A density trial of cashew (Anarcardium occidentale L.) was carried out at the National Research Centre for Cashew, Experimental Station, Shantigodu, India, during 1982-94 with the objectives of studying growth behaviour, root distribution, photosynthesis, weed biomass, dry branch accumulation, and yield under the varying densities. Densities adopted were 156 trees, 278 trees, 625 trees, 1111 trees, and 2500 trees ha-1. The density was reduced to 50% in the treatment with 1111 trees ha-1 and to 25% in the treatment with 2500 trees ha-1 in the seventh year. Density was reduced to 50% by the 11th year in the treatment with 625 trees ha-1. Higher density treatment resulted in deeper root penetration, lower moisture at depths above 60 cm, greater light interception, and as a result, greater mutual shading induced drying of branches. Yield levels were maintained in higher density plots by pruning to 80% light interception or at later stages by diagonal thinning. Maintaining a tree density of 625 ha-1 (4 m x 4 m) for the first 11 years and diagonal thinning thereafter to reduce the population to 50% (8 m x 5. 7 m x 5.7 m) resulted in maximum cumulative nut yield (4.94 ha-1) and profit (U.S. $2464.90 ha-1).