Seed dispersal is a critical ecological process with strong impacts on the distribution of plant populations. Understanding the landscape features that influence seed dispersal, and the consequent influence on future plant populations, is an important goal in ecology. We studied patterns of seed dispersal among patches of Lasthenia californica (California goldfields), a vernal pool plant species, in Southern Oregon, USA. Because seed dispersal is difficult to observe, we examined variation among patches of L. californica using whole chloroplast genomes, which are generally maternally inherited, and therefore represent a record of seed movement. We sampled 20 individuals from 20 populations of L. californica from a 400m2 prairie site maintained by The Nature Conservancy. We targeted the chloroplast genome using a custom capture array, and for each population, sequenced the chloroplasts from one individual and a pool of 20 individuals. We then used a custom programming script to reconstruct and find new haplotypes among the pooled samples, and assess genetic distances between populations. We used Circuitscape to assess the impact of landscape features such as vernal pools, hummocks and bushes on seed dispersal patterns, and also looked at the influence of geographic distance, patterns of vole (Microtus californicus) trails, and L. californica flower density on seed dispersal. We used images captured with a small drone to map the density of L. californica flowers, and satellite images to map vole trails, which could be seen consistently across multiple years. Because L. californica seeds have a small pappus, we expect that the plant is wind dispersed and bushes may act as barriers, although seeds may also move within vernal pools.
Results/Conclusions
We found 29 haplotypes across the 20 populations. We calculated Edward’s genetic distance between populations and found non-significant effects of habitat (defined by low resistance vernal pools, medium resistance hummocks and high resistance bushes) on genetic distance. There was also no significant effect of vole trails on genetic distance. There was a correlation between geographic and genetic distance (MRM, p=0.042), and a strong significant effect of flower density on genetic distance (MRM, p=0.013). Dispersal in this species appears to be somewhat limited as evidenced by the significant IBD and strong local genetic structure. However there was also evidence of haplotype sharing across the length of the prairie. The strong effect of flower density on dispersal may reflect a pattern of micro-habitat specific preferences by this vernal pool plant species.