Sunday, December 2, 2012

Humble beginnings for the mighty diapsids: the Araeoscelids and Orovenator

The amniotes (those initially four-legged creatures that produce "amniotic eggs", i.e. eggs adapted for land life) are traditionally divided into a few branches depending on some key characteristics of their skull, and the diapsids appear to be the most successful of all these branches. Virtually all living vertebrates that we commonly name "reptiles" are  diapsids: crocs, lizards, snakes, the whole lot of them... Diapsids also include the birds by way of their forebears, the dinosaurs. In contrast, mammals and their ancestors belong to the synapsid branch of the amniotes. The distinction between diapsid and synapsid lays in the number of holes (scientist called those "fenestrae" which means windows) in the skull just behind the eye socket. Diapsids typically have two, the supratemporal (or upper temporal) fenestra on top and the infratemporal (or lower temporal) fenestra below. Synapsids only have one,  the bottom infratemporal fenestra, simply called temporal fenestra. Both diapsids and synapsids evolved from more primitive reptiles with no opening behind the eye: the anapsids. The distinction between the different type of skulls are depicted in Figure 1. 

Fig 1.- Different type of skulls among reptiles. Top left: anapsid skull of Procolophon trigonoceps (after Romer, 1956) top right: diapsid skull of Petrolacosaurus kansensis (after Reisz, 1981), bottom: synapsid skull of Eothyris parkeri (after Reisz et al., 2009).

The distinction seems simple enough but only reflects the primitive initial condition. Evolution indeed loves playing tricks. The same way that we know that dolphins are really mammals and not some strange air-breathing fish, scientists figured that everybody's favorite marine mesozoic  reptiles, the plesiosaurs and ichthyosaurs, are really diapsids in disguise. Their skulls only have one opening behind the eye socket, the supratemporal fenestra, a condition which is called 'euryapsid'. The euryapsids used to be considered as a fourth branch of reptiles, but carefull examinations of the fossils show they evolved from diapsid ancestors and that the loss of the infratemporal fenestra is only a secondary characteristics. More tricky are the turtles. These do not have any opening behind the eye which classified them as "anapsids". However, nowadays, it seems quite firmly established that they too are highly modified diapsid reptiles, although it took scientists quite a while to figure this one out. Now look a the highly modified skulls of birds and snakes: you will have hard time recognizing any of the original temporal openings in them, but since their ancestors were diapsid reptiles, by way of basic phylogeny law (you shall belong to the same clade as your ancestors, ... i.e the monophyly principle), they too are diapsid reptiles. 

That being said, the diapsids then really are the most successful group of amniotes with some 18,000 species, including birds, alive today (compared to the 5700 species of mammals representing the only survivors of the synapsid branch). But what is their origin? They probably evolved during the Late Pennsylvanian from a group of anapsid ancestors to which such lizard-like creatures as Paleothyris and Hylonomus belong to.  But this is not at all very clear because of the scarcity of the fossil record. The first true diapsids are a family of small superficially lizard-like creatures called Araeoscelidia, to which Petrolacosaurus and Araeoscelis are the best known representatives. The fossil record of the Araeoscelids extends from the Late Carboniferous to the Early Permian. 

Fig 2.- a reconstruction of Petrolacosaurus kansensis.

The most ancient known diapsid is Petrolacosaurus kansensis from the Late Pennsylvanian (the North American 'Missourian' stage which corresponds to the ICS Kasimovian stage, ~305 MYA) of Kansas. The generic name means "rock lake reptile" in reference to the "Rock Lake Shale" in which the type specimen (a nearly complete hind limb) was found. Although originally described in 1945 as a pelycosaur (thus a synapsid), it was not until 1977 after new specimens including the skull with its two characteristic openings, were thoroughly described, that it was realized to be the earliest known diapsid, raising the little critter from relative obscurity to paleontological stardom. In a world dominated by giant arthropods and fearful amphibians,  Petrolacosaurus was indeed relatively small, measuring probably about 70 to 80 cm in length, accounting for the long tail.

Fig 3.- Reconstruction of Spinoaequalis schultzei.

Also from the Late Pennsylvanian of Kansas but a bit later (from the Calhouns Shale formation dated to the North American 'Virgilian' stage roughly corresponding to the ICS Gzhelian, ~ 300 MYA), comes Spinoaequalis schultzei. This early diapsid is a bit smaller (30 cm) and is only tentatively placed among the araeoascelids. The interesting note about this critter is the tail:  the tall and equal size neural and haemal spines of the caudal vertebra (thus the generic name) which gives the distinct tall and laterally compressed shape to the tail is viewed as an adaptation for swimming, a good indication that Spinoaequalis was an aquatic animal. This is supported by the fact that its fossil was discovered in freshwater deposits among remains of spiny sharks (acanthodians) and other fully aquatic animals. However, the long and slender limbs, are those of a terrestrial animal, an indication that it wasn't fully aquatic. In any case,  this makes Spinoaequalis the first amniote to have ever returned  to water since their epic conquest of the dry lands.

Fig 4.- Skulls of Araeoscelis gracilis (after Reisz et al., 1984).
Araeoscelis dates from the Early Permian and was originally described in 1910 as a lizard. It has the same body plan than Petrolacosaurus but the skull was more massive with strong teeth, ideal for crushing the heavy exoskeleton protecting some of the arthropods of that time. Araeoscelis was about the same size too, with an estimated length of 70-80 cm accounting for the unknown tail. As a probable adaptation of its specialized diet, the lower temporal fenestra has closed making the skull more robust. Araeoscelis, has therefore this 'euryapsid' condition that will be common to the large marine predators of the Mesozoic.  Two species have been  described, A. gracilis from the Arroyo Formation of Texas (Kungurian age, ~275 MYA), known from several fairly complete specimens, and A. casei from the Admiral Formation of Texas (Artinskian age, ~285 MYA), known from at least seven individuals. The two are virtually indistinguishable and the separation into distinct species seems only to be justified by their difference in age, A. casei being from slightly (~ 10 millions years) older rocks than A. gracilis

Fig 5.- Reconstruction of Araeoscelis gracilis.

The other Araeoscelids are poorly known and all date from the Early Permian. The 70 cm long Dictybolos tener from the Wellington formation of Oklahoma (~280 MYA) is known from isolated bones from numerous individuals. This one was presumably semi-aquatic and a fish eater. Zarcasaurus tanyderus from the Cutler Formation of New Mexico is known from a partial disarticulated skeleton. Characterized by its rather long neck vertebrae, it was a close relative of Araeoscelis. In Europe, the rather dubious Aphelosaurus lutevensis from the Tuilières formation of South Central France and first described in the 19th century,  is another possible Araeoscelid, though it is hard to tell without any knowledge of a crucial piece of fossil information: the skull. Similarly, Kadaliosaurus priscus from the Rotliegend of Germany is only known from a postcranial skeleton and its classification among the araeoscelids is only tentative.

Fig 6.- Reconstructed skull of Orovenator mayorum (after Reisz et al., 2011).

Besides the Araeoscelids, there is one additional diapsid dating from the Early Permian, Orovenator mayorum discovered in one of the fissure fills of the Richards Spur locality of Oklahoma, and known from two partial crushed skulls. Orovenator has the distinction of being the earliest known and most primitive of the Neodiapsids, a clade that contains all the known diapsids except for the araeaoscelids. This was a small animal with an elongated skull half the length of those of Petrolacosaurus and Araeoscelis. The Richards Spur locality with its distinct Early Permian fauna of some 30 taxa of fully terrestrial vertebrates, is thought to originate from an upland ecosystem, which would only fossilized in very exceptional cases. This is in sharp contrast to the Araeoscelids which were all found in lowland swampy habitats, with better chances for fossilization. The hypothesis is therefore that the initial split of the early diapsids into the Araeoscelids and the Neodiapsids is the result of adaptation to two different habitats, with the Araeoscelids in the lowlands and the Neodiapsids in the uplands.

In conclusion, the early diapsids show a surprising degree of diversity with some forms that became at least partially aquatic (Spinoaequalis, Dictybolos) while others adapted to the harsher conditions of the uplands (Orovenator).  There is also evidence of a quite specialized diet for some (Araeoscelis). However, the remains of these animals are quite rare and there is a rather long gap in the fossil record before we see them appear again in the Late Permian. Their number would not significantly increase before the Early Triassic. It is somewhat tempting to imagine that like the mammals of the Mesozoic dominated by the dinosaurs, the early diapsids lived in the relative shadow of the other reptiles for some 50 million years,  when the world was dominated by  larger synapsids, anapsids and  amphibians and that it will take the massive Permian-Triassic extinction event to see the diapsids finally take the upper hand.


Brinkman, D., Berman, D., & Eberth, D. (1984). A new araeoscelid reptile, Zarcasaurus tanyderus from the Cutler Formation (Lower Permian) of north-central New Mexico. New Mexico Geology, 34–39. 

Debraga, M., & Reisz, R. (1995). A new diapsid reptile from the uppermost carboniferous (Stephanian) of Kansas. Palaeontology, 38(1), 199–212. 

Lane, H. (1945). New mid-Pennsylvanian reptiles from Kansas. Transactions of the Kansas Academy of Science (1903-), 47(3), 381–390. 

Olson, E. (1970). New and little known genera and species of vertebrates from the Lower Permian of Oklahoma. Fieldiana: Geology, 18(3), 359–434. 

Reisz, R. R. (1977). Petrolacosaurus, the oldest known diapsid reptile. Science, 196(4294), 1091–3. 

Reisz, R., Berman, D., & Scott, D. (1984). The anatomy and relationships of the Lower Permian reptile Araeoscelis. Journal of Vertebrate Paleontology, 4(1), 57–61. 

Reisz, R. R., Modesto, S. P., & Scott, D. M. (2011). A new Early Permian reptile and its significance in early diapsid evolution. Proceedings. Biological sciences / The Royal Society, 278(1725), 3731–7. 

Williston, S. (1910). New Permian reptiles: rhachitomous vertebrae. The Journal of Geology, 18(7), 585–600. 

Williston, S. (1913). The skulls of Araeoscelis and Casea, Permian reptiles. The Journal of Geology, 21(8), 743–747.

Original artworks on Paleoexhibit are copyrighted to Nobu Tamura. Do not use without permission (Email: nobu dot tamura at yahoo dot com)