Several skip-row planting systems, designed to make soil water from fallowed rows available to plants in adjacent planted rows, are used in many of the cotton growing regions of the USA and the world. Three planting designs were established in 1980 and 1981 by planting every row (solid), two of three rows (2 .times. 1), or two of four rows (2 .times. 2) to determine the effectiveness of skip-row planting systems for cotton development and yield. Soil water extraction from planted and fallow row was determined by neutron scattering, while soil cores were used to determine root length density (RLD) in planted and fallowed beds. Cotton in the 2 .times. 2 and 2 .times. 1 patterns yielded 71 and 47% more lint than the solid design when determined on a planted area basis. No differences in yield, however, were detected on a total area basis. Cotton in the solid pattern had significantly higher micronaire, while the skip-row designs resulted in greater staple length. Seasonal water extraction was similar for the planted rows of all designs and for the fallowed row of the 2 .times. 1 pattern. Fallowed rows of the 2 .times. 1 pattern exhibited the least water removal. Water use efficiency (WUE) was not influenced by pattern in 1980, with all designs having values of about 3.0 kg lint ha-1 mm-1 water. Water use efficiency decreased to 2.1 for the solid and 2 .times. 1 patterns and to 1.6 kg ha-1 m-1 for the 2 .times. 2 design in 1981. Cottin in the solid design generally exhibited the greatest RLD at 0-150 mm. No significant differences between treatments were observed at lower soil depths. Root length density tended to decrease between sampling dates in this study. Rooting activity was significantly correlated with water use. Based on yields, fiber characteristics, and WUE, skip-row systems were not more efficient than planting every row.