the objective of this work was to determine the repeatability and the number of the required evaluations to obtain determination coefficient greater than 90% in the productive and qualitative variables of forage and cattle excreta on a signal grass (urochloa decumbens) pasture. the utilized variables were: fecal production, feces mineral composition, herbage mineral composition around dung pads, feces biometry, herbage mass and herbage rejection around dung pads, urinary volume, and urine chemical composition. a fixed rotational stocking was used, with three days of occupation and 32 or 67 days of resting period, during the rainy and dry seasons, respectively. repeatability analyzes were obtained with the software genes, using the principal component method based on the covariance. the repeatability coefficients (r) had high estimates, and the r2 were equal to or greater than 90%, except for herbage mass at 20-40 cm, for the difference between pre-and post-grazing herbage height of herbage grown on dung pad, and for nitrogen concentration in the urine for the treatment with 3,2 animal units per hectare (r<0.50). for r2 above 90%, the number of evaluations required would be 11, 9, and 9, for the variables forage, feces, and urine, respectively.