[Above: Photo of the type specimen of Juravenator under UV light. From Chiappe & Göhlich, 2010.]
This month, Luis Chiappe and Ursula Göhlich published the first real English-language follow-up paper on the small German compsognathid Juravenator starki (named for its discovery at Stark Quarry near the Jura mountains, which also gave their name to the Jurassic period). You may remember this critter causing a stir when it was first described in 2006. Juravenator was the first non-avialan theropod found from the limestone deposits of Germany and France (which have also yielded Compsognathus and Archaeopteryx) to preserve clear, fairly extensive impressions of its skin and other soft tissue.
The specimen was discovered in 1998, but as far as I know news of this discovery first hit the Internet back in 2001, by way of a German-language news story reported to the DML. The new fossil was nicknamed "Borsti", from the German borstig, meaning "bristly." As some early-release photos show, at this point much of the fossil was not yet prepped, and the bulk of the skeleton was still encased in rock. I remember getting my hands on some photos from the early 2000s showing only the skull exposed, with no trace of soft tissue. Nevertheless, being a compsognathid, scientists fully expected that in life, this animal would have been covered in short, bristly stage 1 or 2 (in Richard Prum's model of feather evolution) protofeathers like its close relative Sinosauropteryx prima.
By 2006, the whole skeleton had been exposed, along with unexpected soft-tissue traces. While limestones from this area are world famous for their preservation of feathers, this is usually limited to the large, vaned feathers present on the wings and tails of aviremigian birds (those with feathered wings like Archaeopteryx). Small theropods like Compsognathus usually preserve very little, if any, soft tissue traces, and even the down-like or proto-feathery body covering of Archaeopteryx is only very rarely preserved, and then only as the faintest wisps in the rock. For this reason, even though the two known specimens of Compsognathus itself didn't preserve any feathers, it wasn't necessarily scaly all over (only a few hints of possible scales have been noted from the tail of one specimen, and even interpretation of those has been ambiguous).
So, when Göhlich and Chiappe described Juravenator, they may have been surprised to find extensive and well-preserved soft tissue surrounding the tail and part of the legs, showing very clear impressions of small, bumpy scales like those known of more primitive theropods and most sauropod and ornithischian dinosaurs. [See photo at right, from Chiappe & Göhlich 2010]. This caused a bit of a scandal, and rendered the name Borsti ironic: here was a specimen which phylogenetic bracketing methods predicted feathers, but the prediction failed. A number of explanations were offered for this. It could be that our phylogenetic analyses were off: that is, Juravenator was not a compsognathid at all, but something more primitive, having arisen before the origin of feathers. Alternately, since impressions were known only from the tail, it could have been feathered elsewhere on its body.
The first suggestion was complicated by the fact that the relationships of primitive coelurosaurs are notoriously poorly understood (some later analyses even found Juravenator to be more advanced than compsognathids), not to mention that the only known specimen came from a juvenile, so testing its relationships are a tricky proposition to begin with. The second suggestion, that Juravenator was only partly feathered, sounded a bit like special pleading given that there was only absence of evidence to go by.
The first hint that the second explanation may have been correct came in a little-known German-language follow-up paper published by the original authors later in 2006 in the journal Archaeopteryx. Apparently, this paper reported that, on closer examination, very faint, thin impressions of some kind of filament were present on the top edge of the tail. But that was all we had to go on until this month, when the complete osteology of the specimen was published.
Examination of the specimen under UV light (performed by H. Tischlinger, one of the co-authors of the paper in Archaeopteryx) has revealed more soft tissue than reported in the description. Additional impressions of scales can be seen under UV on the snout and lower legs, as well as the visible-light impressions on the tail. This new paper confirms the 2006 reports of proto-feather-like filaments on parts of the tail. As expected for these deposits, the feather remains are very poorly preserved, and only the tips are evident. But their size and arrangement seems to closely match those of Sinosauropteryx [see diagram at left, from Chiappe & Göhlich 2010]. Additionally, these impressions lie above the level of the clear, in-tact scale impressions. Impressions of internal tissues, including what may be collagen, can also be seen under UV below the scales and between the vertebrae. This makes the standard interpretation of the filaments as frayed collagen fibers by the birds-are-not-dinosaurs crowd pretty much impossible.
So what does all this mean? Clearly, the second explanation for the surprisingly "featherless" Juravenator seems to have been correct. As the authors note, it's possible that at this early stage of evolution, feathers and scales co-existed across the body of dinosaurs like Juravenator, and possibly even Sinosauropteryx and Dilong, where no scales are preserved but feathers are still found only in certain parts of the body. So far, this kind of co-existence of widespread scaly skin with fringes of feathers has only been known in the ornithischian Psittacosaurus but, they point out, it's not inconsistent with theoretical models of feather development and evolution. [Accompanying image: my life restoration of Juravenator starki from 2006. I'll have to update this to reflect the position of filaments on the tail.]
Of course, it is also possible that explanation #1 is also correct. The authors noted that compsognathids have sometimes been found to be an evolutionary grade, not a natural grouping. In studies which have found a natural, monophyletic Compsognathidae, only a few species (usually of Compsognathus, Sinosauropteryx and Huaxiagnathus) have been included in the analysis. More testing, with more included taxa, are needed to suss out where on the dinosaur family tree other supposed compsognathids belong.
Goehlich, U.B., Tischlinger, H., and Chiappe, L.M. (2006). "Juravenator starki (Reptilia, Theropoda) ein nuer Raubdinosaurier aus dem Oberjura der Suedlichen Frankenalb (Sueddeutschland): Skelettanatomie und Wiechteilbefunde." Archaeopteryx, 24: 1-26.
Chiappe, L.M. and Göhlich, U.B. (2010). "Anatomy of Juravenator starki (Theropoda: Coelurosauria) from the Late Jurassic of Germany.Neues Jahrbuch für Geologie und Paläontologie - Abhandlungen, 258(3): 257-296.