Measuring maternal line selection driven by differential survival in ex situ collections for plant conservation

Ex situ plant collections may be increasingly needed to protect representative samples of threatened or rare species and provide plant material for conservation translocation. A primary problem in these ex situ collections is loss of intraspecific variation due to plant death. Different management actions would be required depending on whether such loss results from demographic stochasticity or selection. We developed a method to test for maternal line selection according to provenance, given that maternal line and provenance serve as convenient and important partitions of the intraspecific variation represented in ex situ plant collections. The method tests for provenance selection (i.e., the differential survival of plants in maternal lines sourced from different geographic regions) and opportunity for selection within provenances (i.e., differential survival among plants in maternal lines sourced from the same geographic region). We used data from an ex situ collection of 243 saplings sourced from eight provenance regions and 42 maternal lines of Arkansas oak (Quercus arkansana), a vulnerable tree species restricted to small, scattered populations. We found no statistically significant provenance selection or opportunity for selection among the maternal lines during a 10.5-month period, despite low overall survival rate (0.531) and large variation in survival rate among provenances (0.250-0.833) and maternal lines (0-1). By distinguishing survival patterns expected solely under demographic stochasticity from those likely resulting from selection, our method can be used to inform management of a wide range of ex situ collections for plant conservation.