Convergent evolution... tell me more

Do you mean convergence, or should we say parallelism?

In the discussion of convergence it is usual to draw a distinction with parallelism, which is defined as the evolution of similar features in groups that are not directly related to each other but have a shared ancestor.

Thylacosmilus atrox skullConsider, for example, the striking convergence, as we’ll presently call it, of the sabre-tooth mode, most famously seen in placental felids (such as the smilodontids) and the marsupial thylacosmilids. Both are equipped with dramatic stabbing canines, but one has to observe that not only are they both mammals but they have also inherited much the same dentition from a common ancestor that lived some time in the Mesozoic. Is it too surprising, therefore, to see repeated evolution of the sabre-tooth? Parallelism can be said to apply to the placental felids, in as much as they are a closely related group in which the sabre-tooth mode has evolved independently several times during the Tertiary. The case of the thylacosmilids may, however, be more ambiguous, depending on the nature of the common mammalian ancestor and the exact route of evolution taken to produce a thylacosmilid (as against any placental sabre-tooth). As far as we are aware this question has never been studied in detail. We need to understand how the associated functional complex that permits the sabre-tooth mode of life (which recall involves a great deal more than the giant canines, including musculature, and behaviour) evolved in comparison to the placentals? What are the historical pathways that link the thylacosmilids to the early marsupials, and how do these compare in detail to the equivalent pathway in the placentals?

Smilodon fossilIt is apparent, therefore, that distinctions between convergence and parallelism even in quite closely related groups may be more fuzzy than expected: from the point of view of evolution the pathway is just as important as the end-point. The usual approach taken in attempting to distinguish between convergence and parallelism is based on identifying the degree of common ancestry, a method often adopted by cladists. This, however, turns out to be less helpful and in any event is ambiguous because ultimately all life shares a common ancestry. We would suggest that rather than agonizing as to whether it is convergence as against parallelism is hardly worth the effort: frankly it seems to be a rather sterile exercise.

Convergence and parallelism form a continuum and even the former has limits: the octopus and vertebrate eyes are convergent, but that may be almost as far as comparisons can go. We do find such camera-eyes in box jellyfish, but to find them in plants or bacteria would be most surprising, but if found would be simply more extreme examples of convergence (notably, a structure much like a camera-eye has evolved in one group of dinoflagellates). So to the first approximation if you wish to dither between labelling a shared trait as convergence or parallelism the best procedure is to see where the examples are placed in terms of relative phylogenetic distance. This follows the lucid and sensible analysis offered by Kester and Berger (1977). Further attempts to pinpoint a definition may be less useful than actually documenting them!

Go to the top of the page