RESULTS

Larval Survival

The proportion of A. macrodactylum surviving in ponds with trout was significantly lower than in ponds without trout (Tables 6.1 and 6.2). All A. macrodactylum larvae in control ponds survived the 30 d experiment (Table 6.1). Maximum number of surviving A. macrodactylum larvae in ponds with trout was 4. Ambystoma gracile survivorship was also significantly lower in ponds with trout than in control ponds (Tables 6.1 and 6.2). However, survivorship of A. gracile larvae was more variable than that of A. macrodactylum in both control ponds and ponds with trout (Table 6.1). Survivorship of A. gracile and A. macrodactylum larvae were not statistically compared because experiments were not run concurrently and the results could be confounded by experimental conditions.

Table 6.1. Mean number and proportion (p) of Ambystoma macrodactylum (AM) and A. gracile (AG) surviving in artificial ponds with trout and without trout.

^aNumber of ponds used for survivorship analysis; one fishless tank with A. gracile was omitted from statistical analysis due to disease outbreak.

Table 6.2. Summary of ANOVA on proportion of larval Ambystoma macrodactylum (AM) and A. gracile (AG) surviving 30 d experiments in artificial ponds with trout and without trout.

Larval Growth

Both A. macrodactylum and A. gracile in ponds without trout had significantly greater SVLs than larvae in ponds with trout (A. macrodactylum, P = 0.0071; A. gracile, P < 0.0001, Wilcoxin tests; Table 6.3). Prior to the start of the experiments, SVLs were not significantly different between treatment groups (A. macrodactylum, P = 0.83; A. gracile, P = 0.27, Wilcoxin tests). Mean larval body mass also was not significantly different between treatment groups at the beginning of the experiments (A. macrodactylum, P = 0.17; A. gracile, P = 0.24, Wilcoxin tests; Table 6.3). At the end of the 30 d experiment, larval A. gracile in ponds without trout had significantly greater body mass than did larvae in ponds with trout (P = 0.0081, Wilcoxin test). However, larval A. macrodactylum did not differ in body mass between treatments (P = 0.11, Wilcoxin test; Table 6.3).

Table 6.3. Mean snout-vent lengths (mm) and weights (g) of larval A. macrodactylum and gracile in ponds with fish and without fish at the start of each 30 d experiment. Snout-vent lengths were measured for each larva. Mean weights were calculated by dividing the total weight of all larvae by the number of larvae in each pond.

Substrate Use

Substrate location of larval A. macrodactylum on day 30 was significantly influenced by the presence of trout (P < 0.0001, X² = 693.3, d.f.=4, Drop in Deviance test). Larval A. macrodactylum surviving 30 d in ponds with fish were found only in rock substrates (Table 6.4). In contrast, larval A. macrodactylum in control ponds were present in all substrates and in the open at the conclusion of the experiment. Substrate location of larval A. gracile was also influenced by trout on the final experiment day (P < 0.0001, X² = 632.08, d.f.=4, Drop in Deviance test). Larval A. gracile in ponds with trout were found either in rock or woody materials (Table 6.4). Ambystoma gracile in control ponds were found primarily in wood and vegetation, but also occurred in rock substrates and in open areas. At the termination of the laboratory experiments no larvae of either species in ponds with trout were found in vegetation or in the open.

Table 6.4. Mean (range) of surviving larval A. macrodactylum (AM) and A. gracile (AG) associated with each substrate at the conclusion of laboratory experiments. Larvae classified as open were exposed on sand substrate or in the water column.

^aSubmerged vegetation.