I, for one, think that basal paravians and avialans were not highly
arboreal. But there are at least three features that suggest that
something other than running fast on the ground had become a new
priority in the biology of basal paravians.
One is the very long leg wings, which Xu hypothesized were in
conflict with fast running. This is hard to test.
Two are the longer penultimate phalanges, seen also, I believe, in
Xiaotingia.
Three is the evolution of the reversed hallux. It is at least
slightly reversed by Jeholornis and more so, as well as longer and
stronger, by Changchengornis. I do assume that this indicates that
being good at gripping branches became important to their survival
early in the evolution of the modern flight apparatus.
I guess then a testable prediction is that, if the forms in which
incipient flight evolved were not using tree habitats at all (if
strict ground up is correct), the 1st digit should begin its long
sequence of adaptations AFTER the flight apparatus is assembled.
Instead I interpret the descent of the hallux to lie close to the
other toes in Microraptor, and more so in Epidendrosaurus and
Xiaotingia to possibly indicate the first steps of selection for
branch walking. After these incipient stages there is a long and
complex radiation of avialans with increasing rotation of the
hallux.
The small body size, large wings, showy tails, feathered legs, long
penultimate phalanges, and descending hallux of basal paravians could
be seen as consistent with animals that were using tree habitats some
of the time and adapting to those habitats. If the descending hallux
and penultimate phalanges have nothing to do with branches, and small
animals can run just fine with long wings on their legs, then this is
not evidence for branch walking at all. But maybe then the burden
lies with others to explain the very long and diverse series of
morphological changes in Mt I.