Various studies have suggested that female anurans select oviposition sites based on factors such as water depth (Seale, 1982; Caldwell, 1986), water temperature (Howard, 1978), or absence of predators (Howard, 1978; Resetarits and Wilbur, 1989). Selection pressure should be high for females that can assess environmental factors leading to increased survival of offspring. Crump (1991) found that Hyla pseudopuma females chose oviposition sites based on assessments of egg loads already present. When females oviposited at a site occupied by tadpoles, eggs were cannibalized by more developed, conspecific tadpoles. Using the leptodactylid frog, Pleurodema borellii, we further test the hypothesis that females assess the presence of tadpoles in a pond before ovipositing. This assessment by females should be important in terms of maternal fitness. Alford and Crump (1982) and Petranka (1989), for instance, found that at higher densities larval development took longer and tadpoles metamorphosed at smaller sizes, thus diminishing their chances of survival.

Pleurodema borellii lives in subtropical forests and shrubby dry chaco areas of northwestern Argentina (Cei, 1980). Adults measure 40–55 mm SVL, and males are smaller than females. Breeding occurs throughout the summer, from mid-August until March. As females oviposit in open standing water, males use their hind legs to whip up a foam nest that protects the eggs from desiccation.

Materials and Methods

Oviposition site selection was investigated in P. borellii during summer 1996–1997. We studied a free-ranging population at two backyard pools, 15 m apart, located on the outskirts of the city of San Miguel de Tucumán, Province of Tucumán, in northwestern Argentina. One of them (the “small” pool, or SP) was kidney shaped and measured about 2 × 3 × 0.75 m (3.5 m³ of water). The other pool (the “big” pool, or BP) had a rectangular shape and measured 8 × 4 × 1.5 m (48 m³). Although the pools were different in size, we considered the study appropriate because our aim had not been to create a controlled situation (i.e., use same pool sizes) but to use pools selected by the frog. However, when comparing the two pools, differences in sizes and volumes were considered.

The pools were filled with rain water and fallen leaves and branches all year long except for BP that was drained that summer and filled with clear chlorinated water during two and a half months (December, January, and two weeks in March) for human use. Each night, from mid-August to the end of March, we checked number of frogs present at each pool, number of calling males, amplecting pairs, and number of nests that were produced. We also monitored nest development and hatching time of larvae. We counted hatching larvae from 11 foam nests. We checked for metamorphosing tadpoles (“froglets” still with a tail stub, approximately stage 44 of Gosner, 1960) as they came out on the side of SP. We were unable to check for metamorphosing froglets at BP because of the draining of that pool.

We conducted two experiments: (1) foam-nest-transfer experiment; and (2) tadpole-density experiment. In the foam-nest-transfer experiment, we determined the fate of recently oviposited foam nests in the presence of more advanced conspecific tadpoles (see further for size), by taking 10 foam nests from BP (which, by visual inspection, showed a low density of tadpoles) and placing them on the surface of SP (which showed a high density of tadpoles, meaning tadpoles were abundant and collided into each other continuously). The tadpoles from SP were of about 10–15 mm snout–vent length with hind legs visible (stages 33 to 37, Gosner, 1960). Foam nests were transferred on seven different evenings (in Feb.) and left until the next morning. The nests were observed hourly for a five-hour period and then checked again the following morning.

In the tadpole-density experiment, we studied effects of density and presence of large tadpoles on the development of the small ones. We used 12 green plastic bowls (30 cm diameter). Each one was identified with a number and filled with 1.5 liters of water from an outdoor, rain-fed, artificial pond. Each day, water level was checked for evaporation, and water was added accordingly. Every two to three days, about a third of the water was replaced with fresh water from the outdoor pond. Tadpoles were fed boiled lettuce ad libitum.

We divided the bowls into four treatments, each replicated three times. The small tadpoles used in this experiment were siblings all from the same foam nest, and they were four days postoviposition (about 4 mm snout–vent length, stages 24 to 25, Gosner, 1960). We counted days postoviposition instead of posthatching because larvae hatched one to three days postoviposition, and we did not know hatching time of individual larvae. The large tadpoles that we used in this experiment were older conspecific tadpoles from different foam nests (size range 12 to 15 mm snout–vent length, hind legs fairly well developed, stages 35 to 37, Gosner, 1960). We separated the bowls into the following treatments. Control treatment—we placed 10, four-day-old tadpoles in each of three bowls [we selected 10 tadpoles per 1.5 liters of water as our baseline following Lavilla and Rougés (1992), who kept 10 tadpoles of this species per liter of water, the tadpoles reaching metamorphosis in 29 days]. Large-tadpoles-present treatment—we placed 10, four-day-old tadpoles in each of three bowls with five large conspecific tadpoles. Density treatment—we placed 20, four-day-old tadpoles in each of three bowls. Density plus large-tadpoles-present treatment—we placed 20, four-day-old tadpoles in each of three bowls with five large conspecific tadpoles. (Although variation in total biomass resulting from the large tadpoles may be a factor in this treatment and in the large-tadpoles-present treatment, we did not consider it here because we were interested in reporting whether large tadpoles preyed on small ones and on inhibition of growth resulting from high density in number of tadpoles present. However, biomass may be an important factor and needs to be considered in a future study.)

Once the experiment began, we checked bowls every two days. We recorded number of tadpoles present, size, and developmental stage. The experiment was terminated when the first of the small tadpoles reached stage 42 (Gosner, 1960), that is, front limbs had emerged. This occurred on the 19th day after starting the experiment or 23 days postoviposition. At that time, the small tadpoles that remained in each bowl were counted and a χ² goodness-of-fit test, combining replicates, was performed. The remaining small tadpoles were also weighed collectively to the nearest 0.01 mg, and a mean was calculated for each bowl. Tadpoles were blotted with a piece of cloth before weighing. We performed a Friedman Two-Way Analysis of Variance by Ranks to compare among treatments and a multiple comparisons between groups or conditions to compare between treatments (Siegel and Castellan, 1988).

Results

Toward the end of the summer, a pattern of breeding activity in P. borellii emerged, which seemed to depend on the abundancy of tadpoles present in the pools. Breeding activity started at SP in mid-August (end of winter) followed closely with activity at BP. In September, breeding activity dropped to zero in SP and increased considerably in BP (Table 1). During the following months, no breeding activity was observed at SP except for a brief come-back at the end of the summer (eight nests that disappeared by the next morning, see results on foam-nest transfer). Meanwhile, breeding activity continued at BP, except for the months when it was drained and observations were suspended.

Table 1. Number of Foam Nests of Pleurodema borellii in Two Pools during the Spring and Summer 1996 and 1997. NSP, number of nests in “small” pool; NBP, number of nests in “big” pool. The number of weeks observations were conducted is indicated in parentheses. When there is only one number, it is the same amount of time for both pools. When there are two numbers, the first corresponds to SP and the second to BP

View Fullscreen

The drop in activity at SP was associated with the increasing numbers of growing tadpoles in SP, whereas in BP, because of its large volume of water, there was no apparent immediate limit to the number of eggs that could be laid by breeding pairs. In fact, there were, on average, about six times more nests per week in BP than in SP, BP being 5.3 times larger in surface than SP and 13 times larger in volume. Each foam nest had about 1300 hatching tadpoles (for n = 11, x̄ = 1276.55, SD = 351.94). A total of 388 froglets emerged from SP for the whole breeding season.

Tadpole-density experiment

With respect to number of tadpoles remaining at the end of the experiment, the greatest impact was suffered by small tadpoles in the large-tadpoles-present and the density plus large-tadpoles-present treatments, the treatments that included large tadpoles (Table 2: χ² = 21.24, df = 3, P < 0.001). In these treatments, several small tadpoles were eaten by conspecific large tadpoles, especially during the first few days of the experiment.

Table 2. The Effect of Density and/or Presence of Larger Tadpoles on Recently Hatched Sibling Tadpoles of Pleurodema borellii during 19 Days of Observations. Treatments were: I (10 small tadpoles per replicate); II (10 small tadpoles + 5 large tadpoles per replicate); III (20 small tadpoles per replicate); IV (20 small tadpoles + 5 large tadpoles per replicate). N is the number of small tadpoles left in each bowl at the end of the experiment (23 days postoviposition) with % given in parentheses. Mass is the average mass in grams of a small tadpole at the end of the experiment for each replicate. See Methods for definitions of small and large tadpoles

View Fullscreen

Large tadpoles significantly affected the mass of small tadpoles (Table 2; Friedman two-way ANOVA, F_r = 8.2, n = 3, k = 4, P < 0.05). Based on a multiple comparison test, there was a significant difference (P = 0.05) between the control treatment and the combined density with large tadpoles (the density plus large-tadpoles-present treatment). The control treatment and the large-tadpoles-present treatment did not differ (P = 0.07). Consequently, the combination of density and large tadpoles had the major impact on survival of small tadpoles.

Discussion

Our observations showed that frogs did not necessarily oviposit even when offered essential characteristics for anuran breeding, such as standing water, abundant vegetation, and favorable climatic conditions. We suspect that tadpole presence may have been an important factor in the selection of oviposition sites. As this study and others have shown (e.g., Licht, 1967; Wilbur, 1976; Alford, 1989), high tadpole densities affect their development and their chances of survival. In addition, our study showed that more developed tadpoles cannibalize smaller conspecific tadpoles. It is not clear how females detect the presence of tadpoles at a particular site, but they seem to avoid such sites in spite of male choruses. Crump (1991) found that larger tadpoles of Hyla pseudopuma ate conspecific eggs which might explain why females did not oviposit at locations where there were already many eggs. Our observations of cannibalism among conspecific tadpole P. borellii and on growth inhibition favor the hypothesis that female P. borellii may select an oviposition site depending on the presence or absence of tadpoles.

Future studies should explore mechanisms, such as chemical and mechanical, that might be involved in females'choices of oviposition sites based on tadpole presence. The mechanisms affecting tadpole growth also need further investigation, including those of biomass effect, chemical cues, and/or mechanical cues.