Stylosanthes guianensis (Aubl.) Sw. varieties were evaluated for shoot characteristics and associated variability within and between varieties under field conditions at Fort Pierce, Fl., USA during the summer of 1989. Shoot dry weights, stem diameters, taproot diameters were measured on 14 varieties 75 days after transplanting. Coefficient of variation (CV) and percentage sums of squares from the analysis of variance (ANOVA) were used as a measure of variation. Minimal variations for shoot characteristics were attributed to replications or plant samples. The variations associated with experimental and sampling errors (ANOVA) were 22.4 - 29.3 and 30.7 - 34.1 %, respectively, for taproot diameter and shoot characteristics. The large experimental error suggests that a high proportion of variability in the ANOVA model was unknown, whereas the large sampling error suggests that within-variety variability was also considerably large. Mean shoot weights had a nine-fold difference between varieties with CV values ranging from 36 to 88%. This study suggests that a large inherent variation for shoot characteristics occurs between and within Stylosanthes varieties. Despite the large unknown variation, differences in shoot characteristics between varieties may be exploited by plant breeders.