To bulk or not to bulk? Intense genetic differentiation of seed dormancy among Stipa lagascae (Poaceae) genotypes accessed and multiplied in arid Tunisia

Abstract

Many ecological restoration projects involving wild dryland grasses work with bulk‐harvested seed, yet bulking may lead one to overlook and lose genetic variation that turns out to be crucial for restoration success, including specific dormancy breakdown requirements (conditional dormancy) and genetic variation in these. This article reports on the genetic variation of conditional seed dormancy of Stipa lagascae Roem. & Schult. (Poaceae), a winter‐growing bunch grass with high‐grazing value, native to the circum‐Mediterranean areas but threatened with extinction. This grass is a cleistogamous inbreeder whose population remnants exhibit a fine‐grained genetic structure in many ecological traits. Seed from different lineages was obtained at the end of a seed increase program that chose not to bulk seed during seed sourcing, seed increase, nor seed harvesting for final use. Over a period of 3 years, three germination trials were run with seed coming from 15 to 20 different seed lineages and harvested over different years in two related seed increase plots. Seed lineage was the single most important factor explaining differences in conditional dormancy. Seed weight did not add much to explain these differences. The singularity of behavior of each lineage was conserved over the different harvests. Young seed exhibits the strongest differentiation in conditional dormancy. To restore genetically diversified populations of this inbreeding grass, restoration practitioners should avoid to bulk seed during sourcing and seed increase, because unique seed lineages get easily lost unseen. However, seed for final use could be bulk harvested if seed is used when still conditionally dormant.