Data from "Competitive exclusion and species coexistence in a 25-year grassland field experiment"

I conducted a field experiment designed to test coexistence and competitive exclusion by maximizing the probability that competitive exclusion would occur. I planted 32 species from a range of dry to wet habitats into a 16-year-old mesic prairie restoration, creating a wetness adaptation (niche) gradient of species. I expected the species from very different habitats (wet and dry) than the target one (mesic) to be lost first and the most.<b>Methods</b><br><i>Site</i>The Fermi National Accelerator Laboratory Environmental Research Park (Fermilab), located 75 kilometers west of Chicago, Illinois, (41 50’ N, 88 10’E) encompasses 2600 hectares of previously cultivated land in the midwestern United States. It is located in the prairie peninsula and has rainfall adequate to support temperate forests in the absence of fire. Over fifteen years, two hundred hectares were seeded in an attempt to re-create tallgrass prairie typical of the region prior to European settlement. Planting began in 1975 with over 70 species of seeds. Seeding was done annually in 15 different tracts. The site is burned on approximately a 5-year rotation as weather permits. Tract four, 1.7 ha in size and planted in the fall of 1977, was selected for the experiment based on soil moisture, location, and established plant species. Plots were located in Wauconda silt loam soil, a somewhat poorly drained, non-hydric soil with 6-24 inches to the water table with no flooding. Most of these areas are used to grow corn and soybeans and native vegetation was probably mixed hardwoods and prairie grasses.<br><i>Layout and planting</i>The mesic grassland plantings at Fermilab have been shown to contain restorations with low species richness because of grass over-dominance compared to old-growth mesic grasslands. Old-growth remnant grasslands in the region average 14.2 species per m². Tract four averaged 6.16 species per m² when sampled in 1995. The 16 years prior to seeding would have allowed the vegetation to stabilize beyond early successional interactions, resulting in the most competitive species to become abundant. As part of a supplemental seeding experiment, 32 species were overseeded by hand into nine 10 x 10 m plots in 1993 following a fall burn. The species were selected to represent a wide range of moisture regimes from dry to wet and flowering phenology from spring to fall. Moisture regimes were determined from existing databases including the USDA Plants Database and the Illinois Wildflowers website. Most moisture regimes were similar between the two sources.<br><i>Sampling and Statistical Analysis</i><br>In 1995, data were collected on the existing vegetation. To document the relative abundance of the two dominant grasses, tillers were counted in each of five 1m² randomly located quadrats in each plot. Five aboveground biomass samples were collected for each of the nine plots using 5 circular 616 cm²samples after vegetation became dormant in the fall, separating forbs and grasses and calculating biomass/m². Meander searches documented the number of individuals of each seeded species. For long-term survival, site visits were made sporadically over the next 22 years, documenting the presence of seeded species. In 2020, meander searches were made biweekly both within and around original plots because most species had spread beyond plot boundaries, documenting the presence of seeded species. Survival of the percent of species in each moisture regime was tested using chi square.Six species failed to establish, most from the wetter end of the moisture spectrum: 1 facultative Upland, 2 Facultative, and 3 Obligate Wetlands species. It is possible that germination occurred but seedling were quickly out-competed and were undetected. Five species were found earlier but not in 2020: Baptisia bracteata, Bromus kalmii, Elymus Canadensis, Helenium autumnale, and Polytaenia nuttallii, leaving 21 species that established and persisted. There were no significant differences in the percent of persistent species from each moisture regime (KW test statistic = 7.063, df=4, p=0.133) with upland species the most persistent and obligate wetland species the least persistent. There was a non-significant trend for dry species to do better than wet species at both establishing and persisting.