pterosaurs in trees:

[Nick Longrich <longrich@phoenix.Princeton.EDU>]

        Much has been made about the length of pterosaur wings, e.g.
Rhamphorhynchus and Pteranodon, and it has been claimed that pterosaurs
were too long-winged to work in forests. I doubt this is true for a number
of  reasons. For one, reasonably large-sized birds and bats live in
forests today. Harpy eagles and monkey-eating eagles, for example, the
Harpy having an 8 ft. wingspan, and the large flying foxes are big- up to 
6 feet. Granted, these may not generally fly through the canopy, but my
understanding of the
world's largest "microbat" a beastie with a 3 ft. wingspan, is that it
does when hunting for birds and mammals at night. And I have seen some
pretty big microchiropteran fruit bats (Artibeus) which are probably
pretty close if not within the size range of a small pterosaur (about a
foot or a little more in wingspan) and seen them caught in dense tropical
forest. 
        Was the wing really so long and narrow? Why not attach it to digit
#5? that would greatly increase your wing surface. I am not arguing all
rhamphorhynchs did this, but some could have and it would have created a
much broader wing. Also, we may not be able to generalize our known
pterosaurs to forest pterosaurs- the marine forms we see may be the
equivalents of long-winged frigatebirds, pelicans, and albatross.
        As for grasping branches: looking at advanced pterosaurs may not
be the best way to understand their evolution. Geococcyx, the roadrunner,
may be pretty darn effective on the ground, but it is descended from
arboreal forms, which is why they have two toes pointing backwards.
Perhaps these forms did not live in trees but that doesn't prove that
their ancestors didn't. Pterosaurs do lack a hallux, but so do bats.    
        Cursoral theories are still plagued with problems. One criticism I
was offered was : "take your hypothetical cursorial bird, leaping into the
air. At any given stage on its evolution to flight, it will be better
adapted to glide downwards than go upwards". 
        Gliding downwards is just more energetically efficient, too. It
hasn't been demonstrated to my knowledge that the flutters and leaps of an
cursorial protobird would be anything but wasted energy. On the other
hand, it has been demonstrated that gliding is an effective and efficient
way to move through the forest (and by Mother Nature, not just by math),
and once you have developed wings, you need only add thrust: the
downstroke- to stay aloft longer and longer. 
        About the only exception to all this is the flying fish. I think,
however, that they are not directly applicable, at least, they can't be
taken at face value as demonstrating how you can leap upwards and develop
flight from the ground. The reason here is that flying fish may find
moving through the air more efficient than moving through the water at
early stages in their evolution because of the great resistance offered by
water. 
        For these reasons, probably all three evolutions of vertebrate
flight began in the trees.