Agriculture in the Texas High Plains depends heavily on irrigation with water withdrawn from the Ogallala aquifer at nonsustainable rates. Our hypothesis was that integrating crop and livestock systems would reduce irrigation water use, maintain profitability, and diversify income compared with a cotton (Gossypium hirsutum L.) monoculture. Thus, from 1998 to 2002, two large-scale systems, with three replications in a randomized block design, compared water use, productivity, and economics of (i) a cotton (var 'Paymaster 2326RR') monoculture with terminated wheat (Triticum aestivum L.) and (ii) an integrated three-paddock system that included cotton in a two-paddock rotation with grazed wheat and rye (Secale cereale L.) and the perennial 'WW-B. Dahl' old world bluestem [Bothriochloa bladhii (Retz) S.T. Blake] for grazing and seed production. All paddocks were irrigated by subsurface drip. Angus crossbred beef steers (Box taurus L.; initial body weight = 249 kg; standard deviation = 26 kg) grazed from January to mid-July. During the 4 yr of this experiment following the establishment year, cotton lint yield was 1036 and 1062 kg ha -1 for the cotton monoculture and the integrated system, respectively. Bluestem seed yield averaged 24 kg pure live seed ha-1. Steers gained 153 kg on pasture and 0.82 kg d-1. Per hectare, the integrated system used 23% less (P < 0.001) irrigation water, 40% less N fertilizer, and fewer other chemical inputs than the cotton monoculture. Profitability was about 90% greater for the integrated system at described conditions. Integrated production systems that are less dependent on irrigation and chemical inputs appear possible while improving profitability.