Description:
<p>It has been repeatedly argued that inclusion of ecological characters into the data set in phylogenetic analysis and subsequent derivation of an evolutionary ecological hypothesis from the resulting tree leads to tautological underestimation of the number of homoplastic character-state transformations. As it is argued here, these characters should be included in the tree-building character matrix because the null hypothesis ("the analysed characters are historically contingent and need no ad hoc adaptive explanation") is either equally strong or stronger than before their inclusion. The hypothesis that the Caribbean insular red species of Dysdercus (Hemiptera: Pyrrhocoridae) form two mimetic complexes is tested against data sets including or excluding ecologically meaningful chromatic characters into the total data set.</p>