Does phylogenetic inertia explain the evolution of ineffective antipredator behavior in a sunfish-salamander system?

The streamside salamander, Ambystoma barbouri, exhibits ineffective antipredator behavior (high emergence rate from refuge, and high activity while out of refuge) and thus suffers heavy predation in stream pools with sunfish. A. barbouri evolved relatively recently from an ancestor that closely resembled a sister species, A. texanum, which breeds in fishless, ephemeral ponds. Sunfish thus represent a relatively new selection pressure for A. barbouri. Phylogenetic inertia predicts that (1) A. texanum should be very poor at coping with fish and (2) because it has only recently been exposed to fish, A. barbouri should still be poor at avoiding fish, but due to its recent exposure to fish, A. barbouri should be better than A. texanum at coping with sunfish. Experimental results provided mixed support for these predictions. As predicted, A. texanum suffered heavy sunfish predation. Compared to A. texanum, A. barbouri showed a greater tendency to initiate alarm moves that enhanced escape success from fish. However, in both the presence and absence of fish, A. barbouri showed higher emergence rates from refuge and higher movement while out of refuge than A. texanum. These behaviors tend to increase exposure to sunfish, i.e., for these key behaviors, A. barbouri apparently evolved in the wrong direction as far as fish predation is concerned. Due to these offsetting effects (increased exposure to fish, increased escape success), A. barbouri is no better at surviving with sunfish than A. texanum. A possible explanation for the high activity of A. barbouri is its use of highly ephemeral habitats (relative to A, texanum) that favor the evolution of higher activity, feeding, and developmental rates for A. barbouri relative to A. texanum.