Plant species richness is an important property of ecosystems that is altered by grazing. In a semiarid environment, we tested the hypotheses that (1) small-scale herbaceous plant species richness declines linearly with increasing grazing intensity by large ungulates, (2) precipitation and percent sand interact with grazing intensity, and (3) response of herbaceous plant species richness to increasing intensity of ungulate grazing varies with patch productivity. During January to March 2012, we randomly allocated 50, 1.5-m x 1.5-m grazing exclosures within each of six 2,500 ha study sites across South Texas, USA. We counted the number of herbaceous plant species and harvested vegetation in 0.25-m² plots within exclosures (ungrazed control plots) and in the grazed area outside the exclosures (grazed treatment plots) during October and November 2012–2019. We estimated percent use (grazing intensity) based on the difference in herbaceous plant standing crop between control plots and treatment plots. We selected the negative binomial regression model that best explained the relationship between grazing intensity and herbaceous plant species richness using the Schwarz Bayesian Information Criterion. After accounting for the positive effect of precipitation and percent sand on herbaceous plant species richness, species richness/0.25 m² increased slightly from 0 to ~ 30% grazing intensity and then declined with increasing grazing intensity. Linear and quadratic responses of herbaceous plant species richness to increasing grazing intensity were greater for the least productive patches (<15.7 g/0.25 m²) than for productive patches (≥15.7 g/0.25 m²). Our results followed the pattern predicted by the intermediate disturbance hypothesis model for the effect of grazing intensity on small-scale herbaceous plant species richness.

We sampled vegetation in the study sites during October and November (autumn) 2012-2019 inside and outside 1.5 m x 1.5 m grazing exclosures. Fifty grazing exclosures were randomly allocated with each of six 2,500 ha study sites during January - March 2012. Grazing exclosures were rerandomized each year after vegetation sampling. Vegetation sampling consisted of counting the number of herbaceous plant species with a 0.25 m² sampling frame in the exclosure and in the grazed area outside the exclosure and then harvesting herbaceous vegetation inside and outside the exclosures. Harvested plant material was dried at 45 C to a constant mass and weighed. Data were entered in an excel spread sheet and statistical analyses were performed using Statistical Analysis Systems (SAS) software.