Relationships among 70 North American Oat Germplasms: I. Cluster Analysis Using Quantitative Characters

Abstract

Cluster analysis can be used to identify cultivars with similar adaptation, which can be useful for sampling in subsequent studies and parental selection in breeding programs. To test these applications in oat (Avena sativa L. and A. byzantina C. Koch), relationships among 70 current and historically important North American oat cultivars and germplasm accessions were measured using cluster analysis of 13 quantitative characters. Cultivars were planted in replicated field trials at two locations for 2 yr and evaluated for heading date, maturity date, height, tillers per mete row, aboveground biomass, grain weight, test weight, 100‐kernel weight, and culm diameter. Four derived characters (harvest index, growth phase index, grain‐filling period, and a straw stiffness index) were also estimated for each germplasm. Heading date was the most important source of variation among genotypes. Cluster analysis produced four large groups that generally corresponded to latitude of origin or adaptation with significant differences between groups for all characters. Three late‐heading spring oat cultivars (Banner, White Tartar, and Cornellian) were placed together with five late‐heading winter cultivars (Hairy Culberson, Winter Turf, Walken, Norline, and Wintok), despite crop history and pedigrees indicating that these two groups of cultivars are divergent. The importance of heading date as a source of variation and the false grouping of divergent, lateheading oats as genetically similar sugges that clusters based on quantitative characters are derived from a biased genetic‐relationship measure. Clustering by morphology may, however, be valuable for identification of genotypes with similar adaptations for breeding and agronomic research programs.