If homologies could be recognised as such, then the relationships between species could be inferred from their shared homologies. Unfortunately, however, homologies and analogies cannot always be unambiguously distinguished in practice. The risk of confusion is especially great when closely related species are compared, because similarities in their morphology and ecology make the parallel evolution of analogous features in separate lineages quite likely. As with other statements concerning history, homologies must themselves be inferred. In some cases, this may seem easy enough. Figure 2 shows the structure of a human arm, a bird's wing and an insect's wing. We readily recognise the first pairing as being homologous and the second as being analogous, but why?
What aspect of the human arm and the bird's wing would suggest that they are homologous?
In spite of the differences of their superficial form, they share the same basic construction: corresponding bones, with the same spatial relationships, though with differing proportions, may be recognised (indicated by different shadings in the figure). The similarity of the wings of the bird and the insect, in contrast, is only superficial (reflecting their common adaptation to flight); they are of markedly different construction, the insect's wing, of course, having no bones at all.
Thus the mode of construction seems to offer a clue, and a useful concept in this respect is that of the information content of features. The more numerous the points of resemblance between structures being compared, in terms of the elements making them up, their positional relationships with respect to each other and their pattern of development, the more likely they are to be homologous. In other words, there would be an improbably large amount of detailed similarity to explain away as coincidental convergence. The appendages in Figure 2 are all features of high information content, and so the numerous structural similarities of the first pairing strongly imply homology, while the lack of them for the second pairing implies analogy.
Many other examples are less easy to resolve, however, and continue to create systematic problems to this day. This is a common problem with many fossil taxa. Many of the morphological features of simple fossil shells, for example, have a very low information content, and so homology and analogy are readily confused.
Other approaches to the problem of distinguishing homology from analogy can be adopted, but these vary according to the different systematic methods, which must be considered.