The Osteology of the New Zealand Geckos and its Bearing on Their Morphological Status

Transactions and Proceedings of the Royal Society of New Zealand, Volume 84, 1956-57, Page 341

The Osteology of the New Zealand Geckos and its Bearing on Their Morphological Status By N. G. Stephenson* University of Sydney. and Elsie M. Stephenson� N.S.W. University of Technology [Received by the Editor, September 8, 1955.] Abstract Osteological investigations of the endemic New Zealand geckos, Naultinus and Hoplodactylus, were carried out by means of alizarin transparencies, serial sections and X-ray photographs. As a result, a description of the entire cranial and post-cranial skeleton in both forms is given. It is concluded from this study and from a comparison with other geckos that the New Zealand genera, in their osteological features at least, represent the most primitive known members of the Geckonidae. Introduction Some geckos are unique among the Lacertilia in possessing amphicoelous vertebrae and persistent intercentra. From the viewpoint of vertebral structure the family of geckos therefore includes what have been recognised as the most primitive of living lizards in spite of any specialised characters they may possess. Boulenger (1885) placed the Geckonidae first amongst his Lacertilia Vera. Camp (1923) placed the section Gekkota first in his division Ascalabota of the Sauria, and this classification of lizards has remained a basic scheme for many years. An implication of Underwood's (1954) proposed classification of the geckos, which is discussed more fully below, is that the procoelous vertebral condition is primitive in geckos and that the amphicoelous condition is secondary. Underwood has subsequently indicated that in a further paper he is withdrawing this view, but its acceptance, which initially was a logical outcome of his placing of geckonid genera, would have had considerable effect on the position of the Gekkota as a whole. At the present time there is much need for a revised classification of the entire Squamata. Shute and Bellairs (1953) point out that “further evidence, based on the widest range of morphological characters, is clearly required before the complex interrelationships of the different families of lizards can be elucidated.” McDowell and Bogert (1954) in referring to this matter stress that the project of a revised classification “would require a careful study of the anatomy of more different species than have been examined by workers up to this time.” Morphological studies on members of the primary family Geckonidae might well be expected to have an important bearing on contemporary relationships as well as on the evolutionary history of lizards. It is with this possibility in view that the present study of the New Zealand geckos has been undertaken. Material and Methods Geckos are represented in New Zealand by two endemic genera, Naultinus Gray and Hoplodactylus Fitzinger. Naultinus has a single species, whereas four species of Hoplodactylus are recognised by Malcolm Smith (1933). In this study, specimens of Naultinus elegans and of two species of Hoplodactylus (H. duvaucelii and H. pacificus) have been used. In view of the similarity between the latter two species and of the fact that the morphological features discussed are

mainly at the generic level, no specimens of H. granulatus and H. maculatus were obtained for study. In any case, McCann (1955), who has carried out a systematic study of the New Zealand lizards, claims that H. maculatus is at the most a synonym for H. pacificus. The bulk of the material for study consisted of alizarin transparencies which were prepared by standard methods. Care was exercised when clearing in caustic potash, and in the case of one large, spirit-hardened specimen the clearing process took over six months. In all cases the results were satisfactory for a clear distinction between bone, calcified cartilage and uncalcified cartilage. The transparencies were first studied in toto and then dissected out under a binocular microscope. Portions of the head and axial skeleton of Hoplodactylus pacificus were embedded and serially sectioned in order to clarify doubtful points. X-ray photographs were used in the study of the postcranial skeleton. In order to provide comparative material, alizarin transparencies of three Jamaican and one Australian gecko were prepared. These were Aristelliger praesignis Hallowell, Sphaerodactylus parker Grant, Gonatodes fuscus Hallowell, and Phyllurus (Gymnodactylus) platurus White. Acknowledgments Our thanks are due to Mr. E. G. Turbott, Dr. C. A. Fleming and Mr. C. McCann for assistance with New Zealand geckos, and to Dr. G. Underwood for Jamaican material. We are furthermore indebted to Professor P. D. F. Murray for helpful criticism and suggestions. The Skull (Figs. 1 and 2.) Of the median dorsal skull elements, the nasals and parietals remain separate throughout life in both Naultinus and Hoplodactylus. The premaxillae are more or less fused in the adults and contribute to a wedge-shaped prenasal process which extends up between the two nasals. The frontals also are fused, but the degree of fusion, as with the premaxillae, is less in Naultinus than in Hoplodactylus. Similarly, the occipital elements (supraoccipital, exoccipitals and basioccipital) are separate from one another in Naultinus but indistinguishably fused to form a more solid ring of bone in Hoplodactylus. As is characteristic of recent geckos (Camp, 1923), there is no pineal foramen in either genus. In Naultinus the parietal roof is largely membranous towards the middle line, and a large membranous area may also occur medially in the floor of the cranium, between the basioccipital and the basisphenoid. In a general comparison of Naultinus and Hoplodactylus it is noted that adults of the former genus show a lack of ossification in many regions of the skeleton and less tendency towards the fusion of paired elements. Such features might well be interpreted as being neotenic. The geckonid tendency towards reduction of membrane bones, which has resulted in the loss of skull arches, is the most striking feature of the skulls. The consequent incompleteness of the temporal region in geckos has prevented these otherwise primitive lizards from throwing valuable light on the much debated question of modified parapsid or diapsid origin of the lacertilian skull (Watson, 1914; Williston, 1914; Broom, 1925; Versluys, 1936). The temporal space is neither divided into upper and lower temporal fossae nor demarcated from the orbit. Distinct bones, shaped like arrow-heads, are situated at the sides of the parietals and frontals, at their sutures with one another. These, by their relationships in geckos and by comparison with the elements present in other lizards, are normally regarded as postfrontals, though possibly they represent a fused postfrontal and postorbital on each side. Elements similar to the postfrontals of the New Zealand geckos are noted in alizarin transparencies of Phyllurus platurus, Aristelliger praesignis, Sphaero-

dactylus parkeri and Gonatodes fuscus (Sphaerodactylidae). Postfrontals have also been described in Gekko gecko (Boulenger, 1912). Hemidactylus flaviviridis (Mahendra, 1949); Aristelliger lar and Coleonyx variegatus (Eublepharidae) (McDowell & Bogert, 1954). However, they are not universally present for Brock (1932) notes that postfrontals are present in Pachydactylus maculosa and Pachydactylus bibroni but absent in Lygodactylus capensis. Brock states that in Pachydactylus bibront only a small strip of the postfrontal lies along the margin of the frontal, but that the part of it in contact with the Text-fig. 1—A, Skull of Naultinus elegans, dorsal view B, Skull of Hoplodactylus duvacelii, dorsal view C, Skull of Hoplodactylus duvauceli, lateral view ept, epipterygoid; exoc, exoccipital; fr, fused frontals; jugal; mx, maxilla; na, nasal; pa, parietal; pal, palatine; pinx, premaxilla; poc, paroccipital process; poj, postfrontal; prf, pre-frontal; pro, prootic; pt, pterygoid; q, quadrate; ; supraoccipital; stm, supratemporal; tr, transpalatine.

parietal is expanded into a broad plate of bone. This feature, which is particularly accentuated in Aristelliger lar, indicates a reasonable possibility that the broad plate of bone bordering the parietal on each side represents a postorbital fused with the postfrontal but not yet greatly reduced. In none of the above mentioned species is there a prefrontal-postfrontal junction above the orbit. Boulenger (1885) used this feature as one of two points distinguishing the skull of a gecko from that of a pygopod, but, as McDowell and Bogert (1954) point out, this distinction will not hold. Postorbitals, at least as separate elements, are lacking in the adults of both Naultinus and Hoplodactylus, as also are the squamosals. Closely applied to the postero-lateral border of the supratemporal process of the parietal (parietal process) on each side, there is present in both genera a fairly distinct, splint-like element, the supratemporal or tabular. Against the inner side of this supratemporal the head of the paroccipital process abuts, while the dorsal head of the quadrate is directed towards both of them from below. Brock (1932) notes in Lygodactylus a bony splint on the outer surface of the quadrate which is present in embryos as a definite membrane bone. This element, present also in Pachydactylus, is regarded by her as being a vestigial quadratojugal. Examination of detached quadrates of adults of Naultinus and Hoplodactylus reveals no trace of a quadratojugal, but in this respect no information is available regarding developmental stages in the two genera. On the other hand, Brock states that a jugal is altogether lacking in Lygodactylus and merely a microscopic vestige in Pachydactylus where it is seen in adult transverse sections. In Naultinus and Hoplodactylus, however, although the jugal is a slender bone, it is a definite macroscopic element between the maxilla and the transpalatine. This is also the case in Hemidactylus flaviviridis (Mahendra, 1949); Aristelliger lar and Coleonyx variegatus (McDowell & Bogert, 1954): and Gekko gecko (Lakjer, 1927). From our own observations also, a jugal was noted in Sphaerodactylus parkei and Aristelliger praesignis, but could not be distinguished in an alizarin transparency of Gonatodes fuscus. Where it occurs in geckos, the jugal is reduced to a horizontal splint and does not exhibit a more vertical bony extension towards the postfrontal. In this position a ligament extends upwards and slightly backwards from the jugal to the postfrontal in the New Zealand genera and represents a remnant of the postorbital arch which is normally lost in geckos. Lakjer (1927) has noted a similar ligament in Gecko gecko. In Naultinus and Hoplodactylus, the orbit is bounded anteriorly by a conspicuous prefrontal. No lachrymal is visible in the adult skulls, and this is also the case in Hemidactylus flaviviridis (Mahendra, 1949). Brock (1932), however, identifies a small lachrymal in sections of a newly hatched Lygodactylus. Where it occurs in geckos, this bone is crowded within the orbit and, according to Camp (1923), lost to view externally. According to Walls (1942), most lizards usually have fourteen scleral ossicles surrounding the eyeball, although in Sphenodon the scleral bones number sixteen or seventeen. Until more recent years, the highest known number in living animals has been in birds which may have as many as eighteen. An examination of scleral ossicles in the New Zealand geckos reveals that Naultinus has twenty, while Hoplodactylus has twenty-five. The number in this latter genus was first noted by Underwood (1951) to whom specimens of Hoplodactylus pacificus were sent in connexion with his work on reptilian retinas. This count has subsequently been confirmed by us in other specimens of H. pacificus, and also in H. duvaucelii. Underwood's (1954) count for Naultinus is eighteen and twenty-two. Furthermore, he notes that in the case of the Australian gecko Oedura lesueurii there are thirty scleral ossicles. This number is not exceptional for the Australian representatives, for we have counted the same number in Phyllurus platurus.

The greater number of scleral ossicles has been regarded as the more primitive condition on the grounds that Stegocephalians usually had twenty to thirty-two scleral ossicles set in several rows Underwood (1954) indicates that there may be individual variation in the number of scleral ossicles, but at the moment there seems inadequate evidence to support his contention that there has been a multiplication of ossicles to account for the “supernumerary” condition in geckos. The floor of the orbit is formed partly by the jugal, palatine, transpalatine (ectopterygoid) and pterygoid bones. Like the palatines, the pterygoids are very widely separated from one another. Between them in the middle line, the basisphenoid and Text-fig. 2.—A, Skull of Naultinus elegans, ventral view B, Skull of Hoplodactylus duvaucelii, ventral view. C, Lower jaw of Hoplodactylus duvaucelii, left side. ang, angular; art, articular; boc, basioccipital; bs, basisphenoid; bpt, basipterygoid process; c, occipital condyle; cor, coronoid; d, dentary; ec, extra-columella; ept, epipterygoid; h, hyoid; j, jugal; mx, maxilla; pal, palatine; pmx, premaxilla; poc, paroccipital process, ps, parasphenoid, pt; pterygoid; pv, prevomer; q, quadrate; sa, surangular; st, columella auris; tr, transpalatine.

parasphenoid extend forwards. Each pterygoid is connected to the parietal (anterior superior) process of the prootic by an upwardly extending rod, the epipterygoid (columella cranii). The quadrate process of the pterygoid extends back to articulate with the quadrate proper. Supporting the floor of the nasal capsules are the prevomers. These are separate bones lying side by side in the middle line of the palate. Between each prevomer and the more lateral maxilla of its side lie both the choana and, more anteriorly, the foramen of Jacobson's organ. The septomaxillae lie inside the olfactory capsules and are exposed only by the removal of the nasals. Each septomaxilla forms a roof over Jacobson's organ of its side and rests on the prevomer. The relationship of bones in the region of the auditory capsules is best seen in Naultinus, where there is less fusion of skull bones and where, except for the exoccipital and opisthotic on each side, the various auditory and occipital elements are separate. The prootic, forming the anterior part of the auditory capsule is large and irregularly tri-radiate. Dorsally, it meets the supraoccipital behind and extends up as the parietal process in front; at a mid-lateral level its posterior process reaches back almost to the level of the paroccipital process; ventrally it extends down to flank both the basioccipital and the basisphenoid. The opisthotic is much smaller and although distinct in Naultinus from the adjacent supraoccipital, prootic and basioccipital bones, it is indistinguishably fused with the exoccipital. It is close to this region of fusion that the stout paroccipital process arises and extends out laterally. Below the paroccipital process, and just in front of it, is the fenestra ovalis, which is bounded partly by the prootic and partly by the opisthotic. Fitting into the fenestra ovalis is the expanded foot plate of the stapes. In the main the stapes consists of a narrow rod, the columella auris, bearing distally the extra-columella. Slightly below the fenestra ovalis and bounded largely by the opisthotic, though bordered partly by the basioccipital, is the fenestra rotunda. The Lower Jaw As in Hemidactylus flaviviridis (Mahendra, 1949), the rami of the lower jaw are united by a ligament and not suturally. The five bones comprising each ramus are the dentary, splenial, coronoid, surangular and articular. With the latter is fused not only the prearticular but also the angular (Cope, 1892; Camp, 1923). Hyobranchial Skeleton (Fig. 3.) Cope (1892) first pointed out the evolutionary importance of the branchial arches among lizards. Camp (1923), after a study of 26 genera in various Lacertilian families, assessed the relative importance of various primitive characters and concluded that the characters of branchial arches are foremost in phylogenetic significance. In Naultinus and Hoplodactylus, the hyoid and first and second branchial arches persist. Of these three, the hyoid and second branchial arch are attached to the skull. The arches in both genera are similar, except that in Hoplodactylus a definite, though sometimes slight gap appears in the second branchial arch. This occurs between ceratobranchial II (= basibranchial II, Noble, 1921) and epibranchial II, so that the mesial end of the latter lies free in the surrounding muscle. Naultinus shows an even more primitive condition than Hoplodactylus in that no such gap accurs in the second branchial arch. Thus two complete arches reach the skull. In this respect Naultinus agrees with the condition in the procoelous Eublepharine geckonid, Coleonyx variegatus, described by Noble (1921) as having the most primitive type of lacertilian hyoid apparatus. Apart from the identity.

Text-fig. 3.—The hyoid apparatus. A, Naultinus elegans. B, Hoplodactylus duvaucelii cbr I; ceratobranchial I, cbr II; ceratobranchial II; ebr I; epibranchial I, ebr II, epibranchial II; h, hyoid arch hc, body of hyoid; pnt, processus entoglossus.

and relationship of parts, there is a remarkable similarity in the shape of the hyobranchial skeleton in Coleonyx and Naultinus, as also in Hoplodactylus. The hyoid arch in Naultinus and Hoplodactylus is attached distally to the paroccipital process of the skull, whereas the second epibranchial is attached close to the base of this process. The incomplete first branchial arch in both genera consists of a bony ceratobranchial I, to which is attached the epibranchial I as a curved, bluntly ending piece of cartilage. From the hyoid body itself, a uniformly narrow and elongated processus entoglossus extends forward into the tongue. At the anterior curve of the hyoid arch in both genera, there projects a short piece of cartilage continuous anteriorly with a further piece at right angles to it. This latter ends in one or more irregularly shaped, blunt projections on each side. Noble (1921) does not illustrate any structure similar to this, but Camp (1923, p. 445) figures the hyoid of Coleonyx variegatus and shows processes from the hyoid arch which are similar to those of Naultinus and Hoplodactylus. Goodrich (1930) illustrates a widening of the arch at this point in Lacerta. Whatever the significance of their development in the above genera, these processes correspond in position to the anterior processes found on the ceratohyal of certain frogs. Vertebral Column (Fig 4.) The vertebrae in the New Zealand genera are of the primitive amphicoelous type found only in certain geckos among the Lacertilia. Persistent intercentra are present as more or less kidney-shaped elements. The vertebral columns of all the New Zealand species examined show only minor differences, but in view of the relatively few individuals of each species examined, the amount of variation found is striking. The types of variation appear to have no evolutionary significance. In normal specimens, the number of presacral vertebrae is 26. Occasionally 27 occur. The Atlas In Hoplodactylus, a clear suture is shown mid-dorsally between the neurapophyses, although they are quite closely apposed. In Naultinus a slightly wider gap occurs, and this condition is described also in Hemidactylus, where it is regarded by Mahendra (1950) as indicating a more primitive state. The neurapophyses are fused ventrally with the hypocentrum and do not conform with the unique condition in Hemidactylus where the atlantal neurapophyses and hypocentrum are connected by a ligament. A transverse ligamentous band, stretching horizontally between the neurapophyses, divides the cavity enclosed by the atlas into dorsal and ventral portions. The odontoid process of the axis fits into the latter. The Axis The postzygapophyses are better developed than the prezygapophyses. The odontoid process and anterior hypapophysis are more prominent than in Hemidactylus (Mahendra, 1950). Remaining Presacral Vertebrae The succeeding presacral vertebrae bear a pair of foramina close to the median line on the ventral surface of each centrum (Fig 4 B). Mahendra (1950) states that these openings, serving for the passage of blood vessels, are well developed in Geckonidae and Xantusiidae, and probably represent a primitive feature. Apparently such apertures in presacral vertebrae were recorded for the first time in any animal by Mookerjee and Das (1933), who in investigating Typhlops braminus

found a single opening in each vertebra as far back as the pelvic region. Mookerjee and Das indicate that each ventral aperture passes through the centrum and up to the neural canal. They state that branches of the vertebral artery, instead of passing through the intervertebral spaces along with the spinal nerves, enter the spinal canal through these openings. Ribs are not present on the atlas or axis, and not normally as in some geckos, on the third vertebra. The fourth, fifth and sixth vertebrae, however, have short bony ribs, somewhat expanded at both ends. Attached to their distal ends, these ribs bear delicate cartilaginous processes which, being bifurcated terminally, have a “fish-tail” appearance (Fig. 4 D) These processes increase in size from before backwards so that the last which are on the sixth vertebra are much longer than the rest. The first of the series, which is on the fourth cervical rib, can usually be seen anteriorly to the suprascapula. The others are covered by the latter. It is difficult in alizarin transparencies to see much muscular detail, but the first process on each side appears to have a muscular connexion with the suprascapula. In Naultinus, although the anterior two cartilaginous processes on each side are forked, the most posterior one is single, slender and elongated. Calcification of these structures may occur in all species. Mahendra (1935, 1950) describes similar structures in Hemidactylus occurring on the fourth, fifth, sixth and sometimes seventh cervical vertebrae as being presumably uncinate processes. He duly records them for the first time in a living Lacertilian. Their occurrence in geckos, however, is by no means exceptional, for we have subsequently noted them in alizarin transparencies of Jamaican and Australian genera. Furthermore, the term “uncinate” should not be used for these processes without reservation, and it must be pointed out that these geckos do not possess such structures on the sternal and post-sternal ribs where uncinate processes are normally found; that they would be the only living animals possessing uncinate processes on cervical ribs; that their cervical ribs are in two portions, proximal and distal. The processes do not arise from the proximal portion of the rib as is typically the case with uncinate ribs, but from the distal portion. The seventh and eighth vertebrae have long, slender, dorsal ribs, each jointed distally with a short, blunt, ventral piece of cartilage which does not reach the sternum. These ventral pieces of cartilage are indistinguishable from the sternal ribs, though they are much shorter. In general, it seems clear that the seventh and eighth vertebrae are in the same series as those posterior to them, rather than with the more anterior vertebrae of the column On this basis it is reasonable to follow Camp (1923) in relegating these two vertebrae to the dorsal series, thus regarding the cervical vertebrae as six in number. In Hoplodactylus, the ninth and tenth vertebrae are connected directly with the sternum by sternal ribs, which are jointed with the dorsal ribs. The sternal ribs of the eleventh and twelfth vertebrae are connected with the xiphisternum. while the sternal ribs of the thirteenth vertebra almost meet in the mid-ventral line, but do not quite connect with the xiphisternum Sometimes only the twelfth vertebra is connected with the xiphisternum. In Naultinus, the ninth and tenth vertebrae are connected to the sternum proper, while the eleventh, twelfth and thirteenth are connected with the xiphisternum. In general, the sternal ribs are better developed in Naultinus than in Hoplodactylus and they approach the mid-ventral line more closely and for a long distance posteriorly. The sternal ribs are always cartilaginous, but may be calcified, even heavily so in the case of Hoplodactylus duvauceli. In the fourteenth to twentieth vertebrae inclusively, the sternal ribs are free ventrally and diminish in size posteriorly. The twenty-first to twenty-fifth vertebrae have their dorsal ribs decreasing in size and have lost almost all trace of the sternal portions. The twenty-sixth vertebra, normally the last presacral vertebra, has no ribs.

Text-fig. 4.—Vertebrae of Naultinus elegans. A, Second and third cervical vertebrae, lateral view. B, Eighteenth presacral vertebra, ventral view. C, Eighteenth presacral vertebra, lateral view. D, Fourth vertebra, anterior view, buf, foramina for blood vessels; cn, centrum; cp, distal cartilaginous process of rib, cpf, capitular facet; cpt, capitulum of rib; cv, cervical rib; hyp, hypapophysis; ic, intercentrum; od, odontoid process; poz, postzygapophysis; prz, prezygapophysis; r, rib; tb, tuberculum of rib.

Noble (1921) was surprised to find in dorsal ribs of Sphaerodactylus a cartilaginous or fibro-cartilaginous band which extended from the angular portion of the head of the rib to the middle of the neural arch of the vertebra. He suggested that this structure, which was distinct from the capitular head of the rib was a last vestige of the tuberculum found in the double headed ribs of primitive reptiles but believed to be lost in Lacertilia. A similar band of tissue representing the tuberculum occurs in the New Zealand geckos. When sections of dorsal vertebrae and ribs of Hoplodactylus pacificus were examined, it was noted that the tuberculum was distinctly bony. (Fig. 4C.) Camp (1923) did not find parasternal ribs, (abdominal ribs or gastralia) in any of the geckonid genera at his disposal. On the other hand, Mahendra (1950) notes that alizarin preparations of Hemidactylus flaviviridis show the presence of a few delicate parasternal ribs in the superficial parts of the rectus muscle just behind the sternum. No traces of such gastralia have been observed in any of the New Zealand geckos. The sacrum is found normally from the twenty-seventh and twenty-eighth vertebrae, the “sacral processes” of which are widely expanded and connected with the ilia. A great deal of variation occurs in this region, and more posterior vertebrae may take part in the formation of the sacrum if one of the anterior processes fails to develop. In this connexion there may be asymmetry on right and left sides. The first vertebrae of the caudal series have transverse processes diminishing in size from before backwards, and may show traces of caudal ribs as short, blunt, much reduced structures. These caudal ribs are calcified and are particularly well seen in Naultinus, where some four or more occur on each side. El-Toubi and Khalil (1950), in a comparative study of the osteology of some Egyptian geckos, drew attention to the presence of caudal ribs, which as far as they were aware had not previously been recorded in living Lacertilia. These small caudal ribs, articulating with the transverse processes of the first four or so caudal vertebrae were observed by El-Toubi and Khalil on one side only in alizarin transparencies. In the New Zealand geckos they are clearly seen on both sides. It has been noted that the position of the sacrum in the New Zealand geckos is not fixed and that a caudal vertebra may function as a sacral vertebra. In such a case, both the caudal transverse process and its distal rib are expanded for articulation with the ilium. This variation is of interest considering the uncertainty that has existed for some time regarding the nature of sacral processes in lizards. Moodie (1907), investigating the sacrum of the Lacertilia, claimed from embryological evidence that “It can thus be very definitely stated that the Lacertilia occupy an isolated place among all other known reptiles in not having any sacral ribs whatever.” On the other hand, Kamel (1951), in studying the development of ribs in the sacral region of Chalcides ocellatus found that the sacral ribs are quite distinct. They chondrify and become united with the transverse processes. Typically the processes of normal sacral vertebrae in New Zealand geckos are much expanded, and each bears distally a wide, articulating flange of uncalcified cartilage. This fuses very often with the corresponding structure on the neighbouring sacral process. In the caudal region, Y-shaped chevron bones are attached to the ventral region of the intervertebral joints, beginning usually between the third and fourth or fourth and fifth caudal vertebrae. Autotomy occurs very commonly in the fifth caudal vertebra. The first four vertebrae of the series are usually pygals and lack the transverse unossified septum which passes through the body of the centrum of the succeeding caudal vertebrae. In the autotomous region of the tail the vertebrae become extremely elongated.

Pectoral Girdle (Fig. 5.) In the New Zealand genera the sternum is a more or less rhomboidal plate which in Naultinus is either only slightly calcified or not at all. It is not fenestrated. According to Camp (1923), sternal fenestrae, which are present in most families of lizards except the Anguimorpha, are probably sometimes of secondary development, perhaps connected with certain relationships of the pectoralis musculature. Two ribs are attached to each postero-lateral side of the sternum. One or more of the succeeding pairs of ribs, though not attached directly to the sternal plate, join two parallel rods forming the xiphisternum. The number of pairs of ribs doing so varies and is greatest in Naultinus. The xiphisternal rods run close to the middle line and join the sternum at its posterior end. The interclavicle has an irregular, sub-rhomboidal shape in Naultinus, but is distinctly sub-cruciform in Hoplodactylus, especially in H. pacificus. According to Camp (1923), with increasing reduction from the sub-rhomboidal and sub-cruciform, only a splint-like, longitudinal element remains in Paragonatodes, while a mere nodule occupies the position of the interclavicle in Uroplates. The clavicles in Naultinus and Hoplodactylus are slightly dilated mesially. Those of Naultinus are imperforate, while those of Hoplodactylus may have one or two fenestrae. Noble (1921) has shown in a series of geckos how the clavicle might have been gradually expanded and in the extreme stage thinned out until a foramen was formed. On the other hand, Camp (1923) expresses his belief that modern Sauria are derived from ancestors with broadly expanded, non-perforate clavicles, and that simple rounded clavicles have been shaped from these. The distal end of each clavicle articulates with a notch or depression on the antero-ventral region of the suprascapula, just as in Hemidactylus (Mehendra, 1950). This relationship to the suprascapula, which far from being confined to certain geckos occurs also in such forms as Iguana tuberculata (Goodrich, 1930), contrasts with the condition in Sphenodon and in many Lacertians where the clavicle abuts against the scapula which usually has a well-developed acromial process to receive it. Fenestration in the scapulo-coracoid region is not as extensive as in most geckos. Hemidactylus is figured (Mahendra, 1950) to have four fenestrae on each side apart from the supracoracoid foramen. These are presumably scapular, scapulo-coracoid, lateral coracoidal and median coracoidal fenestrae. In both Naultinus and Hoplodactylus, the coracoid and scapula are often indistinguishably fused together, but the line of fusion is sometimes clear in Naultinus This indicates that of the two fenestrae present, one is the scapulo-coracoid fenestra and the other a coracoidal fenestra. A fairly large supracoracoid foramen is present. The two epicoracoids are flat plates, convexly curved along their inner borders. They are well calcified in Hoplodactylus but uncalcified in Naultinus The epicoracoids in both genera overlap in the middle line, the left ventrally to the right, and they are able to do this because the sternum is more posteriorly placed in relation to them. This contrasts with Hemidactylus (Mahendra, 1950), where although the sternum is in similar position the epicoracoids do not meet in the mid-ventral line, and with such lizards as Calotes versicolor (Iyer, 1942) where the epicoracoids are kept far apart by the presence of a broad sternum between them. Pelvic Girdle (Fig. 6 A, B and C.) In Naultinus, calcification is less complete than in Hoplodactylus, and the pubis and ischium can often be seen as separate elements.

Text-fig. 5.—Pectoral girdle and sternum of Naultinus elegans, ventral view. (Only five pairs of ribs are shown posteriorly to the xiphisternum.) cf, coracoidal fenestra, cl, clavicle, co, coracoid; epc, epicoracoid; h, humerus, ic, interclavicle; sc, scapula; scf, scapulo-coracoid fenestra; sf, supracoracoid foramen; ssc, suprascapula; st, sternum; str, sternal rib; xst, xiphisternum.

The foramen cordiforme (ischio-pubic fenestra) is a large space bounded by the pubes and ischia. In a number of lizards such as Hemidactylus (Mahendra, 1950), Uromastyx (El-Toubi, 1949) and Calotes (Iyer, 1942), it is divided by a ligament into right and left halves, but in Naultinus and Hoplodactylus such a ligament is lacking. The epipubis is uncalcified in Naultinus but in the two species of Hoplodactylus examined it showed varying degrees of calcification. A metischial process (spina ischii) on each side is prominent in both genera. The prepubic (pectineal) processes are also well developed, especially in Naultinus. A ligament stretches from each prepubic process to the corresponding ischium, where it is fanned out to its attachment on to the ventral portion of this bone and on to the metischial process. A hypoischium is present in all cases and its form is simple and rod-like, tapering behind. It remains an uncalcified cartilaginous structure in Naultinus. In describing the os hypoischium, Camp (1923) lists the lizards in which it is absent. Among these he includes Hemidactylus, but Mahendra (1950) clearly shows that a small structure of this type is present. Considerable controversy has occurred concerning the derivation of the os hypoischium. Camp (1923) considers that the hypoischium is only an “epiphysial calcareous deposition in the ligamentum hypoischium” rather than that it is a reduced primitive element (Mehnert, 1891). Camp points out that it is absent as far as is known in the extinct amphibians and reptiles. He also states that it is absent in Sphenodon, but Goodrich (1930, p. 201) shows a cartilaginous hypoischiatic process quite clearly in a diagram of the pelvic girdle of Sphenodon. Furthermore, Camp's statement that it is not represented in most primitive geckos does not seem to be a true generalisation now. Fore and Hind Limbs (Fig 6 D and E.) In both fore and hind limbs, the arrangement of carpal and tarsal bones, and of phalanges, is similar in all the New Zealand species examined. In the fore limb, a bony patella ulnaris is present at the proximal end of the ulna. Nine bones are present in the carpus, arranged in three rows. The proximal row consists of a radiale, ulnare and a very small pisiform. The middle row is represented only by a centrale. The distal row consists of five bones, of which the first is smaller than the rest. An intermedium is absent. In the hind limb, two rows of tarsal bones occur, comprising four bones altogether. The proximal row consists of a large bone expanded across the width of the tarsus. Though often described as a tibio-fibulare, it is probably composed of a fused fibulare and intermedium, as El-Toubi (1949) in his discussion on the Lacertilian tarsus and review of the literature suggests. The distal row consists of three bones. As in Hemidactylus (Mahendra, 1950), the tarsal articulating with the first metatarsal is merely a vestige or nodule. A larger bone articulates with the third metatarsal, while the cuboid or fourth tarsal, which is the third distal tarsal present, is still larger. Of the metatarsals, the fifth is shorter than the others and is irregular and massive. It has the characteristic hook-shape typical of the Squamata. The phalangeal formula of the New Zealand geckos is 2, 3, 4, 5, 3 for the fore limb, and 2, 3, 4, 5, 4 for the hind limb. In both cases this represents the primitive reptilian formula (Romer, 1949). Mahendra (1950) describes the phalangeal formula of Hemidactylus as being 2, 3, 3, 4, 3 for both hand and foot. This represents a loss of one phalanx in each of the third, fourth and fifth toes as he correctly states, but the phalangeal formula of the manus is not primitive as is claimed in the summary of his paper.

Text-fig. 6.—A-C—Pelvic girdle, vential view, ilium not shown. A, Hoplodactylus duvaucelii. B, Naultinus elegans. C, Hoplodactylus pacificus. D, Carpus of Hoplodactylus duvauces. E, Tarsus of Hoplodactylus duvaucelu ac, acetabulum. c, 1–5, distal carpals 1–5; ce, centrale; cu, cuboid; ep, epipubis; fc, foramen cordiforme (ischio-pubic fenestra); fi, fibula; hi, hypoischium; is, ischium; mt, I-V, metacarpals I-V; mp, metischial process: mt I-V, metatarsals I-V; of, obturator foramen; p, pubis; pa, patella ulnaris; pi, pisiform; pp, prepubic (pectineal) process: r, radiale: rd, radius; tf, tibiofibulare; ti, tibia, t 1–3, distal tarsals. u, ulnare; ul, ulna.

Cloacal Bones. (Post-anal Bones.) Noble (1921) notes the occurrence of cloacal bones in several geckonid genera. He comments on the fact that very little reference appears in the literature in regard to cloacal bones and suggests that, lying free near the hemipenes and below the skin, they have often been overlooked. Malcolm Smith (1935) states that post-anal bones and sacs are peculiar to the Geckonidae. The sacs are present in both sexes, but the bones only in the males. In those species in which the sacs are absent the bones are also absent. The cloacal or post-anal bones themselves are readily observed in alizarin transparencies of males of both Naultinus and Hoplodactylus and resemble those figured by Noble (1921) for Coleonyx variegatus and Paragonatodes dickersoni. They are paired structures and each is a short, strongly curved, rod-like element, slightly rounded at its inner end. Discussion In his classification of lizards, Camp (1923) attempts to evaluate palacotelic characters, that is, characters reliable as an indication of relative primitiveness. He emphasises that a species exhibiting a few characters of high antiquity may be considered more ancient than one having many characters of lesser palaeotelic value (comparative rank), and takes the important step of assigning comparative rank to some 34 selected characters. This is necessitated by the fact that high specialisation in a group may obscure an archaic position that can only be traced by recognition of satisfactorily primitive characters. Camp's list for lizards generally is accordingly headed by such features as (a) three complete branchial arches; (b) vertebrae amphicoelous; and (c) two complete skull arches. Underwood (1954) draws extensively on the resources of his studies on reptilian eye structure to propound a classification of geckos. His proposed classification implies that the amphicoelous vertebral condition is secondary to the procoelous condition, and this implication he initially accepted. Underwood has recently indicated that he is withdrawing this view because of palaeontological evidence. A considerable radiation of definitive lizards occurred in the Triassic, and all of these are amphicoelous. The rise of definite lizards long antedated the fixation of the procoelous condition. The position of any group of lizards, or of members within a group, can only be ascertained after a careful survey of all the available data. Where there is a clash as in this instance between eye and vertebral structure in geckos, then Camp's procedure of assigning comparative value to these characters must surely be adopted. A similar situation arises when considering the morphological status of the New Zealand geckos. On the one hand they are osteologically primitive, and on the other they are the only geckos in the world known to be ovoviviparous. H. Claire Weekes (1935), in reviewing placentation among Australian scincs and snakes, suggests that cold is the most likely external factor associated with high altitudes that may influence either directly or indirectly the adoption of ovoviviparity. In referring to the ovoviviparity of the New Zealand geckos she observes that the New Zealand climate, approaching that of the British Isles, is the coldest in which geckos occur. Weekes concludes from her study that placentation among reptiles has arisen independently many times in the course of evolution, and that the phenomenon of parallel development of similar types of placentas is common. Mary M. M. Boyd (1942), commenting on Weekes' findings, concludes that since oviparous and ovoviviparous species occur in the same genus, placentation must have arisen fairly recently, after the divergence of the species. She states that the simple type of placenta in Hoplodactylus is according to expectation “since viviparity must have been assumed after the divergence of the species”.

Under such circumstances, although the New Zealand genera may be regarded as unique among geckos in their ovoviviparity, the simple type of placenta as evidenced by the one genus investigated can hardly be regarded as a major or fundamental specialisation off-setting their osteological primitiveness amongst the Gekkota. Morphological data of the various genera of geckos are extremely incomplete, but a comparison of Naultinus and Hoplodactylus with their known forms indicates that at least osteologically the New Zealand genera are the most primitive known members of the family. Even on Camp's (1923) criteria, the condition of the hyobranchial skeleton coupled with their amphicoelous vertebrae would indicate this, for Coleonyx, which is the only known geckonid with a comparably primitive hyobranchial skeleton, is well removed because it has procoelous vertebrae. The osteological features of the New Zealand geckos are perhaps not surprising for primitiveness is increasingly being shown in various representatives of the New Zealand fauna. Sphenodon and the primitive frog Leiopelma are further examples of this feature from among the lower tetrapods. The general characters which indicate the status of the New Zealand geckos are included in the summary below. Summary 1. The cranial and post-cranial osteology of the New Zealand geckonid genera, Naultinus and Hoplodactylus, is studied from alizarin transparencies, serial sections and X-ray photographs. 2. In both genera, there is a remarkable similarity in structure throughout, except that in general Naultinus is neotenic when compared with Hoplodactylus. 3. Median skull elements are entirely separate except for the frontals and premaxillae, which tend to be more fused in Hoplodactylus than in Naultinus. 4. The postfrontal, supratemporal (or tabular) and jugal are present, but the postorbital and quadratojugal are lacking. A ligament represents a remnant of the postorbital bar. 5. Naultinus has twenty scleral ossicles while Hoplodactylus has twenty-five. 6. The hyobranchial skeleton is of a most primitive type with three arches, two of which are attached to the skull. 7. Normally twenty-six amphicoelous, presacral vertebrae occur. Occasionally there are twenty-seven. 8. The ribs of cervical vertebrae bear processes similar to, but not strictly homologous with, uncinate processes. 9. Ribs in their attachment to vertebrae have a tuberculum distinct from the capitulum. 10. Caudal ribs are present. 11. Clavicles are dilated, imperforate in Naultinus but perforate in Hoplodactylus. 12. The pelvic girdle is not divided into right and left halves by a ligament. A hypo ischium is present. 13. The phalangeal formulae for the fore and hind limbs respectively are the primitive reptilian formulae, 2, 3, 4, 5, 3 and 2, 3, 4, 5, 4. References Boulenger, G. A., 1885–7. Catalogue of the Lizards in the British Museum (Natural History). London. —— 1912. Vertebrate Fauna of the Malay Peninsula Reptiles and Batrachians. London. Boyd, Mary M. M., 1942. The Oviduct, Foetal Membranes, and Placentation in Hoplodactylus maculatus Gray Proc. Zool. Soc. Lond. 112, 65. Brock, G. Truda, 1932. Some developmental stages in the skulls of the geckos, Lygodactylus capensis and Pachydactylus maculosa, and their bearing on certain important problems in lacertilian craniology. S. Af. Sci. 29, 508.

Broom, R., 1925. On the Origin of Lizards. Proc. Zool. Soc. Lond., 1925, I. Camp, C. L., 1923. Classification of the Lizards Bull. Amer. Mus. Nat. Hist. 48, 289. Cope, E. D., 1892. The Osteology of the Lacertilia. Proc. Amer. Philos. Soc. 30, 185. El-Toubi, M. R., 1949. The Post-cranial Osteology of the Lizard, Uromastyx aegyptia Forskal. J. Morph. 84, 281. —— and Khaul, A., 1950. Caudal Ribs in Geckos. Nature, Lond. 166, 1120. Goodrich, E. S., 1930. Studies on the Structure and Development of Vertebrates London. Iyer, M. N. N., 1942. The Habits, External Features and Skeletal System of Calotes versicolor, Part I J. Mys. Univ. B 3, 153. Kamel, A. H., 1951. Sacral Ribs of Lacertilia. Nature, Lond. 168, 660. Lakjer, T, 1927. Studien über die Gaumenregion bei Sauriern im Vergleich mit Anamniern und primitiven Sauropsiden. Zool. Jb. 39, 57. McCann, C., 1955. The Lizards of New Zealand. Geckonidae and Scincidae. Dom. Mus. Bull. 17, I. McDowell, S. B. & Bogert, C. M., 1954. The Systematic Position of Lanthanotus and the Affinities of the Anguomorph Lizards Bull. Amer. Mus. Nat. Hist. 105, I. Mahendra, B. C., 1935. The Presence of Uncinate Processes on the Ribs of a Lacertilian. Curr. Sci. 4, 37. —— 1949. The Skull of the Indian House-gecko, Hemidactylus flaviviridis Ruppel. Proc. Zool. Soc. Bengal. 2, 29. —— 1950. The Osteology of the Indian House-gecko, Hemidactylus flaviviridis Ruppel Proc. Zool. Soc. Bengal 3, 49. Mehnert, E., 1891. Untersuchungen uber die Entwicklung des Os hypoischium (Os cloacae aut.) Os epipubis und Ligamentum medianum pelvis bei den Eidechsen. Morph. jb. 17, 123. Moodie, R. L., 1907. The sacrum of the Lacertilia. Biol. Bull. Wood's Hole 13, 84. Mookerjee, H. K. & Das, G. M., 1933. On the Peculiar Apertures in the Vertebral Centa of Typhlops braminus. Proc. Zool. Soc. Lond. 1933, 283. Noble, G. K., 1921. The Bony Structure and Phyletic Relation of Sphaerodactylus and Allied Lacertilian Genera, with the Description of a New Genus. Amer. Mus. Novit. 4, I. Romer, A. S., 1949. The Vertebrate Body. Philadelphia. Shute, C. C. D. & Bellairs, A. d'A., 1953. The Cochlear Apparatus of Geckonidae and Pygopodidae and Its Bearing on the Affinities of These Groups of Lizards. Proc. Zool. Soc. Lond. 123, 695. Smith, Malcolm A., 1933. Remarks on Some Old World Geckoes. Rec Ind. Mus 35, 9. —— 1935. The Fauna of British India. Reptilia and Amphibia. Vol. II—Sauria. London. Underwood, G., 1951. Reptilian Retinas. Nature, Lond 167, 183. —— 1954. On the Classification and Evolution of Geckos. Proc. Zool. Soc. London. 124, 469. Versluys, J., 1936. Kranium und Visceralskelett der Sauropsiden. I. Reptilien. Bolk's Handb vergl. Anat. Wirbeltiere, 4, 699. Walls, G. L., 1942. The Vertebrate Eye and Its Adaptive Radiation. Michigan: Cranbrook Inst. Sci. Watson, D. M. S., 1914. Pleurosaurus and the Homologies of the Bones of the Temporal Region of the Lizard's Skull Ann. Mag. Nat. Hist. 14, 84. Weekes, H. Claire, 1935. A Review of Placentation Among Reptiles with Particular Regard to the Function and Evolution of the Placenta Proc. Zool. Soc. Lond. 1935, 625. Williston, S. W., 1914. The Osteology of Some American Permian Vertebrates. Contrib. Walker Mus. Univ. Chicago I, 107. Dr. N. G. Stephenson, Department of Zoology, University of Sydney, Sydney, N.S.W. Dr Elsie M. Stephenson, Department of Biological Sciences, N.S W. University of Technology, Sydney, N.S.W.