Evolution of Birds - Adaptive Radiation of Modern Birds

Adaptive Radiation of Modern Birds

Modern birds are classified in Neornithes, which are now known to have evolved into some basic lineages by the end of the Cretaceous (see Vegavis). The Neornithes are split into the paleognaths and neognaths.

The paleognaths include the tinamous (found only in Central and South America) and the ratites, which nowadays are found almost exclusively on the Southern Hemisphere. The ratites are large flightless birds, and include ostriches, rheas, cassowaries, kiwis and emus. A few scientists propose that the ratites represent an artificial grouping of birds which have independently lost the ability to fly in a number of unrelated lineages. In any case, the available data regarding their evolution is still very confusing, partly because there are no uncontroversial fossils from the Mesozoic.

The basal divergence from the remaining Neognathes was that of the Galloanserae, the superorder containing the Anseriformes (ducks, geese and swans), and the Galliformes (chickens, turkeys, pheasants, and their allies). The presence of basal anseriform fossils in the Mesozoic and likely some galliform fossils implies the presence of paleognaths at the same time, in spite of the absence of fossil evidence.

The dates for the splits are a matter of considerable debate amongst scientists. It is agreed that the Neornithes evolved in the Cretaceous and that the split between the Galloanserae and the other neognaths - the Neoaves - occurred before the Cretaceous–Paleogene extinction event, but there are different opinions about whether the radiation of the remaining neognaths occurred before or after the extinction of the other dinosaurs. This disagreement is in part caused by a divergence in the evidence, with molecular dating suggesting a Cretaceous radiation, a small and equivocal neoavian fossil record from Cretaceous, and most living families turning up during the Paleogene. Attempts made to reconcile the molecular and fossil evidence have proved controversial.

On the other hand, two factors must be considered: First, molecular clocks cannot be considered reliable in the absence of robust fossil calibration, whereas the fossil record is naturally incomplete. Second, in reconstructed phylogenetic trees, the time and pattern of lineage separation corresponds to the evolution of the characters (such as DNA sequences, morphological traits etc.) studied, not to the actual evolutionary pattern of the lineages; these ideally should not differ by much, but may well do so in practice.

Considering this, it is easy to see that fossil data, compared to molecular data, tends to be more accurate in general, but also to underestimate divergence times: morphological traits, being the product of entire developmental genetics networks, usually only start to diverge some time after a lineage split would become apparent in DNA sequence comparison - especially if the sequences used contain many silent mutations.