- Golden paintbrush (Castilleja levisecta Greenman) historically inhabited the prairies of the Willamette Valley, Oregon. However, this Pacific Northwest endemic is currently restricted to eleven sites in the Puget Trough of Washington and British Columbia. Recovery criteria call for the establishment of new populations throughout the species historic range, including the Willamette Valley. In order to facilitate reintroduction to this region, we examined: (1) habitat characteristics of potential C. levisecta recovery sites in the Willamette Valley and compared them to remaining populations in the Puget Trough, (2) a suite of ecological and genetic factors likely to contribute to reintroduction success using common garden experiments, and (3) how the
performance of this rare hemiparasite was affected by the availability of different host combinations in the greenhouse and in the field. Potential C. levisecta reintroduction sites in the Willamette Valley had distinct vegetation communities and soil characteristics compared with remaining populations in the Puget Trough. This disparity was likely related to regional differences in geology, climate, ocean proximity, and land-use history. Many of the species indicative of remaining populations in the Puget Trough were native pereimials, while those of
potential reintroduction sites in the Willamette Valley were introduced annuals. Soil characteristics of C. levisecta sites were also distinct among the two ecoregions. Puget Trough sites were located on sandy soils with generally high levels of magnesium and sulfur, while Willamette Valley sites were found on silty-clay soils with high concentrations of potassium and phosphorous. Differences in soil texture, and magnesium and potassium concentrations were associated with plant community divergence among the two regions. Decisions regarding seed source and recovery site selection are especially difficult in portions of a species range that are uninhabited, as no reference populations exist with which to compare ecological information. Using common garden experiments, we tested hypotheses about how C. levisecta transplants would perform in relation to the ecological similarity between seed source and introduction site, the effective population size and genetic diversity of seed sources, and the habitat quality of the recovery site. We observed significant variation in performance measures among source populations and common garden sites. Plant community characteristics, including the abundance of nonnative
species and the similarity in community structure between source populations and common garden sites, helped explain the variation in these performance measures. Exotic species cover at common garden sites was associated with a reduction in performance of first year C. levisecta transplants. Survival of second year transplants increased with increasing similarity in plant functional groups between source and common garden sites, supporting the idea of a "home-habitat advantage." These results indicate that high quality prairies, dominated by native perennial species with low nonnative abundance, should be targeted for recovery sites. We recommend using plant material from Whidbey Island, WA, whose three populations represented in our study
consistently performed well. Rare, parasitic plants pose an interesting challenge to restoration practitioners. We examined how the performance of C. levisecta was affected by the availability of
different host combinations in the greenhouse and in the field. Castilleja levisecta individuals were grown with two grass individuals (Festuca roemeri; Poaceae), two non-grass hosts (Eriophyllum lanatum; Asteraceae), one individual of each of these species, or without a host. Our greenhouse results provide little support for the complimentary diet hypothesis, which states that parasites grown with multiple host species perform better than individuals grown alone or with a single host. Castilleja levisecta individuals grown with two different species performed better than those co-planted only with F. roemeri, but did not differ from E. lanatum or no-host treatments. In the field, vole activity had indirect effects on C. levisecta survival mediated through host species. Vole tunneling and C. levisecta mortality were strongly associated with host treatments including E. lanatum. Field survival was significantly higher among no-host C. levisecta individuals than those grown with E. lanatum or mixed host treatments. We do not suggest coplanting C. levisecta with E. lanatum in the field. Although no-host C. levisecta
individuals had the greatest first year field survival, we suspect that perennial host plants will be beneficial to future survival. Therefore, we recommend planting C. levisecta in the Willamette Valley, OR with F. roemeri.