PROJECT SUMMARY Low-input high-diversity (LIHD) grasslands are a promising system for biofuel production as they provide additional environmental benefits compared to high- input low-diversity systems (HILD). In theory, LIHD systems are analogous to perennial native hay systems which can maintain high diversity and sustainable biomass production. We report the results of a seven- year hay management experiment in which fertilization and native prairie seed addition were manipulated in a Kansas grassland. In this experiment, we used a seed addition to test the extent that HILD plots, which were dominated by introduced C 3 grasses initially sown for haying, would convert to LIHD. We manipulated haying, fertilization, and seed addition in sixteen 10x20m plots in a split-plot design. Seed addition in non-fertilized (LIHD) plots strongly increased plant diversity and altered plant composition leading to high abundance of native C 4 tall grasses. Removal of biomass by annual haying in the LIHD plots decreased diversity slightly compared to plots without haying but remained substantially higher than non-seed addition plots. Biomass production in the non-seed addition plots without fertilizer decreased 42% relative to fertilized plots, indicating that high inputs are required to maintain high production. Remarkably, four years after seed addition peak biomass in the LIHD plots was equivalent to HILD plots suggesting that this low input system may achieve comparable production to some HILD systems without inputs of fertilizer. Gregory R. Houseman, Bryan L. Foster, & Kelly L. Kindscher, Department of Ecology and Evolution, University of Kansas, Lawrence, KS, USA Conversion from HILD to LIHD grasslands by seed addition: impacts on plant diversity and production under hay management Effect of native seed addition on production in a similar small-scale experiment Changes in biomass over time for fertilized (F) and seed addition plots (S) measured in early summer (black symbols) corresponding to peak C 3 biomass. Following restoration to C 4 communities, peak biomass was measured in Fall 2007 (red, green symbols) for seed added plots (connected by arrows) Hay yields of unfertilized native communities equaled fertilized C 3 communities Seed addition shifted communities to C 4 grassland and increased plant diversity in both hayed and non- hayed plots Production Diversity Effect of seed addition and fertilization on hay yields measured in the Fall of Conclusions Plots undergoing restoration have the potential for high production and native diversity compared to non-native C 3 systems in NE Kansas Restored native C 4 grasslands are an attractive biofuel system because of the potential for high production, low input costs (e.g. fertilization, planting, tilling), and increased conservation benefits such as enhanced native diversity, low N leaching, and increased habitat heterogeneity. METHODS In March of 2000 we established a 4 x 4 grid of 16, 10 x 20 m plots at the site. Plots were separated by 3 m buffer strips. Each 10 x 20 m plot was divided into two contiguous 10 x 10 m subplots (yielding a total of 32 subplots). A 2 x 2 x 2 factorial set of treatments was assigned to the experimental plots and subplots in a split-plot design: two levels of nutrient fertilization (fertilized; not fertilized) applied as the whole-plot factor; two levels of native prairie seed sowing (sown; seed not sown) applied as a whole plot factor, and two levels of haying (hayed; not hayed) applied as the split-plot factor. There were four replicates of each unique treatment combination. NPK fertilizer (29-3-4) was spread by hand to each 10 x 20 m whole-plot at a rate of g N/m2 per year. In 2003 and 2004, seeds of 41 native species were sown evenly by hand across whole-plots at an average rate 47 seeds per m 2. From 2001 to 2007 we sampled aboveground plant biomass annually just prior to haying and during the peak of production for cool- season grasslands. In 2007, we also sampled in September corresponding to peak biomass for native tallgrass prairie communities To sample biomass, two 0.1 x 2 m strips of aboveground plant material were harvested within each subplot with electric clippers. Each sample was separated into live and litter fractions, with the live fraction further sorted to species. All biomass fractions were dried to constant mass at 74˚C. Low-input, high diversity C4 plots High-input, low diversity C3 plots Standing crop (g m 2 ) -F-S -F+S +F-S +F+S -F+S -F-S Spring Fall No FertFert Standing biomass (g/m2) No Seeds Seeds No seeds Seeds -Hay +Hay -Hay +Hay -F-S -F+S +F-S +F+S Shannon Diversity Species richness -Hay +Hay