Balancing selection in a polymorphic salamander population

Color polymorphism within populations has fascinated biologists for generations because theory predicts eventual loss of polymorphism without some form of balancing selection. However, it is challenging to distinguish a balanced polymorphism from one slowly changing via genetic drift. I used a 20-year time series to evaluate balancing selection in a population of salamanders (Plethodon ventralis) that vary in the presence of a dorsal stripe (possibly a form of disruptive coloration). Time series analysis supported balancing selection, which might be caused by negative frequency-dependent selection, heterozygote advantage, selection-migration balance, or environmental fluctuations. Polymorphism in small cryptic animals is generally thought to be stabilized by frequency-dependent predation, and I hypothesized that if selection was driven by visual predators, it would be apparent only during the active season, not when salamanders were hidden underground. Analyzing the time series by season supported that hypothesis. Dynamics over the summer inactive season were consistent with a random walk, but during the winter active season there was a strong tendency to converge toward an equilibrium of approximately 30% striped individuals. This is clear evidence of balancing selection, and I argue that frequency-dependent predation is the most likely explanation, but further work is needed to test alternative hypotheses.