Background/Question/Methods: Many checkerspot butterflies form small, discrete colonies with independent population dynamics, and these insects have become one of the key model organisms for meta-population dynamics. These dynamics, however, have also caused many species to become rare or endangered. In 2011 and 2012, we studied two species from the genus
Euphydryas with similar life-histories and metapopulation dynamics, but whose populations overall are performing very differently. One species, the Baltimore checkerspot (
E. phaeton), is common and wide-spread in Eastern North America. The Taylor's checkerspot (
E. editha taylori), by contrast, is rapidly declining, currently found at only a handful of sites in the Pacific Northwest, and recently declared endangered. At two sites in the summer of 2011, 43
E. phaeton were tracked for 15 minutes, recording behavior, intraspecific interactions, and location (using hand-held GPS units) every 15 seconds. In late spring of 2012, the same protocol was used to measure movement and behavior of
E. e. tayloriflying in two adjacent fields in western Oregon. The data were analyzed using mixed-effects models, and step-length distributions were fit to bounded power-law, exponential, mixed-exponential, and Weibull distributions using ML methods. Subsequently, we developed a semi-discrete metapopulation model with a novel density-dependent emigration function affected by male harassment. We used this model to simulate metapopulation stability over 100 generations under variable intensities of male harassment to assess the effect of harassment on long-term population stability.
Results/Conclusions: We found strong differences in the rates of harassment between E. phaeton and E. e. taylori. E. phaeton females experienced little or no harassment while E. e. taylori females often experienced unrelenting harassment by courting males. Female movement patterns were strongly affected by harassing males, and often attempted to hide or flee. Fleeing females were frequently chased out of suitable habitat patches. In E. e. taylori, fleeing movements caused the overall step-length pattern to be best fit by a bounded power-law as compared to an exponential distribution, while an exponential best fit the step-lengths of the largely unharassed E. phaetonfemales. When compared against a broader set of distributions, both Weibull and and mixtures of exponential distributions were favored in both species. Meta-population models revealed that, when populations become isolated, intense male harassment increases extinction probability. Models incorporating male harassment indicated a long-term extinction probability of 17.5 to 61%, depending upon population size, compared 1.2 to 27% for the same population sizes using a typical density-depend emigration function.
