WHAT is it really like being a fish out of water? Well, you can鈥檛 breathe for one thing. Your skin dries up. So does the rest of your body because your kidneys aren鈥檛 designed to conserve fluid. And without legs you can鈥檛 head for home. Even if you could, how would you find your way? You can鈥檛 see or hear properly, and neither can you smell. Then there鈥檚 the problem of lunch. Your jaws aren鈥檛 made for catching and eating terrestrial fare. Anyway, your digestive system couldn鈥檛 cope with it. And if, against all the odds, you survive and find an equally stranded mate, reproduction is unthinkable.
Given all these obstacles, how on earth did anything emerge from the water to roam dry land? Until quite recently, the evolution of land animals was seen as a matter of necessity. In an increasingly arid world, so the thinking went, fish were forced to haul themselves out of pools that were drying up and go in search of new ponds. In the process of coping with drought, these resourceful creatures evolved the limbs, lungs and senses that made a permanent move to land possible. But this now seems implausible-these days we accept that evolution is blind. As Mike Coates of University College London notes wryly, it credits our fishy ancestors with remarkable foresight to launch themselves out of their natural habitat and go in search of waters new.
Over the past decade, new fossil finds have been used to decipher what really led animals to colonise the land. It鈥檚 a tale of fish fingers, rear-wheel drive and changing brains. Sometimes things were not quite what they seemed-like Ichthyostega, the missing link that wasn鈥檛. The puzzle is still far from complete, but it is becoming clear that most of the major changes needed for life on dry land happened in the water. The popular image of the first land animal has always been some sort of fish making its way up the beach, says Coates. 鈥淏ut I think if we looked at it, we鈥檇 see it and think `giant salamander鈥�.鈥�
Advertisement
The epic journey from water to land began about 400 million years ago. During the Devonian a new world emerged as complex plant ecosystems formed on land for the first time. 鈥淭he key step during the mid-Devonian was the appearance of wood, which allowed land plants to grow much larger than previously. This led to the evolution of the first trees,鈥� says Per Ahlberg of the Natural History Museum in London. The trees created shady areas and proper canopy cover on land, but they also changed the character of freshwaters-particularly the shallows. 鈥淲ith shoreline vegetation, fallen tree trunks in the water, and maybe areas of flooded forest, the water鈥檚 edge was becoming a new kind of environment that previously didn鈥檛 exist,鈥� says Ahlberg. This halfway house provided a natural staging post for the move onto land. Plants blazed a trail for our amphibious ancestors.
Just after the first full-size trees evolved, a group of lobe-finned fish called the panderichthyids emerged. Like today鈥檚 lobe-finned fish, which include coelacanths and lungfish, their fins were attached to mini-limbs that looked a bit like stumpy paddles. But even at this early stage-long before the evolution of legs and arms-panderichthyids had the body form of a four-legged land animal. In fact, with their flattened heads and long tails they were almost crocodilian in shape, apart from having fins instead of limbs. This body shape would have been ideal for exploiting a predatory niche in shallow water, according to Ahlberg. 鈥淎ll the principal players in this story have lungs already,鈥� he adds. Lungs pre-date the move to land by around 50 million years, and they may have evolved to allow fish to survive in oxygen-depleted waters. In any case, panderichthyids, with their lungs and gills, would have been able to breathe out of water as well as in it. They filled their lungs by forcing air in from the mouth-a technique known as buccal pumping which is still used by amphibians today. Panderichthyids also had their eyes on the top of their heads, rather like a mudskipper鈥檚, suggesting that they would have functioned above the waterline.
There is no doubt, though, that panderichthyids were still fish. So how do we get from them to the first creatures that walked the land on four legs, the tetrapods? In 1932, Gunnar S盲ve-S枚derbergh discovered some fossilised bones in Greenland that date from between 364 million and 354 million years ago. The animal, which was named Ichthyostega, was fishlike in many ways, but turned out to have legs, arms and digits. What鈥檚 more, it had powerful front limbs. 鈥淚t鈥檚 striking when you look at things like Ichthyostega and you see these massive forelimbs,鈥� says Coates. 鈥淭hey do indicate very strongly that this animal was spending some of its time out of water.鈥� These front limbs would also have kept Ichthyostega鈥檚 throat off the ground, improving the efficiency of its buccal pumping breathing system. Ichthyostega was hailed as the missing link.
But in 1987, Ahlberg and Jenny Clack of the University of Cambridge discovered some remarkably complete fossils on the barren shores of Greenland. Acanthostega is around the same age as Ichthyostega and is also a very primitive tetrapod, forcing the palaeontologists to rethink. Years of painstaking laboratory analysis have revealed that Acanthostega looked similar to the panderichthyids, except that it had limbs with digits instead of lobe-fins. The big surprise, however, is that this creature would have spent most of its time in the water. 鈥淲e didn鈥檛 expect to find Acanthostega having such a fish-like gill apparatus,鈥� says Coates, who described the material with Clack. 鈥淲hile it had lungs, it may not have been obliged to use them. The gill skeleton is an important part of our interpretation of Acanthostega as primarily, and primitively, aquatic,鈥� he adds.
Another surprise came when Coates and his team looked at the limbs. Acanthostega鈥檚 hind limbs were obviously robust, although the researchers suspect they were used to paddle through the water rather than walk on land. The forelimbs, however, had not evolved much from the structures found in lobe-finned fish. So, it appears that one of the earliest changes in the evolution of tetrapods was a move from front-wheel to rear-wheel drive. Fish tend to have their motor at the front in the form of powerful pectoral fins, but in Acanthostega and most other tetrapods, the power comes from the hind limbs. 鈥淭etrapods normally walk around with this sort of wheelbarrow effect,鈥� says Coates. 鈥淭he drive is at the rear, the front is the suspension.鈥�
Fast-forward
With the discovery of Acanthostega, there has been a real fast-forward in our knowledge. For a start, the so-called missing link, Ichthyostega, with its massive forelimbs, doesn鈥檛 fit the emerging pattern of development and has been relegated to a side branch off the main evolutionary tree. And other early tetrapods are turning up. 鈥淥nce we realised what Acanthostega鈥檚 jaw looked like, that was the key to finding a whole lot of other Devonian tetrapods that were only represented by jaws,鈥� says Clack. These include Elginerpeton, the most primitive tetrapod we know of, which dates from between 370 and 364 million years ago.
Elginerpeton was unearthed in the 19th century near Elgin in Scotland, but it took Ahlberg鈥檚 trained eye to recognise its significance. It is a monstrous predator-at least 1.5 metres long-with teeth that are much more fishlike than those of most other tetrapods. Comparative anatomy and computer analysis reveal that Elginerpeton is at the base of the tetrapod family tree, but its evolutionary history is not straightforward. The hind limb was 鈥渧ery good as a paddle but pretty lousy as a walking limb鈥�, says Ahlberg. Yet some of the skeleton is adapted for bearing weight, which suggests that at some point it came onto dry land. It seems that Elginerpeton evolved to become more aquatic again after some initial adaptation to life on land.
With more specimens to work from, researchers have now been able to compare early tetrapods-with some unexpected results. Although the limbs are often quite similar, for example, there is a surprising variety in the number of digits these animals possessed. Nowadays, the basic kit of fingers is five, although some animals, such as chickens and horses, lost digits as they evolved. Early tetrapods were better endowed: a Russian specimen called Tulerpeton had six digits on each limb, Ichthyostega had seven on its hind limbs and Acanthostega had eight digits on both arms and legs. This seems to be a time of evolutionary experimentation, before natural selection settled on the number five. 鈥淲hen I started work on the Acanthostega,鈥� says Coates, 鈥淚 didn鈥檛 know if I鈥檇 find any digits at all-let alone eight. Then, following up a lead on obscure Russian papers on Tulerpeton, I started on the Ichthyostega hind limb and found the seven-digit pattern.鈥�
The first fingers may have evolved by a lucky accident of misplaced genes, when a control gene responsible for coordinating brain development was duplicated elsewhere in the genome. Earlier this year, Susan Dymeki and her colleagues at Harvard Medical School reported that mice with a defective version of this gene have normal brains, but hardly any sign of digits. It seems that nature used an old tool to create a new and highly successful invention in body design. But what benefit would fingers have been to our fishy ancestors? Ahlberg points out that the water margins were probably weed-choked, and that digits would have been very handy to manipulate the thick vegetation, allowing early tetrapods to move around more easily.
Clack is looking at skulls and starting to build up a picture of the evolution of the senses, particularly hearing. She has studied the main ear bone-the stapes-in early tetrapods and has followed its development through to the first fully terrestrial forms. The bone that developed into the stapes was originally part of the gill system in fish. 鈥淭he transition between the fish-like condition in panderichthyids and the tetrapod-like condition in Acanthostega must have been really quite rapid,鈥� says Clack, 鈥渁nd it seems to have occurred at the same time as changes to the limbs.鈥�
Our own tiny stapes-the stirrup-shaped bone in the middle ear-helps us hear a pin drop, but in early tetrapods the stapes is rather large and stout. It 鈥渕ay be to do with hearing low frequencies, perhaps in water鈥�, says Clack. 鈥淚t only gradually gets converted into something which is suitable for airborne sound.鈥� Clack has also been looking at the optic region in the skulls of early tetrapods. The trend is for the eyes to get bigger and move from the sides towards the top of the head. Other researchers have found evidence for retractor muscles, which in modern tetrapods pull the eyes into the head and operate the translucent nictitating membrane that forms the inner eyelid.
Overall, what鈥檚 remarkable is the similarity between the various early tetrapods. This may be because they shared the same niche. 鈥淭he tetrapods from the Devonian are largely aquatic,鈥� says Clack.
Mind the gap
All were big predators, measuring over a metre long, with tail fins, more than five digits and the crocodilian body form. The fossil record suggests that tetrapods were fashioned according to this template until the beginning of the Carboniferous, about 354 million years ago. For the next 20 million years the record is silent. This period, known as the Tournaisian, must have seen a huge diversification of the tetrapods, because after this gap in the record we see a variety of body forms that are much more like today鈥檚 land animals.
Fossils from East Kirkton in Scotland date from just after the Tournaisian gap (快猫短视频, 12 February 1994, p 21). They reveal that by about 338 million years ago, tetrapods had conquered land and lived fully terrestrial lives. During this time, evolutionary experiment must have been rife to give rise to such a dazzling array of animals. 鈥淚t鈥檚 telling us something about a major radiation [evolutionary divergence] down there,鈥� says Coates. 鈥淭here鈥檚 more data coming in all the time [and this] will give us clues about what we might expect to find in the Tournaisian.鈥� Clack describes unravelling what was going on at this time as 鈥減robably one of the most difficult problems in early tetrapod palaeontology鈥�.
Luckily, a few clues are emerging. Clack is currently working on a fossil called 鈥淧eter鈥�, from Dumbarton in Scotland. 鈥淲e鈥檝e got this articulated specimen that comes right from the middle of the gap. It鈥檚 quite a large creature, about the same size as some of the Devonian beasties and it looks like it鈥檚 more terrestrial than its aquatic descendants,鈥� she says. 鈥淣ew Devonian tetrapods are cropping up all over the world, now people know what to look for. I鈥檓 hoping that鈥檚 what鈥檚 going to happen to the Tournaisian now.鈥�
The Tournaisian may be the largest gap in the record of tetrapod evolution, but it is not the only one. Ahlberg is looking for missing links between the most primitive tetrapods and their closest fish relatives. He has just published his analysis of a new specimen from Latvia, which fits between the panderichthyids and Elginerpeton. Its jaw is very unusual, with a series of seven rows of sharp pointed teeth. 鈥淚t鈥檚 clearly some sort of early experiment with a new mode of feeding,鈥� says Ahlberg, who has never seen anything like this before.
With the Latvian fossil and Clack鈥檚 new Tournaisian specimen, our understanding of how animals colonised the land is constantly evolving. And there is more to come. Clack and Ahlberg are preparing to re-describe Ichthyostega. For the past 70 years, a single research group in Stockholm has held most of these fossils and much of the material has not been scrutinised. 鈥淭he current reconstructions are wrong in almost every respect,鈥� says Clack. She and Ahlberg will first focus on the brain case. 鈥淲e already know how weird that is and we鈥檝e got some ideas about what it鈥檚 going to tell us,鈥� says Clack. After that they will look at other areas including the neck and pelvic regions, which are hardly known. 鈥淧otentially we鈥檒l have to rethink the whole thing again with the new Ichthyostega study coming up,鈥� she says.
