Bracon cephi and Bracon lissogaster are native parasitoids of the wheat stem sawfly, Cephus cinctus, an important pest of dryland wheat production. This spatial distribution study, using survey data from seven dryland wheat fields at four locations in north-central Montana over two years, examined: (1) the spatial distribution and spatial association of C. cinctus infestation and parasitism, and (2) parasitism in response to the level of C. cinctus infestation. Spatial distribution analyses revealed that C. cinctus infestations were significantly aggregated along field edges in three of five fields, while parasitism followed a random spatial distribution in all five fields. We demonstrate that the level of parasitism is highest for lower levels of C. cinctus infestation, in fields where parasitoids are either scarce or abundant. The combination of (1) a spatial aggregation of C. cinctus infestations along field edges, and (2) a potential negative interaction between parasitism and C. cinctus infestation may lower the level of parasitism at high host densities. This outcome can be viewed as beneficial to the insect pest, because parasitized C. cinctus larvae are at a greater risk of being cannibalized by conspecifics. An equally plausible hypothesis for this negative interaction is a limitation on the successful location of unparasitized hosts at higher host densities, which is also compelling because of the cryptic location of hosts within wheat stems that belie little visible evidence of infestation. Supporting data from other heavily infested fields that were sampled in 1999-2000 are used to illustrate the recurrence of this pattern. These findings are discussed in the context of developing redistribution strategies for introducing these parasitoids at sawfly infested locations where they are currently present at negligible levels.