The Origin of Snakes

Summary 1. Current theories of ophidian evolution suggest that the snakes have been derived either from aquatic, above‐ground terrestrial, or burrowing ancestors. It is generally supposed that the ancestors of snakes, irrespective of their habits, were closely related to the platynotid lizards. In the absence of adequate palaeontological data, the respective merits of these assumptions must mainly be assessed on evidence obtained from study of living forms. 2. The close relationship between the snakes and lizards is pointed out in this article, and the main characteristics of the different families of fossil and living snakes are reviewed. 3. Many important features of the snakes as a group are described and their functional significance is discussed. Comparison with the corresponding condition in lizards is made. Ophidian characters regarded as being of particular significance include the bony investment of the forebrain, the platytrabic chondrocranium with absence of the interorbital septum and endocranial side‐walls, the absence of the epipterygoids and temporal arcades and the extreme mobility of the jaw skeleton in most forms; the presence of zygosphenal articulations throughout the vertebral column; the elaboration of the trunk musculature and ventral scales; the extreme concentration of alkali‐metal cation on the secretory side. (For reviews of evidence, see Crane (1950), Hokin & Rehm (1947), Rehm (1950).) (2) When the two sides of the mucosa are connected electrically, a current of relatively large magnitude can be maintained which requires the continuous performance of electrical work (Rehm, 1943; Rehm & Hokin, 1948; Crane, Davies & Longmuir, 1946, 1948 a ; Crane & Davies, 19486). (3) There is a fall in the p.d. across the mucosa with the onset of secretion, but once secretion is well established further increases in the rate of secretion can occur with only small changes in the p.d. (cf. Crane, 1950; Rehm, 1944; Rehm, 1950). (4) The change from the non‐secretory to the secretory state and back is always associated with concomitant changes in the electrical parameters of the mucosa, whether these changes occur spontaneously or by drug action. In the absence of oxygen and in dead mucosae both p.d. and acid secretion are abolished (cf. Crane, 1950; Rehm & Hokin, 1947; Rehm, 1950). (5) Passage of electric current (c. 1 mA./cm. 2 ) through the mucosa from an external source increases or decreases the rate of secretion of acid by a secreting mucosa according as the applied current enhances or opposes the spontaneous potential measured between calomel electrodes on either side of the mucosa (Rehm, 1945; Crane et al. 1946, 19486). (6) Passage of current does not initiate secretion in the dog (Rehm, 1945) but may do so in the frog (Crane et al. 1948 b ). These results were obtained with intact gastric mucosa, and because of the technical difficulties involved it has not, so far, been possible to obtain direct information regarding the detailed electric changes in and around the oxyntic cells. A difficulty is that these cells are mostly isolated from each other and make up only about one‐tenth of the volume of the mucosa (Davies & Roughton, 1948) and about one‐twentieth of the dry weight (calculated from Engström & Glick, 1950). However, quantitative relations have been estimated (Rehm, 1943; Rehm & Hokin, 1948; Rehm, 1950; Crane & Davies, 1948 b , 1950 b ), and it seems certain that the energy required for secretion by acid‐secreting gastric mucosa is greater than the maximum electrical energy output of non‐secreting gastric mucosa. Since acid secretion is associated with a great increase in the rate of respiration of oxyntic cells (Davies, 1948 a ), and hence in the available metabolic power, care is clearly needed when comparing activities in the secreting and non‐secreting mucosa. It is more difficult to determine the relation between the energy required for the increase in acid secretion during the passage of electric current from nutrient to secretory side, and the energy available from this current. Although the total power expended by the current is greater than that required, the fraction capable of doing useful work is probably less. On the other hand, it seems probable that the gastric mucosa can supply sufficient coulombs of electricity to account for the production of the hydrogen ions, and that this also applies to the increased rate of acid secretion when current from an external source is passed through the mucosa (Rehm, 1950). Before attempting to include these observations in a theory of the mechanism of hydrochloric acid production it is necessary to return to a further consideration of the biochemistry of gastric mucosa. the condition in burrowing lizards, together with the other primitive features displayed, suggest that forms such as Cylindrophis (Fig. iA) and Ilysia have not departed greatly from the basal ophidian type. The late Cretaceous Dinilysia seems unlikely to have been a burrower on account of its relatively large size. From the evolutionary standpoint adopted, this snake may be considered as representing an early venture in above‐ground radiation on the part of this archaic family. The Uropeltidae shows many resemblances with the Anilidae (Haas, 1931 b) , but has acquired certain aberrant features (e.g. modified occipito‐vertebral articulation), and the vestigial hindlimbs have undergone further reduction. The presence of two transverse ventral scale rows to each body segment (possibly a primitive feature and shared by the Leptotyphlopidae and Typhlopidae) suggests that the Uropeltidae have not been directly derived from the Anilidae. The Boidae and the allied Xenopeltidae possess many primitive characters, especially of the viscera and vascular system; at the same time they show certain advanced features and have undergone a minimum of aberrant specialization. The central position accorded to these families in Boulenger's scheme (1893‐6) appears to be fully justified. While zoo‐geographical considerations seem to show that the Boinae are the more primitive of the two boid subfamilies (Darlington, 1948), the occurrence of such features as premaxillary teeth and a supraorbital bone among the Pythoninae suggest that the latter is the older group. The whole family is of particular interest since its different members perhaps illustrate, both in structure and in range of habits, the characteristics of the snakes in general at the end of the Mesozoic, during the early stages in their radiation as an above‐ground group. Summary 1. Current theories of ophidian evolution suggest that the snakes have been derived either from aquatic, above‐ground terrestrial, or burrowing ancestors. It is generally supposed that the ancestors of snakes, irrespective of their habits, were closely related to the platynotid lizards. In the absence of adequate palaeontological data, the respective merits of these assumptions must mainly be assessed on evidence obtained from study of living forms. 2. The close relationship between the snakes and lizards is pointed out in this article, and the main characteristics of the different families of fossil and living snakes are reviewed. 3. Many important features of the snakes as a group are described and their functional significance is discussed. Comparison with the corresponding condition in lizards is made. Ophidian characters regarded as being of particular significance include the bony investment of the forebrain, the platytrabic chondrocranium with absence of the interorbital septum and endocranial side‐walls, the absence of the epipterygoids and temporal arcades and the extreme mobility of the jaw skeleton in most forms; the presence of zygosphenal articulations throughout the vertebral column; the elaboration of the trunk musculature and ventral scales; the extreme reduction or loss of limbs; the great development of the vomeronasal apparatus; many peculiarities of the eyes and their adnexae (study of which led Walls to advance the view that the snakes were derived from forms in which the eyes had degenerated markedly in association with burrowing habits and subsequently became re‐elaborated, permitting return to above‐ground life); the absence of the tympanic membrane and middle ear cavity; the elongation and asymmetry of the viscera and their blood supply; the presence of paired copulatory organs in the male. 4. Comparison with lizards indicates that many of the latter show a close approach to the ophidian type of organization, although the features common to all lizards and snakes are comparatively few. 5. The phylogeny and characteristics of the platynotid lizards are described, and the similarities between the different platynotid groups and snakes, both fossil and living, are discussed. While the Platynota in general, and the Dolichosauridae in particular, show many snake‐like features, they appear to differ markedly from all snakes in the structure of the skull and in certain other characters. It is suggested that the evidence is inadequate to establish close relationship between the snakes and any of the known platynotid types; the origin of snakes from unknown platy‐notids during the early Mesozoic remains a possibility. 6. The very numerous resemblances between burrowing lizard