In addition, the femalereproductive season is typically extended inother lizard species of the same region, such asC. lineatissimus (Ramırez-Bautista et al., 2000), C.communis (Ramırez-Bautista and Pardo-De laRosa, 2002) and other tropical or subtropical te-iids, including Ameiva ameiva (Vitt, 1982) and C.ocellifer (Vitt, 1983).

This study, and others in the same region,suggest that variation in amount and durationof precipitation among years are factors thatpartially explain variation in life history. Severalstudies have pointed out that droughts can de-crease the duration of the reproductive period,clutch size, and egg size (Ballinger, 1977; Ra-mırez-Bautista and Vitt, 1997). Although precip-itation is an important factor in lizard repro-duction, temperature, and photoperiod play im-portant roles in determining the onset of repro-ductive activity of several lizards species fromChamela (A. Ramırez-Bautista and Z. Uribe, un-publ. data; Ramırez-Bautista, 1994, 1995; Ramı-rez-Bautista and Vitt, 1997). A combination ofbehavioral, ecological, and evolutionary factorscould be influencing life-history characteristicsof S. utiformis, as occurs in other species (Ballin-ger, 1977; Dunham, 1981, 1982; Vitt, 1986; Ra-

mırez-Bautista et al. 1995; Ramırez-Bautista andVitt, 1997).

Life-history characteristics of S. utiformis fitone of the patterns identified by Tinkle et al.(1970), including (1) small SVL at sexual matu-rity, (2) SVL at sexual maturity of males andfemales is reached during the first year of life,(3) fast growth after birth (unpubl. data), and(4) one clutch of several eggs (compared withsmall syntopic species). However, more infor-mation is needed on recruitment and the dura-tion of reproductive activity. Although there issome information on the demography of thisspecies (unpubl. data), more data are needed todetermine that S. utiformis is in fact short-lived.A long-term demographic study would greatlyfurther an understanding of the relative roles ofproximate and ultimate variation in life historytraits in S. utiformis.

Acknowledgments.—We wish to thank the Es-tacion de Investigacion, Experimentacion y Di-fusion Biologica de Chamela, UNAM, and thechief, Felipe Noguera, for his support of thestudy, L. J. Vitt for his help with statistical anal-ysis and his teaching, G. Herrera, X. Hernandez-Ibarra and R. Cervantes-Torres for reading themanuscript, Z. Uribe, R. Ayala, D. Pardo, C.Balderas, A. Borgonio, and J. Barba for theirsupport during the field work. We also thanktwo anonymous reviewers for critically readingthe manuscript. Financial support was from agrant from Facultad de Ciencias UNAM as apart of the program for graduate students (PA-DEP) during 1989–1990, and Direccion Generalde Asuntos del Personal Academico (DGAPA)which provided (ARB) a postdoctoral scholar-ship (University of Oklahoma) during 1996–1997.

LITERATURE CITED

ANDERSON, R. A., AND L. J. VITT. 1990. Sexual selec-tion versus alternative causes of sexual dimor-phism in teiid lizards. Oecologia 84:145–157.

ANDREWS, R. M., AND O. J. SEXTON. 1981. Water re-lations of the eggs of Anolis auratus and Anolis lim-ifrons. Ecology 62:556–562.

ASTRONOMICAL ALMANAC OF THE WORLD. 1984. U.S.Government Printing Office, Washington, DC.

BALLINGER, R. E. 1977. Reproductive strategies: foodavailability as a source of proximal variation in alizard. Ecology 58:628–635.

. 1979. Intraspecific variation in demographyand life history of the lizard, Sceloporus jarrovialong an altitudinal gradient in southeastern Ari-zona. Ecology 60:901–909.

BENABIB, M. 1994. Reproduction and lipid utilizationof tropical populations of Sceloporus variabilis. Her-petological Monographs 8:160–180.

BULLOCK, S. H. 1986. Climate of Chamela, Jalisco, andtrends in the South Coastal Region of Mexico. Ar-chives for Meteorology, Geophysics, and Biocli-matology B 36:297–316.


DAVIS, W. B., AND J. R. DIXON. 1961. Reptiles (exclu-sive of snakes) of the Chilpancingo region, Mexico.Proceedings of the Biological Society of Washing-ton 74:37–56.

DUNHAM, A. E. 1978. Food availability as a proximatefactor influencing individual growth rates in theiguanid lizard Scelopous merriami. Ecology 59:770–778.

. 1981. Populations in a fluctuating environ-ment: the comparative population ecology of theiguanid lizards Sceloporus merriami and Urosaurusornatus. Miscellaneous Publications of the Museumof Zoology, University of Michigan 158:1–62.

. 1982. Demographic and life-history variationamong populations of the iguanid lizard Urosaurusornatus: implications for the study of life-historyphenomena in lizards. Herpetologica 38:208–221.

, AND D. B. MILES. 1985. Patterns of covariationin life history traits of squamate reptiles: the effectsof size and phylogeny reconsidered. AmericanNaturalist 126:231–257.

FITCH, H. S. 1970. Reproductive cycles in lizards andsnakes. University of Kansas Museum of NaturalHistory, Miscellaneous Publications 52:1–247.

. 1978. Sexual size differences in the genus Sce-loporus. University of Kansas Sciences Bulletin 51:441–461.

GOLDBERG, S. R., AND C. H. LOWE. 1966. The repro-ductive cycle of the western whiptail lizard (Cnem-idophorus tigris) in southern Arizona. Journal ofMorphology 118:543–548.

GORMAN, G. C., AND P. LICHT. 1974. Seasonality inovarian cycles among tropical Anolis lizards. Ecol-ogy 55:360–369.

GRANT, B. W., AND A. E. DUNHAM. 1990. Elevationalcovariation in environmental constraints and lifehistories of the desert lizard Sceloporus merriami.Ecology 71:1765–1776.

GUYER, C. 1988a. Food supplementation in a tropicalmainland anole, Norops humilis: demographic ef-fects. Ecology 69:350–361.

. 1988b. Food supplementation in a tropicalmainland anole, Norops humilis: effects on individ-uals. Ecology 69:362–369.

LEMOS-ESPINAL, J. A., AND R. E. BALLINGER. 1995.Ecology of growth of the high altitude lizard Sce-loporus grammicus on the eastern slope of Iztacci-hualtl volcano, Puebla, Mexico. Transactions of theNebraska Academy of Sciences 22:77–85.

LEMOS-ESPINAL, J. A., G. R. SMITH, AND R. E. BALLIN-GER. 1999. Reproduction in Gadow’s spiny lizard,Sceloporus gadovae (Phrynosomatidae), from aridtropical Mexico. Southwestern Naturalist 44:57–63.

LICHT, P. 1984. Reptiles. In G. E. Lamming (ed.), Mar-shall’s Physiology of Reproduction. Vol. 1, p. 206–282. Churchill Livingstone, Edinburgh, Scotland.

LICHT, P., AND G. C. GORMAN. 1970. Reproductiveand fat cycles in Caribbean Anolis lizards. Univer-sity of California Publications in Zoology 95:1–52.

MARION, K. R. 1982. Reproductive cues for gonadaldevelopment in temperate reptiles: temperatureand photoperiod effects on the testicular cycles ofthe lizard Sceloporus undulatus. Herpetologica 38:26–39.

MILES, B. D., AND A. E. DUNHAM. 1992. Comparativeanalysis of phylogenetic effects in the life-history

patterns of iguanid reptiles. American Naturalist139:848–869.

PARKER, W. S. 1973. Natural history notes on the ig-uanid lizard Urosaurus ornatus. Journal of Herpe-tology 7:21–26.

PARKER, W. S., AND E. R. PIANKA. 1975. Comparativeecology of populations of the lizard Uta stansburi-ana. Copeia 1975:615–632.

RAMIREZ-BAUTISTA, A. 1994. Manual y claves ilustra-das de los anfibios y reptiles de la region de Cha-mela, Jalisco, Mexico.

. 1995. Demografıa y reproduccion de la la-gartija arborıcola Anolis nebulosus de la Region deChamela, Jalisco. Tesis doctoral, Facultad de Cien-cias, Universidad Nacional Autonoma de Mexico.

RAMIREZ-BAUTISTA, A., AND D. PARDO-DE LA ROSA.2002. Reproductive characteristics and cycle of thewidely foraging lizard, Cnemidophorus communisfrom Jalisco, Mexico. Southwestern Naturalist 47:205–214.

RAMIREZ-BAUTISTA, A., AND L. J. VITT. 1997. Repro-duction in the lizard Anolis nebulosus (Polychroti-dae) from the Pacific coast of Mexico. Herpetolo-gica 53:423–431.

. 1998. Reproductive biology of Urosaurus bi-carinatus (Sauria:Phrynosomatidae) from a tropicaldry forest of Mexico. Southwestern Naturalist 43:381–390.

RAMIREZ-BAUTISTA, A., Z. URIBE-PENA, AND L. J. GUIL-LETTE JR. 1995. Reproductive biology of the lizardUrosaurus bicarinatus bicarinatus (Reptilia: Phryno-somatidae) from Rıo Balsas Basin, Mexico. Herpe-tologica 51:24–33.

RAMIREZ-BAUTISTA, A., L. J. GUILLETTE JR., G. GUTIER-REZ-MAYEN, AND Z. URIBE-PENA. 1996. Reproduc-tive biology of the lizard Eumeces copei (Lacertilia:Scincidae) from the Eje Neovolcanico, Mexico.Southwestern Naturalist 41:103–110.

RAMIREZ-BAUTISTA, A., C. BALDERAS-BALDIVIA, AND L.J. VITT. 2000. Reproductive ecology of the whip-tail lizard Cnemidophorus lineatissimus (Squamata:Teiidae) in a tropical dry forest. Copeia 2000:712–722.

ROSE, B. 1982. Food intake and reproduction in Anolisacutus. Copeia 1982:322–330.

RUBY, D. E. 1981. Phenotypic correlates of male re-productive success in the lizard, Sceloporus jarrovi.In R. D. Alexander and D. W. Tinkle (eds.), NaturalSelection and Social Behavior, pp. 96–107. ChironPress, New York.

SELBY, S. M. 1965. Standard Math Tables. 14th ed.Chemical Rubber Co., Cleveland, OH.

SMITH, G. R., AND R. E. BALLINGER. 1994. Temporaland variation in individual growth in the spinylizard, Sceloporus jarrovi. Copeia 1994:1004–1013.

SMITH, G. R., R. E. BALLINGER, AND B. R. ROSE. 1995.Reproduction in Sceloporus virgatus from the Chir-icahua Mountains of southeastern Arizona, withemphasis on annual variation. Herpetologica 51:342–349.

SMITH, G. R., J. A. LEMOS-ESPINAL, AND R. E. BALLIN-GER. 1997. Sexual dimorphism in two species ofknob-scaled lizards (genus Xenosaurus) from Mex-ico. Herpetologica 53:200–205.

STAMPS, J. A. 1983. Sexual selection, sexual dimor-phism, and territoriality. In R. B. Huey, E. R. Pian-


ka, and T. W. Schoener (eds.), Lizard Ecology:Studies of a Model Organism, pp. 169–204. Har-vard Univ. Press, Cambridge, MA.

TINKLE, D. W. 1969. The concept of reproductive ef-fort and its relation to the evolution of life historiesof lizards. American Naturalist 103:501–516.

TINKLE, D. W., AND R. E. BALLINGER. 1972. Sceloporusundulatus: a study of the intraspecific comparativedemography of a lizard. Ecology 53:570–584.

TINKLE, D. W., AND A. E. DUNHAM. 1986. Compara-tive life history of two syntopic sceloporine lizards.Copeia 1986:1–18.

TINKLE, D. W., H. M. WILBUR, AND S. TILLEY. 1970.Evolutionary strategies in lizard reproduction.Evolution 24:55–74.

VAN LOBEN SELS, R. C., AND L. J. VITT. 1984. Desertlizard reproduction: Seasonal and annual variationin Urosaurus ornatus (Iguanidae). Canadian Journalof Zoology 62: 1779–1787.

VINEGAR, M. B. 1975. Demography of the striped pla-teau lizard, Sceloporus virgatus. Ecology 56:172–182.

VITT, L. J. 1982. Reproductive tactics of Ameiva ameiva(Lacertilia: Teiidae) in a seasonally fluctuatingtropical habitat. Ecology 60:3113–3120.

. 1983. Reproduction and sexual dimorphism

in the tropical teiid lizard Cnemidophorus ocellifer.Copeia 1983:359–366.

. 1986. Reproductive tactics of sympatric gek-konid lizards with a comment on the evolutionaryand ecological consequences of invariant clutchsize. Copeia 1986:776–786.

. 1990. The influence of foraging mode andphylogeny on seasonality of tropical lizard repro-duction. Papeıs Avulsos Zoologia (Sao Paulo) 37:107–123.

. 1992. Diversity of reproduction strategiesamong Brazilian lizards and snakes: the signifi-cance of lineage and adaptation. In W. C. Hamlett(ed.), Reproductive Biology of South AmericanVertebrates, pp. 135–149. Springer-Verlag, NewYork.

VITT, L. J., AND J. D. CONGDON. 1978. Body shape,reproduction effort, and relative clutch mass in liz-ards: resolution of a paradox. American Naturalist112:595–608.

VITT, L. J., AND R. D. OHMART. 1975. Ecology, repro-duction, and reproductive effort in the iguanid liz-ard Urosaurus graciosus on the lower Colorado Riv-er. Herpetologica 31:56–65.

Accepted: 1 April 2002.


Quantitative Assessment of Habitat Preferences for the Puerto RicanTerrestrial Frog, Eleutherodactylus coqui

KAREN H. BEARD,1 SARAH MCCULLOUGH, AND ANNE K. ESCHTRUTH

School of Forestry and Environmental Studies, Yale University, New Haven, Connecticut 06511, USA

ABSTRACT.—We conducted a quantitative analysis of adult and juvenile Eleutherodactylus coqui (coquı)habitat preferences in Puerto Rico. The analysis consisted of two surveys: one to quantify potential habitatand another to quantify habitat use. Coquıs were found to use most habitats available to them; however,adults and juveniles preferred different plant species, habitat structural components, and heights from theforest floor. Adult and juvenile coquıs had opposite associations with many important plant species in theforest (e.g., Prestoea montana and Heliconia carabea) and habitat structural components. Adults had a neg-ative association with leaves and a positive association with leaf litter. Juveniles showed the opposite trend.Adults were more evenly distributed with respect to height than were juveniles, with adults preferringheights around 1.1 m and juveniles preferring heights closer to the forest floor. The quantitative surveytechnique for determining habitat preferences used in this study generally confirmed coquı habitat pref-erences known from qualitative assessments.

Understanding a species ecological role andpredicting the effect of habitat change on a spe-cies requires knowledge of habitat preferences.

1 Corresponding Author. Present address: Depart-ment of Forest, Range, and Wildlife Sciences and theEcology Center, Utah State University, 5230 Old MainHill, Logan, Utah 84322-5230; E-mail: [email protected]

Most studies to date have determined amphib-ian habitat preferences based on qualitative as-sociations (e.g., Cooke and Frazer, 1976; Beebee,1977; Strijbosch, 1979; Pavignano et al., 1990; Il-dos and Ancona, 1994). These studies are usefulbecause they are cost-efficient and easy to con-duct (Margules and Augustin, 1991). However,more labor-intensive quantitative assessmentsgenerally provide a more accurate picture of

11ELEUTHERODACTYLUS COQUI HABITAT PREFERENCES

habitat preference (Arthur et al., 1996; Poole etal., 1996; Mercer et al., 2000). The purpose of thisstudy was to determine whether qualitative andquantitative analyses yield different results re-garding the habitat preferences of a terrestrialamphibian.

The most abundant nocturnal species in thesubtropical wet forests of Puerto Rico is a ter-restrial frog, Eleutherodactylus coqui, known asthe coquı. Natural coquı densities are among thehighest known for any amphibian species in theworld and have been estimated at 20,000 indi-viduals/ha (Stewart and Woolbright, 1996).They are important nocturnal predators andconsume an astounding 114,000 invertebrates/ha/night (Stewart and Woolbright, 1996). In ad-dition, their densities are associated with chang-es in invertebrate densities, herbivory, plantgrowth, and leaf litter decomposition rates(Beard, 2001). Coquı density also increases fol-lowing hurricane disturbances that define thestructure and function of the ecosystem (Scatenaand Lugo, 1990; Scatena et al., 1996; Woolbright,1996; Foster et al., 1997). Therefore, knowledgeof coquı habitat preference contributes to an un-derstanding of ecosystem function.

Using qualitative approaches, researchershave identified coquı habitat preferences for ju-venile and adult coquıs (e.g., Pough et al., 1977;Formanowicz et al., 1981; Townsend et al., 1984;Townsend, 1985; Woolbright, 1985a; Townsend,1989). However, the results of those studieswere not verified with quantitative assessmentsand therefore may not accurately describe hab-itat preferences. In this research, we determinecoquı habitat preferences quantitatively by de-termining the habitat types that disproportion-ately serve as foraging habitat for both juvenileand adult coquıs. The results are compared toresults in previous studies that determined re-lationships between coquıs and habitat typesqualitatively to determine if the two methodsproduce similar results. Because coquı habitatpreferences may be based on a number of dif-ferent habitat characteristics, we quantified hab-itats using (1) plant species, (2) habitat structur-al components (e.g., branches, leaves, and soil),and (3) height from the forest floor (as in Town-send, 1989; Woolbright, 1996).

MATERIAL AND METHODS

Study Area.—Research sites were located inthe Luquillo Experimental Forest (LEF) in thenortheastern corner of Puerto Rico (188189N,658509W). The forest is classified as subtropicalwet (Ewel and Whitmore, 1973). Peak precipi-tation occurs between the months of May andNovember, with average inputs of about 400mm/month during these months, and drier pe-riods occur between January and April when

precipitation inputs average about 200–250mm/month (Garcia-Martino et al., 1996). Meanmonthly air temperatures are fairly constantthroughout the year and average between 21and 248C (Garcia-Martino et al., 1996). Hurri-cane Hugo passed through the forest in 1989and Hurricane George in 1998; both these hur-ricanes caused widespread defoliation andfelled trees.

Study sites were located in the secondary Ta-bonuco forest zone that is dominated by Dac-ryodes excelsa (Vahl.) (tabonuco), Prestoea montana(R. Graham) Nichols (sierra palm), Sloanea ber-teriana (choisy), and Cecropia schreberiana (Brownet al., 1983). The understory contains manyplant species but is dominated by Danea nodosa,Ichnanthus palens, Heliconia carabea, Piper glabres-cens, Pilea inegualis, Prestoea montana, and Scleriacanescens (Brown et al., 1983). Data were collect-ed in three 20 3 20 m plots located in the BisleyExperimental Watershed area in the LEF. Allthree plots were at elevations between 250 and300 m.

Habitat Survey.—During the months of Julyand August 1999, understory composition andstructure was quantified in three 20 3 20 mplots. Each plot was surveyed using three rowsof transects with each row comprised of fiveparallel transects, each 4 m long. Transectswithin rows were 3 m apart, and the rows were1 m apart.

Characterization of the tropical wet forest un-derstory by the density of plant species is dif-ficult for several reasons. Individuals are diffi-cult to identify because many have multiplestems and many plants are vines, so it is inap-propriate to use ground stem counts to assessspecies importance. In addition, equal counts ordensities of different species do not necessarilyindicate that they make an equal contribution topotential habitat because this type of count doesnot account for the volume of the species avail-able as habitat. It has been found that for thistype of forest a volumetric assessment of plantapparency (importance) (Cates, 1980) is a bettermethod for determining vegetation compositionand structure in the understory. The methodused in this study for characterizing habitat inthese plots is discussed in detail in Willig et al.(1993, 1998).

At each transect, apparencies were surveyedat seven evenly spaced heights (0.15 m, 0.46 m,0.76 m, 1.07 m, 1.37 m, 1.68 m, and 1.98 m) tovolumetrically assess habitat characteristics. Ateach height, any object that occurred on thetransect was tallied according to plant speciesand habitat structure. Any object that touchedthe string, extending between the transect end-points, was recorded as a ‘‘foliar hit’’ (Cook andStubbendier, 1986). The apparency of a plant

12 K. H. BEARD ET AL.

FIG. 1. The number of expected and observed ob-servations for (A) adult coquıs and (B) juvenile coquısfor each structural component in the Bisley Water-sheds, Luquillo Experimental Forest, Puerto Rico. BR5 branch, FB 5 fallen branch, FT 5 fallen trunk, FL5 flower, LF 5 leaf, LL 5 leaf litter, PT 5 petiole, PI5 plastic items (PVC pipes and plastic bags), RC 5rock, RT 5 root, SL 5 senesced leaf, ST 5 senescedstem/stalk, SN 5 snag, SO 5 soil, SST 5 stem/stalk/trunk, and TW 5 twig.

species or habitat structure in a particular sitewas estimated as the total number of foliar hitsby that species or habitat structure at any heighton all 15 transects within the site. Categories forhabitat structural components used in the anal-ysis are listed in Figure 1.

Frog Census Survey.—Frog surveys were con-ducted during the same month as the habitatsurveys, occurring between 20 and 29 August1999. Since coquıs are nocturnal, all frog sur-veys were conducted at night between 2000 and2400 h. During these hours, males typically call,and females and juveniles typically forage(Stewart and Woolbright, 1996). The ordering ofthe plots was alternated nightly to control fordeclining frog activity toward midnight (Stew-

art, 1985; Woolbright, 1985a). Three observerssurveyed a plot by walking slowly through theplots in an S-shape for 2 h. Frogs were locatedby visually inspecting soil, rocks, leaf litter, andvegetation up to a height of around 2 m. Theheight was appropriate because frogs can beconfidently observed up to 2 m in height, andthe majority of their activity occurs within 3 mof the ground (Stewart, 1985). For each frog ob-served, its habitat-by-plant species, habitatstructure, and height above ground was record-ed. Habitat structure categories are listed in Fig-ure 1. Height above ground was recorded to thenearest of the seven height categories. Frogswere also identified as either adults, meaningsnout–vent length (SVL) . 5 24 mm, or juve-niles, meaning SVL , 24 mm (Woolbright,1985b).

Statistical Analyses.—Goodness-of-fit G-statis-tics were calculated to assess whether coquıs ex-hibit habitat preferences with respect to habitatavailability for plant species (Sokal and Rohlf,1981). If coquıs have a random distribution withrespect to plant species, then the number of ob-servations on each plant species should be pro-portional to the relative apparency of each plant(ratio of the apparency of a species to the sumof the apparency of all species). For this test, thecalculated expected frequencies of coquı occur-rence, based upon plant apparencies, dictatedthat all but 14 of the most commonly occurringplant species were pooled into a single class tomaximize degrees of freedom. This resulted information of 15 classes, with no classes havingexpected values less than 5.00 (Sokal and Rohlf,1981). Similar tests were run for both juvenileand adult coquıs separately. For the test onadult coquıs, all but six of the common plantspecies were pooled, and on juvenile coquıs allbut 14 of the common plant species werepooled.

The goodness-of-fit G-statistic was also usedto assess spatial distribution with respect tohabitat structural components. If coquıs have arandom distribution with respect to habitatstructural components, then the number of ob-servations on each component should be pro-portional to the relative apparency of each com-ponent. For a more conservative adult and totalcoquı test, around 34% of the component obser-vations at 0.15 m were excluded from the anal-ysis because they occurred on plant species (i.e.,I. palens, S. canescens, P. inegualis, and seedlings)unable to physically support adult coquıs (KHB,pers. obs.). For the test on total coquıs, all but13 of the most commonly occurring habitatcomponent categories were pooled, with no cat-egories having frequencies less than 5.00. For thetest on adult coquıs, all but seven of the mostcommon habitat component categories were


pooled. For the test on juvenile coquıs, all but10 of the common habitat component categorieswere pooled. Habitat structural componentsused for the analysis are listed in Figure 1.

A goodness-of-fit G-statistic was also calcu-lated to assess spatial distribution with respectto height. If coquıs have a random distributionwith respect to height, then the number of ob-servations for each plant part should be propor-tional to the relative apparency of each height.For the statistical analysis, height categorieswere not consolidated. Zero height was not in-cluded in the analysis because it was not mea-surable in the habitat survey. Again, for a moreconservative adult and total coquı test, around34% of the component observations were ex-cluded from the analysis to control for plantsunable to physically support adult coquıs.

To conduct the G-statistic analyses, it was nec-essary to assume that coquı observations wereindependent. Based on the natural history of thespecies, the assumption is likely not to be truein some cases. For example, the assumption maybe violated when male coquıs exhibit territori-ality.

Chi-squared statistics were used to determinepositive and negative associations with partic-ular plant species, plant parts, and heights. Thethree plots were treated as three replications.Observed and expected frog observation prob-abilities were determined for each plot to con-duct the statistical analyses. Observed probabil-ities were determined from the frog census data.For example, the probability of observing a frogon a plant species, habitat component, or heightis the number of times a frog is observed on aspecies divided by the total number of times afrog is observed. Expected probabilities weredetermined from the habitat survey data. For ex-ample, the expected probability of observing afrog on a plant species, habitat component, orheight is equal to the number of hits for the spe-cies, component, or height divided by the totalnumber of hits. Expected probabilities weremultiplied by the number of frogs observed todetermine the expected number of frogs ob-served.

G-statistic and chi-squared statistics were con-ducted using Microsoft Excel for Windows 2000.For all tests, significance was detected if P ,0.05.

RESULTS

During the habitat survey, a total of 1481 dif-ferent habitat components touched the transectlines from the three plots. The apparencies of 60species of herbaceous and woody plants werecalculated based on this data. During the frogcensus survey, 614 coquıs were observed fromthe three plots. Of the total specimens, 267 were

characterized as adults and 347 as juveniles. Ofthe total number of observations, 492 weremade on identifiable plant species.

Coquıs were observed on 51 different plantspecies. The most important plant species weredefined as those that were observed in both thehabitat survey and frog census and that had atleast a total of 15 observations in either. Therewere 17 species that fell into this category (Table1). The relative apparencies of these speciesranged from 0.002 to 0.242. The other specieshad a cumulative relative apparency of 0.108,with each species having an apparency less than0.003 on average. These species do not representa significant portion of the taxonomic or struc-tural components of the understory, and theyare not considered further. Only three specieshad more than five hits in the plant survey andnone in the frog census.

Coquıs exhibited a nonrandom spatial distri-bution in the environment in relation to plantspecies (df 5 14, G 5 110.69, P , 0.001). Morespecifically, adult and juvenile coquıs when an-alyzed as separate groups were found to exhibita nonrandom spatial distribution with respectto plant species (df 5 6, G 5 204.52, P , 0.001;df 5 14, G 5 260.82, P , 0.001). Adult and ju-venile coquıs exhibited opposite associations, ei-ther positive or negative, with important plantspecies, such as Prestoea and Heliconia (Table 1).The negative associations between adults andgrasses were not considered because these spe-cies are unable to physically support adult co-quıs.

Coquıs were observed on 16 habitat structuralcomponents (Fig. 1). Only categories with atleast three observations in either the habitat sur-vey or frog census survey were considered. Therelative apparencies of these habitat componentsranged from 0.00 to 0.48. Unlike their relation-ship to plant species, total coquıs exhibited arandom spatial distribution with respect toplant parts (df 5 13, G 5 21.26, P . 0.05). How-ever, adult and juvenile coquıs exhibited a non-random spatial distribution with respect toplant parts (df 5 7, G 5 106.61, P , 0.001; df5 10, G 5 151.03, P , 0.001). Adult and juvenilecoquıs did not exhibit the same associations, ei-ther positive or negative, with respect to the ma-jority of plant part categories (Fig. 1). Adults ex-hibited a negative and juveniles a positive pref-erence for leaves (df 5 2, P , 0.05). For bothleaf litter and fallen branches, adults exhibiteda positive and juveniles a negative preference(df 5 2, P , 0.05). Other significant associationsincluded positive preference for roots by adults;negative selection for twigs and petioles byadults; and negative preference for stem/stalk/trunk, twigs, and senesced leaves by juveniles(df 5 2, P , 0.05).


TABLE 1. Apparency and coquı preference for the most commonly observed plant species in the BisleyWatersheds, Luquillo Experimental Forest, Puerto Rico. Associations were determined using a chi-squared test(df 5 2, P , 0.05). NS 5 Not Significant, NA 5 Not Applicable.

HabitatSpecies Family

Censusobservations

total prob.

Transect hits

total prob. Difference Total AdultsJuve-niles

Trees:Cecropia schreberianaDacryodes excelsaPrestoea montanaSloanea berteriana

CecropiaceaeBurseraceaePalmaceaeEleocarpaceae

85

10215

0.01780.01110.22710.0334

711

30956

0.00540.00850.23970.4034

0.01240.0026

20.012520.0100

1NS22

111NS

NSNS22

Vines:Marcgravia rectifloraMikonia cordifoliaPhilodendron angustatumRourea surinamensis

MarcgraviaceaeAsteraceaeAraceaeConnaraceae

917

91

0.02000.03780.02000.0022

34772420

0.02630.05970.01860.0155

20.006320.0218

0.001420.0132

NS21NS

NS2NSNS

NS2NSNS

Ferns:Cyathea borinquensisDanea nodosaThelypteris deltoidea

CyatheaceaeMarattiaceaeThelypteridaceae

52217

0.01110.04890.0378

213272

0.01620.02480.0558

20.00510.0241

20.0179

21NS

NSNSNS

21NS

Shrubs:Heliconia carabeaPilea inegualisPiper glabrescensPsychotria berteriana

HeliconiaceaeUrticaceaePiperaceaeRubiaceae

25142813

0.05560.03110.06230.0289

623028

3

0.04800.02320.02170.0023

0.00750.00790.04060.0266

1111

1NS1NS

2111

Grasses:Ichnanthus palensScleria canescens

PoaceaeCyperaceae

771

0.17140.0022

31252

0.24200.0403

20.070520.0381

NANA

NANA

22

Coquıs exhibited a nonrandom spatial distri-bution with respect to height (df 5 6, G 5 31.06,P , 0.001). They had a positive preference forthe forest floor 0.46 m and a negative preferencefor 0.15 m, 1.07 m, and 1.68 m (df 5 2, P ,0.05). Adult coquıs independently exhibited arandom spatial distribution with respect toheight meaning that they were found at mostheights in proportions similar to the amount ofstructure available at those heights (df 5 6, G 55.13, P . 0.05). Adults had a positive preferencefor 0.76 m, and 1.07 m and a negative preferencefor 0.15 m and 1.68 m (df 5 2, P , 0.05). Incontrast, juvenile coquıs exhibited a nonrandomspatial distribution with respect to plant height,meaning that they had height preferences (df 56, G 5 45.75, P , 0.001). Juveniles exhibited apositive preference for 0.46 m, 0.76 m, and 1.37m and a negative preferences for 0.15 m, 1.07m, and 1.68 m (df 5 2, P , 0.05).

DISCUSSION

Coquıs may depend on certain common hab-itat features for reproduction (Stewart andPough, 1983; Townsend, 1989; Woolbright,1996), but they have less specific habitat require-ments for calling and foraging. Although coquıhabitat preferences are strongly partitioned ac-

cording to life stage, as a whole, coquıs are gen-eralists in their preferences for plant species,structural components, and heights from theforest floor. The ecological effects of coquıs aredifficult to observe at large spatial scales (Beard,2001). Their general use of habitat may explainthis phenomenon since effects of habitat gener-alists are often less distinguishable on the land-scape than that of specialists; their effects arenot ‘‘localized’’ and as a result occur at slowerrates (Jeffries et al., 1994).

Coquıs generally have positive associationswith shrubs and negative associations withgrasses, vines, and ferns. Exceptions includePhilodendron angustatum and Danea nodosa, whichboth have unusually broad leaf structures fortheir respective plant classification categoriesand, therefore, are better able to structurallysupport coquıs than other species in those hab-itat categories. The positive association betweencoquıs and particular species, such as Cecropiaand Heliconia, support findings from previouswork that showed a positive relationship be-tween these species and coquı density (Wool-bright, 1996).

The data illustrate ontogenetic shifts in for-aging habitat preferences by adult and juvenilecoquıs for the tree habitat type category and for


some important plant species, such as Prestoeaand Heliconia (Table 1). Results suggest thatadult coquıs selected plant species that werebest able to structurally support them, and ju-venile coquıs selected plant species that were lo-cated close to the forest floor. Juvenile coquıpreferences may result from a combination ofjuveniles satisfying moisture requirements(Pough et al., 1983) and the ability of low-lyingshrubs and seedlings to physically support ju-veniles. Structural support has been found to bea major factor in determining plant species pref-erences for other faunal species in the study for-est (Willig et al., 1998).

Prestoea (sierra palm) provides an interestingexample of an ontogenetic shift in habitat pref-erences in this study. Although adult coquıswere found to have a positive association withsierra palm, when adults and juveniles were an-alyzed together, they had a negative associationwith sierra palm (Table 1). This occurred be-cause juveniles had a negative association withsierra palm and they are more abundant thanadults. This finding illustrates the problems as-sociated with conducting habitat use studies us-ing only one life-history stage without consid-ering the relative spatial distributions of otherstages.

Coquıs, in general, used habitat componentsin proportion to their availability in the environ-ment, although juvenile and adult coquıs greatlydiffered in their preferences for habitat compo-nents. Adult coquıs generally used a muchgreater variety of habitat components than didjuveniles (Fig. 1). Adults positively selected forobjects near the ground, including leaf litter,fallen branches, roots, and soil. Leaf litter hereis distinguished from leaves and senescedleaves. Leaf litter was considered those leavesthat are dead and fallen, and frequently on theforest floor, but also caught on understory com-ponents after falling. Adult coquıs had negativeassociations with leaves and stems/stalks/trunks, even though they used these habitats.These results contradict previous studies sug-gesting that adult coquıs prefer to forage onleaves and trunks (Townsend, 1989; Stewart andWoolbright, 1996). However, this study was con-ducted during one season, and our results maybe the result of specific site conditions, such asmoisture availability, that vary throughout theyear. Juvenile coquıs were found to prefer leavesand to avoid leaf litter as has been found in oth-er studies (Townsend 1985). Juvenile coquıs mayavoid leaf litter because of predator pressures(Stewart and Woolbright, 1996).

Coquıs were generalists in their use of habi-tats at different heights from the forest floor, butsegregation of juvenile and adult coquıs byheight was notable. Adults had a wider range of

preferences for heights from the forest floor thandid juveniles. Adults exhibited negative selec-tion for 0.15 m, large positive selection for 1.1m, and negative selections for greater heights.This supports other findings that indicate thatadult coquı prefer heights around 1 m (Stewart,1985; Townsend, 1985). As expected, juvenile co-quıs tended to use substrates close to theground, with 64% using heights from 0.15–0.45m from the forest floor (Townsend, 1985). How-ever, juvenile preferences for heights did notcompletely agree with this qualitative observa-tion; for example, they had a negative selectionfor 0.15 m, the lowest height category measuredabove the forest floor, although they were fre-quently found there. This is an example of howa quantitative assessment can highlight a habitatpreference not observable with a qualitativeanalysis.

The results should be viewed in light of thesuccessional changes occurring in the forest.Two major hurricanes passed through the forestin the decade before this study. Adult coquıshad a strong positive association with dead, fall-en leaves and early successional species, such asCecropia, Heliconia, and Prestoea. Interestingly, co-quı abundance increases following disturbanceevents when early successional plants are mostabundant (Woolbright, 1991). The results sug-gest that as early successional species, such asCecropia and Heliconia, are replaced coquıs willbe forced to switch from the most preferred spe-cies to less preferred species. Therefore, factorsrelated to forest succession may impact coquıs.One explanation for this pattern may be thepreference to breed in these habitats (Townsend,1989). The mechanisms affecting coquı popula-tions should be further explored.

A number of factors could create juvenile andadult coquı habitat partitioning. It has been sug-gested that coquı moisture requirements serveas the primary factor determining the distribu-tion pattern (Pough et al., 1977; Townsend,1985). Coquı prey limitation lends support forthe hypothesis that intraspecific competition de-termines the pattern (Toft, 1985; Beard, 2001).Studies thus far have shown that differentialpredation by life history stages does not deter-mine the pattern (Formanowicz et al., 1981). Ju-venile and adult coquı habitat partitioning hasbeen explored elsewhere (Townsend, 1985);however, the mechanism determining the pat-tern remains uncertain.

As may be expected for an ubiquitous speciesthat is endemic to an island community sub-jected to frequent disturbances, the coquı is ahabitat generalist. Unlike other Eleutherodactylusspp. in Puerto Rico, which have more specifichabitat requirements, such as the cave-dwellingEleutherodactylus cooki, the coquı has not experi-


enced population declines (Joglar and Burrow-es, 1996; Joglar et al., 1996). It appears that thecoquı is easier to conserve because it requiresfew specific habitat features. Alternatively, thistrait makes it more difficult to manage coquısoutside of their native range, for example, in Ha-waii where coquıs have been introduced recent-ly (Kraus et al., 1999).

The results describing coquı habitat prefer-ences found in this study using quantitativemethods generally support results found in pre-vious studies on coquı habitat preferences usingqualitative methods, although there are excep-tions. This suggests that in some cases, easier-to-conduct qualitative surveys may be used inplace of the more labor-intensive quantitativemethods to assess habitat preferences. Furtherstudies will provide more guidance in the typesof species and habitat where these methods areinterchangeable.

Acknowledgments.—The Yale Institute of Bio-spheric Studies provided funding for SMcC andthe Tropical Resources Institute provided fund-ing for KHB and AKE. Further support wasprovided by the AAUW Educational Foundationand the U.S.D.A. Institute of Tropical Forestryin Puerto Rico. F. Scatena, C. Estrada, and C.Shipek provided field assistance. K. Vogt, A.Kulmatiski, T. Gregoire, and two anonymous re-viewers provided useful comments on earlierversions of this manuscript.

LITERATURE CITED

ARTHUR, S. M., B. F. J. MANLY, L. L. MCDONALD, ANDG. W. GARNER. 1996. Assessing habitat selectionwhen availability changes. Ecology 77:215–227.

BEARD, K. H. 2001. The Ecological Roles of a Terres-trial Frog, Eleutherodactylus coqui, in the NutrientCycles of a Subtropical Wet Forest in Puerto Rico.Unpubl. Ph.D. diss., Yale Univ. New Haven, CT.

BEEBEE, T. J. C. 1977. Habitat selection by amphibiansacross an agricultural land healthland transect inBritain, UK. Biological Conservation 27:111–124.

BROWN, S., A. E. LUGO, S. SILANDER, AND L. LIEGEL.1983. Research History and Opportunities in theLuquillo Experimental Forest. USDA Forest Ser-vice, New Orleans, LA.

CATES, R. G. 1980. Feeding patterns of monophagous,oligophagous, and polyphagous insect herbivores:the effects of resource abundance and plant chem-istry. Oecologia 46:22–31.

COOK, C. W., AND J. STUBBENDIER. 1986. Range Re-search: Basic Methods and Techniques. Society forRange Management, Denver, CO.

COOKE, A. S., AND J. F. D. FRAZER. 1976. Character-istics of newt breeding sites. Journal of Zoology(London) 175:29–38.

EWEL, J. J., AND J. L. WHITMORE. 1973. The EcologicalLife Zones of Puerto Rico and the U.S. Virgin Is-lands. U.S. Forest Service, Institute of Tropical For-estry, Rio Piedras, Puerto Rico.

FORMANOWICZ JR., D. R., M. M. STEWART, K. TOWN-

SEND, F. H. POUGH, AND P. F. BRUSSARD. 1981. Pre-dation by giant crab spiders on the Puerto Ricofrog Eleutherodactylus coqui. Herpetologica 37:125–129.

FOSTER, D. R., M. FLUET, AND E. R. BOOSE. 1997. Hu-man or natural disturbance: landscape-scale dy-namics of the tropical forests of Puerto Rico. Eco-logical Applications 9:555–572.

GARCIA-MARTINO, A. R., G. S. WARNER, F. N. SCA-TENA, AND D. L. CIVCO. 1996. Rainfall, runoff andelevation relationships in the Luquillo Mountainsof Puerto Rico. Caribbean Journal of Science 32:413–424.

ILDOS, A. S., AND N. ANCONA. 1994. Analysis of am-phibian habitat preferences in a farmland area (Poplain, northern Italy). Amphibia-Reptilia 15:307–316.

JEFFRIES, R. L., D. R. KLEIN, AND G. R. SHAVER. 1994.Vertebrate herbivores and northern plant commu-nities: reciprocal influences and responses. Oikos71:193–206.

JOGLAR, R. L., AND P. A. BURROWES. 1996. Decliningamphibian populations in Puerto Rico. In R. Powelland R. W. Henderson (eds.), Contributions to WestIndian Herpetology: A Tribute to Albert Schwartz,pp. 371–380. Society of Amphibians and Reptiles,Ithaca, NY.

JOGLAR, R. L., P. A. BURROWES, AND N. RIOS. 1996.Biology of the Puerto Rican cave-dwelling frog,Eleutherodactylus cooki, and some recommendationsfor its conservation. In R. Powell and R. W. Hen-derson (eds.), Contributions to West Indian Her-petology: A Tribute to Albert Schwartz, pp. 251–258. Society for the Study of Amphibian and Rep-tiles, Ithaca, NY.

KRAUS, F., E. W. CAMPBELL, A. ALLISON, AND T. PRATT.1999. Eleutherodactylus frog introductions to Ha-waii. Herpetological Review 30:21–25.

MARGULES, C. R., AND M. P. AUGUSTIN (EDS.). 1991.Nature conservation: cost-effective biological sur-veys and data analysis. CSIRO, Melbourne, Victo-ria, Australia.

MERCER, R. D., S. L. CHOWN, AND D. MARSHALL. 2000.Mite and insect zonation on a Marion Island rockyshore: a quantitative approach. Polar Biology 23:775–784.

PAVIGNANO, I., C. GIACOMA, AND S. CASTELLANO.1990. A multivariate analysis of amphibian habitatdeterminants in north western Italy. Amphibia-Reptilia 11:311–324.

POOLE, K. G., L. A. WAKELYN, AND P. N. NICKLEN.1996. Habitat selection by lynx in the NorthwestTerritories. Canadian Journal of Zoology. 74:845–850.

POUGH, F. H., M. M. STEWART, AND R. G. THOMAS.1977. Physiological basis of habitat partitioning inJamaican Eleutherodactylus. Oecologia 27:285–293.

POUGH, F. H., T. L. TAIGEN, M. M. STEWART, AND P. F.BRUSSARD. 1983. Behavioral modification of evap-orative water loss by a Puerto Rican frog. Ecology64:244–252.

SCATENA, F. N., AND A. E. LUGO. 1990. Natural Dis-turbance and the Distribution of Vegetation in TwoSubtropical-Wet Steepland Watersheds of PuertoRico. Institute of Tropical Forestry, U.S. Forest Ser-vice, Rio Piedras, Puerto Rico.


SCATENA, F. N., S. MOYA, C. ESTRADA, AND J. D. CHI-NEA. 1996. The first five years in the reorganiza-tion of aboveground biomass and nutrient use fol-lowing Hurricane Hugo in the Bisley experimentalwatersheds, Luquillo Experimental Forest. Biotro-pica 28:242–440.

SOKAL, R. R., AND R. J. ROHLF. 1981. Biometry. W. H.Freeman, Co., San Francisco, CA.

STEWART, M. M. 1985. Arboreal habitat and para-chuting by a subtropical forest frog. Journal of Her-petology. 19:391–401.

STEWART, M. M., AND F. H. POUGH. 1983. Populationdensity of tropical forest frogs: relation to retreatsites. Nature 221:570–572.

STEWART, M. M., AND L. L. WOOLBRIGHT. 1996. Am-phibians. In D. P. Reagan and R. B. Waide (eds.),The Food Web of a Tropical Rain Forest, pp. 363–398. Univ. of Chicago Press, Chicago.

STRIJBOSCH, H. 1979. Habitat selection of amphibiansduring their terrestrial phase. British Journal ofHerpetology. 6:93–98.

TOFT, C. A. 1985. Resource partitioning in amphibi-ans and reptiles. Copeia 1985:1–21.

TOWNSEND, D. S. 1989. The consequences of micro-habitat choice for male reproductive success in atropical frog (Eleutherodactylus coqui). Herpetologi-ca 45:451–458.

TOWNSEND, D. S., M. M. STEWART, AND F. H. POUGH.

1984. Male parental care and its adaptive signifi-cance in a Neotropical frog. Animal Behavior. 32:421–431.

TOWNSEND, K. V. 1985. Ontogenetic Shift in HabitatUse by Eleutherodactylus coqui. Unpubl. master’sthesis. State Univ. of New York, Albany.

WILLIG, M. R., E. A. SANDLIN, AND M. R. GANNON.1993. Structural and taxonomic components ofhabitat selection in the Neotropical folivore, Lam-ponius portoricensis Rehn (Phasmatodea: Phasma-tidae). Environmental Entomology. 22:634–641.

. 1998. Structural and taxonomic correlates ofhabitat selection by a Puerto Rican land snail.Southwestern Naturalist 43:70–79.

WOOLBRIGHT, L. L. 1985a. Patterns of nocturnalmovement and calling by the tropical frog Eleuth-erodactylus coqui. Herpetologica 41:1–9.

. 1985b. Sexual Dimorphism in Body Size ofthe Subtropical Frog, Eleutherodactylus coqui. Un-publ. Ph.D. diss., State Univ. of New York, Albany.

. 1991. The impact of Hurricane Hugo on for-est frogs in Puerto Rico. Biotropica 23:462–467.

. 1996. Disturbance influences long-term pop-ulation patterns in the Puerto Rican frog, Eleuth-erodactylus coqui (Anura: Leptodactylidae). Biotro-pica 28:493–501.



Multiple Clutching in Southern Spotted Turtles, Clemmys guttata

JACQUELINE D. LITZGUS1 AND TIMOTHY A. MOUSSEAU

Department of Biological Sciences, University of South Carolina, Columbia, South Carolina 29208, USA

ABSTRACT.—We examined the reproductive output of spotted turtles (Clemmys guttata) from a populationin South Carolina. We used radio telemetry, palpation, and x-rays to monitor the reproductive condition offemales over two field seasons. We present the first evidence for multiple clutching in a wild population ofspotted turtles. Of 12 females with radio transmitters that became gravid, five produced second clutches,and one produced a third clutch. Average annual clutch frequency was 1.2 per female. Clutch frequencywas independent of body size. We compared reproductive output among three populations: Ontario, Penn-sylvania, South Carolina. Individual clutch sizes varied with latitude. Clutch size was largest in the north(mean 5 5.3 eggs), midsized in the central population (3.9), and smallest in the south (2.9). We suggest thatthis pattern is related to seasonality differences, which result in different selective pressures on body sizeof females. Total annual egg production (the sum of all clutches within a reproductive season) by gravidfemales did not differ between the Ontario (5.3 eggs) and South Carolina populations (4.6). These dataindicate that, although individual clutch sizes differ between northern and southern spotted turtles, totalannual reproductive output is consistent in these widely separated populations.

Latitudinal variation in clutch size has been re-ported for many vertebrates. Studies have ex-amined variation in reproductive output amongconspecific populations of mammals (e.g., Lord,

1 Corresponding Author. E-mail: [email protected]

1960; May and Rubenstein, 1985), birds (e.g.,Ricklefs, 1980; Godfray et al., 1991), amphibians(e.g., Cummins, 1986), fishes (e.g., Healey andHeard, 1984; Fleming and Gross, 1990), and rep-tiles (Moll, 1979; Fitch, 1985; Sinervo, 1990). Typ-ically, clutch size increases with increasing lati-

18 J. D. LITZGUS AND T. A. MOUSSEAU

tude (for summaries of hypotheses proposed toexplain the adaptive significance of latitudinalvariation in clutch and egg size, see Iverson et al.,1993).

Among intraspecific turtle populations, lati-tudinal variation in both clutch size and clutchfrequency have been reported. Clutch frequencyhas been considered one of the most importantaspects of turtle reproduction (Gibbons, 1982).Turtles at northern latitudes experience shortersummers and tend to have lower clutch fre-quencies and increased individual clutch sizesrelative to southern conspecifics (Moll, 1973,1979; Iverson, 1992; Iverson and Smith, 1993;Iverson et al., 1993, 1997; Zuffi and Odetti,1998). Turtles occurring at southern latitudesmay produce several clutches of eggs in a singlereproductive season. For example, female paint-ed turtles (Chrysemys picta; Moll, 1973) and slid-er turtles (Trachemys scripta; Jackson, 1988) fromsouthern locations may produce up to fiveclutches in a season.

Spotted turtles (Clemmys guttata) have neverbeen reported to produce more than one clutchof eggs per year in the wild (e.g., Ernst, 1970,1976, Litzgus and Brooks, 1998), although cap-tive females may produce two clutches per year(Wilson, 1989; Ernst and Zug, 1994). Indeed,spotted turtles at the northern extreme of therange in Ontario (ON), where summers areshort and relatively cool, have a clutch frequen-cy of less than one clutch per year (Litzgus andBrooks, 1998). In a four-year study on the ONpopulation, just over half of the adult femalesbecame gravid each year, most females did notoviposit every year, and 16% of females did notproduce eggs in three consecutive years (Litz-gus and Brooks, 1998). In Pennsylvania (PA),near the central part of the species’ range, fe-male spotted turtles produced one clutch peryear (Ernst, 1970, 1976).

Spotted turtles range from 458N southward to288N (Ernst et al., 1994; Barnwell et al., 1997);however, little is known about the general ecol-ogy of southern populations. We are conductinga three-year radio telemetry study on a spottedturtle population in South Carolina (SC). Ourstudy includes an examination of seasonal activ-ity and habitat use, reproductive behavior, andthe adaptive significance of relationships amongegg size, clutch size, clutch frequency, and fe-male body size within and among populationsof the spotted turtle across its range. The objec-tives of the current paper are (1) to present thefirst evidence for multiple clutching in a wildpopulation of spotted turtles, (2) to examine therelationship between female body size andclutch frequency within the southern populationstudied, and (3) to examine latitudinal variation

in reproductive output among populations ofthis wide-ranging species.

MATERIALS AND METHODS

The spotted turtle ranges from southern On-tario and Maine, southward along the AtlanticCoastal Plain to central Florida, and westwardalong the southern shores of the Great Lakes toIllinois (Ernst et al., 1994; Barnwell et al., 1997).This relatively small freshwater species has beenassigned varying degrees of declining and pro-tected status throughout its range (Lovich andJaworski, 1988; Ernst et al., 1994). The two pri-mary reasons for decline of the spotted turtleare habitat loss and collection for the pet trade(Lovich and Jaworski, 1988; Burke et al., 2000).

This ongoing study is being conducted on theAtlantic Coastal Plain at the Francis Beidler For-est National Audubon and Nature ConservancySanctuary, Harleyville, SC (338N, 808W), in thesouthern part of the spotted turtle’s range. Bei-dler Forest is a 4452-ha sanctuary that includesupland pine (Pinus taeda, Pinus glabra) and me-sophytic hardwood (e.g., Quercus spp., Fagusgrandifolia, Liquidambar styraciflua) forest, season-ally flooded hardwood bottomland swamp for-est (distinguished by the presence of dwarf pal-metto, Sabal minor), and 728 ha of old-growthbald cypress (Taxodium distichum) and tupelogum (Nyssa aquatica) swamp forest. For a de-tailed description of the Beidler Forest habitatsand their associated plant communities seePorcher (1981).

In the 2000 field season (March to December),four females were outfitted with radio trans-mitters (163 MHz, AVM Instrument, Livermore,CA). In 2001 (January to December), 11 femaleswere outfitted with radio transmitters (164MHz, Holohil Systems Inc., Carp, ON); three ofthese 11 females were also tracked in the 2000field season. Turtles were radiolocated using aTelonics receiver and a three-element, folding,aluminum Yagi antenna. Gravid females werelocated twice daily, once during the day andagain at sundown, until they oviposited.

From March to July, the reproductive condi-tion of female turtles outfitted with radio trans-mitters was determined by palpation, change inbody mass, and x-ray photography (Gibbonsand Greene, 1979). Turtles were x-rayed at 52kVpeak and 2.5 mA·s at a local veterinary clinic.Clutch size and clutch frequency were deter-mined from x-rays and observed oviposition.Using our current field data and data publishedin the literature, we compared clutch frequencyand average individual clutch sizes among pop-ulations of spotted turtles. We treated eachclutch as independent for the calculation ofclutch size; thus, for a female that produced twoor more clutches (whether in the same year or

19MULTIPLE CLUTCHING IN CLEMMYS GUTTATA

TABLE 1. Clutch size and clutch frequency data determined from x-rays of spotted turtles (Clemmys guttata)from South Carolina in two field seasons. Data for the year 2000 are shown at the top of the table; data for2001 are shown below the horizontal line. Exact oviposition date (‘‘Lay date’’) was known for 13 nests and wasknown for within five days for five nests. Also shown is the total number of eggs produced in a year by eachfemale.

TurtleI.D.

Clutch 1

X-raydate Lay date

No.eggs

Clutch 2

X-raydate Lay date

No.eggs

Clutch 3

X-raydate Lay date

No.eggs

Totalno.

eggs

61720

5/175/175/17

5/225/265/28

322

6/12 6/20–6/21 2 522

16

20

5/215/25/7

6/8–6/135/195/29

343

6/15 6/23 3 7/10 7/13–7/15 33

103

307080

102402700

5/75/75/25/75/25/21

5/285/215/195/295/216/5–6/7

233343

6/156/156/157/10

?6/246/207/13–7/15

3234

556743

in different years), each clutch was consideredan independent datum. We also compared totalannual egg production per female among spot-ted turtle populations.

RESULTS

In the 2000 field season, three of the four fe-males (75%) with transmitters became gravid;one female produced two clutches of eggs. In2001, nine of the 11 females (82%) with trans-mitters became gravid; four females producedtwo clutches, and one female produced threeclutches. Average annual clutch frequency forboth years combined was 1.2 per female. Of 18nests, exact dates of oviposition were deter-mined for 13 nests and approximate ovipositiondates (within five days) were determined forfive nests (Table 1). The nesting season lastedfrom mid-May to mid-July. Females that pro-duced multiple clutches did not necessarily ovi-posit earliest in the reproductive season, andclutch size did not decrease with clutch number(Wilcoxon 2-Sample Test: Z 5 0.4, P 5 0.7). Infact, in two cases, the number of eggs in thesecond clutch was greater than that in the firstclutch.

We tested whether there was a relationshipbetween female body size (plastron length, PL;independent variable) and clutch frequency (de-pendent variable) in the SC spotted turtle pop-ulation. Although the shape of the regressionline was positive and linear, only about 18% ofthe variation in clutch frequency was explainedby body size (F1,11 5 2.3, P 5 0.2). Both thesmallest (PL 5 84.8 mm) and largest (PL 5 98.5mm) females in the population produced twoclutches of eggs in a single reproductive season.

The female that produced a third clutch in 2001was the second largest (PL 5 96.5 mm) gravidfemale.

We examined geographic variation in repro-ductive output among spotted turtle popula-tions spanning the species’ latitudinal range. Wecompared average individual clutch size amongthree populations; ON (Litzgus and Brooks,1998), PA (Ernst, 1970), and SC (current study).Clutch size differed significantly among thethree populations (F2,50 5 46.9, P , 0.0001). Bon-ferroni a posteriori tests (a 5 0.017) indicatedthat all interpopulation comparisons were sig-nificantly different (ON vs. PA: t 5 3.9, df 5 30,P , 0.001; ON vs. SC: t 5 9.3, df 5 41, P ,0.0001; PA vs. SC: t 5 3.6, df 5 25, P , 0.01).Clutch size was largest in ON turtles, followedby PA turtles, and turtles in SC had the smallestindividual clutch sizes (Table 2).

Differences in average individual clutch sizesamong females from the three populations wasexplained by variation in body size (PL) and lat-itude (ANCOVA with PL as a covariate, F3,50 543.0, P , 0.0001, assumption of homogeneity ofslopes was met). Clutch size among populationswas affected by the population of origin of thefemale (i.e., latitude, F 5 14.5, P , 0.0001) andbody size (F 5 12.6, P , 0.001). Clutch size dif-fered between ON and SC, and between PA andSC when controlling for body size but did notdiffer between ON and PA (lsmeans compared:ON vs. PA: P . 0.05; ON vs. SC: P , 0.0001;PA vs. SC: P , 0.001).

Body size (PL, Table 2) of gravid females dif-fered among the three populations (F2,36 5 46.7,P , 0.0001). Turtles from ON were significantlylarger than turtles from PA (Bonferroni t 5 8.1,


TABLE 2. Body size (plastron length in mm) and reproductive output of female spotted turtles (Clemmysguttata) from three populations (location in 8N latitude) spanning the range of the species. Values are means 6SE. N 5 number of gravid females for the body size and mean total annual egg production comparisons, andN 5 number of clutches for the clutch size comparisons. Means within rows with different superscripts aresignificantly different (see text for statistical values).

Ontario(458N)

Pennsylvania(408N)

South Carolina(338N)

Body size (N)Clutch size (N)Range of number eggs in clutchesMean annual clutch frequencyMean total annual egg production

for gravid females (N)Total annual egg production#

103.4 6 0.9 (19)a

5.3 6 0.2 (24)a

4–70.7†

5.3 6 0.2 (24)a

3.7

89.2 6 1.7 (8)b

3.9 6 0.2 (8)b

3–51*

3.9 6 0.2 (8)b,c

3.9

90.5 6 1.4 (10)b

2.9 6 0.2 (19)c

2–41.2†

4.6 6 0.7 (12)a,c

3.5

† Calculated using two years (1994–1995 in ON, 2000–2001 in SC) of x-ray data.* Data reported by Ernst (1970) include only nesting females, therefore we estimated clutch frequency to be 1.# Calculated by multiplying clutch frequency by clutch size; therefore, this value accounts for sexually mature females that

did not become gravid.Sources: ON (Litzgus and Brooks, 1998), PA (Ernst, 1970), SC (current study).

df 5 25, P , 0.0001, a 5 0.017) and SC (t 5 8.1,df 5 27, P , 0.0001). Body size did not differbetween gravid female turtles from PA and SC(t 5 0.6, df 5 16, P 5 0.6).

We compared the mean total number of eggsproduced by gravid females in a single repro-ductive season among the three populations(ON, PA, and SC; Table 2). The total number ofeggs produced annually differed marginally(F2,43 5 3.25, P 5 0.049). Posthoc comparisonsindicated that most of the variation in this over-all population effect is attributable to the differ-ence in total annual egg production betweenON and PA turtles. Bonferroni a posteriori tests(a 5 0.017) indicated that females from the ON(N 5 24) population produce significantly moreeggs per year (mean 6 SE) than females in thePA (N 5 8) population (5.3 6 0.2 and 3.9 6 0.2,respectively, t 5 4.74, df 5 18, P 5 0.00016). Thenumber of eggs produced annually did not dif-fer between the ON and SC (N 5 12) popula-tions (5.3 6 0.2 and 4.6 6 0.7, respectively, t 51.02, df 5 13, P 5 0.33), nor did it differ be-tween the PA and SC populations (t 5 1.00, df5 13, P 5 0.33). We also calculated the totalnumber of eggs produced in each population bymultiplying clutch frequency by clutch size. Thisvalue accounts for sexually mature females thatdid not become gravid and is therefore a con-servative estimate of population-level reproduc-tive output. Because there is no error term as-sociated with this calculation, we could not sta-tistically compare among the populations; how-ever, a qualitative comparison indicates that thevalues for total annual egg production are sim-ilar (last row of Table 2).

DISCUSSION

This paper presents the first evidence for theproduction of multiple clutches of eggs by wild

spotted turtles. It is not too surprising that spot-ted turtles in SC are capable of producing mul-tiple clutches in a single reproductive season be-cause the duration of warm temperatures andthe length of the growing season (61 months:April to September; pers. obs.) at this southernlocale likely gives ample time for maturation ofa second (or third) compliment of follicles. Incontrast, summers are relatively short (3months: June to August) at the northern extremeof the spotted turtle’s range (Litzgus andBrooks, 2000), where a maximum of one clutchof eggs was produced per year (Litzgus andBrooks, 1998).

In some reptiles, it has been found that largerfemales tend to produce a first clutch of eggsearlier in the season than smaller females, whichthen allows larger females to produce secondclutches (Wallis et al., 1999; Castilla and Bau-wens, 2000). We did not observe such a trend inour spotted turtles in SC: large females did notnecessarily oviposit earlier in the season thansmall females, and both large and small femaleswere gravid more than once. Iverson and Smith(1993) found that, in painted turtles, later clutch-es tended to contain fewer eggs than earlierclutches. In contrast, clutch size did not decreasewith clutch number in SC spotted turtles. Someturtles produced the same number of eggs intheir first and second clutches, and other fe-males produced a greater number of eggs intheir second clutches.

Clutch frequency is predicted to increase withbody size in turtles (Moll, 1979), and indeedstudies have found a significant positive corre-lation between body size and clutch frequency(e.g., Zuffi and Odetti, 1998; Wallis et al., 1999).In the ON spotted turtle population, femalesthat were gravid three years in a row were sig-


nificantly larger than females that never becamegravid and were also larger than females gravidin only one or two of the three years (Litzgusand Brooks, 1998). Within a population, it ispossible that larger females are more likely toproduce multiple clutches of eggs than smallerfemales because larger body size allows for arelatively greater accumulation of energy storesthat can be allocated to egg production. Thus,larger size may allow females to maintain moreregular reproductive cycles (Litzgus andBrooks, 1998). In addition, larger females maybe older and may therefore take greater risks toreproduce (Williams, 1966).

We did not find a significant relationship be-tween body size and clutch frequency in the SCpopulation of spotted turtles. Other studies onchelonians have found no relationship betweenbody size and clutch frequency (e.g., Johnsonand Ehrhart, 1996; Mueller et al., 1998; Chenand Lue, 1999). The production of multipleclutches within a year may be independent offemale body size but may instead be variable ona year-to-year basis based on female foragingsuccess and concomitant energy acquisition andstorage (e.g., Bonnet et al., 2001; Congdon andTinkle, 1982; Gibbons, 1982; Castilla et al., 1992;Ritke and Lessman, 1994). Given that in the SCspotted turtle population, both large and smallfemales were gravid more than once, that largefemales did not necessarily oviposit earlier inthe season than small females, and that clutchsize did not decrease with clutch number, itseems likely that the ability of female spottedturtles in SC to produce multiple clutches is de-pendent upon available resources and not bodysize. Thus, our data suggest that there may beplasticity in clutch frequency. If so, one wouldpredict that, if a northern female turtle experi-enced an earlier spring and/or a longer, moreproductive summer than usual, she would pro-duce two or more clutches of eggs in the follow-ing reproductive season. Future work should ex-amine the potential plasticity of clutch frequen-cy in northern turtle populations, especially inlight of apparent global warming trends.

Individual clutch sizes were bigger in ON tur-tles compared to PA and SC turtles. Increasesin clutch size with increasing latitude have alsobeen observed in other turtle species (e.g., Tin-kle, 1961; Moll, 1979; Congdon and Gibbons,1985; Iverson, 1992; Iverson and Smith, 1993;Iverson et al., 1993, 1997). Clutch size is relatedto body size in turtles, and larger turtles tendto occur at more northern latitudes (Iverson,1992). Differences in individual clutch sizes be-tween northern and southern female spottedturtles may be the result of different selectivepressures acting on body size and can thereforebe discussed in terms of a trade-off between so-

matic growth and reproductive output. Becauseturtles have a hard shell, body size limits thenumber of eggs that can be carried at one time,especially in a small turtle species like the spot-ted turtle. There is likely strong selective pres-sure for large female body size (delayed matu-rity) in the north (Lindsey, 1966; Murphy, 1985;Galbraith et al., 1989) where the short summerlimits the potential for successful incubation ofturtle eggs before the onset of winter, so femalesmaximize clutch size (Tinkle, 1961; Wilbur andMorin, 1988; Willemsen and Hailey, 1999) in theone chance they get to reproduce per year. Inaddition, large body size in the north may beadvantageous for successful overwintering dur-ing the long, cold winter (Murphy, 1985; Gal-braith et al., 1989; Brooks et al., 1992). Southernfemales do not experience the same limitationsof a short growing season. As a result, the cost,in terms of overwinter survival, of early repro-duction in the south may be less than that in thenorth. Therefore, southern females mature atsmaller body sizes, which in turn results in thepartitioning of reproductive output in the formof multiple clutches throughout the growingseason.

Natural selection acts on life-time reproduc-tive success; the product of survival and annualreproductive output. For long-lived species,such as chelonians, only a snapshot view of re-productive success can usually be obtainedwithin the time-frame of a funded research pro-gram. In studies of latitudinal variation in re-productive output, individual clutch sizes aretypically compared. However, total annual eggproduction (whether produced in a single clutchor partitioned across multiple reproductiveevents) may be a more appropriate measure offemale fecundity and, thus, should be the focusof analyses of the adaptive significance of lati-tudinal variation in reproductive output amongconspecific populations. Our data indicate thatfemale spotted turtles in the southern popula-tion produce, on average, the same number ofeggs as more northern females but that somesouthern females partition their annual repro-ductive output across multiple reproductiveevents (i.e., northern females ‘‘put all their eggsin one basket,’’ whereas southern females par-tition their eggs across multiple baskets). Parti-tioning reproductive output across multipleclutches may be advantageous in predatoravoidance. Once a clutch of eggs is discovered,regardless of size, a predator usually consumesthe entire clutch (Obbard and Brooks, 1981; J. D.Litzgus, pers. obs.). Thus, the bet-hedging strat-egy of producing multiple clutches across timeand space of some southern female turtlesseems better for increasing lifetime reproductivesuccess compared to the strategy employed by


northern females. However, options for northernfemales are limited; the summer is too short todo any more than produce one clutch of eggs,and most females reproduce only every otheryear (Litzgus and Brooks, 1998). Clearly, thesedifferences in reproductive ecology will influ-ence population persistence in the face of sud-den natural environmental changes and delete-rious human impacts.

Acknowledgments.—This research was supportedby grants from the National Geographic Society(grant 6630-99), the Santee Cooper power com-pany, SC, the Chelonian Research Foundation,and National Science Foundation (grant0090177). In kind, support was provided by theFrancis Beidler Forest National Audubon andNature Conservancy Sanctuary, Harleyville, SC.We especially thank N. Brunswig (SC State Di-rector, National Audubon Society) for his sup-port of this project. X-rays were conducted atthe Goose Creek Veterinary Clinic, GooseCreek, SC (2001 field season), and at Westbury’sVeterinary Clinic, Summerville, SC (2000 fieldseason). J. D. Congdon (SREL, Aiken, SC) pro-vided a radio receiver and 11 AVM radio trans-mitters. J. Kelling and B. G. Merritt donated bat-teries for transmitters. A. Dandridge, S. E.DuRant, W. J. Humphries, and T. J. S. Merrittassisted with field data collection. T. J. S. Merrittand W. E. Winterhalter provided comments onan earlier draft of the manuscript.

LITERATURE CITED

BARNWELL, M. E., P. A. MEYLAN, AND T. WALSH. 1997.The spotted turtle (Clemmys guttata) in central Flor-ida. Chelonian Conservation and Biology 2:405–408.

BONNET, X., G. NAULLEAU, R. SHINE, AND O. LOUR-DAIS. 2001. Short-term versus long-term effects offood intake on reproductive output in a viviparoussnake, Vipera aspis. Oikos 92:297–308.

BROOKS, R. J., C. M. SHILTON, G. P. BROWN, AND N. W.S. QUINN. 1992. Body size, distribution, and re-production in a northern population of wood tur-tles (Clemmys insculpta). Canadian Journal of Zo-ology 70:462–469.

BURKE, V. J., J. E. LOVICH, AND J. W. GIBBONS. 2000.Conservation of freshwater turtles. In M. K. Kle-mens (ed.), Turtle Conservation, pp. 156–179.Smithsonian Institution Press, Washington, DC.

CASTILLA, A. M., AND D. BAUWENS. 2000. Reproduc-tive characteristics of the lacertid lizard Podarcisatrata. Copeia 2000:748–756.

CASTILLA, A. M., L. J. BARBADILLO, AND D. BAUWENS.1992. Annual variation in reproductive traits in thelizard Acanthodactylus erythrurus. Canadian Journalof Zoology 70:395–402.

CHEN, T. H., AND K. Y. LUE. 1999. Population char-acteristics and egg production of the yellow-mar-gined box turtle, Cuora flavomarginata flavomargina-ta, in northern Taiwan. Herpetologica 55:487–498.

CONGDON, J. D., AND J. W. GIBBONS. 1985. Egg com-

ponents and reproductive characteristics of turtles:relationships to body size. Herpetologica 41:194–205.

CONGDON, J. D., AND D. W. TINKLE. 1982. Reproduc-tive energetics of the painted turtle (Chrysemys pic-ta). Herpetologica 38:228–237.

CUMMINS, C. P. 1986. Temporal and spatial variationin egg size and fecundity in Rana temporaria. Jour-nal of Animal Ecology 55:303–316.

ERNST, C. H. 1970. Reproduction in Clemmys guttata.Herpetologica 26:228–232.

. 1976. Ecology of the spotted turtle, Clemmysguttata (Reptilia, Testudines, Testudinidae), insoutheastern Pennsylvania. Journal of Herpetology10:25–33.

ERNST, C. H., AND G. R. ZUG. 1994. Observations onreproductive biology of the spotted turtle, Clemmysguttata in southeastern Pennsylvania. Journal ofHerpetology 28:99–102.

ERNST, C. H., J. E. LOVICH, AND R. W. BARBOUR. 1994.Turtles of the United States and Canada. Smith-sonian Institution Press, Washington, DC.

FITCH, H. S. 1985. Variation in clutch and litter sizein New World reptiles. University of Kansas Mu-seum of Natural History Miscellaneous Publication76:1–76.

FLEMING, I. A., AND M. R. GROSS. 1990. Latitudinalclines: a trade-off between egg number and size inPacific salmon. Ecology 71:1–11.

GALBRAITH, D. A., R. J. BROOKS, M. E. AND OBBARD.1989. The influence of growth rate on age andbody size at maturity in female snapping turtles(Chelydra serpentina). Copeia 1989:896–904.

GIBBONS, J. W. 1982. Reproductive patterns in fresh-water turtles. Herpetologica 38:222–227.

GIBBONS, J. W., AND J. L. GREENE. 1979. X-ray photog-raphy: a technique to determine reproductive pat-terns of freshwater turtles. Herpetologica 35:86–89.

GODFRAY, H. C. J., L. PARTRIDGE, AND P. H. HARVEY.1991. Clutch size. Annual Review of Ecology andSystematics 22:409–429.

HEALEY, M. C., AND W. R. HEARD. 1984. Inter- andintra-population variation in the fecundity of chi-nook salmon (Oncorhynchus tshawytscha) in its rel-evance to life history theory. Canadian Journal ofFisheries and Aquatic Science 41:476–483.

IVERSON, J. B. 1992. Correlates of reproductive outputin turtles (order Testudines). Herpetological Mono-graphs 6:25–42.

IVERSON, J. B., AND G. R. SMITH. 1993. Reproductiveecology of the painted turtle (Chrysemys picta) inthe Nebraska Sandhills and across its range. Cop-eia 1993:1–24.

IVERSON, J. B., C. P. BALGOOYEN, K. K. BYRD, AND K.K. LYDDAN. 1993. Latitudinal variation in egg andclutch size in turtles. Canadian Journal of Zoology71:2448–2461.

IVERSON, J. B., H. HIGGINS, A. SIRULNIK, AND C. GRIF-FITHS. 1997. Local and geographic variation in thereproductive biology of the snapping turtle (Che-lydra serpentina). Herpetologica 53:96–117.

JACKSON, D. R. 1988. Reproductive strategies of sym-patric freshwater emydid turtles in northern pen-insular Florida. Bulletin of the Florida State Mu-seum of Biological Sciences 33:113–158.

JOHNSON, S. A., AND L. M. EHRHART. 1996. Repro-


ductive ecology of the Florida green turtle: clutchfrequency. Journal of Herpetology 30:407–410.

LINDSEY, C. C. 1966. Body sizes of poikilotherm ver-tebrates at different latitudes. Evolution 20:456–465.

LITZGUS, J. D., AND R. J. BROOKS. 1998. Reproductionin a northern population of Clemmys guttata. Jour-nal of Herpetology 32:252–259.

. 2000. Habitat and temperature selection ofClemmys guttata from a northern population. Jour-nal of Herpetology 34:178–185.

LORD, R. D. 1960. Litter size and latitude in NorthAmerican mammals. American Midland Natural-ist 64:488–499.

LOVICH, J. E., AND T. R. JAWORSKI. 1988. Annotatedcheck list of amphibians and reptiles reportedfrom Cedar Bog, Ohio. Ohio Journal of Science 88:139–143.

MAY, R. M., AND D. I. RUBENSTEIN. 1985. Reproduc-tive strategies. In C. R. Austin and R. V. Short(eds.), Reproduction in Mammals, IV, Reproduc-tive fitness, pp. 1–23. Cambridge Univ. Press, Cam-bridge.

MOLL, E. O. 1973. Latitudinal and intersubspecificvariation in reproduction of the painted turtle,Chrysemys picta. Herpetologica 29:307–318.

. 1979. Reproductive cycles and adaptations. InM. Harless and H. Morlock (eds.), Turtles: Per-spectives and Research, pp. 305–331. John Wileyand Sons, New York.

MUELLER, J. M., K. R. SHARP, K. K. ZANDER, D. L. RAK-ESTRAW, K. R. RAUTENSTRAUCH, AND P. E. LEDERLE.1998. Size-specific fecundity of the desert tortoise(Gopherus agassizii). Journal of Herpetology 32:313–319.

MURPHY, E. C. 1985. Bergmann’s Rule, seasonality,and geographic variation in body size of housesparrows. Evolution 39:1327–1334.

OBBARD, M. E., AND R. J. BROOKS. 1981. Fate of over-wintering clutches of the common snapping turtle(Chelydra serpentina) in Algonquin Park, Ontario.Canadian Field-Naturalist 95:350–352.

PORCHER, R. D. 1981. The vascular flora of the FrancisBeidler Forest in Four Holes Swamp, Berkeley andDorchester Counties, South Carolina. Castanea 46:248–280.

RICKLEFS, R. E. 1980. Geographical variation in clutchsize among passerine birds: Ashmole’s hypothesis.Auk 97:38–49.

RITKE, M. E., AND C. A. LESSMAN. 1994. Longitudinalstudy of ovarian dynamics in female gray tree-frogs (Hyla chrysoscelis). Copeia 1994:1014–1022.

SINERVO, B. 1990. The evolution of maternal invest-ment in lizards: an experimental and comparativeanalysis of egg size and its effects on offspring per-formance. Evolution 44:279–294.

TINKLE, D. W. 1961. Geographic variation in repro-duction, size, sex ratio and maturity of Sternotherusodoratus. Ecology 42:68–76.

WALLIS, I. R., B. T. HENEN, AND K. A. NAGY. 1999.Egg size and annual egg production by female de-sert tortoises (Gopherus agassizii): the importance offood abundance, body size, and date of egg shell-ing. Journal of Herpetology 33:394–408.

WILBUR, H. M., AND P. J. MORIN. 1988. Life historyevolution in turtles. In C. Gans and R. B. Huey(eds.), Biology of the Reptilia. Vol. 16. Ecology B.Defense and Life History, pp. 387–437. Alan R.Liss, Inc., New York.

WILLEMSEN, R. E., AND A. HAILEY. 1999. Variation ofadult body size of the tortoise Testudo hermanni inGreece: proximate and ultimate causes. Journal ofZoology 248:379–396.

WILLIAMS, G. C. 1966. Adaptation and Natural Selec-tion: A Critique of Some Current EvolutionaryThought. Princeton Univ. Press, Princeton, NJ.

WILSON, P. R. 1989. Clemmys guttata (spotted turtle)reproduction. Herpetological Review 20:69–70.

ZUFFI, M. A. L., AND F. ODETTI. 1998. Double egg-deposition in the European pond turtle, Emy orbi-cularis, from central Italy. Italian Journal of Zool-ogy 65:187–189.


Documents

Reproductive Ecology of Sceloporus utiformis (Sauria: … · 2008-06-09 · Reproductive Ecology of Sceloporus utiformis (Sauria: Phrynosomatidae) from a Tropical Dry Forest of Me´xico