Neighbour effects in the genus Avena. 1. Comparison of crop species.
Abstract
A comparison is made of the relative advantages of interspecific mixture experiments using additive and substitutive designs. The results of growing Avena sativa, A. strigosa and the weed species A. fatua and A. ludoviciana in an additive design under glass on fertile, irrigated field soil are described. Although the indicator crop species differed markedly in their pattern of growth, the effects of 4 added species were similar in each. A higher level of infestation merely intensified these effects. The effects of added species were not always quantitative variants of the indicator's own density response and were qualitively different in the cases of stem extension and wt./seed; they were related to the degree of shading suffered by the indicator. For A. sativa, the adaptive value of increased stem extension in the presence of A. ludoviciana was calculated to be about 20% extra wt. in each seed. When canonical variates were calculated using observations of 7 primary characters made on individual plants in each plot the most important variate in each indicator species was identified as wt. of panicles. In A. sativa three axes represented neighbour effects but in A. stigosa only the first and third axes were concerned with these; the second axis represented experimental error. From summary.<new para>ADDITIONAL ABSTRACT:<new para>Oats cv. Blenda and Avena strigosa were grown in glasshouse trials in mixture with each other and with A. fatua and A. ludoviciana using an additive design. Although the 2 indicator species differed markedly in pattern of growth, the effects of infestations of 4 added species were similar in each, and a higher level of infestation merely produced intensified effects. The effects of added species were not always quantitative variants of the indicator's own density response and were qualitatively different where stem extension and weight/seed were considered. The differences were related to the degree of shading suffered by the indicator. In Blenda, the adaptive value of increased stem extension in the presence of A. ludoviciana was calculated to be about 20% extra weight in each seed. Canonical variates were calculated (by means of an ordination technique, described) using observations of 7 primary characters made on the individual plants in each plot. The most important canonical variate in each indicator species was identified as weight of panicles. In Blenda, 3 axes represented neighbour effects, whereas in A. strigosa the 2nd axis represented experimental error.