Application to real swimming vertebrates

1. Introduction

The problem of generating propulsion in water-dwelling vertebrates is being discussed for a long time. Abel (1912) already dealt thoroughly with the swimming locomotion of different vertebrates. In particular, the generation of propulsion in fast swimmers with a pronounced caudal fin was unresolved despite various efforts to find a satisfactory explanation. On page ‘swimphysics’  I have explained at some length the hydrodynamic regularities relevant for swimming locomotion. This investigation includes the clarification of the mechanism of propulsion in fast swimmers. Propulsion in fast swimmers with a pronounced caudal fin is always effected by a bending of the trunk, but contrary to previous presumptions never by the tail fin. This fin only can serve directional control. The consideration of hydrodynamic laws leads to applicable criteria for judging the suitability for fast locomotion, manoeuvrability, and the preferred kind of locomotion in fishes as well as in further aquatic and amphibic vertebrates. This is explained as follows.

2. Locomotion of aquatic vertebrates

Apparently, the greatest former problems were caused by a missing understanding of the mechanism of propulsion generation. Therefore, here the most important statements of my previous investigation are briefly repeated. They dealt with the fundamental diference between flying in the air and ‘flying’ under water. In principle, the generation of propulsion is effected in any case by a rearward acceleration of the surrounding agent, irrespective of how this is achieved.

The air flow along a wing is accompanied by two effects which are basically connected with each other, namely

        -   by an acceleration of the adjacent air and

        -  by the occurrence of a resultant force due to a pressure difference between upper and lower surface respectively left and right flank.

For flying in the air exclusively the pressure difference is utilized for lift, whereas the increased flow velocity along the essentially rigid wing section is negligible because of the low density of air. Propulsion and thus an additional acceleration of air in birds must be generated by pronation and supination of the manus, alternatively it is achieved during gliding on an inclined path by a gravitational component. The wing beat produces a portion of the lift as well as the entire propulsion.

On the contrary, locomotion in water is exclusively based on the high density of the agent. It is effected by a reaction to the rearward acceleration of a water mass. The likewise occurring force due to a pressure difference by alternately bending the trunk is obstructive for the motion in this case and must be compensated. A different kind of producing propulsion is achieved in certain vertebrates by the thoracic fins. It resembles the wing beat of birds.

The wide-spread misunderstandings have led to hardly defensible statements in the past. Various publications by authors with only a superficial knowledge of hydrodynamic laws, dealing with the generation of propulsion were suitable to create even more confusion by using inappropriate comparisons, and thus could not contribute to a better understanding. Hopefully, this can be achieved by the following remarks. The knowledge of the hydrodynamic laws given on page ‘locomotion’ is taken for granted.

Since apart from directional control the only function of the tail fin consists in the compensation of the periodically occurring side force during generation of propulsion, the shape of the tail fin allows unequivocal statements about the preferred kind of locomotion in Recent and fossil fishes, respectively concerning their basical swimming capabilities. Furthermore, presence or absence of the tail fin in modern or fossil fishes as well as in further fossil aquatic vertebrates can be used to determine whether these animals were able to move using propulsion by the trunk or were restricted to less effective methods. In addition, statements as to the maximum speed in comparison with other swimmers are possible. Shape and relative size of the tail fin compared to the total length of the trunk and the size of the other fins can be used as criteria. Fortunately, such statements can easily be tested by observation of modern swimmers. Based on these criteria all swimming vertebrates can be united in different groups:

       -   slowly swimming fishes

       -  slowly swimming tetrapods

       -   fast swimmers in the laminar flow region

       -  fast swimmers in the turbulent flow region

       -  These groups can in part further be divided, according to the different mechanisms of propulsion.


2.1 Slowly swimming fishes

Caudal fin, thoracic and further fins or fin fimbrisses are present in all modern teleostei. Size and shape of the caudal fin are very variable, ranging between a tapering fimbris as in eels and a big semilunate fin as in tunnies. Between these extremes the almost endless variations due to adaptation to the differing requirements of locomotion and milieu can be found. However, obviously there is a strict separation between fishes moving in the laminar flow region on the one hand and the big fast swimmers on the other hand, which are adapted to the requirements of turbulent flow, because a special nutrition did not allow an alternative locomotion respectively only made it possible.

The overwhelming majority of fishes belongs to the slow forms moving in the laminar flow region. Slow swimmers are exemplarily represented by early chondrichthyes, dipnoi such as Neoceratodus, and crossopterygii as shown in fig.1. These forms are well known already from the Paleozoic period (Thenius 2000) and could maintain their ecological niches essentially unchanged up to the present day due to an adaptation to particular environmental conditions with low competition. However, they are restricted to a few genders. In these forms the tail fin is comparatively weak and unspecialized. Certainly, the attainable speed was of little importance to these forms. Their mode of life apparently is determined by requirements of a different kind. The modern Latimeria (Fricke et al. 1987) is indicative of the sluggish locomotion of these forms. Although the caudal fin is present, its unspecialized shape suggests that it had never to compensate significant side forces. Thus, it is likely that the trunk did not play an important part in the generation of propulsion. On the other hand, undulating dorsal and ventral fins, in lung fishes wing-like thoracic fins and a fin fimbris ranging to the rear end, cannot produce a remarkable acceleration of water and, therefore, do allow only low speeds. This is in agreement with observations by Riess (1986) who states a sluggish locomotion in Neoceratodus forsteri. In addition, eels and further eel-like fishes indicate by their weakly developed or missing tail fin that they are restricted to an undulating locomotion and consequently to low speeds. Swimmers primarily living near the sea bottom are not adapted to fast locomotion.



Fig. 1.  Examples of slowly swimming fishes with unspecialized caudal fin. Propulsion is generated without a predominating influence by the tail fin.

a. Neoceratodus sp., b. Coelocanthus sp., c. Platysomus sp., without scale, redrawn from Abel (1912) respectively from Carroll (1993).


The overwhelming majority of fishes belongs to the actinopterygians. Generally these forms possess a well developed tail fin. Nevertheless in most representatives, in particular in reef-dwellers fast locomotion is of subordinate importance, whereas manoeuvrability is emphasized. The type of reef-dwellers is characterized by types such as the Permian Platysomus (fig. 1c). Locomotion and position control in such forms living primarily in calm water are mainly performed by elongated dorsal and ventral fins, whereas the small caudal fin, compared to the trunk height, plays a subordinate role in the short-time fast locomotion, during escape from a danger for instance. All representatives of this group move in the laminar flow area.


2. 2 Slowly moving tetrapods

Aquatic mammals and reptiles also can be subdivided into slow and fast swimmers. Presence or absence of a large tail fin again can be used as a criterion. Slow swimmers are exemplarily represented by modern crocodiles and seals, but can be found as well among fossil forms. Several kinds of locomotion may be employed, and an examination is required from case to case whether propulsion was generated primarily by undulation of the tail only, or additionally by trunk bending, or exclusively by the extremities modified to fins. These are the only basical possibilities of locomotion in forms lacking a well developed caudal fin.

2. 2. 1 Forms with propulsion by tail undulation

Crocodiles and modern monitors move in the water primarily utilizing an undulating movement of the tail, whereas the extremities only allow a rowing locomotion based on drag which is mainly used for manoeuvring. The extremities have to retain another function for going ashore, and therefore represent a compromise. This kind of locomotion precludes high velocities, only a short-time burst for taking prey by surprise.



Fig. 2. Slowly swimming tetrapods, with propulsion generated by the undulating tail.

                                                                     a. Mesosaurus,

                                                                     b. Thadeosaurus

                                                                     c. Hovasaurus,

                                                                     d. Claudiosaurus


There is one group of fossil vertebrates which is characterized by a shape of the skeleton indicating that propulsion was exclusively generated by undulation of the tail. These forms possess a long and flexible tail. The vertebral column shows relatively short and uniform spinous processes, occasionally somewhat elongated ones in the region between shoulder and pelvis (Ebel et al. 1998, Ebel 2000), and thereby suggests a predominating, but not exclusive aquatic lifestyle. The absence of a caudal fin is a clear hint that the wave length of the undulation was definitely shorter than the tail and, therefore, there were no uncompensated side forces (Lighthill 1976). The extremities in these forms do not suggest a remarkable specialization, although the legs may be stronger than the arms.



 Fig. 3. More such forms

                  a. Askeptosaurus,

                   b. Champsosaurus,

                   c. Protosuchus



Several fossil reptiles can be placed within this group, ranging from the Permian age to the Upper Cretaceous. Of course, this arrangement is not meant to represent a generic relationship. Without claiming completeness this group comprises the following forms as examples. Some of them are so far regarded as terrestrial, however, in my opinion they would have been mainly amphibic (fig. 2 and 3):

        -   Mesosaurus          (length approx. 1 m, Permian of  South Africa  und South America)

        -  Thadeosaurus       (length < 0,5 m, Upper Permian of Madagaskar)Hovasaurus          (length 0,5 m, Upper Permian of Madagaskar)

        -  Claudiosaurus      (length approx. 0,5 m, Upper Permian of Madagaskar)

        -  Askeptosaurus     (length approx. 2 m, Middle Triassic of Switzerland )

        -  Ticinosuchus        (length approx. 3 m, Triassic of Switzerland)

        -  Protosuchus         (length approx. 1 m, Lower Jurassic)

        -  Champsosaurus   (length as modern crocodiles, Upper Cretaceous of the western North America), after Carroll (1993)

2. 2. 2 Forms with propulsion by tail undulation and additionally by parts of the presacral vertebral column

A second group is represented by swimmers in which the undulation apart from the tail comprises further parts of the vertebral column. This kind of locomotion is roughly comparable to the motion of slender bottom-related sharks with a weakly specialized caudal fin, for example Echinorhinus brucus or Scyliorhinus stellaris (fig. 4). The tail fin of the latter is apparently adapted to offer a favourable flow pattern of the wake, but not to the compensation of side forces as in lamnid sharks. For example mosasaurs, mesosuchian crocodiles, and ichthyosaurs without a pronounced caudal fin can be placed in this group. These forms show a progressive reduction of the connection between vertebral column and pelvic bones which is a necessary adaptation for an effective undulation of further parts of the trunk. Moreover, this modification suggests an increasing adaptation to marine conditions. Although a well developed tail fin is not yet present in these forms and is not required because of the relatively short wave length of the undulation, nevertheless a lateral flattening of the postsacral vertebral column indicates that for instance in the Middle Triassic Mixosaurus or in the mosasaur Plotosaurus from the Upper Cretaceous presumably more power for an extended locomotion was needed which for good efficiency required a favourable flow pattern near the end of the tail. Because of the modified shape of the tail a somewhat higher speed than in the previous group can be supposed, but still a relatively low one. An increased undulation frequency cannot contribute to a considerably better performance, since a flow separation would be likely.



Fig. 4.  Scyliorhinus sp. as an example of a slow swimmer with propulsion by the undulating trunk and a comparatively weakly developed caudal fin.


These forms present an increasing perfection of an arm transformation to fins. The posterior fins resemble the anterior ones, they may even be larger. Because of the mostly considerable mass of these animals the fins were not suitable for a fast propulsion, but more likely merely for manoeuvring. For example, the total length of Shonisaurus was about 15 m (Carroll 1993), which means a big mass. A skin shape including the dorsal fin, sometimes laid underneath the bones in reconstructions in Liassic ichthyosaurs, as shown for example by Sander (2000) for Mixosaurus and Shonisaurus, is very probably not compatible with the reality. In forms without a large caudal fin remaining side forces were not allowed. Therefore, only low speeds were possible and consequently a large dorsal fin was not required and has never been found in a fossil of these species (Riess 1986). This author regards Shonisaurus as an underwater flyer which used propulsion by the tail fin for a short time at best for escaping. However, from a hydromechanical viewpoint there are no arguments supportingthis assumption. The argumentation of (Riess 1986) was also contradicted by Sander (2000)



                    Fig. 5.  a. Mixosaurus, b. Shonisaurus from M. Sander (2000).


As a whole this group is transitional as to its properties to a further one with a big caudal fin. One or another member may have advanced in this direction as for example the Upper Jurassic metriorhynchids.


2. 2. 3  Forms with propulsion by the extremities

In the third group of slow tetrapods forms can be found which primarily move by means of their arms which have been transformed to fins or better to say wings, some additionally also by means of the posterior ones. Because of this kind of locomotion the members of this group are characterized by strong locomotory limitatations and low speeds. This group mainly consists of the plesiosaurs. The tail of these forms is comparatively short and certainly not involved in propulsion, possibly in directional control. The very variable shape of the skeleton concerning length of the neck and relative size of the extremeties is indicative of a basal precursor specialized to various food sources. Apparently, high manoeuvrability and flexibility of the head in connection with a long neck was more important for food supply than a high speed. A long neck is however not suitable for propulsion by undulation, since the following relatively stiff trunk would strongly hamper the water motion. Certainly, these animals were underwater flyers (Robinson 1975, Riess 1986). But they cannot have been fast swimmers. This follows on the one hand from the considerable size respectively from the big mass, since an increasing absolute mass leads to a decrease in attainable speed, if propulsion is generated by the extremities (penguins are definitely faster than turtles). On the other hand, in comparison to the long and slender wings of penguins, those of plesiosaurs are short. In addition, for judging the attainable speed it does not matter whether the posterior fins could also be used simultaneously for propulsion or whether they worked in step (Lingham-Soliar 2000) or counter step (Riess & Frey 1991). According to Lingham-Soliar (2000) the posterior fins were restricted regarding their capability of producing propulsion. The different design and possibilities of motion of anterior and posterior fins discussed by this author make a separated function likely, that is, propulsion was probably generated normally by the anterior fins alone. If fins or extremities are utilized for propulsion then in general it is done by the anterior ones. Although I do not intend to generally criticize the conclusions of Lingham-Soliar (2000), nevertheless his paper again demon- strates that regularities of hydrostatics, hydrodynamics and aerodynamics are difficult to understand by people not closely familiar with these problems and are hardly ever correctly applied. Correctly Lingham-Soliar (2000) mentions that plesiosaurs were underwater flyers, but incorrectly in his fig. 14 he shows a rowing motion of the anterior fins. However, so far the mechanism of thrust generation has nowhere correctly been presented.



Fig. 6. Generation of propulsion in plesiosaurs and acting forces during underwater flight.

V = propulsive force


Since the acting forces apart from the weight, the buoyancy and the propulsive force must not necessarily pass through the centre of gravity, an additional trim force of variable magnitude is required for the compensation of moments. As indicated in fig.6, this compensation can be done by the posterior fins




Fig. 7. Course of fin motion and hydrodynamic propulsive force generated by thoracic fins, schematic and exaggerated presentation. The course of this movement resembles the wing motion of rays.


A rowing locomotion is never used by swimmers well adapted to underwater flight (Reif 1981b), only for retarding and manoeuvring. Propulsion is based on the acceleration of the adjacent water by the forced flow along a profiled area, but not on the utilization of the drag of a rowing thoracic fin. During the generation of propulsion the angle of attack of the fins is low, which as in propellers is determined by the most favourable ratio of lift to drag. Fig.7 illustrates that the thoracic fins generate a force of variable direction with propulsive components which attains its maximum size during the downstroke. Any time the optimum angle of attack is in conformity with the resultant speed and the direction composed of cruise speed and fin motion. Therefore, the fin must appropriately be rotated during a beat, as indicated in fig. 7. By a differentiated change of the angles of attack of the left and right fins the propulsive direction is variable within a certain margin. In birds with their variable plumage this variation is possible within a large area, as demonstrated during retarding and a following spot landing. The motion of the fins resembles the undulation in rays. It is the same principle of a rearward transportation of water as can easily be observed. This method is technically utilized by the Voith-Schneider propeller, which is mainly employed in faireys and rotates about a vertical axis below the hull, thus making possible various directions, the disadvange being the considerable draught.


Fig. 7a. Course of fin motion in rays

The propulsive force generated by the thoracic fins produces a moment M = a * V/2 at each fin, as shown in fig. 6b. Since the movable shoulder joint cannot transmit a moment, the propulsion must exclusively be transmitted to the trunk by the musculature. With increasing size of an animal and a higher propulsive force required this becomes more and more difficult, because the mass grows with the third power of the length, the cross-section of the musculature, however, only with the second power. Therefore, the attainable speed of big forms must necessarily be low, in comparison to penguins or turtles. It is remarkable at all that certain big plesiosaurs could use a propulsion by the thoracic fins. This configuration appears hardly suitable for a continuous locomotion. It is likely that a locomotion over large distances at considerable speed did never play an important part in the evolution of these forms. Presumably, they were adapted to good manoeuvrability and diving in the moderately shallow water of epicontinental seas. As already mentioned by Abel (1912) the lifestyle was presumably largely comparable to modern seals, and therefore apart from the overall strong skeleton the fins had presumably to maintain an additional function for going ashore. As in turtles this function was probably accompanied by great trouble. The plesiosaurs were prisoners of their extreme specialization and unable to change their lifestyle if required.

With certain restrictions further forms such as the nothosaurs, placodonts or the extremely specialized Tanystropheus (Wild 1973) can be placed in this group. A locomotory specialization is hardly recognizable, neither in the extremities nor in the trunk with a relatively short tail. This leads to the presumption of low speeds. Apparently, as in the plesiosaurs the speed was of low importance for their food supply, however they were water-dwellers. Presumably they were not able to swim faster than Recent crocodiles and adapted to an almost immobile food.

2. 3 Fast swimmers in the laminar flow region

Fast swimmers are characterized by a compromiseless adaptation to the necessaries of hydrodynamics, although speed is not in any case the only feature requiring an optimization. It is not decisive if such swimmers move continuously at the highest possible speed or if the attainable speed is absolutely high, but that they are capable at all to move at high speed. There are fast swimmers in the turbulent flow region as well as in the laminar region. All of these forms, relatively small fishes such as trouts or the biggest whales, are marked by a basically similar streamlined shape indicating low drag, as well as by a well developed caudal fin.

The speed in fishes moving in the laminar flow region is restricted to relatively low values. However, they take strictly account of hydrodynamic limits, that is if they might reach the critical Re-number of 5*105 they must be able to avoid the hazard of separation of laminar flow. Fast locomotion plays a short-time role only, even if the maximum speed is not more than roughly 5 km/h at which for example a trout can stay in the laminar flow region. A trout must be able to move continuously upstream in its environment. This speed also defines the minimum speed of its predators. In these forms the tail fin allows adjusting the position by vertical undulation.



Fig. 8. Fast swimmers in the laminar flow region.                                             

a. pike, b. trout, without scale (after Hertel 1963)


A frequently occurring adaptational type in the laminar flow region is represented by a spindle-like trunk and a distinctly developed large tail fin as in trouts or carps (fig.8) or pikes such as Lepisosteus (fig.8a) in which the beginning decrease of cross-section shows a considerable backward shift. These forms are either adapted to fast streaming water or to short-time attainable high burst speeds. In Lepisosteus, a gender already known from the Upper Cretaceous (Thenius 2000), dorsal and ventral fins are in the rear part of the trunk where they do not disturb the flow on the one hand and can support the acceleration on the other. During hovering the fins can undulating ensure the position.


Remember, please, the realationship between frictional drag and Re-number. The transition area from laminar to turbulent flow marks an upper limit for swimmers such as those shown in fig.8. These forms are unable to cross it. The drag increase by more than 100 % does not allow it, In addition, a further drag increase would follow from a speed increase.







2. 4 Fast swimmers in the turbulent flow region.

Contrary to the forms mentioned above, all modern fast swimmers move in the turbulent flow region. Apart from lamnid sharks and tunnies these are marine mammals which have become completely independent of the land. Big fast swimmers are most interesting for recognizing the physical regularities of swimming locomotion, because these forms demonstrate a perfect adaptation. In most cases the primary evolutionary aim is not the maximum speed but a high continuous speed. Big forms such as the slowly moving Rhiniodon just cannot be so slow that the flow remains laminar everywhere along the body. Since in this case a backward shift of the maximum cross-section can never avoid the flow change to turbulence, the shape of these forms is no more in accord with a laminar section, but a favourably streamlined shape. The change to turbulence occurs relatively early, and therefore the position of the maximum cross-section is no more important.



Fig. 9.  Fast  swimmers in the turbulent flow region.

a. lamnid shark, b. Ichthyosaurus sp., c. Recent dolphin.


a dorso-ventral one in mammals on the other, inherited from the respective ancestors (Klima 1992). However, if propulsion were not generated by the bending of the trunk, but as usually shown by previous authors by the caudal fin it might be vertical also in cetaceans as well as in other swimmers. The large tail fin was only developed as an obligatory organ for balancing the flow direction of the wake in forms performing a continuous locomotion over large distances, which in other words became merely marine and completely independent of the land.

A remarkable feature of all secondarily aquatic vertebrates is the increasing reduction of the pelvis and its separation from the verte- bral column. It can be found in mosasaurs as well as in ichthyosaurs (Riess 1986). This trait is present also in all cetaceans. Although also in plesiosaurs a reduction of the pelvis occurs, nevertheless the vertebral column is still supported by bones. Presumably, the reason can be found in the fact that in contrast to other forms because of the propulsion generation by the extremities and the additional task of being capable to go ashore the ‘base-board’ with shoulder and pelvic joints in connection with strong gastralia in between was strongly stiffened. In purely marine swimmers the reduction of the superfluous rear extremities is not the primary aim of the modification, but a better flexibilty of trunk and vertebral column. It is an imperative necessity to make possible the required bending of the trunk as well as its optimum course in time along the contour. This is the only way for the generation of the required propulsion at a minimum energy supply. This modification is even present in sea-cows (sirenia) which move distinctly slowly (Klima 1992). The presence of this trait is always a reliable indication of propulsion by trunk bending. If rudiments of pelvic bones remain embedded in soft parts as in whales this does not hamper the locomotion and therefore is tolerable. This kind of propulsion is utilized by all marine mammals, even by the sea-otter (Enhydra), but the evolution of a tail fin is only required in permanent swimmers.

Obviously, in most of these forms despite their considerable size large forces are not present at the tail fin, because a particular stiffening is not necessary. Since the side force occurring during generation of propulsion is acting close to the centre of gravity, the resulting moment is low due to the small lever-arm (see locomotion, fig.7). Therefore, the required balancing fin force is small because of the large lever-arm between tail fin and centre of gravity.

The tail strut of whales is slender and remarkably long at dorsal view, and the fluke can therefore be relatively small (Hertel 1963). However, the trunk is rather high in the tail region. This height results from a considerable length of the spinous processes, necessary for a sufficient lever-arm of the dorsoventral musculature for generating propulsion by trunk bending. The validity of the principle of generation of propulsion  by trunk bending is, of course, not affected by the seemingly strange configuration. Analogous to a bird wing propulsion is generated by the flow along the entire body, but not just of the rear part where the bending occurs. Apparently, already small trunk deflections are sufficient to attain the normal cruise speed. This is an indication of extremely low drag at the body surface. This property would presumably not be available to a comparable extent in a technical imitation and unfortunately would not allow a general application of the principle, maybe however for special purposes.

Only at the highest power and at high frequency an obligatory steering of the tail fin is necessary. This fact can best be observed in the vertebral column of tunnies which represent an extreme adaptation to high speeds and in my view are the most fascinating representatives of vertebrates at all in this respect. These swimmers possess two joints near the tail fin (Fierstine & Walters 1968), which in connection with a suitable musculature make possible an exact steering of the caudal fin.


Fig. 10. The rear trunk area of certain tunnies represents a natural analogue to provisions in airplanes for improving the flow in the empennage area.

As previously mentioned, tunnies are equipped with several dorsal and ventral finlets (fig. 10) which probably have the task to improve the flow in the rear part of the trunk and help to convert the spent energy to speed as perfectly as possible. Obviously, the hydrodynamic adaptation of the body of fast tunnies and cetaceans is such a perfect one that relatively small changes of curvature are already sufficient to maintain high speeds, and the required corrections by the caudal fin are low. Locomotion in the turbulent flow region has the invaluable advantage for these swimmers that the hazard of flow separation is low.

Another remarkable feature of big fast swimmers is the presence of a large dorsal fin. In forms with a large vertical tail fin it is present in any case. Although its position on the trunk may be variable, it is however always located in the area where during fast swimming the largest side forces occur. This fin also can only generate a force which is acting perpendicular to the swimming direction. Obviously, its task consists in preventing rolling movements of the body and displacements due to side forces. In cetaceans this fin is largely unnecessary because of the dorsoventral trunk motion, and in big whales not specialized to fish predation and good manoeuvrability the relative size of this fin is rather variable. Contrasting to whales this fin is well developed in dolphins. In these forms the fin is necessary for good manoeuvrability. Apparently the occurring side forces are nevertheless low, since the stiffness of the fin is rather low. For example, if killerwhales (Orcinus orca) jump out of the water it can be observed that the fin tumbles down due to its weight. Maybe, this happens only in individuals in captivity (kind information by Prof. M. Klima, 2000). In the slowly swimming Amazonas-dolphin Inia geoffrensis the fin is reduced to a low keel. In sea-cows it is completely missing. A corresponding purpose may have been linked with the small posterior fins of fast moving ichthyosaurs. They were not required for directional control and hardly suitable for that task, but more suited for an improving influence on the flow, as indicated for tunnies

Apparently, the fast locomotion of big forms was not possible or necessary before the Jurassic period, following a relatively long evolution since the Lower Trias. The development of new capabilities can be observed along general lines in ichthyosaurs. Later, in cetaceans the same aim was attained within a considerably shorter time. This evolution demonstrates the generally observable acceleration of evolutionary modifications. The ichthyosaurs were the only reptiles which developed a semilunate caudal fin.

Probably, early ichthyosaurs discovered the predation on a highly mobile food source below the water surface and progressively specialized to various prey. The construction of the eyes can be used as a clear indication of a changed visibility due to the predominant locomotion under water and thus a changed refractive index. On the other hand, swimming under water is advantageous because in contrast to surface swimmers the bow wave drag is missing. This led to a locomotion at lower energy consumption. However, it is likely that a high speed during predation was not the most important criterion but an increasing endurance. Fresh water dolphins such as Inia geoffrensis move at a relatively low speed which hardly exceeds 4,5 m/s, but generally is restricted to approximately 1,5 m/s (Klima et al. 1980). This speed is entirely sufficient for gaining their prey, since most fishes in the laminar flow region are restricted to lower speeds.

Starting in the Trias, the phylogeny of ichthyosaurs suggests a progressive specialization towards the trunk-propelled locomotion, but judging from the fossil record the evolution was only completed during the Jurassic. Abel (1912) has indicated the evolution using the caudal fins of Mixosaurus, of a juvenile as well as of an adult ichthyosaur with semilunate tail fin (Stenopterygius) from the Lias e, and of a further form (Macropterygius) from the Upper Jurassic (fig.11). Although the evolution probably has taken the outlined course, nevertheless it is questionable about the details. Abel’s ideas suggest a directed evolution, but the skin shapes laid underneath the skeletons are not proven and in Mixosaurus very doubtful..


Fig. 11. Supposed shapes of the tail in (a) Mixosaurus, (b) juvenile, (c) adult Stenopterygius, (d) Macropterygius, redrawn from Abel (1912)

Mixosaurus with a maximum length of about 1,5 m (Sander 2000) is the least adapted to high speeds among the sufficiently known forms. Its pelvic bones are not completely separated from the vertebral column (Riess 1986). For this reason its locomotion was restricted to the undulation of the tail which as in crocodiles did not allow considerable speeds, as presumed already by Riess (1986). The undulation by the tail does not require a large dorsal fin, and it is generally absent in such forms. However, the extremities of Mixosaurus show a good adaptation to aquatic conditions. Brinkmann (1996) has demonstrated that this species was viviparous, which hints at the entire independence of the land. Yet, Mixosaurus is likely to have been a slow swimmer. An extrapolation of skin shapes based on Stenopterygius and Leptopterygius from the Toarcian to Triassic forms does not appear readily allowable.

The same applies to an enlarged skin lobe in the region of the processes of the central caudal vertebrae, which are elongated and inclined against each other (Abel 1912, Sander 2000). Evidence is missing. Apparently, this skin lobe is meant to demonstrate that an initially small fin grew in the course of evolution to the later semilunate tail fin. However, this is rather improbable. Following my remarks on hydrodynamics (page swim physics) an enlarged caudal fin makes sense only if it is located at the very rear end of the trunk, and it has always evolved there. In Mixosaurus a tail fin in the region of the enlarged caudal vertebrae must not be expected. Presumably, this formation of the vertebral column is connected with a high and slender termination comparable to the tail of crocodiles in order to make possible a low drag flow at the end of the body. This development is an indication of increased energy expenditure for swimming and thus to its growing importance, compared to crocodiles for example. The evolutionary course implied by Abel (1912) is not supported by the shape of the vertebrae. In Mixosaurus these are high-oval, short, and laterally compressed, whereas in later forms they are long and roundish. The well developed fins of the extremities are an expression of high manoeuvrability.

Shonisaurus from the Upper Trias is characterized by similar features, but by a higher evolutionary level. The separation of the pelvis from the vertebral column indicates improved locomotory abilities, generated by larger portions of the trunk. The likely twisting motion of the trunk appears best comparable to modern sharks such as Scyliorhinus which live near the bottom. The attainable speed would have been still rather low in such forms, but the efficiency of locomotion increased by a bigger accelerated water mass. The dorsal fin indicated by Sander (2000) is improbable in this case, too. The outlined caudal fin is hypothetical as well. The hindmost caudal vertebrae are also flattened in Shonisaurus (Riess 1986) and suggest a similar mechanism as in Mixosaurus. The extremities reveal an advanced adaptation to high manoeuvrability. Their relatively large size, compared to later forms, may result from a low attainable speed as in the modern Inia. Compared to Mixosaurus such forms appear better suited to a continuous swimming locomotion.

However, the capability of a fast continuous locomotion was only achieved by forms with a pronounced semilunate caudal fin. They have completed the kind of locomotion offering the ability to bridge large distances at a low expenditure of energy. The tail fin of a juvenile ichthyosaur from the Toarcian shown by Abel (1912) cannot serve as evidence that this species developed from Mixosaurus. Transitional forms are unknown. With the evolution of the semilunate caudal fin the achievement of fast continuous swimming is essentially completed in ichthyosaurs. Fast forms generally possess reduced anterior and posterior fins which can be interpreted as an expression of higher attainable speed and less manoeuvrability required. The caudal fin of Macropterygius from the Upper Jurassic of Solnhofen depicted by Abel (1912) has a further reduced skeleton in this area, as it can also be found in modern forms. This fact is an indication of low forces acting on this fin.

After the achievement of the optimum shape for fast swimming this kind of locomotion was maintained until the extinction of ichthyosaurs. As in different forms such as the underwater hunting rhamphorhynchoids (Ebel 1996) or the plesiosaurs modifications predominantly have to do only with adaptations to special kinds of food. For exmple, Eurhinosaurus from the Toarcian appears highly specialized (Riess 1986), with an elongated upper jaw and a relatively short lower one, a ‘rummage snout’ according to (Abel 1927). After Sander (2000) all Liassic ichthyosaurs already possess the characteristic tail kink, and it is by no means certain whether the reconstruction by Riess (1986) of the tail without a semilunate fin in Eurhinosaurus is really correct (Wade 1990). Without a specimen with the skin shape preserved it is not possible to state how propulsion was generated. Maybe Eurhinosaurus was not a specialized fish hunter, living off-shore, as supposed by v. Huene (1922).

Although localities with accumulations of ichthyosaurs in the Middle Trias as well as in the Lower and Upper Lias represent snapshots of the evolution, they make clear general tendencies of occurring modifications. It appears that early ichthyosaurs with amphibian ancestors first lived in rivers and specialized to hunting fish, later with improving swimming capabilies extended their milieu to the sea in accordance with the radiation of fishes, until with the achievement of the perfect swimming configuration life in the open seas had become possible. However, as the example of the Amazonas dolphin (Inia geoffrensis) demonstrates (Klima et al. 1980), a return from mere marine forms to rivers is not precluded and can lead to convergencies with fossil forms, which presumably came from smaller rivers and later adapted to such an environment.

Whereas birds as reptile descendants represent the peak of perfection on land, this applies to the ichthyosaurs regarding the possibilities of evolution in the water. They have perfectly taken advantage of hydrodynamic properties on the one hand, but on the other they were dependent for better or for worse on certain environmental conditions which ultimately turned out not to remain constant.