| Summary: | A comparative analysis has been made of the feeding behaviour and population structure of the Marion Lake, B.C., populations of the salamanders Taricha granulosa (Skilton) and neotenous Ambystoma gracile (Baird). The aims of the study were to provide information relevant to the population dynamics of the two salamander populations and to construct a computer simulation model of salamander feeding. During 1968 and 1969 populations of Taricha and Ambystoma were trapped and observed in Marion Lake. The adult T. granulosa and A. gracile populations were estimated respectively at 2449 and 14,500. It was estimated that there were about 45,500 A. gracile young of the second year in the lake. These populations represent total lake biomasses of about 18 kgm. for Taricha and 409 kgm. for Ambystoma. The numbers of young larvae were not determined for either species. The A. gracile population was dispersed and constant through the spring and summer. Density and size composition varied between areas with different vegetation and substratum. The T. granulosa distribution within the lake was highly contagious and changed through time. Although displaced animals returned to the areas where they had been originally caught the positions of their home ranges changed through time. T. granulosa entered the lake in April and May, showed evidence of "wandering" in mid-summer and left the lake in September and October. Fifty percent of adult A. gracile seen during the breeding season were metamorphosed and were found in the lake only during April and May. Neotenous adults remained in the lake all year, as did larvae of the first and second year. Those larvae which metamorphosed did so in late August and early September. Ambystoma gracile grew rapidly in the lake but no growth was seen in T. granulosa. No source of T. granulosa mortality was found in the lake but A. gracile were preyed on by fish and diving bugs and dead specimens were found on the lake bottom.
Taricha and Ambystoma were found to be nocturnal predators in Marion Lake. Taricha. Ambystoma feeding on the benthos and Ambystoma feeding in the water column had distinct types of feeding behaviour. Taricha were primarily benthic feeders, capable of utilizing scent to locate high densities of prey such as hatching tadpoles. There was no evidence that Ambystoma could do this. Amphipods and dipteran larvae accounted for 64% of the total stomach contents volume of Taricha and benthic feeding Ambystoma. The cladoceran Sida crystallina accounted for 63% of the food volume of water column feeding Ambystoma. The salamanders reacted with feeding behaviour when stimulated with visual, tactile or chemical stimuli. It was concluded that a combination of these stimuli were responded to by animals feeding in the lake.
Both species of salamander had a constant rate of digestion which was independent of the amount of food in the stomach. Male Taricha of 5.9 - 6.3 cm. body length digested horse heart at a rate of 0.27 cc/day. Female Taricha 5.2 – 5.6 cm. long and Ambystoma 6.1 - 7.7 cm. long had a digestion rate of 0.22 cc/day. Taricha were capable of eating six times the food volume that they could digest in one day and Ambystoma could eat five times the amount which they could digest in one day. The mean volume of stomach contents from lake animals was less than the amount which could be digested in one day. Taricha fed in the laboratory behaved in the same way. The animals were therefore hunting at a high level of hunger when they were maximally responsive to prey and had a short refractory period between prey ingestions.
Taricha were presented with 0.042 cc. pieces of horse heart at densities of 8, 30, 60, 100 and 200 /m². The mean daily ration and the time between ingestions were not significantly different between these densities when the salamanders were fed at a given density for six consecutive days. When Taricha were switched to a high prey density (l00/m²), after feeding at low prey density (8/m²) they responded by eating a large volume of food on the first day.
The results of the present study were combined with Ware's information on feeding by the rainbow trout (Salmo gairdneri Richardson) (in prep.) and a computer simulation model of predation was produced. The model was based on Holling’s "disc equation" (1959), which was expanded to include some effects of hunger, prey vulnerability and temperature. Predictions by the model were compared to the results of five experiments in which Taricha fed on tadpoles and pieces of horse heart. The observed and predicted proportions eaten differed by a mean value of 0.9%. When a prey type was present in high proportion and at high density the model tended to slightly underestimate numbers of the more numerous prey eaten and tended to slightly overestimate numbers of the less numerous prey eaten.
Predation by Taricha and Ambystoma in Marion Lake was simulated for the months of May, June, July and August. Data describing seven prey types were used in the model. In Marion Lake these prey accounted for a total of 79% of the volume of Taricha stomach contents and 75% of benthic feeding Ambystoma stomach contents. Predicted proportions of prey types eaten by Taricha were a mean of 5.3 percentage points in error and proportions predicted for Ambystoma were a mean of 19.5 percentage points in error. The reasons for this discrepancy are discussed and predation strategies of the trout, mantid and salamander are compared. === Science, Faculty of === Zoology, Department of === Graduate
|
|---|
