Two methods of determining diet botanical composition, microhistological (MH), and stable carbon isotope ratio (CR) analyses were used to determine botanical composition of ingesta and fecal grab samples in steers grazing rhizoma peanut-mixed tropical grass pastures. Three pastures were used over two grazing seasons, 1992 and 1993, in Brooksville, FL. A weighted-disc double-sampling technique was used to determine forage mass and botanical composition, percentage of rhizoma peanut (Arachis glabrata), grass (Paspalum notatum and Cynodon dactlyon), and forb (primarily Chenopodium ambrosioides) on offer every 28 d throughout the grazing seasons. There was an effect of sampling date (P<.001), sampling date x pasture (P<.001), and sampling date x year (P<.001) on forage mass on offer. There was a pasture x year x sampling date interaction (P<.001) for all botanical components. In 1992 and 1993, using cannulated steers sampled every 56 d, there were interactions with year for rhizoma peanut and forb (P<.05), but not for grass with MH analysis (components: rhizoma peanut, grass, and forb). Ingesta and fecal rhizoma peanut (r = .73 and .92 for 1992 and 1993, respectively) and ingesta and fecal forb (r = .86 and .98 for 1992 and 1993, respectively) were positively correlated (P<.001). Ingesta and fecal grass were positively correlated (r = .52, P<.001), but the correlation was not as high. With the CR analysis (components: Calvin cycle [C3] plants and C4-dicarboxylic acid pathway [C4] plants), ingesta and corrected fecal (corrected for in vitro organic matter digestibility [IVOMD]) C3 plants were positively correlated (r = .62; P<.001). Diet composition of fecal grab samples from noncannulated steers, collected on the same sampling schedule as for hand-clipped pasture samples, differed at times due to the complexity of the sward (both rhizoma peanut and forb constituted a single component, C3, in the CR analysis). Based on these results, if there is a substantial contribution of forb to the diet, fecal microhistological analysis may be more informative than fecal carbon ratio analysis for estimating diet selection by cattle grazing tropical pastures.