Skip to main content
. 2009 Nov 23;4(11):e7957. doi: 10.1371/journal.pone.0007957

Figure 1. In a common morphospace, major axes of morphological diversity may still differ among groups of interest.

Figure 1

Ordered axis plots may not be able to discriminate between patterns of morphological diversity along axes of a multidimensional morphospace because of its reliance upon principal components analysis and the inherent biases of this method. Here the aspects of diversity parallel to PC1 are highlighted with a red arrow for each group of interest. Note that the length of most variable group (pink) is parallel to PC1 because it has the greatest influence over the determination of PC1 in this common morphospace. Other less variable groups (blue, green) have less influence over the trajectory of PC1, but still possess variation that lies parallel to PC1. However the greatest axis of variation within these less variable groups may lie along a vector that differs from PC1. Without knowing a priori whether axes of variation among distinct groups are similar, it is impossible to know the degree to which an ordered axis plots approach will yield misleading results.