In traditional grazing trials, per animal and per hectare productivity are determined, but pasture and animal measurements are generally inadequate to address reasons for different treatment responses. This 2-yr study examined the diet and diet characteristics of steers grazing tall fescue (Festuca arundinacea Schreb.), switchgrass (Panicum virgatum L.), flaccidgrass (Pennisetum flaccidum Griseb.), and bermudagrass (Cynodon dactylon [L.] Pers.) pastures. A randomized complete block design was used with two agronomic replicates. In vitro DM disappearance (IVDMD) of masticates from tall fescue (TF), switchgrass (SG), and flaccidgrass (FG) generally were similar; all exceeded 70% (70.3 to 79.9%), except for bermudagrass (BG), which was lowest (63.5 to 65.0%). Sieving the masticate DM showed BG to contain the highest proportion (20 to 29%) of particles passing a .5-mm sieve and the lowest proportion (6 to 8%) of particles retained on a 2.8-mm sieve. This resulted in BG having the smallest mean (1.29 mm) and median (1.08 mm) particle sizes in yr 1 and a mean (1.10 mm) and median (.91 mm) particle size smaller than SG in yr 2. Except for BG, the IVDMD of the masticate DM was lowest for small particles. Canopies showed BG and SG to have the highest proportion of stem (47 and 52%, respectively); the BG canopy was composed of fractions that were consistently lowest in IVDMD. In general, BG offered the grazing animal a canopy inferior in IVDMD concentration and had canopy characteristics that limited the animals' selection of a diet with quality as high as that of TF, SG, or FG.