The Head of Xenopus laevls.

Transcription

1 The Head of Xenopus laevls. By Nellie F. Paterson, D.Se., Ph.D., Department of Zoology, University of the Witwatersrand, Johannesburg. With Plates 9 to 16. CONTENTS. P A G E INTRODUCTION 161 LATERAL LINE SENSORY ORGANS 163 MUSCULATURE 165 BLOOD-VESSELS THE CHONDROCRANIUM Metamorphosis Olfactory Eegion Nasal Cavities Auditory Eegion THE HYOBRANCHIAL SKELETON THE CRANIAL NERVES 198 Ganglion Pro-oticum Nervus Trigeminus Ramus Mandibularis Ramus Ophthalmicus Profundus 203 NERVUS FACIALIS Truncus Supra-orbitalis Ramus Hyomandibularis Ramus Palatinus NERVI GLOSSOFHARYNGEtTS AND VAGUS Nervus Glossopharyngeus Nervus Vagus 220 SUMMARY OF COMPOSITION AND DISTRIBUTION OF NERVES SUMMARY 227 REFERENCES 228 INTRODUCTION. THE Aglossa, comprising only the genera Xenopus, Pipa, Propipa, Hymenochirus, and Pseudohymenochirus, are characterized among other things by the absence of a tongue and by a pectoral girdle that exhibits considerable deviation from that of typical Anura Phaneroglossa. NO. 322 M

2 162 NELLIE F. PATBESON The Aglossa are usually classified as the lowest of the A n u r a, but as Gadow in his account of the Amphibia in the ' Cambridge Natural History' (1909) indicates, their characteristic features are not necessarily primitive ones. A tongue is lacking in the majority of truly aquatic forms, and in the Aglossa the shoulder girdle and other parts of the body are doubtless specialized in response to their particular habits. It is therefore not surprising to find that the Aglossa present some striking morphological similarities with the aquatic Urodela on the one hand, and with certain genera of the Phaneroglossa on the other, but it is very doubtful if these resemblances are of any consequence. Probably they are merely accidental or superficial likenesses, as for instance the similarities of skeletal structures indicated in the course of the present study as existing between Xenopus laevis and the Phaneroglossangenera Hemisus, Breviceps, and Probreviceps. The resemblances between Xenopus and such Urodela as Siren and Proteus, especially in regard to the arrangement of the nervous system, seem to be more significant, but even this may be due to adaptation to somewhat similar habitats. The anatomical peculiarities of X. laevis Daud., the ' clawed toad' or ' platanna' of Southern Africa, have for many years been favourite subjects of investigation by students of morphology, but despite the valuable contributions that have appeared, there still remain many problematic points. X. laevis abounds in ponds and dams throughout southern and tropical Africa. Eose (1929) in his book 'Veld and Vlei' introduces the reader to a discussion on Xenopus by stating that 'the plathander has an even stronger claim than the springbok to be considered a typically South African animal, for whereas horned antelopes are found in many other countries, clawed toads are found in southern and tropical Africa and in no other part of the world'. The larvae have been described and figured by Bles (1904) and Peter (1930), and are chiefly remarkable for the extreme transparency of the tissues. They may also be recognized by the presence of a pair of long trailing' tentacles' (tent,fig.30 a, PL 16), which disappear at metamorphosis, and by the characteristic

3 HEAD OF XENOPUS 163 attitude they take up in the water; they usually remain more or less stationary with the head directed downwards and with the thin tapering extremity of the tail constantly vibrating and thus enabling them to maintain this position. The present study of X. laevis was begun with the intention of investigating the distribution of the cranial nerves, but as it was necessary to make a somewhat intensive study of other associated cephalic structures and, as certain of these observations have proved of interest, they have been included in the dissertation. For the purpose of this study a series of larvae in progressive stages of development and measuring from 5 mm. to 60 mm. long were sectioned after having been fixed in aqueous Bouin. In order that comparisons between larval and adult conditions might be made possible, and conclusions might be arrived at regarding the changes occurring during the process of metamorphosis, several transforming specimens as well as some very young frogs were also examined. The latter were sectioned after having been decalcified for several days in 3 per cent, nitric acid as prescribed in the eighth edition of Bolles Lee's 'Vade Meeum', 1921, pp The thickness of all the sections was 10/*, and the stains used were Hansen's haematoxylin and eosin. The preparations were found to be very satisfactory for a general study of the various organs and systems in the head, including the cranial nerves. LATERAL LINE SENSORY ORGANS. As Xenopus is one of those interesting Anura in which the lateral line sensory organs persist throughout life, due no doubt to their aquatic existence, a comparison of their arrangement in larval and adult stages is of some interest. Escher (1925), who gives a diagram of the organs in the adult, failed to observe them in the larva. It is to be admitted, however, that, owing to the light sprinkling of somewhat star-shaped patches of pigmentation, and more particularly to the remarkable transparency of the head region, the larval sensory organs are rather obscured and are only determined with difficulty. The ventral organs are easily distinguished as white markings

4 164 NELLIE F. PATERSON on the darkly coloured abdominal region in all larval stages. Even in young tadpoles measuring about 15 mm. in length and in which the hind-limbs have not yet been protruded, the organs of the dorsal and median rows (osl., msl, fig. 30 a, PI. 16) are sharply contrasted with the darker body coloration between the fore- and hind-limb buds. On the dorsal and ventral surfaces of the head in front of this region the organs are indistinguishable macroscopically from the general surface of the body. That they are developed, however, is evident from sections of the anterior part of the body. Certain organs (csl.,fig.30 a, PI. 16) are also obvious in these early stages on the ventral fin anterior to the cloaca, where they at first appear as a lateral row of small rounded dots which show a gradual elongation during subsequent stages of development. In older larvae and in those undergoing metamorphosis these cloacal sensory organs decrease to a few vertically elongate organs on each side of the fin fold, and with the gradual absorption of the tail they come to lie even nearer to the cloacal aperture. Thus in the mature frog they are only observed with difficulty as indistinct ridges on the ventral fold of skin immediately anterior to the cloacal aperture. All the sensory organs become more apparent just before metamorphosis, when the skin is more deeply and more uniformly pigmented and the whitish organs show up by contrast. It is then evident that the arrangement in the larva is essentially similar to that of the adult, and that both conform in general to the plan drawn up by Escher (1925) for U r o d e 1 e Amphibians. The dorsal, median, and ventral rows of trunk sensory organs are strongly developed, the two former systems extending on to the tail (fig. 30 b, PI. 16). The dorsal row may be continued only for a short distance caudally in some specimens, but the median organs are apparent throughout the length of the tail at the dorsal ends of the myomeres. The cephalic organs are arranged into supra-, infra-, and postorbital series above, and oral and gular rows below. In addition, as Escher (1925) has indicated in the adult Xenopus, there are hyomandibular organs (hso., fig. 30 b, PI. 16) which are laterally continuous with the median line on the body, an

5 HEAD OF XENOPUS 165 accessory row (ace, fig. 30 b, PL 16) which arches dorsally, and several small round parietal organs (par.,fig.30 b, PI. 16) situated dorsally midway between the supra-orbitals. The chief difference between larval and adult arrangements is found in the orbital grouping. In the larva the supra-orbitals (sorb.,fig.30 b, PI. 16) and post-orbitals (porb.,fig.30 b, PL 16) are quite distinct, but during metamorphosis they gradually approach the eye, so that in the adult they are arranged into a single circumorbital series. Closer inspection reveals that there is some considerable variation in the arrangement and number of organs in the several rows, not only in different specimens but also in the two sides of an individual. In general, however, each series of sensory organs is as a whole sufficiently constant to allow of their relative positions being easily determined. MtTSCULATUBE. Edgeworth (1929, 1935) has reported upon the cephalic muscles of the larva of X. fraseri in the course of his wellknown dissertations on the muscles of Vertebrates, and the entire musculature of the adult X. laevis has been studied in detail by Grobbelaar (1924). The present series of preparations of both larval and adult X. laevis have made it possible to draw conclusions regarding the changes undergone by the musculature of the head during metamorphosis. The muscles associated with the chondrocranium up to the time of metamorphosis are essentially similar to those of X. fraseri. The levatores mandibulae anterior and posterior are two rather broad muscles arising dorsally on the palatoquadrate and having their insertions on the dorsal surface of the Meckel's cartilage. The levator mandibulae posterior (lev.post,figs.7, 8, PL 11) lies ventrally to the levator mandibulae anterior (lev.ant., figs. 6-8, PL 11), than which it is also much shorter, as it only extends from the posterior part of the processus muscularis of the palatoquadrate, whereas the levator mandibulae anterior originates farther back on the subocular bar of the palatoquadrate. As in X. fraseri these two muscles maintain their relative positions throughout, and there is no reversal of

6 166 NELLIE F. PATEBSON positions such as appears in larval Anura P h a n e r o g l o s s a (Edgeworth, 1924,1935). In addition to these two masticatory muscles, the larval X. laevis has a third horizontally placed muscle (lev.tent, figs. 6-8, PI. 11), which seems to be a portion of the levatormandibulae anterior, as it arises just laterally to the latter muscle on the lateral margin of the subocular bar of the palatoquadrate. It soon separates from the levator mandibulae muscle, and as it passes forwards to the base of the tentacle it appears dorsolaterally above the mandibular muscle in the deep depression between the pronounced processus muscularis (mus.fr., figs. 7, 8, PI. 11) and the commissura quadrato-cranialis anterior. The ventral muscles of the larval head are two broad transverse sheets, the musculi intermandibularis and interhyoideus (int.hy.,figs.6, 7, PI. 11), lying between the Meckel's cartilages and the cerato-hyalia respectively and each joining its fellow mid-ventrally in an obvious median raphe. The levator hyoideus is a strongly developed vertically situated muscle, divisible into two somewhat fan-shaped fasciculi. The outer, larger fasciculus (lev.hy., figs. 7, 8, PI. 11), termed the orbito-hyoideus by Edgeworth (1935), arises in X. 1 a e v i s on the processus muscularis of the palatoquadrate and passes to a distinct ventro-lateral process of the cerato-hyale. This fasciculus partly overlies the second, rather smaller one (dep.man., figs. 6-8, PI. 11), the quadrato-hyoangularis of X. fraseri (Edgeworth, 1935). In X. laevis it originates dorsolaterally on the cerato-hyale anterior to its articulation with the palatoquadrate and spreads out fanwise immediately below the processus muscularis. Edgeworth (1935) finds that in X. fraseri it also originates on the quadrate, but in X. laevis no distinct evidence of this could be observed. Its fibres, however, lie close to the outer angle of the quadrate just below the processus muscularis but seem to have no connexion with it. The muscle passes forwards in front of the insertion of the orbito-hyoideus, below the suspensorial region, and is inserted ventro-laterally on the Meckel's cartilage, to which it acts as a depressor muscle. As in X. fraseri and Pip a, the muscles connected with

7 HEAD OF XENOPUS 167 the branchial skeleton are simpler in their arrangement than are those of the Anura Phaneroglossa. The dorsal muscles consist only of the constrictores branchiales, the levatores arcuum branchialium being unrepresented. The constrictores arise on parts of the chondrocranium and pass to the ventro-lateral region of the cerato-branchialia. In both Pip a and X. fraseri Edgeworth (1935) has recorded four constrictor muscles, those in Pip a all arising on the auditory capsule. In X. fraseri the first originates on the ascending and ventro-lateral processes of the quadrate and the remaining three on the auditory capsule. In X. laevis the constrictores branchiales form an almost continuous lateral sheet of muscle, in which the limits of the four separate muscles can only be made out with difficulty. The first arises on the subocular bar of the palatoquadrate; the second originates on the ventrolateral process of the palatoquadrate; while the third, which is the broadest, is located on the crista parotica and commissura brancbio-cranialis. The fourth is a small muscle situated at the posterior end of the branchial skeleton. The hypobranchial muscles consist simply of one broad transversus ventralis springing from the second branchial arch, meeting its fellow in a mid-ventral raphe, and of four separate subarcuales recti. The latter pass ventrally between the branchial arches, the anterior one running from the first arch to the cerato-hyale, as in larvae of Anura Phaneroglossa (Edgeworth, 1919,1935). In addition to these there is another muscle which does not seem to have been previously recorded. It lies on the inner side of the last branchial arch and forms a broad longitudinal muscle extending backwards beneath the pharynx. At the junction of the latter with the oesophagus the fibres of the two sides become confluent below the oesophagus just anteriorly to the constrictor oesophagi. The muscle continues posteriorly, attached ventrally to the branchial bar and with its dorsal fibres lying just above the dilatator laryngis. Subsequently it is attached to the posterior angle of the chondrocranium close to the origin of the dilatator laryngis. It is also continued obliquely over the posterior part of the branchial skeleton on to the commissura branchialis, immediately above

8 168 NELLIE F. PATEKSON the constrictor branchialis IV, with which it appears to fuse. Edgeworth (1935) states that in Eana the transversus ventralis IV is incompletely developed and that in Xenopus it is absent, probably having atrophied at an early stage. This longitudinal muscle is somewhat suggestive of the transversus ventralis IV of some Pisces and other Amphibia, although it is more dorsal in position and does not meet in a median raphe. Its relation with the constrictor oesophagi seems to indicate the possible origin of the latter from this muscle, and furthermore, the posterior part arching dorsally over the branchial skeleton, may represent a levator arcuum branchialium which has joined up with the constrictor branchialis IV. Metamorphosis of the Anura Aglossa is not accompanied by any such marked changes in the musculature as may occur in other Amphibia. The arrangement of the cephalic muscles in young forms of X. laevis is, therefore, comparatively simple and directly comparable with that of the larva. The main differences between the two conditions are found in the changed positions of the muscle insertions consequent upon the backward migration of the suspensorial region which takes place at metamorphosis. The levatores mandibulae anterior and posterior of the larva are two very strong muscles in the young frog, passing from the fronto-parietale round the lateral wall of the auditory capsule to the lower jaw (lev. man.,figs.18 a, b, PI. 12). They are respectively the musculi pterygoideus and temporalis described in the adult Xenopus by Grobbelaar (1924). The tentacular muscle of the larva has apparently no separate counterpart in the adult. With the complete disappearance of the tentacle at metamorphosis, the muscle merges with the levator mandibulae anterior of the adult. Both fasciculi of the larval musculus levator hyoideus are converted into the strong depressor mandibulae of the frog. In transforming specimens the arrangement is intermediate between that of the larva and adult. The quadrato-hyoangularis, instead of being partly overlapped by the orbitohyoideus, lies anterior to it and has now assumed an entirely different position. It is more vertical than in the larva and,

9 HEAD OF XBNOPUS 169 with the disappearance of the processas muscularis, it moves on to the quadrate, from which it stretches to the end of the mandible, being inserted on its ventro-lateral margin. The fibres of the larger posterior fasciculus still ran to an insertion on the cerato-hyale as in the larva. In the young frog the two fasciculi are still as distinct as they are in the larva. The anterior one, in addition to its connexion with the quadrate, also originates on the annulus tympanicus. The orbito-hyoideus of the larva has, however, now lost its connexion with the cerato-hyale, and stretches between the posterior extremities of the quadrate and mandible. It is continued backwards and upwards over the auditory region to the dorsal fascia, and lies (m.dep.man.,figs.19 a, b, PI. 13) externally to the museuli petrohyoideus and cueullaris (m.pt.hy. and m.cu.,figs.19 a, b, PI. 13). The petrohyoideus extends from the hinder part of the crista parotica to the ventral surface of the crieoid cartilage. The cueullaris arises close to the petrohyoideus, and in Eana (Edgeworth, 1935) and adult Xenopus (Grobbelaar, 1924) it passes backwards to be inserted on the inner surface of the scapula. In the present preparations of young frogs, however, the musculus cueullaris passes over the inner surface of the scapula on to the clavicle near the union of the parts at the glenoid cavity. Neither of these muscles is apparent in the larva of Xenopus and, due to the fact that they appear suddenly during metamorphosis, their evolution from the branchial muscles is difficult of determination. Even in transforming specimens, in which the atrophying branchial arches are still evident as a compressed lateral process (bra.,fig.21 a, PL 13), both muscles are fully developed. At such a stage the remains of the last two constrictores branchiales may still be seen passing between the auditory capsule and the hyobranchial skeleton, just anteriorly to the museuli petrohyoideus and cueullaris. The cueullaris certainly seems to be separated from the dorsal end of the last constrictor branchialis, but the development of the petrohyoideus is not so clear. During metamorphosis the larval muscle interpreted as a fused transversus ventralis IV and levator arcuum branchialium IV atrophies to an irregular

10 170 NELLIE F. PATERSON bundle lying on the inner side of the diminishing branchial skeleton. The petrohyoideus lies more dorsally over this region, and posteriorly runs into this reduced muscle. The whole course of the petrohyoideus at this time is reminiscent of the posterior petrohyoideus of other Anura, and in Xenopus it seems as though it might be a modification of a muscle such as the last levator arcuum branchialium, or even in part of a transversus ventralis IV, rather than simply of the posterior constrictor branchialis which is still visible during metamorphosis. It is therefore assumed that as in both Urodela and the Anura Phaneroglossa, where the cucullaris and petrohyoideus are separated from the posterior branchial muscles (Edgeworth, 1935), the cucullaris in Anura Aglossa has also a similar origin. There is also some evidence that the petrohyoideus may be derived from a muscle that in part resembles the levator arcuum branchialium IV, in which case the single petrohyoideus of Xenopus would be comparable with the posterior petrohyoideus muscle of certain other Amphibia. The larval intermandibularis and interhyoideus muscles are not subjected to any marked alterations at metamorphosis. The former is divisible into anterior and posterior parts, corresponding to the submentalis and submaxillaris respectively of Grobbelaar's description of the adult musculature (1924). The interhyoideus is rather narrower than in the larva, and is referred to as the subhyoideus by Grobbelaar. There is no marked change in the laryngeal muscles at metamorphosis. They are well developed even in young specimens measuring 10 mm. and in which the arytenoid cartilages are just becoming demarcated in the connective tissue at the side of the glottis. The constrictor (con.,fig.20 a, PI. 13) and dilatator laryngis (dil.,figs.20 a, b, PI. 13) are both simple paired muscles inserted on the arytenoid cartilages. The insertion of the dilatator occurs above the constrictor laryngis, but as is demonstrated in figs. 20 a, b, PI. 13, its insertion is located slightly posterior to and not in front of the constrictor as is usual in Anura. While the constrictor is a short, more or less obliquely vertical muscle, which takes its origin on the ventral side of the arytenoid cartilage, with the fibres of the two sides

11 HEAD OF XENOPUS 171 meeting mid-ventrally in some larvae, the dilatator laryngis is much longer and passes forwards from the postero-lateral angle of the chondrocranium. Metamorphosis leaves the constrictor laryngis unchanged, but the dilatator becomes detached from the chondrocranium and migrates on to the postero-medial process of the hyobranchial skeleton {ppm.,figs.20, 21, PL 18). In the young frog it therefore originates from the posterior end of the posteromedial process near the ligamentum hyo-cricoideum. A remarkable and interesting feature is the presence of a pair of muscles which are considered to be homologous with the laryngei dorsalis of Urodele Amphibia. Eidewood (1897) was of the opinion that in Xenopus there is an undifferentiated constrictor, to which he referred as the compressor glottides. In the present preparations of both larval and adult specimens of X. laevis there is a distinct separation of muscles serving to constrict the larynx. The dorsal pair of muscles arises on the arytenoid cartilages immediately above the insertion of the more ventral constrictor laryngis, but as is indicated in figs. 20 a, i, PL 13, they are anterior to the insertion of the dilatator. They arch dorsally over the larynx, meeting in a mid-dorsal raphe. While the arrangement of the laryngeal muscles seems to present some considerable variation in Urodele Amphibia, the Anura appear to be characterized by the general occurrence and constant arrangement of the constrictor and dilatator muscles. Laryngei dorsales have been observed in various Urodela, but it is apparent that they are by no means constant, for Edgeworth (1935) mentions several genera in which they are lacking. The dorsal constrictors observed in X. laevis are strikingly similar to the laryngei dorsales of the Urodela, and their presence may account for the more posterior insertion of the dilatator muscle in this species. These muscles have not been recorded so far in any other Anuran, and it is difficult to account for their presence in Xenopus. They may indicate a more primitive condition, or they may merely represent a more obvious differentiation of the constrictor muscle into dorsal and ventral parts.

12 172 NELLIE F. PATEESON The throat muscles require but little mention for they have been fully investigated by Eidewood (1897). The geniohyoideus consists of internal and external or lateral portions, inserted respectively on the postero-medial and postero-lateral processes of the hyobranchial skeleton. There is no separation, as in the Anura Phaneroglossa, of the lateral portion into internal and external muscles. Eidewood (1897) found that the geniohyoideus of adult Xenopus arose near the mandibular symphysis, the internus being anterior to the externus; but in the present young specimens they seem to originate together at the symphysis. The hyoglossal (hyog.,figs.20 a, b, PI. 13) was not observed to consist of the three portions described by Eidewood (1897). In these young frogs the condition is essentially as in other Anura; it arises from the anterior end of the postero-medial process, near the crieoid; it meets its fellow of the opposite side, and the two run together through the hyoglossal foramen to the floor of the mouth. BLOOD-VESSELS. Only two pairs of blood-vessels are relevant to the study of the chondrocranium and the cranial nerves. These are the bilaterally symmetrical internal carotid artery and the head vein (de Beer, 1926) or vena jugularis s. cardinalis anterior (Gaupp, vol. 2, 1899). The internal carotid artery arises far back in the tadpole, and in coursing forwards it comes to lie laterally to the auditory capsule in a deep groove on the lower side of the crista parotica. In this position it is immediately ventral to the head vein, the two vessels pursuing a more or less parallel course throughout. The head vein has to travel some considerable distance beyond the level of the sinus venosus before joining the ductus Cuvieri, as the latter, owing to its unusual length, instead of being transversely directed, lies obliquely in a postero-anterior direction, and meets the sinus venosus at an angle. In front of the auditory region (fig. 15, PL 12; fig. 27 6, PL 15) the two blood-vessels remain in juxtaposition for a short distance until the carotid artery moves mesially to supply the

13 HEAD OF XENOPUS 173 brain. The main stem of the internal carotid courses forward beneath the chondrocranium, where it is located mesially to the head vein (fig. 11, PL 11; fig. 13, PL 12; fig. 276, PL 15). It supplies the palate, and when passing round the processas muscolaris of the palatoquadrate, it gives off a branch to the levator hyoideus muscle. More anteriorly the artery lies close to the ramus mandibularis V (figs. 7,8, PL 11) and divides to supply the tissues of the snout. The arterial supply to the brain enters the cranium through a definite ventro-lateral carotid foramen, and immediately divides into two (fig. 12, PL 12;figs.27 b, d, PL 15). The inner and rather more posterior branch is the arteria carotis cerebrals, while the outer and stouter branch is the arteria ophthalmica. The arteria carotis cerebralis itself divides into anterior and posterior branches, as may be seen in fig. 27 d, PL 15. The posterior artery (pea., figs , PL 12), the R. posterior of Gaupp (1899) and Francis (1934), gives off a vessel to the lateral wall of the infundibulum, and a more posterior one anastomoses with its fellow directly above the hypophysis. As in Rana (Gaupp, 1899) and Salamandra (Francis, 1934), shortly after this anastomosis the pair of posterior cerebral arteries unite below the medulla oblongata at the level of the cerebellum to form a median artery (ba., fig. 17, PL 12), the arteria basilaris (Gaupp, 1899), which extends along the ventral fissure of the medulla oblongata. The anterior cerebral artery (aca., fig. 11, PL 11; fig. 12, PL 12) does not divide into dorsal and ventral parts as in the Salamander (Francis, 1934), but proceeds forwards over the optic nerve, gradually approaching its fellow (aca.,figs.9, 10, PL 11), with which it unites mid-ventrally at the anterior end of the telencephalon. On its forward course it gives off blood-vessels to the brain; one branch in particular passes latero-dorsally around the anterior end of the dieneephalon and apparently corresponds with the dorsal branch of the cerebral artery of the Salamander. The arteria ophthalmica (opha., fig. 11, PL 11) is directed anteriorly and leaves the cranial cavity through the wide foramen for the nervus oculomotorius without approaching the

14 174 NELLIE F. PATERSON brain (fig. 11, PL 11; fig. 27 d, PL 15). It passes into the orbit {o'pha.,fig.10, PL 11), and, running antero-laterally, it enters the eyeball immediately above the nervus opticus, after havmg divided as in the Salamander (Francis, 1934) to supply the eye-muscles and other tissues in the orbit. The head vein is formed anteriorly by factors from the snout and olfactory regions, where it seems to be homologous with the orbito-nasal vein of Eana (Gaupp, 1899). Proceeding posteriorly it receives the venous blood from the levator hyoideus muscle (l.h.vn., figs. 7, 8, PL 11), and is then located dorsally to the musculi levatores mandibulae in the groove between the processus muscularis and the commissura quadrato-cranialis anterior (h.vn.,figs.7, 8, PL 11). Unlike the carotid artery, it does not pass round the processus muscularis, but pursues a more or less straight course to the orbital region where it is situated close to the subocular bar of the palatoquadrate. In the orbit it is joined by two veins. One, which corresponds with the inferior orbital of Rana, runs from the floor of the orbit and joins the head vein anteriorly. The other, the ophthalmic vein (o'ph.vn., figs. 9, 10, PL 11), is longer and passes backwards from the dorsal part of the eyeball to join the main vein just in front of the foramen for the nervus oculomotorius and near the posterior limit of the subocular fenestra. Behind the orbit the vein passes to the ventral surface of the chondrocranium, where it lies in a groove together with the ramus ophthalmicus profundus V, the truncus hyomandibularis VII, and the ramus palatinus VII (fig. 13, PL 12; fig. 27 b, PL 15). Posteriorly to the carotid foramen the venous blood from the brain is collected into a large vein which passes between the truncus supra-orbitalis and the rest of the ganglion pro-oticum (figs. 15,16, PL 12;fig.27 d, PL 15). This cerebral vein is joined among others by a pituitary vein, which passes forwards from the hypophysis along the mesial side of the ganglion pro-oticum but shows no signs of anastomozing with its fellow. As the component nerves of the ganglion pro-oticum diverge in passing through the wide foramen pro-oticum, the main cerebral vein also makes its exit from the cranial cavity {cm., fig. 15, PL 12; fig. 27fc.PL 15) and joins the main head vein (h.vn.,fig.15, PL 12)

15 HEAD OF XBNOPUS 175 lying in the above-mentioned groove on the lower surface of the chondrocranium. From the foregoing it is evident that, apart from a few minor details, the main course of the internal carotid artery and the head vein of X. laevis is in close agreement with the findings of other workers on both Urodele and Anuran Amphibia. THE CHONDROCBANIUM. Within comparatively recent years information regarding the finer details of the morphology of the head of representatives of the various vertebrate phyla has been accumulating rapidly, and in this connexion the impetus given to comparative cranial osteology, which has resulted in such interesting conclusions regarding the evolution of the parts, has been largely due to the observations of such workers as de Beer, Gaupp, Goodrich, Parker, de Villiers, and his students. The latter have been responsible more particularly for many exhaustive studies of the skulls of South African Anura. It is to be noted, however, that in the Amphibia, while observations have been made on the development of the skulls of Urodela and Anura Phaneroglossa, little information is to be found regarding the condition in the five genera of Anura Aglossa. Edgeworth (1935), in his monograph on the cephalic muscles of the vertebrates, describes the palatoquadrate of Pip a and of a larva of X. fraseri measuring 17 mm. long. Some of his findings seem to be at variance with Kotthaus's lengthy account (1933) of the chondrocranium of X. laevis. The present study, in addition to being an attempt at tracing the transition from the larval to the adult condition in X. laevis, also endeavours to correlate the findings of these two previous investigators with the conclusions herein arrived at concerning the construction of the chondrocranium. Kotthaus (1933) has made a study of five different larval stages, ranging in length from 7-4 mm. to 48 mm., and in age from 5 to 53 days. As the material for the present investigation was collected locally in Johannesburg, where the larvae abound in certain ponds and dams, no attempt was made to record their

16 176 NELLIE F. PATEBSON ages, but the measurements of the preparations selected for study are as follows: Stage Total Length. 5-7S mm Length to Cloacal Aperture. 2-0 mm mm. (metamorphosing) Figure. 22 a a & 6 In addition to the above, serial sections were also made of a larva which had just emerged from the egg, and of several young frogs after decalcification of the skeletal elements. Diagrammatic reconstructions such as those sketched in figs. 22a,&, 23, PI. 13; figs.24, 25a,b, PI. 14;figs. 26a,b, 27a-d,Pl. 15 were plotted graphically from drawings of the serial sections, and in some cases, as for instance in the chondrocrania shown in fig. 25 a, PI. 14; figs. 27 a-d, PL 15, wax models were also constructed to check the accuracy of the graphical reconstructions. It is apparent from the above short table of measurements that the first two stages at present under observation are both younger than Kotthaus's first stage, while Stage 3 seems to be intermediate between his first and second larvae. There are also some slight variations in the sizes of subsequent stages, but these differences are, however, hardly of any consequence to a study that endeavours to survey the whole development of the chondrocranium. The chondrocranium of Xenopus is of the more primitive platybasic type described by Gaupp and Goodrich (1930). This is clearly demonstrated in the very early stages depicted in figs. 22 a, b, PI. 13, but is also evident even in larvae measuring 10 mm. in length (fig. 23, PI. 13). Serial sections of a larva which had just emerged from the egg, and which like Stage 1 also measured 5 mm. in length, revealed that the chondrocranium was not sufficiently differentiated from the connective tissue to allow of the determination of the parts. Therefore, no stage showing the separate trabeculae cranii and parachordals was

17 HEAD OF XENOPDS 177 found, even although there was little difference in the external appearance of the recently hatched larva and that used for fig. 22 a, PI. 13, in both of which the mouth was still closed. In Stage 1, therefore, the trabecnlae (trab.) and parachordals (pcli.,figs.22 a, b, PL 13) have already fused, and the primitive fenestra hypophyseos (basicranial fenestra of Kotthaus, 1933) is very broad and filled by the hypophysis. Even at this early stage, the fenestra is bordered anteriorly by the broad intertrabecular plate (Up., fig. 22 a, PL 13), which itself is continuous with the flat ethmoidal region in front. In larvae measuring 10 mm. long (Stage 3, fig. 23, PI. 13) the fenestra is somewhat smaller, and in these measuring 16 mm. it was observed that the membrane stretching across its floor was completely invaded by a single layer of cartilaginous cells. The fenestra, therefore, closes rather earlier than is indicated in Kotthaus's preparations, in which traces of it were still found in larvae measuring 18-2 mm. The carotid foramina are indicated for the first time by Kotthaus in the 18-2 mm. long larva, so that it is not evident if they have already been demarcated in his first stage larva. Among the present preparations the carotid foramina are established in larvae measuring only 10 mm. in length (i.e. Stage 3). These are only slightly older than Kotthaus's first stage, which measured 7-4 mm., but as is evident in fig. 23, PI. 13, there is a considerable bridge of cartilage intervening between the fenestra hypophyseos (fb.) and the carotid foramen (/<?.). The parachordals are rather short in Xenopus, and they unite early to form the broad basal plate enclosing the notochord. The anterior end of the latter has already been reduced a little even in Stage 2 (fig. 22 b, PL 13), but its diminution is not nearly so rapid as is supposed bj Kotthaus (1933), for as is shown infig.27 d, PL 15, the notochord still extends well forwards and has the dorsal curvature described by Goodrich (1930) for typical forms. Even in older forms with a total length of as much as 60 mm., the notochord, although considerably reduced, may still be traced up to the level of the posterior acustic foramen. In early larval life (figs. 22 a, b, PL 13) the brain lies fully exposed above the broad ethmoidal and intertrabecular plates. NO. 322 N

18 178 NELLIE F. PATERSON There are no indications of lateral walls, which develop later (Stage 3,fig. 23, PL 13) as more or less vertical cartilages directly continuous with the trabeculae cranii. Thus, in these first tnree larval stages the brain is not at any point roofed over by cartilage. Anteriorly it lies in a shallow depression in the intertrabecular plate, and the nervi olfactorii are continued forwards over the dorsal surface of the flat ethmoidal plate to the olfactory sacs. As development proceeds, the side walls gradually arch over the brain, but as they only meet anteriorly and posteriorly, a single large fronto-parietal fenestra remains dorsally. The arrangement is, therefore, somewhat different to that obtaining in R a n a, where Gaupp (1896) describes the presence of taeniae tecti transversalis and median's, and the resultant division of the dorsal opening into a fenestra frontalis and a fenestra parietalis. In Xenopus, due to the absence of the taeniae tecti transversalis and medialis, the single fenestra represents a combination of those featured in E ana. It would also appear from the studies of de Villiers and his students that, with few exceptions (Bufo, Schoonees, 1930), the Ranid arrangement of this part of the chondrocranium is not evident in South African A n u r a. The posterior fusion of the cartilages dorsally results in the formation of the tectum posterius (tp., fig. 24, PI. 14; fig. 27 a, PL 15), which was first observed in larvae measuring 28 mm. (Stage 5). As is consistent with the condition found in other Anur a (Gaupp, 1896; Goodrich, 1930), there is also an anterior fusion of the cartilages forming a dorsal arch (to.,fig.24, PL 14; figs. 27, a, d, PI. 15) over the anterior part of the brain. This anterior tectum is continued forwards, but never extends as far as the larval olfactory sacs. It occupies a position similar to that of the epiphysial and paraphysial bars of Ami a and other Pisces (Goodrich, 1930), which, however, are transverse bars developed from the supra-orbital cartilages. The palatoquadrate of X. laevis is essentially similar to that of X. fraseri (Edgeworth, 1935). In all larval stages there is a pronounced processus muscularis {mus.fr., figs. 7, 8, PL 11; figs. 22 a, b, PL 13; fig. 24, PL 14; figs. 27 a, c,pl. 15) continuous with the commissura quadrato-cranialis anterior (c.q.c.a.) which is directed mesially to join the ethmoidal region.

19 HEAD OF XEKOPUS 179 The palatoquadrate passes backwards along the floor of the orbital region as a long subocular bar (s.o.b.,fig.9, PL 11; figs. 22 a, b, 23, PL 13; figs. 27 a-d, PL 15) which ends in a prominent lateral process, the ventro-lateral process in Edgeworth's description ofx.fraseri. In Stages 1 and 2 the ventro-lateral process is faintly indicated and the subocular bar, although somewhat slender posteriorly, nevertheless effects a fusion with it, the result being that the subocular fenestra is completely formed between the subocular bar and the trabecula cranii. Kotthaus (1933) has found that in his earliest stage (7-4 mm. long) the subocular fenestra remains open posteriorly. It is therefore of some interest to note that, although the subocular bar is fully established in Stages 1 and 2, in Stages 3 and 4 of the present series, with the greater elongation of the chondrocranium, the posterior part of the subocular bar becomes very slender. The cartilaginous cells are only seen with difficulty at this point, and it almost appears as though the subocular bar and ventrolateral process were united by connective tissue only. It is considered, however, that the subocular fenestra is in all stages completely enclosed. The processus ascendens {ascpr.,fig. 11, PL 11; fig. 13, PL 12; fig. 24, PL 14; figs. 27 a, c, PL 15) effects a fusion with the pila antotica, so that there is a continuous cartilaginous wall between the oculomotor and pro-otic foramina. The larval otic process, which is hardly evident in young larvae (fig. 23, PL 13), develops rather later and fuses with the auditory capsule (l.o.p.,figs.14, 15, PL 12; fig. 24, PL 14; figs. 27, a, c, PL 15). As there is no basal process (vide fig. 27 b, PL 15) the palatine nerve (pal.) and head vein Qi.vn.), which in the larva run parallel for some distance, lie freely in a ventral groove of the chondrocranium, and are not at any point underlain by a cartilaginous process of the palatoquadrate. The quadrato-ethmoidal cartilage (q.e.c,fig.6, PL 11; fig. 23, PL 13;fig.24, PL 14;figs.27 a, c, PL 15) is a slender bar extending between the quadrate and the outer angle of the ethmoidal region. It is not developed in the first two stages, but becomes evident in Stage 3 (10 mm. long), in which the rudiment of the tentacle makes its first appearance. About a third of the way

20 180 NELLIE F. PATERSON from the anterior limit of the quadrato-ethmoidal cartilage the cartilaginous axial support of the tentacle projects laterally (e.tent,fig.23, PI. 13; fig. 24, PI. 14; figs. 27 a, c, PI. 15), a condition which corroborates Edgeworth's view (1935) that the axial cartilage is an outgrowth of the quadrato-ethmoidal cartilage. In his second stage larva, Kotthaus (1933) finds that the connexion between the quadrate and ethmoidal regions is incomplete, and he has concluded that the cartilage directed forwardly from the quadrate is the support for the tentacle. It follows, however, that if there is an incipient tentacle in the larva represented in fig. 23, PL 13, there should be a slightly longer tentacle in one measuring 12*3 mm., and the connexion between the quadrate and ethmoidal cartilage should be effected. As the tentacle gradually elongates during larval life, its axial cartilage becomes concomitantly longer (e.tent.,fig.24, PI. 14; figs. 27 a,c, PI. 15). The present preparations have, however, not given full support to Edgeworth's statement (1935) that 'subsequently the posterior part of the quadrato-ethmoidal cartilage disappears leaving the tentacle attached to the ethmoidal region'. Parker (1876) has depicted this condition in the larva of X. lae vis, but these preparations of the same species indicate that the quadrato-ethmoidal cartilage persists as a slender bar which is still apparent just prior to metamorphosis in larvae measuring 60 mm. (Stage 6). During metamorphosis the quadrato-ethmoidal cartilage and also the tentacular cartilage are both absorbed, and neither is apparent at the end of this process. It seems logical to suppose that, during metamorphosis when the jaws elongate, the connexion between the ethmoid cartilage and quadrate will disappear, so that for a time the tentacle will remain attached to the outer corner of the ethmoidal cartilage as Edgeworth maintains. Authorities such as Gaupp ( ), de Beer (1926,1937), and Goodrich (1930) have shown the importance of the position of the various foramina in the elucidation of the parts of the chondrocranium. As a detailed study of the cranial nerves has been made during the course of the present investigation (p. 198), some reference may here be briefly made to certain discrepancies between the present findings and those of Kotthaus. In his

21 HEAD OF XENOPUS 181 account of the chondrocranium of X. laevis Kotthaus has dealt more particularly with the positions of the nerves in the cranial cavity and their exits therefrom. The fact that he has not followed the entire course of each of the nerves may account for certain erroneous conclusions. He finds that in all stages, except the last, the nervi opticus and oculomotorius emerge through a common foramen, but that in the last larval stage they are separated. The preparations used for the present study have revealed that these two foramina are separated from each other very early in larval life. Almost as soon as the lateral walls of the chondrocranium are developed the pila metoptica becomes apparent. In one series of sections of a larva measuring 10 mm. (fig. 23, PI. 13) the oculomotorius passed out through the narrower posterior part of the foramen opticum, but in another series of a similar larval stage a narrow pila metoptica intervened between the two foramina. In all subsequent stages examined each of these nerves left the cranial cavity by its own wide foramen. This arrangement is clearly shown in fig. 11, PI. 11, where the nervus oculomotorius is seen passing out posteriorly to the nervus opticus, and is also evident in figs. 27 a, c, d, PI. 15, which are reconstructions of a younger larva than that showninfig. 11, PI. 11. Kotthaus's account and figures of the elements arising from the ganglion pro-oticum are somewhat at variance with the present findings. It seems that he has overlooked the presence of the truncus supra-orbitalis, and has thereby been misled in his interpretation of the other branches of the trigeminal and facial nerves. He thus shows in his fig. 10 a ramus maxillaris V, which, as will be explained later below, is considered to arise quite differently in X. laevis than that of typical Anura. He also represents the ramus palatinus VII as a lateral nerve passing dorsally over the ventro-lateral process of the palatoquadrate, whereas it maintains a ventral position throughout. The nervus abducens, which is an extremely delicate nerve passing from the cranium with the ramus ophthalmicus profundus V, is represented as lying mesially to the two branches of the trigeminus and as a nerve subequal in thickness to the rami of V and VII. It is therefore apparent that there is little agreement between Kotthaus's views and the

22 182 NELLIE F. PATERSON present conclusions regarding the cranial nerves and their foramina. It is the opinion of the present writer that, immediately on passing out through the wide foramen pro-oticum, the truncus supra-orbitalis and ramus mandibularis V diverge from the rest of the ganglion pro-oticum (vide figs. 14,15, PI. 12;figs.27 a-d, PI. 15) and pass out dorso-laterally through what may correspond with the superior trigeminal foramen of Kotthaus. The rami ophthalmicus profundus V, hyomandibularis VII, and palatinus VII pass out together through the wide ventro-lateral portion of the foramen pro-oticum (fig. 14, PI. 4; fig. 27 d, PL 15). The dorsal and ventral parts of the foramen pro-oticum are directly continuous, and as in other Anura (Goodrich, 1930) there is no prefacial commissure. The nervus acusticus penetrates the mesial wall of the auditory capsule through two foramina (acus., fig. 27 d, PI. 15), a condition which is evidently typical of the Anura, judging by the observations of Goodrich (1930), Bchoonees (1930), du Toit and de Villiers (1932), du Toit (1933), and de Vos (1935). Occasionally, as in Hemisus (de Villiers, 1931), only one acustic foramen occurs, while Hylambates (du Toit, 1930) is exceptional in having three. Confluent with the acustic foramina at an early stage is the endolymphatic canal (ez/.+ acus., fig. 23, PI. 13), the three becoming separate passages in somewhat older larvae (fig. 27 d, PI. 15). The glossopharyngeal, vagus, and posterior lateral line nerves after arising separately from the brain, unite within the cranial cavity (figs. 27 a, d, PI. 15). They therefore emerge together through the postero-ventrally situated foramen jugulare (J.jug., fig. 27a,PI. 15), immediately lateral to which they form the large complex ganglion (g.v.), from which the several branches are later separated. Little importance attaches to the jaw region. The Meckel's cartilages are slightly curved rods (Meek., figs. 22 a, b, PI. 13; fig. 24, PI. 14;figs.27 a, c, d, PI. 15), which project forwards from the quadrate below the superior labial cartilage formed by the anterior margin of the ethmoidal cartilage. The Meckel's cartilages are joined by a single median inferior labial cartilage (ilc, figs. 22 b, 23, PI. 13; fig. 24, PI. 14), which is not present until

23 HEAD OF XENOPUS 183 the mouth opens (fig. 22 b, PL 13). It is noticeable in the diagrams that, compared with the whole chondroeranium, the jaws of the larva are very short. They are situated far forwards, articulating with that part of the quadrate in front of the orbit. 1. Metamorphosis. Apart from the ossification of the parts, the chief differences in the conformation of the chondrocranium attendant on metamorphosis are brought about by the rapid development of the nasal capsules and the backward movement of the jaws, so that the quadrate is located at the anterior level of the auditory capsule. As was previously mentioned, there is a wide dorsal fenestra in larval Xenopus, which at metamorphosis becomes roofed. over by the development of the fronto-parietal bones. In describing the skull of Phrynomerus, de Villiers (1930) discusses the question of this dorsal fenestra in relation to the fronto-parietalia, and finds that in the adult Phrynomerus there is a large fontanelle between the two small lateral frontoparietalia. That this is not an isolated occurrence is borne out by the observations of C. A. du Toit (1930, 1931) on Heleophryne, de Villiers (1931) on Cacosternum and Breviceps, and de Vos (1935) on Spelaeophryne, in each of which there is a dorsal bridge of fibrous connective tissue joining the pair of small fronto-parietalia. The larger fronto-parietalia of Phrynobatrachus are observed by du Toit (1933) to be separated mid-dorsally by a narrow strip of connective tissue, and he considers that, as this condition obtains in so many different genera, as for instance, in Arthroleptella, Anhydrophryne, Heleophryne, Bufo, and Hyperolius, it is typical of the Anura. InHemisusde Villiers (1931) finds that there is a mid-dorsal fusion of the fronto-parietalia, and it is of interest to note in this connexion that this somewhat atypical condition also obtains in Xenopus. The fronto-parietalia, which may be seen towards the end of larval life as paired dorso-lateral bones (fp., figs. 9, 11, PI. 11;figs.12,14, PI. 12) coalesce mid-dorsally in the young frog, forming a continuous thin bone which in more mature specimens

24 184 NELLIE F. PATERSON (fig. 26 a, PI. 15) acquires a sagittal crest. It extends forwards from the tectum posterius as a single bony sheet over the whole surface of the brain, and at the nasal region is directly continuous with what Gilchrist and von Bonde (1919) have termed the supraethmoid, a slightly arched membrane bone (seth.,fig. 3, PL 9; fig. 4, PL 10; fig. 26 a, PL 15) which overlaps the paired nasals. A short distance behind the supra-ethmoid is a minute median aperture (f.par., fig. 26 a, PL 15) through which, as transverse sections reveal, the tractus pinealis passes. For this reason Winterhalter (1931) has identified it with the foramen parietale of Pisces, Stegocephalia, and Eeptilia. The os en ceinture, a term applied by the majority of recent workers to the bone encircling the sphenethmoid region, characteristically extends from the optic foramen to the nasal capsule laterally and sometimes occurs ventrally between the parasphenoid and the solum nasi (de Villiers, 1931; du Toit, 1933). In sections of young Xenopus the sphenethmoid region is occupied by a very thin bone which forms the lateral wall of the cranium anterior to the optic foramen (o.e.c, figs. 26 a, b, PL 15). It joins the fronto-parietalia with the cartilaginous floor of the cranium in front, and rather more posteriorly it is confluent with the parasphenoid. It is obviously homologous with the os en ceinture, but it does not possess the large marrow cavities that characterize it in other Anura (de Villiers, du Toit). The lateral wall of the cranium between the optic foramen and the pro-otic is evidently an ossification of the larval pila antotica, and is therefore to be identified as the pleurosphenoid (pzs., fig. 26 a, PL 15), a term suggested by Goodrich for this region (1930). In the majority of Anura the larval palatoquadrate develops into the adult processus pterygoideus, supporting the quadrato-maxillare and the paraquadratum. In the young Xenopus the processus pterygoideus (p.ptg., fig. 25, PL 14; figs. 26 a, b, PL 15) is a cartilaginous bar passing backwards from the antorbital process beneath the ventro-lateral surface of the eyeball, occupying a position similar to that of the subocular bar of the larval palatoquadrate. It extends towards the quadrate and is roughly oval in transverse sections. In the subocular

25 HEAD OP XENOPUS 185 region only its mesial and dorsal surfaces are covered by bone, but behind the orbit it may be almost completely invested by the pterygoid, which posteriorly has a ventral projection parallel to and almost meeting the processus coronoidens of the lower jaw (cor.,fig.18 a, PI. 12). In Xenopus the processus pterygoideus merges posteriorly with the quadratum (fig. 18 b, PL 12), a relatively broad postorbital process lying beneath the auditory capsule (quad., figs. 18 a-c, PL 12). Duringmetamorphosisthequadratumisproduced dorsally into a prominent otic process (otp.,fig.25 b, PL 14) and ventrally into an equally well-developed quadrate process (quad.). The otic process projects as far as the crista parotiea, but at this stage there is no fusion of these two parts, although in a lateral aspect (fig. 25 b, PL 14) the crista is seen to overhang the otic process laterally. When metamorphosis is completed, the otic process articulates with the ventro-lateral wall of the auditory capsule, and anteriorly has a dorsal extension which proceeds to the crista parotica, but its synostosis with the latter is not so marked as in the majority of the Anura Phaneroglossa (de Villiers, 1930, 1931; du Toit, 1933; de Vos, 1935). The remaining elements belonging to the palatoquadrate in the majority of Anura are mainly of interest in Xenopus on account of their negativeness. The quadrato-maxillare (quadrato-jugal of Gilchrist and von Bonde, 1919), which appears to be a feature of many Anura Phaneroglossa (Gaupp, 1896; de Villiers, 1930 et seq.) is absent in Xenopus. Gilchrist and von Bonde state that ' they (the quadrato-jugals) are very inconspicuous' and do not represent them in their diagrams. The maxillare (max.,figs. 26a, b, PL 15) iscontinuedfor some distance beneath the antorbital process, but it ends abruptly and has no connexion with the quadrate, so that it is quite evident that the quadrato-maxillare is lacking inxenopus. de Beer (1926,1937) and Goodrich (1930) have explained the importance attaching to the articulations between the palatoquadrate and the chondroeranium in the Tetrapoda, and in this connexion the presence of a basal process is of interest. The former author has shown that in the Anura Phaneroglossa there is no true basal process, but that there is a pseudo-basal

26 186 NELLIE F. PATERSON process posterior to the ramus palatinus VII. In the foregoing account of the chondrocranium of the larval Xenopus it has been shown that there is no basal process. During metamorphosis the ascending process of the palatoquadrate disappears, a feature which is consistent with the development of typical Anura (de Beer, 1926, 1937). The adult otic process (otp., fig. 26 a, PL 15) is a short broad bone which passes forwards and inwards from the quadrate to meet the anterior margin of the pro-otieum (pro.). During and after metamorphosis the anterior end of the auditory capsule is connected to the basal plate by a ventral cartilage (ppc, figs. 18 a-c, PL 12; fig. 26 b, PL 15) which underlies the large ganglion pro-oticum (g.pro.). It passes downwards and outwards, coming into contact not only with the pterygoideus but also with the inner margin of the quadratum, near the junction of the latter and the processus pterygoideus (figs. 18, a-c, PL 12). The head vein Qi.vn.,figs.18 a, b, PL 12) and the ramus hyomandibularis VII Qvym.) lie dorsally to it, and as it lies posteriorly to the ramus palatinus VII, and consequently does not affect the course of this nerve, there can be no doubt that it is comparable with the postpalatine commissure of other Vertebrates, and therefore with the pseudobasal process described by de Beer (1926, 1937). As in other Anura (de Beer, 1926, 1937), the pterygoideus of Xenopus has a medial portion (m.ptg., fig. 18 c, PL 12; fig. 26 b, PL 15) which projects as a thin horizontal plate below the eustachian tube, and anteriorly is in contact with the post-palatine commissure. The squamosum (paraquadratum of Gaupp and de Villiers) appears during metamorphosis as a slender membrane bone (prq., figs. 18 a-c, PL 12;figs.26 a, b, PL 15) lying lateral to the auditory capsule and immediately above the plectrum (plec). It forks posteriorly over the middle ear, so that in sections (figs. 18a-c,P1.12) dorsal and ventral parts are evident and the ventral part becomes closely apposed to the ventral part of the annulus tympanicus. Gilchrist and von Bonde (1919), in their macroscopic examinations of the skull of Xenopus, have also remarked on the absence of the palatinum. This, however, is a variable quantity even in the Anura Phaneroglossa, in some species of

27 HEAD OF XENOPTJS 187 which it has also been observed to be lacking (de Villiers, 1933; de Vos, 1935). In regard to certain of these negative points, it is of interest to note that the quadrato-maxillare and palatine are also absent in Hemisus, Breviceps, and Probreviceps (de Villiers, 1931, 1933). These, and other features that these species may have in common with X e n o p u s, cannot, however, indicate any possible affinity, nor can they even be regarded as the results of adaptations to similar habitats or habits, for as de Villiers (1931) has shown^ Hemisus is very markedly modified for terrestrial life, while, as is well known, X en op us is a purely aquatic form. Xenopus is somewhat exceptional among Annra in having a single median vomer (= praevomer, Broom, 1903, 1935). The parasphenoid, which is regarded by Broom as the homologue of the mammalian vomer, is one of the first bones to be observed in Xenopus. It appears in larvae measuring 28 mm. (Stage 5), in which its pointed anterior end occurs below the bulbi olfactorii, while its posterior extremity is located near the anterior limit of the notochord. It constitutes the somewhat thin-walled floor of the cranial cavity, and as it maintains practically the same width throughout it does not possess the postero-lateral processes that are so conspicuous in the Eanid type of skull (jps.,fig.26 b, PL 15). During metamorphosis the praevomer is developed along the ventral surface of the septum nasi, and from its inception is directly continuous with the parasphenoid, no line of demarcation being observed between the two. In Teleostei, where a median praevomer (Goodrich, 1930) is a typical condition, it is thought to represent a fusion of a pair of embryonic structures, but inxenopus no indication of a paired origin was found even during metamorphosis. In fact, in sections it merely appears to be a forward prolongation of the parasphenoid along the floor of the olfactory capsule, and even though it is such a distinct bone (vom.,fig.26 b, PL 15) in the adult, there is a possibility that the median praevomer of Xenopus is merely an anterior part of the parasphenoid. The lower jaw is typical of the Anura in the investiture of the Meckel's cartilage by only two membrane bones, the

28 188 NELLIE F. PATEBSON angulare and the dentale. The former is the posterior crescentic bone (ang.,fig.18 c, PI. 12) which leaves the cartilage exposed dorsally and which is produced mesially into a well-marked processus coronoideus (cor.,fig.18 a, PI. 12) for the insertions of the musculi levatores mandibulae (lev. man.,fig.18 a, PI. 12). The dentale, as in the majority of A n u r a, is a slender edentulous bone. 2. Olfactory Eegion. One of the most striking occurrences at metamorphosis is the rapid, almost sudden transformation of the olfactory region. Throughout larval life the nasal sacs lie on each side immediately above the lateral part of the ethmoidal cartilage, and just behind the junction of the ethmoid and quadrato-ethmoidal cartilage. Not even in older larvae, such as Stage 6, were the cartilages of the adult nasal capsule foreshadowed. In the larvae the internal nares tend to lie at a level anterior to the widely open external nares, so that the former are located in front of the union of the palatal process of the palatoquadrate with the ethmoidal cartilage. In the very young larvae, before the mouth opens, neither of these apertures is established. The three cavities of the adult may be recognized, the laterally placed cavum medium (cav.med.,fig.6, PL 11) communicating widely with the cavum principale (cav.princ). The recessus medialis of the cavum inferius is remarkable at this time for its size, and for the fact that, as in the majority of adult A n u r a, it is anteriorly situated and appears in sections even before the cavum principale. This may be due to the anterior position of the internal nares in the larva, for in the adult Xenopus the internal nares and also the recessus medialis are more posteriorly situated. The glandula nasalis medialis occurs between the dorsal wall of the recessus medialis and the cavum principale as they do in the young frog. The section represented in fig. 6, PI. 11, does not pass through the recessus medialis, but the glands may be seen lying immediately ventral to the cavum principale. During metamorphosis the whole of the anterior part of the skull becomes complicated by the development of the cartilages and additional chambers in the olfactory region. This region,' as well as the auditory capsule, has been thoroughly examined

29 HEAD OP XENOPTJS 189 in various genera of South African Anura by de Villfers and his students, and although he concludes (1930) that there is, on the whole, little variation in the size and shape of the nasal cavities, yet, as is to be expeeted, there are some slight differences that seem to characterize the individuals. In Xenopus the cartilages (fig. 25 a, PL 14) are broadly arranged on a plan similar to that described by Gaupp ( ) in Eana, which has formed the basis of all more recent work. The whole nasal region of Xenopus is projected upwards and forwards in front of the lower jaw, and is remarkable for the size of the cartilago alaris, which is outlined in fig. 25 a, PI. 14 (cart.al.). The cartilago alaris, the tectum and septum nasi all appear in sections anterior to the nasal cavities, and as the cartilago alaris extends as far back as the planum terminate, it constitutes the main lateral cartilaginous wall of the nasal region. Anteriorly it has a very narrow connexion with the solum nasi just in front of the crista intermedia, and behind this its inner concave surface is curved over the lateral surface of the cavum medium, the naso-lacrimal duet (n.l.d.,fig.2, PL 9) passing between it and the septomaxillare (sem.). There is only one prenasal cartilage, and as this is an antero-ventral extension of the cartilago alaris, it must be regarded as being comparable with the superior prenasal cartilage of Eana. It is, however, ventral in position, and differs from that of Eana and apparently from the majority of Anura in that it is flexed under the nasal capsule and is continued posteriorly to fuse with the solum nasi. This arrangement partly resembles that obtaining in Hemisus (de Villiers, 1931), in which there is likewise no inferior prenasal cartilage and in which the cartilago prenasalis superior arises postero-ventrally from a large cartilago alaris. In Hemisus, however, it does not seem to establish a connexion with the solum nasi, so that in Xenopus it might even be regarded as a fusion of the two prenasal cartilages present in Eana. As in Hemisus, the prenasal cartilage (spc, figs. 1-3, PL 9;fig.25 a, PL 14) acts as the support for the premaxillare, which like the maxillare is toothed. It extends as far back as the glandula intermaxillaris (g.i.m.,fig. 3, PL 9) being situated dorso-laterally to it, but never imbedded in it. The lamina

30 190 NELLIE F. PATBESON superior and inferior are not quite normal when compared with the Eanid formation. Following Gaupp's description (1904), the crista intermedia should separate the superior and inferior nasal cavities and the cavum medium should lie between its two laminae. In Xenopus the crista intermedia itself is normal, but rather lateral in position, so that it is located (eris., fig. 2, PI. 9) between the cavum medium and the recessus medialis of the cavum inferius. The crista intermedia is not produced into a conspicuous lamina superior above the cavum medium as in the majority of Anura. In Hemisus de Villiers (1931) has observed an unusually small lamina, superior, which, however, is normal in position. In Xenopus the usual lamina superior must be considered as being absent, a conclusion which is supported by the position of the septomaxillare in this region (sem., fig. 3, PI. 9). It may be noted, however, that at this point a thin inwardly directed cartilage (he, fig. 3, PI. 9) is given off from the crista intermedia in front of the planum terminale. It lies between the cavum principale and the recessus medialis, and is, therefore, atypical of the lamina superior. Furthermore, the lamina superior is usually directed laterally from the crista intermedia, but in this case it is mesially directed. These peculiarities are perhaps to be correlated with the fact that in Xenopus the recessus medialis of the cavum inferius is more posteriorly situated than in other Anura previously investigated. Posteriorly to this horizontal cartilage is another mesially directed cartilage. This is a thin horizontal rod occurring dorso-laterally over the isthmus near its junction with the recessus lateralis, and is evidently to be compared with the processus lingularis of Eana (Gaupp, 1904). The lamina inferior is a small posterior cartilage which fuses with the planum terminale of the cartilago obliqua. The lamina inferior (lam.inf.,fig.25 a, PI. 14) is dorso-laterally flexed, somewhat as in Phrynobatrachus (du Toit, 1933) and Hyperolius (du Toit and de Villiers, 1932). Together with the planum terminale it forms in section a somewhat horseshoe-shaped cartilage passing below the postero-lateral portion of the cavum medium and the naso-lacrimal duct. The septo-

31 HEAD OF XBNOPUS 191 maxiilare forms a dorsal arch over this region and almost meets the lateral part of the lamina inferior. Bather more posteriorly, and after the disappearance of the cavum medium from sections, the ductus nasolacrimalis is for some time almost encircled by the septomaxillare. The nasal bones (nas., figs. 2, 3, PL 9) are large in Xenopus and almost meet mid-dorsally above the septum nasi. Anteriorly they are narrowed and lie above the tectum nasi. They then extend laterally over the cartilago obliqua, and behind the planum terminale they rest on the planum antorbitale, bridging the gap between the anterior projection of the latter and the tectum nasi. 3. Nasal Cavities. Some idea of the arrangement of the nasal cavities of the young frog may be gained from an examination of figs. 1-3, PL 9; fig. 4, PL 10. The superior and median cavities are large, but the inferior in comparison is rather small. There is a small vestibular region supported laterally hy the cartilago alaris. The two 'Wulste' described by Gaupp (1904), and to which recent workers seem to attach some importance, are both represented. Immediately behind the vestibulum there is a recessus sacciformis communicating with both the infundibulum and the cavum medium, its relations with the latter being somewhat similar to those observed in Cacosternum by de ViUiers (1931). The recessus sacciformis (recsac, fig. 2, PL 9) is smaller than that of Ban a, and appears to be closely associated with the lateral wall of the ductus nasolacrimalis (n.l.d.,fig.2, PL 9). The two appear almost together in sections immediately posterior to and below the vestibulum, and both seem to effect a wide communication with the infundibulum. in Bana the two open close together into the cavum medium, but in X en o - pus they seem to be continuous with the infundibulum just anterior to the cavum medium, the aperture of the ductus nasolacrimalis being immediately behind that of the recessus sacciformis. The ductus nasolacrimalis (n.l.d.,ftg. 3, PL 9) passes backwards over the outer surface of the planum terminale, with the postero-

32 192 NELLIE F. PATBESON lateral portion of the cavum medium (cav.med.,fig. 3, PI. 9) situated laterally to it. It passes laterally over the planum antorbitale, and during metamorphosis communicates with Jie exterior in front of the lower eyelid by two apertures, the anterior of which is very inconspicuous. The presence of two external apertures is, however, not an unusual occurrence, for two openings have also been recorded in Hyperolius by du Toit and de Villiers (1932). In the young frog the anterior aperture disappears and the terminal part of the duct then forms the lumen of the so-called 'tentacle' of the adult (n.l.d., fig. 5, PL 10), the aperture being located at its free extremity. As in E a n a, the plica terminalis is similarly associated with a horizontal fold behind the vestibulum, but the plica obliqua, which is well' developed and more posteriorly located than in Eana, depends from the cartilago obliqua and the septomaxillare. It has also been observed by de Villiers (1930,1931) to be similarly associated with the cartilago obliqua in P h r y n o- merus and Oacosternum. The cavum medium (cav.med.,fig.2, PL 9) is large and is more or less horizontal in anterior sections. It has a large rather vertical postero-lateral extension (cav.med.,fig.3, PL 9), the dorsal and ventral walls of which are very thin. There is little to note regarding the cavum inferius. In comparison with the other two chambers it is small, and its recessus medialis (rec.med.,fig.3, PL 9), which in other Anura is well developed and anteriorly situated, is smaller and more posterior in position in Xenopus. A narrow recessus lateralis (rec.lat, fig. 3, PL 9) lies above the maxilla and continues as far back as the choane, where it forms a deep bay (smp.,figs.4, 5, PL 10) at the lower end of the cavum principale, thus corresponding with the sulcus maxillopalatinus of Eana (Gaupp, 1904). As may be observed fromfigs.4 and 5, PL 10, Xenopus differs from Eana and other Anura in having no marked eminentia olfactoria. A certain degree of interest attaches to the glands in the nasal region. The glands of Bowman and the glandula nasalis medialis are arranged similarly to those of Eana (Gaupp, 1904). In Xenopus the latter occurs between the cavum

33 HEAD OF XENOPUS 193 principale and the recessus medialis of the cavum inferius. Posteriorly to the recessus medialis it is located among the branches of the nervus olfactorius, between the cavum principale and the septum nasi. It seems to consist of two parts, a more anterior part opening into thefloorof the cavum principale (d.g.m.,fig.1, PL 9), and a posterior portion which communicates with the dorsal wall of the recessus medialis close to the level of the section shown in fig. 3, PI. 9. The glandula nasalis lateralis of E a n a is lacking in Xenopus, but the presence of an intermaxillary gland is of interest, de Vos (1935) has indicated the absence of this gland in Hemisus, Spelaeophryne, and in the Pipidae, its absence in the Anura Aglossa being correlated with the aglossal condition. In Xenopus,however, there can be no doubt as to the identity of the well-developed gland (g.i.m., fig. 3, PI. 9; figs. 4, 5, PI. 10) which occurs between the premaxillae. It opens on each side near the choane, about 80-90/x. behind the section represented in fig. 4, PL 10. Miiller (1932) has enumerated his reasons for considering that this is an intermaxillary rather than a pharyngeal gland, and his conclusions are fully corroborated by the present observations Auditory Eegion. In his youngest larval stage, measuring 7*4 mm., Kotthaus did not observe a cartilaginous ear-capsule, but he found a small lateral protrusion of the parachordal which he describes as follows: ' Am Ende des ersten Drittels ihrer Lange zeigen die Parachordalia einen kleinen, mit der Spitze nach aussen gerichteten Auswuchs; von diesem Fortsatz nimmt die Verknorpelung der Ohrkapsel ihren Ausgang und wiirde der " Com- 1 Since the above observations were made the writer has been able to procure a copy of Foske's account (1934) of the nasal cavities of X. laevis, of which none was obtainable in South Africa while the present studies were being prosecuted. The above description of the nasal cavities is, on the whole, in agreement with that of Foske, the main difference being one of terminology. In describing the glands associated with the cavities, Foske has identified the one between the pre-maxillae as a palatal rather than an intermaxillary, which he considers is absent, and he has termed the one opening into the cavum principale the glandula oralis interna, a gland not previously recorded in Amphibia. NO. 322 O

34 194 NELLIE F. PATERSON missura basi-capsularis anterior" (Gaupp, 1893) bei Eana entsprechen.' In the present series of preparations the paraehordals of Stage 1 do not seem to be so long as those of Kotthau^'s first stage, and no trace of the commissura basi-capsularis was found, although the auditory sacs (as.,fig.22 a, PL 9) were conspicuous. In slightly older larvae (Stage 2) the floor of the auditory capsule (fac, fig. 22 b, PI. 13) is established and is in contact with the more elongate parachordal, presumably at a point comparable with the commissura basi-capsularis anterior. In Stage 3, which is only 2-6 mm. longer than the youngest specimen examined by Kotthaus, the auditory capsules are fully developed and have effected their fusion with the parachordal region (fig. 23, PI. 13). In this and all subsequent stages the inner ear was found to be completely enclosed in cartilage except, of course, at such points as the fenestra ovalis, the endolymphatic and perilymphatic canals, and the two foramina acustica. The large anterior and posterior dorsal openings depicted by Kotthaus in his second stage were not observed. The wall of the capsule is admittedly thin between the ridges for the semicircular canals, but there is no trace of a gap in either the anterior or posterior cupula even in larvae measuring 10 mm. in length. The middle ear, which has been fully investigated in the adult Xenopus byde "Villiers (1932), makes its appearance during metamorphosis. An interesting stage in its development was observed during this process, when the eustachian tube was seen to develop as a narrow passage between the quadrate and the pro-otic. The eustachian tube has, at this time, a lumen of almost subequal diameter throughout; the vault described by Parker and confirmed by de Villiers (1932) is not, as yet, evident but may be seen in the young frog. The plectral apparatus seems to arise at two separate centres and becomes continuous in young forms. As is shown in fig. 25 b, PI. 14, the pars externa, which in the young frog consists of obvious dorsal and ventral parts, has not yet been clearly differentiated. In sections (videfig.18 b, PI. 12) they are faintly indicated in the tissues above and below the pars media plectri. Applied to the blind end of the middle ear (me.,fig.25 b, PI. 14)

35 HEAD OF XENOPUS 195 is a rounded cartilage which must be considered as the anterior end of the pars media plectri (pnip., fig. 25 b, PL 14). At this stage it is not connected with the rest of the plectrum, which is composed of a fairly long cartilaginous rod divisible into the pars media and pars interna of the adult. The latter is rather broader than the median part and is applied to the fenestra ovalis (J.o.,fig.25 b, PL 14). de Villiers (1932) finds that in the adult the pars media plectri passes through the middle ear, dividing it into dorsal and ventral chambers, an occurrence which is unusual among South African Anura. In the present preparations of metamorphosing larvae and young frogs this condition does not obtain. In all of them the pars media plectri (jmip.,figs.18 b, c, PL 12) lies externally to the middle ear, which remains undivided. Xenopus is known to present many variable morphological features, so that it may be suggested that the condition described by de Villiers is either not constant or is only to be observed in older frogs. The annulus tympanicus is poorly developed in metamorphosing forms. Its ventral rim passes backwards some distance, but its dorsal rim is inconspicuous. Even in young frogs (fig. 18 a, PI. 12) the dorsal rim (dat.) is less marked than the ventral (vat), and although it becomes rather better defined as development proceeds, the whole structure (seefigs.26 a, b, PL 15) never becomes quite annular as it is in Eana; the squamosum lies close to the dorsal lip of the annulus, and in part serves as a boundary in this region. This condition probably explains Parker's description of the annulus tympanicus as a crescentshaped structure. In sections of a metamorphosing tadpole and of a young frog shortly after metamorphosis full confirmation has been obtained of de Villiers's conclusion (1932) that in Xenopus a pars interna plectri and also an operculum are present as in many other Anura. During metamorphosis the pars media plectri expands posteriorly into a broader pars interna, which almost fills the fenestra ovalis. A small cartilaginous operculum is faintly indicated 40/x. posterior to the extremity of the pars interna, the intervening space being occupied by connective

36 196 NELLIE F. PATEESON tissue. In a very young frog which had just metamorphosed, the operculum (ope, fig. 19 b, PI. 13) was more obvious as a cartilaginous piece between the upper and lower margins of the fenestra ovalis. At this stage the operculum is more obviously fused with the upper part of the auditory capsule, but its dorsal and ventral limits are still apparent. Between the pars interna plectri and the operculum the connective tissue, which is just beginning to appear in the section represented in fig. 19 a, PI. 13 (cnt.), has been somewhat reduced and now covers only 20/t. In rather older specimens the operculum becomes more closely incorporated into the wall of the auditory capsule and the connective tissue joining it to the pars interna plectri is not so obvious. The condition of the sound-conducting apparatus in young specimens of Xenopus is therefore essentially similar to that of E a n a (Gaupp, 1904), and fully justifies de Villiers's (1982) conclusions regarding the identity of the structures in the fenestra ovalis of the adult frog. THE HYOBEANCHIAL SKELETON. With the exception of the two early stages, the hyobranchial skeleton of the various larvae used in the present study differs but little from that described and figured by Kotthaus (1933). Some variation in the rate of development is naturally to be expected in these very early stages, so that while Kotthaus finds incipient developments of all four branchial arches in larvae measuring 7-4 mm., the third and fourth being merely slight lateral protrusions, in the first stage of the present series only the first and second arches are established. In Stage 2 (7 mm. long) the development of the arches has proceeded farther than Kotthaus's first stage, with the result that all four arches have united and the branchial skeleton, although small, is actually in a condition similar to that of older larvae. At metamorphosis certain of the parts of the hyobranchial skeleton undergo somewhat radical changes, but as these have been fully elucidated by Eidewood (1897, 1900) in his studies of this and other species of the Anura Aglossa, it is only necessary to recapitulate certain of the more salient points. The features that have been emphasized by Eidewood as being

37 HEAD OF XENOPUS 197 characteristic of the Aglossa include the large hyoglossal foramen (kf., fig. 21 b, PL 13), which seems to appear after the absorption of the basihyale and part of the basibranchiale. It is bounded anteriorly by the anterior cornua, which in the metamorphosing stage (fig. 21 a, PI. 13) are joined by connective tissue only (cnt), but which in the young frog (fig. 21 b, PL 13) effect a secondary cartilaginous fusion. The large posterolateral processes are also characteristic of the Aglossa, and are considered by Eidewood (1897) to develop as secondary outgrowths of the antero-lateral region of the branchial plate and not from the remains of the branchial skeleton. This conclusion is based on the comparative study of the development in Pip a and X e n o p u s; but, on the other hand, a comparison of the two hyobranchial skeletons represented in figs. 21 a, b, PL 13, suggests the possibility of the origin of the postero-lateral process from the reduced branchial skeleton (bra., fig. 21 a, PL 13). The relative arrangement of the parts, and especially the position of the thyroid foramen (th.f.,figs.21 a, b, PL 13) are constant in the two. One might, therefore, conclude that in Xenopus there is a direct correlation of the compressed branchial skeleton and the postero-lateral process of the adult, although this does not seem to be substantiated by Eidewood's observations on P i p a. During metamorphosis the thyroid foramen (ih.f.) gradually becomes an open cleft, which in the young frog lies between the elongate postero-median (ppm.) and postero-lateral processes (ppl, figs. 18 a-c, PL 12; fig. 21 b, PL 13). The incipient posteromedian processes are recognizable during metamorphosis (fig. 21 a, PL 13), and their union with the larynx described by Eidewood (1897) probably relates to later adult life, for in the young frogs herein examined they were only attached posteriorly to the ericoid, the connexion seemingly corresponding with the ligamentum hyo-cricoideum of Eana (Gaupp, 1904). The larynx agrees with Eidewood's description (1897). Its cartilages are noticeable even in early larval stages, and in young frogs the antero-ventral part of the annular ericoid cartilage (Ian., fig. 21 b, PL 13) becomes adherent to the posterior wall of the hyoid. The bronchial cartilages (br. pr.) are

38 198 NELLIE F. PATEBSON comparatively longer than those observed by Eidewood in older specimens. THE CRANIAL NERVES. In demonstrating the anatomy of Xenopus to students the writer has been at a loss to explain the innervation of the upper jaw. The maxillary and mandibular nerves are strongly developed and subequal in thickness in typical Anura Phaneroglossa. In macroscopic examinations of Xenopus, however, while the mandibular nerve is easily traced, it has not been possible to demonstrate the occurrence of a maxillary nerve similar to that of more typical genera. It was, therefore, in an endeavour to elucidate this point that the present microscopical observations of sections of both larvae and young frogs were undertaken. As metamorphosis and its resultant morphological changes has little effect on the distribution of the cranial nerves, the following account is based chiefly on the condition in the larva. Certain changes are, of course, to be expected after metamorphosis, but as most of the remarks concerning the larval arrangement are in general applicable to the young frog, only the more striking variations between the two stages are indicated in the course of the discussion. As a point of interest it may also be mentioned that a consideration of the larval arrangement of the nerves is significant in that it clearly indicates the similarities between the cranial nerves of Xenopus and those of certain Urodela (Norris, 1913; Escher, 1925; Benedetti, 1933; Francis, 1934) as well as those of the larva of Eana (Strong, 1895). In the reconstructions shown in figs. 28, 29, PI. 16, such nerves as the opticus, oculomotorius, trochlearis, and abducens have been omitted, as their inclusion would render the diagrams unnecessarily complicated. All these nerves except the abducens, which on account of its proximity to the ramus ophthalmicus profundus V is difficult to indicate separately, are shown in figs. 27 a, c, d, PI. 15, and also in various diagrams of transverse sections.

39 HEAD OF XENOPTJS 199 Ganglion Pro-oticum. Approximately one-third of the way from the anterior end of the medulla oblongata the facial and auditory nerves arise together as a large lateral mass. This close association of these two nerves is characteristic of Pisces and Amphibia, and is therefore to be expected in Xenopus. The two are not easily recognizable as separate nerves even in sections of larvae, but the fibres of the facialis issue more ventrally from the brain than those of the acusticus, and as they pass antero-laterally the former remains more mesial in position. Furthermore, as van der Horst (1934) has shown in a larva slightly older than Stage 6 of the present series, they may be distinguished from each other due to the fact that the fibres of nervus facialis are not so coarse as those of nervus acusticus. Immediately in front of this mass, the anterior lateralis (ant.lat.,figs.27 a, c, d, PI. 15;figs.28, 29, PL 16) emerges dorsolaterally by a single root, and runs forwards parallel and dorsal to the facialis nerve. In Eana the anterior lateralis (dorsal VII) emerges with the facialis even in the tadpole, but in the Urodela, as Strong (1895) has indicated, the two nerves are separate, so that in this respect Xenopus shows a closer approximation to the Urodele condition. The nervus trigeminus arises just posteriorly to the separation of the facialis from the acusticus, and the three bundles of fibres, anterior lateralis, trigeminus, and facialis, run anteriorly parallel to each other, remaining quite distinct until about the level of the anterior part of the mesencephalon, where they become approximated to form the large ganglion pro-otieum (g.pro.,figs.27 a, b, d, PI. 15;figs. 28,29, PI. 16). At the anterior border of the cerebellum, and just before the ganglion prooticum is formed, the anterior lateralis divides into two (d.al., v.al.,fig.17, PI. 12). The fibres in the dorsal part (d.al.) are not joined by any otherfibresfrom the ganglion pro-oticum, but remaining distinct, they continue anteriorly as the truncus supraorbitalis (ts., fig. 16, PI. 12), which, therefore, in Xenopus is composed of only lateral line fibres. The fibres in the ventral portion of the anterior lateralis

40 200 NELLIE F. PATERSON (v.al., figs. 16, 17, PI. 12) have a closer association with the other components of the ganglion pro-oticum. They extend laterally along the nervus trigeminus and intermingle with the fibres of the nervus facialis, just prior to or coinciding with the merging of these two nerves in the ganglion pro-oticum. A similar division of the anterior lateralis nerve (dorsal VII) into dorsal and ventral parts has been described in Eana by Strong (1895) and in Siren by Norris (1913), in both of which, as in Xenopus, the dorsal half maintains a dorsal position above the trigeminus while the ventral half passes into the nervus facialis, supplying the lateral line fibres to the ramus hy omandibularis. In transverse sections the ganglion-pro-oticum (g.pro., figs. 15, 16, 18 c, PI. 12) appears as a large oval mass, in which it becomes increasingly difficult from behind forwards to distinguish between the different fibres. In anterior sections of larvae, and especially after the separation of the truncus supraorbitalis from the ganglion, the fibres are indistinguishable, a fact also noticed by van der Horst (1934) in his preparations. It must be remarked, however, that in young frogs, where the ganglion pro-oticum is large but relatively not so elongate as in the larva, the various component fibres are more easily traced, probably due to the fact that the nerves separate sooner and almost simultaneously from the ganglion. Nervus Trigeminus. 1. Eamus Mandibularis. Norris (1913) has observed that in Siren the ramus mandibularis V emerges through a foramen which is common to it and to the anterior lateral line nerve. In Xenopus a slight difference was observed in this connexion between larval and adult conditions. Whereas in the larva the ramus mandibularis V and the truncus supra-orbitalis pass out through separate foramina, which, however, are situated close to each other, in the young frog (fig. 18 a, PL 12) the nerves, after separating from the ganglion pro-oticum, are directed anteriorly in the cranial cavity and again become approximated, so that they

41 HEAD OF XENOPUS 201 emerge together through a single foramen nearly opposite the origin of the nervus opticus from the brain. Strong (1895) has shown that in E a n a the rami ophthalmicus profundus V and maxillo-mandibularis are ganglionated as they leave the Gasserian ganglion, and it is considered that in Xenopus an essentially similar condition obtains. The ramus mandibularis V is the first to separate from the ganglion prooticum after the dispersal of the anterior lateral line fibres, and it is slightly ganglionated as it proceeds dorso-laterally to its foramen. The ganglion of the ramus ophthalmicus profundus V remains in longer association with the nervus facialis, with which it emerges ventrally from the chondrocranium {of.-\-hym.-{-'pal.,fig.14, PI. 12). Thefibresof the three nerves, ramus ophthalmicus profundus V, ramus hyomandibularis VII, and ramus palatinus VII, are, however, easily determined and separate shortly after issuing from the chondrocranium (fig. 13, PI. 12). In the larva the ramus mandibularis V (r.man.,fig.15, PL 12) separates from the ganglion pro-oticum almost in the same transverse plane as the division of the truncus supra-orbitalis into the ramus ophthalmicus superficialis and truncus infraorbitalis. Shortly after emerging from the chondrocranium it gives off a branch (md^,fig.12, PI. 12;fig.28,P1.16) which passes out laterally in the direction of the truncus infra-orbitalis (io.). It applies itself closely to the latter, and passing beneath it, proceeds ventro-laterally over the dorsal end of the constrictor branchialis II muscle and innervates some of the neuromasts. In the larva of Eana, Strong (1895) observed three accessory trigeminal branches which effected temporary fusions with the dorsal anterior lateral line nerve. These accessory nerves of Eana are stated to arise from a 'few large ganglion cells in the dorsal and mesial side of the trunk of the V, constituting the apex of the Gasserian ganglion' (p. 111). In Xenopus larvae this ramulus of the mandibularis V seems to be directly comparable with the middle accessory nerve of Eana, for not only does it come into contact with the lower ramus of the lateral line nerve as in Eana, but it also arises dorsally from the ramus mandibularis V which itself constitutes the upper part

42 202 NELLIE F. PATEESON of the ganglion pro-oticum after the truncus supra-orbitalis has separated from it. In young frogs a similar nerve (md. y,fig.29, PL 16) is closely associated with the truncus infra-orbicalis, to which it runs parallel and then ends in the tissues above the middle ear. The main stem of the ramus mandibularis V (r.man.,fig.13, PL 12) continues forwards and outwards in a deep lateral groove of the chondrocranium, and opposite the optic foramen it gives off another small branch (md. 2,fig.28, PL 16) which on its course anteriorly passes laterally and ventrally to the floor of the orbit. It is located immediately below the truncus infra-orbitalis, with which, however, it does not seem to anastomose. Owing to its position and distribution it may nevertheless be comparable with the innermost accessory trigeminal branch of Ban a. In front of the orbit the ramus mandibularis V broadens out over the dorsal surface of the musculus levator mandibulae anterior. More anteriorly, in the deep lateral groove between the commissura quadrato-cranialis anterior and the processus muscularis of the palatoquadrate, the ramus mandibularis V (r.man.,fig.7, PL 11) lies between the musculus levator mandibulae posterior and the levator muscle of the tentacle. While in this position the nerve supply to the levatores mandibulae is given off from the ramus mandibularis (cf. fig. 7, PL 11). Anteriorly, when it is situated between the quadrato-ethmoidal (q.e.c.) and Meckel's cartilages (Meek.), the ramus mandibularis V (r.man., fig. 6, PL 11) divides into several branches. One branch (md. 3,fig.28, PL 16) passes ventro-laterally to the skin, and is perhaps to be compared with the ramulus labialis of Siren (Norris, 1913). Two other small branches later re-unite to form the nerve (md.^ fig. 28, PL 16) to the tentacle. Of the remaining two branches, md. 6 passes latero-ventrally round the margin of the Meckel's cartilage and supplies the musculus intermandibularis, while md. 5 continues forwards lying dorsolaterally to the Meckel's cartilage. It divides anteriorly into several branches which pass over the inferior labial cartilages and innervate the tissues of the lower lip. These two nerves are identical with the rami mandibularis inferior and superior of P r o t e u s as described by Benedetti (1933), and also bear some

43 HEAD OF XBNOPUS 203 resemblance to the ramuli intermandibularis and mandibniarfs externus in Norris's account (1913) of the nerves of Siren. In the young frog the ramus mandibularis V is a very thick nerve, which after leaving the cranium is directed laterally and ventrally, and in passing over the museuli levatores mandibulae on its ventro-lateral course it supplies both these muscles. Its anterior branches (md.g_ s,fig.29, PL 16) are similar to those of the larva, with the exception that md. 5 and md. 6 are given off after the main ramus has passed to the ventral surface of the mandible, with the result that md. 5 is more ventrally situated than it is in the larva. From the foregoing it is evident that the ramus mandibularis V of X e n o p us is composed largely of general cutaneous fibres together with some motor fibres, and that it corresponds with the mandibular portion of the maxillo-mandibular stem in other Amphibia. As is shown in figs. 28, 29, PL 16, there is no separation of a ramus maxillaris from this stem as is usual in A n u r a, the supply to the upper jaw being drawn mainly from the ramus ophthalmicus profundus V. 2. Eamus Ophthalmicus Profundus V. After the ramus mandibularis V has been given off from the ganglion pro-oticum, the mixed fibres that remain form a large bundle which passes out ventrally from the chondrocranium (op.+hym.-\-j>al., fig. 14, PL 12). In this mass the large ganglionated cells of the ramus palatinus VII are distinguished ventrally, while the ramus hyomandibularis VII is lateral and the ramus ophthalmicus profundus V is located dorso-mesially. The three ganglia therefore emerge together, and in passing anteriorly are situated in a marked ventral groove (fig. 13, PL 12) in which they almost immediately separate from each other. In the young frog the arrangement is otherwise. All the nerves are dispersed almost simultaneously from the ganglion pro-oticum; the ramus palatinus VII (pa?.,figs.18 a, b, PL 12) passes directly through its own ventral foramen; the ramus hyomandibularis VII (jkym.) is directed antero-laterally between the floor of the auditory capsule and the quadrate. There then remain at different levels in the cranial cavity the ramus

44 204 NELLIE F. PATERSON ophthalmicus V (op.), the ramus mandibularis V (r.man.), and the truncus supra-orbitalis (ts.), which all emerge rather more anteriorly. The two latter, as has been explained previously, pass out through a common foramen; the ramus ophthalmicus profundus V has a separate foramen which is located immediately antero-ventrally to that for the ramus hyomandibularis VII. Immediately on leaving the cranial cavity, and before entering the orbit, the ramus ophthalmicus profundus V of the young frog becomes associated with the nervus oculomotorius in such a way that it passes between the two branches of III. In the larva this association occurs more anteriorly in the orbit as it also does in the Salamander (Francis, 1934). The additional anastomoses between the ramus ophthalmicus profundus V and the nervi oculomotorius and trochlearis, shortly to be described, are more easily followed in sections of larvae than in preparations of young frogs. The ramus ophthalmicus profundus V of the larva is at first situated ventrally as in Eana (Strong, 1895). It is a stout nerve composed of general cutaneous fibres, and on its course anteriorly it gradually assumes a more dorsal position. At first on entering the orbit (op.,figs.9, 10, PI. 11) it may be observed above the subocular fenestra and in more anterior sections it occurs near the commissura quadrato-cranialis anterior. The position of this nerve is essentially similar to that described in Siren by Norris (1913), where it at first runs in the gap between the orbito-sphenoid cartilage and the base of the petrosal, and more anteriorly it lies between the orbito-sphenoid cartilage and the temporal muscle. It also resembles the profundus nerve of the Salamander, which is described by Francis (1934) as lying mesially to the levatores mandibulae muscles as it passes forwards to the orbit. In the larval Xenopus, while it is situated over the subocular fenestra and still some distance behind the eyeball, the ramus ophthalmicus profundus V becomes closely connected with the nervi oculomotorius and abducens. The latter, which as in most Amphibia arises ventrally in the same transverse plane as the IX-X complex, is in close contact with the ramus ophthalmicus profundus V when it leaves the ganglion pro-

45 HEAD OF XBNOPUS 205 oticum. They run forward together, only separating in the orbit, where the abducens is observed as a very fine nerve (VI, figs. 9, 10, PL 11) which proceeds for some distance ventrally to the rectus externus eye-muscle before entering it. Francis (1934) and Edgeworth (1935) both mention that in Amphibia the abducens also supplies the museulus retractor bulbi. In Xenopus, owing to its fineness, the abducens is difficult to trace beyond the rectus externus muscle, but in older larvae and young frogs a few of its fibres seem to extend as far as the retractor bulbi, so that it is distributed similarly to that of more typical forms. After emerging through its own separate foramen the nervus oculotomorius of the larva travels outwards and forwards until it lies immediately dorsal to the ramus ophthalmicus profundus V. Asinmany other Amphibia it then divides (III,fig.10, PL 11; fig. 27 a, PL 15), and its two branches apply themselves closely to the ramus ophthalmicus profundus V lying in its bifurcation. The superior branch of the oculomotorius passes to the superior rectus eye-muscle (r.s.,fig.9, PL 11) entering it ventrally. Immediately this branch of the oculomotorius has passed over the dorsal surface of the ramus ophthalmicus profundus V, the latter gives off a short dorsal twig which anastomoses with a downwardly directed branch of the nervus trochlearis. The inferior branch of the oculomotorius is the stouter and associates with a ventral branch of the ramus ophthalmieus profundus (pp. v fig. 28, PL 16) with which its fibres intermingle before the two nerves again separate. The inferior ramus of III then supplies the recti internus and inferior and the obliquus inferior eye-muscles. The delicate branch of the ramus ophthalmicus profundus V mentioned above (op. x ) divides into two even finer branches, one of which anastomoses with a twig from the nervus trochlearis, while the other, following the course of the nervus opticus, enters the eyeball. Norris (1913) has also described a similar association of the oculomotorius and a branch of the ramus ophthalmicus profundus V, the ramulus ophthalmicus profundus minor, and suggests that it probably constitutes a superior ciliary nerve. In Siren, however, this nerve is a dorsal branch of the ramus ophthalmicus profundus V, whereas

46 206 NELLIE F. PATEKSON in Xenopus the branch which anastomoses with the oculomotorius passes off ventrally from the main stem. This portion of the larval nervous system is perhaps more consistent with the arrangement observed in Eana (Gaupp, 1899) or in the Salamander, in the latter of which Francis (1934) has described a ramus communicans from the oculomotorius to the ramus ophthalmicus profundus V, and close to its point of origin a ciliary ganglion is formed. Francis further states that a twig from the ramus inferior III passes among the fibres of the musculus retractor bulbi and may continue to the eye, entering the ' sheath surrounding the optic nerve'. In the Salamander this nerve constitutes an inferior ciliary, a superior ciliary nerve also being present. In Xenopus the superior ciliary nerve was not observed, and although a ciliary ganglion is not evident along the course of the nerve that enters the eyeball, it is concluded that it is comparable with the inferior ciliary of the Salamander and of typical Anura. Shortly after entering the orbit, the ramus ophthalmicus profundus V gives off a thin dorsal branch (op. 2,fig.28, PI. 16). This divides into at least three very fine twigs, the ventral one uniting with the nervus trochlearis, while the other two supply the subcutaneous tissue behind and above the eye. From its origin and distribution it would seem that this nerve corresponds with part of the ramulus ophthalmicus profundus minor of Siren, where according to Norris it supplies the skin around the eye and also comes into contact with the nervus trochlearis. The main ramus ophthalmicus profundus V pursues an antero-dorsal course, and in front of the anterior limit of the brain it divides into at least four thin but quite obvious branches. The most dorsal of these (op. A, fig. 28, PL 16) proceeds laterodorsally over the chondrocranium and enters the nasal capsule as shown in fig. 8, PI. 11. This constitutes the ramulus nasalis internus, which in the larva passes dorsally over the olfactory nerve towards the middle line. Anteriorly it passes to the inner side of the olfactory sac and gives off a small ventral branch which pierces the ethmoidal cartilage and ends in the connective tissue below. The main ramulus nasalis internus is continued

47 HEAD OF XENOPTJS 207 anteriorly in front of the nasal region and terminates in the tissues of the snout. The ramulus nasalis internus arises similarly in the young frog. It may be seen entering the nasal capsule in fig. 5, PL 10 (int.nas.), and shortly afterwards it divides into two. The thinner lateral branch passes forwards and outwards over the dorsal surface of the cavum principale and leaves the nasal capsule (int.nas.!, fig. 3, PL 9), passing between the cartilago obliqua and the nasale. Anteriorly to this it lies dorsally to the septomaxillare and ends in the plica obliqua. The medial branch (inf.n«s. 2,fig.3,P1.9; fig. 4, PL 10;fig.29,P1.16) is a stout nerve, which after passing dorsally and mesially around the cavum principale lies between the septum nasi and glandula nasalis medialis. As in Eana (Gaupp, 1899) it makes its exit from the nasal capsule by the fenestra naso-basilis and comes into contact with the glandula intermaxillaris. It passes over the superior pre-nasal cartilage and premaxilla (int.nas. 2,figs.1, 2, PL 9) before dividing into several branches which terminate in the glandular skin at the extremity of the snout. In some young frogs in addition to the above, the glandula nasalis medialis was observed to receive a supply from a fairly conspicuous ventrally directed branch of the ramulus nasalis internus before it divided into its lateral and medial branches. The other three terminal branches of the ramus ophthalmieus profundus V of the larva (op. 3, op. 5, and op. s, fig. 28, PL 16) pursue a more or less parallel course anteriorly, there being temporary associations between them as is indicated in fig. 28, PL 16. Judging by their distribution they are to be correlated with the terminal ophthalmic branches of Siren (Norris, 1913) and Proteus (Benedetti, 1933) rather than with those of the Anura. In comparing them with the nerves of Siren, op. 3 and op. 4 of the latter are respectively op. a and op. 5 of Xenopus, in which case op. 6 is identified as a ramulus nasalis externus. A ramulus nasalis externus has been observed by Benedetti (1933) to be similarly separated from the ramus ophthalmieus profundus V in Proteus, and in all three genera it innervates the skin lateral to and in front of the nasal region. It is to be noted, however, that in the majority of

48 208 NELLIE F. PATBRSON Amphibia the ramulus nasalis externus gives off a branch which enters the nasal capsule, but this was not observed inxenopus. After the anastomosis between the ramulus nasalis xternus and op. 5 has been effected, a thin nerve (op. 7, fig. 28, PI. 16) separates ventrally, and passing directly downwards over the anterior margin of the commissura quadrato-cranialis anterior, it becomes associated with an anterior branch of the ramus palatinus VII below the ethmoidal cartilage. This nerve is probably merely an anterior contmuation of op. 5, and the whole nerve is to be regarded as homologous with op. 4 of Siren, that is, with the ramulus palatinus profundus. This anastomosis between the ramus ophthalmicus profundus V and the ramus palatinus VII is even more obvious in Xenopus during metamorphosis, when the ramulus (op. 7, fig. 29, PL 16) forms a stout nerve passing obliquely through the antorbital cartilage to join the ramus palatinus VII. In sections of young frogs the ramulus communicans ad VII was more difficult to trace. A thin nerve was given off as in metamorphosing specimens, but its connexion with the ramus palatinus VII was not determined with certainty. The remaining branch (op. 3, fig. 28, PL 16) is considered to represent the ramus maxillaris. In Anura this branch is usually a well-defined nerve originating from the maxillomandibular trunk, but in Urodela its origin is variable. In Siren Norris (1913) has ascertained that the truncus infraorbitalis is composed of general cutaneous fibres in addition to the lateral line fibres. The general cutaneous fibres which comprise the ramus maxillaris are therefore separated from the truncus infra-orbitalis in Siren; the lateral linefibres innervate the infra-orbital neuromasts as does the truncus infra-orbitalis in Xenopus. The arrangement in Xenopus seems to approach more closely that of Proteus, in which Benedetti (1933) has found that the truncus infra-orbitalis consists only of lateral line fibres, the general cutaneous fibres which form the ramus maxillaris being derived from the ramus ophthalmicus profundus V, in a manner similar to the course of op. 3 in Xenopus. The arrangement in the larval Salamander is in essentials similar to that of Siren, the ramus buccalis

49 HEAD OF XENOPUS 209 being connected with the lateral line nerve, but in the adult (Francis, 1934) the ramus maxillaris resembles that of Anura. It is therefore evident that in the larval stage the arrangement of the branches of the ramus ophthalmicus profundus V of Xenopus is intermediate between that of Siren and Proteus. All the main ophthalmic profundus branches described by Morris are present in Xenopus, but in addition the ramus maxillaris occurs as a branch of the profundus as it is in Proteus. A comparison between the larval arrangement and that occurring during metamorphosis is also of some interest. From fig. 29, PL 16, it is apparent that the ramuli ophthalmicus profundus minor (op.^), the nasalis intemus (op.4), and the ramulus commum'cans ad VII (op. 7 ) arise from the main ophthalmic trunk much as they do in the larva. The ramulus nasalis extemus (op. e ) and the ramus maxillaris (op. 3 ) are situated closer together, the latter separating in front of the branch anastomosing with VII, and not posterior to it as in the larva. The Harderian glands which develop during metamorphosis are innervated by a special nerve (op. 8,fig.5, PL 10; fig. 29, PL 16) which passes laterally from the main ramus immediately anterior to the ramulus nasalis intemus. Nervus Facialis. As in many Anur a, the geniculate ganglion is not separated from the Gasserian ganglion, but it becomes closely associated with the trigeminus and forms the ventral part of the conspicuous ganglion pro-oticum. As has been previously observed, the nervus facialis emerges from the medulla oblongata with the acusticus, from which it separates as a small but distinct bundle passing forwards between the brain and the inner wall of the auditory capsule (VII,fig.17, PL 12). In sections of both larvae and young frogs the nervus facialis (VII, figs. 28, 29, PL 16) may be traced for at least 6G0-700JU. as a separate nerve lying immediately ventrally to the nervus trigeminus. It eventually applies itself to the ventral surface of the latter after receiving some lateral line fibres from the anterior lateralis nerve. Just as the latter is identical with the dorsal VII of the larva of NO. 322 P

50 210 NELLIE F. PATEESON Eana, the ventral nerve which emerges with the nervus acusticus corresponds with the ventral VII in Strong's description of Eana (1895). Eegarding the derivation of the general cutaneous fibres of the facialis nerve in Amphibia, it is a generally accepted fact that these enter the facialis by way of the ramus communicans. This is probably also their usual origin in Xenopus, for a ramus communicans occurs in both larvae and adults. It has also been suggested by van der Horst (1934) that 'in the ganglionic complex fibres from the trigeminus might join the hyomandibular nerve, and the same might occur where the greater part of the sensory facialis fibres pierces through the descending trigeminus, though it would be difficult to prove this'. Norris (1913) finds that in Siren the ramus jugularis VII may receive general cutaneous fibres from the ramulus malaris of the ramus mandibularis V, and also that there is a small bundle of general cutaneous fibres entering the brain near the motor root of VII. It has already been proved by van der Horst (1934) that the latter does not occur in Xenopus, and during the course of the present investigation no anastomoses were observed between the ramus mandibularis V and the ramus hyomandibularis VII. It therefore seems that in Xenopus, apart from van der Horst's suggestion, the only apparent source of the general cutaneous component of the nervus facialis is through the ramus communicans. 1. Truncus Supra-orbitalis. Xenopus, in common with other Anura, has only a single root to the anterior lateralis nerve, and this is regarded by van der Horst (1934) as being comparable with the ventral root of the lateral line nerve of Pisces and Urodela. The anterior lateralis nerve is quite distinct, and in the foregoing remarks attention has already been drawn to its division into dorsal and ventral parts in the vicinity of the ganglion prooticum. While the fibres of the ventral half lose their identity in merging with the nervus facialis, the dorsal division remains distinct and is easily traced anteriorly. In the ganglion prooticum of both larva and adult it maintains a dorsal position

51 HEAD OF XENOPUS 211 above the trigeminus, and there is no indication of its being joined by any cutaneous fibres from the latter. The supra-orbital trunk is the first to separate from the ganglion pro-oticum. It emerges through a wide dorso-lateral foramen just anterior to the mesencephalon (ts.,fig.16, PL 12). As it passes outwards it divides sharply into dorsal and ventral nerves. The dorsal branch (so., figs. 9-11, PL 11; figs , PL 12) represents the ramus ophthalmicus superficialis of the larva of Eana (Strong, 1895) and of Siren (Morris, 1913). It passes anteriorly maintaining a dorsal position throughout. It gives off short posterior and anterior nerves to the skin behind the nasal region, and anteriorly it occurs above the terminal branches of the ramus ophthalmicus profundus V, innervating the supra-orbital series of sensory organs. The other branch of the supra-orbital trunk (io.,figs.12-15, PL 12) passes in a ventral direction until it reaches the lateral neuromasts. It then pursues an antero-lateral course immediately below the eye, giving off a series of twigs to the infraorbital sensory organs and terminating in the neuromasts of the upper jaw. There is no doubt as to its identity with the lateral line part of the truncus infra-orbitalis of Siren (Norris, 1913). It is also homologous with the ramus bucealis of the larva of Eana, which has been shown by Strong (1895) to be composed of lateral line fibres innervating the infra-orbital series of sensory organs. 2. Eamus Hyomandibularis VII. After the ramus mandibularis V has separated from the ganglion pro-oticum, there remain mixed fibres received from the anterior lateralis, the trigeminus, and facialis nerves. The lateral line fibres received from the ventral division of the anterior lateralis are difficult to trace in the ganglion pro-oticum of the larva, but in the young frog their passage is more easily determined. After spreading laterally over the trigeminus {v.al.,fig.16, PL 12) they are directed medially and eventually pass into the ramus hyomandibularis VII, which, therefore, as in Eana (Strong, 1895), in addition to its motor and general cutaneous fibres, also receives fibres from the lateral line

52 212 NELLIE P. PATERSON system. The rami ophthalmicus profundus V and palatinus VII receive no fibres from the anterior lateralis nerve, the former being composed of general cutaneous fibres and the lattei of communis fibres. In the larva the ramus hyomandibularis VII is a broad nerve (hym., fig. 13, PI. 12) which separates laterally from the ganglion pro-oticum. As it assumes a more lateral position (hym.,fig. 11, PL 11;fig.12, PI. 12;fig.27 b, PL 15) the head vein (h.vn.) runs between it and the more medially placed ramus palatinus VII (pal.). The ramus hyomandibularis VII proceeds in an anterolateral direction, lying at first immediately below the subocular bar of the palatoquadrate where it is joined by the ramus communicans from the IX-X complex. In front of this it takes up a position ventro-lateral to the subocular bar (hym.,fig.9, PL 11), and when the latter joins the processus muscularis, the ramus hyomandibularis VII is located above the lateral part of the cerato-hyale. While in this position, it lies mesially to the orbito-hyoideus muscle, to which it sends at least two small branches. In passing round the articular surface of the ceratohyale, the ramus hyomandibularis VII divides into two. The more ventral branch (r.jug.,fig.28, PL 16) corresponds with the ramus jugularis of Siren (Norris, 1913) and of Salamandra (Francis, 1934). It passes ventrally close to the inner surface of the orbito-hyoideus muscle to which it supplies a few fibres. It also gives off a nerve which passes forwards to the quadratohyoangularis muscle, the main stem (r.jug., figs. 7, 8, PL 11) then proceeding ventrally between the orbito-hyoideus muscle and the lateral surface of the cerato-hyale, eventually applying itself to the ventral surface of the musculus interhyoideus. The dorsal nerve (r.ment, figs. 7, 8, PL 11) lies just ventrolaterally to the processus muscularis of the palatoquadrate (mus.pr.), between the orbito-hyoideus and quadrato-hyoangularis muscles. On a level with the origin of the former muscle on the processus muscularis, it divides into two. One branch (m.ext.,fig.6, PL 11) passes dorsally along the quadratohyoangularis muscle (dep.man.) and divides anteriorly to supply the oral sensory organs near the base of the tentacle. The other branch (mini.,fig.6, PL 11) runs out laterally until it lies beneath

53 HEAD OF XENOPUS 213 the skin, to which it gives off a few fibres. As it proceeds forwards it gradually becomes more ventral in position and subsequently enters the ventral surface of the intermandibular muscle, after having supplied the gular series of sensory organs. Prom the foregoing it is evident that in Xenopus the ramus hyomandibularis VII is composed of lateral line, motor, and general cutaneous fibres, the homologue of the ramus mandibularis internus VII described and figured in Ban a by Strong (1895) being apparently unrepresented. The arrangement of the nerves at this stage is perhaps to be compared with that obtaining in Siren, in which Norris (1913) has observed an obvious division of the truncns hyomandibularis VII into an anterior lateral line ramus mentalis and a posterior ramus jugularis composed of motor and general cutaneous fibres. This separation of the fibres is perhaps a little less obvious in Xenopus, but, judging by the distribution of the branches, there can be no doubt that the more ventral one (r.jug.) which supplies the orbito-hyoideus, depressor mandibulae, and ultimately the interhyoideus muscles, corresponds with the ramus jugularis of Siren. The more anterior and dorsal nerve is homologous with the ramus mentalis of Siren, and of its main branches the one supplying the oral sensory organs (m.ext.,fig.6, PI. 11; fig. 28, PL 16) represents the ramus mentalis externus, while the rather more postero-ventral branch (m.int., fig. 6, PL 11;fig.28, PL 16) ending in the intermandibular muscle is comparable with the ramus mentalis internus of Siren. In this connexion it may also be noted that Escher (1925) in describing the sensory organs and their innervation in Anura, although agreeing with Strong's interpretation of the composition of the nerves, has referred to the anterior and posterior parts of the ramus mandibularis externus of Bana as the rami mentalis externus and internus respectively, thus bringing them more into fine with the Urodele condition. The arrangement of the ramus hyomandibuiaris VII of metamorphosing larvae and young frogs (fig. 18, PI. 12;fig.29, PL 16) is perhaps more easily correlated with that of the larva of Bana. As in the larva the ramus hyomandibularis VII of young Xenopus is found to be composed largely of lateral-line

54 214 NELLIE F. PATEBSON fibres. Owing to the migration of the suspensorial region at metamorphosis, it passes backwards for a short distance along the lateral wall of the auditory capsule, and is there joiued by the communicating nerve from the IX X complex. It is then directed ventrally and is located mesially to the depressor mandibulae muscle, its relation with this and other muscles being essentially similar to the larval nerve. It divides into a stout ramus jugularis to the subhyoideus (interhyoideus) muscle and a ramus mentalis, the distribution of which differs markedly from that of the larva. After a short anterior course the ramus mentalis gives off a branch which supplies the skin (cut, fig. 29, PI. 16). Almost immediately it divides into anterior and posterior nerves, the arrangement being similar to that of the larva of Ran a as described and figured by Strong (1895). The anterior nerve of the adult Xenopus corresponds with the ramus mentalis externus of the larva, while the longer posterior branch (m.int, fig. 29, PI. 16) is the ramus mentalis internus of the larva. Both of them innervate the anterior and posterior ventral sensory organs, the motor fibres observed in the ramus mentalis internus of the larva, not being observed in the young frog. 8. Ramus Palatinus VII. In Xenopus the fibres comprising the ramus palatinus VII become markedly ganglionated and form the ventral part of the ganglion pro-oticum. In the larva the nervus palatinus VII, however, after separating from the ramus hyomandibularis VII and ramus ophthalmicus profundus V (fig. 11, PI. 11; fig. 12, PL 12; fig. 27 i, PL 15) just posterior to the subocular fenestra, is a very delicate nerve. In the adult it is much thicker and passes directly downwards through its ventral foramen. An indication has previously been given of the important relation between the course of the palatinus VII and the conformation of the subocular region of the chondrocranium, and reference has been made to the fact that, as there is no basal process, the palatinus VII does not become enclosed in the subocular shelf. On leaving the chondrocranium it lies in a marked ventral groove together with the ramus hyomandibularis VII and the ramus ophthalmicus

55 HEAD OF XENOPUS 215 profundus V. It passes below the chondrocranium and over the dorsal surface of the internal carotid artery (pal, fig. 27 b, PL 15), but when the latter courses mesially to give off the cerebral artery, the ramus palatinus YII is situated laterally to it. The nerve pursues a more or less straight course anteriorly, giving off twigs to the dorsal wall of the pharynx and mouth. It is not located near any of the masticatory muscles as is the similar nerve in Siren (Norris, 1913). As in this latter genus, the ramus palatinus VII divides into two main branches, one of which is medial (jpal. v figs. 28, 29, PL 16) and the other lateral (pal. 2 ) in position. The latter is located just beneath the subocular fenestra in the larva, while the medial branch runs below the floor of the chondrocranium and supplies the anterior part of the mouth. In Siren both nerves communicate with a branch of the ramus ophthahnicus profundus V, but, as has been mentioned in connexion with the distribution of the ramus ophthahnicus profundus V, in Xenopus there is only one such anastomosis. In both larva and adult this fusion is effected with a branch of the lateral palatine nerve immediately behind the internal naris. In Ran a Bender (1906) describes a similar anastomosis taking place between the ramus median's, however, and the ramus maxillaris superior V. In the sections of Xenopus larvae the medial branch was observed to terminate behind the lateral, but in young frogs it was traced anteriorly to the floor of the nasal capsule (jpal.^fig.4, PL 10), where some of its fibres innervate the glandula intermaxillaris (g.i.m.). Possibly there is also a further fusion of the ramus ophthahnicus profundus V and ramus palatinus VII fibres in the vicinity of this gland, for both nerves send off short branches which ramify among the gland. Such a contact would be comparable with the anastomoses of the medial palatine nerve of the Salamander described by Francis (1934). The other palatine branches observed by Norris in Sir en are apparently unrepresented in Xenopus, but this is not remarkable, for, as Francis (1934) shows, they are not always present even in Urodela. The palatine nerve seems to be one of the least variable of the Amphibian cranial nerves, and its distribution inxenopus

56 216 NKLLIE P. PATBESON is quite typical of the phylum. It shows some resemblance to that of Eana (Strong, 1895; Gaupp, 1899), but its ramifications are less complicated. It differs from that of Prot JUS, investigated by Benedetti (1933), in having two and not one main branch, but it shows a close approach to the arrangement observed in Siren by Norris (1913). Brief mention has already been made of the absence of communis fibres in the truncus hyomandibularis of Xenopus. Most investigators have recorded their presence in one of the branches of this nerve. In Urodela they constitute the ramus alveolaris, which is generally identified with the ramus mandibularis internus of Anura (Strong, 1895). This latter author and others such as Bender (1906) and Francis (1934) have considered' the possible homologies of this nerve and have concluded that it corresponds with the chorda tympani of mammals. It is, therefore, rather remarkable that such a nerve, which is so universally found, should not have been observed in Xenopus. There is no accounting for this absence of communis fibres in the ramus hyomandibularis VII of Xenopus, but it is to be noted that, taking the nerves as a whole, and especially those of the V-VII complex, there is a very obvious segregation of the different types of fibres. Thus, in each of the nerves certain fibres predominate. The supra-orbital trunk is a derivative of the lateral line nerve; the ramus ophthalmicus profundus V is composed of general cutaneous fibres; the ramus mandibularis V is a mixture of cutaneous and motor fibres; and the ramus hyomandibularis VII, which is largely composed of lateral line fibres, also contains motor and cutaneous elements. These points are perhaps of little significance, but the fact nevertheless remains that as far as can be ascertained from the present series of preparations of Xenopus, the ramus palatinus VII is the only source of communis fibres in the nervus facialis. Nervi Glossopharyngeus a d Vagus. The posterior lateralis, the glossopharyngeus, and vagus emerge separately from the brain, about two-thirds of the way along the medulla oblongata. The first of these has been described and figured by van der Horst (1934), who shows that

57 HEAD OF XENOPUS 217 it enters the brain by many separate bundles which therenpon break up into separate fibres. It lies dorsally in about the same horizontal plane as the lateralis anterior. It passes outwards and ventrally, joining the glossopharyngeus before emerging from the chondrocranium. With these two nerves the vagus roots also unite, and the large bundle of fibres which results passes out through the wide foramen jugulare below the posterior border of the auditory capsule. Outside the cranium the fibres form a large ganglionie complex (g.v.,figs.27 a, c, PL 15; figs. 28, 29, PL 16) in which it is difficult to determine the various fibres. The posterior lateralis fibres remain dorsal in position in the glossopharyngeal-vagus ganglion, and in both larva and adult are the first to separate from the ganglion. The posterior lateral-line nerve divides into anterior and posterior branches to supply the trunk series of sensory organs, the anterior rami extending to the neuromasts posterior to those innervated by the truncus infra-orbitalis. In both larva and adult the posterior branch divides into two main lateral-line trunks (figs. 28, 29, PL 16), those in the larva being continued into the caudal region. One small dorsally directed lateral-line nerve (r.temp., figs. 28, 29, PL 16) is doubtless comparable with the supratemporal nerve of Urodeles (Norris, 1913; Escher, 1925). 1. Nervus Glossopharyngeus. After passing into the glossopharyngeus vagus ganglion, the fibres of IX are difficult to distinguish from those of X. They are somewhat finer, and separate from the vagus shortly after the lateralis posterior fibres have been given off. The nervus glossopharyngeus is directed anteriorly and is situated in a ventro-lateral groove of the auditory capsule below the crista parotica. It soon becomes ganglionated (g.gl., figs. 27 a-c, PL 15; figs. 28, 29, PL 16), and from the ganglion the ramus communicans to the ramus hyomandibularis VII (IX ad VII, fig. 27 a, PL 15; figs. 28, 29, PL 16) separates dorsally. It takes a medial course and unites with the main stem of the ramus hyomandibularis VII before the latter divides into the rami jugularis and mentalis. This arrangement seems to be

58 218 NELLIE F. PATERSON typical of the Anura, but in the Urodela (Norris, 1913; Benedetti, 1933; Francis, 1934) the ramus communicans joins the ramus hyomandibularis VII very close to its dh ision into its main branches. In Urodela, therefore, the general cutaneous fibres pass almost directly into the ramus jugularis, but in Anura they mix with the other fibres composing the main stem of the ramus hyomandibularis VII. As Strong (1895) has shown in E a n a, the general cutaneous and motor fibres remain ventral and pass into the ramus jugularis, while the ramus mentalis, although receiving some motor and cutaneous fibres, is usually largely composed of lateral line and communis fibres. A short distance in front of the separation of the ramus communicans to the ramus hyomandibularis VII, the nervus glossopharyngeus in the larva divides into two main branches (IX, figs. 13, 14, PL 12; fig. 27 a, PL 15; fig. 28, PL 16) which run parallel for some distance anteriorly, and on reaching the thymus gland (th.gl.,figs. 13,14, PL 12) are arranged one on each side of it. They are directly comparable with the pre- and posttrematic branches of the glossopharyngeus of Urodela, but their homologies with the nerves of Anura are rather more obscure. The dorsal one (jo.a.s., fig. 28, PL 16) resembles the ramus pharyngeus of the larva of Eana, but the ventral branch (f.a.i., fig. 28, PL 16) shows some considerable differences from the ramus lingualis IX, the course and composition of which has been studied by Strong (1895). Bender (1906) disagrees with Gaupp's interpretation (1899) of the branches of the nervus glossopharyngeus in the adult Eana, and recognizes a pharyngeus dorsalis IX, and a ramus lingualis, comparable with the pharyngeus dorsalis and post-trematic nerves of Pisces. Bender's findings are therefore in agreement with those of Strong, and it seems likely that the two nerves present in the larva of Xenopus are identical with those of Eana. Some differences in the distribution of the ventral nerve are, of course, to be attributed to the absence of a tongue in Xenopus. Apart from this, it is also to be noted that there is less mixing of the fibres in the glossopharyngeus of Xenopus than there is in Eana, a feature also noticed in connexion with the nervi trigeminus and facialis. Thus the dorsal or pre-

59 HEAD OF XENOPUS 219 trematic branch is composed of only communis fibres, agreeing in this respect with the ramus pharyngeus of Eana, and the post-trematic branch consists of motor fibres; in Eana the latter also contains general cutaneous and communis fibres. The post-trematic branch (p.a.i., fig. 28, Pi. 16) pierces the ventro-lateral process of the palatoquadrate (fig. 13, PI. 12), thus coming to lie (p.a.i., fig. 12, PL 12) ventrally to the truncus infra-orbitalis (io.), and with the constrictor branchialis muscle between them. The nerve then extends antero-ventrally, lying between the ventro-lateral process and the constrictor branchialis muscle, and passing through the fibres of the latter becomes more ventral in position. It continues anteriorly beyond the constrictor branchialis muscles and is located dorsally to the cerato-hyale. More anteriorly it passes into a more median position (IX, fig. 7, PL 11) above the musculus interhyoideus, into the dorsal surface of which it eventually enters. On its antero-ventral course this nerve also innervates the constrictor branchiales I and II, and the musculi subarcuales recti I and II. In the larva the dorsal branch of the nervus glossopharyngeus passes gradually in a medial direction supplying the roof of the pharynx. It gives off a branch which applies itself to the ventral surface of the ramus palatinus VII, but does not fuse with it. Its association with the palatinus is only of a temporary nature for it separates from it almost immediately and ends in the roof of the mouth. Norris (1913) has reviewed the possible communis anastomoses between the glossopharyngeus and facialis nerves in Urodela, and has shown that a Jacobson's anastomosis between the ramus pre-trematicus IX and the ramus palatinus VII is frequently established. A similar union of these nerves has also been described in Eana by Bender (1906). Owing to the fineness of the nerves in the larva of Xenopus, it could not be determined with certainty if the association of these nerves should be regarded as a Jacobson's anastomosis. There is certainly a very striking similarity between the arrangement in Xenopus and that of such Urodela as Siren (Norris, 1913), Proteus (Benedetti, 1933), and Salamandra (Francis, 1934), but, as no communication between the glossopharyngeus and facialis was observed in

60 220 NELLIE F. PATERSON young frogs, it seems doubtful if the larval connexion can be interpreted as a Jacobson's anastomosis. At metamorphosis the nervus glossopharyngeus is subjected to certain changes. The post-trematic branch (p.a.i., fig. 29, PI. 16) becomes reduced and passes over the thymus gland and crista parotica, ending in a few muscle-fibres, which are probably the remains of the musculus constrictor branchialis II. In the young frog it could not be found, and the musculus interhyoideus (= subhyoideus) is innervated by the ramus jugularis VII alone. The one main ramus of the glossopharyngeus in the young frog (IX,fig.29, PI. 16) represents the pre-trematicus of the larva, but its distribution is somewhat different. It is directed antero-ventrally, passing mesially to the cornu of the hyoid, and on reaching the lateral wall of the pharynx (IX, figs. 18 a-c, PL 12) it divides into three main branches. The distribution of these nerves seems to vary; in some specimens all three branches pass to the ventral wall of the pharynx and mouth: none of them innervates the dorsal wall of the pharynx, so that the communis fibres to it are furnished by the ramus palatinus VII only. In slightly older specimens, which are probably more representative of the adult condition, two of the three rami run to the dorsal wall of the pharynx, while the remaining branch is distributed ventrally. 2. Nervus Vagus. In both larval and adult Xenopus the roots of the vagus are quite distinct from those of the lateralis posterior and the nervus glossopharyngeus. It was observed that in the larva the vagus arose by a series of seven roots passing out laterally from the medulla oblongata. The roots are obviously arranged into three pairs, with the dorso-lateral root much more conspicuous than thefiner ventral root. The fourth dorsal root has no ventral counterpart. In young frogs there are only three dorsal and two ventral roots, the second pair of the larva, which is thinner than the others, having probably fused with the third. As in the larva there is no ventral root opposite the last dorsal root. The presence of these roots is of some significance, for it involves a consideration of the controversial question of the segmentation

61 HEAD OP XENOPUS 221 of the head. Kingsbury (1926), who has reviewed the whole problem, maintains that serial repetition of the branchial arches does not coincide with the somatic metamerism, and that the fifth to the tenth cranial nerves are not to be regarded as segmental nerves. None of the larval stages of Xenopus herein described was young enough to make a comparative study of the somites and cranial nerves possible, so that no attempt has been made to study the possible metamerism in Xenopus. There is, however, an obvious similarity between the arrangement of the nerve roots and those described by Goodrich (1918) and de Beer (1922) in Blasmobranchs. Both of these authors subscribe to the view of a segmentally arranged occipital region of the skull, de Beer concluding that in Squalus there are nine somites. In this connexion these authors show that the vagus nerve possesses four dorsal roots and that the ventral roots in the four vagus somites are drawn from the nervus hypoglossus, which in Blasmobranchs usually has no dorsal root. In Squalus de Beer (1922) has observed three roots belonging to somites 7-9 and states that 'since the eighth somite is the last of the vagus segments, the 9th is morphologically the 1st spinal or post-vagal' (p. 467). While it is beyond the scope of the present discussion to expatiate on the acceptance or otherwise of the theory of a segmentally arranged head, it is felt that some significance attaches to the similarity of the nerve roots in Xenopus and Elasmobranchs. This is further strengthened by the presence in young frogs of a connexion between the vagus and the hypoglossus. Tensen (1927) has observed that in Pip a the nervus hypoglossus has no dorsal root, but that there are two ventral roots. In Xenopus the hypoglossus arises similarly to Pip a in both larval and adult stages, the anterior ventral root being considerably weaker than the posterior. In tracing the distribution of the vagus in sections of a young frog, one branch was observed to communicate with the ganglion of the hypoglossus nerve. This connexion was not observed in the larva, not even during metamorphosis. It seems to develop after metamorphosis, and was too conspicuous to be in the nature of an individual variation. Strong (1895) remarks on a similar communicating branch between the vagus

62 222 NELLIE F. PATERSON and hypoglossus nerves in Eana. Following Addens's (1933) observations it seems to obtain in many of the lower chordates, and it is the general consensus of opinion that the hypoglo, sus may supply some of the motor elements to the vagus nerve, probably those to the last ventral root. The nervus vagus is a stout nerve, its distribution in larva and adult differing mainly in connexion with the branchial supply in the former. The branchial nerve (br.x., fig. 27 c, PI. 15; fig. 28, PL 16) also gives off a branch which passes outwards through the commissura branchio-cranialis and then runs ventrally, innervating the subarcuales recti III and IV, the transversus ventralis II, and also the constrictor branchiales III and IV. The visceral branch (vis.,figs.28, 29, PL 16) divides, as in other Amphibia, into branches to the alimentary tract (gas., fig. 29, PL 16), the lungs and the heart (cd.,fig.29, PL 16). The recurrent laryngeal nerve (rln.,fig.29, PL 16) is directed forwards and divides into branches which end in the laryngeal muscles. In addition to these, certain motor-nerves are observed in the adult to pass off anteriorly to the musculus cucullaris (mus. v fig. 29, PL 16), to the petrohyoideus (mus. 2 ), and to the muscles and also the skin in the region just anterior to the shoulder (mus. 3 ). The motor nerve to the musculus cucullaris is generally interpreted as the nervus accessorius. Addens (1933) has made a careful investigation of this nerve, and has found that in Gasterost e u s it arises from the beginning of the spinal motor column and not from the caudal end of the vagus column. The musculus trapezius (= cucullaris) is, therefore, innervated by the XI nerve, a fact that naturally is rather at variance with the assumption that this muscle is a derivative of one of the branchial muscles of the larva. Addens refers to an article by Vdlker (1908), who has shown that in Larus ridibundus 'the trapezius is formed by the fusion of split-off portions of the occipital and first cervical myotomes' (p. 342). Addens, therefore, supports the view that the trapezius is not a branchial muscle, and furthermore indicates that the nervus accessorius is derived from the first spinal nerve. The latter point is, perhaps, not always so clearly demonstrable as it is in G a s t e r - osteus, but apart from this, as has been previously indicated,

63 HEAD OF XBNOPUS it is the generally accepted opinion that some of the motor fibres in the vagus may originate in the anterior spinal nerve. In lower chordates the first spinal nerve is usually termed the hypoglossus; it may have no dorsal root, so that the second spinal is the first complete nerve of the spinal column. The nervus accessorius in most cases is not separable from the vagus as a distinct nerve. It seems to have merged with the vagus and merely appears as a motor branch passing off anteriorly from the main visceral branch of X. The interpretation of the composition and distribution of the main cranial nerves ofxenopus described in the foregoing may be summarized as follows: Nervus Trigeminus. 1. Eamus mandibularis: motor and general cutaneous fibres. (a) 1st ramulus (md.j): anastomoses with truncus infraorbitalis; comparable with middle accessory trigeminal branch of E a n a. (b) 2nd ramulus (md. 2 ): passes below truncus infraorbitalis and supplies floor of orbit. (c) 3rd ramulus (md. 3 ): to skin, comparable with the ramulus labialis of Siren. (d) 4th ramulus (md.^): to tentacle in larva. (e) 5th ramulus (md. 5 ): is the ramulus mandibularis superior to the lower jaw. (f) 6th ramulus (md. 6 ): is the ramulus mandibularis inferior to the musculus intermandibularis. 2. Eamus Ophthalmicus Profundus: general cutaneous fibres. (a) Eamulus ciliaris inferior (OJJ.J: associates with III and enters eyeball. (b) Eamulus ophthalmicus profundus minor (op. 2 ): innervates the skin above and behind the eye, and associates with IV. (c) Eamus maxillaris (op. 3 ): to upper jaw.

64 224 NELLIE F. PATBESON (d) Eamulus nasalis internus (op.^: (i) Lateral branch (int.nas.) passing into the plica obliqua. (ii) Medial branch (int.nas.) passing ventro-mediallj through nasal capsule and ending in skin of snout. (e) Eamulus communicans ad VII (op. s and op.j: anastomoses with lateral branch of ramus palatinus VII. (/) Eamulus nasalis externus (op. e ): to skin lateral to and in front of nasal region. Nervus Facialis. 1. Truncus supra-orbitalis (fe.): lateral line fibres. (a) Eamus ophthalmicus superficialis (so.): to all the supra-orbital sensory organs. (b) Truncus infra-orbitalis (io.): to the infra-orbital sensory organs. 2. Eamus Hyomandibularis Qiym.): lateral line, motor and general cutaneous fibres. (a) Eamus jugularis (r.jug.): to orbito-hyoideus, quadrato-hyoangularis, and interhyoideus muscles. (b) Eamus mentalis (r.ment.): (i) Eamus mentalis internus (r.int): to skin, gular sensory organs, and larval intermandibularis muscle, (ii) Eamus mentalis externus (r.ext.): to oral series of sensory organs, (iii) Eamus cutaneus (cut): as a separate nerve in the adult. 3. Eamus Palatinus (pal.): communisfibres. (a) Eamus palatinus medialis (pal.j): to roof of mouth. (b) Eamus palatinus lateralis (paz. 2 ): establishes a connexion with ramus ophthalmicus profundus V and ends in roof of mouth. Nervus Glossopharyngeus: communis and motor fibres. 1. Eamus communicans ad VII (IX ad VII): anastomoses with stem of ramus hyomandibularis VII.

65 HEAD OF XENOPUS Eamus pre-trematicus (p.a.s.): supplies roof of pharynx in larva, and dorsal and ventral walls of pharynx and mouth in adult. 3. Eamus post-trematieus (p.a.i.): present in the larva, innervating musculi constrictor branchiales I and II, subarcuales recti I and II and interhyoideus. Nervus Vagus-Nervus Accessorius: lateral line, communis, general cutaneous and motor fibres. 1. Lateralis posterior (gost.lat.): divides into anterior and posterior branches to the trunk sensory organs, the former giving off a ramus supra-temporalis. 2. Eamus branchialis (br.x): to branchial region and also to musculi constrictor branchiales III and IV, and subarcuales recti III and IV. 3. Truncus intestino-accessorius (vis.): (a) Eamus accessorius (wms.j): to musculus eucullaris. (b) Eamus muscularis (mus.^): to musculus petrohyoideus. (c) Eamus muscularis (m«s. 3 ): to skin and muscles of shoulder. (d) Main Stem dividing into: (i) Eamus laryngeus recurrens (rln.): passes forwards to laryngeal muscles and also to transversus ventralis IV. (ii) Eamus intestinalis (gas.): to alimentary tract, also gives off a pulmonary branch. (iii) Eamus cardis (cd.): to heart. NO. 322

66 226 NELLIE F. PATERSON TABULAE SUMMARY OF COMPARISON OF DISTRIBUTION OF MOTOR NERVES IN AMPHIBIA. MUSCLE. Levatores mandibulae anterior and posterior Orbito-hyoideus and quadrato-hyoangularis Intermandibularis Interhyoideus Constrictores branchiales: I and II III and IV Subarcuales recti: I and II III and IV Transversus ventralis: I IV Cucullaris Petrohyoideus Geniohyoideus Hyoglossus Laryngeal muscles Anura (Strong and Gaupp) R. mandibularis V R. hyomandibularis VII R. mandibularis V R. hyoideus VII R. accessorius IX and X Hypoglossus Hypoglossus II. laryngeus INNERVATION. Urodela Xenopus. (Norris and Francis) R. mandibularis V R. mandibularis V. R. jugularis R. jugularis VII. VII R. intermandibularis V R. jugularis VII R. recurrens X R. accessorius Hypoglossus Hypoglossus R. laryngeus recurrens R. mandibularis inferior and R. mentalis internus in larva. R. jugularis VII and R. post-trematicus IX in larva. R. post-trematicus IX. R. branchialis X. R. post-trematicus IX. R. branchialis X. R. branchialis X. R. laryngeus recurrens X. R. accessorius. Truncus intestinoaccessorius. Hypoglossus. Hypoglossus. R. laryngeus recurrens.

67 HEAD OP XENOPUS 227 SUMMARY. The foregoing account deals with some microscopic observations on the main anatomical features of the head of X. 1 a e v i s, in both larval and young adult stages. The arrangement of the lateral line sensory organs of the larva has been compared with that of the frog, and both are in general agreement with the distribution of similar organs in the Urodela. The larval musculature is found to be similar to that of X. fraseri, described by Edgeworth. Notes regarding the changes occurring at metamorphosis are given. The course of the head vein and internal carotid artery has been followed, as both blood-vessels are relevant to the study of the chondrocranium and nerves. A complete study of the chondrocranium in different stages of larvae from the time of hatching up to metamorphosis reveals several points at variance with Kotthaus's findings. Metamorphosis results in marked changes in the auditory and nasal regions. The former region has been studied by de Villiers in the adult Xenopus, and is herein only briefly reviewed. In regard to the nasal region, an account is given of the cavities and cartilages, both conforming in essentials to the typical Anuran arrangement. The hyobranchial skeleton of the larva and young frog are in close agreement with that of mature specimens investigated by Eidewood. The cranial nerves have been studied in some detail, and are observed to depart in some respects from those of typical A n u r a. The composition of the various nerves is in agreement with that of both Anura and Urodela, but the arrangement of the nerves approaches that of such Urodela as Siren (Norris), Proteus (Benedetti),and Salamandra (Francis). This is perhaps most noticeable in the innervation of the upper jaw, for a maxillary nerve is not separated from the maxillomandibular stem as in most Anura; the general cutaneous supply to the maxillary region is derived in Xenopus from the ramus ophthalmicus profundus V as in some Urodela.

68 228 NELLIE F. PATEKSON BEFERENCES. Addens, J. L., "The Motor Nuclei and Roots of the Cranial and Krst Spinal Nerves of Vertebrates", 'Zeitsehr. f. d. ges. Anat.', 1. Abt., 101. de Beer, G. R., "The Segmentation of the Head in Squalus acanthias", 'Quart. Journ. Micr. Sci.\ "Studies on the Vertebrate Head. II. Orbito-temporal Region of the Skull", ibid., 'Development of the Vertebrate Skull.' Clarendon Press, Oxford. Benedetti, E., "II Cervello e i Nervi Cranici del Proteus anguineus", 'Mem. d. 1st. Ital. di Speleologia', Ser. Biol., Mem. III. Bender, 0., 'Die Sehleimhautnerven des Facialis, Glossopharyngeus und Vagus.' Jena. Bles, E. J., "Life-history of Xenopus laevis Daud.", 'Trans. Roy. Soo. Edin.'j 41, pt. iii. Broom, 11., "Mammalian and Reptilian Vomerine Bones", 'Proe. Linn. Soc. N.S. Wales', pt "The Vomer-Parasphenoid Question", 'Ann. Transv. Mus.', 18, pt. 1. Edgeworth, F. H., "Development of the Hyobranchial and Laryngeal Muscles in Amphibia", 'Journ. Anat. London', "Autostylism of Dipnoi and Amphibia", ibid., "Masticatory and Hyoid Muscles of Larvae of Xenopus laevis", 'Journ. Anat. Camb.', 'Cranial Muscles of Vertebrates.' Camb. Univ. Press. Escher, K., "Das Verhalten der Seitenorgane der Wirbeltiere und ihrer Nerven beim Ubergang zum Landleben", 'Acta Zool.', Arg. VI. Eoske, H., "Das Geruchsorgan von Xenopus laevis", 'Zeitsehr. f. Anat. u. Entwicklungsgesch.', 103. Francis, E. T. B., 'Anatomy of the Salamander.' Clarendon Press, Oxford. Gaupp, E., 'Anatomie des Prosches', 1-3. Gilehrist, J. D. E., and von Bonde, C, 'Dissection of the Platanna and the Frog.' Univ. of Cape Town. Goodrich, E. S., "Development of the Segments of the Head of Scyllium", 'Quart. Journ. Mior. Sci.', 'Studies on the Structure and Development of Vertebrates.' Macmillan & Co., London. Grobbelaar, 0. S., 'Beitrage zu Einer anatomischen Monographic der Xenopus laovia Daud.' Berlin, van der Horst, 0. J., "Lateral Line Nerves of Xenopus", 'Psyohiat. en Nourol. Blad.', no. 3 en 4. Kingsbury, B. F., "Branchiomerism and the Theory of Head Segmentation",' Journ. Morph. and Physiol.', 42, no. 1.

69 HEAD OF XENOPtfS 229 Kotthaus, A., "Entwieklung des Primordial-Craniums von Xenopus laevis bis zur Metamorphose", 'Zeitgchr. f. wiss. Zool.', 144. Muller, E., "Unters. u. d. Mundhohlendriisen der Anuren Amphibien", 'Morph. Jahrb.', 70. Norris, H. W., "Cranial Nerves of Siren lacertina", 'Journ. Morph.', 24, no. 2. Parker, W. K., "Structure and Development of tie Skull in Batrachia", 'Phil. Trans. Roy. Soe.', 166. Peter, K., "Development of the External Features of Xenopus laevis", 'Journ. Linn. Soc. Zool.' Ridewood, W. G., "Structure and Development of the Hyobranchial Skeleton and Larynx in Xenopus and Pipa; with Remarks on the Affinities of the Aglossa", 'Journ. Linn. Soc. Zool.', "Hyobranchial Skeleton and Larynx of a New Aglossal Toad, Hymenochirus Boettgeri", ibid., 27. Rose, W., 'Veld and Vlei.' Speciality Press of South Africa, Capetown. Schoonees, D. A., "Skedelmorphologie van Bufo angostieeps (Smith)", 'S. Afric. Journ. Sci.', 27. Strong, O. S., "Cranial Nerves of Amphibia", ' Joum. Morph.', 10. Tensen, J., "Einige Bemerkungen u. d. Nervensystem von Pipa pipa", 'Acta Zool.', 8. du Toit, C. A., 1930, "Skedelmorphologie van Heleophryne regis", 'S. Afric. Joum. Sci.', 27 and 28. G. P., "Cranial Characters of Phrynobatrachus natalensis (Smith)", ibid., 30. G. P., and de Villiers, C. G. S., "Die Skedelmorphologie van Hyperolius horstoekii as Voorbeeld van die Polypedatidae ", ibid., 29. de Villiers, C. G. S., "Development of a Species of Arthroleptella from Jonkershoek, Stellenbosch", ibid., "New Aspects of Anuran Osteology and Osteogeny", ibid., "Cranial Characters of the South African Brevicipitid, Phrynomerus bifasciatus", 'Quart. Journ. Micr. Sci.', "Cranial Characters of the Brevicipitid Genus, Caeostemum (Boulenger)", ibid., "Further Notes on the Genus Cacosteroum including an Account of the Cranial Anatomy of Caeostemum namaquense Werner", 'S. Afric. Joum. Sci.', "Some Features of the Cranial Anatomy of Hemisus marmoratus", 'Anat. Anz.', 71, no. 14/ "VS. d. Schadelbau der Brevicipitidengattung Anhydrophryne Hewitt", ibid., 71, no. 14/ "U. d. Schadelbau des Breviceps fuscus", ibid., 72, no. 6/ "VS. d. Gehorskelett der Aglossen Anuren", ibid., 74, no. 4/5.

70 230 NELLIE F. PATEESON De Villiers, C. G. 8., "Breviceps and Probreviceps: Comparison of the Cranial Osteology of two closely related Anuran Genera", ibid., 75, no. 12/14. de Vos, C. M., "Spelaeophryne and the Bearing of its Cranial Anatomy on the Monophyletic Origin of the Ethiopian and Malagasy Miorohylids", ibid., 80, no. 13/16. Winterhalter, W. P., "Unters. ii. d. Stirnorgan der Anuren", 'Acta Zoologica', 12. EXPLANATION OF PLATES All the illustrations were drawn with the aid of a Leitz drawing apparatus adjusted to the magnifications given. The following table is supplemental to that on p. 176, and summarizes the figures illustrating the anatomical features of the various larval Stage Fig a Not figured , 24, ^,21 a, 25,29 Young Frog. 5, 18, 19, 20,21 6. EXPLANATION OF LETTERING. etc., auditory capsule; aca., anterior cerebral artery; ace., accessory sensory organs; acus., acustic foramina; aim., anterior intermandibular muscle of young frog; ang., angulare; ary., arytenoid cartilage; as., auditory sac of young larva; asc.pr., ascending process of palatoquadrate; asm., anterior semicircular canal; at., annulus tympanicus; ha., basilar artery; bas.hy., basihyale; bra., remains of branchial arches; br.pr., bronchial cartilage; car., internal carotid artery; cart.al., cartilago alaris; cart.obl., cartilago obliqua; cart.obl.+pl., cartilago obliqua and planum terminate; cav.med., cavum medium; cav.princ., cavum principale; c.br., musculi eonstrictores branchiales; cer., cerato-hyale; cl. cloacal aperture; cnt., connective tissue; con., musculus constrictor laryngis; cor., processus coronoideus; c.q.c.a., commissura quadrato-cranialis anterior; eric., cricoid cartilage; cris., crista intermedia; crp., crista parotica; csl., cloaeal sensory organs; c.tent., cartilage of tentacle; cm., cerebral vein; dat., dorsal part of annulus tympanious; dep.man., musculus depressor mandibulae of larvae (quadrato-hyoangularis) j d.g.rn., ductus glandula nasalis medialis; dil., museums dilatator laryngis; ec, ethmoidal cartilage; elf., endolymphatic foramen; elf.+acua., endolymphatic+acustic foramina; en., external nares; eus,, eustachian tube; exo., exoccipital condyle; fac, floor of auditory capsule in young larva; /&., fenestra hypophyseos; /c, carotid foramen; f.jug., foramen jugulare; fl. fore-limb ;/o., fenestra ovale; f.m., foramen magnum; fp., fronto-parietale; f.par., foramen parietale; f.pro., foramen

71 HEAD OF XBNOPUS 231 pro-oticum; gen., musculus geniohyoideus; g.i.m., glandula intermaxillaris; gm., glandula nasalis medialis; Har., Harderian gland; Ac., horizontal cartilage from crista intermedia; hf. hyoglossal foramen; Id., hind limb; hsc., horizontal semieular canal; hm., hyomandibular sensory organs; h.vn., head vein; hyog., museums hyoglossus; ilc., inferior labial cartilage; int.hy., musculus interhyoideus; int.man., museulus intermandibularis; iorb., infra-orbital sensory organs; Up., intertrabecular plate; lam.inf., lamina inferior; Ian., larynx; for., musculus laryngeus dorsalis; lev.ant., musculus levator mandibulae anterior; lev.hy., musculus levator hyoideus (= orbito-hyoideus); lev.post., musculus levator mandibulae posterior; lev.man., musculus levator mandibulae in young frog; lev.tent., museulus levator tentaculi; l.h.vn., vein from levator hyoideus; Lo.p., larval otic process; max., maxillare; m.cu., musculus eucullaris; m.dep. man., musculus depressor mandibulae of young frog; me., middle ear; Meek., Meckel's cartilage; m.pt.hy., musculus petrohyoideus; m.ptg., medial part of pterygoideus; msl., medial sensory organs; mus.pr., processus muscularis of palatoquadrate; nas., nasale; n.l.d., duetus nasolaerimalis; not., notochord; o.e.c, os en ceinture (orbitosphenoid); ope, operculum; opha., ophthalmic artery; oph.vn., ophthalmic vein; osl., dorsal sensory organs; otp., otic process; par., parietal sensory organs; pea., posterior cerebral artery; pch., parachordal; pip., pars interna plectri; pl.ani., planum antorbitale; plee., plectrum; pis., pleurosphenoid; pl.term., pla,num terminate; pm., premaxillare; p.met., pila metoptica; pmp., pars mediaplectri; porb., post-orbital sensory organs; ppc, post-palatine commissure; ppl., posterolateral process of hyobranchial skeleton; ppm., posteromedian process of hyobranchial skeleton; p.ptg., processus pterygoideus; pro., pro-oticum; prq., paraquadratum; ps., parasphenoid; pt., plica terminale; ptg., pterygoideus; pt.vn., pituitary vein; quad., quadratum; r.e., musculus rectus externus; rec.lat., recessus lateralis; rec.med., recessus medians; rec.sac, recessus sacciformis; r.s., musculus rectus superior; sem., septomaxillare; sep., septum nasi; seth., supra-ethmoid; smp., sulcus maxillopalatinus; s.o.b., subocular bar of palatoquadrate; sof., subocular fenestra; sorb., supra-orbital sensory organs; sp. 1 and 2, foramina for 1st and 2nd spinal nerves; s.p.e., superior prenasal cartilage; spir., spiraculum; to., tectum anterius; tee., tectum nasi; tent., tentacle; th.f., thyroid foramen in hyobranchial skeleton; th.gl., thymus gland; thyr., thyroid gland; tp., tectum posterius; trab., trabecula cranii; utr., utriculus; vat., ventral part of annulus tympanicus; vest., vestibulum; vlp., ventro-lateral process of palatoquadrate; vom., vomer (= praevomer); vsl., ventral sensory organs. The main cranial nerves are indicated by Roman numerals (I to X); other abbreviations connected with the nervous system are as follows: ant.lat., lateralis anterior; br.x, branchial nerve; cd., ramus cardis X; cf., nerve to levator mandibulae posterior; cut., ramus cutaneus VH; d.al., dorsal part of lateralis anterior; gas., ramus intestinalis X; g.gl., ganglion glossopharyngeus; g.pro., ganglion pro-oticum; g.v., ganglion IX-X; hym., truncus hyomandibularis; io., truneus infra-orbitalis; inf., infundibulum; int.nas., ramulus nasalis internus (op. 4 ); md. 1-6, branches of ramus mandibularis V; m.ext., ramus mentalis externus VII; m.int., ramus

72 232 NELLIE F. PATEESON mentalis interims VII; mus. x, ramus accessorius; mus. 2, nerve to petrohyoideus muscle; mus. 3, nerve to skin and shoulder; op., ramus ophthalmicus profundus V; p.a.i., ramus post-trematicus IX; pal., ramus palatinus VII; p.a.s., ramus pre-trematicus IX; pin., pineal body; pit, hypophysis; post.lat., lateralis posterior; r.jug., ramus jugularis VII; rln., ramus laryngeus recurrens X; r.man., ramus mandibularis V; r.merit., ramus mentalis VII; r.temp., ramus supratemporalis; so., ramus ophthalmicus superflcialis; ts., truncus supra-orbitalis; v.al., ventral part of lateralis anterior; vis., truncus intestino-accessorius X. PLATE 9. Figs Transverse sections through the nasal capsule of a tadpole towards the end of metamorphosis. X 30. PLATE 10. Fig. 4. Transverse section through posterior region of nasal capsule of a metamorphosing specimen. X 25. Fig. 5. Transverse section posterior region of nasal capsule of a young frog, passing through external opening of nasolacrimal duct. X 25. PLATE 11. Figs. 6 and 7. Transverse sections through the olfactory region of a tadpole measuring 60 mm. long, x 30. Fig. 8. Transverse section of tadpole showing the entrance of the ramulus nasalis internus into the olfactory capsule. X 25. Figs. 9 and 10. Transverse sections through the telencephalon of a 60 mm. long larva, showing the disposition of the nerves. In fig. 10 the oculomotorius is dividing before encircling the ramus ophthalmicus profundus V. X 25. Fig. 11. Transverse section through the oculomotor foramen in the tadpole. The ophthalmic artery can be seen passing out with the nerve. X25. PLATE 12. Fig. 12. Transverse section through the carotid foramen of a larva. The artery is just dividing into two within the cranial cavity. X 25. Figs. 13 and 14. Transverse sections through the thalamencephalon of larva, showing the separation of the nerve trunks from the ganglion prooticum. x 25. Fig. 15. Transverse section to show the exit of the cerebral vein from the cranial cavity on the right, and the junction of the pituitary and cerebral veins on the left. X 25. Fig. 16. Transverse section passing through the ganglion pro-oticum after the separation of the truncus supra-orbitalis from it. x 25. Fig! 17. Transverse section passing through anterior region of cerebellum before formation of the ganglion pro-otioum. X 25. Figs. 18 a-c. Transverse sections passing through the auditory capsule of a young frog. The post-palatine commissure, part of the middle ear, and the plectral apparatus are visible. X Fig. 18 a is 90/x in front of Fig. 18 6, and there are 380/A between the latter and Fig. 18 c.

73 Quart. Journ. Mier. Scl Vol. 81, N. 8,, PL 9 ma s.p.c S iv - P. Paterson, del.

74 Quart. Jomm. Mkr. 8d. V6L 81, if. 8., PI. 10 so Id sett, int. maitf rm.ent. aim.. 'ck. F. Paterson, del.

75 Quart. Joum. Mier. Set. V6L 81, N. 8,, PL 11 cer- N. F. Paterson, del,

76 Quart. Jmirn, Micr. Sci. Vol. 81, N. 8., PI. 12 an Aym. 18 (b) {f tevman. N. F. Paterson, del.

77 Quart. Jaurn. Micr. Set. Vol. 81, N. 8., PI 13 20(a) nt. er. sc tent. ocas. mcu. ot N. F. Pateraon, del.

78 Quart. Journ. Micr. 8ei* Vol. 81, N. 8., PL ec. xk...tent. tp. ca.rt.cbl -plant. term. fleck quad e/: N. F. Paterson, del.

79 Quart. Journ. Micr, Set. Vol. 81, N. 8., PI. 15 cq.c.a. N. F. Paterson, del. acus.

80 . Jottm. Micr. 8oi. Vol. 81, N. 8., PL rvjlr a, 30 (b) 50 29^ cntnas.,. <f s >- f* pp.,. md: next rme. 30 (a) N. JF. Paterson, del.

81 HEAD OF XBHOPTJS 233 PLATE 13. Fig. 19 a. Transverse section of young frog, passing through the auditory capsule near the end of the pars intema pfectri and just anterior to the operculum. x Fig Transverse section of voting frog, 40/A behind fig. 19 a. This section passes through the operculum, the lower end of which has not fused with the wall of the auditory capsule. X46-5. Figs. 20 a and b. Transverse sections through the larynx of a young frog, showing the laryngeal muscles. Fig. 20 a is 110/x anterior to Fig Both X50. Fig. 21 a. Diagrammatic reconstruction of the hyobranchial skeleton of a metamorphosing specimen. Ventral view, x 8-4. Fig Diagrammatic reconstruction of the hyobranchial skeleton of a young frog. Ventral view, x 8-4. Figs. 22 a and 6. Diagrammatic reconstructions of the chondrocrania of very young larvae, measuring 5 mm. and 7 mm. long respectively. The smaller larva had only recently emerged and its mouth was still closed. Dorsal views, each x 50. Fig. 23. Diagrammatic reconstruction of the chondrocranlum of a larva measuring 10 mm. long. Dorsal view. X 25. PLATE 14. Fig. 24. Diagrammatic reconstruction of the ehondrocranium of a larva measuring 60 mm. long. Dorsal view, x 8-4. Fig. 25 a. Diagrammatic reconstruction of the cartilages in the nasal capsule of a metamorphosing larva. Side view. X Fig Diagrammatic reconstruction of the suspensorial region of the same larva as in Fig. 25 a. With the exception of the fronto-parietale, the ossifications have been omitted. Side view. x8-4. PLATE 15. Fig. 26 a. Dorso-lateral view of the skull of adult Xenopus laevis. Xapprox Fig.266.VentralviewofskullofadultXenopus laevis. xapprox.2-5. Fig. 27 a. Dorsal view of a reconstruction of the ehondrocranium and cranial nerves of a larva measuring 28 mm. long. On the right side the cartilages have been removed to expose the ganglia and roots of the nerves. X In this and subsequent diagrams of the nerves no attempt has been made to indicate the different kinds of fibres. Fig. 27 b. Ventral view of the same reconstruction as in Fig. 27 a. Only the articular regions of the palatoquadrate are given in order to show the relative positions of the blood-vessels, some of the cranial nerves, and the parts of the palatoquadrate. X Fig. 27 c. Side view of the same ehondrocranium. The nerves are seen issuing from their respective foramina. X Fig. 27 d. Diagram representing an inner view of the right half of the same ehondrocranium as in Figs. 27 ct-c. In this diagram the positions of the nerves and blood-vessels within the ehondrocranium are shown. X16-65.

82 234 NELLIE F. PATEESON PLATE 16. Fig. 28. Side view of a diagrammatic reconstruction of the main cranial nerves in a larva measuring 60 mm. in length. X Fig. 29. Diagrammatic reconstruction of the main cranial nerves in a metamorphosing specimen. Side view, x Fig. 30 a. Side view of tadpole, showing positions of lateral line sensory organs observed in young stages. 30 mm. in total length; 11 mm. to cloacal aperture; tentacle, 2-5 mm. long. X 4-5. Fig. 30 b. Dorsal view of a metamorphosing larva to show the arrangement of the lateral line sensory organs. 46 mm. in total length; 15-5 mm. to cloacal aperture; tentacle, 3-0 mm. long. x4.