Adaptive Male Parental Care in the Giant Bullfrog, Pyxicephalus adspersus

Abstract

-Male giant bullfrogs Pyxicephalus adspersus exhibit paternal care through construction of channels that guide tadpoles to larger bodies of water. We found the following. These channels (which may exceed 15 m in length) make available cooler water to broods in the rapidly drying breeding puddles that frequently reach critically high temperatures. Eggs of P. adspersus exposed to temperatures above 38?C died, a temperature reached in a third of the puddles measured. In contrast, tadpoles readily survived temperatures above 38?C, and experiments showed that tadpoles can survive the highest temperatures recorded at the puddles. Males actively defended their offspring against predators and were sometimes killed while performing this behavior. Survival of eggs and larvae in territorial broods was roughly twice that of broods in nonterritorial breeding arenas. The large cost of paternal care in the giant bullfrog is therefore offset by strong fitness gains. We suggest that channel construction and predator defence, crucial for tadpole survival, can be performed most efficiently by large-bodied parents, explaining why males (not the smaller females) perform parental care. Parental care has been reported in a number of frogs (Townsend, 1986; Summers, 1989; Kok et al., 1989), most of these from terrestrial breeders in the tropics (Wells, 1981). Either males or females can perform parental care. Paternal care is more likely in frogs than in other vertebrates because the cost of paternal care is usually low (Wells, 1981). This is exemplified in territorial species where eggs or tadpoles are deposited in a male's territory and where he cares for them while continuing to attract females (but see Townsend, 1986). Parental care may include 1 Corresponding Author: E-mail: jwhferguson@ zoology.up.ac.za. guarding of eggs or tadpoles against predation (Kluge, 1981), preventing desiccation (Townsend et al., 1984), maintaining access to water (Kok et al., 1989), and carrying tadpoles, either to water (Summers, 1990) or until development is completed (Wells, 1981). Many anurans breed in ephemeral puddles (Wells, 1977; Sullivan, 1989). The hypothesized advantages of this habit is the absence of resident predators and reduction of competition for the tadpoles compared to those in permanent bodies of water (Wassersug, 1975), as well as a rapid rate of development (Heatwole et al., 1968; but for a counterargument, see Abe and Neto, 1991). Giant bullfrogs, Pyxicephalus adspersus (adult 310 This content downloaded from 157.55.39.124 on Wed, 20 Jul 2016 04:21:13 UTC All use subject to http://about.jstor.org/terms PATERNAL CARE IN GIANT BULLFROGS mass 160-970 g), breed in temporary ponds following the first heavy summer rains. These frogs are rare and breed irregularly (on average in the order of once in four years at a particular site) and little is known of their breeding biology. Oviposition typically occurs immediately after heavy rain, either in a male's territory or communally in a nonterritorial breeding aggregation (unpubl. data). Hatching occurs 48 h following oviposition, and active schooling and feeding has been observed 24 h later (Balinsky and Balinsky, 1954). The territorial puddles often are exposed to high temperatures, which could result in the death of eggs and tadpoles. Adult males construct channels that tadpoles use to move from small isolated peripheral puddles to larger bodies of water Kok et al. (1989). We suggest that, in addition to this role, the channels also serve to make cooler water accessible to the broods in warm, rapidly drying breeding puddles. We studied the effects of channel construction on offspring survivorship. The questions we asked were (1) What are the water temperatures in channels compared to those in puddles? (2) What is the thermal tolerance of bullfrog tadpoles? (3) What effect does paternal care have on survival of bullfrog broods? MATERIALS AND METHODS Study Area.-We studied bullfrogs during two breeding seasons (1992-1994) at Glen Austin Bird Sanctuary Midrand (25?58'S; 28?10'E) South Africa (elevation 1500 m). The site contains a temporary lake (called a pan) with surface area of 9200 m2, which fills to a maximum depth of 30 cm after heavy summer rains (October-January). The area around the pan comprises dry grassland (Bankenveld, see Acocks, 1988) with sedges (Scirpus spp.) and cat-tails (Typha capensis) along the edge of the pan. A second study site at Witpoort comprised a small human-made dam with a relatively constant water level, containing at least 80 m of edge of which at least 60 m were suitable for bullfrog breeding. Behavioral Observations and Removal Experiment.-Following the appearance of frogs after the first summer rains, we made daily observations from 0700-1800 h throughout the breeding season (approximately 30 days) with additional nocturnal observations on 11 days while channels were constructed. Frogs were sexed based on size and coloration (Passmore and Carruthers, 1995). Immatures had green and black colors, whereas adults were green with pale yellow markings. Adults were dimorphic in color and snout-urostyle length. Males have yellow-orange axillary regions and measured larger than 13 cm (determined from pairs in amplexus, as well as from 18 dissected specimens, 12 males, four females, which had been killed by predators or by motor cars). Males were classified as territorial or nonterritorial, based on the presence or absence of behavior such as site fidelity and male-male combat. Males were individually branded using a portable branding iron (Jennions, 1992). Brands were treated with antibiotic cream. Frogs were released within 5 min of branding, and they immediately resumed breeding behavior. Territories of males were marked using a range of unobtrusive numbered markers. If channels were constructed between a territorial puddle and the pan, its width, minimum depth, and length were measured and the date and time of construction noted. Positions of males as well as movements of tadpoles were recorded after the construction of a channel. Clutch sizes were determined by direct enumeration or from color photographs (Cooke, 1983). Clutches were classified as surviving if any of the tadpoles entered metamorphosis. Mortality of clutches by dessication, predation, or other causes was recorded. We performed a removal experiment to determine the effects of a territorial male on desiccation and predation of tadpoles. Males were removed from 10 randomly determined territories with similar-sized two-day-old clutches (16002400 eggs) and within 5 m of the pan. Fifteen similar territories situated within 20 m of territories of removed males served as controls for the experiment. To ensure comparable brood age and size as well as vegetation cover in experimental and control territories, these were distributed as duos (removal + control) or trios (removal + two controls) throughout the part of the pan used for breeding. Predators observed to prey on bullfrogs or their broods were identified to as low a taxonomic level as possible. Predation also was inferred if an entire clutch disappeared, with marks remaining in the mud adjacent to the breeding puddle. Puddle Characteristics.-The water temperatures of 48 puddles with channels were recorded using a Sensortek BAT-12 digital thermometer, accurate to 0.1?C. To determine the daily temperature variation in the puddles, temperature measurements were made at hourly intervals from 0700-1800 h, 1 cm below the water surface in the center of the aggregate of tadpoles. Body temperatures of the tadpoles in the field were assumed to be the same as those of the surrounding water. Maximum depths of these puddles were measured using a plastic ruler. Temperature Tolerance Experiment.-Six hundred tadpoles from 18 breeding puddles were kept in aquaria at 25-28?C for 24 h. Each tadpole was used only once, with 30 randomly chosen 311 This content downloaded from 157.55.39.124 on Wed, 20 Jul 2016 04:21:13 UTC All use subject to http://about.jstor.org/terms