A model for cultivar classification and identification with reference to oats (Avena). II. A probabilistic definition of cultivar groupings and their Bayesian identification

Using analysis of variance, four sequentially obtained characters were chosen from 21 to discriminate among 14 groupings (demes) proposed elsewhere. The 14 estimated sets of 36 multinomial probabilities obtained therefrom, together with the estimated probability of occurrence of each deme, were used to construct a Bayesian identification procedure. Computer simulation of identification suggested that 50 plants are sufficient for identification. The merits of Bayesian identification schemes for infraspecific categories are described.