Presumably, from a physiological standpoint, early season water stress should be avoided to ensure early fruit initiation, good fruit retention, and optimum yield potential of cotton (Gossypium spp.). This study was conducted to determine the optimum timing of the initial post-plant irrigation and the long term effect of postponement on subsequent plant growth patterns, fruit retention, and yield. A short-season Upland variety, (G. hirsutum L.), DPL 20, was planted on 19 April in Marana, AZ, elevation 1970 ft., on a Pima clay loam (Typic Torrifluvent) soil. Plots (experimental units) consisted of eight 40 in. rows and extended the full length of the irrigation run (600 ft.). Experimental design was a randomized complete block with four replications. Initial post-plant irrigations, designated T1, T2, and T3, were applied when the midday leaf water potential ( (x)) of the uppermost, fully-developed leaf reached -15, -19, and -23 bars, respectively. All treatments received the same irrigation regime following the initial post-plant irrigation. Basic plant measurements were taken weekly from each experimental unit. These included plant height, number of mainstem nodes, location of first fruiting branch, fruit retention, number of nodes above the uppermost white bloom, bloom count within a 166-ft(2) area, and percent canopy cover. Soil-water data at seven 25 cm depth increments was collected from a total of 36 access tubes located within the field study, with three tubes per plot. Lint yields (lb. lint acre^-1) were 1112, 1095 , and 977 for T1, T2, and T3, respectively. Yields were significantly lower when the initial post-plant irrigation was applied after (x) dropped below -19 bars, confirming the results of a previous study conducted in 1992. Throughout the growing season, height-node ratios (HNR) of T1 and T2 plants were at or above the upper threshold established for DPL 20, while T3 HNR remained close to the expected baseline. Fruit retention was low for all three treatments due to season long insect pressure from Lygus bug. The low fruit retention data also reflects the effects of high HNR. Future work will include efforts to separate changes in (x) due to day-to-day climatic variations from those caused by soil-water depletion. A second objective will be to incorporate the data obtained from the neutron moisture meter probe into the study results in an effort to better describe the complete soil-plant-atmosphere continuum as affected by the various treatment regimes employed in this study.