Measurement of herbage‐on‐offer in grazing trials is useful for monitoring seasonal changes in herbage availability, for defining the grazing pressure, and may serve as the basis for adjusting rates of stocking. Generally, a large number of samples must be harvested per pasture in order to obtain reliable estimates of herbage yield. However, the amount of labor and time needed to harvest and process samples may limit the number of samples which may feasibly be taken per pasture. In this experiment, our objective was to maintain precision with a low number of harvested samples by first determining the mean bulk‐height of the pasture with a weighted disk measure and then harvesting sites of the mean bulk‐height (± 1/2 cm). Every 5 days, 25 measurements of bulk‐height were made at random locations per paddock with a weighted disk meter having a 0.2 m2 disk weighing 2.5 kg and dropped from a height of 80 cm. Mean bulk‐height was calculated and 10 quadrats (0.25 m2) harvested per paddock from locations having the mean bulk‐height. The data obtained in pastures of tall fescue (Festuca arundinacea Schreb.) were influenced by period of sampling (spring, summer, and fall) and grazing treatments (1‐, 2‐, or 4‐paddock rotation). The coefficients of variation for fresh herbage yield of 10 0.25 m2 quadrats at the mean bulk‐height within grazing treatments, were about 5%. For combined data from the grazing treatments, correlation coefficients (r) between herbage mean bulk‐height and dry herbage yield as determined from the harvested quadrats were 0.714, 0.822, and 0.709 for spring, summer, and fall growth. However, the correlation coefficients varied both in spring and fall according to different grazing treatments. In spring grazing, r's were 0.399, 0.918, and 0.670; and in the fall grazing, r's were 0.300, 0.848, and 0.580 for the 1‐, 2‐, and 4‐paddock systems, respectively. Correlation coefficients varied less in summer with r's of 0.904, 0.914, and 0.772 for the 1‐, 2‐, and 4‐paddock systems. Trampling of forage and accumulation of mature residue in some grazing treatments appeared responsible for the low correlation between mean bulk‐height and harvested herbage yield. Use of the disk‐meter appears to have greatest usefulness in pastures where there is not excessive accumulation of plant residues due to undergrazing or trampling. Under such conditions, sites located by mean bulk‐height allow the experiments to quickly obtain reliable herbage yield estimates from fewer sampling sites.