Figure 4. Biases in the total versus dimorphic part of the data in three exemplar species. Biases, seen as sex-wise correlation of principal component space, is less when analysing the dimorphic part of the data only as opposed to the total data, likely because the signal/noise ratio is higher in the former: A–D, Seminatrix pygae; E–H, Cyperideis torosa; I–L, Mus musculus. Blue is male and red is female (true sex).