>> Whales, before getting wet, didn't just decide to lose their
>> hind limbs, sprout flukes, and head for the sea, right?
>
> Right. IMHO stem-cetaceans provide a good analogy, because
> raoellids and pakicetids have both been interpreted as terrestrial
> quadrupeds that headed to water only occasionally, such as a
> refuge, or to feed on fish. Raoellids and pakicetids show minimal
> adaptations to an amphibious/aquatic lifestyle. The limb bones of
> both groups are pachyostotic ('osteosclerotic'), a trait that has
> been tied to increased submersion by these animals, with the
> heavier bones providing ballast. In all other respects, the
> postcranial skeletons of these stem-cetaceans are thoroughly
> terrestrial and cursorial.
> To return to theropods, it is possible that the ambiguous
> scansorial characters seen in the pes are indeed nascent arboreal
> characters in otherwise dedicated terrestrial/cursorial animals.
> I tend to think so. But I admit that this assertion comes from
> viewing these characters through the prism of what came after.
> Arboreality had to start somewhere, and the elaborate plumage
> suggests to me that the paravian integument had a big headstart on
> the osteology.