THE EVOLUTION AND GENETICS OF HERBICIDE RESISTANCE IN WEEDS

The importance of various factors influencing the evolution of herbici de resistance in weeds is critically examined using population genetic models. The factors include gene mutation, initial frequency of resis tance alleles, inheritance, weed fitness in the presence and absence o f herbicide, mating system, and gene flow. Where weed infestations are heavy, the probability of selecting for resistance can be high even w hen the rate of mutation is low Subsequent to the occurrence of a resi stant mutant, repeated treatments with herbicides having the same mode of action can lead to the rapid evolution of a predominantly resistan t population. At a given herbicide selection intensity, the initial fr equency of resistance alleles determines the number of generations req uired to reach a specific frequency of resistant plants. The initial f requency of resistance alleles has a greater influence on the evolutio nary process when herbicides impose weak selection, as opposed to very strong selection. Under selection, dominant resistance alleles increa se in frequency more rapidly than recessive alleles in random mating o r highly outcrossing weed populations. In highly self-fertilizing spec ies, dominant and recessive resistance alleles increase in frequency a t approximately the same rate. Gene flow through pollen or seed moveme nt from resistant weed populations can provide a source of resistance alleles in previously susceptible populations. Because rates of gene f low are generally higher than rates of mutation, the time required to reach a high level of resistance in such situations is greatly reduced . Contrary to common misconception, gene flow from a susceptible popul ation to a population undergoing resistance evolution is unlikely to s low the evolutionary process significantly. Accurate measurements of m any factors that influence resistance evolution are difficult, if not impossible, to obtain experimentally. Thus, the use of models to predi ct times to resistance in specific situations is markedly limited. How ever, with appropriate assumptions, they can be invaluable in assessin g the relative effectiveness of various management practices to avoid, or delay, the occurrence of herbicide resistance in weed populations.