The development of optimal weed management strategies that rely, in part, on crop interference will require an understanding of how weeds compensate for limitations in above- and belowground resources. Trade-offs in the leaf morphology and biomass partitioning of rice and late watergrass were investigated under glasshouse conditions in 1999 and 2000. Both species responded to shade with increased height, reduced biomass, greater partitioning of biomass to leaves, and greater leaf area ratios. At the lowest light level (18% sunlight), plants of both species showed little response to nitrogen (N). However, height, tillers, biomass, and leaf area increased for plants grown at 50% and full sunlight as N increased from 0 to 224 kg N ha−1. Late watergrass exhibited more plasticity in specific leaf area and root weight ratio than rice in response to shade. This plasticity contributed to the ability of late watergrass to maintain a higher percent of its tillers and total dry weight than rice when sunlight was reduced by 50%. These results support the hypothesis that except at low light levels, limited N further reduces the growth of shaded late watergrass plants. Thus, weed management strategies that limit the plasticity of late watergrass by manipulating light and N availability are likely to be more effective than strategies that rely on manipulating a single resource.

Late watergrass is a serious weed of California rice that has evolved resistance to molinate, thiobencarb, fenoxaprop-ethyl, and bispyribac-sodium. To obtain an insight into the origin and spread of resistant (R) late watergrass in California rice fields, we evaluated similarities in morphological traits and amplified fragment length polymorphism (AFLP) fingerprints among 15 R strains compared with susceptible (S) strains. All strains were derived by inbreeding from accessions collected in rice fields of the Sacramento Valley, CA. In the field, R plants were shorter than S plants; they also had narrower and shorter flag leaves and thinner culms. Spikelets also appeared smaller and more slender in R plants. There was greater morphological similarity among the 15 R strains than among the eight S strains. The mean coefficients of variation for morphological traits were much smaller among R strains, which in a cluster analysis (Ward's method) were grouped morphologically apart at early clustering stages from the more variable S strains. AFLP electropherograms also showed greater similarity between R strains. R strains were grouped separately from the S strains in a cluster analysis based on calculated Nei and Li coefficients used in an unweighted pair group method using arithmetic means. However, small genetic differences also existed because the R strains were grouped into six clusters, suggesting that R strains were not samples from an identical strain. It was concluded that R strains originated from a preexisting and preadapted mutant late watergrass population in the Sacramento Valley. This study establishes that resistance moved by spikelet dispersal, not independent mutation events, most likely defined the geographical distribution of R late watergrass in California. Prevention and control of this dispersal combined with elimination of seed-producing survivors after herbicide treatment should be relevant components of the integrated management of herbicide-resistant late watergrass in California rice.