80 Christopher King Beachy 



DISCUSSION 



The potential role of variation in offspring size has long been 

 a topic of debate. Kaplan and Cooper (1984) suggested that egg size 

 variation in amphibians enables a female to produce viable offspring 

 in unpredictable environments. In addition, Kaplan and Cooper (1984) 

 submit that there are two levels at which variation in offspring size 

 may be introduced: within a single clutch and among successive clutches 

 of an individual female (i.e., at the intraclutch and interclutch levels). 

 I tested these hypotheses by comparing the amount of variation in 

 egg size observed in five species of salamanders that use larval 

 environments ranging (in terms of safety from desiccation) from 

 permanent to ephemeral. 



These data add to the existing evidence showing that variation 

 in egg size in amphibians is extensive. In addition, it appears that 

 variation in offspring size differs among individual clutches and that 

 the degree of variation differs among species. The question to be 

 asked is whether the degree of variation is related to different life 

 history strategies employed by species using different environments. 



Kaplan and Cooper (1984) proposed that variation in amphibian 

 egg size might reflect different reproductive strategies. In their attempt 

 to model parental investment, they included a consideration of the 

 extensive variation in propagule size seen in amphibians, insects, 

 and plants. Earlier models of parental investment lacked this aspect 

 (e.g., Smith and Fretwell 1974). A study by Crump (1981) emphasized 

 the potential role of variation in amphibian egg sizes. Crump found 

 no significant differences in egg size variation among species of 

 treefrogs that use habitats of differing variability. However, among 

 species that use temporary ponds, individual females produced clutches 

 that tended toward a platykurtic distribution of egg size (bet 

 hedging). Those species that breed in permanent ponds tended 

 toward a leptokurtic distribution (canalization). 



However, Crump's evidence has been criticized as unconvincing 

 (McGinley et al. 1987). In a mathematical consideration of Smith and 

 Fretwell's and Kaplan and Cooper's models, McGinley et al. (1987) 

 suggested that variable environments do not necessarily select for 

 variable parental investment in offspring. Parental fitness can be maxi- 

 mized, even in heterogeneous environments, by investing equally in 

 all offspring. Is variation in egg size adaptive? Or are there factors 

 that prevent a female from investing equally in all offspring? 



If one were to consider only the intraclutch variation that I 

 present here, the supposition that variation in egg size is correlated 

 with desiccation risk appears to garner little support. The data on 

 interclutch variation, however, suggest that species that use ephemeral 



