Feathered vs. scaly skin in dinosaurs

The recent discovery of the compsognathid Sinosauropteryx prima fossil with feather impressions once again started debate, on the Dinosaur Mailing list, of whether dinosaurs should be expected to have feathers. For most of the history of dinosaurs, most paleontologists and lay-people alike concluded that dinosaurs, as reptiles, must have scaly skin as modern reptiles have scaly skin. However, this argument has never made sense to me.

Extinct elephants, with the exception of the "wooly" elephants of the ice ages, are almost always reconstructed with naked skin since modern elephants have naked skin. Birds, not lizards or other reptiles, are the dinosaurs' closest living relatives; indeed, under modern classification schemas, birds are dinosaurs. I would think, therefore, it would be logical to assume dinosaurs, except the large forms, as indicated by skin impressions and large modern mammals, would be feathered.

The argument that dinosaurs are reptiles and therefore must have scaly skin makes even less sense when you consider that a major branch of reptiles, the birds, have scales only on their feet (see the DOL Dictionary for the definition of the Reptilia).

Anybody who has read any of the many articles I've written for Dinosauria On-Line that touch on this subject knows that I have concluded that, at the very least, the small, bird-like dinosaurs had some sort of integument, probably feathers. Unfortunately, I appear to be in the minority. Most scientists and lay-people alike seem to continue to follow the illogic of applying living-relative analogues to some animals, like elephants, but not others, like dinosaurs, and thus argue that dinosaurs in general should be assumed to be featherless.

Gregory Paul, author of Predatory Dinosaurs of the World and world class artist, has this to say about it (with a brief comment on the implications of feathers for dinosaurian endothermy):

I have over the years been concerned that paleontology is not always practiced as a rigorous science (some nonpaleontologist scientists have expressed the same concern to me). The issue of feathered dinosaurs well illustrates the problem.

Back in the mid 70's Disco was King, and at that time the [first suggestions were presented] that the probable ancestors of birds, the possibly endothermic theropods, were feathered. This inspired some artists, including me, to illustrate small dinosaurs with some form of pelage [insulatory covering], usually some form of feathers. This is were the science problem began.

A number of paleontologists vigorously objected to restorations of feathered dinosaurs. Sometimes they did so derisively and harshly -- odd how emotional people can get about a little dinofluff. The basic scientific argument was something like this. Dinosaur "mummies" show that dinosaurs had scales, like ectothermic reptiles, not feathers like endothermic birds. No feathers had been found adorning small dinosaurs. Parsimony therefore favored scales in dinosaurs; feathers were too speculative. Some researchers had somewhat more complex and subtle arguments as to why

This argument was always false. The scaly skin was and is limited to large dinosaurs. If mammals were extinct, and only the skin of fossil elephants and rhinos were preserved, then some form of insulation would be ruled out in small mammals by the above logic. Parsimony would favor scales in small dinosaurs only if both their ancestors AND descendants were scaly. As it is, with scaly ancestors, and feathery relatives, parsimony was neutral on the issue. The question could be answered only by direct fossil evidence. The evidence was also neutral because no integument had been clearly preserved on any small dinosaur. Example -- claims that the lack of feathers on the famous Compsognathus specimen contradicted their presence were spurious because scales were not found either! Nor are body feathers present on all but one of the Archaeopteryx specimens from the same sediments. Viewed objectively, restoring Compsognathus and other small dinos with feathers, scales or nonscaly skin was equally speculative and plausible! Yet the majority opinion always nonobjectively presumed that scales were less speculative than feathers. I took a gamble and consistently restored small theropods with feathers, a bet that seems to have paid off at least in part.

What do we currently know about the integument of small dinosaurs? I have only seen a rather poor color photo of the Chinese theropod [Sinosauropteryx]. However, Currie has seen the original and says that it clearly has feather-like structures similar to those of birds from the same sediments. Although the feathers are entirely peripheral to the body on the slab, this is typical of bird fossils. A full body covering is very possible if not probable. There is some evidence that the structures on the underside of the tail base are intermediate to scales and feathers. There may be more specimens from the same sediments, [and] certainly more can be expected [which] will help solidify the data base one way or another. The rest of the discussion assumes the specimens do have short ratite-like feathers, but no contour or long feathers.

The E. [Early] Cretaceous theropod appears to be broadly similar in form to L. [Late] Jurassic Compsognathus, with short arms, and a long rather heavy tail. If so it is a basal tetanuran, much less advanced than more bird-like theropods which show flight adaptations.

A partial small theropod from the E. Cretaceous of South America was reported to have naked nonscaly skin in Nature (1996, 379:32). No clear photos of the skin surface were published, and it is not clear whether the outer skin layer had been lost. More data is needed.

Also in Nature (1994, 370:363) it was reported that the E. Cretaceous ornithomimid Pelecanimimus has an integument of uncertain nature. Again more data is needed.

The only small ornithischian skin reported was a nonscaly, punctured surface on Thescelosaurus earlier in this century, but the integument was not figured and is not accessible.

The combined data suggests but does not prove that feathers evolved in terrestrial theropods well before the evolution of birds, at least by the L. Jurassic and in tetanuran theropods. How much earlier feathers may have evolved is not clear, so Triassic and earlier Jurassic theropods may or may not have been insulated. Therefore artists should feel free to restore Coelophysis with or without feathers. If someone tries to tell you otherwise just subtly roll your eyes, or smile enigmatically yet knowingly - its what I used to do. Feathers are highly probable in the advanced bird-like theropods. What small ornithischians were wearing is up for grabs.

The probable presence of feathers in at least some small theropods has important metabolic and thermoregulatory implications. Recently it has been asserted that the presence of feathers does not necessarily indicate endothermy, contrary to a long opinion otherwise. This is based on [several] arguments. [One argument is that] early birds could only have powered flight with the small reptile-type muscles their small sternal plates were able to support. This is false because some modern flying birds have flight muscles much smaller than the avian norm, and early bird arms and shoulder girdles were large enough to support them. Another argument is that early bird bones show they grew slowly like reptiles. However, the thin walled bones of the fossil birds do not preserve the period of juvenile growth, which may have been rapid (the growth rings only indicate that the birds continued to grow slowly as adults). Last, some birds bask. This, however, only shows that endotherms bask with insulation. The observed temperature rise is too modest for reptiles.

Which brings us to the point that all the many thousands of insulated tetrapods are endothermic. Even furry insects tend to be endotherms. Insulation prevents ectotherms from rapidly absorbing the large amounts of heat they need from the environment. This problem is so serious that even desert reptiles cannot use insulation as a solar screen. Only endotherms that need to retain body heat are insulated. Ergo, insulation is always excellent evidence for endothermy. Also for resting metabolic rates above the reptilian level. Therefore, ecothermy is effectively falsified in insulated dinosaurs and birds.