Australia: The Land Where Time Began

A biography of the Australian continent 

Acanthostega

This was the second tetrapod from the Devonian from East Greenland to be described (Jarvik, 1952), what could be determined from 2 partial skull roofs was all that was known for many years, though this information did demonstrate that it was a very different animal from its contemporary. The original description of the of Acanthostega was based on a specimen that shows the right side of the midline bones of the skull that was more or less complete in which the maxilla bearing teeth was to some extent detached from the remainder of the cheek. The skull roof bones showed a pattern that differed in some respects, at the back edge of the skull table in particular, where in Acanthostega there was a deep embayment and a pronglike tabular “horn,” formed by the tabular bone, which is a distinctive structure that gave it the name “spine armour.”

Spectacular new material of this animal was found when an expedition followed up on some accidental finds in 1970 of by a sedimentologist of Acanthostega material. Among the specimens collected by the 1987 expedition were the bodies of the animals whose skulls had been collected in 1970. Following on from these accumulated finds in 1970 and 1987 the nearly complete skeleton of Acanthostega is now known, the descriptions of these skeletons being the basis of a series of papers. According to Clack¹ the picture of the animal that has emerged from these descriptions is much more like an animal that would be expected to be a primitive tetrapod than is Ichthyostega. Some ideas about what early tetrapods were like have been changed by study of this animal and this study has also led to some new ideas being proposed concerning the development of limbs, the evolution of the ability to breath directly from the air, as well as the tetrapod radiation that has subsequently continued from the Devonian to the present.

Clack¹ suggests Acanthostega and Ichthyostega have almost nothing in common, apart from primitive similarities. The skull of Acanthostega was quite shallow, and it had a snout that was rounded bluntly (Clack, 1994a; 2002a, 2003a), as is the case in Ichthyostega. A bony sheet formed the palate in Acanthostega, as was also the case in Ichthyostega, though not attached as solidly to the remainder of the skull as it did in Ichthyostega. The external nostril was situated close to the edge of the mouth, as it did in Ichthyostega, and the lateral line organs were in most cases housed in tubes that threaded through the bone, though these tubes merged into grooves at some places, the pores allowing them access to the outside are more conspicuous than they are in Ichthyostega. There was a sculptured surface of pits, ridges and furrows in both animals on the skull and dermal shoulder bones, in common with other early tetrapods, as well as many contemporary fishes. It is unknown what the function of this ornamentation was, though similar surface ornamentation is seen on certain extant fishes, especially catfishes.

Acanthostega, that was smaller than Ichthyostega, had a skull measuring about 200 mm at maximum known size, and a different diet from that of Ichthyostega, as suggested by its teeth. The marginal teeth of the outer row were much smaller than those of Ichthyostega, while on the palate the inner row of teeth were smaller toothlets and denticles, but with some large fangs. The contrast in the upper and lower teeth is less noticeable in Acanthostega than in Ichthyostega. Acanthostega appears to have been completely aquatic, based on nearly all its features, Clack¹ suggesting that its teeth, that resembled those of many contemporary lobe-finned fishes, were probably adapted for feeding in water.

Many of the Acanthostega skulls recovered in Greenland had fallen apart along the midline. When in 2002 Ventastega material was collected it became obvious that there was little if any contact between the 2 nasal bones, or between them and the premaxilla. In Acanthostega there was an internasal fontanelle, though it was somewhat narrower, but it was overlooked until the Ventastega skull roof was discovered. The suturing of the bones to each other was particularly secure in other parts of the roof of the skull, e.g., the skull table, and interdigitations interfingering in 3 dimensions. There are broad overlapping areas on either side of the snout where the sutures have been designed to resist twisting of the snout during feeding (Clack, 2002a, 2003a). The types of sutures present in different parts of the skull have been investigated by recent work (Markey & Marshall, 2007).

Several of the specimens of Acanthostega have been preserved in such a good state that some of the more delicate parts of the skeleton that are not usually preserved have been found. The gill Skelton of Acanthostega has proven to be remarkably fishlike (Coates & Clack, 1991), with each branchial element being strongly ossified and having a deep groove running down the back. Elements from at least 3 and probably 4 of the gill arch pairs have been recovered. When these are compared with those of modern fishes for which the gill arches are known the gill bars of Acanthostega resemble most closely those of Neoceratodus, the Queensland lungfish (Australian). This fish has lungs for breathing air, though it still retains internal gills that it uses for breathing in water as do most other fish. This fish has 3 pairs of functional gill arches, as is found in Acanthostega. In the lungfish the afferent branchial artery which takes blood to the gills for oxygenation is housed by the groove along the back of the individual elements, thus providing 1 piece of evidence that suggests Acanthostega breathed with internal gills similar to those of Neoceratodus as part of its breathing mechanism.

The shoulder girdle provides more evidence supporting this conclusion. The shoulder was comprised of a stout cleithrum that was bonded to the endoskeletal scapulocoracoid in Acanthostega as it does in Ichthyostega. The dermal clavicle sheathed the entire unit across the ventral surface and up the lower part of the leading edge of the dermal clavicle. In Acanthostega there was an interned flange, the postbranchial lamina, along the leading edge of the cleithrum that was slightly hollowed out from the front, a structure present in most fishes (Coates & Clack, 1991), which has as number of functions. The shoulder is given vertical strength by its angled cross section, and it also forms the back of the gill chamber, where its shape helps direct water out after it has passed over the gills. This adds to the evidence from the gill elements to strongly suggest Acanthostega still used internal gills for breathing, though Clack¹ suggests there is no doubt it would also have used lungs as does Neoceratodus, as lungs or their equivalent appear to have been a common heritage in all bony fishes. The presence of the anocleithrum as part of shoulder girdle was another fishlike character that is a remnant of the more extensive dermal shoulder girdle of fishes (Coates, 1996). Ichthyostega appears to have lacked a postbranchial lamina in the shoulder girdle, and an anocleithrum has not been found, though Clack¹ suggests this may have more to do with the degree of ossification of the elements or the preservation state of the specimens than with its lack in the living animal.

The surface of the glenoid facet of Acanthostega was not strongly contoured, being a relatively flat oval that was directed to the side of the animal, which contrasts with the situation of in Ichthyostega, which suggests its forearm was also directed almost entirely sideways, though Clack¹ suggests it may have had more freedom of movement to the front than did that of Ichthyostega. A number of the most remarkable new facts about the tetrapods of the Devonian have been revealed by the forearm of Acanthostega (Coates & Clack, 1990).

In Acanthostega the humerus is shaped like a broad L, as occurs in most other tetrapods, though a number of foramina and processes were present, as is the case in Ichthyostega, that are lost in the evolution of tetrapods as a whole. It was found, in the same study that considered the growth patterns of the humerus of Ichthyostega, that specimens of Acanthostega also represented a range of sizes from subadult to presumed adult, or again, the smaller examples were 2/3 as large as the known maximum. In Acanthostega there was no evident change of shape or proportion, and no reduction of foramina size, the bone getting larger instead, which differs from the case in Ichthyostega, though the bone exhibited a number of features that included a latissimus dorsi process and a deltopectoral crest for the attachment to the shoulder. These, as well as an almost total lack of a ventral ridge in Acanthostega, which are features that are shared with tetrapods of the Carboniferous and Tulerpeton, though they are lacking in Ichthyostega and tetrapodomorph fishes.

In Acanthostega the bones of the lower arm, the ulna and radius, differ in several ways from any other known tetrapod, the radius being an elongate element that is cylindrical in section near the point of articulation with the humerus, though at the distal end it is flattened and spatulate, and was almost twice as long as the ulna. The ulna was short and flattened and had an articulation that was strap-shaped where it joined the humerus, though it had a narrow articulation where it joined the wrist. The leading edge that faced the radius was straight and the back edge was curved, and there was no prominent olecranon as occurs in Ichthyostega.

These shapes are unique among tetrapods, though they reflect very closely those found in lobe-finned fishes such as Eusthenopteron and Tiktaalik. Clack¹ suggests that the most plausible interpretation is therefore that the shapes and proportions are those of a primitive forelimb that has retained several characters that are fishlike, and not a forelimb that has been adapted from the forelimb of a tetrapod that is more “conventional.” This conclusion is important for the interpretation of many aspects of Acanthostega morphology, including the hand and digits, which is probably the most surprising aspect of Acanthostega.

The intermedium, which is attached to the ulna, is the only true wrist bone that is known, though a number of other possible elements have been discovered that may be wrist bones, but when they were found they were jumbled up making it impossible to know where they were positioned in life. Because of the discrepancy in length between the radius and the ulna indicates that the wrist must have broad and inflexible, the overall impression being that it is difficult to imagine how it could have functioned as a weight-bearing joint.

One specimen that has been found has 8 digits in full articulation at the end of the arm, each being complete, and none being a duplicate of any of the others. They are in a regular array, at the leading edge there are 3 small ones and at the rear are 2 slender ones. It is uncertain if the digits had been enclosed in any type of webbing in life, though Clack¹ suggests they might have been as they are still in articulation in the fossil. The individual phalanges are basically a simple cylinder that is slightly flattened with a small amount of expansion at each end where it attaches to its neighbours. These appear to be unquestionably adapted for use as a paddle and not a walking leg when the structure of the joints and the arrangement of the digits are taken into consideration. There are several ways in which this discovery has proven to be sign significant, as well as the original function of limbs with digits, also in connection with ideas concerning embryonic development of limbs.

A number of Acanthostega specimens have preserved the hind limb and pelvic girdle (Coates, 1996). When compared with Ichthyostega and some Carboniferous species, the pelvic girdle was relatively small for a tetrapod, though it was larger than those of fishes, as well as being more tetrapod-like than fishlike in other respects. As with the pelvic girdle of Ichthyostega it was a single element, and similarly had 2 dorsal processes, one of which was short and stout and was directed upwards and a long slender one that was directed to the rear, this rear one being at an angle of about 45^o to the plane of the vertebral column. The process that was more dorsally directed would have been attached on its inner surface to the vertebral column, but with no facet for articulation of a sacral rib. It therefore appears that ligaments must have formed an attachment. The sacral rib which linked the girdle to the vertebrae had been modified to a small degree for this function, though its end was slightly expanded to provide greater surface area for attachment.

Acanthostega had a femur that was somewhat longer than the humerus on the proximal surface of which was a large rectangular adductor blade, this flange being the attachment surface for the adductor muscles that pulled the leg inward and backward. It is suggested by the size of the flange that the adductor muscles were large and powerful, but it is suggested by of their position they were used in swimming. The tibia and fibula, the bones of the lower leg, were flattened and there was a slight overlap between them along their length, as is also seen in Ichthyostega, and as also in Ichthyostega, flattened elements formed the ankle. Though the hind limb is known only from a disrupted specimen that they are likely to have had 8 digits on the hind limb as are present on the forelimb, based on a judgement that Clack¹ says is rather conservative. On this limb a small digit is located at each edge of the foot and a second small one contributing to the leading edge.

In both Acanthostega and Ichthyostega this structure strongly suggests the hind limb was largely used as a paddle, the powerful action produced by the adductor muscles being used for swimming rather than walking. This idea is supported by study of some modern salamanders that have a large adductor muscle blade that is placed further along the shaft of the femur than is the case with the terrestrial ones (Coates, 1996). The vertebral column of Acanthostega has been preserved almost completely, which includes the most anterior elements that form the neck and those in the region of the attachment of the pelvis (Coates, 1996). Many of the ribs from the neck and the vertebral column just behind the pelvis have also been preserved. In one respect a striking thing about the vertebral column is the small amount of variation along its length, which is similar to the situation in fish such as Eusthenopteron, or an aquatic tetrapod from the Carboniferous such as Greererpeton. In Acanthostega the neural arch processes, the zygapophyses were not well developed, though there were a few cases in which there were additional articulations between them instead. Essentially the centra were husks that surrounded the notochord, being comprised of 2 halves that didn’t fuse in the midline, with the exception of the atlas and the sacral intracentra. For an animal in which the head and body were usually supported by water, and not the limbs and girdles, that have specialised muscles to transmit the forces to the vertebral column,  this lack of differentiation and lack of zygapophyses this is consistent.

Most of the trunk ribs of Acanthostega are missing or not articulated, though the cervical ribs are well preserved and are still articulated when found. The neck muscles and those supporting the shoulder were attached to broadened, scooped out ends of cervical ribs that were about the same length as the trunk ribs, being short and straight with their heads being only weakly differentiated into upper and lower processes for articulation to the vertebrae. They differed from the heavy, flanged, overlapping ribs seen in Ichthyostega, and these, together with the vertebral column that is highly specialised, are probably the clearest evidence of the difference that existed between the 2 early amphibians with regard to their bodies and lifestyles.

Acanthostega had a substantial tail fin that has been compared to that of a lungfish, with fin rays that were longer and more numerous, reaching further around under the tail. Ichthyostega had a much less substantial tail fin that was relatively modest, and with fin rays that were shorter and less numerous, and not reaching around underneath the tail as in Acanthostega. The tail fin of Acanthostega was that of an ambush predator, with the fin rays making it capable of producing and controlling fine rippling movements enabling it to remain stationary in the water until prey came within reach when it could dart forward rapidly and powerfully with rapid acceleration to catch the prey. This tail would not be suitable for an animal that regularly ventured onto the land. Clack¹ says more details of the postcranial Skelton are available in Coates (1996).

There are many features of the postcranial skeleton of Acanthostega that clearly mark it as an aquatic animal that rarely if ever ventured onto dry land, and it has been estimated that its legs would have been unlikely to be capable of supporting its weight if it did. Clack¹ suggests the more important question is whether its aquatic adaptations were of primary or secondary type, are they relicts of its fish ancestry or derived from an ancestral terrestrial animal that returned to the water (Clack & Coates, 1995; Coates & Clack, 1995).

There is another key region, in addition to the hydrobranchial and postcranial skeleton which provides insight into the events of the transition from fish to tetrapod, the braincase and associated structures. The Acanthostega braincase is directly comparable with that of fishes such as Eusthenopteron and Panderichthys, as well as those of tetrapods from the Carboniferous. There are several similarities to the case of Eusthenopteron that are seen in the otoccipital region, especially in the dorsal and posterior parts. The side wall of the braincase is, however, penetrated by a large hole, the fenestra vestibuli, which partially incorporates the vestibular fontanelle of a fish such as Eusthenopteron. A crucial feature of Acanthostega is that it lacks the lateral commissure of the fish, which is part of the side wall. In Acanthostega, rather than having a hyomandibula that pivots against the commissure, the stapes, the equivalent bone, fits into the hole. In Acanthostega the stapes is relatively much smaller than the hyomandibula of Eusthenopteron, though it is robust when compared to the stapes of later tetrapods, with the fenestra vestibuli being almost filled by a large foot plate, and it was a flattened platelike element more distally.

Acanthostega is implied to be a primitive aquatic form by several features of its anatomy which indicate it had no ancestral forms more terrestrial than itself. The forearm bones of Acanthostega are such a feature, which is the most similar to those of Eusthenopteron than those of any other known vertebrate. Only 1 tetrapod group from the Mesozoic that were secondarily specialised aquatic tetrapods, some pachypleurosaurs that were related to plesiosaurs, have a radius that is substantially longer than the ulna, but it is the rule in tetrapodomorphs. Therefore it is suggested by this that its condition in Acanthostega is much more likely to be primary (primitive) than secondary.

The presence of a large, deep tail webbed with long lepidotrichia is another of these features. Clack¹ says it seems unlikely that once the dermal fin skeleton had been lost it would be re-evolved, though it is equally likely that such a fin would be lost in any form that was more terrestrially adapted. It can also be argued that a finned tail would be subjected to erosion or tearing by contact with the ground and to drying out as a result and would be likely to become susceptible to infection.

According to Clack¹ to imagine a terrestrial animal retaining a tail fin like that of Acanthostega is at least as difficult as to imagine one that was primitively aquatic evolving a large pelvis and femur. A conspicuous feature of terrestrial tetrapods is the possession of a pelvis that is in contact with the vertebral column, and it has been argued that as Acanthostega had these it therefore must have evolved from a terrestrial ancestral form. In the Devonian an early tetrapod would presumably produce young that lived in water, for at least the early part of their lives, regardless of the status of the adults. An argument can be made that even fully terrestrial adults might have retained fin rays until they metamorphosed; therefore it is not impossible to imagine a paedomorphic juvenile of such an adult that was fully terrestrial that evolved into an animal like Acanthostega that had a finned tail. Clack discusses this in more detail in Chapter 6 of her book¹.

The growth of the humerus of Ichthyostega may display something like this. A relatively long period of growth as an aquatic form might be indicated by the more primitive fishlike nature of the humerus in subadults that develop the terrestrial capabilities later. Acanthostega, in contrast to this, shares some characters of the humerus with tetrapods that came later and were terrestrially adapted, which were never developed by Ichthyostega.

The hind part of the skull of Acanthostega where it has lost the operculogular bone series is another region of its anatomy that might indicate the existence of terrestrial ancestors. It has been argued that these bones would still be present and functional in the gill movements associated with ventilation if Acanthostega was as aquatic as a fish. According to this view the skeleton of the gill and associated shoulder girdle characters present in Acanthostega would be primitive, though relict, being no longer functional.

A similar situation can now be seen in Tiktaalik, as far as the loss of the operculogular series in Acanthostega is concerned. As its gill skeleton is now known in some detail it has been found to differ very little from that in more basal tetrapodomorphs (Downs et al., 2008), though the opercular bones appear to have been lost completely. The situation in lungfishes is also comparable, as they have lost the series while retaining the use of the gills, and complementing gill-breathing with lungs they use for breathing air, a situation that has been envisaged for Acanthostega. Clack¹ suggests that as in lungfishes Acanthostega and other early tetrapod probably gulped air, raising the snout above the surface of the water and swallowing air. A variety of modern fishes have lost the operculogulars, some of which, though not all, breathe air, while others, such as the moray eel, do not.

A number of imaginative scenarios have accommodated the evolution of a pelvis and femur. One such suggestion is a predator concealed in water weeds that uses its hind paddle to force it through the plant growth, in fish such as the Sargassum frog fish, or one of the gurnards, in which the pelvic girdle is attached to the shoulder girdle; increased power and direction changes that are more accurate may be needed, and these could be produced by enlarged hind limbs. The muscles that controlled the tail in swimming had their site of origin on the pelvis; therefore elaboration of the girdle may have occurred to provide increased use of the tail for rapid bursts of speed required by an ambush predator to capture prey.

A primitive tetrapod such as Acanthostega provides clues useful in solving the problem of how limbs with digits may have evolved for an original use that did not include walking on land. It suggests the limbs and digits evolved at a time when tetrapods were still mostly aquatic, so their main function was one that used them in the water. A number of such uses have been suggested based on the presence in some modern fishes that have fins that are digitlike. These suggestions include clinging to weed to help maintain position in vegetation-rich water, and slow walking on the seafloor, as by the Tasmanian hand fish, the Spotted Hand Fish, Brachioninichthys hirsutus. Other proposed uses include moving through weed-choked swamps or feeling among the plants stems and debris. One of the possibilities that have been suggested arose from the comparison between spawning behaviour of fishes and amphibians and involved amplexus among the amphibians, the males legs are wrapped around the body of the female to ensure fertilisation is completed.

The anatomy of Acanthostega can provide much information concerning the order in which events occurred over the course of the transition. An example is limbs with digits that apparently evolved prior to walking. The wrists of Acanthostega differ from those of later tetrapods in the lengths of the radius and the ulna which were different from each other with the result that the distal ends could not have formed an effective bearing surface on which to balance the weight of the animal. Clack¹ says this can be said with some confidence, though the actual wrist bones were not ossified. Acanthostega had digits, though they were somewhat unusual ones, therefore the digits must have evolved before the wrists. Recent developmental genetic studies have provided some evidence from the fins of lungfish (Johanson et al., 2007b). The angle joint was also not suitable as a weight-bearing joint for similar reasons, as it is rather inflexible, so as with the digits on the forelimbs, those on the hind limbs also evolved before the ankle joint.

A number of the characters that are supposed to be those of tetrapods that are found in Acanthostega, with the exception of limbs with digits, and are often linked with terrestriality, can actually be present among modern fishes that are not known to have ever walked or lived on land. In the Amazon Basin the Arapaima that breathes air is one such fish, which has a substantial endoskeletal coracoid portion to its shoulder girdle, an enlarged pelvis, though it is not attached to the vertebral column, and in the vertebral column that is well ossified there are zygapophyses and large ventral ribs, and its tail has a similar shape to that of Acanthostega.

See Acanthostega, a Stem Tetrapod – Life History Revealed by Synchrotron microtomography

Sources & Further reading

1. Clack, JA, (2012). "Gaining Ground: The origin and evolution of tetrapods", Indiana University Press