Wheat stem sawfly (WSS), Cephus cinctus Norton (Hymenoptera: Cephidae) is the most damaging insect pest to Montana's $ 1 billion dollar per year grain industry. Current WSS control methods are either expensive, reduce wheat yields, or are not effective. Our objective was to compare burning, grazing, tilling, trampling and clipping wheat stubble fields on over-wintering WSS larval populations. Treatments were evaluated in three experiments using a randomized complete block design and four replications at each site. Eight, six, and two sites were used for Experiments 1, 2, and 3, respectively. Contrast statements were used to make pre-planned comparisons among treatments. For Experiment 1, treatments were fall tilled, fall grazed, spring grazed, fall and spring combined (Fall/Spr) grazed, and an untreated control. Five mature ewes were confined with electric fence to 111 m 2 plot for 24 h in the fall and spring grazed treatments resulting in a stocking rate of 452 sheep d/ha. For Fall/Spr, the stocking rate was 904 sheep d/ha. For Experiment 2, treatments were fall grazed, fall burned, fall tilled, and an untreated control. In Experiment 3, treatments were fall trampled, spring trampled, Fall/Spr trampled, hand clipped to a stubble height of 4.5 cm, and an untreated control. Trampled treatments were done at the same stocking rates as grazing treatments but sheep were muzzled to prevent intake. Wheat stem sawfly larval numbers were collected in the fall and spring, pre- and post-treatment, respectively, by collecting all plant material from three, 0.46 m lengths of row and counting the number of live larvae present. In Experiment 1, WSS mortality was greater ( P < 0.01) for the mean of all grazed treatments (68.4%) than either control (43%) or tilled (47%) plots. Mortality did not differ ( P = 0.75) between fall (67%) and spring (64%) grazed plots but was greater ( P = 0.02) for Fall/Spr (74%). In Experiment 2, larva mortality was greater ( P < 0.01) for fall grazed (63%) than burned plots (52%). In Experiment 3, WSS mortality was greater ( P < 0.01) for the mean of all trampling treatments (57%) than either control (33%) or clipped (32%) plots. Mortality did not differ ( P > 0.25) between fall (54%) and spring trampling (47%) but was greater ( P = 0.01) for Fall/Spr (70.6%). No differences ( P > 0.25) were detected for WSS mortality when grazing was compared to trampling. These results indicate the potential for using grazing sheep to control wheat stem sawfly infestations in cereal grain production systems.