You probably have never heard of Janavis finalidens. It was a bird with webbed feet and a body vaguely reminiscent of a wild hen. It would have been the size approximating to a grey heron and had a mass estimated to be 1.5 kg. I use the past tense because it lived in the late Cretaceous, so birds had evolved away from the therapods well before the extinction. This bird was, in shape, very similar to modern birds, which is hardly surprising because flight puts limits on them, but there is one notable difference: the beak was lined with teeth. We do not know a lot about them, in part because their bones are rather fragile and more difficult to fossilize, but maybe also because they are not so spectacular as the monster dinosaurs of the time. This particular example was found a couple of decades ago in a quarry, and when it was worked out what it was, it was filed.
However, more modern equipment, specifically micro-computed tomography, has re-examined the samples. Originally, they thought they had a handful of bones from the spine, wings, shoulders and legs. However, one of the bones they thought had been a shoulder bone was a pterygoid, a bone from the bony palate of the skull.
Most current birds belong to a group called neognaths, which means “new jaws”. The key bones here are mobile, and they allow the birds to move the upper beak independently of the skull. There is a small group of birds, the emu, cassowary, ostrich, kiwi and tinamous (47 species, ground dwelling, but some can fly) have the bones in the upper palate fused together. These are also called paleognaths, or “ancient jaws”. You will probably suspect from this naming that it was believed that birds originally came with these fused jaws, but most subsequently evolved the ability to move the upper beak. In this context, non-avian dinosaurs also have fused palates, and the last common ancestor of all modern birds lived some 80 million years ago, so it would be reasonable to assume that it had a fixed palate like the other dinosaurs. Unfortunately, this is one of those theories that is hard to test because the small delicate pterygoid is usually missing from the fossils.
However, a recent article in Nature (Benito et al., vol 612, pp 100 – 105) indicated that Janavis‘ pterygoid “probably formed part of an unfused bony palate”. That means the upper beak was probably mobile. Note the uncomfortable “probably”. The resemblance of the pterygoid to that of modern chickens now suggests that the mobile upper beak evolved first, and the fused beaks arose later. That, of course, raises the question, how did it evolve, and why did some birds revert to the fused palate.
How the beak functions is crucially dependent on the bones of the upper palate. By unfusing it, it increases the flexibility of the beak and improves the use of the beak. However, fused palates are not necessarily a drawback, and might give beaks of larger birds additional support. For the kiwi, the beak is extremely long compared with the bird, and fusing the upper beak to the skull might give it more strength as it probes logs for food (often grubs in decaying logs). It might also be of interest that these birds, being flightless, tend to get most of their food from the ground, including plant material, but then again so do hens.
Accordingly, if you are concerned with the evolution of birds, admittedly not a common concern, you now have a problem, and the question is, how do you solve it? One way is to find plenty of fossils, but the difficulty is first, they are rare, and secondly, we have many samples from times when both were present, including now. How do you know you are not being misled? An important aspect of science is that once you have a reasonably well-defined problem and a possible solution, you can arrive at ways of testing it. One of the peculiarities of evolution is that as an egg grows to the adult, it often gives clues as to the evolutionary path it took. The most obvious example is the frog, first going through the tadpole stage. In the case of modern paleognaths, one approach being considered is look at their development stages. If there are differences, this would be a clue that the trait arose independently more than once.