In many areas of western North America, populations of elk are believed to harm production of cattle by competing with them for limited supplies of native forage. We examined effects of variation in the population density of elk (Cervus elaphus canadensis) during winter on growth and reproduction of cattle during spring, using a randomized complete block experiment conducted in sagebrush grassland during four years. We manipulated elk numbers to achieve four levels of population density (0, 9, 15, and 31 elk/km2), replicated each level three times, and observed responses of cattle to these manipulations. Mean birth dates of calves born to cows in the intermediate (9, 15 elk/km2) treatment levels were delayed by 5 d relative to birth dates of calves born to cows in controls (0 elk/km2), but these trends were not statistically significant (quadratic effect F1,6 = 3.2, P = 0.13). Calf body mass at birth was not significantly influenced by treatment, but calf body mass at the end of spring declined linearly (F1,6 = 7.9, P = 0.03) with increasing elk population density, from a mean of 80.2 kg in the control to 73.0 kg in the 31 elk/km2 treatment. Calf body mass at weaning was weakly depressed by treatment, with the largest treatment effects occurring at the 9 elk/km2 level (F1,6 = 8.8, P = 0.02). Body mass of cows at the end of the spring grazing season tended to decline linearly with treatment, but these tendencies were not statistically significant (linear contrast F1,6 = 1.9, P = 0.22). We did not find significant effects of treatment on cow body mass at time of weaning (control vs. other contrast F1,6 = 3.3, P = 0.12) or on natality rates (control vs. other contrast F1,6 = 1.2, P = 0.31), although, in both cases, values for the control tended to exceed the treatments. Body mass of cows (t1 = ‐3.9, P = 0.0003) and calves (t1 = ‐3.9, P < 0.0001) at the end of the spring grazing season were quadratically related to the biomass of available herbaceous forage during spring. We observed a threshold in effects of forage supply on cattle production at about 45 g/m2 of live and dead herbaceous biomass. Cattle production declined with declining forage biomass when forage supply fell below this threshold. Total cattle production (kilograms per cow per year), was quadratically related to elk population density (F1,6 = 5.8, P = 0.05). Average cattle production in the control (X = 248 kg·cow‐1·yr‐1) exceeded the mean of the other treatment levels (X = 224 kg·cow‐1·yr‐1, F1,6 = 6.7, P = 0.04). Quadratic responses in cattle performance were apparently caused by compensatory growth after the spring grazing season: growth rates of cows during spring were inversely related to their subsequent growth rates during summer and fall (F1,46 = 33.5, P < 0.0001, r2 = 0.50). We conclude that elk grazing caused reductions in cattle production, but the magnitude of the effects were not proportionate to elk population density. Our studies revealed that competition between cattle and elk operates in a threshold fashion. If sufficient forage is available to cattle following elk grazing, elk populations will not harm cattle performance. In the system we studied, this threshold occurred at ≈ 45 g/m2 of live and dead herbaceous biomass.