A 2-year field experiment was conducted in central Greece (Platykampos, Larissa) to investigate productivity parameters of cotton under conditions of water stress. A Latin square split-plot design with three replications was used to evaluate the effect of three irrigation levels (250, 350, and 450 mm) and three fertilization rates (60, 110, and 160 kg N ha(-1)), where irrigation level was the whole-plot factor and the fertilizer was the split-plot factor. The results showed that irrigation level had no significant effect on soil chemical properties, but these only changed with fertilizer application. Concentration of soil nitrates increased in proportion to the amount of applied fertilizer in early July. The associated rise in electrical conductivity (EC) was not sufficiently high as to adversely affect salt-tolerant cotton. The soil acidity produced during formation of nitrate was evident by a soil pH decrease of 0.2 units in the high fertilizer application. A great decline of nitrate N and EC and a rise of pH in all treatments in early August indicated rapid N uptake by the crop during the late stage of vegetative growth. In contrast, cotton yield was not affected by the rate of fertilizer application but by the level of irrigation. This is the reason that correlations between soil properties and yield were insignificant in early July and August. It appears that there was sufficient N available to the crop from sources other than fertilizer N (soil-derived N and irrigation N). Preplant soil nitrates were greater than residual nitrates in the second growing season and indicated depletion of soil mineral N pools of the order of 36 kg N ha(-1) in the 0- to 25-cm depth. Significant negative correlations between soil properties and cotton yield appeared only at the end of the season and indicated that depletion of soil mineral N increased with increasing crop N requirement or irrigation level.
