Abstract
Scoring for morphometric characters was conducted on both herbarium material (including artificially produced hybrids) and specimens collected across Canada and Alaska and studied in situ. The data from herbarium material were analysed by means of various principal component analyses and by a series of discriminant analyses. The morphometric data from field material were obtained by repeated observations of every spikelet along the one side of each spike selected for analysis. A representative sample of spikes was collected from every population visited, with emphasis on hybrid populations. These repeated observations within spikes were also made in order to evaluate the reliability, representability, and taxonomic value of every spikelet and the degree of homologization within a spike. The field data were analysed by means of discriminant analyses and various cluster analyses.The conclusion reached is the recognition of three species for Canada and Alaska, H. jubatum L., H. caespitosum Scribn., H. brachyantherum Nevski, and a hybrid H. jubatum × brachyantherum representing F1 and F1-like phenotypes. Classification function coefficients were computed and are provided in this paper in order to effect both the identification of the three taxa and the hybrid and their circumscription.All the herbarium specimens were identified by means of these coefficients and, as a result, previously unsuspected hybrids were revealed. This study indicated that an analysis of herbarium specimens alone can be misleading and, consequently, that field data are necessary. It is also demonstrated that the spikelets within a spike are not uniform enough for taxonomic reliance to be placed on observations made on one spikelet alone.
Publisher
Canadian Science Publishing
Cited by
12 articles.
订阅此论文施引文献
订阅此论文施引文献,注册后可以免费订阅5篇论文的施引文献,订阅后可以查看论文全部施引文献