The ribs of Belodon differ from those of true Crocodilia by stronger forking at the proximal end; this is due to the deeper position of the parapophysis. These details agree more with the Dinosauria. The abdominal ribs are also different from those of true Crocodilia.
Both the Parasuchia and Pseudosuchia possess a clavicle and interclavicle, whereas these are generally unknown in the Theropoda and Dinosauria (1). True Crocodilia lack the former, but the interclavicle is all the more strongly developed. Clavicles are known from Mystriosuchus and Aetosaurus. In Erpetosuchus a true acromion is developed on the scapula for articulation with the clavicle. In interclavicle in the same species is distally drawn out into a long spine and proximally expanded. In the other forms from which the interclavicle is known (Belodon, Mystriosuchus, Stagonolepis, Aetosaurus), the proximal expansion is comparatively smaller, and a considerable swelling follows in front of the pointed distal end.
The scapula and coracoid have been found in Belodon, Mystriosuchus, Stagonolepis, Aetosaurus, and Erpetosuchus. The scapula has the familiar elongated structure with a broad proximal end, where the disc-shaped, round coracoid with a notch on the side is placed; the latter is most considerable in Belodon, it is only arched in Stagonolepis, Aetosaurus, and Erpetosuchus. This is also the form in which the coracoid appears in Triassic Theropoda. The scapula of these is like that of Belodon, only even slimmer. The scapula of Erpetosuchus differs from the others by its slimness, stronger curvature, and the already-mentioned clear acromion. In this whole group of forms, the scapula and humerus are about of equal length.
The structure of the humerus varies rather more. The humerus of Belodon is expanded at both ends, and constricted in the middle. the strongly upwardly moved radial crest stands at nearly a right angle to the broad surface of the proximal end. The entepicondyle can often be almost confused with that of the former. The shovel-like proximal end of the humerus of Stagonolepis is much broader; the crest is placed rather more distally; the epicondyles are not present at the distal end as in Belodon (and Mystriosuchus). The similarity of the humerus of Thecodontosaurus and Stagonolepis is strikingly large. In a manner of speaking, this form is also similar to the single humerus from the Rhaetic of Linksfield in Morayshire that SEELEY described as Saurodesmus, a crocodile (1). The humerus of Episcoposaurus horridus COPE (2), from the New Mexico Triassic, is, like Belodon, very broad at both ends, with a strongly developed ectepicondyle, so that the contour of the medial side forms a nearly straight line, that of the lateral an approximately 45° angle. The humerus of Aetosaurus, with a rather broad proximal end and narrow distal end, shows great agreement with the Zanclodontidae. The proximal part of the humerus of Erpetosuchus is extremely narrow.
The ulna in Parasuchia and Pseudosuchia is quite crocodile-like with its broad and curved proximal part, which possesses an olecranon. The radius is narrow and has a concave proximal articular surface. In Parasuchia (Belodon, Stagonolepis, Episcoposaurus), the ulna is broader than in Pseudosuchia, whose forearm recalls the Theropoda even more, although the latter possess no olecranon.
Quite clear differences between the Parasuchia and Pseudosuchia are shown in the pelvic bones, and here, as usual, the latter approach the Theropoda much more than the Parasuchia. The figures speak better than words. The completely closed, thus crocodile-like, acetabulum in Belodon and Stagonolepis is characteristic. Aetosaurus only exists in poor preservation. In all these forms the main bulk of the ilium is found behind the acetabulum; there the ilium is extended into a long posterior point, just like Thecodontosaurus; only Stagonolepis is shorter than the others. The ilium of Theropoda is different from the former principally only by the completely open acetabulum.
The ischium is of similar form in all these genera and is very different from that of true Crocodilia. It does not stand transversely as in these, but almost entirely sagittally. Distally is runs into a long, stalk-shaped process; the proximal part is a broad plate that meets the pubis anteriorly and closes the acetabulum ventrally together with this. In the Pseudosuchia the stalk is more slender.
The pubis is most interesting. In Belodon it is a roundish plate perforated (1) by the obturator foramen; the medial edge is thin, the lateral thick. Aetosaurus has a flat plate as a pubis, which articulates on the ilium with a thick stem, the extreme part of which is missing. Ornithosuchus has a long, quite theropod-like, plate-shaped pubis, whose thick, narrow proximal end is turned over. The pubis of Stagonolepis also seems to be developed dinosaur-like, at least in the proximal parts. HUXLEY has figured a pubis (2) that appears quite crocodile-like, but which lacks the upper end. A piece never figured or mentioned in the literature lies in the collection of the Geological Survey in the Jermyn Street Museum, London, which SEELEY (private communication) holds as definitely the missing proximal part. In any case, Mr. E. T. NEWTON also assured me that the piece belongs to Stagonolepis, without however wishing to interpret the bones. I can hardly hold the bones as such for anything else than SEELEY. I figure the pieces here in outline, both of which I have handled, and give in addition a reconstruction according to SEELEY’s view (Fig. 73a and b).
The bones of the hind limb, particularly the femur, stand nearer the Crocodilia than the Dinosauria, although both show no great differences. The femur is always clearly S-shaped; it is only entirely straight in Episcoposaurus and lacks the fourth trochanter, which is otherwise always present, at least in outlines, and stands above the middle. The distal end of the fibula of Episcoposaurus recalls many Theropoda, with a concave and a strongly convex side.
The metatarsals and phalanges of the Parasuchia and Pseudosuchia are developed crocodile-like. In Theropoda they are interesting and important by the different degree of reduction of metatarsals I and V in the separate forms.
The armament of Triassic Crocodilia agrees with Mesozoic Crocodilia in that the back is covered with only two rows of long, decorated plates. Bernissartia fagesi DOLLO from the Wealden of Bernissart is the first crocodile that has four rows of dorsal scutes like those now living. The dorsal armor of Belodon, Mystriosuchus, Stegomus, Stagonolepis, and Episcoposaurus is adorned with two rows of strong dorsal spines; in Episcoposaurus they are up to 8.5 cm tall. the abdominal armor of all Parasuchia and Pseudosuchia consists of square scutes; in the best specimens of Aetosaurus, Mystriosuchus, and in Dyoplax one sees a large number of small plates. In block no. XXI of the Aetosaurus group, eight rows are recognizable. Teleosaurus temporalis E. DESLONGCHAMPS from the Upper Lias has only six rows of abdominal plates. No armor at all is known from Triassic theropod Dinosauria.
From the previous statements, I think it has been shown sufficiently that the Parasuchia and Pseudosuchia belong neither to true Crocodilia nor to Theropoda. However, they perhaps stand rather nearer the Crocodilia than the latter because of:
1) the armament.
2) the number and particularly the shortness of the cervical vertebrae,
3) the missing struts of the diapophyses of the dorsal vertebrae,
4) the closed acetabulum,
5) the structure of the femur, ulna, and metatarsals.
The form of the coracoid and the structure of the ischium, and to some extent pubis, certainly again indicate the Dinosauria very much.
The Parasuchia and Pseudosuchia on the one hand and the Theropoda on the other seem to me to be parallel branches of one stem, from which one may deduce what one wishes. I will go into this question in more detail elsewhere. The one cannot be the progenitor of the other, for they appear simultaneously.
The question now arises as to how the two groups Parasuchia and Pseudosuchia are related to each other. It is to be considered that the latter are less specialized in many respects, although the quite crowded anterior dentition of Erpetosuchus does not entirely correspond to this feature. In the Parasuchia which are gavial-like in appearance, the excessive length of the premaxillae as well as the posterior placement of the narial openings is a high-degree specialization that, as far as it concerns the last point, has never been adopted by later true Crocodilia, i.e. the same is achieved there by the posterior placement of the choanae (1). In this the forms classified as Parasuchia prove to be secondarily evolved. For all that, both these groups have so much in common that links them to each other and contrasts them to the Eusuchia and Dinosauria, that it seems proven that this should be expressed in the systematic nomenclature, so that both groups should be placed in a single suborder. Because the name Parasuchia is the older, the precedence will also belong to it. However, the differences between both groups should not be underestimated in this. They can be distinguished as the families Belodontia and Pseudosuchia.
I may now indicate in a few words the enormous distribution of the Parasuchia in the uppermost Triassic.
Belodon has been made known sufficiently from the Upper Keuper of Swabia and France by H. von MEYER, GÜMBEL, etc.; it also occurs in the Brunswick Rhaetic Keuper. GRESSLY and RÜTIMEYER have collected very fine pelvic bones, teeth, etc. in the Knollenmergel of the uppermost Keuper of Schönthal near Basel. Teeth are known from the Keuper of Warwickshire, Somerset, and Devon in England. LYDEKKER has found Belodon in the Indian Maleri Sandstone, and COPE, EMMONS, LEIDY, LEA, etc. have shown teeth, skulls, skeletal parts in great distribution in the North American New Red. Mystriosuchus is known from the Swabian Stubensandstein, Stagonolepis only from Elgin (Sandstone), Parasuchus from the Indian Maleri Sandstone. Rytiodon (Rutiodon) EMMONS from North Carolina and Connecticut probably stands nearest Mystriosuchus; only the upper temporal fenestrae are closed posteriorly by the squamosals, and the quadrate is directed obliquely anteriorly instead of posteriorly. MARSH has named Stegomus from a dorsal armament from the Connecticut River Sandstone, which is perhaps identical with Belodon. Saurodesmus SEELEY (2) from the Rhaetic of Morayshire is based on a single humerus. Episcoposaurus COPE (3) is found in the Triassic of New Mexico. Other isolated occurrences are the finds of Aëtosaurus O. FRAAS in the Schilfsandstein, Dyoplax O. FRAAS in the Schilfsandstein of Württemberg, and Ornithosuchus and Erpetosuchus E. T. NEWTON in the Elgin Sandstone.
This otherwise well-known presentation shows the great distribution of Parasuchia in the Upper Triassic. Only Belodon can be called really abundant, both in Europe and America, where a series of species are distinguished. Stagonolepis was certainly only found in the neighborhood of Elgin, but not infrequently there.
NEW AND MISINTERPRETED PARASUCHIA
I. E. FRAAS in “Festschrift über die Schwäbischen Trias-Saurier” 1896, p. 18, has mentioned a piece of a lower jaw from the Schilfsandstein of Feuerbacher Haide as Zanclodon arenaceus n. sp. without however figuring it. He says about it: “there is no doubt about the affiliation to the Dinosauria.” It is incomprehensible to me how such an expert on the Triassic Reptilia could say this. In the above sentence I might replace the word “Dinosauria” with “Parasuchia” or “Belodontia”. The figure will show anyone that this is a case of Belodon or Mystriosuchus. On the old label, dated 1869, the piece is also named Belodon kapffi. This find is only interesting in that Belodon thus dates back in Württemberg to the Lower Keuper.
II. Rileya bristolensis n. gen. et sp. Another find of Parasuchia that I may mention was made as long ago as the 1830s by RILEY and STUTCHBURY in the Magnesian Conglomerate of Bristol, but hitherto not recognized as such.
There are two vertebrae and a humerus which were found among a great mass of Thecodontosaurus and Palaeosaurus bones and together with them. The humerus was figured by RILEY and STUTCHBURY (1) and has never since been opposed. SEELEY (2) said that at first glance it probably recalls Belodon, but since he places the bones reversed, as I attempt here, he comes however to the conclusion that it is a dinosaur humerus, although as he himself says (3), no similar theropod humerus (Saurischia) is known to him. The humerus (No. 95 and 96 in the Bristol Museum) has a length of 17 cm, is 2 cm broad in the middle, and 5.5 cm at both ends. It is very slimly built. The strong distal condyles and particularly the ulnar are striking. The proximal part is unusually narrow, and the radial crest only weakly developed. It seems to me that the bone shows the posterior side and then must be a right humerus, however I cannot maintain this with certainty. Compared with Belodon, the median longitudinal side also appears here curved inward; the lateral side is straight. However, the radial crest forms a peculiar acute angle above, and the ectepicondyle is missing, as in all other Parasuchia and the Theropoda. In Erpetosuchus the proximal and distal ends are also as broad or narrow as here.
Also the two centra cannot belong to a dinosaur. They were later obtained for the Yale Museum, New Haven, Conn., U.S.A., by exchange. From a photograph taken at natural size, for which I thank Prof. BEECHER, one is 3 cm long and 2.5 cm tall, the other 4.8 cm long and 2.3 cm tall, thus very elongated, more than the caudal vertebrae of Thecodontosaurus. In general they are too long, too round in cross-section, and far too constricted in the middle for Dinosauria. The articular surfaces are slightly hollowed. All this indicates Parasuchia. The vertebrae of Belodon are not so long, but those of Stagonolepis are very similar. since, among the extremely numerous bones of Theropoda, only the humerus and two vertebrae belong to Parasuchia, it is very likely that they come from the same animal; so I combine them both here provisionally under one name. The locality is the dolomite conglomerate (Middle Keuper) of Durdham Down (Redland), near Bristol.
TRUE (?) CROCODILIA FROM THE LOWER MUSCHELKALK
I) In the Tübingen Collection are two caudal vertebrae from the Wellen dolomite of Merklingen and from Hausen on the Würm. Both lack the neural arch.
The larger specimen from Merklingen (pl. VIII, fig.5), a middle caudal vertebra, is 2.5 cm long, 2.2 cm high, and 2.0 cm broad. On the underside it has a clear groove, and on the posterior edge below, the two attachment surfaces for the haemapophyses. The edge of the rather flat articular surface is upturned and curved, particularly on the posterior side. The neural canal sinks rather deeply into the middle of the centrum and is broader posteriorly than anteriorly.
The second piece is one of the last caudal vertebrae (pl. VIII, fig. 6). It is 1.5 cm long and 7 mm high. It also possesses a groove on the underside. The articular surfaces are slightly hollowed. The sides show indistinct longitudinal ledges, which probably served for the attachment of the strong caudal tendons. The neural canal is quite deeply sunk into the centrum.
I think that such caudal vertebrae seem to agree best with a crocodile, but I cannot prove this.
II) The Geological Collection at Halle possesses a fine large caudal vertebra from the Schaumkalk of Wolfsburg at North Harz. It belongs without doubt to a true crocodile. Despite its length of 4.5 cm, it must come from the posterior half of the tail, for the entire neural arch is so extraordinarily small, one could say rudimentary, and the transverse process is completely missing. Therefore it must have been a huge animal. In the place of a more detailed description I refer to the sketches reproduced here. This isolated find is not sufficient for further conclusions.
TRUE CROCODILIA FROM THE UPPER MUSCHELKALK
Several pieces from the Upper Muschelkalk still lie before me which I can only place in the Eusuchia:
I) Procerosaurus cruralis n. gen. et sp. (pl. IX, fig.1)
A find that may be of interest is a left femur from the Upper Muschelkalk bonebed of the Oelmühle, Württemberg (1). Unfortunately it lacks the femoral head. The whole piece is rather compressed anteroposteriorly. The fourth trochanter is not developed in this species. The distal condyles are just as in Crocodilia and Theropoda. The preserved length of the bone is 30.5 cm, the breadth in the middle 4 and at the distal end 6 cm; but the compression is to be taken into account in these, for the thickness in the middle is only 1.8 cm. The medial condyle is compressed from the medial side, which causes the great breadth of the distal end. The femur is distinguished in particular by its slim, rather S-shaped structure and the lack of the fourth trochanter. The latter in particular seems to me to prove that it belongs to a dinosaur (2). In the Parasuchia and Crocodilia the fourth trochanter is usually only weakly indicated. An edge is present at the place in question in the Crailsheim piece. The bone is much slimmer than all parasuchian femora. The femur of Episcoposaurus is certainly also quite narrow but should be straight. Therefore I hold Procerosaurus for a eusuchian. It has the most similarity with a gavial femur.
II) Pectenosaurus strunzi n. gen. et sp. (pl. VI, fig. 5)
There are two caudal vertebrae that come from the main Muschelkalk of Bayreuth and were collected by STRUNZ. The more complete piece is a vertebra from the middle of the caudal vertebral column, whose neural arch refers it to the Eusuchia. The spinal process lies far back; a narrow lamella is placed on it on the anterior side as in the gavials; its contour is rather curved in the middle where the bone fibers are directed vertically toward the edge, on the posterior side they run more upward and parallel to the spinal process, on the anterior toward the prezygapophysis. For the anterior caudal vertebrae this fact seems to me to indicate similar features to those in Rhacheosaurus, Geosaurus, or Gavialis (gangeticus) in that a spine is found anterior to the spinal process. This is unknown in any Parasuchia. The other vertebra belongs to the same species; its articular facets are almost entirely flat; the centrum is rather flattened on the underside.
III) A left articular + angular from Busendorf, Lorraine (pl. VI, fig. 3) has great similarity with that of the gavial; it is certainly not different. The angular is extended longer anteriorly and relatively narrower. The articular surface for the quadrate is inclined more outward than is the case in Recent Crocodilia and seems to indicate a more steeply placed quadrate, also the edge is not so high behind the true articular surface in Recent Crocodilia. The piece shows the contact surface with the splenial and is seen from the lateral side. On the figure, the right side is behind and the left is in front; the most posterior end of the articular is missing.
The important facts about the Triassic Dinosauria have been mentioned in the foregoing section. Details with be found in a monograph in preparation. The large Dinosauria are confined to the Upper Triassic, the younger Keuper and its equivalents. The huge bones of Euskelosaurus are also found in South Africa in the uppermost beds of the Karroo. Smaller forms occur already in the Muschelkalk; they are still undescribed (I refer to the monograph). Tanystrophaeus, which stands near the Dinosauria in many points, and which is not yet fully known, is also found in the Lower and Upper Muschelkalk. The Triassic Dinosauria of the entire Earth are divided into about 24 genera that may be grouped in four families.
Isolated remains of doubtless Pterosauria have only been found in the very youngest beds of the Triassic. Only a few such occurrences are known to me from the literature.
The first probably flying or fluttering reptile, Tribelesodon longobardicus, has been described by BASSANI from the Raibl beds above Besano (Atti della Soc. Ital., vol. 29, 1866, 25-30, without figs.). It is confused by another skeleton mixed up with it. Depending on the interpretation of the bones, it may or may not have possessed flying fingers. I can say nothing certain on the systematic position. A description of this by F. von NOPCSA is awaited soon.
DEFFNER and O. FRAAS (1) found poorly preserved flight fingers of a pterosaur named by them Pterodactylus primus in the sand marl of the bonebed of the Galgenberg near Malsch (near Wiesloch) in Baden, close above the Rhaetic sandstone.
In 1857 OPPEL found three flying finger pieces in the sandstone of the contorta zone at Birkengehren near Esslingen. H. von MEYER has described and figured two of them (one is lost) (2). These remains are not enough to support conclusions on the nature of the animal.
ANDLER (3) and H. von MEYER (4) found a split flight finger phalanx, which should be specifically different from that of OPPEL, in the “limestones” of Aichscheiss near Esslingen, which ANDLER assigns to the angulata beds and H. von MEYER to the contorta beds.
West of Wachet, Somerset, DAWKINS found a hollow Pterodactylus bone in the Rhaetic with Avicula contorta and Hypsiprymnopsis (5).
GERVAIS (6) also knew a few broken pieces of pterosaurian phalanges from the only slightly younger, certainly Liassic, sandstones of Hettingen, Lorraine.
All these fragments unfortunately throw no new light on the oldest Pterosauria. The only indicate that the adaptation of the forelimbs was already complete.
In the Natural History Cabinet, Stuttgart, I recently discovered the neural arch of a cervical vertebra that could possibly belong to Pterodactylus (pl. IX, fig. 3); it comes from the Rhaetic bonebed and bears the figure E.51, but unfortunately no locality description. The Tübingen University Collection has one like it from the Rhaetic sandstone of Schlösslesmühle near Steinbronn; it is only 3 mm shorter than the better-preserved Stuttgart specimen. The latter is 28 mm long. The spinal process is broken off, its base is 10 mm long. The zygapophyses are flat, however not quite horizontally placed; the anterior are long and extended forward and rather upward/outward, the posterior corresponding; the articular surface is 8 mm long and 4 mm broad; its outer edges are 13 mm apart, the breadth of the vertebra in the middle measures only 9 mm. Below the prezygapophyses a sharp edge runs downward. The height of the neural arch, which seems to be preserved up to the neurocentral suture, measures 9 mm up to the base of the missing spinal process. The length of the centrum seems to have been 16-17 mm. This vertebra shows great similarity with Rhamphorhynchus and Campylognathus, the long-tailed forms.
Among the MILLER-ENDLICH material from the Rhaetic bonebed of Bebenhausen, in the Tübingen University Collection, are found two smaller short phalanges that could very probably belonged to the hind limbs or the claw-bearing digits of the manus of a Pterodactylus. The larger piece, a fragmentary distal end, measures 4 mm in length, but could have been 1.5-2 cm long when complete.