Re - Feathers and Flight

[J M V Rayner <J.M.V.Rayner@bristol.ac.uk>]

Like Kevin I have preferred to lurk through much of this discussion.
There's been some great stuff here. Richard, you needn't think your
display hypothesis is lost in the literature: I always cite it as one of
the two good models for the evolution of a feather indendent of the
evolution of flight (the other is Dyck's proposal on feathers for water
repellence).

One recent thread worries me, and that is the problem of the tail.

It has been argued recently that a long tail is a bad design for a flying
bird and therefore it is a poor strategy for a primitive flier. I do not think
this argument should be accepted uncritically, for several reasons.

First, the slender-wing theory which led to this result is only a prediction,
which is not yet tested in real or model birds. In particular, it neglects the
effects of wing flapping (and therefore of oscillating air currents in the
region in front of the tail), and it neglects the presence of the body,
although the airflows generated by the body will determine the flows
at the (hypothetical) apex of the tail, which are the critical influence
on the forces generated by the whole tail.

Second, is the evidence from modern birds.  Long feathered tails have
evolved in various groups of modern birds, always associated with life or
roosting among clutter (good examples: magpies, some cuckoos (coucals,
roadrunner, etc., cormorants). In each case the tail is associated with
short, rounded wings with a low aspect ratio. In strict aerodynamic terms
such a bird might not be the `best designed for flight' - it may for
instance be relatively uneconomic - but there is no doubt that it is the
best design for inclusive fitness for birds exploiting certain habitats. I
have argued that the reason for this is that birds with long wings cannot
live in confined spaces for risk of damage, and the large tail area
certainly _does_ compensate for the reduced lift from the small wings. 

At first sight this seems to have a lot to do with Archaeopteryx. Heinroth
showed years ago that the Archaeopteryx body form is very similar to that
of some tropical cuckoos. This has often been cited as evidence that Arch.
was arboreal, with similar habits to these modern animals. This is
persuasive, but it relies on the implicit (and false) assumption that
Jurassic woodland/forest had similar topography to modern, and offered
similar food sources, _and similar avian community structures_. I fear
this close parallel suggesting an arboreal Arch., and therefore an
arboreal origin of flight (note the non-seq.) may be illusory.

However, there is another advantage of short, rounded wings. They are
found also in galliforms and some gruiforms which make rapid climbing
take-offs, often from the ground, but are not always associated with
modern arboreal habitat. For aerodynamic reasons, short rounded wings are
effective at producing large climbing forces at rather low flight speeds. 
That sounds a familiar requirement for a cursor! Indeed, if one was
designing a small bipedal cursor for flight, needing a rapid take-off, this 
is pretty well how one would do it. 

There is, however, more to the tail perhaps than this. In a modern bird,
in flight, just as in an aircraft, the tail is primarily a lifting and control
organ. There are very few birds for which the tail is totally unimportant
mechanically in flight, and these all appear to be highly derived. The mean
force the tail generates in flight is primarily upwards. Since the
(upwards) force from the wings is in front of the centre of mass, the bird
can be in a mean pitching equilibrium. But consider a running biped. In
theropods, like modern kangaroos, bipedal lizards, etc., the tail is also
an organ of stability, but stabilizes the animal about the pelvis: the
centre of mass is in front of the point of balance. It is the _weight_ of
the tail which is important. It is not surprising that in many of these
animals we find substantial bony structures in the tail, often distributed
towards the tail tip. (Before anyone complains, I know the tail also
functions as attachment for some of the pelvic and hindlimb musculature, but
that doesn't explain all of the weight in the biped tail: think of it that
the presence of this musculature facilitated the evolution of the sort of
balance required in bipeds, and therefore is a partial explanation for the
prevalance of bipedality in higher tetrapods.)

The sum of this is that, in the course of the evolution of flight
from a theropod to what we would now recognize as a bird, the function of
the tail had to switch from a downwards _weight_ force to an upwards
_lift_ force. We don't know precisely how or when this happened, but Arch.
probably gives us an important clue: one of its unique features is the
bony and feathered tail. It is conceivable that aerodynamically this
structure was more-or-less neutral, with the weight of the tail vertebrae
balanced by the lift from the feathers. We know from the `next' birds in
the fossil record (various small early Cretaceous species) that after
Arch. selection acted primarily on the flight organs, and we see major
modifications to the wings and tail. In particular, the tail vertebrae
have reduced and fused to the pygostyle. This is exactly what we would
expect if the limited static stability of Arch. as a weakly-derived
flier were the foundation for rapid improvement of flight capabilities,
including dynamic control by a combination of a lighter, more manouevrable
tail and a more fluent forelimbs. 

Some caveats to this argument:

1. It says little about whether Arch. (or the proto-bird) was arboreal or
cursorial. Much the same logic would apply to both cases. It remains more
closely compatible with the cursorial/theropod relationships of early
birds which Kevin, among others, has put forward.

2. It makes it more tempting to claim Arch. as _THE_ first bird, the
ancestral bird, or whatever. This is a dangerous and potentially very
misleading misconception. Arch. is the oldest and least-derived bird
(member of Avialae) currently known and accepted. Selection after Arch.
was apparently rapid, and presumably selection before it was as well.
Arch. may be secondarily specialized compared to the true (conjectural)
proto-bird (arguing this to be the first animal able to use a feathered
wing to be able to support its own weight in air by flapping wings). (NB:
a weaker definition that modern flight in birds. NB also I am not calling
it Proto-avis.) It may have its bony feathered tail as a derived character,
for instance, which proto-bird and later radiations did not share. If it
was well adapted to its environment, it may have derived characters which
the available material does not help us to identify.

3. The comparison with bats is useful here, and does shed some tentative
light. Bats had tetrapod ancestors, and therefore different organization
for supporting the weight in static stability. They did not therefore need
a large weighty tail (Cynocephalus (Dermoptera) does not have an
extended tail, although the homologous flying squirrels and phlangers usually
do) in the non-flapping ancestor. Hence the different organization of the
bat pelvis and tail region. Suggestive, but perhaps no more, that unlike
bats birds did not have an arboreal origin.

4. Arboreal vs cursorial. These are often produced as direct opposites.
This has become a source of much confusion. They are variously (and
differently) hypotheses about type of habitat and behaviour, and about
types of locomotion. In modern small animals (<1 kg) skeletal design for
arboreality is not always evident, for selection for mechanical adaptation
on bones, skeletons, etc. to optimise locomotion is not strict when body
mass is low. In mechanical terms, it is more constructive to think of the
origin of theropod flight as being _with_ or _against_ gravity (downwards,
probably from gliding), or upwards, probably from running/jumping/short
glides). The question of the habitat or environment in which the primitive
flight adaptation evolved is in this sense secondary to the fundamental
mechanics. Not to say it is important: the habitat facilitates the
adaptation. This is the reason for emphasizing the small size: small
tetrapods have much greater flexibility in habitat and environment use than
do larger ones. The proto-bird's ancestor may have used what was for it an
unusual habitat, but this of course is so speculative as to be a useless
hypothesis. It does however emphasize Kevin's point that comparisons with
modern taxa or faunas are unlikely to be very informative.

Jeremy Rayner


+++++++++++++++++++++++++++++++++++++++++++

Dr Jeremy M. V. Rayner 
School of Biological Sciences
University of Bristol
Woodland Road
Bristol BS8 1UG   
U.K.

tel. 0117 928 8111, messages 0117 928 7476, fax 0117 925 7374

e-mail J.M.V.Rayner@bristol.ac.uk