Genetic variability for loci governing enzyme/morphological variants and for herbicide response was determined in 10 populations of the slender wild oat (Avena barbata Pott. ex Link ♯ AVEBA), six populations of wild oat (Avena fatua L. ♯ AVEFA), and three populations of godetia (Clarkia williamsonii Lewis & Lewis). The enzyme loci were identified by starch gel electrophoresis and included peroxidase, 6-phosphogluconate dehydrogenase, esterase, and leucine aminopeptidase for the slender wild oat; peroxidase, esterase, leucine aminopeptidase, and malate dehydrogenase for the wild oat; and esterase, phosphoglucoisomerase, leucine aminopeptidase, acid phosphatase, and glutamate oxaloacetate transaminase for godetia. Morphological loci included lemma and leaf sheath hairiness for the oats. For both the enzymatic and morphological loci, levels of genetic variation for each population were quantified by a polymorphic index. The herbicide barban (4-chloro-2-butynyl 3-chlorophenylcarbamate) was used on the wild oats; bromoxynil (3,5-dibromo-4-hydroxybenzonitrile) was used on godetia. Genetic variation for herbicide response was based on genetic variances calculated from phytotoxicity scores. Populations were ranked from highest to lowest for the polymorphic indices and the genetic variances. Concordance between the rankings was tested by rank correlation. Statistically significant relationships were found between the enzyme/morphological characters and herbicide response in the slender wild oat and the wild oat. For some species, the level of genetic variation for response to herbicides appears to be associated with genetic variation for enzymatic and morphological loci.

The morphological and physiological response of the slender oat (Avena barbata Pott ex Link # AVEBA) to the herbicides barban (4-chloro-2-butynyl 3-chlorophenylcarbamate) and difenzoquat (1,2-dimethyl-3,5-diphenyl-1H-pyrazolium) in conjunction with decreased water availability was determined for seven populations, representing three ecotypes, under greenhouse conditions. Generally, within the range of sublethal herbicide doses, with increasing herbicide rates, phytotoxicity ratings increased, but plant dry weight, tiller height, and number of spikelets were decreased. Also, the number of juvenile tillers decreased, while that of fertile tillers increased. Flag leaf area increased and flowering was delayed. The ratio of number of spikelets to plant dry weight and seed weight was influenced the least. Under the highest rates of difenzoquat, the within-plant variance of spikelet number decreased, indicating that there may have been a more equal partitioning of resources amongst tillers for spikelet production. The general influence of water stress was to amplify the effect of the herbicide. For example, the dry treatment reduced dry weight and tiller height, and delayed flowering. Within a particular herbicide treatment, the effect of the water stress was to cause reduced within-plant variance for days to flowering, flag leaf area, and number of spikelets. Three reactions were observed that could have helped buffer decreases in spikelet production: 1) An increased fraction of the dry weight of the plants was partitioned into the spikelets at the expense of other vegetative matter, 2) the increased leaf area of the primary tiller may have helped counterbalance any reduction in photosynthesis caused by herbicide action, and 3) an increased number of juvenile tillers was converted into fertile tillers resulting in an increased number of mature tillers. These data indicate that the slender oat has a remarkable “phenotypic plasticity,” which enables it to maintain reproductive structures under sublethal herbicide doses.

The level of genetic variation for tolerance to herbicides was quantified in populations of slender wild oat (Avena barbata Brott. # AVEBA), wild oat (Avena fatua L. # AVEFA), and godetia (Clarkia williamsonii Lewis & Lewis) that had not been previously exposed to herbicides. Seedlings of wild oat and godetia were treated with barban (4-chloro-2-butynl-m-chlorocarbanilate) and bromoxynil (3,5-dibromo-4-hydroxybenzonitrile), respectively. The plants were rated for phytotoxic effects following treatment. A one-way analysis of variance on arcsin-transformed phytotoxicity ratings showed significant amounts of inter- and intrapopulation variability for herbicide reaction. Furthermore, the amount of genetic variance for herbicide reaction is higher than expected on the basis of mutation alone, suggesting selection favoring genes conferring herbicide tolerance occurs in natural populations.