Switchgrass (Panicum virgatum) has been identified as a model dedicated energy crop species. After a perennial grass is established, the major variable costs are for nitrogen (N) fertilizer and harvest. Prior to establishing switchgrass on millions of ha in a particular agro-climatic region, it would be useful to determine switchgrass yield response to N and its response to harvest frequency relative to alternative grass species. The objective of this research is to determine biomass yield response to N for four perennial grass species and to determine the species, N level, and harvest frequency that will maximize expected net returns, given the climate and soils of the U.S.A. Southern Plains. Yield data were produced in an experiment that includes four species (switchgrass, bermudagrass (Cynodon dactylon), weeping lovegrass (Eragrostis curvula), and carostan flaccidgrass (Pennisetum flaccidum)), four N levels, and two harvest levels. Linear response plateau (LRP), linear response stochastic plateau (LRSP), and quadratic response (QR) functions are estimated. For all combinations of biomass and N prices considered, the optimal species is switchgrass. For most price situations, it is economically optimal to fertilize established stands of switchgrass with 69 kg N ha(-1) yr(-1) and to harvest once yr(-1) after senescence. (C) 2010 Elsevier Ltd. All rights reserved.