This thesis examines breeding site attributes and egg size in Ambystoma maculatum, the spotted salamander, in Algonquin Park, Ontario. In Chapter 1, I tested the hypothesis that the presence/absence of fish was the most important factor determining the suitability of breeding sites. In Chapter 2, I investigated the prediction of the optimal egg size theory that when individual offspring survival is low relative to survival in another habitat, females are selected to lay larger eggs. In Chapter 3, I investigated the hypothesis that bet-hedging by producing variably-sized eggs in clutches laid in unpredictable habitats is a superior strategy to laying one size of eggs. The first hypothesis was supported, and the other two were rejected. Female phenotype proved to be far more important to egg size than did predatory environment (Chapter 2), and bet-hedging in egg size appeared to be pre-empted by bet-hedging in reproductive effort (Chapter 3).