Aposematic organisms have warning signals advertising their defenses to potential predators. Warning signal efficacy is better in higher densities, therefore, positive frequency-dependent selection is expected to select against less common signals. This should lead to a reduction in warning signal diversity within a given area, but could promote the formation of multiple locally adapted signals. Many of the studies on locally adapted warning signals have been carried out in tropical systems, while fewer studies have explored locally adapted warning signals in northern areas. The wood tiger moth (Parasemia plantaginis) is found across the Holarctic and has conspicuous hindwing coloration that serves as a warning signal. In addition to its hindwings, P. plantaginis also has conspicuous black and white forewing patterns that may function as a warning signal. Two predominant forewing patterns are observed in Alaska. One pattern is dominant around Fairbanks, while the other is dominant around Anchorage. The differences in which pattern is most and least abundant in each geographic location raises the possibility that local predators contribute to maintaining the pattern differences through positive frequency dependent selection in each area. We hypothesized that the forewing pattern differences are due to predator-driven selection.
Results/Conclusions
To test our hypothesis, we placed artificial moths with both local dominant and foreign forewing patterns in each of the two regions to see if predators selected against the foreign forewing types, which would suggest the forewing patterns have a warning signal function. We found that the locally predominant forewing pattern in Fairbanks was avoided more by predators than the less abundant forewing pattern. However, both forewing patterns were avoided equally around Anchorage. Population genetic analyses from specimens caught during fieldwork showed that separate populations existed in Fairbanks and Anchorage with restricted gene flow between them. This result is consistent with the hypothesis that selection removes recent migrants with dissimilar forewing patterns. Our predation experiment results suggest that positive frequency dependent selection may be partially responsible for maintaining warning signal differences between Anchorage and Fairbanks, although predators seem to avoid both forewing patterns in Anchorage. Restricted gene flow between the two populations could be attributed to a combination of selection against foreign morphs in Fairbanks and physical barriers, which both likely contribute to forewing pattern differences in Alaska.