RESUMO
A standard conservation strategy for plant genetic resources integrates in situ (on-farm or wild) and ex situ (gene or field bank) approaches. Gene bank managers collect ex situ accessions that represent a comprehensive snap shot of the genetic diversity of in situ populations at a given time and place. Although simple in theory, achieving complementary in situ and ex situ holdings is challenging. Using Trifolium thompsonii as a model insect-pollinated herbaceous perennial species, we used AFLP markers to compare genetic diversity and structure of ex situ accessions collected at two time periods (1995, 2004) from four locations, with their corresponding in situ populations sampled in 2009. Our goal was to assess the complementarity of the two approaches. We examined how gene flow, selection and genetic drift contributed to population change. Across locations, we found no difference in diversity between ex situ and in situ samples. One population showed a decline in genetic diversity over the 15 years studied. Population genetic differentiation among the four locations was significant, but weak. Association tests suggested infrequent, long distance gene flow. Selection and drift occurred, but differences due to spatial effects were three times as strong as differences attributed to temporal effects, and suggested recollection efforts could occur at intervals greater than fifteen years. An effective collecting strategy for insect pollinated herbaceous perennial species was to sample >150 plants, equalize maternal contribution, and sample along random transects with sufficient space between plants to minimize intrafamilial sampling. Quantifying genetic change between ex situ and in situ accessions allows genetic resource managers to validate ex situ collecting and maintenance protocols, develop appropriate recollection intervals, and provide an early detection mechanism for identifying problematic conditions that can be addressed to prevent further decline in vulnerable in situ populations.