Geographic Variation in the WarningSignals of the Wood Tiger Moth (Parasemia plantaginis; Arctiidae)EsitetÀÀn JyvÀskylÀn yliopiston matemaattis-luonnontieteellisen tiedekunnan suostumuksella julkisesti tarkastettavaksi yliopiston YlistönrinteellÀ, salissa YAA303 marraskuun 8. pÀivÀnÀ 2013 kello 12.Academic dissertation to be publicly discussed, Warning signals are not expected to be diverse because they are under stabilizing selection. This dissertation aims to study historic and contemporary selection mechanisms underlying the geographic warning signal diversity in the wood tiger moth (Parasemia plantaginis). In the first study, I explored the variation in hindwing warning color frequencies across the distribution of P. plantaginis and the phylogeography of the species. Males have polymorphic hindwing color (yellow or white) across much of their distribution but turn reddish in the Caucasus and mostly white in Japan, which coincide with mitogenetic divergence. Females vary continuously in hindwing warning color between yellow and red, becoming yellow east of the Ural mountain range, but does not correspond to mitogenetic divergence. Post-glacial isolation may be one contributing factor to the genetic divergence and the hindwing warning signal differences in the Caucasus region. In the second study, I found that thermoregulation benefits of melanin can trade-off with a larger more effective warning signal, which contributes to variation in melanization within Europe. In the third study, I found that forewing patterns function as warning signals and that generalization by predators might play an important role in warning signal diversity along with frequency dependent selection. In the fourth chapter, I found three populations in Japan differing in warning signal traits with some gene flow suggesting both selection and isolation may be underlying causes. Overall, my results suggest that historical and abiotic factors contribute to the observed warning signal diversity in P. plantaginis, rather than predation alone. This adds to our understanding of how historical factors along with environmental heterogeneity in biotic and abiotic factors can promote warning signal diversity. My results highlight the need to take an integrated approach to studying geographic diversity in warning signals.