Clustering of strains belonging to the Black Aspergilli by multivariate analysis. Taxonomic Studies on Japanese Industrial Strains of the Aspergilli (Part 31)

Abstract

Each of 20 kinds of mycological character were graded into 6 ranks on 69 representative strains and tabulated in Table 95 based upon the categories in Table 89 in Part 27, Table 90 in Part 28 and Tables 92 and 93 in Part 30 and they were analysed by a computer using the present table and Dr. C. Hayashi's method, “A Factor Analytic Method for Qualitative Data or Quantification on Response Pattern”, and then eigenvalues (ρ²) and sample scores in 5 components were shown in Table 96.
The strain numbers and sample scores in each of 5 components were noted and plotted on 4 planes named Case I, II, III and IV, as seen in Figs. 37, 38, 39 and 40, respectively, in which strains keeping a short distance and/or keeping a similar relation to one another were obtained by lines. Thus, 6 clusters, (a), (b), (c), (d), (e), and (f) were obtained including some subclusters such as (d)', (d)'(f)' and (f)'.
The results are statistically significant when they agreed with one another in more than 2 Cases. The clusters (d) and (f) were more or less overlapped and some strains belonged to some different clusters when the Cases varied, although most of the strains belonged to a constant cluster even in different Cases. The relations between the clusters and Cases were shown in Table 97.
In order to know distribution of mycological characters in a cluster, some principal characters along with their grade-numbers were selected and the selected strains were shown with large black spots in the Case I, as seen in Figs. 41 to 48.