||<p>Because species respond individually to climate change, understanding community
assembly requires examination of multiple species from a diversity of forest niches.
I present the post-glacial phylogeographic history of an understory, parasitic herb
(<italic>Epifagus virginiana</italic>, beechdrop) that has an obligate and host specific
relationship with a common eastern North American hardwood tree (<italic>Fagus grandifolia</italic>,
American beech). The migration histories of the host and parasite are compared to
elucidate potential limits on the parasite's range and to understand their responses
to shared climate change. Two chloroplast DNA regions were sequenced and 9 microsatellite
loci genotyped from parasite specimens collected throughout the host's range. These
data were compared with available cpDNA sequences from the host (McLachlan et al.
2005) and host fossil pollen records from the last 21,000 years (Williams et al. 2004).
Analyses of genetic diversity reveal high population differentiation in the parasite's
southern range, a possible result of long term isolation within multiple southern
glacial refuges. Estimates of migration rates and divergence times using Bayesian
coalescent methods show the parasite initiating its post-glacial range expansion by
migrating northward into the northeast from southern areas, then westward into the
midwest, a pattern consistent with the development of high density beech forests.
This result is strongly confirmed through spatial linear regression models, which
show host density plays a significant role in structuring parasite populations, while
the initial migration routes of the host are irrelevant to parasite colonization patterns.
Host density is then used as a proxy for the parasite's habitat quality in an effort
to identify the geographic locations of its migration corridors. Habitat cost models
are parameterized through use of the parasite's genetic data, and landscape path analyses
based on the habitat map show a major migration corridor south of the Great Lakes
connecting the northeast and midwest. Host density was the major determinant controlling
the parasite's range expansion, suggesting a lag time between host and parasite colonization
of new territory. Parasites and other highly specialized species may generally migrate
slower due to their complex landscape requirements, resulting in disassociation of
forest assemblages during these times. From these results, the low migration capacities
of highly specialized species may be insufficient to outrun extirpation from their