Three areas in Northern Michigan; Vanderbilt, Pellston and Cross Village were compared for climate differences; host plant phenology differences and Papilio canadensis Rothschild and Jordan (Lepidoptera: Papilionidae) oviposition preference differences. The ’cold pocket’ hypothesis predicted that these sites were climatically distinct, with Vanderbilt the coolest site and Cross Village the warmest; that phenology in the ’cold pocket’, Vanderbilt area, would be delayed; and that oviposition preference by ’cold pocket’ P. canadesis butterflies would be for white ash. Every year with regard to total degree-day accumulations, Vanderbilt was the coldest site. Vanderbilt was cooler than the other sites only sixteen times in twenty-nine years during the time period when P. canadensis butterflies were actively selecting host plants (March 1st - July 5th). In the years of this study, 1996 and 1997, P. canadensis butterfly populations from across Northern Michigan did not show oviposition preference differences. In 1996 and 1997, butterflies from the ’cold pocket’ did not show an oviposition preference for white ash. These results indicated a greater depth and complexity to climate/plant/herbivore interactions than previously assumed by the ’cold pocket’ hypothesis.