Return to World Chelonian Trust Main Page for more Turtle and Tortoise Information
LONG NECKS, FLAT HEADS AND THE EVOLUTION OF PISCIVORY .
- Scott Thomson
Applied Ecology Research Group and CRC for Freshwater Ecology University of Canberra, ACT, 2601, Australia.
Copyright © 2003 - Scott Thomson, All rights reserved
Questions and comments related to this article should be directed to Scott Thomson
Eastern Long Neck Turtle Care - Chelodina longicollis - Scott Thomson
The long neck chelid species Chelodina colliei is examined to determine if there is morphological evidence to support its placement within the genus Chelodina. Chelodina colliei was found to share morphological characters with the Chelodina longicollis group and those characters that were similar to Macrochelodina were found to not be homologous. Macrochelodina group was found to have a series of synapomorphies that separates this group from the Chelodina and Chelodina colliei. The most significant synapomorphy was the development of retrahens capitus collique muscles for a more powerful strike. Those characters that identify Chelodina colliei are were found to be autapomorphic.
Keywords: strike and gape, Chelidae, Chelodina, Macrochelodina long-neck turtle.
Securing fish is problematic for terrestrial and semi-terrestrial vertebrates. Fish are optimised for locomotion through water whereas the body forms of many predators are a compromise between the demands of terrestrial or aerial locomotion and the need to secure aquatic prey. Some predators use stealth, or stealth combined with mimicry, such as the alligator snapping turtle (Macroclemys temmincki) with a lure on the tongue to attract fish (Pritchard, 1989). Some use an aerial attack at high speed such as many sea birds, develop high speed and maneuverability in water such as penguins and seals, or simply trap fish in vulnerable locations such as rapids in rivers, eg. Bears. Others use a strike and gape mode of feeding where by the head and mouth attain sufficient speed, against an essentially inert body, to capture fish in there own medium. Among the birds, darters and pelicans have evolved down this line. Many species of freshwater turtles have also evolved strike and gape feeding such as the emydids Dierochelys and Emydoidea, the trionychids Chitra and Pelochelys and the chelids Chelus, Hydromedusa and Chelodina. A common feature of this mode of feeding is a long neck and streamlining of the skull, with elongation in the plane of strike.
Although the obvious feature is the elongation of the neck, a sweep of characters is required for a species to develop the strike and gape mode of feeding. First is a large body relative to the head and neck to overcome the action/reaction physical constraints of objects moving in water. The body must have high inertia so that when the head and neck strikes the body does not react by moving backwards. In this area turtles are at an advantage as the shell gives them a naturally high body to head weight ratio. Another requirement is the musculature to strike and in many species this means enlargement and elongation of the longissimus dorsi muscles. In turtles these muscles run down the thoracic vertebrae between the rib heads and the shell (Bojanus, 1817), hence a secondary feature is the enlargement of this area by expansion of the neurals, enlargement of the plastral buttresses or both (Thomson and Georges, 1996). Another feature is the small, elongate and flattened skull with eyes positioned forward common to many strike feeders. This is necessary so that at least the head and neck offers minimal resistance to high speed travel through water. This feature makes strike mode piscivores recognisable at a glance, they have a flat head with binocular vision, most visible among turtle species such as Chitra indica and Macrochelodina expansa.
The family Chelidae (Testudines: Pleurodira) contains three genera which are classified as long-necked piscivores, Hydromedusa and Chelus from South America as well as Chelodina and Macrochelodina from Australasia (Gaffney, 1977; Pritchard, 1984). It was determined using the cladistic method that these were related to each other, [Chelus [Hydromedusa: Chelodina – including Macrochelodina]], and that they formed the clade Chelod which was the sister of all short-necked turtles excluding Pseudemydura (Gaffney, 1977). Pritchard (1984) who used a functional approach to demonstrate that the three genera were independently derived disputed this. Characters that supported this were the different ways in which the head and neck were accommodated by the shell, ie scute differences, and the different directions of the strike (Pritchard, 1984). In fact it could be argued, on the basis of neck skeletal structure, that Chelus is not a long-necked turtle at all as it has the same atlas axis structure as Phrynops (Williams, 1950). The independent derivations were confirmed when DNA sequencing was utilized to develop turtle phylogenies which found that the clade Chelod was polyphyletic with respect to the short necks (Seddon et al., 1997; Shaffer et al., 1997; Georges et al., 1999). It was also found that Chelus was the sister to the clade containing Phrynops and Acanthochelys (Georges et al., 1999).
It has been demonstrated that the genus Chelodina could be divided into two functionally different groups called the Chelodina longicollis group (true Chelodina) and the Chelodina expansa group (Macrochelodina) (Burbidge et al., 1974; Legler, 1981; Rhodin, 1994). Chelodina group contains the species Chelodina longicollis, Chelodina steindachneri, Chelodina novaeguineae, Chelodina reimani, Chelodina mccordi, Chelodina canni and Chelodina pritchardi whereas Macrochelodina contains the species Macrochelodina expansa, Macrochelodina rugosa, Macrochelodina burrungandjii and Macrochelodina parkeri. That there are two functional groups, with differing feeding behaviour, in what is currently considered a single genus makes this an interesting case study for the evolution of the strike and gape mode of feeding.
One of the major difficulties in the relationships of these two groups has been the placement of the Western Australian species Chelodina colliei. When the functional groups were first proposed Chelodina colliei (formerly Chelodina oblonga sensu Thomson, 2000) was considered to belong to Macrochelodina (then the Chelodina expansa group). This was done on the grounds that it had an extremely long neck and flat head (Goode, 1967) a proposal followed by many subsequent authors (eg Legler, 1981; Cann, 1978; Rhodin, 1994). The first attempt at a non-morphological approach found that Chelodina colliei did not fit into either group and was considered to be a third lineage and sister to the other two groups (Burbidge et al., 1974). It was not until the use of electrophoresis came about that evidence that Chelodina colliei was the sister of all other Chelodina longicollis group species (the true Chelodina) came to light (Georges and Adams, 1992). This was backed up by independent studies using DNA sequencing which also found this relationship (Seddon et al., 1997; Georges et al., 1999). It has become apparent that the traditional morphological approaches used by Goode (1967) and others since were not finding the phylogeny of this group. A clash between morphology and molecules.
It is important within this large genus to have all species placed within a group, not only for convenience, because as it has been proposed on numerous occasions that the Chelodina represent two genera (Legler, 1981), Chelodina for the C. longicollis group and Macrochelodina for the C. expansa group (Iverson et al. 2001), a taxonomy being followed in this paper. That there is disagreement on this requires that further analysis is done to determine this placement and the evolutionary implications of this are born out. In this study an analysis of the morphology of the Chelodina and Macrochelodina is presented with a discussion of the evolutionary implications for piscivory. It was determined that, based on morphology, the species Chelodina colliei is in fact a highly derived member of the Chelodina group which converged on the Macrochelodina condition.
It was found that the two functional groups differed in their morphology in regions of the skull, neck and shell. Macrochelodina is typified by the enlargement of the anterior bridge strut and buttresses to the extent that the strut continues from the juncture of the peripherals and the pleurals postero-medially to a point covering greater than half the distance to the vertebrae. The suture with the plastron is on a raised buttress that is formed from the first pleural. There is enlargement of the retrahens capitus collique muscles and subsequent scaring on the ventral surface of the carapace. This is most extreme in the Arnhem Land species Chelodina burrungandjii. Further to this the longissimus dorsi muscles are expanded, at least in the anterior, with expansion of the rib arches common, increasing to an extreme in Chelodina burrungandjii where the neural bone series is contiguous and exposed. There is a narrow plastra and high bridge arches with well developed buttresses. All members of Chelodina and Macrochelodina have a completely fused and elongated atlas/axis complex, hence, this unit can be ignored below this level. The rest of the vertebrae are large and elongated, robust with a large degree of musculature. The transverse processes are large, triangular and well developed. The postzygapophyses are fused into a single unit giving a semi-lunar appearance on all vertebrae. The total length of the vertebrae when added to head length is usually longer than the carapace length. Within the skull the jugal is wide, ie it approximately half the diameter of the eye. The region of the squamosal into which the digastricus maxillae insert is large and extends towards the skull. The region of the squamosal dorsal to the tympanic cavity is narrowed, to form a ridge, with no flattening or flaring.
Contrary to this the Chelodina has a small anterior bridge strut without a significant progression on to the carapace in most species. The suture is flush with the surface of pleural one. Chelodina novaeguineae has a variation on these conditions, approaching that of Macrochelodina. The retrahens capitus collique muscles are small and close to the rib heads as are the longissimus dorsi muscles. A contiguous series of neurals has never been observed in any member of the Chelodina, however, isolated neurals commonly appear in old individuals of Chelodina longicollis. This group tends to be flattened with low bridges, small buttresses and wide plastrons, the extreme of this condition observed in Chelodina steindachneri and Chelodina longicollis. The vertebrae are small and elongated, delicate in structure with a small amount of musculature. The transverse processes are large and triangular. The postzygapophyses are isolated on most vertebrae, with the exception of the eighth in larger individuals. The total length of the vertebrae when added to head length is usually shorter than the carapace length. The skull has a narrow jugal which is less than one quarter of the diameter of the eye. The region of the squamosal into which the digastricus maxillae insert is small and does not extend towards the skull. The region of the squamosal dorsal to the tympanic cavity is widened and flattened, does not form a ridge, with considerable flattening or flaring in the direction of the parietal.
In Chelodina colliei a small anterior bridge strut is present, with small buttresses, which does not significantly extend beyond the pleural/ peripheral juncture. The suture does not progress on to the pleurals at all, with the exception of large individuals. The retrahens capitus collique muscles are small and close to the vertebrae. The longissimus dorsi muscles are enlarged far beyond that of any other species of Chelid, except perhaps Chelus fimbriatus. A more or less complete series of neurals is present, and has been reported many times in the literature. The carapace is flattened and the plastron narrow with increased height in the bridges present, however, buttresses are small and do not add strength or structure to the shell. The vertebrae are large and elongated, thin but with a large degree of musculature. The transverse processes are large, triangular and well developed. The postzygapophyses are isolated on most vertebrae, with the exception of the eighth in larger individuals. The total length of the vertebrae when added to head length is by far longer than the carapace length. The skull has a narrow jugal which is less than one quarter of the diameter of the eye. The region of the squamosal into which the digastricus maxillae insert is small and does not extend towards the skull. The region of the squamosal dorsal to the tympanic cavity is widened and flattened, does not form a ridge, with considerable flattening or flaring in the direction of the parietal.
Chelodina colliei shared many characters with the Chelodina. In fact those characters which it did not share with this group and where clearly derived were all associated with the strike and gape mode of feeding and were subsequently shared with the Macrochelodina. These particular characters although superficially similar to the Macrochelodina often differed in their structure and where clearly homoplasic. For example, neck length was extreme in both Macrochelodina and in Chelodina colliei. However, the individual vertebrae of Chelodina colliei shared primitive states with the Chelodina longicollis group such as the structure of the postzygapophyses. A single derived character for the clade containing both the Chelodina longicollis group and Chelodina colliei was also found. Any phylogenetic hypothesis developed from these data would separate the Chelodina expansa group from the clade containing both the Chelodina longicollis group and Chelodina colliei, once the necessary steps were taken to account for the homoplasy (eg stepmatrices or separate the characters). A result in agreement with the hypotheses presented using allozymes by Georges and Adams (1992) and that using sequence data by Seddon et al. (1997).
With the addition of morphological data that supports the most recent hypothesis’ we believe that there is now clear evidence from many approaches that demonstrate the placement of Chelodina colliei in the clade containing the Chelodina. Wether the two clades represent two monophyletic but closely related genera is not the subject of this paper and I am following the nomenclature of Iverson et al., (2001), however, based on this study and that of several others I recommend that Chelodina colliei be placed with the Chelodina longicollis group.
Many of the morphological characters that were determined by Gaffney (1977) to be synapomorphies for the Chelod are in fact character sweeps all related to the strike and gape mode of feeding. When these characters are looked at in detail it can be seen that they are not in fact that similar at all. This is a clear demonstration of convergent evolution, all the species involved converging on high speed head movement.
If we examine these characters from a functional approach reasons for the convergence become apparent. If you want high speed in a given direction (in this case forwards) then development of the required musculature is inevitable. Expansion of the longissimus dorsi muscles, used for the forward movement of the head, is understandable. Second to this is the expansion of the retrahens capitus collique muscles at the side of the cervical and thoracic vertebrae. If both Chelodina colliei and the members of Macrochelodina are aiming for a high-speed strike why have both not secondarily utilised these muscles. Their strike methods must be different, as they are not utilising the same muscle groups for the strike. The retrahens capitus collique muscles are used to move the cervical vertebrae sideways and this can, in pleurodires, be used to add additional strength to the strike. If the posterior cervicals are bent sideways to the left (from above) and the retrahens capitus collique muscles on the right side are contracted then the neck will be moved from left infection to right inflection at great speed. Combine this with rapid contraction of the longissimus dorsi and the head will shoot forwards at great speed. In other words, Macrochelodina does not strike like a snake. Chelodina colliei on the other hand does. It only uses its greatly enlarged longissimus dorsi muscles for the strike and hence straightens the neck at high speed. Chelodina has no ability at a fast strike lacking any musculature that could be used for it. They are suck and gape feeders, swimming after their prey.
The utilisation of the retrahens capitus collique muscles is a synapomorphy for the Macrochelodina, this character not being present in any other turtle lineage with the possible exception of Chelus. Interestingly, the development of the strike and gape mode of feeding has evolved on four occasions within the Chelidae, far greater than in any other group of turtles. Pritchard (1984) demonstrated that this had occurred three times (Chelodina [including Macrochelodina], Hydromedusa and Chelus) and this study has demonstrated that it actually developed twice in the Chelodina. One possible explanation for this is the differences in the cervical structure between pleurodires and cryptodires. Cryptodires cannot elongate the eighth cervical due to height restrictions within the carapace and also can only utilise the longissimus dorsi muscles for the strike, as there is virtually no sideways inflection of the neck possible. It would appear that it is not possible for Cryptodires to be as efficient as piscivores.
Turning to the presence of buttressing and neural bones, both these characters are also heavily linked with the strike and gape mode of feeding. Rapid contraction of the retrahens capitus collique muscles will cause transverse stress on the carapace. This in turn would cause torsion of the carapace unless some sort of buttress was available. Hence, the size of the buttress and thickness of the shell under it is linked to the development of the retrahens capitus collique muscles. This can be demonstrated by looking at the different species within the Macrochelodina, where we find that the Arnhem Land species, with the largest retrahens capitus collique muscles, also has the highest degree of buttressing. Exposure of the neural bones is dependent upon the development of the longissimus dorsi muscles (Thomson and Georges, 1996). With enlargement of the longissimus dorsi the requirement of space between the rib heads and the thoracic vertebrae is increased. This requires the spreading of this entire area, hence, expansion of the neural bones. In Chelodina and Macrochelodina this is a secondarily derived state as they are one of the many Australian lineages to have lost exposed neurals in the past (Thomson and Georges, 1996).
One of the most striking similarities of strike and gape mode feeders is the development of elongated and flattened heads. Pritchard (1984) discussed this in some detail. Within the Chelodina clade (Macrochelodina + Chelodina) there are two basic skull patterns, flat heads and rounded heads. The Macrochelodina group has flat heads, as does Chelodina colliei. Whereas the Chelodina longicollis group has a rounded head, a feature that has also been used to place Chelodina colliei with the Macrochelodina. A flat head basically means that the skull has widened, elongated and flattened to some degree. How the skull accomplishes this may be different among lineages. Once again homoplasy for the flat head character can be demonstrated by the presence of the derived characters. There are a number of characters that are synapomorphies for the Macrochelodina where Chelodina colliei and the Chelodina longicollis group share the primitive state. In fact in those parts of the skull flattened Chelodina colliei possesses autapomorphies which place it uniquely from any other Chelodina clade member. Hydromedusa and Chelus also possess flattened skulls. Hydromedusa is elongated in a similar fashion to the Macrochelodina, whereas Chelus is shortened and widened. Pritchard (1984) discusses this and demonstrates, from a functional approach, the error of assuming homology. Yet again the characters which support the group Chelod of Gaffney (1977) are a part of the character sweep which is a natural requirement for a piscivore.
Fish and aquatic invertebrates are a high value food source for many species of animals. Be they bird, mammal or reptile, terrestrial animals must develop the necessary architecture to be a successful piscivore. Although many mammalian piscivores are undoubtedly successful, only the birds and reptiles have developed the strike and gape method of catching aquatic prey. Turtles, being semi to totally aquatic have advantages over birds in that most piscivorous birds must still be able to fly, hence, compromises must be minimised. Turtles have natural inertia, a heavy shell, and many lineages of turtles have turned to piscivory. Most piscivorous turtles use the strike and gape mode of feeding.
Since Gaffney (1977) published his phylogeny there have been repeated publications that refute the hypothesis. However, it is the quality of the dataset that should be questioned not the methods of comparative morphology. Gaffney (1977) was missing many key taxa, such as Rheodytes and Elusor (yet to be described), Elseya dentata and any member of the Chelodina. Considering the limitations of his raw data, it’s not surprising that the phylogenetic hypothesis was wrong. Using this paper as an example of the superior nature of molecular data is a straw man argument. They have assumed that the quality of the data is good without checking it against a good data set.
Ashley, L.M. 1962. The laboratory anatomy of the turtle. W.M.Chelodina Brown Co., Dubuque. 48pp.
Bojanus, L.H. 1819. Anatome Testudinis Europaeae. Vilnae. 178pp. 1970 SSAR Reprint.
Burbidge, A.A., Kirsch, J.A.W., & Main, A.R. 1974. Relationships within the Chelidae (Testudines: Pleurodira) of Australia and New Guinea. Copeia 1974:392-409.
Cann, J. 1978. Tortoises of Australia. Angus and Robertson, Sydney. 94pp.
Gaffney, E.S. 1977. The side-necked turtle family Chelidae: a theory of relationships using shared derived characters. American Museum Novitates 2620:1-28.
Gaffney, E.S. 1979. Comparative cranial morphology of recent and fossil turtles. Bulletin of the Museum of Natural History. 164(2):65-376.
Georges, A. & Adams, M. 1992. A phylogeny of Australian Chelid turtles based on Allozyme Electrophoresis. Australian Journal of Zoology 40:453-476.
Georges, A. & Adams, M. 1996. Electrophoretic delineation of species boundaries within the short-necked freshwater turtles of Australia (Testudines: Chelidae). Zoological Journal of the Linnaean Society. 118:241-260
Georges A., Birrell J., Saint K., McCord W.P. and Donnellan S. (1999). A phylogeny for side-necked turtles (Chelonia: Pleurodira) based on mitochondrial and nuclear gene sequence variation. Biological Journal of the Linnean Society, London 67, 213-246.
Goode, J. 1967. Freshwater tortoises of Australia and New Guinea. Lansdowne Press: Melbourne.
Iverson, J.B., Thomson, S.A. and Georges, A. (2001). Validity of taxonomic changes for turtles proposed by Wells and Wellington Journal of Herpetology. 35(3):361-368
Legler, J.M. 1981. The taxonomy, distribution, and ecology of Australian freshwater turtles (Testudines: Pleurodira: Chelidae). National Geographic Society Research Reports 13:391-404.
Pritchard, P.C.H. 1984. Piscivory in turtles, and evolution of the long-necked chelidae. in Ferguson, M.W. (ed) The structure, development and evolution of reptiles. Zoological Society of London, Symposium. 52:87-110.
Pritchard, P.C.H. 1989. The Alligator Snapping turtle (Macroclemys temmincki) Conservation and Biology. Milwaukee Public Museum. 104pp.
Rhodin, A.G.J. 1994a. Chelid turtles of the Australian archipelago: I. A new species of Chelodina from southeastern Papua New Guinea. Breviora 497:1-36
Seddon, J., Georges, A., Baverstock, P. and McCord, W. 1997. Phylogenetic relationships of chelid turtles (Pleurodira: Chelidae) based on mitochondrial 12S rRNA gene sequence variation. Molecular Phylogenetics and Evolution. 7:55-61.
Shaffer, H.B., Meylan, P. and McKnight, M. 1997. Tests of turtle phylogeny: Molecular, morphological, and palaeontological approaches. Systematic Biology 46:284-303.
Thomson S.A. (2000). On the identification of the holotype of Chelodina oblonga (Testudinata: Chelidae) with a discussion of the taxonomic implications. Chelonian Conservation and Biology 3:745-749.
Thomson, S. and Georges, A. 1996. Neural bones in Australian chelid turtles. Chelonian Conservation and Biology. 2(1):82-86.
Thomson, S., White, A. and Georges, A. 1997. A re-evaluation of Emydura lavarackorum: Identification of a living fossil. Memoirs of the Queensland Museum. 42(1):327-336.
Williams, E. 1950. Variation in the cervical articulation in recent and fossil turtles. Bulletin of the American Museum of Natural History.
Distribution of characters among the three groups. 0 = primitive state; 1 = derived. Polymorphisms shown.
|2||bridge strut suture||0||0||1|
|3||retrahens capitus collique||0||0||1|
|10||insertion of digastricus maxillae||0||0||1|
Return to World Chelonian Trust Home Page Return to Taxonomy and Morphometrics
World Chelonian Trust
PO Box 1445