A cotton irrigation study was conducted at Lubbock, Texas in 1995 using continuously measured canopy temperature to automatically apply water. The objective of the experiment was to determine the number of irrigations applied by different time thresholds and measure the response of cotton to periods of either water deficit or excess applied at different times during the season. Three irrigation levels were created by using different stress times, accumulations of either 4, 6, or 8 hours above a canopy temperature threshold of 28°C, to produce irrigation signals. The irrigation signals scheduled irrigation in a system that used a base irrigation interval of 3 days. In addition to three normal water levels established by time thresholds of 4 h, 6 h, and 8 h, periods of water deficit and water excess were superimposed on each water level. Total irrigations of 39, 35, and 33 cm were applied by the 4 h, 6 h, and 8 h time thresholds, respectively, between DOY 180 and DOY 243. The number of irrigations applied by the time threshold treatments were similar except, during the late irrigation season between DOY 218 and DOY 243, when the 4 h and 6 h time threshold treatments applied 7 irrigations compared to 6 irrigations for the 8 h threshold treatment. The number of irrigations applied in 1995 varied from 18 to 15 with no radiation level limitations on the accumulation of stress time. Restricting stress time accumulation to the daylight period when radiation was above 200 wm^-2 improved the uniformity of irrigation control between 1995 and previous tests conducted in 1991 and 1992. Daily stress time and leaf water potential were linearly related among time threshold irrigation treatments. Daily stress time values were correlated with single, daytime leaf water potential measurements taken when the canopy was under maximum daily heat stress. There was no difference between the lint yields of 1404 and 1435 lbs/acre for the 4 h and 6 h time threshold treatments which were greater than the 1271 lbs/acre yield of the 8 h time threshold treatment.