Species of Trichoderma and the closely related genus Gliocladium are ubiquitous, active components of the soil microflora. As biocontrol agents, these fungi can protect a wide range of agronomic, horticultural, and ornamental plants against a variety of plant pathogens (6, 13), and a few strains have been patented for commercial use. Some strains enhance plant growth or rooting of cuttings even in the absence of pathogens (4, 14). Trichoderma spp. have also been studied extensively as potential sources of cellulases for utilization of lignocellulosic materials (10, 17). The taxonomy, genetics and population composition of these fungi are poorly understood. Limited knowledge of variability in these fungi and infrequent culturing of their sexual stages make delineation of narrowly defined species difficult (16). Additionally, little is known about such genetic components as ploidy levels, frequency of heterokaryosis, or the prevalence of parasexual events. Finally, methods are lacking to differentiate among strains for patent purposes, or to determine the variability and abundance of strains in natural ecosystems. Isozyme analysis by starch gel electrophoresis is a powerful method for taxonomic, genetic and population studies. Descriptions of general techniques and of numerous enzyme assay systems have been published (9, 12, 18, 20, 21). However, for studies of specific organisms, suitable enzyme assays and corresponding electrophoretic buffers must be determined empirically (8, 12, 18). For an enzyme assay to be useful, enzyme activity must be reliably detected and activity bands must be clearly resolved. In addition, polymorphism must exist among isolates of the study population. Also, the usefulness of an enzyme assay for a particular study often depends on the number of loci detected. Assays detecting products of single loci can be particularly useful because polymorphism among individuals reflects allelic differences, and electrophoretic variants can be scored as allozymes. The genetic basis of an isozyme phenotype usually can be inferred from the complexity of the activity band pattern visualized (9, 12, 18). In this study, 63 enzyme assays were tested in combination with four electrophoretic buffers to assess detection of enzyme activity, resolution of activity bands, and complexity ofphenotypes visualized. This report summarizes the most useful systems for studies of Trichoderma and Gliocladium spp. and identifies those which apparently visualize products of single loci, i.e., allo-