In many years, cotton producers in the Midsouth delay the first irrigation to allow time for other field operations such as pesticide and fertilizer application. This practice is especially common on clay soils that require several days to dry after furrow irrigation. However, the cost to the producer of delaying irrigation is not well understood. The objective of this research was to estimate the impact of delaying the first irrigation for cotton on clay soil to help producers make more informed decisions regarding irrigation timing. Cotton irrigation studies were conducted at the University of Arkansas Northeast Research and Extension Center at Keiser during the 2001 through 2003 growing seasons, with the cultivar PM 1218 BG/RR planted on a Sharkey silty clay (Chromic Epiaquerts) precision graded to approximately 2 mm m-1 slope. All plots contained four rows approximately 180 m long with a 97 cm row spacing and four-row border area left between each pair of plots. A well watered treatment was irrigated at a 50 mm estimated soil water deficit (SWD) based on the Arkansas Irrigation Scheduler. Irrigations for two delayed-irrigation treatments were initiated on the date of the second irrigation or third irrigation of the well watered treatment and then irrigated at a 50 mm estimated SWD. A nonirrigated check was included. Irrigations were ceased when open bolls were observed. Three-year-average yields decreased with delaying irrigation. There was a consistent trend for lower yield for each delay in the first irrigation; however, in 2003, the differences among all four treatments were not significant. The three-year-mean irrigation water use efficiency was higher for the well watered treatment than for either delayed-irrigation treatment; however, in two of the three years (2001 and 2003), the differences among the treatments were not significant. Gross revenues for the well watered treatment were numerically greatest each year; however, when the costs of irrigation were included, estimated net revenues for the well watered treatment were not always highest. A number of scenarios were investigated (e.g., different water sources, rented land), and even though two of the three years were wetter than normal, the delayed-irrigation treatments always had significantly lower estimated net revenues than the well watered treatment. Furthermore, the estimated net revenues for the delayed-irrigation treatments in each case were not significantly different from those of the nonirrigated treatment. When all other factors were held constant for rented farmland, the well watered treatment had higher estimated net revenues than the nonirrigated treatment for diesel cost < $0.65 L-1 ($2.47 gal-1). A treatment with a delayed first irrigation, a fairly common practice on Midsouth cotton farms, was only more profitable than the nonirrigated treatment when diesel cost < $0.19 L-1 ($0.74 gal-1).