[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Subject Index][Author Index]

Re: Fw: Dinosaurs and birds

By way of pre-amble: apologies to all who might have prior art for lack of 
citation. Assumes bipedal animals that have attained the ability to generate 
some theoretical minimum of aerodynamic effect with the forelimbs.

Quick re-cap of the debate (started w/ a resolved miscommunication) as I 
understand it (corrections welcome); I say ground-up flight evolution scenarios 
that pre-suppose flat terrain and advantage conveyed by increases in maximum 
speed generated by "wing-assisted" (=forelimb assisted) aerodynamic thrust are 
feasible, and even likely, due to the presence of other potential advantages 
such as acceleration and maneuverability, all on a variety of substrate 
scenarios. Mike H. says increasing maximum speed w/ aerodynamic thrust on flat 
ground is not possible because either the hind limbs will fail before the 
minimum threshold of advantage is reached, or use of the 'wings' while running 
will result in a reduction of overall speed. He is skeptical of such advantage 
factors as acceleration and maneuverability on probabilistic/relative 
efficiency grounds. I think this is a correct summation of positions...


----- Original Message ----
From: Michael Habib <mhabib5@jhmi.edu>
To: dinosaur@usc.edu
Sent: Sunday, April 8, 2007 6:00:37 PM
Subject: Re: Fw: Dinosaurs and birds

> ============================
> 1). Huh? 'Cant-go-any-faster-than-the-hind-legs-can-run" is the 
> model/analogy _you_ use to "refute" the idea that forelimb assistance 
> can convey advantage by increasing maximum speed. The assumptions I 
> listed are inherent to _your_ argument that fore-limb assistance can't 
> increase maximum speed in the absence of incline.

Constant thrust force and parallel thrust force are not assumptions of 
the hind limb limited running mechanic.  They are not assumptions 
inherent to my argument, and I thus I never meant to imply them.  The 
animal can point the thrust in any direction it wants, and apply it in 
any series of pulses or continuous push that it wishes.  If the hind 
limbs are the limiting factor in velocity, then aerodynamic thrust 
still does not speed them up.  

A). A running bipedal animal can use an incremental increase in forward thrust 
to increase stride length, w/out decreasing stride frequency, therefore 
increasing forward velocity. If thrust continues to be increased, failure of 
the hind limbs will indeed eventually occur, but the rate of forward motion 
will be greater at the failure point than in the initial state. The speed 
difference between the initial state and the fail point is the window of 
advantage, in the evolutionary context. Window-size is proportional to the 
relative thrust-generating capabilities of the fore and hind limbs. If the net 
thrust is appropriate in direction, timing, and force, it doesn't matter 
whether it comes from stronger leg muscles, a rocket pack, a car, tail wind, 
gravity, or "wings". Just don't turn up the juice too much, or you will wreck, 
perhaps from the inside out. As sometimes happens to steroid users.

I take the gist of your comments on "friction coefficient" to indicate you 
calculate that an upward thrust vector (or any reduction of the gravitational 
vector) reduces the ability of the feet to impart thrust to the substrate, and 
therefore reduces speed. 

1). Consider a running bipedal animal w/ clawed hind feet on hard flat ground. 
At some point, imparting more thrust w/ the feet becomes counter-productive w/ 
respect to speed because traction is broken, and energy is wasted. If forward 
thrust from the forelimbs is applied at this point, the foot thrust required to 
maintain speed is reduced, reducing slippage and reducing energy waste. In 
other words, the hind leg is returned to it's optimal state relative to thrust 
vs slippage, which obviously varies by substrate. 

If net horizontal 'non-hind foot' forward thrust is further increased 
incrementally, while hind limb thrust is decreased such that speed is constant, 
there will come a point where all that is required of the legs to maintain 
speed is counteract gravity. Note-- at this point, the stride frequency is the 
same as the initial state, as is the stress placed on the leg when the foot is 
placed on the ground at the 'front' of the stride cycle. Unless the max stride 
frequency while the hind limbs are 'idling' w/ respective thrust generation is 
exactly equal to that attainable while the legs are imparting thrust, and the 
workload of the skeleton is exactly that required to handle a sprint w/ out 
failing, the way forward to increased speed is clear, including in the 
evolutionary sense. 

2). Now consider the substrate grades from hard rocky soil through swamp, and 
that roots, vines, etc ,are usually present. Claws are mechanical devices, and 
claw-assisted traction is not necessarily gravity dependent, which is why they 
are so widely used. In fact, reduction of the gravity vector on the hind foot 
can _increase_ the thrust potential of same, rather than reducing it. (Anyone 
who doubts this should run a zip-line (or 'zip-rail", to coin a term) across a 
plowed field such that some body weight can be supported by arms, and compare 
the time of crossing the field to an un-assisted crossing.) Gravity-support 
responsibilities can be shifted from the tail (see comment C) to the hind feet 
and back as necessary while under way, and the tail-friction passively relieves 
the hind legs of most directional control responsibilities (in the maintaining 
a straight-line sense). This is somewhat analogous to the way a cable-skidder 
works when winching it's way through a swamp,
 the 'wings' being the cable/winch assembly, and it is an efficient means of 
transport in "sub-optimal substrate".

This is not only supported by the 
physics of the situation;

B). See comment A. 

 we can gather data from a wide range of 
birds, and the vast majority do not use the wings at all while running. 
  Those species that utilize the wings during running do so briefly, 
during acceleration or turning.  This indicates an advantage in balance 
or turning radius.  It also indicates a lack of utility in wing 
assistance to raise maximum speed (even though living birds can point 
aerodynamic force vectors in a number of directions).

C). We agreed that adult living birds are not good models. They can fly, and 
the behavioral phenotype reflects that. They don't _need_ wing-assisted 
running, why would they seek to use it? When they need more speed, they launch. 
Their behavior is irrelevant when evaluating a ground-up scenario, which is 
populated by animals that can gererate aerodynamic thrust with their forelimbs, 
can't fly, but would certainly like to. The mechanics and behavior of massive 
flightless birds like ostriches are irrelevant, period. Ground-up scenarios 
involve small animals, almost surely w/ tails.

 BTW-- the tails are important to the traction/sub-optimal substrate debate 
(see comment 1), and also (it seems to me) allow (when used in conjuncture w/ 
the fore-limb "wings') the hind limbs to be more free/effective/important in 
(prey) capture/fighting... 

> In any case: (another) major flaw in the model that is used to 'prove' 
> that inclines are required for fore-limb assist to be useful in 
> forward locomotion is yet another underlying assumption; that of hard 
> smooth ground. An animal that can create fore-limb thrust is _highly _ 
> unlikely to
> restrict itself to a theoretical parallel-to-earth-surface direction
> for that thrust, and directing thrust slightly upward can increase
> stride length and height w/out necessarily decreasing stride frequency.

I only assumed that the ground was firm enough that the feet do not 
slip.  If they slip, then producing lift towards the substrate can be 
effective for some animals; this would then be effectively WAIR all 
over again.  Producing both lift and thrust such that the animal 
increases height on each stride does not increase forward velocity 
during running; it will actually slow the animal.  

D). Maybe. maybe not. It depends on how much thrust and how it is applied.

It will clear higher obstacles, however. 

E). Which in many environments increases its rate of forward motion, in the 
non-theoretical sense.

Thus, I did not really assume a parallel-to-earth 
vector; I simply investigated the problem using that vector because it 
would be the most advantageous from a speed gain scenario, if such 
maximum velocity increases were possible.  I simultaneously considered 
other vectors; but they slow the animal down intrinsically.  The 
exception being a downward vector to increase friction coefficient, 
which is usually an incline situation and is the WAIR dynamic.

> Conditions that reduce hindlimb traction even slightly such as muddy 
> ground, shallow water, and certain types of vegetative cover therefore 
> alter the benefit profile of fore-limb thrust considerably. I call 
> this the 'tread-lightly factor'... and it is easy to construct 
> scenarios where _maximum forward speed_ is increased through 
> _fore-limb assistance_. "Sinking in is not a nimble thing to do" -- a 
> song someone should have written, but didn't.

That's a reasonable point; though the vector would be generally be 
pointed down into the substrate to increase the frictional force on the 
feet (ie. WAIR dynamic).  Only if the animal is sinking appreciably, or 
wading rather deeply, would producing a positive lift force on each 
stride be helpful.  Deep wading will interfere with the forelimbs 
unless it is taxon with long legs, which then reduces the need for help 
from the forelimbs.  Thus, there is a window of possible help there, 
but it is limited.  Modern wading birds, even those that run on muddy 
ground, are rarely seen running while using the wings, unless they are 
launching.  This indicates that the window of advantage is small among 
modern birds in those environments.  However, they might not be good 
models for basal forms.

> A model wherein the simplifying assumptions are hard, smooth, flat 
> ground and a constant thrust vector at precise right angle to the 
> gravitational vector is interesting as a first step, and indeed 
> inevitably results in a faceful of dirt or reduction of velocity for 
> the poor creature required to operate under those conditions. However, 
> when the results are applied to the real world where other conditions 
> exist, it is a clear-cut case of garbage in, garbage out.

I really think you are under the impression that the physical analysis 
I discussed was a much more limited than it is.  The only thing I 
assumed was that the animal isn't slipping heavily.  That is hardly 
"garbage in"; it is pretty realistic since most running animals don't 
slip constantly.  It was not a narrow computer simulation, it was just 

F). Sure they do. Every step they take they slip, their entire lives. I bought 
a bottle of Snapple last night; the factoid on the label said, "Did you know
that walking on hard ground uses 7% more calories than walking on
pavement?" _Hard_ ground. Does concept that open up the possibility of an 
energetics 'advantage window' in some circumstances (ie, substrates and body 
plans) that might operate at various speeds/gaits? Yes. How big the window is 
depends on the relative efficiency of the front limbs (aero-thrust) vs hind 
limbs (foot thrust). 

> By contrast, I quite agree that generating lift forces in other
> directions, or in short pulses, might have advantages for agility or
> maneuverability.
> =============================
> 2. "Might"?
> =============================

Yes, might.  If there was a consistent advantage, then we would see 
more wing use by running birds.  In addition, we need more information 
on the mechanics of the basal forms we're discussing to know if they 
could produce the advantages suggested above.

> If the animal wants to run faster,
> it is more efficient to speed up the hind limbs than to add thrust.
> ==================================
> 4). That depends _entirely_ on the ambient environment, bodyplan and 
> lifestyle of the animal in question.

But it is true for the vast majority of known body plans, lifestyles, 
and ambient environments in which avian locomotion has been studied, 
because rather unusual circumstances have to be in place for it not to 
be true.  I am obviously not certain of the precise body plans or 
lifestyles of basal birds, but what we know of their structure so far 
suggests to me that my statement would hold for them as well.  Until 
evidence shows otherwise, it seems the more strongly supported 
hypothesis.  Again, we can observe quite easily that volant running 
birds do not flap while running except during fast starts and launches.

> Again, there is the implied assumption is that the process of evolving 
> flapping forward flight through forelimb-assistance begins w/ some 
> sort of bipedal cheetah analog running on hard smooth ground

That assumption is not made, actually.  The results are the same with a 
modest runner on any reasonably level surface (ie. not a tree trunk) 
where the animal doesn't slip a great deal.

> =================================
> Thus, regardless of the gait, aerodynamic thrust is not helpful for
> faster speed.
> =====================================================
> 5). Again, that depends on the capabilities of the hind limbs, the 
> direction of thrust, and the physical qualities of the substrate. If 
> by 'chance,' some benefit occurs, evolution toward flight continues.
> =====================================================

My statement does not depend on the capabilities of the hind limbs, or 
the direction of thrust.  It only depends on substrate to the degree 
that they animal is not slipping heavily.  I suppose flight evolution 
could have been closely tied to locomotion on slick mud surfaces, but I 
doubt it, based on the range of environments in which near-avians and 
basal-birds are preserved.  In addition, a "slick-running" model has 
the same problems as the WAIR model with regards to flight apparatus 

> It can be useful for increasing _acceleration_, and thus
> getting to a given speed more rapidly.
> =====================================================
> 6). Which, btw, is "a cursorial mechanism by which forward progress is 
> directly enhanced by wing oscillation", and significant from the 
> standpoint of a selective process.
> =====================================================

Yes; I noted this exception to my original statement in my last post.

> They are not required for mathematical analysis, actually, and I did
> not mean to imply such assumptions.
> ==================================
> 7). Huh? If you are using a model, you have to take responsibility for 
> the underlying assumptions.
> =================================

True, I am.  But I did not make a number of underlying assumptions that 
you original thought I might have made.

> Is there a reason that you
> separate mathematical/mechanical analysis from evolutionary analysis?
> I am generally used to melding the two together.
> ===============================================
> 8). You just have to pay attention to which form of analysis is 
> subjugated to which. For instance, I think you sometimes forget that 
> the relative efficiency between competing iterations of a given system 
> at a given time is what is relevant in the competitive context, as 
> opposed to the theoretical efficiency of the total system relative to 
> the ambient environment.

I keep both in mind; I'm trained as an evolutionary biologist, first 
and a biomechanist, second.  Obviously, what is important is whether 
the animal has improved relative efficiency compared to its 
competitors.  However, this can be evaluated by asking whether or not 
there is an overall increase in efficiency from a given behavior, 
or if a given hypothesized dynamic is mechanically feasible.  
If a dynamic is not feasible, or can never be efficient compared to ancestral 
then it will not have an advantage as a competitive variant over a 
given iteration, regardless of the playing field.

> Advantage is binary relative to construction of a specific 
> evolutionary scenario, ie, it is "helpful", or it is "not helpful".

I'm not sure this is actually true, but it's a reasonable 
simplification for the case at hand.  Really, advantage is probalistic.
G). Exactly. Qualitative ranking of advantage can be doable _among_ exploits, 
given knowledge of environment, allowing quantification of probability. Or the 
process can be reversed, gaining knowledge of environment, given knowledge of 
morphology. Quantifying talent w/in a given exploit is not doable, when 
constructing evolutionary scenarios. The dataset is unattainable, for reasons 
that include the ones you mention below, and the results are opinions. You run 
the 40 in <10 seconds, I run it in <30 minutes. We are both well above the 
survival benchmark (I hope). Who can quantify how that affects our relative 
chances in our real lives? There is no way.

  If the probability of advantage for each individual is high, then the 
probability of overall expansion of the trait in question is high.  If 
the probability of advantage is low per individual, then the trait 
might still expand (bet-hedging traits, frequency dependence, etc. 
might help it along) or it might disappear.  This is the case even if 
the trait is sometimes "helpful".  However, a binary state is probably 
a good approximation here because we're discussing limits.

> If you find that it is helpful, it must be included in consideration 
> of the effects of selection on heritable morphological variance. If 
> you say "not helpful to increase stride length", then it seems to me 
> inescapable that you are using
> the unrealistically limiting assumptions: 1)  thrust is always at 
> exactly right angle to
> the gravitational force, and 2) substrate conditions are optimal for 
> hindlimb traction.

I am saying that it will not be helpful to increase stride length (at 
least in a manner that increases forward speed) in most realistic 
cases.  I am not assuming anything about thrust direction with that 
statement (although thrust is actually defined as horizontal; the term 
we should be using there is the 'resultant force'), nor am I assuming 
that substrate conditions are optimal.  I merely assumed that they were 
not extremely sub-optimal.

> And why would I leave out leaping? If it works, it works. Prey capture 
> scenarios come to mind, as do refugia.

I don't leave them out, either.  They just weren't relevant to my 
initial comments.  I'm not arguing against a "ground-up" scenario on 
the whole, just a maximum speed increase advantage.

> We must be careful, however, because even juvenile galliforms are not
> particularly good models.
> ================================
> 11). But probably the best living models we have, yes?
> ================================

Probably not, actually.  The juveniles of other groups might be better. 
  In particular, the juveniles of species that are not burst takeoff 
specialists would be more informative, for a number of reasons.

> ===============================================
> 12). The wing kinematics changes have been studied through the entire 
> maturation process? Or are you extrapolating from adult birds? This 
> comment (unlike other sections of this post) is not in any way 
> adversarial... I am genuinely curious, and your expertise is 
> impressive, to say the least.
> ===============================================

The flight-related anatomy has been studied throughout the maturation 
process for galliforms, rails, geese, and a few other groups.  Wing 
kinematics have been studied in juveniles of galliforms to some degree, 
at various ages, though a reappraisal might be helpful.  In any case, I 
was using data from actual juveniles, albeit mostly qualitative data 
with regards to the kinematics.  The structural data is more amenable 
to quantitative analysis, and happens to give the same answer in this 
case.  I'll dig up some references for you.  Thanks for the compliment, 
btw, that was very kind of you.

> =============================================
> 13). In my opinion, that ("feasibilty" of increasing max run speed) 
> cannot be determined by theoretical analysis; once the 
> possible/impossible threshold is breached, subject to reasonable 
> assumptions, it is time to take data.
> =============================================

The analysis for modern avian running dynamics isn't quite as 
theoretical as you may have guessed.  There it is more known mechanics 
and computation involved in my original appraisal than theoretical 
biology.  In any case, getting data is always a good step, and one 
reason that I am confident of my conclusions is that the behavioral 
data and structural data from modern birds supports those conclusions.  
However, which observations and calculations are relevant to basal 
birds is a tricky business, and then we are in theoretical analysis.

> =================================================
> 14). So your opinion is that within a cohort of same-age quail, those 
> whose feathers are continually pruned will continue to arrive at point 
> B at the same time as those who have begun to receive thrust 
> assistance from decreasing 'limb-load'? Obviously, divergence will 
> occur. But you feel it will not be measurable before the un-pruned 
> birds attain flight?
> =================================================

No, I think those with feathers intact will tend to get to point B 
faster, but only because of increased fast start acceleration.   Thus, 
the further the animals run, the less relative time will be spent 
flapping and the smaller the difference in time to get to the finish.  
This should be measurable, though the differences might be small and 
large samples therefore might be required.  I would also recommend 
using a continuous vortex gait flyer, if possible.  Juveniles from 
precocial ducks might be a better model than young galliforms.  
Incidentally, decreasing limb load would be disadvantageous to running 
speed, because it would mean the hind limbs were producing less force 
in their interaction with the ground.  Did you mean to say 
"increasing", or were you referring to some other loading?

> ============================================
> 15).  I have not "suggested" anything. I have clearly and 
> unequivocally stated, from the get-go, inclines are not necessary for 
> valid fore-limb assisted (= "wing-assisted"), ground-up evolutionary 
> scenarios. I paste in the original statement and your original 
> response for those who might be confused about what started this:
> I wrote in reply to one of Scott's posts: "Not sure I understand, from 
> the perspective of a 'ground-up' selective
> process that can transform a terrestrial mud-lover into a barn swallow,
> where the line between volancy and various forms of wing-assisted
> running is (inclines are NOT necessary, in my opinion)."
> Mike H. replied-- "The incline is necessary, because without it there 
> is no requirement to
> produce a lift force towards the substrate, which is the critical
> aspect of wing assisted running."
> My position-- Inclines are not necessary for constructing valid 
> ground-up scenarios wherein forward flapping flight evolves; that 
> includes, but is not limited to, scenarios involving selective 
> advantage conveyed by forelimb generated thrust that increases the 
> maximum speed of the animal.
> ============================================

Good call going back to the beginning (think we've lost people yet?)  I 
misinterpreted your original comment about inclines not be necessary as 
saying that the animal would attain better forward speeds even if there 
were not an incline.  I did not mean that inclines are required for the 
evolution of flying forms from cursorial ancestors, though I see 
exactly why you thought that's what I meant.  Sorry for the confusion.

> and that maximum
> speed can rarely be enhanced by wings.
> ====================================
> --- Sigh. "Rarely"?
> ===================================

Well, yes, because the animal would need to be running under rather 
unusual conditions

H). A lot of unusual conditions occur in geologic time...

 to create a situation wherein the forelimbs can 
enhance maximum running speed.  Under most conditions, running speed is 
not enhanced.  And maximum running speed is never truly enhanced, since 
gaining traction on mud or a tree trunk only allows the animal to reach 
a speed closer to its normal maximum, and does not allow the animal to 
exceed its normal maximum speed.