A 4-year study to evaluate and further define nitrogen (N) management for cotton utilizing multiple split applications was initiated in 1991 on a Dundee silt loam at the Delta Research and Extension Center. Four N rates (60, 90, 120, and 150 lb N/A) and six N application systems involving combinations of preplant (PP), pin-head square (PH), and/or early bloom (EB) applications were evaluated. Lint yields were quite different between years yet the responses were similar (no year by treatment interaction). In three of four years, there was no significant yield response to N rates above 90 lb/A. Of the application systems evaluated, the PP50-PH50 system produced the highest yields averaged over N rates followed closely by the PP100 and PP25-PH50-EB25 systems. The lowest yields occurred with the PP25-PH25-EB50 system where only 25% of the total N was applied prior to initial fruit set and development. There was no effect of application systems on second harvest yields, only on first harvest and total yield. Multiple split N applications at both pin-head square and early bloom did not increase yields above those obtained with a single split application at pin-head square. Data indicated that having sufficient N available during vegetative growth and early reproductive growth was the most important consideration in timing.