Annual crop fields typically are simple habitats dominated by a few plant species where pesticides play a major role in managing weed and insect infestations. Recently, there has been significant interest in the potential to reduce reliance on pesticides by manipulating plant species and communities to benefit natural enemies of insects and weeds. Such efforts aim to enhance natural enemy impact by providing appropriate food, shelter, and hosts, and efforts typically are accomplished by manipulation of plant species, populations, or communities. Habitat management is generally viewed as an important factor in maintaining stable insect and natural enemy populations in agricultural systems and may have a similar function in increasing weed seed predation. Crop and noncrop habitats provide resources to natural enemies either directly through floral nectar and pollen, indirectly by increased host or prey availability, or through emergent properties of the habitat such as by moderating the microclimate. These critical resources for natural enemies can be provided in agricultural ecosystems at several scales: within fields, at field margins, or as a component of the larger landscape. Because individual natural enemy species may require quite specific resources at different times and spatial scales, not all attempts to manipulate habitat diversity are equally effective. We review the role of plant resources, including weeds, in supporting natural enemy communities and provide case studies of how varying plant diversity at different spatial scales can influence the effectiveness of biological control in agricultural landscapes.

Avena fatua (wild oat) populations with resistance (R) to one or more herbicides have been described in numerous cropping systems worldwide. We previously reported that the R3 and R4 wild oat populations from Montana, USA, were resistant to four herbicides representing three different modes of action: tralkoxydim [acetyl-CoA carboxylase (ACCase] inhibitor), imazamethabenz and flucarbazone [acetolactate synthase (ALS) inhibitors] and difenzoquat (growth inhibitor). We now quantify resistance levels of these populations to triallate [very long chain fatty acid (VLCFA) biosynthesis inhibitor], pinoxaden (ACCase inhibitor) and paraquat (photosystem I inhibitor). Glasshouse doseresponse experiments showed that, compared with the means of two susceptible (S) populations, the R3 and R4 populations were 17.5-and 18.1-fold more resistant to triallate, 3.6-and 3.7-fold more resistant to pinoxaden, respectively, and 3.2-fold (R3) more resistant to paraquat. Pre-treatment of R plants with the cytochrome P450 inhibitor malathion partially reversed the resistance phenotype for flucarbazone (both populations), imazamethabenz (R4), difenzoquat (R4) and pinoxaden (R3), but not for tralkoxydim, fenoxaprop-P-ethyl or triallate. Target site point mutations known to confer resistance to ALS or ACCase inhibitors were not detected via DNA sequencing and allele-specific PCR assays in R plants, suggesting the involvement of non-target site resistance mechanism(s) for these herbicides. Together, our results complete the initial characterisation of wild oat populations that are resistant to seven (R3) or six (R4) herbicides from five or four mode of action families respectively.