Paul's comments about Feduccia's bird digit paper

Gregory Paul

I have now read the Burke-Feduccia Science paper (Science vol. 278, pg. 666) on the alleged nonhomology of theropod and bird hands -- a subject I have worked upon (1984a) -- and the supporting viewpoint by Hinchliffe (Science vol. 278, pg. 596). As I suspected, it gives no reason to challenge the fact that birds are as much glorified dinosaurs. Instead, we are seeing the last gasps of a dying hypothesis.

To start with, the fossil evidence clearly shows that the outer digits are lost in theropods, leaving only I-II-III in avetheropods, with I always being a strong, big clawed thumb weapon (except when the arm is hyper-reduced). In herrerasaurs V is a wee splint and IV is just two small bones. There is no practical way to make IV into a strong finger with five bones including a large claw, while losing the big thumb, etc.

The anti-dinosaur group keeps stressing that digits are normally reduced symmetrically, and that this should be true of birds as well. Of course in the next breath they acknowledge that dinosaurs lost only the outer fingers, so there is no reason the other group of bipedal archosaurs, the birds, could not have done the same.

The basic problem is that even in the earliest bird embryos there are only four digits, the fifth is entirely lost. If there were five digits and we could watch which ones were lost there would be no problem. As it is there is currently no way to reliably number the digits in bird embryos. Doing so requires a number of untestable assumptions.

At the same time, we do not have any avetheropod embryos to examine. It is quite possible that they grew their fingers in exactly the same manner as baby birds, with the well developed digit opposite the ulna being III, rather than IV as is common in other tetrapods. After all, in adult avetheropods the digit opposite the ulna IS number III.

Why the avetheropod-bird clade would initially emphasize the development of III rather than IV is obvious. Loss of digit I in even the embryos would leave a big gap between the pisiform in the side of the wrist and metacarpal IV, unless the other digits shifted laterally. So IV would be were V was, and III would be were IV was. If the shift is not made from the get go, it is only going to have to occur at some point later. Also, IV will be entirely lost. To follow the usual tetrapod finger growth pattern would require IV to grow large in embryos, then be completely lost later on (Burke & Feduccia say that some lizards sharply reduce the size of IV, but it is not completely lost. Does anyone know what happens to digit IV in horse embryos?) This would be a waste of growth energy, and natural selection does not work to make finger buds convenient for embryologists to count, but to maximize efficiency of growth. The severe asymmetry of finger growth in theropods-birds should have forced them to reconfigure the growth pattern, so that III is initially emphasized rather than IV, and the latter is never more than a stub before it is eliminated.

The problem is that some embryologists expect digit IV to be large because it is so in animals with symmetrical finger reduction, and some want it to be IV, so they say it is, even though strong asymmetric finger growth could be expected to result in important changes in embryonic growth. As it is, there is no conclusive evidence that birds retain digits II-IV rather than I-III, there is no way to compare avian and dinosaur hand embryology, and so the problem is untestable.

Another thing the anti-dinosaur group does not have is a fossil record that in any way supports symmetrical reduction of fingers in protoavians. Of course this is because nondinosaurian ancestors of birds did not exist! There is a wonderful fossil trail of asymmetrical finger reduction in dinosaurs leading to the avian condition.

Science is always partly political, and it is important to understand the deep bias of some against birds being dinosaurs. For example, Feduccia (1996) claims that the hand of Archaeopteryx "does not closely resemble that of a theropod dinosaur." This when the [bird]'s long raptorial hand with a semi-lunate carpal block, and three gracile digits, is clearly a diminutive version of a dromaeosaur manus! Feduccia cannot point to any nondinosaur that has a hand anything at all like those of birds.

There are other errors in the B&F paper, and Hinchliffe's note. In defining Aves, B&F cite the presence of feathers, opisthopubic pubis with hypopubic cup and scapula articulating with coracoid at 90 degree angle.

The most strongly reflexed coracoids to be found outside Aves are in dromaeosaurs, troodonts and oviraptors, in which the coracoid is large, vertical, the outer surface faces anteriorly, and articulates with the anterior edge of the broad sternal plate via transverely long articulations (Barsbold 1983, Paul 1988, Russell & Dong 1993, Norell et al 1997). This is the avian condition, and the possession of a large sternal plates in these "dinosaurs" makes them more avian than Archaeopteryx!

I used to illustrate Archaeopteryx with a strongly retroverted pubis, but all the new specimens show it ain't so, the pubis was nearly vertical. The pubis was more strongly retroverted in dromaeosaurs (Barsbold 1983, Paul 1988). The pubic shafts of dromaeosaurs and Archaeopteryx are thin plates than meet at a shallow angle, a feature found in no other archosaurs.

As for the feathers, Feduccia in 1996 (under one of my illustrations) said that "there is no evidence that any dinosaur possessed feathers" which was true at the time, but then there was no evidence that they did not. He continued "feathers are absolutely unique to birds." That was an unsubstantiated opinion based on a lack of fossil evidence. If and when it is shown that theropods did have simple but true feathers then all the fuss about digit counts will fade away.

Hinchliffe dredges up the old saw that theropod arms were too short for them to be protobirds. Never mind that the arms of Deinocheirus were 2 meters(!) long, or that in dromaeosaurs and oviraptors the long arms are as long relative to the hindlimbs as in Euparkeria and other small early archosaurs [which the anti-dinosaur crowd claim are the true ancestors of birds -- ed.]. Never mind that the only long armed obligatory bipeds other than birds are theropods.

Hinchliffe then states that semi-lunate carpals are rare in dinosaurs. Actually they are known in Coelurus, dromaeosaurs, troodonts, oviraptors, and therizinosaurs. Even if they were rare, what does that have to do with the price of tea in China? That would only mean that those few theropods are the most probable relatives of birds. Notice how Hinchliffe takes an avian feature of theropods, the lunate carpal, and manages to use misleading rhetoric to turn it into a reason to think that dinosaurs are not close to birds! Using that kind of illogic, they will soon have birds allied with flying fish!

Then Hinchliffe repeats that the most bird-like dinosaurs are Cretaceous forms that appear after the Late Jurassic Archaeopteryx. Well, there is Coelurus with that "rare" semi-lunate carpal in the Late Jurassic. Troodont teeth have been identified from the same period (Chure 1994). A few dromaeosaur-like bones are known from the same time (Jensen & Padian 1989). Possible dromaeosaur teeth have been found in the Middle Jurassic (Evans & Milner 1994). Gosh, there could have been a whole bunch of little dino-birds skittering around the last half of the Jurassic for all we know. But small predators near the top of the food chain are rather scarce and difficult to preserve -- that old fossil record is spotty, after all [not to mention the fact that the fossil record is not a true chronological representation of the evolution of life -- ed.].

As the last subject points out, it is apparent that those who oppose dinosaurs as bird ancestors are not sufficiently familiar with the record and anatomy of theropods. I suggest that they compare the paroccipital process of Archaeopteryx to Dromaeosaurus, and note the extreme similarity in the twist of the process and the subrectangular distal expansion that results in a deep auditory meatus not observed in other archosaurs (Currie 1995). This is just convergence? No, Archaeopteryx is a small, flying dromaeosaur.

On a related subject, I just read Hecht's article in New Scientist, Oct. 18, on the proof that the short armed (relative to Archaeopteryx) dino-bird was feathered, and possibly secondarily flightless. Considering that it had contour feathers (that may have first evolved for flight) but was flightless makes this a good possibility. Of course, moi was the first one to argue that some Cretaceous dinosaurs were secondarily flightless (1984b, 1988).

Jonathon Woolf writes, in response to Greg's question about horses:

According to MacFadden' Fossil Horses, horses never had digit I, except in Hyracotherium's forefoot where it was a simple splint. Hyracotherium's hindfoot doesn't seem to have had either I or V. Digit V was also lost from the forefoot over time. Even in Hyracotherium, digit III is the main one. It became more so over time, and II and IV deteriorated. In modern Equus, II and IV are both represented by simple splints of bone, except in atavistic individuals where one or both grow into reduced hooves. What happens in horse embryos I'm not sure, but I would guess that II and IV simply stop growing at some early stage of development.

1. Barsbold R 1983 Joint Soviet-Mongolian Palaeont Exp 19:5
2. Chure D 1994 BYU Geology Studies 40:11
3. Currie P 1995 J Vert Paleo 15:576
4. Evans S & Milner A 1994 p 303 in Frase & Sues eds, In the Shadow of the Dinosaurs
5. Feduccia A 1996 The Origin and Evolution of Birds
6. Jensen J & Padian K 1989 J Paleont 63:364
7. Norell M et al 1997 Nature 389:447 see fig 1
8. Paul G 1984a Nature 310:732
9. Paul G 1984b p 175 in Reif & Westphal eds, Third Symp Mesozoic-Terrestrial Ecosyst
10. Paul G 1988 Predatory Dinosaurs of the World
11. Russell D & Dong Z 1993 Canadian J Earth Sci 30:2163 see figs 1a & 2a