This manual provides a reference guide to permanent sample plot (PSP) techniques in mixed tropical forests. It is intended as an update to <i>Synnott's Manual of permanent plot procedures for tropical rainforests</i>. The manual considers the objectives of PSPs as that of providing data for growth and yield models which will be used to assist forest management. The PSPs are a means of measuring tree growth, mortality, and regeneration in relation to stand density. They may be laid down either as a network of sampling plots, passively sampling existing forest management practices; or as measurement plots within an experimental design. Both approaches are properly required for the optimum capture of data. There is no satisfactory method of calculating numbers of PSPs required, but approximate indications are that one plot per 1000 ha of forest, laid in a stratified random sample, will normally be adequate. Experimental designs should concentrate on the effects of extremes of stand density, from controlled logging or poisoning, and always include untreated and unlogged forest as a control to define maximum basal area and growth under maximal competition. Randomized block design is recommended as simple, robust and statistically rigorous. Replicates should be placed on different sites.