Restoring grassland legumes using overseeding can depend on seed spatial structure

Abstract Introduction Although many grassland restoration experiments address initial planting strategies, relatively few test whether overseeding of replanted native grassland can improve legume and forb establishment. Additionally, most overseeding experiments thoroughly mix seed to ensure that each species reaches all areas of a site. However, natural dispersal is unlikely to exhibit this “uniform” spatial seed structure but more closely approximates “aggregated” seed arrival. Objectives I tested whether forb species had higher abundance when using aggregated rather than uniform overseeding and whether the response varied with sowing density. Methods I overseeded 15 forb species into 3 × 5 m plots using either uniform or aggregated spatial arrangements at three seeding rates and quantified cover after six growing seasons in a south‐central Kansas restored prairie. Results Legume frequency and cover were higher than other overseeded forbs. At the highest seeding rates, Amorpha canescens cover was 15 times higher in aggregated than uniform sowing treatments. In contrast, Desmanthus illinoensis benefited most from uniform sowing, while Lespedeza capitata was insensitive to spatial overseeding treatments. The benefit exhibited by A. canescens under aggregated sowing occurred despite application of seeds to only 6.7% of the area occupied in uniform plots (0.75 vs. 11.25 m 2 ). The effect of overseeding on plant species richness increased with seed density; however, at the highest rate, uniform sowing counteracted the positive effect of overseeding on plot richness. Conclusions The results suggest that adding spatial structure to overseeding may increase the establishment of some forb species in replanted grasslands.