It's time once again to look at squamates…
Caption: I've mentioned before how inspirational the wonderful illustrations of Alan Male (from Philip Whitfield’s Reptiles and Amphibians: An Authoritative and Illustrated Guide of 1983) were to a young me, and here are the several cordylids he illustrated. It's sometimes obvious that Male only had access to photos, and photos that weren't especially useful at that, and hence made mistakes (the head of the Chamaesaura here is inaccurately short, for example). Images: Alan Male, from Whitfield (1983).
That’s snakes and lizards… though, given that snakes are a specialized group of lizards, maybe I should stop using ‘snakes and lizards’ as a descriptor. In fact, maybe we should get into the habit of calling all squamates ‘lizards’ and hope eventually that tree-based thinking wins the day. Ha, yeah right.
Here, we’re going to embark on a whistlestop tour of the cordylids, or girdled lizards: a diurnal, endemic African group, mostly associated with dry rocky places and hilly slopes, and well known for the armour-plated, sometimes spiky, covering of many of its species. The use of burrows and crevices is common across the group, and the majority are viviparous, giving birth to between one and twelve babies. These animals are generally between about 12 and 40 cm long in total.
Caption: some of the books I consulted while compiling this article. Even in the age of the internet, books are essential and you still have to do everything you can to accrue them. Remind me why I have no money. Image: Darren Naish.
Within squamate phylogeny as a whole, cordylids appear part of Scincoidea, the large clade that also includes the hyperdiverse skinks, night lizards (Xantusiidae) and plated lizards (Gerrhosauridae). To use the most general generalization possible, scincoids are mostly smooth-scaled, insectivorous, ancestrally terrestrial lizards that often have small bony plates (osteoderms) embedded in the skin and use their well-clawed limbs to clamber, run and dig. Within Scincoidea, both anatomical similarities and molecular data show that cordylids and gerrhosaurids are close, the two forming the clade Cordyliformes (Lang 1991). An old-fashioned view is that gerrhosaurids should be subsumed into Cordylidae. That’s a bad idea (come on, they’re easily ‘distinct enough’ to warrant ‘family’ status) and it’s not currently in fashion.
How, then, to discuss this moderately (but not especially) speciose group, about 70 extant species of which are currently recognized? Our coverage of cordylids here will revolve around taxonomy, for reasons that’ll be obvious once we get into it. For girdled lizards, you see, were, until recently, grouped into four genera: the extremely flat-bodied flat lizards Platysaurus, the flattish crag lizards Pseudocordylus, the deeper-bodied, often spiny, girdled lizards, zonures or sungazers Cordylus and the serpentiform grass lizards, snake lizards or ground lizards Chamaesaura. Let’s get to it, starting with flat lizards.
Caption: the ‘original’ girdled lizard… the Cape girdled lizard Cordylus cordylus, type species of the genus Cordylus and thus the groups Cordylinae and Cordylidae. This individual was photographed in Stony Point Nature Reserve, Western Cape, South Africa. Image: Bernard Dupont, CC BY-SA 2.0 (original here).
Flat lizards are very special. As might be obvious, flat lizards are extremely flattened, both in the head and body and also the thighs and tail base. 16 species are currently recognized, the most recently named of which is Attenborough’s flat lizard P. attenboroughi Whiting et al., 2015 of southern Namibia (yup, species number 42 or something named after Sir David).
Bright colours and bold patterns are present across Platysaurus species, with blues, oranges, yellows and reds being present on the head, flanks, limbs and especially the undersides of the head and belly. The lizards flash these ventral markings by tilting or lifting themselves, and their ventral distribution of course means that they aren’t obvious at all times. These colours are typical of adult males; females across most species are generally blackish-brown with three white longitudinal stripes. However, the females in some species are brightly coloured too.
Caption: a montage that gives some idea of how variably coloured and patterned, and how spectacular, some flat lizards are. From left to right, we’re seeing a Common flat lizard P. intermedius in Zimbabwe, a Broadley’s flat lizard P. broadleyi in Augrabies National Park, South Africa, and an Imperial flat lizard P. imperator in Zimbabwe. These three species represent three of the four main clades within Platysaurus. Images: Sputniktilt, CC BY-SA 3.0 (original here); Bernard Dupont, CC BY-SA 2.0 (original here); © i_c_riddell, CC BY 4.0 (original here).
Flat lizards are mostly insectivorous, and climb, leap and chase in their efforts to catch flies, butterflies and other insects. They’ve even been seen to dive into pools and grab insects from below the water (Alexander & Marais 2007). Some amount of herbivory and frugivory is present too, some species eating flowers, leaves, seeds, berries and – in the Augrabies flat lizard P. broadleyi – figs (Pianka & Vitt 2003, Alexander & Marais 2007). The lizards mostly eat ripe figs on the ground but will also climb intro trees to get them, and will “use birds as cues to find fruiting fig trees” (Pianka & Vitt 2003, p. 229). P. broadleyi, incidentally, is another recently named species (it was only scientifically named in 1997, having previously been included within the Cape or Namaqua flat lizard P. capensis). Herbivory in lizards is generally associated with large size, but it appears that these lizards break that rule.
Caption: flat lizard be flat. This view of a Broadley’s flat lizard gives an excellent indication of what platysaurine proportions are like. Numerous individuals of this species are readily visible in Augrabies National Park in South Africa where they can be observed leaping to catch black flies. Image: Bernard Dupont, CC BY-SA 2.0 (original here).
We’ve seen already that cordylids are overwhelmingly viviparous. Platysaurus is oviparous, females producing two elongate, soft-shelled eggs. The remarkable flat bodies of these lizards have apparently resulted in the evolution of low clutch size and unusual egg shape (Pianka & Vitt 2003). Flat lizards are highly gregarious, tens of individuals typically feeding and hanging out in the same rocky areas, and sharing the same crevices after sundown. Their nesting is often communal too.
Wherefore art though, Platysaurus? The general view across much of the 20th century was that flat lizards are nestled somewhere in among the cordylid assemblage, in which case their lack of both osteoderms and spininess would be advanced traits. This hypothesized phylogenetic position (Lang 1991) also required that ovipary re-evolved in these lizards from viviparous ancestors… which is unlikely albeit not impossible (I’ve discussed this phenomenon a few times before, most recently in my 2020 article on the reproductive habits of Mesozoic mammals).
Caption: this is cordyliform phylogeny as hypothesized by Lang (1991). Note that the serpentine chamaesaurs are outside of Cordylinae, so that the primary split within Cordylidae is between Chamaesaurinae and Cordylinae. That was a 'traditional' view that had existed since the 1930s. Within Cordylinae, flat lizards are highly derived (this term has now been ruined, since some take it to mean 'very recently evolved within a group', and others to mean 'highly modified relative to the ancestral condition'... here, it can mean either or both). Molecular studies have shown these key propositions to be wrong. It should be added that Lang (1991) is a good example of a morphology-based phylogenetic study done well: Mathias Lang described and illustrated relevant character states, explained and described his methodology and approach, considered alternative hypotheses, and went to considerable trouble to obtain and observe relevant specimens. Image: Lang (1991).
The hypothesis that flat lizards – which, in some ways, seem ‘more typical’ than other cordylids – are the most modified (‘most derived’) cordylids always did seem a bit odd… couldn’t it be that things were the other way round? Molecular work published this century appears to confirm exactly this, since Platysaurus has more recently been recovered as the sister-group to the rest of Cordylidae. This, combined with anatomical distinctiveness, means that Cordylidae is now generally considered to consist of the two clades Platysaurinae and Cordylinae (Pyron et al. 2014).
Caption: a few cordylid anatomical characters of special interest, as illustrated by Lang (1991). At left, we're looking at the left side of the palatal surface of the skull of Cordylus cordylus, the unusual feature being the subdental shelf: the hollowed-out area, lingual to the teeth, and marked at its edge by the crista dentalis. The function of this shelf is unknown but it might have a role in holding food items in place while they are crushed. At right, a montage showing osteoderm distribution in (A) Cordylus, ventral (left) and dorsal views, (B) Pseudocordylus, ventral (left) and dorsal views, (C) Platysaurus, dorsal view and (D) Chamaesaura, dorsal view. All from Lang (1991).
Cordylus, the break-up. Within Cordylinae specifically, molecular work published in the 1990s and early years of the 21st century indicated that Pseudocordylus (the crag lizards) is not monophyletic, with different species being in different places among Cordylus (Frost et al. 2001). In addition, Cordylus seemed paraphyletic not just to Pseudocordylus, but to the snake-like Chamaesaura species too. One solution to this problem might be to sink Pseudocordylus and Chamaesaura into Cordylus (Frost et al. 2001, du Toit et al. 2004). A criticism of the relevant studies, however, is that they didn’t sample a sufficiently broad range of species to really get to grips with their phylogeny (and hence taxonomy).
Stanley et al. (2011) opted to sort things out via much more representative sampling. They found that several clades traditionally included within Cordylus are actually outside a clade that included Chamaesaura, Pseudocordylus and ‘core’ Cordylus (the type species is the Cape girdled lizard Co. cordylus (Linnaeus, 1758) of South Africa). A solution could have been to give up on the genus-level distinctions and just lump everything together, but Stanley et al. (2011) instead opted to devise a new taxonomy that reflected this phylogeny.
Caption: a fairly messy effort to depict cordylid phylogeny as recovered by Stanley et al. (2011), with photos depicting genus-level diversity within the group. It should be obvious immediately that Smaug is big and very spiky, that Ninurta is small and delicate, that Chamaesaura is serpentine, that Pseudocordylus lacks spines and big, plate-like scales, that Ouroborus curls up into a ring, that Karusasaurus and Namazonurus are broad-bodied, that Hemicordylus is a bit generic, and that Cordylus in the strict, modern sense is speciose. Images: Smaug: Eric Johnston, CC BY-SA 3.0 (original here); Ninurta: Amada44, public domain (original here); Chamaesaura: Alandmanson, CC BY-SA 4.0 (original here); Pseudocordylus: Bernard Dupont, CC BY-SA 2.0 (original here); Ouroborus: Handré Basson, CC BY-SA 3.0 (original here); Karusasaurus: Tony Rebelo, CC BY-SA 4.0 (original here); Namazonurus: © Tim Brammer, CC BY-NC 4.0 (original here); Hemicordylus: Abu Shawka, CC0 (original here); Cordylus: Bernard Dupont, CC BY SA-2.0 (original here).
Describing everything they did would take too long to explain, but the takehome is that those ‘Cordylus’ clades recovered outside of ‘core Cordylus’ required new names. So, hello Smaug, Ninurta, Ouroborus, Karusasaurus and Namazonurus, and hello again to a resurrected Hemicordylus (originally named in 1838). A ‘restricted’ version of Pseudocordylus (‘core’ Pseudocordylus = the clade that includes the type species P. montanus and its closest relatives) is now in use.
This new taxonomy makes it easiest to talk about cordyline diversity, since the members of the different clades mostly look obviously different and also do different things. It won’t be lost on you that some of these names are recognizable: Ouroborus was named for the mythical symbol whereby a dragon-like reptile is consuming its own tail, while Smaug was named for the dragon of Tolkien’s 1937 book The Hobbit. By the way, this new taxonomy corresponds quite closely to distribution and morphology (Stanley et al. 2011).
Caption: name an animal – any animal, I think – after a famous character from popular fiction, and you’re gonna win some coverage in the popular media. That’s exactly what happened when Smaug the cordylid was published in 2011. Nine Smaug species are currently recognized, the best known of which is the Giant girdled lizard S. giganteus, shown here. Images: Unwin Paperbacks; Wilfried Berns, CC BY 2.5 (original here).
Giant, terricolous Smaug, ring-forming Ouroborus. The Smaug species are extremely large (SVL can be 20 cm and total length 40 cm), robustly built, have an especially spiny tail, and are associated more with burrows and crevices in soil than rock (they’re terricolous, rather than rupicolous). The burrows of the Giant girdled lizard or Sungazer S. giganteus, excavated in fine soil, can be at least 1.8 m long and are sometimes home to a mother as well as her young (Branch 1988). They’re notably deep-headed relative to rupicolous cordylines.
Caption: a slightly sketchy effort to depict a Giant girdled lizard in life, and which depicts some of the remarkable features of this large lizard (it can reach 40 cm in total length). The tail spines are especially large, and also slightly curved, and the spines around the back of the head are especially prominent. Image: Darren Naish.
Ouroborus contains the somewhat Smaug-like Armadillo girdled lizard O. cataphractus, the spiny species famous for grabbing hold of its tail in its mouth and forming a spiky ring. While most cordylines are (apparently) solitary, this species is gregarious, with groups of up to 30 sharing the same refuge site and colonies consisting of hundreds of individuals (Alexander & Marais 2007). Loners exist though and tend to be adult males; these differ from females in reaching larger sizes and in having proportionally larger heads and longer tails (Mouton et al. 2006).
Caption: the Armadillo girdled lizard is another incredibly charismatic, handsome animal… its machine-perfect plates give it such an awesome look. And of course it’s famous for rolling into a defensive circle. This species has been over-collected for the pet trade in the past, though this is now illegal. Images: Kevin Murray, CC BY 4.0 (original here); Handré Basson, CC BY-SA 3.0 (original here).
The Pseudocordylus crag lizards are flat-headed, rock-dwelling animals with spiny tail whorls but a head, neck and body mostly covered in granular, plate-like or slightly convex scales. They mostly eat arthropods but small lizards and leaves, flowers and berries are eaten by some species too (Alexander & Marais 2007). They’re mostly associated with the Drakensburg and Cape Fold Mountains of South Africa, Swaziland and Lesotho, and some (like the Drakensburg crag lizard P. subviridis) live at sufficient altitude that they can sometimes be subjected to freezing temperatures. So long as they’re sufficiently deep in crevices and cavities, they avoid exposure and are ok.
Caption: another fine reminder of what spectacular beasts some of these cordylids are. This is a Drakensburg crag lizard Pseudocordylus subviridis of South Africa and Lesotho. An obvious feature of these lizards is that the head and tail possess osteoderms whereas the body and limbs do not. Populations included within this species are variable and there's some controversy over how their taxonomy should be resolved. Image: Amada44, CC BY 3.0 (original here).
Namazonurus, Karusasaurus, Ninurta. Moving now to more of the new genera named by Stanley et al. (2011), we come to Karusasaurus, a group of just two species that occur across South Africa and southern Namibia. The best known of them is the widely distributed Karoo girdled lizard K. polyzonus (the other is the Namibian or Jordan’s girdled lizard K. jordani). Karusasaurus species recall Smaug and Ouroborus in having spiky tail whorls but their bodies are notably broad and flattened and mostly covered in small, smooth scales arranged in transverse rows. Karoo girdled lizards are mostly mottled in brown and cream but melanistic populations occur on the cool Atlantic coast of South Africa (Broadley & Branch 2002, Mouton et al. 2002). Namazonurus is Karusasaurus-like, also relatively flat-bodied, but is evidently more closely related to Cordylus (Stanley et al. 2011, Pyron et al. 2014). Both genera (Namazonurus and Karusasaurus) share a transparent lower eyelid, a feature not present in Cordylus.
Caption: Karoo girdled lizard, exemplar of Karusasaurus. The generic name incorporates the Khoisan word karusa, meaning dry, barren, thirstland. Image: Tony Rebelo, CC BY-SA 4.0 (original here).
Ninurta contains only N. coeruleopunctatus of South African’s southern Cape region, a mid-sized (SVL 8 cm), lightly built cordylid with a flattish head. It grouped close to the Chamaesaura species in Stanley et al.’s (2011) phylogeny but a position close to Hemicordylus has been found elsewhere (Pyron et al. 2014).
Caption: a Blue-spotted girdled lizard, sole representative of the genus Ninurta. The general appearance of this lizard – it’s small, skinny, long-bodied and with small scales across its body that aren’t big and plate-like – is consistent with phylogenetic studies that find it close to the ancestry of the long-bodied Chamaesaura species. Image: Amada44, public domain (original here).
Finally, Cordylus itself – even with several lineages now removed and placed in their own genera – is still a large clade with a complex phylogenetic structure. This is the most widely distributed cordylid clade, occurring across a large part of southern and eastern Africa as far north as Ethiopia. There are dwarf species within this group (like the Rooiberg girdled lizard C. imkeae and its relatives; SVL can be 4 cm, total length 10-12 cm) and arboreal ones too (like the Tropical girdled lizard C. tropidosternum). These live in hollow trees and stumps and hide under loose bark and dead leaves.
Caption: a montage of species still included in Cordylus. At left, the strikingly attractive Black girdled lizard Co. niger, here photographed at Cape of Good Hope Reserve, Western Cape, South Africa. At right, Cape girdled lizard Co. cordylus – the ‘original’, or OG, cordylid – photographed in Mountain Zebra National Park, Eastern Cape, South Africa. Images: Bernard Dupont, CC BY SA-2.0 (originals here and here).
Snake-like cordylids! Most of us who know even a bit about squamates are aware that a snake-like form evolved multiple times within these animals. Dibamids, pygopodids, amphisbaenians, skinks (several times), anguids…. But poorly known even among specialists is that cordylids include snake-like species too, and here I’m taking about the Chamaesaura species, the five or so species of which are variously dubbed grass lizards, snake lizards or ground lizards. ‘Grass lizard’ is a not especially useful name, given that it's used around the world for various entirely different types of lizard.
These lizards are so weird compared to other cordylids – come on, the idea of snake-like cordylids is wild! – that they have, at times, been considered the sister-group to the remainder of the group (Lang 1991). They’re widespread across Africa’s southern half, occurring along the coastal strip of South Africa, throughout Angola, and (on the eastern side of the continent) from Zimbabwe and Mozambique to Rwanda, Tanzania and Kenya.
Caption: at left, Chamaesaura skulls illustrated by Lang (1991). (A) Ch. macrolepis in dorsal view, (B) Ch. anguina in right lateral view, and (C) the lower jaw of Ch. anguina in lingual (above) and labial views. These skulls are shallow at the snout end (as is typical for rupiculous cordylines) and more pointed (in dorsal view) than is the case in many other cordylids. At right, a range map for the Chamaesaura species as depicted by Lang (1991). If you’ve thought of these animals as mostly limited to the south of the continent, it’s obvious that they’ve actually staged a decent invasion of the continent’s southern half. Images: Lang (1991).
They’re mid-sized to long (with SVLs of between 14 and 17 cm, and total lengths of as much as 50 cm), extremely long-tailed (the tail might be three or four times the body length), and snake-like in movement. While they do have a ‘cordylid look’ to their integument (the body scales are sharply keeled), they differ from most other cordylids in that osteoderms are restricted to the head, a condition which has convergently arisen in Pseudocordylus. The head is also flattened, this being consistent with phylogenies that find Chamaesaura within a section of the tree where the surrounding lineages are rupicolous (Stanley et al. 2011).
The Chamaesaura species are further interesting in that they seem to represent a continuum, with one species (Ch. aenea) having five clawed digits on all four limbs, another (Ch. anguina) having only one or two clawed digits on its much reduced limbs, and a third (Ch. macrolepis) lacking forelimbs entirely, and having only a single clawed digit on its tiny, spike-like hindlimbs (Branch 1988). Patterns of variation like this are seen in some other reduced-limbed, serpentiform lizard groups (like certain skinks). Broadley & Branch (2002) noted that the Ukinga girdled lizard Co. ukingensis approaches the Chamaesaura species anatomically in having a reduced osteoderm compliment and in lacking the lateral groove that cordylines normally have along the side of the body. However, molecular work finds it to be part of Cordylus in the new, restrictive sense, and thus not phylogenetically close to Chamaesaura. Any similarities are therefore convergent.
Caption: the snake-like Chamaesaura species are often larger than I'd assumed previously (reaching 50 cm in total), and I think that that's conveyed in this photo. They aren't specialized climbers but, as shown here, are obviously capable of clambering in shrubs, bushes and grass clumps. This is the Cape grass lizard Ch. anguina, and specifically part of the type population Ch. a. anguina (the subspecies-level taxonomy of this species is controversial and some studies have raised the subspecies to species level). Image: Chris Vynbos, CC BY-SA 4.0 (original here).
As suggested by one of their common names (‘grass lizard’), the Chamaesaura species are strongly associated with grassy areas, mostly in hillside and plateau environments, and they don’t regularly use rocky retreats as their cousins do. They pursue grasshoppers, beetles and other insects in these habitats and ‘swim’ rapidly through grass via horizontal undulation. Alexander & Marais (2007) noted that their reliance on grass patches makes them vulnerable both to isolation (since grassland fragmentation and the construction of roads, walls and channels is a problem across the region), and to the loss or deterioration of these places by fire. Both issues could worsen across time, especially in a warmer, less climatically predictable world.
Caption: this fairly remarkable photo of a mother and baby Armadillo girdled lizard creates the impression that giving birth to a proportionally big, scale-covered baby might be a traumatic, perhaps painful event. The photo is (c) Gary Fogel and used to be present on a page dedicated to these lizards. It’s no longer findable online.
We’ve mentioned viviparity a few times now. The Chamaesaura species are no exception and can have among the largest litters – up to 12 babies – of any cordylid. Viviparity in squamates has often been linked with evolution in cool environments. We might not think of southern Africa as cool, but there are climatic models in which it was when cordylines initially diversified, this being the Oligocene (between about 35 and 25 million years ago) (Mouton & Van Wyk 1997, Stanley et al. 2011). Some studies have therefore proposed that cordylines “are a cool-adapted lineage and that viviparity developed in the family during cold conditions” (Stanley et al. 2011, p. 68).
And that’s where things must come to an end. This article is another rescued and revamped one from the archives, the original version appearing back in 2008 (a wayback machine version is here). More squamates are coming!
Caption: screengrabs of the original version of this article, from 2008. I managed to obtain and use some different images of Chamaesaura back then.
For previous Tet Zoo articles on squamates, see…
In Quest of Anguids, May 2006
Pompey and Steepo, the World-Record-Holding Champion Slow-Worms, May 2007
Cambodia: now with dibamids!, May 2011
The Tet Zoo Guide to Mastigures, August 2018
The Remarkable Basilisks, May 2023
Do Lizards Really Have ‘Mite Pockets’?, March 2024
Meeting Lake Zacapu’s Garter Snake, March 2024
Ray Hoser, Number 1 Taxonomic Vandal, May 2024
The Mysterious Dibamids, May 2024
The Rehabilitation of Günther’s Black Cameroonian Snake, May 2024
Ikaheka and Other ‘Palatine Draggers’, Cryptozoic Elapid Snakes of Melanesia, June 2024
Arboreal Alligator Lizards of Mesoamerica... and Beyond!, July 2024
Grayiids: Africa Has Water Snakes Too, July 2024
Refs - -
Alexander, G. & Marais, J. 2007. A Guide to the Reptiles of Southern Africa. Struik Publishers, Cape Town.
Branch, B. 1988. Field Guide to the Snakes and Other Reptiles of Southern Africa. New Holland, London.
Broadley, D. G. & Branch, W. R. 2002. A review of the small east African Cordylus (Sauria: Cordylidae), with the description of a new species. African Journal of Herpetology 51, 9-34.
du Toit, A., Mouton, P. le F. N. & Flemming, A. F. 2004. Aseasonal reproduction and high fecundity in the Cape grass lizard, Cordylus anguinus, in a fire-prone habitat. Amphiphia-Reptilia 24, 471-482.
Frost, D., Janies, D., Mouton, P. le F. N. & Titus, T. 2001. A molecular perspective on the phylogeny of the girdled lizards (Cordylidae, Squamata). American Museum Novitates 3310, 1-10.
Lang, M. 1991. Generic relationships within Cordyliformes (Reptilia: Squamata). Bulletin de l'Institut Royal des Sciences Naturelles de Belgique, Biologie 61, 121-188.
Mouton, P. le F. N., Flemming, A. F.& Kanga, E. M. 2006. Grouping behaviour, tail-biting behaviour and sexual dimorphism in the armadillo lizard (Cordylus cataphractus) from South Africa. Journal of Zoology 249, 1-10.
Mouton, P. le F. N., Nieuwoudt, C. J., Badenhorst, N. C. & Flemming, A. F. 2002. Cordylus polyzonus (Sauria: Cordylidae) populations in the Western Cape, South Africa: relics or ecotypes? Journal of Herpetology 36, 526-531.
Mouton, P. LeF. N. & Van Wyk, J. H. 1997. Adaptive radiation in cordyliform lizards: an overview. African Journal of Herpetology 46, 78-88.
Pianka, E. R. & Vitt, L. J. 2003. Lizards: Windows to the Evolution of Diversity. University of California Press, Berkeley.
Stanley, E. L., Bauer, A. M., Jackman, T. R., Branch, W. R. & Mouton, P. LeF. N. 2011. Between a rock and a hard polytomy: rapid radiation in the rupicolous girdled lizards (Squamata: Cordylidae). Molecular Phylogenetics and Evolution 58, 53-70.
Whitfield, P. 1983. Reptiles and Amphibians: An Authoritative and Illustrated Guide. Longman Group Ltd, Harlow, UK.
