The wheat stem sawfly, Cephus cinctus Norton, is a major insect pest in dryland wheat (Triticum L. spp.; Poaceae) fields in the northern Great Plains of the United States and in southern regions of the prairie provinces of Canada. Field infestations by this pest commonly show a distinct "edge effect", with infestation levels highest at the field edge and declining inwards. In this study, we characterized the edge effect of C. cinctus infested wheat fields in Montana at four locations in two separate years. "Infestation level" refers to the average proportion of wheat stems infested with C. cinctus in samples collected in a given wheat field. The gradual decrease in C. cinctus infestation towards the centre of wheat fields is referred to as the "infestation slope", and it was presented as a proportion of infestation level at field edges. We showed that (i) for most of the fields bordering grassland or fallow wheat, a simple exponential decline curve provided a significant fit to the infestation slope; (ii) the infestation slopes in different fields and in different directions within fields were similar and appeared to be independent of the overall infestation level; and (iii) the relationship between infestation level and the proportion of samples infested followed an asymptotic curve, which indicates a high level of dispersal of ovipositing C. cinctus females. The general perception has been that the edge effect of C. cinctus infestation is a result of ovipositing C. cinctus females being both weak flyers and short-lived, but our results do not fully corroborate this perception. Currently, the only reliable way to detect C. cinctus infestations is to dissect individual stems and look for developing instars. However, this type of sampling is extremely time consuming and labor intensive and therefore impractical for wheat growers. Increasing the understanding of the spatial patterns in C. cinctus infestations is a first step towards development of an optimal sampling plan for this important field pest. Thus, the results presented are relevant for researchers involved in sampling designs and integrated pest management procedures for field pests. We discuss behavioral and evolutionary factors that may contribute to the edge effect of C. cinctus infestations.