Plants from four species' of barley, H. agriocrithon, H. spontaneum, H. distichum, H. intermedium, and their F 1 hybrids were grown together in a mechanical' diallel and competitive effects were investigated by the application of the diallel analysis technique of DURRANT (1965) .
There was no evidence of hybrid vigour when the hybrids were grown as pure stands but in binary mixtures the hybrids, considered overall, were more vigorous than their parents . Although this relationship was not complete the hybrids grown with the parents formed the most productive class, exceeding by 24 % the combined mean grain yield of the parents grown alone and in parental mixtures .
In the eight characters investigated there was a tendency for compensation to occur between the mixture components but this was far from complete (i .e. an increase in one component was not precisely balanced by a decrease in the other) . Those genotypes with large pure stand values tended to become smaller and, in compensation, the smaller genotypes larger. Since, however, the compensation was not complete the mixture means tended to be lower than the mean of the components in pure stand . The converse was found to be true for yield of grain .
The principal response to competition is for each genotype to be either uniformly suppressed or enhanced in the presence of an associated genotype, the strength and direction of this response being given by the a value calculated for each genotype . The a values were negatively correlated with their pure stand values (e .g . plants with a high dry weight when grown alone were depressed more than plants with a low dry weight when grown alone, and vice-versa) . There was, however, no correlation between the pure stand value and the a value in respect of yield and 100-seed weight .
Although both H. agriocrithon and H. spontaneum were found to be poor competitors, only those hybrids derived from crosses with H. spontaneum have this poor competitive ability transmitted : hybrids involving H. agriocrithon are intermediate in competitive ability between their parents . Again, the hybrid formed between the two advanced barleys H. distichum and H. intermedium produces a hybrid of lower competitive ability than either parent .
The poorer competitors tend to produce fewer tillers and, as a consequence have a lowered seed output per plant when associated with other genotypes in mixtures . Grown in pure stand the weak competitors tend to be multi-tillering and to have a high output of seeds .
The plastic responses of other characters associated with yield, together with those such as ear emergence and height are also discussed . One hybrid, (s x d), a cross between H. spontaneum and H. distichum is of particular interest in that although a poor competitor it is nevertheless very high yielding when grown in pure stand . Furthermore, its suppression in mixtures is not sufficiently great as to reduce its overall ranking compared with its associates when aggregate yield is considered . Analytical techniques such as those described here could be of use to the plant breeder in his search for high yielding, compatible genotypes, of low competitive ability. The technique may also be used to test breeding procedures .