The argumentation that interprets deformities as atavism is inconsistent.
Deformities taken as indications of an assumed phylogeny (that is, interpreted as atavism) if they demonstrate similarities with presumed ancestors of the organism in question, are in fact used very selectively. Almost all deformities, however, cannot be interpreted as evolutionary returns. Deformities such as forked ribs, hare lip, polydactyly, the development of two heads or the presence of a fifth leg, for example, are for a certainty not indications of earlier phylogenetic stages. Interpreting deformities as atavism is thus only possible if a particular evolutionary development is already presupposed. From this is then concluded which deformities can be seen as returns to earlier stages. As the matter to be proved is thus already assumed, atavism cannot be regarded as proof of phylogeny. The fact that a small number of deformities are reminiscent of similarities in assumed ancestors of a specific organism is not particularly remarkable, and due to the similarity of many developmental forms it is also not surprising. Added to this is the fact that when atavisms are scrutinised more closely, they are found to be in no way similar in all aspects. The argumentation with atavistic structures themselves is again also selective. For instance, there are horses that have two-toed feet (Fig 201, see above), yet amongst horse fossils only three or four-toed varieties are known.
Pursued consistently, the atavistic interpretation would lead to ridiculous conclusions, as the following example shows. Four-winged fruit fly mutants (Fig. 57, below right) are regarded as an indication that the normally two-winged insects (diptera) have descended from four-winged types. The development of four wings is interpreted as atavism. There are, however, also fruit fly mutants with four halteres and no wings – a pointless construction which certainly cannot be regarded as an indication of phylogenetic ancestors.