Fossil and molecular data provide complementary insights into late Quaternary population history. I compiled fossil pollen and packrat midden macrofossil data for Douglas-fir (Pseudotsuga menziesii) from over 500 sites in western North America to identify glacial refugia and postglacial changes in distribution. I then used those paleoecological inferences to develop a set of competing hypotheses that I tested using molecular phylogeographic analyses. 

Results/Conclusions

Fossil data suggested two coastal refugia and several Rocky Mountain refugia. However, the possibility of fewer refugia could not be ruled out because of gaps in sampling. Moreover, it is unclear whether or not very low pollen percentages in some places such as Yellowstone indicate Douglas-fir’s presence. In the Southwest, elevation shifts up to 1000 m were observed. Molecular phylogeographic analyses showed that the Rocky Mountains likely harbored multiple distinct glacial populations and lend support to the refugium near Yellowstone. Molecular data also suggested possible refugia not identified in the fossil record in eastern Oregon and the northern U.S. Rockies, which could be tested in future paleoecological studies. Consistent with a history of isolation and elevation shift, Southwestern populations did not contribute to the colonization of Canada. The number of coastal refugia could not be tested for lack of molecular variation there. Molecular data also revealed a unique contact zone among coastal and Rocky Mountain varieties in Canada, in which hybridization occurred during Holocene colonization by pollen-mediated gene flow but not by seed. Combined fossil and molecular data suggest postglacial migration rates of 50 - 165 m/yr, lower than previous estimates and lower than estimates of rates required to track future climate warming in some regions. Fossil and molecular data largely agree, but each provides unique insights.