1. Fish to amphibian

AcanthostegaEusthenopteron (fish) → Tiktaalik (fish) → Acanthostega (tetrapod)

The transformation of fish into the kind of amphibians we see in the mid Palaeozoic is not inconceivable, but if it did happen, it is difficult to conceive of its happening by any random process. Water and air are very different environments. Water is liquid H2O, but the oxygen component is unavailable for life; fish depend on the oxygen dissolved in the water, which varies with temperature and depth. At a temperature of 15° C close to the surface, it is around 10 mg/L (milligrams oxygen per 1000 g water) – thus 1 part per 100,000 by weight. The concentration in air is 23 parts per 100 by weight. In addition, animal bodies are buoyant in water because they themselves mainly consist of water; in air the full force of gravity is apparent. On land, animals need legs to move around, or, in the case of snakes, which no longer have legs, muscular scales. A shift from water to air therefore involves a radical re-organisation of embryonic development and final design.

Modern theory prefers to designate the new animals as ‘tetrapods’ – vertebrates with limbs and feet – rather than ‘amphibians’. That is because, contrary to the expected story, the first four-footed animals appear to have acquired their feet while living entirely in the water. Hence tetrapods include four-legged aquatic animals such as newts and salamanders as well as exclusively terrestrial animals. The term does not of itself establish that all animals with four limbs have a common ancestor.

Background: fish living and extinct

By the time the first tetrapod fossils appear, nine-tenths of Earth’s preserved geological time had already passed, and of that nine-tenths, fish fossils occur only in the last twenty-fifth. As with all the animals that characterise the Cambrian Explosion, the origins of fish are obscure. Even if we confine ourselves to relationships within the group, fishes take so many different forms that untangling their relationships is a challenge, and schemes for classifying them have changed over the years. Currently, living fish are divided into 5 classes:

The pike Esox - a ray-finned fish

  • Hagfish (Myxini)
  • Lampreys (Petromyzontida)
  • Cartilaginous fish (Chondrichthyes)
  • Lobe-finned fish (Sarcopterygii)
  • Ray-finned fish (Actinopterygii)

As the exact number is not crucial, we will assume here that each of these classes stems from a common ancestor: that the fish in the seas today arise from no more than five lineages. That is of course a vast amount of evolution, and of a most wonderful kind. Whether the five classes were themselves interrelated is impossible to test because the fossil material prior to their appearance is too poor.

Classes subdivide into orders. The hagfish and lamprey classes consist of only 1 order. The cartilaginous fish comprise 14: chimaeras, sharks of various kinds, skates, sawfish and rays. The lobe-finned fish consist of 2 orders: coelacanths and lungfish. The ray-finned fish comprise over 40 orders and now encompass the vast majority of fish in the sea, from seahorses to eels, being more diverse in number of species than all mammals, birds and reptiles put together. Astonishingly, most of that diversity arose late in geological time, during the Late Cretaceous and Cenozoic, when oceans began to cool and deepen. The delay, followed by such a dazzling display of creativity, suggests that they were programmed to diversify at that point.

The gaps between classes are greater than the gaps between orders within a class. Early on, there were more classes than there are now, for several became extinct, notably:
  • Conodonts
  • Pteraspidomorphs (including the heterostracans)
  • Anaspida
  • Osteostracans
  • Placoderms
  • Acanthodians

Acanthodians probably descended from fish ancestral to the cartilaginous group (Coates et al. 2018). The general evolutionary scheme, which is probably correct, is that fish were initially finless and jawless, gradually evolving fins and jaws as competition for resources grew.

From earliest to latest, the first fossil occurrences of these eleven classes, some living, some extinct, are as follows (dates in millions of years, radioisotope chronology):

 Class  Period  Date  Type  Genus
 Conodonts  Cambrian  500 jawless various
 Pteraspidomorphs  Ordovician  480 jawless Arandaspis
 Osteostracans  Silurian  440 jawless Ateleaspis
 Acanthodians  Silurian  440 jawed various
 Anaspids  Silurian  435 jawless Pharyngolepis
 Lobe-finned fish  Silurian  420 jawed Guiyu
 Placoderms  Devonian  416 jawed various
 Cartilaginous fish  Devonian  409 jawed Doliodus
 Ray-finned fish  Devonian  390 jawed Cheirolepis
 Lampreys  Devonian  365 jawless Priscomyzon
 Hagfish  Carboniferous  300 jawless Myxinikela

These are not necessarily the times when the class originated. Isolated teeth, scales and bone fragments often occur somewhat earlier, shark scales, for example, are known from the Late Ordovician. Not all fish have teeth, scales or bone, and it should be remembered that jawed bony fish are easier to fossilise than jawless cartilaginous fish (such as lampreys). The earliest fish were certainly jawless. Jawed fish did not become dominant until the second half of the Palaeozoic. If the ten classes excluding acanthodians are not interrelated, then jaws evolved independently five times. Independent, parallel evolution (convergence) is a common phenomenon in phylogenetic trees.

The existence of more classes in the past than in the present also does not conform to the theory of evolution, which predicts that morphological disparity should increase with time.

The general pattern is that fish fossils were initially rare, reflecting a world where fish were few and far between and unstable coastal environments tended to erode and destroy fragile carcasses. The colonising fauna came from the open sea, where the floors beneath them were hot, basaltic wastelands. Since the only plentiful food available was algae and microscopic plankton, they had no use for teeth and jaws. Only as near-shore environments became more habitable did they acquire equipment enabling them to nibble seaweeds, filter sediment, crack open shells and indeed capture other fish. It wasn’t a process of natural magic; they were programmed to do so, each according to their kind. As the available niches became more various, as environments began to be vegetated and to host a range of invertebrate life, so the fish moving into them became more specialised.

Lobe-finned fish
The class that presents the smallest apparent gap between fish and tetrapod, at the appropriate time, was the lobe-fins. They had fleshy pectoral and pelvic appendages that might, potentially, have developed into limbs, and they were, in the Devonian, the most diverse fish group, offering a range of forms from which to choose the most tetrapod-like. Eusthenopteron and Panderichthys were two such fish:

Eusthenopteron is sometimes figured as crawling out of the mud of a Devonian lake, apparently with the intention of finding another pool to swim in. Because Eusthenopteron was once cast in the role of the ‘ancestor’ of tetrapods, tetrapodlike behavior was attributed to it. However, taking the whole morphology of the fish, with its streamlined torpedo shape, and dorsal, anal, and pelvic fins placed near the back of the body, it seems that the lifestyle of Eusthenopteron was much more like that of a modern pike (Esox), a fully aquatic lurking predator.

Jennifer Clack, Gaining Ground, p. 63 (2002)

In the absence of better candidates, Eusthenopteron was was regarded as well on the way to becoming a tetrapod, and was portrayed as such in illustrations. Now, however, it is no longer seen as a convincing intermediate and has been relegated to a side branch. So has its contemporary, Panderichthys. ‘Palaeontologists didn’t previously have a decent fossil representing the intermediate between finned fish and four-footed land animals,’ admitted Bob Holmes in the New Scientist. But the admission was made only after the discovery of Tiktaalik.

Tiktaalik roseae
This fossil, a large Late-Devonian fish, undoubtedly narrowed the gap between the fish Panderichthys and the tetrapod Acanthostega. The scientific report appeared in the journal Nature (6 April 2006) and news of it was immediately promulgated by press and radio across the world. The Guardian went so far as to say it was one of the most important fossil finds in history: a missing link between fish and land animals showing how creatures first walked out of water and onto dry land. Nature itself was scarcely more phlegmatic. In the covering news and views item Jennifer Clack and Per Ahlberg said they expected it to acquire the same iconic status as Archaeopteryx. Was the hype justified?
The authors, Edward Daeschler, Neil Shubin and Farish Jenkins, introduced their find as follows:

The relationship of limbed vertebrates (tetrapods) to lobe-finned fish (sarcopterygians) is well established, but the origin of major tetrapod features has remained obscure for lack of fossils that document the sequence of evolutionary changes. Here we report the discovery of a well-preserved species of fossil sarcopterygian fish from the Late Devonian of Arctic Canada that represents an intermediate between fish with fins and tetrapods with limbs, and provides unique insights into how and in what order important tetrapod characters arose. Although the body scales, fin rays, lower jaw and palate are comparable to those in more primitive sarcopterygians, the new species also has a shortened skull roof, a modified ear region, a mobile neck, a functional wrist joint, and other features that presage tetrapod conditions.

Nature 440:760Tiktaalik was a lobe-finned fish, in the same broad group as coelacanths and lungfish. In the Devonian period lobe-finned fish were more diverse than today. The new genus/species added to that diversity. And one may say at the outset, although it is still a fish, Tiktaalik does permit interpretation as an animal transitional to the earliest fossilised aquatic or semi-aquatic tetrapods. Shared features include: a lengthened snout (measured from the eyes to the tip of the skull), a mobile neck (facilitated by the loss of certain bones at the back of the skull), overlapping (‘imbricate’) ribs and a pectoral girdle that may have given it an ability to lift the front part of its body by its fins. Stratigraphically, it appears after Pandericthys and before the first four-limbed animals, Acanthostega and Ichthyostega, though the gap is only (in Darwinian terms) around 5 million years. Since these two tetrapods were very different from each other, presumably their Tiktaalik-like ancestor evolved in two separate directions, with Acanthostega on one branch and Ichthyostega on the other.

Imbricate ribs do not occur in any other fish, nor in Acanthostega, but do occur in Ichthyostega. Since even that animal was primarily aquatic, they are unlikely to have played a role in supporting the weight of the animal. The stiffening of the spine produced by this arrangement would have inhibited horizontal flexion (the wiggling motion of most fishes) just as it would have done in Ichthyostega (Clack 2006), and this fact alone is sufficient to explain the mobile neck and fin joints: without some compensating mobility the animal would not have been viable as a predator. In overall morphology, the ribcages of Tiktaalik and Ichthyostega are unlike each other.

The fin bones terminated in the fanned spines and rays known as lepidotrichia, typical of fish fins. The pectoral fin differed substantially from the forelimb of Acanthostega and more closely resembled that of the lobe-fin Sauripterus than it did the forelimb of Acanthostega (the forelimbs of Ichthyostega are incomplete). Shubin et al. comment:

A Devonian rhizodontid, Sauripterus, is known to possess digit-like radials, but phylogenetic analyses indicate that this group is not the closest relative of tetrapods.

The pelvic fin was not capable of bearing significant weight. In contrast to both tetrapods and lungfish, pelvic girdle’s two halves were unfused and they were not attached to the spine. On the other hand, whereas in Panderichthys the pelvic girdle was smaller than the pectoral girdle, in Tiktaalik their size was about the same. The corresponding fins may or may not have been similar in size: the preserved lepidotrichia suggest that they were, whereas the pectoral fins of Elpistostege, Tiktaalik’s closest relative, were three times bigger than the pelvic fins. In the early tetrapods the hindlimbs were bigger and more powerful than the forelimbs.

Diagram adapted from p 747, showing presumed lineage from fish to tetrapod. Skull roofs (left) show loss of the gill cover (blue), reduction in size of the postparietal bones (green) and gradual reshaping of the skull. Drawings not to scale.The animal does not fall plumb in the middle between Panderichthys and Acanthostega. It is a fish, albeit an unusual one, and while there is only a relatively small gap between Panderichthys and Tiktaalik, a big gap between Tiktaalik and Acanthostega remains. In a significant number of features, body scales, fin rays, lower jaw, palate, the fossil resembles sarcopterygians that are considered to be evolutionarily less advanced. As Ahlberg and Clack say, ‘we have almost no information about the step between Tiktaalik and the earliest tetrapods, when the anatomy underwent the most drastic changes.’ It may be reasonable enough to link the first pair into a group that had a common ancestor (the ‘elpistostegalian fish’), but it requires a lot more extrapolation to link the second pair together.

There is, moreover, a large gap after Ichthyostega. It often fails to be mentioned that, so far as the fossil record is concerned, both Acanthostega and Ichthyostega are dead-ends. They shed no light on the origin of modern tetrapods, aquatic or terrestrial.

Did Tiktaalik walk on its fins?

According to Nature‘s announcement of the fossil, Tiktaalik was a fish that crawled out of the water, but if so, the lineage must have crawled back again, for Acanthostega ‘rarely, if ever, made forays onto dry land and its legs would have almost certainly been incapable of supporting its body had it done so’ (Clack 2002, p 128f). That’s less than was claimed for this animal. We were told in one breath that Tiktaalik had a long snout that would have been suited to catching prey on land, in the next that it hauled itself onto land only to escape predators.

Walking is not an unknown behaviour with certain kinds of fish. In her monograph Clack mentions several examples, including the frogfish, which uses its jointed, grasping fin-rays to cling to the weeds amongst which it lurks for prey. Handfish walk along the bottom using their leg-like fins and are so-named because the extremities of their rear fins resemble human hands. Epaulette sharks will even crawl out of the water and in search of prey use their fins to propel themselves over exposed coral rock. Here is a report of a monkfish 380 metres below sea-level (Laurenson 2005):

One of the monkfish was observed to move along the seabed for several metres at a time by walking. Several ‘walks’ were observed. The gait involved both the pelvic and pectoral fins and the body and tail were lifted clear of the seabed. The pelvic fins appeared to be the main weight bearing fins lifting the body up from the seabed. They also seemed to be responsible for a considerable proportion of the forwards propulsion.

Frogfish, handfish and monkfish are members of the order Lophiiformes or anglerfish, within the subclass of teleost fish, within the class of ray-finned fish. The evolutionary history of these walkers is obscure but we may presume that their particular forms are the result of immense diversification over time within the order. Certainly their appendages are quite unrelated to those of lobe-finned fish. Anglerfish are an example of how wonderfully unconventional life can be, as if whoever designed them was a non-conformist who delighted in subverting stereotypes.

Lophius piscatorius. Click on image for source and larger version of photo.Thinking that was less ideologically driven might not go amiss in the present case. While locomotion may have been a function of Tiktaalik’s front fins, habitat indicated by the geological setting was a floodplain, dissected from time to time by channels. Other fossils recovered from the locality include placoderms and a variety of other lobe-fin species. Out of water, Tiktaalik could, at best, only have dragged its 2½-metre-long body. Monkfish, which like Tiktaalik have a wide mouth, flattened skull and eyes on top of the skull rather than on the sides, again alert us to other possibilities, for in addition to walking they use their fins to scoop out sediment from under their bodies and make a hollow, in which, having camouflaged themselves to match their surroundings, they lie in wait. At the right moment, they strike the prey by suddenly pushing with their fins upwards and forwards. The feeding strategy of Tiktaalik won’t have been the same as the monkfish’s, but it may have been similar. Several lineages of fish have even evolved an ability to crawl out of water, including mudskippers, rockskippers and blennies; but there is no reason to think they are on the way to leaving the water altogether.

More like a red herring
As discussed elsewhere on this site, the question of how one fills the gap between Panderichthys and Acanthostega is ultimately a side issue for the thesis that lobe-finned fish evolved into land-dwelling tetrapods. Apart from the problem of identifying Acanthostega’s descendants, crucial questions include:
  • The characteristics shared by the Devonian tetrapods and certain lobe fins were those that equipped them for their shallow-water environment. They differed from each other because they were each optimised for their own particular life-style (i.e. diet), not because they were evolutionarily on their way to somewhere else.
  • The first known tetrapods with limbs and digits lived wholly (Acanthostega) or primarily (Ichthyostega) in the water, so the ‘acquisition’ of such appendages does not support the story of how ‘our fishy ancestors began hauling themselves onto dry land’. Their splayed limbs were designed for paddling, not walking.
  • There remains a large morphological gap between Tiktaalik and the first tetrapods.
  • Trackways attesting the existence of tetrapods predate Tiktaalik, let alone Acanthostega and Icthyostega, by 10-20 million years, so Tiktaalik cannot be ancestral to the first tetrapods.
  • How does one fill the gap between Tiktaalik, which was a fish with no legs, and an aïstopod such as Lethiscus, which within 20 million years had supposedly acquired legs and limb girdles and then lost them again, and changed from a fish to something more like a snake than any tetrapod?
As Clack remarked in a paper a few months before Tiktaalik, Lethiscus suggests that ‘a great deal happened in the course of tetrapod evolution that we know very little about’. This is one of palaeontology’s trade secrets to which Darwinians do not draw public attention. Ahlberg (2019) expressed the problem in these terms:

There is an odd sense of unease pervading the research community. It has been brought about by a head-on collision between an established interpretative scenario for the origin of tetrapods, seemingly well-grounded in the body fossil record, and two sets of tetrapod trace fossils from Poland and Ireland that appear to flatly contradict it.

Tiktaalik therefore could not have been an intermediate ancestral to the first tetrapods. In a radical revision of the orthodox scenario, he argued that tetrapods and Tiktaalik-like lobe-fins must have

originated no later than the beginning of the Middle Devonian, and coexisted as ecologically separated radiations for at least 15 million years; … that all Devonian and many later tetrapods … retained a permanent dependence on the aquatic environment, allowing only relatively brief terrestrial excursions; that this continued dependence explains the repeated evolution of ‘secondarily aquatic’ forms among them; and that all fully aquatic tetrapods, including Acanthostega, are likely to represent reversals from a somewhat more terrestrial ancestry.

In other words, the fossils of the Late Devonian were evidence of an earlier movement towards terrestrialisation that the fossil record failed to capture. Not only was that process never completed, but by the Late Devonian/Early Carboniferous it had gone into reverse. Consistent with the evidence that lepospondyls such as Lethiscus are further examples, Ichthyostega and Acanthostega represented a loss of terrestriality.

Another view

The phrase ‘tetrapod-like’ implies that we are seeing something evolving towards the tetrapod state, but the significance of such features is as much ecological as evolutionary: they were what a fish needed in order to exploit the new ecosystems of the vegetated shallows. The floodplains were transitional environments between sea and land, and it could be that some fish species, amongst the huge range of forms arising in the Devonian, acquired tetrapod-like features simply because that is what the exploitation of such environments required. The genetic programs which produced the innovations were engineered with the future environments already in mind. Such evolution would be quite distinct from the fumbling-in-the-dark scenario of random mutations, and biologists are beginning to get glimpses of how genetically this might have occurred. In the light of Tiktaalik and similar fish it seems more likely than not some lobe-fins did evolve into aquatic tetrapods. However, this would mark the limit of the program’s ability to innovate in that direction. What has never been demonstrated is how aquatic tetrapods evolved into terrestrial tetrapods (i.e. amniotes). Acanthostega and Icthyostega are both dead-ends.

It is not impossible. Since evolution does not happen by magic, such a path could have been directed by a genome programmed to make it happen. However, in the absence of any evidence for such a transition, it seems better to suppose that amphibians and amniotes had separate ancestries – what one might term the Noah’s Ark scenario. Aquatic tetrapods could not have been on the ark, since the purpose of the ark was to preserve reproducing pairs of terrestrial animals. By the Devonian those animals were beginning to spread abroad. Just like the terrestrial invertebrates, terrestrial vertebrates had an entirely independent origin. The aquatic tetrapods therefore must have originated from fish in the sea. However, they could not always have been tetrapods since, before the Devonian, the environments to which they were adapted – swamps, lagoons, estuaries, wetlands inhabited by other fish – did not exist. In any case, they would have been destroyed in the Hadean Cataclysm.

Unfortunately, the National Academy of Sciences is intent on blurring the differences. Tiktaalik, it says, had both the ‘features of fish (scales and fins) and features of land-dwellers (simple lungs, flexible neck, and fins modified to support its weight). The bones in the limbs of this fossil, named Tiktaalik, resemble the bones in the limbs of land-dwelling animals today.’ This is not a truthful statement. Lungs, simple or otherwise, were not mentioned in the scientific report. A mobile neck (‘flexible’ is too strong, for even Acanthostega did not have a flexible neck) might have helped the animal to gulp air, or even raise its head above water and target prey by the water’s edge, but it was not a uniquely tetrapod feature. The 2-metre-long lobe-fin Mandageria also had a mobile neck (an example of ‘parallel evolution’), as does the modern eel catfish. And Tiktaalik’s fins had no ability to support its weight out of water. Is it only by making misleading if not manifestly false statements that the Darwinian story can be maintained?