Herpetological Review Herpetological Review - Doczine

ARTICLESHerpetological Review, 2008, 39(2), 151–154.© 2008 by Society for the Study of Amphibians and ReptilesOophagy and Larval Cannibalism withoutPolyphenism in Tadpoles of the Great BasinSpadefoot (Spea intermontana)SUE FOXP.O. Box 68, Cedarville, California 96104, USAe-mail: suefox@hughes.netPolyphenism associated with cannibalism has been reliably documentedin two species of North American spadefoot toads (familyPelobatidae: Spea bombifrons and S. multiplicata) (Bragg 1956,1964; Bragg and Bragg 1959; Pfennig 1989; Pomeroy 1981). Spealarvae occur as two morphologically distinct phenotypes: 1) carnivoresare cannibalistic, have beak-shaped, keratinized mouthparts,and hypertrophied jaw musculature; and 2) omnivores haveflat, keratinized mouthparts, and feed primarily on detritus (Bragg1965; Pfennig 1992; Pomeroy 1981). Pomeroy (1981) noted polymorphismin pools containing S. multiplicata 2–4 days after feedingcommenced. In S. multiplicata, consumption of anostracanshrimps or other tadpoles induces the carnivore morphology andmorph determination is reversible based on diet (Pfennig 1990;Pomeroy 1981).Reports of polyphenism in Spea intermontana have been limitedto mouthpart characteristics (Acker and Larsen 1979; Black1973; Orton 1954; Turner 1952). The description provided by Tanner(1939) of the mouthparts of S. intermontana collected in Utahhas been interpreted as indicative of a carnivore morph (Hall et al.2002; Pfennig 1992). Subsequent workers have assumed that larvaeof S. intermontana are potentially morphologically variable(Hall 1993; Hall et al. 1997; Hall et al. 2002). Hall (1998) includestwo photographs of the carnivorous and herbivorous morphologicaltypes. However, the two types have not been describedin detail and there are no reports whether the carnivorous morphis cannibalistic. I report field observations of cannibalism andoophagy by larvae of S. intermontana and the lack of polyphenismassociated with cannibalism.Methods.—I conducted this study at permanent, semi-permanentand temporary ponds in Mono County, California, USA(118.965°N, 38.086°W), during the breeding seasons April throughJune of 1984–1989. Permanent and semi-permanent pools are createdby artesian wells. Semi-permanent pools are present yearround,except in drought years. Temporary pools formed fromground water and surface run-off, and were present only duringwet years. Study pools varied in size from less than one metersquare to several hectares. Additional observations and collectionswere made from three small pools formed from springs near thesouth shore of Mono Lake, Mono County, California (119.053°W,37.940°N).Tadpoles were observed in the field for 1380 h on 260 daysbetween April 1984 and May 1989. The duration of observationsranged from 3–7 h per day. I identified a tadpole as a cannibal if itwas observed eating all or part of a conspecific, while a non-cannibaldid not eat a conspecific. For cannibalistic encounters forwhich an entire sequence was observed I recorded: the length oftime it took a cannibal to consume its prey; the number of cannibalsfeeding on a single tadpole; and the snout–vent length (SVL;tip of the snout to the junction of the posterior body wall and cloaca)and developmental stage (DS, Gosner 1960) of the cannibal(s)and prey.Tadpoles were collected for morphometric analyses mid-Aprilto mid-June in 1984 through 1987 and during May in 1988 and1989. Tadpoles were randomly collected with a dip net once aweek from the main study pool and less frequently from otherpools that contained fewer tadpoles. While phenotypic differencesbetween morphs of S. bombifrons and S. multiplicata are readilydetermined by visual inspection (Bragg 1965; Pfennig 1990;Pomeroy 1981; Storz 2004), there were no obvious morphologicaldifferences among tadpoles of S. intermontana in my studypools. For more detailed morphometric comparisons, cannibalsand noncannibals were identified by offering field-collected tadpoles(stages 25–36) at least one pre-feeding stage conspecific for24 h. Tadpoles were housed individually in 2.4-liter round plasticcontainers (16.8 cm x 12.5 cm) filled with approximately 1680 mlwater to a depth of 8.75 cm. Cannibalism was inferred if a tadpolewas missing or its partially consumed remains were present. Tadpolesthat did not eat conspecifics were labeled non-cannibals.Cannibals (N = 34) and non-cannibals (N = 36) were preservedfor morphometric analysis.All tadpoles were cold-killed and preserved in 10% formalin. Iexamined the external morphology of 1089 tadpoles of differentsizes and developmental stages. Using dial calipers and a dissectingmicroscope, three characteristics were measured: SVL, totallength (TL; tip of the snout to the tip of the tail), DS, and numberof posterior and anterior labial teeth rows (PLT and ALT) usingAltig and McDiarmid’s (1999) terminology. The criterion for labialtooth row presence was at least three teeth on a tooth ridge. Aqualitative description of the keratinized jaw sheaths also was recorded(e.g., serrations on jaw sheaths, thick, thin; see Altig andMcDiarmid 1999).Gut length (GL) and musculus orbitohyoideus length (OH) weremeasured for the 70 experimental tadpoles identified as cannibalsand non-cannibals and for 157 tadpoles collected from a singlepopulation over the course of their development. These two traitsare diagnostic of the carnivore morphotype for S. multiplicata(Pfennig 1989; Pomeroy 1981). A dissecting microscope with ocularmicrometer was used to measure OH to the nearest 0.1 mm.The relationships of GL and OH to body size were analyzed usinganalysis of covariance.I tested whether metamorphs were cannibalized by conspecificlarvae by placing metamorphs in the water of ponds 1–2 m fromshore. This forced the individual to swim to shore above feedingaggregations of tadpoles. I conducted 35 trials with individualmetamorphs in 1984 and 45 metamorphs in 1986.Results.—Cannibalism was observed on 8 of 260 days of fieldobservations: 6 days in 1984 (April 15–18, 21, 22) and 2 days in1987 (May 8, 15). A total of 41 occurrences of cannibalism wereobserved in the 8 days over a period of 27.5 h, in three differentponds. Based on the number of egg clutches recorded in the threeponds and the average number of eggs per clutch (mean = 812,SD = 297), the number of tadpoles present on each day the cannibalismwas recorded ranged from 4872 to more than 10,000 tad-Herpetological Review 39(2), 2008 151