Selection in resemblance matrices

Looking ahead to Section 5, when the active window is a (triangular) resemblance matrix, selection can take place just as for a (rectangular) datasheet, by Select>Highlighted or Select>Samples>(•Sample numbers) or (•Factor levels). Another option is provided in that case: selection of only the rows and columns containing at least one value above or below a specified threshold by, for example, Select>Samples>(•Values>0.95), or selecting only rows and columns containing at least one Undefined! resemblance entry, by Select>Samples>(•Undefined values). These are mainly used for picking out, in the first case, collinear environmental variables from a large correlation matrix (values > 0.95 or <–0.95 say), Section 13. In the second case, this might more easily identify similarities that are undefined because neither sample contains any species at all, in cases where the similarity measure (such as Bray-Curtis) treats such samples as uninformative, Section 5.