Bird flight, how did it really start ?
On the Evolution of Flight in Archaeopteryx as well as in Pterosaurs
So far, there was no reasonable hypothesis to explain the origin of flight in vertebrates. Several amateurish models have been debated by palaeontologists. In 1984, a conference took place in Eichstaett, Bavaria, which dealt exclusively with Archaeopteryx, the oldest known animal with feathered wings, and with the evolution of flight ability. The various presentations by the participants were published (M. Hecht et al. 1985). I was very interested in these papers, since the origin of flight appeared rather mysterious to me, and my expectation was to find a convincing explanation. However, the views concerning the evolution of flight adopted in these articles were completely incredible and not based on the principles of flight physics. It is not as easy as palaeontologists with a missing special qualification think it is. On the contrary, some expert knowledge is required to understand the actual path of development. After all, I got some insight into the poor knowledge of the involved workers and the available fossil material, from which clumsy conclusions had been derived. Unfortunately, in the mean time there has been no progress concerning the abilities of these “experts”.
The main weakness of the old models consists in the fact that they emanate from the forbidden presumption that a primordial wing in the ancestors of Archaeopteryx was already present. However, the problem is just how a wing could develop from an unspecialized arm or forefoot. It cannot suddenly have been present, rather its development is based on a long-lasting evolution, on mutation and natural selection.
Can a moderately critical reader believe that the bird ancestors with their primitive wings had only to learn to use them for flying in the air? According to the opinion of palaeontologists this should have been an easy task. In 1995, I found the explanation of the real path which resulted in the ability to fly high in the air. Although the solution was not simply obvious, in most cases the hard parts of fossils yield sufficient informations for a reliable reconstruction of the former lifestyle. In this case it was the long bony tail of Archaeopteryx, and additionally the same feature in long-tailed pterosaurs, in both forms stiffened, which offered the decisive hints. This feature is very obstructive for a flight high in the air, nevertheless it must have served a special purpose. In addition, in both forms the finger claws are still present. In later birds such as Eoalulavis from the lower Cretaceous of Spain and in all Recent birds it is replaced by the important alula. Thus, these early forms can hardly have been flyers high in the air.
Fig.1. The well preserved Berlin specimen of Archaeopteryx. Particularly conspicuous are the long bony tail and the sharp claws of the hands as well as teeth.
The first find of the Urvogel opened a path to the evolution of birds, since this animal presented old reptile features on the one hand, on the other bird features pointing to a close relationship between reptiles and birds. In particular, the long bony tail and the sharp claws of the manus are conspicuous. These traits cannot be found in modern birds, and they are indicative of aggravating differences. Therefore, many workers have correctly considered that Archaeopteryx cannot yet have been a true bird. However, the mystery persisted how the wing could have evolved, and this fact was used for attacks against Darwin’s idea of evolution, because the wing was seemingly suddenly present. Darwin supposed correctly that the wing would previously have had a different function, though he did not know which one. Archaeopteryx did not lead to unequivocal statements about the evolution of active flight, as little as did the previously known pterosaurs. However, certainly there is a path leading from reptiles to flying vertebrates.
The glider hypothesis as well as the runner hypothesis could not at all convince me because of their imminent weaknesses. Maybe, I am better qualified to develop a new theory than anybody else, since I have expert knowledge in this field. I was interested in flight from my earliest years, have designed model airplanes, was an active glider pilot for many years, have studied flight physics, and have worked in the department of aerodynamics and flight mechanics at Dornier Aircraft Company since 1965. On the other hand, I have discovered my passion for fossils several decades ago, obviously a favourable coincidence.
When I began to deal with the evolution of flight I was very disappointed by the amateurish ideas presented so far. Admittedly, I did not immediately have an answer either, but I noticed that the presented ideas could not be correct. Up to now, biologists, palaeontologists, and further -gists have used very poor lines of argumentation, needless to say they are completely wrong. Experts in anatomy or in ornithology are by no means qualified by this knowledge to reconstruct succcessfully the origin of flight. Using their simple arguments they do Darwin an injustice and demonstrate that they do not have understood the principles of evolution. Darwin has well formulated his ideas and has well recognized the real problem. In any case, in birds, pterosaurs, as well as in bats, there must be an uninterrupted path from unspecialized arms and hands to a flying wing by means of mutation and natural selection. To narrate such simple stories as those heard in Eichstaett and still today by so-called experts would have been an easy task to Darwin, yet he resisted the temptation. He was a true scientist!
The flight movement has its origin in forms that were underwater fish hunters.
Flying under water is very common among many modern seabirds which use their wings for the production of propulsion when hunting under water, for example auks or puffins, and the penguins are completely restricted to hunting fish under water.
Archaeopteryx cannot really contribute to the clarification of the problem of flight evolution, nor can the well-known pterosaurs. None of the models presented so far is beyond doubt, they all are weak and wrong. The argumentation does not at all touch the heart of the problem.
My starting point concerning the evolution of flight is based on the application of principles of flight physics. Of course, I do so not only, because I have a thorough knowledge about it, but because this is the only reasonable method to arrive at plausible and provable results. I am accustomed to utilize objective criteria! So far, any new find of a coelurosaur from the Cretaceous of China with more or less well preserved feathers is praised by the palaeontologist as another key to understand the evolution of flight. Although a certain pride of discoverers is understandable, nevertheless I cannot see what could be expected beyond earlier finds. The ability to fly in the air is proved already by the find of Eoalulavis hoyasi (Nature, vol. 382, 1996) from the Lower Cretaceous of Spain. The decisive feature is definitely the presence of the alula, an inevitable pre-condition for flying in the air, which in Archaeopteryx is missing. On the other hand, the presence of a long bony tail clearly contradicts the ability to fly high in the air. Apparently, there were feathered theropods during the Jurassic and Cretaceous periods, but not all of them were able to fly. The flight feather is a special modification of protective feathers.
2. Old erroneous ideas as to the evolution of flight
Before I explain my theory I would like to have a critical look at the old hypotheses. The glider idea originates from the 19th century. At that time nobody had an idea about the principles of active flight. Accordingly, this idea is, unfortunately, extremely weak. Nevertheless, it persist being propagated, although there has been a lot of progress concerning flight physics in the meantime, however, obviously not available to all people dealing with this problem. The old ideas are in a way based on Darwin’s idea of a modified arm function, but it cannot have worked this way. There are many valid arguments against both hypotheses marking them as incorrect as well as completely naive.
2.1 The cursor hypothesis
This idea was created as an alternative of the glider model, because several workers had strong doubts as to the previously formulated glider hypothesis. The cursor hypothesis was mainly supported by the late J. Ostrom, assuming that an Archaeopteryx-precursor had ‘primordial’ wings with feathers which it used as a net for catching insects. Other workers had the idea that the primordial wings functioned as stabilizers during ‘hectic running movements’. The missing sternum in Archaeopteryx was used as an argument that the animal was unable to perform the upward wing beat. Meanwhile the missing breastbone has been discovered with the seventh specimen of Archaeopteryx, and this problem now seems to be resolved.
Fig. 2. Archaeopteryx-precursor as a very hypothetical runner
Certainly, somewhen a first bird could take off from the ground and fly away. But the line of argumentation presented by the supporters is not at all convincing. This hypothesis cannot explain why in an animal living on the ground wings should evolve and the ability to fly in the air. There is no physical principle on which such an idea might be based. Lift cannot be generated just by flapping the arms up and down, even if they are equipped with unspecialized feathers. To be able to produce lift as well as propulsion the wing must possess a well developed section, a profile, apart from the exact sequence of movements. Such a profile cannot be created by witchcraft. For ground-dwelling animals there is no need to fly in the air. In Archaeopteryx the maximum running speed was approximately 10 km/h. This speed would have been much too low to generate sufficient lift and to make lift-off possible, since the animal could not yet have learned to perform the exact wing movement, even if a wing was present.
Rather for the successful evolution of new capabilities the presence of an attractive new food source must be postulated. There must be a need for a behavioural change. But why should a reptile fly for this reason? Insects for example could also easily be caught on the ground. Recent birds having lost their flight ability in part or completely demonstrate that the ability to fly in the air is per se not an attractive aim. Our domestic chicken (Gallus) can still fly a short distance, however it does not like to do so. In journals such as Nature you can occasionally read statements of the kind that nature experimented with new possibilities. It is frustrating to read about such ideas again and again. In my view, this is a naive argumentation, since in any case animals are individuals which do not have time for experiments, but have to look for food all day long. Any animal must perfectly work any time during the entire history of its species. There is no superior authority controlling the evolution, it is just mutation and natural selection by chance, probably alternatively in connection with cybernetic processes which result in an acceleration. Taking account of physical principles the running hypothesis simply does not make sense, it cannot seriously be called a hypothesis. It appears entirely unbelievable that a running animal can catch flying insects in this way. This hypothesis does not represent an improvement. As an endeavour it may be justified but not more. The same applies to the following glider hypothesis.
2.2 The glider hypothesis
As a modification of the original arm function it was considered that precursors of Archaeopteryx were tree-dwellers which became gliders. This process was assumed to have started by leaping from branch to branch using the ‘simple wings’ for generating some lift. Lateron this glide was believed to be extended more and more, until finally the animal could fly. It rings quite reasonable, however it is completely unrealistic. The false idea is that the presence of a wing producing some lift is taken for granted. But it remains entirely unclear how it evolved. This idea goes back to the nineteenth century, when nobody had an idea how flying in the air is feasible. There is even no definition of the kind of gliding involved in these gliders. It is well known that there are two possibilities of safely coming down to the ground, namely parachuting and real gliding. There is a fundamental difference between these alternatives. Parachuting is not a gliding process, but a retarded fall as performed by parachuters. It is exclusively based on the drag of a falling body, without any lift being produced. In contrast to this possibility real gliding including the generation of lift is represented by gliding birds, sailplanes, and even by modern re-entry vehicles. The supporters of the glider hypothesis use to leap between the two alternatives, just as it appears opportune to them for their argumentation. No knowledge in flight physics is ascertainable, only mere fantasy and speculations by laymen.
Parachuting is comparatively easy. It represents a balance condition between the mass of a falling matter and its aerodynamic drag. With his closed parachute a parachuter attains a speed of approximately 200 km/h. After having opened the parachute, the speed is reduced to about 20-25 km/h. Although this speed is still high and corresponds to a free fall from roughly 2 m, it is acceptable to a trained jumper, in particular if he comes down on a soft ground. To retard a fall effectively a large parachute area is required.
Fig. 3. Descent by parachute.
In addition, the braking effect is dependent on the shape of the parachute. The best effect is attained by a cup-shaped parachute. The direction is always nearly vertical. It can be modified within certain limits as in a steerable parachute by diverging the flow at the trailing edge which results in a movement in the opposite direction. However, the glide angle is rather steep and cannot be compared with the angle obtained by real gliders. Alternatively a forward component can be established by a jump with a forward component which in any case is more and more reduced during parachuting, as in ski-jumpers.
The problem in this line of argumentation consists in the gain of sufficient wing area. The wing area of Archaeopteryx which had already almost perfectly developed wings has been estimated as well as its weight by former workers. Using these values I have estimated the rate of descent during parachuting. It would amount to nearly 33 km/h, considerably more than in a human parachuter. This value corresponds to a free fall from 4 m, completely unacceptable to an animal with delicate bones, since it would lead to a serious injury at least, presumably to death. Unfortunately, the idea of gliding as a pre-condition of flight ability does not appear tenable. A stable gliding condition cannot be achieved before the final speed is reached. Therefore, it remains entirely mysterious how a development towards a lower speed could start, if high speeds in the precursors had inevitably to lead to death.
Fig. 4. A flying squirrel parachuting with a forward component
Because of the rapidly increasing speed the duration of such a glide would be restricted to less than one second, if it should not be lethal. Too little time would be available to learn an effective glide control. Without aerodynamic means flight control would have to be done by shifting the centre of gravity as required, in any case a very rough method which was employed by the German flight pioneer Otto Lilienthal in Berlin in the late 19th century and led to his death in a crash. As one can see, there are too many physical problems. Objections are inappropriate that modern birds such as hawks are able to descend very steeply and stop their descent shortly before reaching the ground by a few wing beats. The defenders of the glider hypothesis admit that the powered flight only evolved later. According to this line of argumentation Archaeopteryx and her precursors are believed to have been mere gliders, unable to use their wings for powered flight during a parachuting stage.
On the whole, I can definitely state that the ability to fly in the air would not have been possible using the idea of a parachuting stage. Any attempt to come down by parachuting would have ended in a lethal crash. Modern parachuting animals such as flying squirrels have no chance to become real flyers in the future, that is, to develop a wing beat. It is impossible. It is possible that in certain parachuting forms such as Draco volans due to a favourable shape there is also some lift in addition to drag, comparable to an airplane fuselage which also produces some lift, but nevertheless the evolution of an active flight is definitely precluded.
By the way, birds do not use this kind of descent, even if it may occasionally appear as parachuting. Birds move using aerodynamic principles; their flight is any time exactly controlled.
2.2.2 Real gliding
Real gliding is completely different from parachuting. Whereas parachuting is based on drag which requires a large braking area and no lift is present, real gliding is based on the generation of lift in connection with minimum drag. To be able to carry the animal in the air the wings must be equipped with a perfectly designed profile.
Fig. 5. Generation of lift by a profile in birds or in airplanes
When passing the profile the air has to cover a longer distance on the upper surface than on the lower one. To say so, the air must hurry up on the upper side in order to arrive simultaneously with its neighbouring particles from the lower side at the end of the wing. This results in a stretch and thus in a lower pressure on top of the profile. Lift is produced, because there is a pressure difference between upper and lower side. This pressure difference carries the animal in the air. For the generation of lift, of course, a certain air speed is required.
A perfectly developed profile in birds cannot evolve by chance or by witchcraft. However, it is an inevitable pre-condition for flying in the air. It must be able to produce sufficient lift as well as the lowest possible drag. Here another dilemma of the supporters of the glider hypothesis becomes evident, since an explanation for the evolution of a profiled wing cannot be found by the idea of precursors jumping from branch to branch. The evolution of a wing and of active flight cannot be explained in this way.
Fig. 6. The cambered profile of a bird wing and the alula
An alleged advantage of this kind of flight evolution was seen by certain workers in the fact that gravity could be used for propulsion. This ludicrous statement would mean that an increasing gravity would be advantageous which is paradoxical. Flying is possible despite gravity. Maybe, these workers intended to express that a gliding animal is moving on an inclined plane and does not need additional propulsion. Nevertheless, for example designers of sailplanes strive to make the glide angle as small as possible by reducing the drag of new designs more and more, that is, to resist the effects of gravity as much as possible and to obtain increasing distances from a given altitude.
Fig. 7. Stable glide in a modern bird. The occurring forces are balanced, forming a closed triangle.
In comparison to parachuting the angle of attack during real gliding is rather small. Though it is dependent on the shape of the profile, generally it is between 5° and 8°, the values for maximum range respectively maximum endurance. To be able to glide on an inclined plane an animal must be able to control its flight path exactly. Gliding is only possible, if it can exactly maintain the allowable range of angles of attack which yields sufficient lift on the one hand and no excessive drag on the other. This range is very narrow and comprises only a few degrees.
Fig. 8. Dependency of lift and drag on the angle of attack
The highest allowable angle of attack is roughly 15° to 20°. If it is exceeded a completely uncontrollable flight situation occurs. This is the most frequent cause of airplane desasters, if the flight speed falls short of the minimum allowable speed unintentionally or unnoticed. Flying in the air is not dangerous because of too high speeds, but of too low ones. Most crashes happen during takeoff and landing. Low speeds include the biggest problems also for birds regarding steering and power consumption. Only the tiny humming-birds can hover in the air, because their small size allows a symmetrical profile which generates the same lift and propulsion during the upward beat as well as during the downward beat, and they are able to compensate their weight in an inclined attitude. Birds are able to vary the direction of the force generated by the wings within a wide range. However, the shaking wing motion of a kestrel (Falco tinnunculus) is not a real hover, since it utilizes headwind to stay above a certain location.
On the other hand, if the angle of attack is too small then not enough lift is produced and the animal cannot maintain its altitude for this reason. Gliding makes great demands concerning the steering capabilities, much greater ones than considered by the advocates of this hypothesis, if they did at all. I cannot see how a gliding precursor of birds could have managed these problems. In order to avoid a lethal fall, initial glides had to be shorter than one second, as also in parachuting. This time is neither sufficient to gain the speed required nor to learn an effective glide control. Crashed gliders were unable to pass on their experiences to successors; another dilemma which has not been considered or quickly passed over in silence.
Apart from the existing control problems real gliding is connected with considerable speeds. No glider uses to fly at a speed near the maximum lift because of the imminent hazard of stall by flow separation. A safety margin is generally considered. A reasonable gliding speed of Archaeopteryx would be approximately 45-50 km/h. Even in case of a relatively low rate of descent the total dynamic energy had to be annihilated at arrival on the ground. This speed corresponds to a free fall from a height of approximately 8,5 m and would certainly have been lethal. Remember that at the supposed start of evolution there were no means available to retard the gliding speed, as modern birds can do using a few wing beats. Recent birds do not only use their wings for lift and propulsion, but can generate a force of almost any direction, thus also for retardation. Following the argumentation of the glider hypothesis the wing beat should have developed only later. There are only woolly ideas about this evolutionary process. An impression as to the problems of lacking means of retardation can be obtained by an albatross which has to land in calm air, but often tumbles over. This is an exceptional problem, since normally albatrosses can expect high wind speeds to which these birds are adapted, but no more to low speeds. In addition, we must take into account that at the beginning of flight evolution the wings must have been considerably smaller than those of Archaeopteryx. Therefore, the precursors were unable to do all the clever tricks of modern birds.
It is noteworthy that obviously the wing of Archaeopteryx was not yet adapted to low speeds, because the decisive feature of birds flying in the air was still lacking, namely the alula. Three fingers of the manus were still equipped with claws which were only modified in the later real bird Eoalulavis. The alula functions as a small additional wing which is used to prevent or postpone flow separation of the outer wing and probably as a sensor of imminent flow separation. The missing alula would be very strange in a flying bird, which of course would have been confronted with the problems of take-off and landing from the very beginning.
As a whole, the glider hypothesis also is not at all convincing, despite the continuous efforts of its supporters to find a link between Archaeopteryx and modern parachuting reptiles or mammals which certainly are no gliders, but parachuters. The latter mammals have not the slightest chance to learn the active wing beat. Remember that their descent is based on drag, not on lift.
Fig. 9. Flying squirrel making an impossible attempt of a wing beat. Right arm shows the normal angle of attack, left arm an increased angle of attack due to the downward beat.
Again and again the supporters of the glider hypothesis have maintained that the transition to powered flight would be comparatively easy, if only the gliding flight works. In fact this is completely untrue.
Let us assume that the squirrel of fig. 9 is gliding as a bird, capable to generate some lift with its patagium. If it makes a downward wing beat then the aim is to increase the produced force and to reduce the glide angle. However, the wing movement results in a considerably increased angle of attack, since for the generation of propulsion a slow beat is insufficient. The wing motion would result in an excess of the allowable angle of attack; instead of a gain of lift it would completely collapse and the animal fall down. On the other hand, at arrival at the lower turning point of the wing motion the angle of attack would become negative, leading to a complete loss of lift. During the wing beat lift and propulsion must simultaneously and continuously be generated. This is a very complicated matter which can only be done by a real wing which can be rotated as required.
Real flyers are equipped with the so-called propatagium. A modified finger or another bone of the manus serves for its control. Its function corresponds to the alula of birds. Parachuters do not have a propatagium, which they do not need at all. The exact observance of a certain angle of attack is not necessary since the glide angle is steep in any case. The evolution of real flight from (parachuting!) gliders is just impossible.