protofeathers

[Alan Brush <brush@uconnvm.uconn.edu>]

The description of Beipiaosaurus has rekindled discussion on protofeathers. Another bit of excitement in an already wonderful summer.

When the term protofeather was introduced into the mix regarding the integumentary structures on the fossils from the Yixian formation it focuses on Sinosauropteryx. But there were clearly wider implications. The structural evidence indicated, at least to my eye, that these filaments were epidermal, and consisted exclusively of single, unbranched structures. Without any additional morphological evidence, such as branching, it was impossible to say that they were modern feathers. I deal specifically with this problem in my Dinofest manuscript. In order to establish homology with modern feathers (which includes those found on Archaeopteryx) additional information such as internal structure (eg are filaments hollow as are the parts of feathers), were they produced in a feather-type follicle, and did they consist of feather beta-keratin (or some alternative). The implication was that while there are feather structures that resemble these filaments (eg single barb), we knew nothing of their production and growth or their chemical composition. However, they may represent a morphological precursor produced by the same machinery, and consisting of the same material as modern feathers.

We also have evidence that once the machinery existed the entire array of modern feather morphologies, and by implication function, evolved very rapidly. Changing feather structure appears to be simple. There is no direct way to establish the morphology (eg downy, pennaceous etc) of a primitive feather. When we have feathers in the fossil record essentially all modern types are present. It is difficult to tell which morphology is derived, and which is primitive. But, all are made in precisely the same way and consist of exactly the same material, with similar supramolecular structures (eg beta-helical filamentation patterns).

There is no doubt in anyones mind that these structures extended beyond the body surface and were subject to wear. It is possible that in the early stages the filaments grew continuously, unlike feathers (which grow only during a short part of their cycle, and then are replaced annually or more frequently by molt. Molt, by the way, is also another characteristic of all feathers. Various stages of wear could easily produce fraying at the tips, unraveling of filamentous infrastructure, splitting, or any number of other conditions that may resemble a tuft of branches. But until they can be studied, it is not easy to distinguish among these possibilities. 

I consider the term protofeather to apply to all morphological feather types. This includes contour. all stages of down, bristles, filoplumes, etc as they are all made from exactly the same type of follicle and consist of the same proteins. It makes no sense to try to distinguish among feather generations (on a single birds) or different morphologies, as they probably differ only in the information that directs their construction. It is futile to use feather generation morphology to try to establish relationships, as once the potential to produce feathers is established, the potential to produce all types of feathers follows automatically.

The structures on Sinosauropteryx and Beipiaosaurus can certainly be referred to as protofeathers. The should not be categorized as primitive feathers. Ron Orenstein (posting of 29 May 99) makes these arguments very well. 

How modern feather morphology was derived from these starting conditions (a functional follicle, with cells capable of feather keratin synthesis) is still debated. I have manuscripts in both the SICB and Yale symposia that deal with exactly this themes. It's not simple, but it sure is a lot of fun.

Cheers,

Alan

Alan H. Brush
92 High St.
Mystic, CT 06355
Brush@uconnvm.uconn.edu