On the Head of the Liopelmid Prog, Ascaphus truei. I. The ......On the Head of the Liopelmid Prog,...

On the Head of the Liopelmid Prog,Ascaphus t ruei .

I. The Chrondrocranium, Jaws, Arches,and Muscles of a partly-grown Larva.1

By

H. K. Pusey, M.A.,

Department of Zoology and Comparative Anatomy. Jenkinson MemorialLecturer in Embryology in the University of Oxford.

With Plates 6 to 14 and Text-figures 1 to 7.

CONTENTS. . P A G E

1 . F O R M E R W O E K . . . . . . . . . 1 0 6

2 . M A T E R I A L A N D M E T H O D S . . . . . . . 1 0 7

3 . A C K N O W L E D G E M E N T S . . . . . . . . 1 0 8

4 . A D E S C R I P T I O N O F T H E C R A N I A L A N A T O M Y O F A S I N G L E P A R T L Y -

G R O W N L A R V A O F A S C A P H U S T R U E I . . . . 1 0 9

( a ) T h e A n t e r i o r E n d o f t h e C r a n i u m 1 0 9(b) The Supra-rostral System . . . . . . 1 1 0(c) The Nasal-sac and Preoral Buccal Cavity . . . 112(d) T h e Side wal l of t h e N e u r o c r a n i u m a n d i ts F o r a m i n a . 113(e) T h e F l o o r of t h e N e u r o c r a n i u m . . . . . 1 1 5( / ) T h e A u d i t o r y Capsule 115(g) T h e Car t i lage Cranial Roof . . . . . . 117(h) T h e P a l a t o q u a d r a t e 118

i. T h e Commissu ra Quadra to -c ran ia l i s An te r io r . . 118ii. T h e Pos te r io r S p u r of t h e Q u a d r a t e . . . 1 1 8

iii. T h e P t e r y g o i d B o n e - r u d i m e n t ( = l i g a m e n t u m q u a d -ra to -e thmoida le ) . . . . . . 1 1 9

iv. T h e ' Pos te r io r B a s a l P r o c e s s ' of t h e Q u a d r a t e . . 1 1 9v . T h e Ascending Process a n d L a r v a l Ot ic Process . 120

vi . T h e Muscu la r Process . . . . . . 121vi i . A n Ar te r i a l T u n n e l t h r o u g h t h e Q u a d r a t e . . 121

(i) T h e L o w e r J a w Sys t em . . . . . . . 121(j) T h e H y o - b r a n c h i a l A p p a r a t u s . . . . . 122

5. A C O M P A R I S O N O F C E R T A I N C R A N I A L M U S C L E S O P A S C A P H U S ,

U R O D E L A , A N D O T H E R A N U R A 126

(a) Preface . . . . . . . . . 126(&) T h e M a n d i b u l a r Muscles 127(c) T h e H y o i d Muscles 131(d) TheLeva tores Arcuum Branchialium Muscles . . . 1 3 5

1 This papa- was writ ten in the spring of 1941 and has no t been

altered since.

NOS. 334-5 T

106 H. K. PUSEY

PAGE(e) T h e C o n s t r i c t o r e s B r a n c h i a l e s M u s c l e s . . . . 1 3 6( / ) T h e S u b a r c u a l e s R e c t i M u s c l e s 1 3 7(g) T h e S u b a r c u a l e s O b l i q u i M u s c l e s 1 3 9(h) T h e T r a n s v e r s i V e n t r a l e s M u s c l e s . . . . . 1 4 1(i) D i s c u s s i o n . . . . . . . . . 1 4 2(j) T h e D i a p h r a g m a t o - b r a n c h i a l i s M u s c l e s . . . . 1 4 4

6 . T H E H Y P O B R A N C H I A L S P I N A L M U S C L E S O F A S C A P H U S . . 1 4 5

( a ) T h e G e n i o h y o i d e u s M u s c l e s . . . . . . 1 4 5(6.) T h e R e c t i C e r v i c i s M u s c l e s . . . . . . 1 4 5

7 . D I S C U S S I O N . . . . . . . . . . 1 4 6(a) T h e Pos i t ion of t h e Sp lanchn ic S t r u c t u r e s of A s c a p h u s

a n d o the r F r o g L a r v a e . . . . . . 146(b) Is the Cranial Ground-plan of Ascaphus really Primitive ? 148(c) Extrapolation to the Larval Ancestor of the Frogs . . 153(d) The 'Anterior Basal Process' ( = Gaupp's Commissura)

and the 'Pos te r io r Basal Process ' . . . . 154(e) The Postulated Ancestral Larva and its Evolution to the

Modern-type Tadpole . . . . . . 157(/) A Criticism of Save-S6derbergh's Theory of the Nature of

the Commissura . . . . . . . 161(g) Changes in the Auditory Regions of Anuran Skulls . . 164(h) The Articular Region and the Body of the Quadrate . 166(i) The Otic Processes 167(j) The Muscular Process 168(/<•) The Origins of the Adductor J a w Muscles . . . 170(1) The Ascending Process of the Palatoquadrate . . . 1 7 1(m)The Larval Otic Process of X e n o p u s . . . . 171(n) The Roof of the 'Quadrate Tunnel ' . . . . 1 7 3(o) The Supra-rostral System . . . . . . 1 7 3

8O T H E CHARACTERS WHICH A s c A P H u s SHARES WITH THE URODELES 175

9. T H E CHARACTERS WHICH A S C A P H B S SHARES WITH D I S C O -

G L O S S U S . . . . . . . . 177

10. SUMMARY 178

11. LITERATURE CITED . . , . . . . . 1 8 0

12. ABBREVIATIONS USED I N THE FIGURES . . . . . 181

13. DESCRIPTION OF PLATES 6 TO 14 . . . . . . 183

1. FORMER WORK.

NOBLE (1931, p. 486) has published a classification of theorder Anura in which he places the primitive North Americanfrog, A s c a p h u s t r u e i , Stejneger, with the New Zealandgenus L i o p e l m a , in a separate family, 'Liopelmidae', which

HEAD OP ASCAPHUS 107

lies at the base of his evolutionary tree. He thus believes thatthe Liopelmidae show more primitive characters than anyother living family of frogs. With this view the writer is in fullagreement, for much new evidence is given below of the primi-tive nature of the larval skull, jaws, gill arches, and muscles ofthe Ascaphus tadpole. De Villie'rs (1934) has giyen anaccount of various other attempts to classify this animal; healso is in agreement with Noble's classification.

The adu l t skul l of Ascaphus was described by deVilliers (1934) and by Wagner (1934) from sections of one indivi-dual. These papers are mentioned in various places by de Beer(1937) and by me (Pusey, 1938), where reconstructions are givenfrom their published figures together with criticisms of certainof their findings. Noble (1931) had already figured a palatalview of the adult bony skull (his fig. 81 a) and had given someaccount of its structure. The l a r v a l skull has neverbeen t he subjec t of de t a i l ed work, so far as thewriter is aware, although Noble (1927, fig. 9) has publishedfigures of isolated sections of larvae to illustrate particularpoints. Valuable accounts, however, of the habits of both thelarvae and the adults are given by Gaige (1920) and Noble(1927). Gaige states that Van Denburgh (1912) has given someaccount of the (? adult) skeleton, but I have not had access tothis paper.

Both de Beer (1937) and Pusey (1938) haves t res sed the i m p o r t a n c e of a knowledge of thes t r u c t u r e and deve lopmen t of the skul l of Asca-phus for the i n t e r p r e t a t i o n of the s t r u c t u r eand evo lu t ion of the a n u r a n skull in genera l .I hope that this will be the first of a series of papers aimed atfilling this important gap in our knowledge of Anuran evolution.

2. MATERIAL AND METHODS.

The specimens of Ascaphus in my possession were collectedfrom Dick Creek, in the Carbon Eiver valley of the EainerNational Park, Pierce County, Washington State, U.S.A. Thework is based on a single series of transverse sections of a partly-grown larva of Ascaphus t r u e i . The overall length of the

108 H. K. PUSBY

animal was 28 mm.; the length of the tail, measured ventrally,was 17 mm. and measured dorsally, 19 mm.; there were novisible hind legs. By comparison with the size ratios of E a n at e m p o r a r i a and Discoglossus p i c tus tadpoles andwith metamorphic stages of other A s c a p h u s tadpoles, it isprobable that this specimen had completed from one-half to two-thirds of its larval life. It was, therefore, a well-formed larvawith all its structures completely laid down; very little youngcartilage is present in the sections.

The gut and the lenses of the eyes were removed and the speci-men was stained in bulk with borax carmine. It was impregnatedwith 2 per cent, celloidin and was then embedded in wax and cut,as transverse sections, at 15 t̂. The sections were counterstainedwith picro-nigrosin. Drawings were made from every thirdsection with a microprojector, and .from them the various recon-structions wrere made, either by the ' contour' or the ' projection'method described in another paper (Pusey, 1939); these aregraphical methods. Selected sections were also drawn with theaid of the microprojector. Towards the end of the work twoadditional tadpoles of Ascaphus were cut as transversesections, and one post-metamorphic animal was similarlysectioned.

For comparison with the Ascaphus material, I usedpreviously prepared sections of various stages of Discoglossusp i e t u s , Bombina v a r i e g a t a , E a n a t e m p o r a r i a ,and S a l a m a n d r a macu losa . I also had access to furthersections of E a n a t e m p o r a r i a , Bufo v u l g a r i s , andvarious Urodeles, which are in the departmental collection ofthe Oxford University Department of Zoology and ComparativeAnatomy.

3. ACKNOWLEDGEMENTS.

I wish to repeat my sincere thanks to Prof. James E. Slaterand to Mr. John W. Slipp, of the College of Puget Sound,Washington, for the time and effort which they gave to collect-ing the material of Ascaphus and for their generosity insending it to me; to Prof. Carl L. Hubbs and to Mrs. HelenT. Gaige, of the University of Michigan, who helped me in myearly efforts to obtain material and sent me an adult specimen


of A s c a p h u s ; to Mr. E. Maxwell Savage, for his generosityin supplying me with a wealth of young larvae of D i s c o g 1 o s -sus p i c t u s and B o m b i n a v a r i e g a t a ; to Drs. E. S.Kussell and C. H. Waddington, for their gifts of Discoglossustadpoles; to Dr. Nellie F. Patterson, of the University ofWitwatersrand, who made and sent to me a wax-plate modelof part of the chondrocranium ofXenopus l a e v i s ; and toProf. D. M. S. Watson, F.R.S., for welcoming me back to hisDepartment to use the microprojector there, and for giving memuch valuable information from his own researches. The work,as a whole, was carried out in Oxford, at the Department ofZoology and Comparative Anatomy. I therefore wish to endby thanking Prof. E. S. Goodrich, F.E.S., for his interest in mywork and for reading the manuscript for publication.

4. A DESCRIPTION OF THE CRANIAL ANATOMY OF A SINGLE

PARTLY-GROWN LARVA OF A S C A P H U S T R U E I .

(a) The A n t e r i o r End of the Cran ium.

The cranial cavity is not shut off by cartilage anteriorly, butopens freely forwards (oec, figs. 15 and 16, PL 11; and fig. 18,PI. 12). In this region, the side wall of the brain-case is piercedby the foramina for the olfactory nerves (fol, fig. 1, PL 6, andfig. 19, PI. 13); these foramina open laterally and not forwardsas in other Anurans and are at the anterior end of the skull, withvery little cartilage in front of them. The trabecular horns inR a n a and the modern-type frogs project from the floor of theskull a long way fo rwards i n t o the s n o u t , in frontof t he o 1 f ac to ry foramina . In A s c a p h u s , however,the cartilage which is equivalent to the horns of the other frogs,hangs down from the trabecular region of the front of the skull asa vertical flange placed below and la rge ly beh ind theo l f a c t o r y foramen (d, figs. 1 and 3, PL 6; fig. 15, PL 11;fig. 19, PL 13). The lower edge of this flange is indistinguishablyfused to the upper edge of the medial part of the supra-rostralsystem (dsm); t h e r e is t h e r e f o r e no e x a g g e r a t e da n t e r i o r d e v e l o p m e n t of t he horn reg ion of t het r a b e c u l a such as is t y p i c a l of o the r frogs. In

110 H. K. PUSBY

the absence of evidence from younger larvae, the actual limitsof the horns, as opposed to the supra-rostral cartilage, must forthe present remain in doubt. In passing forwards from behind,the floor of the cranium becomes thicker. In the thicknessof the cartilage a pit is excavated from in front, ending blindlybehind; this is well shown as ap, in fig. 16, PL 11; fig. 17, PL 12 ;fig. 19, PL 13. It is also shown in fig. 2, PL 6, where the floor ofthe pit appears as a cartilage bridge, cb, joining the upper edgesof the median part of the supra-rostral system; the typical floorof the cranium makes the roof of the pit.

The n a s a l - s a c l ies l a t e r a l l y to t h e o l f a c t o r yforamen and somewha t beh ind i t and no t inf ront of i t as in o t h e r frogs ; it is partially roofed inby a projecting ledge from the orbital cartilage and is supportedbelow on a similar ledge, d, from the region where the trabecularhorn passes into the supra-rostral.

(b) The S u p r a - r o s t r a l S y s t e m .

This system is figured from in front in fig. 16, PL 11, and frombehind in fig. 18, PL 12; it is also seen in the various reconstruc-tions, fig. 15, PL 11; fig. 17, PL 12; and fig. 19, PL 13; and insection in figs. 2, 3, PL 6; and fig. 4, PL 7. It consists of a single,median part (clsm) fused to the trabecular flanges (ct), and a pairof separate, lateral wings (clsl). The median part is a flat platefacing ventrally, with the typical V-shaped notch (nsr) cut outof its anterior edge, as in other larval frogs (fig. 17, PL 12). Onits dorsal surface there are flanges rising up on each side of themiddle line which are fused to the trabecular flanges mentionedabove (fig. 3, PL 6; fig. 18, PL 12). Thus, with the assistance ofthe cranial floor, this medial piece encloses a tunnel, open infront and behind, which houses a preoral buccal cavity (pm,fig. 8, PL 6; fig. 4, PL 7). The V-shaped notch is continued back-wards on the dorsal surface of the ventral plate as a groove,which thus probably indicates that this median structure isreally of paired origin in development (see fig. 18, PL 12).

The lateral wings of the supra-rostral system, are independentstructures. They articulate with the lateral edges of the medianpiece (fig. 3, PL 6) and are loosely held to it, and to the epithelial


roof of the sucker in front, by ligaments (figs. 2,3, PI. 6). The freeedges, both of the lateral pieces and of the median piece, aremutually sheathed behind in dense mesenchyme of the natureof procartilage (dm, fig. 4, PI. 7). This mesenchyme is underlainby very deep epidermis, whose outer part is hardened into thehorny tooth-blades of the upper 'lip' (de and htp, fig. 4, PI. 7).These blades, as Noble (1927) has shown, fail to reach theposterior edge of the lip, so that a band of less cornified epidermisstretches for a short way behind them Ind forms the anteriormargin of the mouth (ul and mo, fig. 5, PL 7).

The ? levator mandibulae posterior profundus muscle (Impm,figs. 4, 5, PI. 7; figs. 6, 7, PL 8; figs. 8, 9, 10, PL 9; figs. 11, 12,13, 14, PL 10) is mainly inserted on the lower jaw. Its mostmedial fibres pass into a tendon which is inserted on the roofof a diyerticulum of the mouth cavity underlying the lower jaw(fig. 4, PL 7, left side, unlabelled). Its most antero-dorsal fibrespass into a diffuse tendon (1st, fig. 3, PL 6) which concentratesin front into a weak strand inserted partly on the dorso-medial edge of the supra-rostral wing (1st, fig. 3, PL 6, right side)and partly on a septum of a lymph space of the snout.

The ? levator mandibulae posterior superficialis muscle (hnsm,fig. 3, PL 6; figs. 4, 5, PL 7; figs. 6, 7, PL 8; figs. 8, 9, 10,PL 9; figs. 11,12, 13, PL 10) passes wholly into a similar diffusetendon (It, fig. 3, PL 6, right side) which makes rather vagueconnexions with the supra-rostral Aving and the lymph spaceseptum; the muscle has no well-organized tendon at all.

Neither of the tendons from these two muscles to the supra-rostral wing is strong or well marked and it is uncertain whetherthey could transmit an effective pull from their muscles to movethe cartilage. Perhaps, when the lower jaw is tightly closed,the lateral flange on the posterior jaw cartilage may lever up thetrabecular-quadrate ligament and so pull the upper end of thesupra-rostral wing backwards and upwards. The middle-pieceof the larval upper jaw system is rigid and is unsupplied withmuscular insertions.

Two ligaments are also attached to the larval upper jawsystt.n (see fig. 19, PL 13, tql and Iq). The 'trabecular quad-rate ligament' (tql, = ligamentum cornu-quadratum laterale,

112 H. K. PUSEY

of Schiilze, 1892) runs from the antero-lateral tip of the quadrateto the postero-dorsal tip of the lateral supra-rostral wing. InE a n a t e m p o r a r i a this same ligament passes further for-wards to the antero-lateral tip of the trabecular horn and isindependent of the supra-rostral (Pusey, 1938, fig. 8, PI. 9). InDiscoglossus p i c t u s , as in A s c a p h u s , it is attachedto the wing of the supra-rostral, but no longer to its postero-dorsal tip, but more anteriorly on its outer side nearer to thetip of the trabecular horn (personal observations). ThusDiscog lossus , in this as in many other points, is inter-mediate between Ascaphus and E a n a . I n P e l o b a t e s i tis attached, at the same time, both to the trabecular horn and tothe adrostral prong of the supra-rostral (see Luther 1914 , fig. 76).

The second is the ligamentum quadrato-ethmoidale (Iq, fig.19, PI. 13). This runs forwards, embracing the lower edge of thepterygoid process; where this process ends in front, the ligamentcontinues freely forwards below the internal nostril (fig. 4, PL 7),and is inserted on the upper part of the posterior pillar (clsm,fig. 18, PI. 12), where the medial part of the supra-rostral is fusedto the trabecular flange. In both E a n a and Discoglossusthis ligament is inserted on the under side of the trabecularhorn, mid-way along its length, in the latter, to a special ventralflange which is thus probably homologous to the trabecularflange of Ascaphus (personal observation, and see alsoLitzelmann, 1923, figs. 12 and 13, for B o m b i n a t o r and vanSeters, 1922, fig. 3, PI. 8, for Aly t e s ) .

(c) The Nasa l - sac and P r e o r a l Bucca l Cav i ty .

The external nostril opens on a projecting funnel (Noble,1927, fig. 8) at the level of the olfactory foramen. This funnelopens into a simple, undivided nasal-sac lying largely behindthe olfactory foramen. The sac tapers behind into a mem-braneous ' posterior narial tube' (pnt, fig. 3, PI. 6; fig. 4, PI. 7)which runs backwards and turns medially through the notch(tint, fig. 18, PI. 12; fig. 19, PI. 13) to join the buccal cavity.This notch is closed below by the quadrato-ethmoidal ligament.The internal nostril (fig. 4, PI. 7) opens by a valvular openinginto a preoral buccal cavity which is housed between the cranial

HEAD OF ASCAPHUS 113

floor and the median part of the supra-rostral system (pm, fig. 3,PI. 6; fig. 4, PL 7; see also fig. 18, PL 12). Each wall of thispreoral cavity is thickened into a longitudinal band of deep*ciliated epithelium (eb, fig. 3, PL 6; fig. 4, PL 7), supplied by abranch of the olfactory nerve (Ib, fig. 3, PL 6) which runs back-wards with the posterior narial tube and passes with it throughthe notch, nnt. Noble (1927, pp. 65 and 66) has described howthe tadpoles of A s c a p h u s feed on particles taken in with acurrent of water through the nostrils and nasal-sacs and passedout at the median ' spiracle'. It is possible, therefore, that theseciliated bands may play a part in this method of feeding.Attempts to find further traces of a ciliary mechanism, how-ever, have failed.

(d) The Side Wall of the N e u r o c r a n i u m and i t sF o r a m i n a .

In the absence of embryological evidence from earlier stages,the cranial wall only calls for a simple description: the' variousfigures show the arrangement very clearly. The cartilage isthick and heavily built for so young a stage and the side wall iscomplete and without membraneous tracts. The olfactory fora-men, at, at its anterior end, has already been mentioned (p. 109).The optic and oculomotor foramina (foil and focn, fig. 19, PL 13)open into a common external pit in the side wall where thecartilage is thinner; a lightly built pila metoptica separatesthem. There is no separate foramen for the IVth nerve, sinceit passes out via the oculomotor foramen, which is a uniquestate of affairs, not found in the other frogs nor in the Urodeles.The pila antotica and parts of the oculomotor and trigeminal(ft) foramina are covered, in lateral view, by the broad ascendingprocess of the palatoquadrate, which is fused to the orbitalcartilage above the pila (see fig. 19, PL 13, where the limits ofthe trigeminal foramen and the anterior border of the auditorycapsule are shown by broken lines, - - -). Fig. 8, PI. 9, shows therelations in this region. The trigeminal foramen is an oval,vertically-lengthened opening between the auditory capsule andthe si, e wall of the skull. Its anterior outlet is partially ob-structed by two sheets of cartilage which lie across it from the

114 H. K. PUSEY

palatoquadrate to the cranial wall. These are the ascendingprocess (pa) above and the commissura quadrato-cranialisanterior (cqa) below, both of which, of course, belong to thepalatoquadrate. Thus t h e r e are t h r e e ex i t s from thetrigeminal foramen: (1) an upper outlet (ftd, fig. 9, PL 9; fig. 15,PI. 11; figs. 19, 20, PI. 13; fig. 22, PI. 14) between the capsule,the orbital cartilage, and the posterior edge of the ascendingprocess; nerves V 2 and 3 and a lateral line branch of VII (fig. 9,PI. 9; fig. 23, PI. 14) pass through this opening; (2) an anterioroutlet pt, fig. 17, PI. 12) into a tunnel between the ascendingprocess above, the pila antotica medially and the commissureand palatoquadrate ventrally and laterally (fig. 8, PI. 9); theanterior end of this tunnel is seen in fig. 20, PL 13, marked pt.The tunnel transmits the profundus branch of the trigeminalnerve and a branch of the lateral head vein (fig. 8, PL 9); (3) alarger, ventrally-facing, pear-shaped opening, seen in fig. 17,PL 12, as a darkly shaded space on the left side, bounded behindby the front edge of the basitrabecular process (bt), medially bythe trabecula, antero-laterally by the posterior edge of the com-missure (pcqa), and laterally by the 'posterior basal process'(bpr). This l a s t open ing is the only r e p r e s e n t a -t i v e , in the l a r v a l skul l of A S C A P H U S , of the hugesubocu l a r v a c u i t y of t he l a r v a l skul l s of allo t h e r f rogs. It is important for the subsequent discussion(p. 154) to realize that no s t r u c t u r e s , such as ne rvesor blood vesse l s , pass u p w a r d s or downwardst h r o u g h th i s small v a c u i t y and t h a t , t h e r e f o r e ,the p o s t e r i o r edge of t he commissure (pcqa)could have been d i r e c t l y app l ied to the wholel e n g t h of the b a s i t r a b e c u l a r / p r o c e s s (bt), in t h eea r l i e r a n c e s t o r , w i t h o u t a f fec t ing any o t h e ra n a t o m i c a l r e l a t i o n s h i p s . In such a case t hev a c u i t y would have been e n t i r e l y a b s e n t .

From behind the notch (nnt, fig. 19, PL 13) for the internalnostril, to the posterior edge of the commissure (pcqa, fig. 15,PL 11; fig. 17, PL 12) the ventro-lateral edge of the trabecularcartilage is carried down into a flange which is fused (1) to theupper edge of the pterygoid process of the palatoquadrate


anteriorly (fig. 5, PI. 7; fig. 6, PI. 8) and (2) to t h e body ofthe q u a d r a t e i t se l f more p o s t e r i o r l y (fig. 7, PL 8;fig. 8, PI. 9). This fusion represents the commissure of modern-type frogs, but in a very much extended and probably moreprimitive condition.

(e) The F l o o r of the N e u r o c r a n i u m .Even in so young a larva the floor of the cranium is fully

chondrified into quite a thick sheet of cartilage; it is piercedby the two pairs of foramina typical of all frog larvae. Eachforamen caroticum primarium (fcp, fig. 17, PI. 12) admits thecarotid artery of its side to the braincase, and each foramencranio-palatinum (fc) transmits the palatine branch from it tothe palate (figs. 2, 3, PI. 6; figs. 4, 5, PI. 7; figs. 6, 7, PI. 8;fig. 8, PL 9). Behind the primary carotid foramina there is aslight pituitary fossa (pitf, fig. 20, PL 13), behind which againthe cranial floor is thickened. In this thicker part the notochord(n, fig. 17, PL 12) is largely encased; however, its anterior tipis not surrounded by cartilage below, and this also holds truein the occipital region (fig. 17, PL 12). Elsewhere, the notochordis sheathed in cartilage both above and below, so that in thischaracter A s c a p h u s is unlike both the Urodela and theother Anura (de Beer, 1937, p. 460); see figs. 12 to 14, PL 10.The cranial floor in the auditory region will be described in thefollowing section.

(j) The A u d i t o r y Capsu le .

The relations of the anterior end of the capsule are wellshown in section in figs. 9,10, PL 9; figs. 11,12, 13, PL 10; andare reconstructed in figs. 22 and 23, PI. 14, among others.

The auditory region of the adult skull of A s c a p h u s hasbeen described by de Villiers (1934). A redescription from hisaccount was given by Pusey (1938) with a reconstruction madefrom his original figures. However, some doubt remained aboutthe truth of my interpretation given there; this doubt is now-removed by the present research, which shows that interpreta-tion to be correct.

It is undoubtedly correct that there is a complete basitra-

116 H. K. PUSEY

becular process (bt), projecting laterally from the cranial base,in front of the foramen [pf) for the palatal branch of the facialnerve. This process projects beyond the floor of the auditorycapsule as a free 'cartilage ledge of the auditory capsule'(de Villiers; see fig. 10, PI. 9; fig. 11, PL 10; and fig. 22, PL 14,of this paper). To its outer end is fused the 'posterior basalprocess' (bpr); see discussion on pp. 154 et. seq. Part of the pro-jecting plate lies behind the exit of the palatine nerve and is thusa post-palatine commissure (ppc, fig. 12, PL 10). Both partsof the plate form the effective floor of the anterior part of theauditory capsule. The t r u e floor of t h e c a p s u l e maybe seen a b o v e i t , h o w e v e r , as a m e m b r a n e o u sshee t (mfac, figs. 11 and 12, PL 10) covering the branches andganglion of the facial nerve and parting them from the auditorynerve above. More anteriorly, the cartilagenous wall and floorof the capsule overhang and are fused with the basitrabecularprocess in front of the hyomandibular branch of the facial nerve.This nerve, therefore, runs out through its own lateral foramen,above the process (fig. 23, PL 14; figs. 11,12, PL 10). In the sameway the antero-medial section of the vertical capsular wall isalso absent and the deficiency is made good by the absorptioninto the capsule of the.prefacial commissure (pfc, fig. 10, PL 9;

.fig. 11, PL 10; fig. 15, PL 11; fig. 22, PL 14) which lies in frontof the joint exit of nerves VII and VIII. How far forwards thecommissure may extend and at what point it is fused to the truecapsular wall cannot be determined at this late stage of develop-ment. Its presence, of course, is proved by the fact that itseparates the root of the Vth from the root of the VHth nerveand makes the VHth nerve appear to cross the apparent capsu-lar cavity lateral to it. These characters show a marked simi-larity to the condition in Urodeles, as was pointed out byde Villiers (see Goodrich, 1930, pp. 259 and 278, and de Beer,1987).

Only a short statement will be made about the remainder ofthe auditory capsule, as this has already been described in theadult condition by de Villiers. Its ventro-lateral wall is piercedby the fenestra ovalis [jo), which is closed by a plate of denseirieaencliyme (figs. 13, 14, PL 10; fig. 17, PL 12; fig. 19, PL 13).


No operculum has yet been formed and there is no trace of anycolumella system; this latter, of course, is also absent from theadult skull (de Villiers), although the operculum is present. Themedial wall of the capsule is very well formed for a young larva.It is well shown in fig. 15, PI. 11. From the front backwards,it is pierced by: (1) an anterior acustic foramen (foa) (or moreexactly a joint opening to this foramen and to the tunnel forthe facial nerve); (2) two sma l l med i an a c u s t i c fora-mina {jam); de Villiers finds only one in the adult skull; (3) anendolymphatic foramen, more dorsally (fen); (4) a superiorperilymphatic foramen (fps); and (5) an inferior perilymphaticforamen (fpi). Just antero-laterally to this last foramen, thetrue floor of the auditory capsule is pierced by a large opening,the fenestra rotunda. However, there is no through way outof the apparent capsular cavity to the subcranial space below,because the cranial basal plate is carried out as a ventral floorwhose outer edge is fused to the lateral wall of the capsule.Thus the fenestra rotunda opens into a small extra-capsularspace, from which, however, there is no ventral outlet. Thiscavity has the false appearance of being a part of the truecapsular cavity, and the inferior perilymphatic foramen openslaterally into it. There is no preoccipital arch in front of thejugular foramen, whilst the occipital arch is well developed andis fused to the side wall of the auditory capsule in the usual way.The relations of the auditory capsule to the palatoquadratewill be dealt with later on (pp. 119 to 121).

(g) The C a r t i l a g e Cran i a l Eoof.There is no complete tectum synoticum as yet, although the

anterior border of it is probably represented at the sides bythe two cartilage projections, ts, figs. 12, PI. 10, and fig. 15,PI. 11; a complete arch is present in the adult (de Villiers).There is no taenia tecti medialis nor t. t. transversalis; theseare also absent in the adult. The sides of the roof are formedby the slightly overhanging upper edges of the orbital cartilages,which pass back into the heavily built flanges of the taeniaeuecti marginales (tm, fig. 10, PI. 9; fig. 11, 12, PL 10), whichproject inwards from the upper edges of the capsules.

118 H. K. PUSBY

(h) T h e P a l a t o q u a d r a t e .In all modern-type Anuran larvae, the articular region of the

quadrate lies far forwards in the snout, below the olfactoryforamen, but in Ascaphus it is no further forwardthan the region of the pila metop t ica , betweenthe optic and oculomotor foramina. All the parts ofthe quadrate bar lie correspondingly far back and all are veryheavily built, whilst the attachments to the cranium show a mostexaggerated autostylism. The absence of any obvioussubocular vacui ty between the cranial wall andthe p a l a t o q u a d r a t e also di f ferent ia tes Ascaphusfrom all other frogs.

(i) The Commissura Quadra to-cran ia l i s Anter ior .This quadrato-cranial connexion (cqa) is present as a grea t ly

extended flange fused to the ventro-lateral edge of thetrabecula, from the level of the notch (nnt, fig. 19, PI. 13) forthe internal nostril to a position close in front of the anteriorcapsular face; this posterior limit is shown as a darkly shadedline in fig. 17, PI. 12, and as a dotted line ( ) in fig. 15, PL 11,and is marked in each case as pcqa. Pig. 5, PI. 7; figs. 6, 7, PL 8 ;fig. 8, PL 9; and fig. 17, PL 12, show clearly that the anteriorhalf of the commissure acts as a support between the trabecula(t) and the dorsal surface of the pterygoid process of the palato-quadrate (ptc). Such a clearly marked pterygoidprocess is a unique feature in Ascaphus; in allother frog larvae the process is obscured in the body of thequadrate owing to the more anterior position of the parsarticularis (Pusey, 1938, p. 545 and Text-fig. 7E of this paper.)The poster ior half.of the commissure joins thet rabecula to the body of the quad ra t e , behindthe pars a r t i cu la r i s . About half of this posterior sectionof the commissure is covered from above by the ascending pro-cess (pa, fig. 15, PL 11) and it is this fact which, in part,accounts for the apparent absence of any subocular vacuitywhen the skull and jaws are viewed from above.

(ii) The Poster ior Spur of the Quadra te .A solid spur of cartilage projects ventrally from the under side


of the quadrate bar, behind and somewhat medially to the parsarticularis. The lower jaw abuts against it from in front, andthe tip of that part (saq, fig. 21, PI. 14) of the ceratohyal (ch),which articulates with the quadrate, abuts against it frombehind (fig. 19, PI. 13). This spur is presumably the homologueof a similar spur described in E a n a (Pusey, 1938, p. 504).

(iii) The P t e r y g o i d B o n e - r u d i m e n t ( = l igamen-tum q u a d r a t o - e t h m o i d a l e ) .

Starting posteriorly, close to the inner side of the spur (fig. 7,PL 8) and continuing forwards along the under side of thepterygoid process (fig. 5, PI. 7; fig. 6, PI. 8) there lies a band ofdense ligamentous tissue (Iq) which is almost certainly therudiment of the pterygoid bone, although this will require finalconfirmation in older stages; it is also certainly the equivalentof Gaupp's quadrato-ethmoidal ligament of E a n a . Where thepterygoid process ends in front, the ligament becomes free andforms the ventro-lateral border of the notch (nnt) for theinternal nostril. It passes forwards to its insertion on theconjoined trabecular horn and supra-rostral system. Pig. 19,PI. 13, shows the full length of this ligament (Iq), part of whichis dotted, where it passes behind the lower jaw.

(iv) T h e ' P o s t e r i o r Basa l P r o c e s s ' o f the Quad r a t e .

Behind the spur, the body of the quadrate is drawn downinto a ventral keel, whose under surface is grooved with al o n g i t u d i n a l notch (nc, fig. 17, PI. 12), into which the upperedge (saq, fig. 21, PI. 14) of the ceratohyal is fitted; in otherfrogs this same groove lies t r a n s v e r s e l y . As this keel passesbackwards, it becomes more and more detached from the bodyof the quadrate and finally projects posteriorly as a free process,the 'posterior basal process' (bpr, fig. 10, PI. 9; fig. 11, PI. 10;fig. 17, PI. 12), which is fused behind to the ventral side of theantero-lateral, free end of the basitrabecular process (bt, figs.11,12, PI. 10; fig. 17, PL 12; fig. 19, PL 13). It will be seen fromfig. 17, PL 12, that the posterior edge of the commissure (pcqa)passes laterally and becomes the inner border of this posteriorbasal process and therefore delimits and encloses the small

120 H. K. PUSEY

subocular vacuity in front. An a r t i c u l a t i o n b e t w e e n ap o s t e r i o r b a s a l p roce s s of t h e p a l a t o q u a d r a t eand t h e o u t e r end of a c o m p l e t e b a s i t r a b e c u l a rp roces s from t h e c r a n i a l floor i s , in Ascaphus ,a f e a t u r e u n i q u e a m o n g f rogs . In D i scog lo s sussuch a basal process is still present in miniature (Pusey, 1938,p. 580, footnote, and Text-fig. 6, bpr, of this paper), but only theo u t e r end of the basitrabecular process is present in the formof a pseudobasal process, from metamorphosis onwards. Allo t h e r f rogs , i n c l u d i n g t h e o t h e r m e m b e r s of t h efami ly D i s c o g l o s s i d a e , have los t t h i s p o s t e r i o rba sa l p r o c e s s .

(5) The A s c e n d i n g P r o c e s s and L a r v a l Ot ic P r o -cess .

From the dorsal side of the body of the quadrate the ascend-ing process (pa) rises up as a broad, flat strap of cartilage whichslopes inwards and upwards to become fused to the orbitalcartilage at the top of the pila antotica (pia, fig. 8, PI. 9; fig. 15,PL 11; fig. 19, PL 13); it forms the roof of a 'profundus tunnel',as explained on p. 114. Nerves V 2 and 3 pass out behind andabove it, from the dorsal division of the trigeminal foramen(fig. 9, PL 9; fig. 23, PL 14) whilst V 1 passes forwards below it,in the tunnel. The posterior borders of both the ascending pro-cess and the general body of the. quadrate are fused to theanterior tip (tac) and to the side-wall (/) of the auditory capsule(see particularly figs. 9,10, PL 9; figs. 11,12, PL 10; fig. 22, PL 14);oac, in fig. 15, PL 11, and/, in fig. 19, PL 13, show the line alongwhich the quadrate is fused tQ the capsule. This fus ionis no t the l a r v a l o t i c p rocess b u t is a u n i q u ef e a t u r e in Ascaphus, helping to give an increased auto-stylic support to the quadrate bar. A similar state of affairs,however, is probably present in the larval skull of X e n o p u s ,though this has not previously been recognized (see p. 171). InD i s c o g l o s s u s , the back of the quadrate fits on to the tipand the side of the capsule by a loose, ball and socket joint; yetthere is no fusion of cartilage between them. Farther back atthe side of the auditory capsule, in A s c a p h u s , this fusion is


interrupted at one point by a foramen (fvd, fig. 22, PL 14, andin the other reconstructions) which allows a branch (vclb) ofthe lateral head vein to pass downwards from the overlyingjaw muscles to join the main venous trunk (vd) below (see fig. 13,PI. 10). That part of the quadrate which lies behind this fora-men and which is again fused to the capsular wall, is the larvalotic process (pot, all reconstructions). See discussion, p. 167.

(vi) The Muscu la r P r o c e s s .The muscu la r process of the q u a d r a t e , so

t y p i c a l of all o t h e r frogs, is scarce ly p r e sen tin Ascaphus as a we l l -deve loped s t r u c t u r e . It is,in fact, represented by the whole extent of the lateral, upperborder of the body of the quadrate from the quadrate tunnel(at, left side of fig. 6, PL 8) to the larval otic process, inasmuchas it is this part which gives origin to hyoid muscles (ohm) tothe ceratohyal, and (sam) to the jaw (fig. 5, PL 7; figs. 6,7, PL 8;figs. 8, 9, 10, PL 9; figs. 11, 12, 13, 14, PL 10). These facts fullybear out the argument previously given (Pusey, 1938, p. 523,&c.) that the whole outer border of the larval quadrate bar,from the larval otic process behind, to the front border of themuscular process in front, is but an elongation of the shorterprocess seen in less modified Vertebrate types. The moreobviously upraised anterior part of the border of the quadrateof Ascaphus is marked as pmq, in figs. 19 and 20, PL 13,and in the sections shown in fig. 7, PL 8, and fig. 10, PL 9. Seediscussion, p. 168.

(vii) An A r t e r i a l Tunne l t h r o u g h the Q u a d r a t e .Finally, attention must be drawn to a tunnel (at), which

passes through the body of the quadrate from behind forwards,below the muscular process. This is well shown in the recon-structions, particularly in figs. 19 and 20, PL 13, and in thesections drawn in figs. 6 and 7, PL 8. It transmits a branch(cbr) of the carotid artery, which supplies the anterior ends of theadductor jaw muscles. It is a unique feature in Ascaphus . Seediscussion, p. 173.

(i) The Lower J a w Sys t em.The lower jaw is seen in lateral view in fig. 19, PI. 13, inKOS. 334-5 K

122 H. K. PUSBY

dorso-iateral view in fig. 21, PI. 14, in anterior view in fig. 20,PL 13, and in section in figs. 4, 5, PL 7, and figs. 6, 7, PL 8.As in most other frogs, the larval jaw is divided into fourcartilages. The anterior jaw cartilages (qjc) call for no specialdescription; they are small rods held together medially by acopula of dense mesenchyme and their outer ends articulatewith the inner borders of the posterior cartilages. In front ofthem the epidermis is cornified into a single, minute tooth-blade. The posterior cartilages, on the other hand, are large,heavily built plates. The notable features about them are theirgreat antero-posterior extension and the postero-ventral spurswhich they carry on their under sides. These two features assistin increasing the degree to which the floor of the pharynx iswalled in by cartilage and are probably adaptations helping toresist the pressures set up when the sucker system is in use.

The huge ventral adductor jaw muscles (Impm) have mostextensive insertions on the whole concave upper faces of thesecartilages; assisted by the ventral spurs, they probably playan important part in the peculiar method of progression usedby tho larva, whilst it remains adhering to the substratum byits sucker. The development of the excessive autostylism ofthe whole upper jaw system is probably an evolutionary responseto meet the strains set up by these heavy muscles.

(j) The H y o - b r a n c h i a 1 A p p a r a t u s .

In the head of all modern-type frogs so far described manystructures, of which the hyo-branehial apparatus is one, havemoved forwards into the snout, see pp. 146. Consequentlyin these frogs the posterior ends of the gill bars, which lie wellin front of the occipital region of the skull, are wholly hiddenby the skull in dorsal view. In the Urodeles, however, the endsof the bars are clearly visible for some distance behind theocciput. This is t r u e also of Ascaphus, as will be seenif lig. 15, PL 11 is superimposed on fig. 21, PL 14, so that thelower jaw in ouch is made to coincide. It is thus true to sayfJiiit tho whole h y o - b r a n c h ial a p p a r a t u s lies muchl'iU'i.hci' back in Asuaphus t h a n in any o t h e r f rog .

A comparison of %. 'M, PL 14, of this paper with figures of


Gaupp's reconstruction of the larval hyo-branchial apparatusof E a n a t e m p o r a r i a (Gaupp, 1904, and as copied in manytext-books, e.g. Goodrich, 1930, fig. 471) brings out a numberof differences and points of interest. Of great importance isthe distortion which must have taken place progressively in theevolution of the Ranid and modern-type condition, from themore primitive condition which Ascaphus has retained andishares with the Urodeles. This distortion extends to the cera-tohyal and to each of the branchial bars behind it. The evolu-tionary change has entailed the movement forwards, in successiveontogenies, of the centre of each cartilaginous arch relativeboth to its lateral (dorsal) end and to its medial (ventral) attach-ment to the hypobranchial plate (hbp). Thus, in A s c a p h u s ,each arch is somewhat concave along its anterior edge, whilstin Eana , each is convex anteriorly. This change of shapebecomes more marked on passing from the last branchial archbehind to the ceratohyal in front. In fact, the change is sogreat in the ceratohyal, that whereas, in A s c a p h u s , thisbar projects postero-laterally at an angle of about 45° to thelong axis of the head, in Ran a it has become a short bar,lying transversely across the pharynx floor, with its axis almostat right angles to the axis of the head. In Ascaphus theupper edge of the bar (saq) articulates with a l o n g i t u d i n a lgroove in the under side of the quadrate (nc, fig. 17, PL 12),whilst in E a n a , this groove has become transversely placed(Pusey, 1938, nc, fig. 1, PL 33). This rotation in the modern-type frog is correlated with a shortening in length, which pre-vents the outer end of the ceratohyal from projecting far outbeyond the overlying quadrate bar in a way that would upsetthe stream-lining of the whole larval head. It is shown on p. 168that this rotation and shortening is also correlated with thedevelopment of a tall muscular process on the quadrates ofmodern-type frogs; thus the presence of a muscular process isbut a further aspect of the general forward migration of thesplanchnic structures.

Fig. 21, PL 14, should be self-explanatory. It is only necessaryto & 'ess those characters in which Ascaphus differs fromRan a. In Ascaphus the whole hyo-branchial apparatus

124 H. K. PUSBY

is more robustly built and more strongly consolidated. Theinner borders of the ceratohyals overlap the basibranchialcopula (bbc, figs. 9,10, PI. 9; fig. 11, PI. 10), whilst their posteriorprocesses are fused to the bases of the 1st branchial arches(fu, fig. 21, PI. 14). The whole basibranchial copula (bbc) ismuch more heavily built, and the two hypobranchial plates arefused to it and are not separated from one another in the middleline behind. All these characters make for solidity.

The ventral end of the first ceratobranchial joins the tip ofthe posterior process of the ceratohyal, and together these tworun back to join the base of the second ceratobranchial; allthree are then fused to the edge of the hypobranchial plate.The third ceratobranchial is separately fused to this platefarther back, whilst the fourth arch is entirely independentof it, its ventral (anterior) tip ending freely in a ventralposition (see right side). This freedom of the fourth bar isnot found in other frogs, but is a character shared with theUrodeles.

Dorsal spicula (sp II and III) are present on the edges of thehypobranchial plate, overlying the points of fusion of thesecond and third ceratobranchials; the first pair of the series,as found in E a n a, is absent. What appears to be the fourthof the series is a pair of long, strong, rounded cartilages, firmlyfused to the hypobranchial plate and reaching back to thetiny glottis. Personal observations on victoria-blue preparationsand sections of the hypobranchial apparatus of E a n a t em-p or a r i a show that the fourth spicula make up a considerablepart of the thyroid processes of the adult, being combined withwhat remains of the hypobranchial plates and being added toby some new cartilage growth. It is likely, therefore, that themuch larger fourth spicula of A s c a p h u s will also be found tomake up the thyroid processes of the adult. The adult ' hyoid'apparatus has been figured by Frazier (1924, fig. 11, PL 2).It possesses well developed thyroid processes, each, however,with a forked end. L i o p e l m a h o c h s t e t t e r i has equallywell developed, but unforked, processes (Trewavas, 1933, fig. 4).Thus the terms 'thyroid process' and 'fourth spieulum' areprobably largely synonymous.


But this is not all. In A s c a p h u s each fourth spiculumis supplied with a subarcualis rectus muscle (just as the IVtharch is) and with a subarcualis obliquus muscle (like the IVth,Illrd, and Ilnd arches). T h e r e f o r e , the possibi l i ty-shou ld be kep t in mind t h a t the 4 th sp i cu lum isr ea l l y a Vth b r a n c h i a l a r c h , s t i l l q u i t e welldeve loped in Ascaphus and supp l i ed wi th musclesl ike the o the r a r c h e s . In Discoglossus and E a n athere is no larval muscle supply to the fourth spiculum and thewhole structure is relatively insignificant.

I hope to make further studies and to publish figures of therelations of this probable Vth arch to the surrounding gill clefts,muscles, and nerves. Should this suggestion prove to be correct,Ascaphus would become further noteworthy as being theonly living tetrapod to retain a recognizable Vth branchialarch.

Each arch has a solid, rounded ventral part, which expandson passing backwards, into a broad, flat plate, which is quiteheavily built and has (at least as yet) a smooth outline, withoutthe ragged branchial rays typical of E a n a and other frogs.The under sides of the Ilnd and Illrd arches are not joinedby any processus branchialis, such as is found in E a n a .This junction is also absent in fully formed larvae of Dis-cog los sus , although here a long process reaches forwards fromthe Illrd arch towards the Ilnd, with which, however, it is stillunfused. In this small point, too, then, Discoglossus isintermediate between Ascaphus and E a n a .

Each gill bar passes dorsally and Nos. I to III have bifurcatedand inrolled, dorsal ends which overlap one another. Theposterior prong of each anterior arch overlaps the anteriorprong of the arch behind; in these places the arches are fusedto give terminal commissures (tc, I-II, Text-fig. 4, p. 140),though this is not apparent from fig. 21, PL 14, except in thecase of the Illrd and IVth arches. The dorsal end of the 1starch carries a long, forwardly directed dorsal process (dpba I)underlying the otic region of the palatoquadrate (figs. 13, 14,PL 10, and Text-figs. 1 and 2, pp. 132 and 133). (Note: theIllrd bar, on the right side, is presumably atypically developed

126 H. K. PUSEY

and is so deeply bifurcated that its dorsal end is split into twoindependent parts.)

The under side of the basibranchial copula (bbc) is carrieddownwards and backwards into a wide ventral keel, whichdivides behind into two slightly diverging prongs {ufbc), whichbecome free of the plate and underlie the hypobranchial region.These prongs are seen in section in fig. 12, PI. 10, and in Text-fig.1, p. 132. They a re t h e h o m o l o g u e of t h e ' u r o b r a n -ch ia le ' , a p rocess of t h e b a s i b r a n c h i a l copu la(copula II) of t h e Urode 1 es and have many of the samerelations to structures such as muscles, thyroid glands, andventral aorta. Thus, their tips embrace the conus region of theheart and the vertically placed ventral aorta (va), whilst theirouter borders give attachment to the subarcuales obliqui(saom II to V) and the recti ceryicis muscles (ran); the thyroidglands (thg) lie above them. Such a w e l l - m a r k e d u ro -b r a n c h i a l e l emen t is u n i q u e among f rogs ; usuallythere is a small, undivided knob on the under side of the copula.

5. A COMPARISON OF CERTAIN CRANIAL MUSCLES OF

ASCAPHUS, UBODELA, AND OTHER ANURA.

(a) P r e f a c e .It was not my original intention to give here an account of

the cranial muscles of the larval A s c a p h u s . A study of thejaw and branchial cartilages, however, showed many uniquefeatures in the related musculature and revealed that, in certainrespects, A s c a p h u s shows a considerable similarity to theUrodeles, whilst, in other respects, it is more primitive eitherthan that group or the Gymnophiona. I hope, at a laterdate, to be able to publish figures of the cranial muscles ofA s c a p h u s , compared with the muscles of Urodeles, Dis-coglossuH and l i ana , but, in view of the war situation,I have decided to set down the facts already worked out andgive here a simple account of the muscles of A s c a p h u scompared with those of certain other Amphibians.

I have not had time to study, in detail, all of the ratherextern-five and involved literature on Amphibian musculature.1 have relied, therefore, largely on the account given in Edge-


worth's (1935) monograph on 'The Cranial Muscles in Verte-brates', particularly as his tables of synonyms give great helpwith the earlier literature, to much of which, however, I havereferred for special points. For convenience I have usedEdgeworth's nomenclature, so far as is possible, but this doesnot mean that his interpretations have necessarily been accepted.It is clear from what follows that certain of his generalizationsmust be considerably modified, in view of the findings fromAscaphus . In addition to. Edgeworth's book I have founda more recent paper by Eaton (1936), on the Urodele Di-c a m p t o d o n e n s a t u s , to give a clear statement of thecranial muscles of the Amblystomid Salamanders comparedwith the muscles of other Urodeles.

Chiefly, however, the following accounts are based on personalobservations on various developmental stages of Ambly-s toma t i g r i n u m , Sa 1 a m a n d r a macu 1 osa, Ascaphust r u e i , Discoglossus p i c t u s , Bombina v a r i e g a t a ,and R a n a t e m p o r a r i a . No detailed discussion of themuscles will be given at the present time, but points of particularsignificance will be stressed as they arise.

(b) The Mand ibu l a r Muscles.

Modern - type Frogs in Genera l .Edgeworth (1935), following earlier work, identifies ten pairs

of mandibular muscles in the typical, modern-type frog larva.They are given in the following diagram, which aims at showingtheir origin in development (after Edgeworth's account).

(1) /levator_-! ma

Mandibularmuscleplate

masticatory 1muscles /

(2) 1. m. extemus

(3) 1(4) 1(5) 1

a. subexternusa. lateralisa. articularis

«,(intermandibularis 1I muscles j -

(6) 1. m. p. superfieialis(7) 1. m. p. profundus

- (8) intermandibularis ant.• (9) mandibulo-labialis

-(10) intermandibularis post.

(This list takes no account of the levator bulbi muscle which isa further outgrowth from the 1. m. anterior muscle.)

128 H. K. PUSBY

A s c a p h u s .In A s c a p h u s only five pairs of muscles are independently

represented; these are probably nos. 5 (or ? 2), 6, 7 (or 7 ? + 1),8, and 10. This absence of subdivision of the muscle-blocksinto small muscles which, in the modern-type frogs, are insertedon the supra-rostral cartilage (nos. 2, 3, and 4) and on thelarval lips (no. 9) would seem to be a primitive character inA s c a p h u s . There is, however, some difficulty in establishinghomologies when the early development is not known, whilstnames, which are devised for muscles in a subdivided condition,often cannot be accurately applied to the muscles of a simplersystem. Consequently some of these names have been usedguardedly for the muscles of A s c a p h u s and these are pre-ceded by a question mark. The choice of names has been guidedby the general topography of the muscles and by the nature oftheir origins and insertions. As an example of the type ofdifficulty, we may take the case of no. 5, the ? 1. m. p. articularismuscle. This muscle, in A s c a p h u s , has just the position,origin, and insertion of the muscle of that name in the advancedfrogs, and so this name has been used for it; but it may well bethat it is actually also the rudiment of muscles nos. 3 and 4 aswell as 5. There is also a further alternative. It seems to menot impossible that this muscle may prove to be no. 2, the1. m. externus muscle, which, in development, has grown backto the otic process (= anterior edge of the muscular process)below ne rve V 3, i n s t e a d of above i t ; this couldhave possibly come about through nerve V 3 being carried farup dorsally on the huge mass of the adductor muscles of thejaw. The advantage of this theory, which can only be proved bya study of the early development, is that it brings A s c a p h u sinto line with the Urodeles, all of which possess a large 1. m.externus muscle but none of the muscles 3 to 5, which aretypical of the frogs. But again it must be stressed that therelations of the muscle in question to nerve V 3 are not those ofthe 1. in. externus of the Urodeles. The relations in A s c a p h u sare clearly shown in lig. 5, PI. 7, and fig. 6, PI. 8 (Imam and V 3).Also against the theory is the fact that the 1. m. externus muscle


is absent in Bombina and is very poorly developed inDiscoglossus of the Discoglossidae.

All the other muscles, except no. 8, the intermandibularisanterior, are well shown in fig. 3, PI. 6; figs. 4, 5, PL 7; figs. 6, 7,PI. 8; figs. 8, 9, 10, PI. 9; figs. 11,12, 13, 14, PI. 10, and call forlittle description. The very great development of ? 1. m. p.profundus should be noted; this is probably in part due to itsuse by the tadpole in connexion with the special sucker mechan-ism and mode of larval progression. Some part, however, ofits great bulk is presumably due to the fact that there is noindependent 1. m. anterior muscle, whose substance is thereforeprobably represented in this muscle, thus increasing its size.The absence of an independent 1. m. anterior is surprising,for such a muscle is present in all other Amphibia. It may benoted, however, that the most medial of the fibres making upthe huge muscle block take origin from the orbital cartilageabove and in front of the ascending process and therefore infront of the joint exit of nerves V 2 and 3; this can be seen infig. 7, PL 8, and fig. 8, PL 9. The remaining fibres of the blockarise from the quadrate and the roof of the auditory capsulepostero-laterally to the nerves. These media l f ibreshave j u s t the r e l a t i o n s , t h e n , of the 1. m. a n t e r i o rmuscle of the Urode les , and though not inde-p e n d e n t , p r e s u m a b l y r e p r e s e n t th i s muscle fusedto the 1. m. p . p ro fundus muscle . The fibres con-cerned remain medial to the rest of the muscle block and areinserted on the roof of the mouth cavity and the median sideof the jaw. They do not pass under the 1. in. p. profundusmuscle to attain a lateral position anteriorly, as is customaryin all modern-type frogs except the Pipidae. In this respectagain Ascaphus is like the Urodeles. Ascaphus is alsounique among frogs in having such 1. m. anterior fibres arisingin the primitive position high up on the orbital cartilage infront of the two branches of the trigeminal nerve. In otherfrogs, where a large subocular vacuity has been evolved, thehead of the muscle has moved downwards through the vacuityand arises on the front face of the auditory capsule below theascending process and on its under side.

130 H. K. PUSEY

Another point of interest is found in the origins of the 1. m. p.superficialis and profundus muscles, which lie far back over theauditory capsule. In all modern-type frogs these muscles arisewholly from the quadrate and therefore in a more anteriorposition. This posterior origin, in A s c a p h u s, is presumablyrelated to the relatively posterior position of the lower jawsystem, but whether it is a primitive character or not is hardto decide; see also p. 170.

The hinder edge of the intermandibularis posterior muscleunderlies the front edge of the interhyoideus muscle (see fig. 9,PI. 9). In this small point Ascaphus is like the Urodelesand unlike the other frogs.

Urode les .The mandibular muscles of Urodeles are present as follows

(Edgeworth):

(i(1) { mamlibulac } •< (2) 1. m. externus

I levntor \j mamlibulac } •< I(anterior J

I masticatory 1I muscles /*""-—-__ |levator

..(3) J mandibulaeMandibular\ / (posteriormuscle 1/plate | \ I intermandibularis^

... , , ^ ^ ' I anteriorintormandibular I _^— I in somemuscles / O r < \ ^/intermandibularis genera only.

t 5 ' (posterior J

E a n a t e m p o r a r i a .E a n a possesses all the muscles listed on p. 127; the 1. m. a.

lateralis, however, only develops just before metamorphosisand is essentially an adult muscle.

Discoglossus p i c t u s .All ten pairs of muscles are • present in Discoglossus .

L. in. externus is small and is absent on one side in one specimen.L. m. a. lateralis lies laterally, and not medially, to nerve V 3and 1. m. a. subexternus is divided into two slips. Thus, Dis-coglossus is in no way primitive in its mandibular muscles.

B o m I) i n a v a r i e g a t a.In Bom bin a 1. in. externus is absent; 1. m. a. subexternus

is in two slipH, and 1. m. a. articularis lies medially to V 3.

flevator 1-I hyoidei |~-~—^

(6) interhyoideus oi

_^(2)

~^(4)

/(6)r \ ( 7 )

orbitohyoideus \

suspensorio-hyoideus Isuspensorio-angularis 1qiuulrato-angulariahyo-angularis )

interhyoideus posterior J

i Depressor[ mandibula&I of adulti

in some


(c) The Hyoid Muscles.

Modern-type Frogs in General.

Edgeworth's description of the origin and'form of thehyoidmuscles of the modern-type Anura may be expressed as follows:

Hyoid Imuscle !<plate j

Asca'phus.

Of the above muscles Ascaphus possesses only four pairs,which, from their relationships, may be identified as nos. 1(?+2), 3, 4,-and 6; they are all clearly seen in fig. 5, PI. 7;figs. 6, 7, PI. 8; figs. 8, 9,10, PL 9 ; figs. 11,12,13,14, PI. 10; andText-figs. 1, 2, and 4 (pp. 132, 133, and 140), and are labelled as:(I)ohm, (3) sam, (4) qdm, and (6) Him. In a d d i t i o n Asca-phus possesses a pai r of muscles homologous wi thno . 3 of t h e U r o d e l e s , i .e . the b r anch io -hyo ideuse x t e r n us muscles (bhem); see pp. 132 and 134 below.

There is no independent suspensorio-hyoideus muscle, thoughsome of the postero-dorsal fibres (shm) in the block marked asthe orbitohyoideus run a more vertical course from the larvalotic process to the tip of the ceratohyal; they do not constitutea separate muscle, however. The hyo-ahgularis muscle isentirely absent, but the suspensorio-angularis muscle appears tohave two heads, arising from the quadrate, one above and onebelow the hyomandibular branch of the Vllth nerve and abranch of the carotid artery (VII hm and cbr, fig. 7, PI. 8;figs. 8, 9, PI. 9), yet the fibres form one muscle block only; if,however, the lower head shifted to the ceratohyal (fig. 9, PI. 9)and the muscle divided into two, an independent hyo-angularismuscle would result.

A large orbitohyoideus muscle (ohm) is present, a r i s ingall a long the u p p e r , ou t e r edge of the q u a d r a t e

132 H. K. PUSEY

b a r , f r o m i t s a n t e r i o r t i p b e s i d e t h e q u a d r a t et u n n e l ( a t ) , a l m o s t t o i t s o t i c r e g i o n (figs. 6, 7,PI. 8; figs. 8, 9, 10, PI. 9; figs. 11, 12, 13, PL 10); it passesdownwards and backwards, to be inserted on the end andouter side of the ceratohyal (Text-figs. 1 and 2). Similarly, thesuspensorio-angularis muscle (sam) takes origin from the underside of the quadrate, r i g h t b a c k t o i t s l a r v a l o t i c

Labm I

• ,„ '., saomll \saomllltoV\ sarml

Is ohm

TEXT-FIG. 1.

Section 11-1-7, through line vv in fig. 21, PI. 14. For key to letteringsee p. 181.

p r o c e s s . T h u s , n e i t h e r t h i s musc le nor the or-b i t o h y o i d e u s is r e s t r i c t e d to a loca l i zed q u a d r a t em u s c u l a r p r o c e s s , as in t h e more a d v a n c e d f rogs .See discussion, p. 168.

Of some importance is the presence of a pair of branchio-hyoideus externus muscles. Each is a moderately large bundleof fibres inserted on the posterior tip of the ceratohyal (bhem,Text-fig. 2), largely on its median side, behind and above theinterhyoideus muscle and passing upwards and backwards, toarise on the outer side of the first ceratobranchial, near itspoBtero-dorsal border and behind the insertion of levator I


(Text-fig. 4). These are the relations of this same muscle inthe Urodeles, with the difference, that in Ascaphus theinsertion of the muscle lies very far back on the tip of theceratohyal and not along its whole under side as in the Urodeles.This muscle has not been described before inthe An ura, though both Discoglossus and Bombina

bal Imm.l abml

shm 'ch

ohm Is

TEXT-FIG. 2.

Section 11-4-6, through line ww in fig. 21, PI. 14. For key to letteringsee p. 181.

retain it as a small bundle throughout the larvalperiod; in this respect again the Discoglossid frogs are inter-mediate between Ascaphus and the more advanced membersof the group.

The interhyoideus muscle does not extend back as far as theposterior tips of the ceratohyals, and there is no separate i.posterior muscle in the opercular wall.

The relative simplicity of the hyoid muscles on the one handand the presence of the branchio-hyoideus muscles on the otherare obviously primitive characters in Ascaphus.

134 H. K. PUSEY

U r o d e 1 e s.

The hyoid muscles, in form and origin, are as follows(Edgeworth):

(1) levator hyoidei (= depressor mandibulae)_(2) hyo- (or branchio-) mandibularis (in some only)

Hyoid ) ^-ii-^^_-^(3) branchio-hyoideus externusmuscle Jplate J

(6) subhyoideus (in some).interhyoideus or«dH~ —ffl interhyoideus.

" -(5) interhyoideus posterior.

Muscle no. 2 above seems to be wholly unrepresented inA s c a p h u s and the other frogs.

D i s c o g l o s s u s p i c t u s .

Muscles 1, 2, 3, 4, and 6 of the anuran series are present inD i s c o g l o s s u s . No. 5, the hyo-angularis muscle, is absentas in A s c a p h u s , though it is present in B o m b i n a . Theinterhyoideus posterior muscle is absent as in A s c a p h u sand R a n a . The suspensorio-angularis and the suspensorio-hyoideus muscles take origin from the auditory region of thequadrate as in A s c a p h u s (where the orbitohyoideus — thesuspensorio hyoideus; see p. 131, above); this is unlike the con-ditions in the modern-type frogs, where the origins lie in theregion of the muscular process. The quadrato-angularis musclearises medially to the ceratohyal-quadrate articulation as inA s c a p h u s and not laterally to it as in E a n a.

A small branchio-hyoideus externus muscle is present oneach side as in A s c a p h u s and the Urodeles; it is also presentin B o m b i n a . Of this muscle Edgeworth (1935, p. 106) says:' The Branchio-mandibularis externus' (sic, which from hisp. 103 is clearly a misprint for ' brancfiio-h y o i d e u s externus')' of larvae of Urodela has no homologue in Dipnoi, Apoda, andAnura' and 'appears to be a secondary larval structure'. Thisstatement clearly now needs revision.

From the above it is clear that D i scog lo s sus retainsmany small primitive characters which it shares with Asca-phus and sometimes with the Urodeles.


R a n a t e m p o r a r i a .

Ran a possesses all the muscles nos. 1 to 6 of the anuranseries, but no. 7, which is present in other genera, e.g.P e l o b a t e s and Bufo, is absent.

(d) The Leva to res Arcuum B r a n c h i a l i u mMuscles .

A s c a p h u s .Leva tor I (labm I, figs. 13, 14, PL 10; and Text-figs. 1

and 2) takes origin from the outside of the membraneous skullroof, over the otic process of the quadrate and runs downwardsand a little backwards, to be inserted on the outer side of thefirst arch and on the base of its dorsal process (dpba I). Itextends about a quarter of the way down the arch, in frontof the efferent artery and the attachments of the branchio-hyoideus externus muscle and the first constrictor muscle.

Leva to r I I arises from the cartilage wall of the lateralsemicircular canal and from the sheath of the neck muscles andruns backwards and downwards to be inserted on the outerside of arch II, a quarter of the way down its posterior face infront of the efferent blood-vessel (labm II, Text-fig. 4).

L e v a t o r I I I arises postero-ventro-medially to L. II, fromthe wall of the posterior ampulla of the capsule and from thesheath of the neck muscles. It passes backwards horizontallyand is then folded forwards as a < , round the back of arch III,and passes a little way along its under side.

L e v a t o r (-{-constr ictor) IV takes origin wholly fromthe sheath of the neck muscles, under the back of the auditorycapsule. It passes backwards horizontally to loop forwardsagain behind the last gill pouch and be inserted for somedistance along the inner, posterior face of arch IV (labm IV,Text-fig. 5, p. 141).

Urode les .Edgeworth (1935, p. 131) says: 'Four Levatores arcuum are

present in larvae. Levator I arises from the auditory capsuleor the dorsal fascia, Levatores II, III, and IV from the dorsalfascia. They are inserted into the dorsal ends of the kerato-

136 H. K. PUSBY

branchialia.' All four arise from the back of the capsule inTri ton alpestr is (Litzelmann, 1923, fig. 9).

Modern-type Anura.The levatores (of which no. IV in all and nos. I to IV in the

Pipidae, are undivided levators+constrictors) take origin farforward on the wall of the auditory capsule (Pipa andDiscoglossus), or on this and the quadrate (Xenopus,Bufo, and Eana) , or all on the quadrate (Pelobates; Edge-worth, p. 133 and personal observations).

Thus Ascaphus shares the more pr imi t ivearrangement of the Urodeles, with only a slightchange towards the advanced anuran type . Dis-coglossus is also more pr imi t ive than most otherfrogs so far described.

(e) The Constrictores Branchiales Muscles.

Ascaphus.Constr ictor I is a muscle of about five fibres attached to

the ventro-medial end of the under side of arch II (cbm I,Text-figs. 2 and 3). It passes' laterally and comes to underlieand run with the 1st afferent artery under arch I (Text-figs.3 and 4); it then loops backwards and upwards to its dorsalattachment on the extreme postero-dorsal tip of arch I, wherethis is in fusion with arch II at the terminal commissure. It isnot in continuity with levator I.

Constr ictor I I . The size and the relations of C. II toarches II and III are just those of C. I to arches I and II(Text-figs. 3 to 5).

Constr ictor I I I is a muscle of one fibre, lying in a loopunder the arteries of arch III. It ends freely against the wallsof the arteries both above and below, is not inserted on anycartilages and does not reach so far dorsally or ventrally asCs. I or II. See Text-fig. 5, p. 181.

There is no separate constrictor IV, though levator+con-strictor IV reaches relatively further ventrally than the levatorsdo in arches I to III. There is no constrictor V.


Urodeles.

There are no constrictor muscles in Urodeles, unless theyare represented by the small, dorsally lying depressores branchi-arium muscles to the external gills; such occur in arches I to III,like the anuran muscles.

In this character, then, Ascaphus and the other frogs areunlike the Urodeles.

Discoglossus .

Constrictor I is just as in Ascaphus . Cs. II and III areboth attached to the incipient processus branchialis from thebase of arch III. As in Ascaphus , C. Ill ends against theefferent blood-vessel at its dorsal end and is not inserted onthe cartilage arch; it is made up of two fibres only. C. II isattached dorsally to arch II.

E a n a .

Cs. I and II are attached to the complete branchial processventrally. C. I l l ends below against the blood-vessel and notagainst the cartilage, but it is attached to arch III dorsally.

(/) T h e S u b a r c u a l e s E e c t i Muscles.

Ascaphus .Suba rcua l i s r ec tus I arises on the under side of arch I

(sarm I, Text-fig. 1), and runs forwards and inwards, to beinserted on the posterior edge of the ceratohyal, just laterallyto its posterior process (fig. 12, PI. 10).

Suba rcua l e s ree t i IV and ? V. There is no S. rectusII. Small bundles of fibres arise from the under sides of spiculumIV (? Vth gill arch) and arch IV (sarm V and IV, Text-fig. 4).The two bundles run forwards and outwards and combine theirfibres into a larger, single bundle which passes under arches IIIand II without receiving fibres from them,1 to be inserted onarch I, at a point just behind the origin of S. rectus I (Text-figs.3 and 2).

1 From transverse sections it is not easy to be absolutely certain thatno fibre, are added to the muscle from arch III; certainly none are addedfrom arch II.

NOS. 334-5 L

138 H. K. PUSEY

These are p r e s u m a b l y s egmen ta l e l ements ofb r a n c h i a l segments V and IV, combined a n t e r i -or ly in to a s ingle musc le , i n s e r t e d on arch I .

Urodeles.S. rectus I is present as in Ascaphus . There is, of course,

no Vth cartilage arch in the Urodeles. The remaining S. rectimuscles (S. rectus IV only, of Edgeworth) have differentrelations from those of Ascaphus , although, undoubtedly,

''• c b m l

ra

TEXT-FIG. 3.

Section 12-1-7, through line xx in fig. 21, PL 14. For key to letteringsee p. 181.

from their position and relations, they are the homologues ofthe anuran muscles. Edgeworth calls them all S. rectus IVand Eaton (1936) uses this same nomenclature. Eaton writes,on p. 65: 'The commonest and most primitive condition' inthe Urodeles ' is for this muscle to originate on the fourth archand insert on each of the three in front of it, continuing, however,as a single muscle. In Amblys toma larvae and those of theSalamandridae this condition is maintained.' I have checkedthe accuracy of this statement on both transverse and longi-tudinal sections of A m b l y s t o m a , S a l a m a n d r a , andN e c t u r u s , and there is no doubt that all the f ibresare attached to the IVth arch and that they are then sharedout in smal le r s l ips which are attached to arches III, II,and I. This arrangement differs from that in Ascaphus and


Discoglossus , where the fibres are attached in smallerbundles on the pos t e r io r arches and join to a singlela rger bundle , all of whose fibres are a t t a chedto the 1st a rch .

Discoglossus .S. rectus I is as in Ascaphus . There is no S. rectus II,

or V. There is, however, a H. rectus III which is probablyabsent from Ascaphus . S. recti III and IV arise on theunder sides of arches III and IV and, mingling their fibres, areinserted as one bundle on arch I, like the Vth and IVth musclesof Ascaphus .

E a n a (in the fully formed larva).S. rectus I passes from arch I to the ceratohyal. ? S. rectus II

passes from the Ilnd arch and the anterior face of the branchialprocess to the ceratohyal, joining with S. rectus I in front.There is no S. rectus III, whilst S. rectus IV runs from theIVth arch to the posterior face of the branchial process.

(g) The Suba rcua le s Obliqui Muscles.

Ascaphus .Small bundles of fibres arise on the under sides of (1) spiculum

IV (=?arch V), (2) arch IV, (3) arch III, and (4) a largerbundle on arch II (saom V-II, Text-figs. 4, 3, and 2). Bundles1 and 2 fuse together on passing forwards (Text-fig. 2, right side)and later fuse again with bundle 3 (Text-fig. 1); the resultingtwo bundles pass forwards and inwards (ventrally to the arterialarches, but dorsally to the subarcuales recti muscles mentionedabove) and are inserted, one behind the other, on the outer sideof the urobranchial prong near its tip (Text-fig. 1).

Ascaphus thus possesses four pai rs of sub-a rcua les obl iqui muscles , one pair in each ofb ranch i a l segments I I to V.

Urodeles .There are only two pairs of oblique muscles in Urodeles;

these arise on the under sides of ceratobranchials II and III

140 H. K. PUSEY

and pass forwards and inwards, fuse together and insert by singletendons, either on the urobranchial prongs, or on the sheathsof the recti cervicis muscles in this region.

Mode rn - type Annra .

In the Anura, including the Discoglossidae, there is but a

bci- bhem

tram

TEXT-FIG. 4.

Section 12-3-1, through line YY in fig. 21, PL 14. For key to letteringseep. 181.

single pair of oblique muscles, present in branchial segment II,corresponding to the largest of the four pairs in A s c a p h u s .Arising on the under sides of the Ilnd arches, the muscles passinwards and slightly forwards to meet one another in a medianraphe which is attached loosely by ligaments to the posteriortip of the undivided urobranchial keel of the basibranchialcopula. This keel, it may be noticed, is better developed and isdeeper in Discoglossus than in other frogs.

Edgeworth (1935) gives just such an account of this musclepair on pp. 161 and 162 of his monograph, yet he calls each musclea 'transversus ventralis II ' . He thus fails to recognize the


obvious homology with the oblique muscles of Urodeles (andAscaphus), his failure probably being due to his theoreticalconceptions of the primitive, ancestral content of muscles inany one branchial segment (see discussion on p. 142).

In respect , therefore, of the subarcuales obli-qui muscles, Ascaphus is the most primitive ofthe living t e t r apods . The Urodeles retain a half

LabmlV t y m I V

<iba II! :

dbm-"'

TEXT-FIG. 5.

Section 12-4-8, through line zz in fig. 21, PI. 14. For key to letteringsee p. 181.

and the modern-type frogs a quar ter of thisancient inher i tance .

(h) T h e T r a n s v e r s i vent rales Muscles.Ascaphus.

The only transversus ventralis muscle present in Ascaphusis that in branchial segment IV. (The transversus ventralis II,of Edgeworth's nomenclature for the Anura, has been describedabove as part of the subarcuales obliqui system.) The fourthtransverse muscle arises in mesenchyme close to the inner,posterior border of the IVth arch, as a poorly developed sheetof fibres which passes inwards and upwards, to meet its fellowin a median raphe in the middle line, just in front of the verysmall glottis; further back, the glottis divides the two muscles(Text-fig. 5, tvm IV).

Jlodern-type Anura.Such a muscle pair begins to develop in young, modern-type

142 H. K. PUSBY

larvae, but disappears again very early, even in the Discoglos-sidae.

Urodeles.The muscle pair is represented in Urodeles by a very large

sheet of muscle lying between the IV arches and even spreadingto other attachments.

The r e t e n t i o n of a pai r of t r a n s v e r s u s v e n t r a l i sIV muscles in to the wel l -deve loped la rva is afu r the r c h a r a c t e r which Ascaphus alone shareswith the Urode les . (A dilator laryngis, with which thistransversus ventralis IV might be confused, is also present inAscaphus . )

(i) Discuss ion .On page 162 of his monograph Edgeworth writes: 'The

primitive condition of the Subarcuales recti was probably one inwhich each was attached to the branchial bar of its segment oforigin and extended forwards to the next anterior bar, and thatof the Transversi ventrales one in which each passed to a medianraphe. Dipnoi, Teleostomi, and Amphibia probably had aSubarcualis rectus and a Transversus ventralis in each branchialsegment.' On p., 156 he writes:' In Urodela Subarcualis rectus Ipasses from the 1st branchial bar to the hyoid bar. The other,more posterior, muscles are modified in association with thedevelopment of the urobranchial outgrowth of the basibran-chiale. In genera with four branchial bars Subarcuales II andIII diverge inwards and unite in a common tendon which isinserted into the urobranchiale, forming Subarcuales obliquiII and III. Subarcualis rectus IV grows forwards beyond thenext bar . . . in three fasciculi inserted into the 1st, Ilnd, andIllrd bars.' Clearly, then, Edgeworth's view is that a branchialsegment may contain a Transversus ventralis and a Subarcualismuscle and this latter can be either in the 'rectus' or in the'obliquus' form, but presumably both subarcualis types cannotbe present at once. He fits the facts to this theory by assumingthat his 'Subarcualis rectus I V muscle of the Urodeles is theproduct only of the IVth muscle plate, in spite of its threefasciculi to the Illrd, Ilnd, and 1st arches. Eaton (1936),


however, writes on p. 65: 'The muscle is built embryonicallyout of part of the ventral material of the second, third, andfourth branchial muscle plates', and Driiner, to quote fromEdgeworth, p. 157, found that in the Urodeles there is 'anadditional innervation of the anterior part of Subarcualis rectusIV corresponding to the segments into which it grows'; thisinnervation is from the branchial branches of the vagus nerve.These facts, then, surely indicate that the so-called Subarcualisrectus IV is really a combined form of Subarcuales recti IV, III,and II, where possibly the IVth has grown forwards to the1st arch and III and II have grown backwards to the IVth. If,however, we adhere to Edgeworth's idea that a branchial seg-ment cannot contain both a S. obliquus and a S. rectus at thesame time, we are immediately in difficulties in the case of theUrodeles, for they certainly contain oblique muscles in segmentsII and III, and we are now suggesting that these segments alsopossess recti muscles as well. Edgeworth himself overcame thisdifficulty by calling this rectus complex simply 'S. rectus IV .There was a further difficulty for him in the case of the frogs.He found that the advanced frogs possessed simple S. rectimuscles in each of segments I to IV in early development; butthey also possessed, on each side in segment II, the musclewhich has been called S. obliquus II on p. 140 of this paper.Edgeworth was compelled to overlook the clear homology whichthis muscle shows to the Ilnd S. oblique muscle of the Uro-deles and was forced to call it a Transversus ventralis II. Withthese reservations Edgeworth's system covered all the animalsdescribed up to 1935.

A knowledge of the muscles of Ascaphus introduces a new-difficulty for Edgeworth's theory. In this animal there are S.obliqui muscles in each of branchial segments II, III, IV, and? V, and there are S. recti muscles in segments I, IV, and ? V.This difficulty could be resolved by extending Edgeworth'ssystem of calling the S. obliqui muscles of the frogs the Trans-versi ventrales muscles and disregarding their obvious similarityto the Urodele S. obliqui muscles, if only there were no persistentTia.nsversus ventralis muscle in segment IV. This IVth segment,however, contains at one and the same time a S. rectus pair,

144 H. K. PUSEY

a S. oblique pair, and a T. ventralis pair. Consequently the S.obliqui muscles of Ascaphus cannot be disposed of as T.ventrales muscles, but must remain as S. obliqui. This means,then, that in segments IV and ? V there are both S. recti andS. obliqui muscles. A realization of this possibility allows usalso to call Edgeworth's 'T. ventralis II ' of the modern-typefrogs the S. obliquus II.

Certain changes must, thus, now be made. (1) We mustbe willing to face the probabi l i ty tha t a pr imi t ivebranchial segment could contain s imultaneouslya S. rec tus , a S. obl iquus, and a Transversusventral is muscle pair. (2) In future the term ' Transversusventralis II ' must be abandoned in favour of 'S. obliquus II ' inthe Anura. (3) The term ' Subarcualis rectus IV' in the Urodelashould probably be altered to 'S. recti IV, III, and II', especiallyas the ? Vth S. rectus muscle in Ascaphus and the Illrd S.rectus in Discoglossus have no connexion with arch IV at all.

Clearly, then, this whole subject requires further carefulresearch into the origins of these various muscles in the early de-velopmental stages of Urodeles, Ascaphus, and the modern-type frogs. It would also be most valuable to investigate thearrangement in L i o p e 1 m a. The attachments of the posteriorS. recti muscles are different in Ascaphus from those in theUrodeles and it is necessary to find out how these differencesarise in development.

(j) The Diaphragmato-branchia l i s Muscles.In Pelobates Schiilze (1892) described a small muscle

running from the posterior face of the IVth arch, near its top, tothe postero-lateral side of the pericardium (' diaphragm') just infront of the glenoid region of the pectoral girdle. This muscle isalso present in Ascaphus (dbm, Text-fig.5), Discoglossus,and R a n a ; yet, in the well-formed larvae of each of thesegenera, it is attached to the Illrd arch, near the terminal com-missure with the IVth and not to the IVth arch as Schiilzeheld (personal observation).

By description, this muscle would seem to be the homologueof the omo-urcualis muscle of certain Urodeles.


6. THE HYPOBRANCHIAL SPINAL MUSCLES OP ASCAPHUS.

(a) The Geniohyoideus Muscles.

The geniohyoideus muscles, in Ascaphus , pass from thepostero-median ends of the anterior jaw cartilages to the outeredges of the basibranchial copula and the hypobranchial plateswhere the 1st and Ilnd arches pass into them; parts of themuscles are attached on the bases of the 1st arches themselves.These relations are more or less those of the other Anura. Onthe right side of this particular specimen, however, (but noton the left) two fibres detach themselves from the inner borderof the muscle (unlabelled, figs. 9, 10, PI. 9; figs. 11, 12, PI. 10)and run an independent, backward course, resting on the inter-hyoideus muscle, to become attached, with the tip of the-rectuscervicis muscle, to the outer side of the right urobranehial prong,near its tip. In this small point this specimen imitates theUrodeles whose geniohyoideus muscles are attached along thefront border of the forked end of the urobranehial prong (e.g.Sa lamandra ) or on the urobranchiale and the recti cervicismuscles (e.g. Amblys toma) .

(b) The Rect i Cervicis Muscles.

In Ascaphus each rectus cervicis muscle is attached infront on the ventro-lateral tip of the urobranehial prong (raw,Text-fig. 1) and passes sharply downwards and backwards,round the antero-lateral wall of the pericardium, to be attached,behind, on the ventral membraneous sheets below the heart(Text-fig. 2). The muscles do not reach back to the girdle.The lower, front border of each muscle is made up of thin, youngfibres (rcsm, Text-fig. 1) which may represent a superficialbundle. The muscles are very much smaller than in the Urodeles,and do not fill up the space above the urobranchiale and belowthe hypobranchial plates. They are also smaller than in themore advanced frogs, including Discoglossus. Thea t t a c h m e n t of the muscles to the u robranch ia leis a cha r ac t e r which Ascaphus shares with theUrodeles, although in the Urodeles the huge muscles have

146 H. K. PUSEY

more extensive attachments spreading beyond the urobranchialeto the basibranchiale and to one or two hypobranehialia.

In the other Anura, the muscle is attached to the bases of theIlnd and Illrd arches (in Discoglossus) , or to the base ofthe Illrd arch and the processus branchialis (in Eana), whilst(in both genera) some fibres start from mesenchyme near thesubarcualis rectus IV muscle. The fibres are not associated withthe small urobranchial keel of the basibranchial copula.

7. DISCUSSION.

(a) The P o s i t i o n of the Sp lanchn ic S t r u c t u r e sof Ascaphus and o the r Frog L a r v a e .

The foregoing descriptions make it clear that the splanchnicstructures of the larval Ascaphus have an arrangementrather dissimilar to that found in any other of the survivingfrogs. For this reason the term ' modern-type' anuran has beenused throughout this paper to sum up all the other frogs takentogether, in contrast to the more primitive type shown byAscaphus (and presumably by L iope lma , though thisform has not been described in its larval condition). The smallerdifferences between the larvae of the modern-type frogs areinsignificant when measured against the differences which anyone of them shows when compared with Ascaphus . Thehead structures of all modern-type frogs show but slight varia-tions from a single basic plan, but the structures of Ascaphuscannot be fitted to this plan without considerable reservation.

Following Noble (1924,1927,1931, &c.) and de Villiers (1934),it is customary now to look upon the Liopelmidae, which is thefamily containing A s c a p h u s , as the most primitive familyof living frogs, and the facts set out in this paper give strongsupport to this view. Consequently, a detailed knowledge ofthe anatomy of Ascaphus , throughout its life history, maybe expected to be of first-rate importance to the understandingof the origins and relations of the isolated and somewhat ano-malous group, Anura, with its peculiar larval specializations.

Luther (1914, p. 94 et seq.) attributed these larval anuranspecializations to a single cause. Edgeworth (1935) is in agree-


ment with Luther's thesis, and writes as follows (p. 237):'Luther was of the opinion, and I think rightly, that the causesare to be found in the great length of the intestine—an adaptationto vegetable food—which brought about a ballooning of the perito-neal cavity and its forward extension. This produced a forwardmigration of all organs lying in front of the peritoneal cavity andbelow the chondrocranium. . . . The influence of the elongationof the intestine is probably not direct and mechanical but heredi-tarily fixed.' Whether these writers have picked on the right causeor not, and there seems to be no good reason to dispute their view,there can be no doubt that such an extensive forward movementof all the splanchnic structures has taken place in phylogeny, andit is this forward movement which has given to the larval, modern-type anuran head its very peculiar arrangement. What follows onp. 168 of this paper makes it clear too that even the possession ofthe peculiar muscular process on the quadrate by larvae ofmodern-type frogs is a direct result of this general forward move-ment, though perhaps this had not been realized before.

If a Urodele is taken as a standard of the more typicalVertebrate for the purposes of comparison, it will be seen thatthe following structures in the modern-type frogs have all movedforwards from their ancestral position:

(1) The trabecular horns (and their probable derivatives,the supra-rostral system).

(2) The nasal sac and its surrounding capsule.(3) The lower jaw system.(4) The whole quadrate bar, particularly:

(a) the articular region;(b) the commissura quadrato-cranialis anterior (= the

'anterior basal process', see p. 154);(c) the pterygoid process;(d) the muscular process (= the specialized anterior end

of the ancestral otic process).(5) The position of the mouth opening.(6) The columella system (and annulus) of the ear apparatus.(7) The hyobranchial apparatus as a whole and the centre

of each arch in particular.

148 H. K. PUSEY

(8) The Eustachian tube ( = hyoid gill pouch) and subsequentgill slits.

(9) The mandibular, hyoid, and branchial muscles and thehypobranchial spinal muscles, their nerves and blood-vessels.

(10) The heart and arterial arches.(11) Certain bone rudiments, particularly of the following

bones:(a) pterygoid;(b) quadratojugal;(c) premaxillary;(d) maxillary.

In e v e r y one of t h e s e c h a r a c t e r s (exceptno.6,andnos.8 (in part) and 11) t h e l a r v a l head of A s c a p h u s is cer-tainly less modified from the primitive ver tebra tep lan than are the larval heads of the modern- typefrogs. There is no evidence in relation to characters 6 and 8 (inpart), since the adult Ascaphus has no middle ear cavity ortube, nor columella apparatus or annular cartilage (de Villiers,1934). I have found no certain trace of any of these charactersin the single larval specimen studied, but see p. 173. Further,this particular specimen is too young to show traces of the rudi-ments of the squamosal, premaxillary, and maxillary bones;however, since the cartilages to which these bones are related liefurther back in Ascaphus, the bones themselves must also liemore posteriorly when they come to be formed, in later stages.The pterygoid rudiment (= the quadrato-ethmoidal ligament)is certainly farther back, and this is true also of the ' trabeeular-quadrate ligament', from part of which the quadratojugal boneis ossified in Ban a ; however, no quadratojugal bone is formedin Ascaphus (Noble, 1931, and de Villiers, 1943).

It would thus seem that Ascaphus is a pers is ten t lyprimitive frog which supplies a most valuable link in the historyof the evolution of the Anura.

(/;) Is the Cranial Ground-plan of Ascaphusreally Pr imi t ive ?

It is well at this point to face a legitimate criticism. Thetadpoles of Ascaphus live in fast-flowing water and in this


connexion possess a powerful sucker formed from the larval' lips'; they also employ a peculiar method of progression whilstthey are attached to the substrate and a specialized method ofparticle feeding (Noble, 1927). Such adaptations might beexpected to be correlated with considerable changes in the ana-tomical ground-plan of the larval head. It could be argued that,if Luther is right in believing that a s ingle cause can havemodified a large number of structures away from the primitivevertebrate plan, some o t h e r s ingle cause might befound responsible for a general reversal of these modifications.It could be suggested that perhaps the requirements of a suckersystem had s e c o n d a r i l y forced back all the splanchnicstructures to their primitive position. If this were true, theanatomy of the larval A s c a p h u s would only simulate theprimitive condition, whilst actually being secondary in nature,and this animal would not be a true guide to an understandingof the evolution of the advanced frogs.

A number of small points from the anatomy of Ascaphuscould be brought forward in apparent support of this view. Forexample: (1) the medial piece of the supra-rostral system seemsto be too far back to be just a simple downturning of theprimitive trabecular horns; (2) the membraneous 'posteriornarial tube' perhaps suggests a backward migration; (3) theposterior origins of the adductor jaw muscles from the auditorycapsule could be interpreted on this theory; (4) the lower jawlies transversely across the throat and does not take the formof a U with its anterior end well forward in the snout, in spiteof the fact that its articular facets are extraordinarily far backfor a frog; (5) the absence of an independent levator mandibulaeanterior (pterygoid) muscle may be suspicious, and could beexplained as due to the forcing back of a more anterior com-missure; (6) the presence of a pre-oral buccal cavity wouldagree with this view; (7) and finally, Save-Soderbergh's idea ofthe nature of the anuran commissure would fall into line withthis possible theory.

It seems to me to be likely that a slight backward movementof structures has taken place in the f ront of the snou tin connexion with the consolidation of the supra-rostral system,

150 H. K. PUSEY

but, for the reasons set out below, I am convinced that therehas been no g e n e r a l b a c k w a r d m o v e m e n t of al l t h es p l a n c h n i c s t r u c t u r e s in the evolution of the skull ofA s c a p h u s . It is worth noticing here, however, that thereprobably are characters in A s c a p h u s which are correlatedwith the sucker mechanism and the mode of progression. Suchprobably are: (1) the general heavy build of the head cartilages;(2) the great strength of the adductor jaw muscles; (3) theexaggerated autostyly of the palatoquadrate with its extensivecommissura and with the additional fusion of the auditorycapsule to the posterior border of the ascending process; (4)the extensive and rigid fusion of the central part of the supra-rostral system to the ethmoid region of the skull; (5) the greatantero-posterior width of the posterior jaw cartilages, with theirposterior spurs; and (6) the consolidation of the whole hyo-branchial apparatus and the width of the individual arches.

Against the view that the cranial plan of A s c a p h u s hasbeen derived from the modern-type plan by a secondary returnof its structures to an apparently more primitive posteriorposition, several points may be brought forward. (1) F r o me v i d e n c e l a r g e l y u n c o n n e c t e d w i t h i t s h e a ds t r u c t u r e s , A s c a p h u s has been placed with L i o p e l m ain the most primitive family of living frogs. The order Anurais marked off from the other vertebrate groups, among otherthings, by the larval specializations of its members. Thesespecializations extend particularly to the splanchnic headstructures, as shown above, and must have been the result ofvery great evolutionary changes. It would be surprising, there-fore, if the most primitive living member of the order failed toshow some trace of the ancestral plan. It would be more sur-prising still if, in fact, that imposing array of apparentlyprimitive characters which A s c a p h u s admittedly possesseswas due to some secondary cause and was not a simple retentionof the ancestral inheritance. In view of the other evidence tobe given below, the general contention that A s c a p h u s ispersistently primitive seems to me to hold good in spite of thefact that such 'link-animals' usually retain a footing in themodern world only at the exf>ense of specialized modifications

HEAD OF ASCAPHUS 15]

to fit some particular niche. The modifications which haveallowed A s c a p h u s to survive are perhaps those related tothe sucker apparatus; this is better developed than in any otherfrog, and its evolution has permitted Ascaphus to colonizethe fast-flowing streams in which its tadpoles develop.

It thus becomes important to disentangle the primitive fromthe specialized in A s c a p h u s . This perhaps can best be doneby extending anatomical studies to embrace L i o p e 1 m a whoselarval adaptations are those to suit a terrestrial development(Archey, 1922) and not a development in fast streams. Thosecharacters, then, which A s c a p h u s shares with Liope 1 mamight be accepted as part of their common ancestral inheritance.But even this test may prove to be not wholly critical, since boththese Liopelmid frogs are without a middle ear apparatus (deVilliers, 1934, and Wagner, 1934, 1 and 2), although such was,of course, present in the ancestor and even in so close a relativeas Protobatrachus (Piveteau, 1937). Whether this apparatuswas lost by the more recent common ancestor of the Liopelmidae,or whether each of the modern genera has lost it separatelycannot be known at present. The same type of difficultymay therefore be expected to concern other characters showncommonly by the two frogs, in view of the great Amphibiantendency to parallel evolution which has been brought out byWatson's researches. Investigations of the larval modificationsof other frog tadpoles adapted to fast water may give help inthis important problem. Unfortunately, I have not had accessto material either of L i o p e l m a , nor of any of the mountain-brook tadpoles.

(2) In many of those characters in which Ascaphus differsfrom the modern-type frogs it approaches more nearly to theUrodeles; see the list on p. 175. This is significant and militatesagainst the contentions put forward by certain Scandinavianworkers (Holmgren, 1934, and Save-Soderbergh, 1934 and 1935)that the Anura and the Urodela are separately derived fromthe Fish. A knowledge of the Urodele head is of great importancein the interpretation of the head of A s c a p h u s .

(3) The family Discoglossidae is the group of frogs whichNoble believes to be the next most primitive after the

152 H. K. PTJSEY

Liopelmidae. My own unpublished studies of the larval, cranialstructures of D i scog los sus p i c t u s show that this frog is, inmany small points, intermediate between A s c a p h u s and sucha modern-type frog as E a n a t e m p o r a r i a . Now the tad-poles of D i s c o g l o s s u s live in pond water and are not taintedwith the specializations connected with a sucker mechanism,yet their cranial structures are best explained by assumingthat there is some retention of the plan shown in A s c a p h u s .This would seem to prove that A s c a p h u s is truly primitivein its head form and not merely secondary. A list of thesimilarities of A s c a p h u s and Discog lossus is given onp. 177.

(4) The hyo-branchial apparatus, the jaw and the branchialmuscles may be picked out for special mention. In respect ofthese structures A s c a p h u s is almost more like a Urodelethan it is like the modern-type frogs, whilst in some points it isperhaps even more primitive than the Urodeles. This couldhardly be so if A s c a p h u s had passed through the modern-type evolutionary stage, to return, by later specialization, to asecondary state of apparent primitiveness.

(5) Finally, it is perhaps worth noticing that a forecast wasmade in a previous paper (Pusey, 1938) of many of the charactersto be expected in the larva of any ancestral frog. This forecastwas based on a study of the tadpoles of E a n a and on thepublished accounts of the larval and adult skulls of A l y t e sand on the adult skulls only of A s c a p h u s and L i o p e 1 m a.At tha t time nothing per t inent was known of thelarval s t ruc tures of Ascaphus. Yet these s t ruc-tures , as set out here, bear out t ha t forecast ina remarkable way and in a way hardly possibleif they are secondary and not pr imi t ive .

Prom the above evidence, therefore, it will be taken forgranted in what follows tha t larval Ascaphus suppliesa link in the chain of evolut ionary stages leadingfrom the ances t ra l to the specialized larva ofthe modern- type frog, of which, for convenience, Eanat e m p o r a r i a may be taken as an example.


(c) Ex t rapo la t ion to the Larval Ancestorof the Progs.

Speaking figuratively, our present knowledge of anuranarvae supplies us with three main points on the ' evolutionarygraph' which should extend from the primitive pre-anuranancestor at the one end to the modern-type anuran, such asE a n a, at the other. These three points are not evenly spacedalong the graph. (1) The point supplied by Eana (and theother modern-type genera) lies at the extreme specialized endof the graph. (2) The point 'Discoglossus ' lies close to it,out a little nearer to the ancestral end; Bombina andAlytes would lie close to Discoglossus, but on the sideof Rana. (3) Half-way down the graph lies Ascaphus,parted from Discoglossus and Eana by a big gap. Thepoint representing the ancestral pre-anuran would lie far beyondAscaphus on the side away from Eana . From our presentknowledge we are unable to fix this last point with any certainty,especially from the larval point of view, although the adultanatomy of the pre-anuran ancestor is becoming clearer fromthe recent work of Piveteau (1937) and Watson (1940); perhapsthe anatomy both of the larval and adult Urodele may be takenas a generalized guide.

If we think in this graphical way of the smooth set of changeswhich must have taken place in phylogeny, modifying theancestral pre-anuran plan to the plan of the modern-type frogs,we can think also of extrapolation. Our curve can be drawnbackwards, with fair certainty, through Eana , Discoglos-sus, and Ascaphus, and this is sufficient to give its generalcourse. By ext rapola t ion into past time we canbuild up for ourselves an idea of the ancestralorganizat ion, by re-versing the changes whichhave taken place along the known part of thecurve. We can thus obtain a fairly exact idea,also, of the homologies of the anatomical partspresent in the modern-type frog, and this islargely tLe purpose of the following analysis .

The difficulty in understanding larval anuran head structuresNOS. 334-5 M

154 H. K. PUSBY

has always lain with the interpretation of the parts of thepalatoquadrate. The supra-rostral system also supplies diffi-culties. Most other structures, however, though modified, areidentifiable in terms of, say, Urodele structure, and thereforecall for less attention.

(d) The ' A n t e r i o r B a s a l P r o c e s s ' ( = G a u p p ' s C o m -m i s s u r a ) and t h e ' P o s t e r i o r Basa l P r o c e s s . '

The two parts of the modern-type anuran palatoquadratesystem which have been the most difficult of interpretation arethe comniissura quadrato-cranialis anterior (of Gaupp) and thepseudobasal process (of de Beer, = basal process of Gaupp).The issue is further complicated (or perhaps simplified) by anadditional process—the 'posterior basal process' of this paper—which is present in A s c a p h u s , and which still just survivesin D i s c o g l o s s u s , but which is totally absent from all othermodern-type frogs, including A1 y tes and B o m b i n a , of thefamily Discoglossidae.

It will be unnecessary to enter into any discussion of thenature of the pseudobasal process of modern frogs until sectionshave been cut of the metamorphic stages of A s c a p h u s , forthe larva of A s c a p h u s throws no new light on the problem,since the process is not yet formed at the stage here studied.But, in connexion with this structure, it may be pointed outhere that the larval A s c a p h u s possesses two processes, eachof which, in part, represents the typical Craniate basal process(see below). Neither of these, however, is the same as thepseudobasal process of modern frogs, so that A s c a p h u s sup-ports the contention of a previous paper (Pusey, 1938), that,whatever may be the nature of the pseudobasal process, it is notthe quadrate basal process, as Gaupp and others have believed.

In the m o d e r n - t y p e frog t a d p o l e s , t h e co minis-su ra is a r e l a t i v e l y n a r r o w band of c a r t i l a g e ex-t e n d i n g from the leve l of the o l f a c t o r y fo ramenin f ron t , to a p o s i t i o n m i d - w a y be tween t h i s andt h e o p t i c fo ramen beh ind (see Pusey, 1938, fig. 8, andparticularly Parker's papers, 1876 and 1881, and also Text-fig.7 ]<i of this paper). Between its posterior edge and the front of

HEAD OP ASCAPHUS

Ct

155

pmcj ^

sov . ._

TEXT-FIG. 6.

Discoglossus p ic tus . Reconstruction of the neurocraniumand palatoquadrate of a young larva in ventral view, to show thepresence of a small 'posterior basal process', bpr. Overall lengthof larva 15-2 mm. and length from anus to tail tip, 8-7 mm.; thesupra-rostral and lower jaw systems have been removed. For thekey to the lettering see p. 181.

the auditory capsule lies the extensive subocular vacuity whichis such a striking feature of the modern-type tadpole's skull.In A s c a p h u s , however, the anterior border of the commissurebegins where the posterior border of the modern tadpole's com-missure ends, whilst the posterior border lies far back under the

156 H. K. PUSEY

ascending process and close in front of the auditory capsule.Thus the commissure in Ascaphus is a much widerband and lies much f a r t h e r back in the head . Asa result, there is no subocular vacuity visible in dorsal view;this space is only a small pear-shaped opening lying just in frontof the auditory capsule; the relations are clearly shown in fig. 17,PL 12, and Text-fig. 7 D. If, then, in the ancestor of A s c a p h u s ,the commissure was still farther back, as our extrapolation ideawould suggest, we must envisage the possibility that the sub-ocular vacuity was then entirely absent and that the hinderborder of the commissure abutted against the whole front borderof the basitrabecular process, which in Ascaphus is incor-porated in the floor of the auditory capsule. A result of thiswould be that the commissure would have formedone c o n t i n u o u s band wi th wha t has been ca l ledthe p o s t e r i o r basa l p rocess (bpr, fig. 17, PI. 12, andText-fig. 7 c and B), and this single cartilage boss so formedwould have projected backwards and inwards from the bodyof the quadrate, to articulate with the basitrabecular process.This boss would then have been the equivalent of the basalprocess of the Urodeles and other Craniates, except that it wouldhave been fused to the trabecula in front of the basitrabecularprocess, as well as simply abutting against this process behind.It is from this line of argument that I have been led to call thecommissure the ' anterior basal process' and the process markedin the figures as bpr the ' posterior basal process', for I considerthese two structures to be the products of the division of one,ancestral, basal process (Text-fig. 7 A). This argument waspreviously set out in another paper (Pusey, 1938), but at thattime it appeared that the commissure was the entire basal pro-cess shifted forwards, for the very small posterior basal processof Discoglossus had only been found after the text waswritten (see footnote on p. 530 of that paper), and, of course,nothing was then known of the large posterior basal process ofAscaphus .

Watson (1940) has shown that the position of the basi-trabecular process (as indicated by the bony basipterygoid pro-

has shifted backwards in the phylogeny of the frogs'


ancestors. In E u g y r i n u s w i l d i , of the Lower Coal Measuresof Lancashire (his fig. 13), it lay in front of the auditory capsuleand apparently independently of its floor. In M i o b a t r a c h u sromer i , of the Coal Measures of Mazon Creek, Illinois (hisfig. 4), it appears to have been a semi-independent part of thecapsular floor, still, however, partially separated from it by anotch. In P r o t o b a t r a c h u s m a s s i n o t i , of the BasalTrias of Madagascar (his fig. 19), it is the effective floor of thecapsule, as in A s c a p h u s . See also Watson's comparativesketches shown in his figs. 17 and 21. These comparisons bringout the fact that, in addition to shifting backwards, the basi-pterygoid process has also spread more and more laterally.

(e) The P o s t u l a t e d A n c e s t r a l L a r v a and i t sE v o l u t i o n to t h e M o d e r n - t y p e T a d p o l e .

With a knowledge of the tadpole of Ascaphus , coupledwith the findings of Watson's recent researches, we can hazarda guess at the l a r v a l conditions found in the ancestors of thefrogs. It is obvious that in the earliest forms, before larvalspecialization had been evolved, a single basal process from thepalatoquadrate met a basitrabecular process which projectedfrom the cranial base some way in front of the auditory capsuleand quite independently of it (Text-fig. 7 A). With the onsetof larval specialization in phylogeny two things seem to havetaken place. (1) The basal process gained a new and increasingfusion directly to the trabecula, in front of the basitrabecularprocess; this would have been the incipient commissure (Text-fig. 7 B). (2) The basitrabecular process developed further andfurther back in each ontogeny, and yet always retained anarticulation, or fusion, with the posterior corner of the quadratebasal process (posterior basal process) (Text-fig. 7 A to c). Withincreasing specialization these movements went in oppositedirections, the one forwards and the other backwards. As aresult, we may assume that the subocular vacuity opened up asa gap in the articulation between the now broad and dividedbasal process, on the outer side, and the retreating basitrabecularprocess on the inside (Text-fig. 7 c). These two forms of speciali-zation would then have continued steadily.


D ASCAPHUS

TEXT-PIG. 7.

The suggested evolution of the larval anuran palatoquadrate system.A, B, and C are hypothetical ancestors representing larval stagesof Eugyr inus , Mioba t rachus , and P ro toba t r achus(at least in so far as the position of the basitrabecular process isconcerned). D, larval Ascaphus, andE,larvalDiscoglossus.The ascending process in A, B, C, and D, parts of the otic processin B, C, and E, and the fusion of the ascending process with the tipof the auditory capsule in D have been cut away; cut surfaces aremechanically stippled.a, 'anterior basal process' = commissura quadrato-cranialisanterior; b, 'posterior basal process'; c, pterygoid process; d,ascending process; e, body of quadrate;/, fusion of pars articulariswith the under side of the pterygoid process; g, basitrabecular pro-cess and post-palatine commissure; h, trabecula; i, cut cartilagesurfaces. For the key to the lettering see p. 181.

160 H. K. PUS BY

Let us take the commissura first. Judging from the conditionin A s c a p h u s, we must assume that the anterior basal process(= commissura) rapidly increased its extent of fusion with thetrabecula, so that finally the whole body of the quadrate andalso the inner, upper border of the pterygoid process came to befused to the whole length of the opposing trabecular cartilage(see fig. 17, PI. 12, and Text-fig. 7 D). In this way the upperedge of the pterygoid process came to be masked in the com-missure, but in A s c a p h u s , in contrast to the modern-typetadpole, its lower edge still indicates a lengthy process otherwiseindependent of the body of the quadrate. The lower edge ofthis pterygoid process, which is shod all along its under sidewith the rudiment of the pterygoid bone (= the quadrato-ethmoidal ligament), bears to the mouth cavity, the trabecula,and the body of the quadrate the typical relations which arefound, for instance, in the fully formed Urodele larva.

With further specialization towards the modern-type tadpolecondition we must assume that two changes took place in thecommissure (Text-fig. 7 E). Firstly, that the position in whichit is attached to the trabecula became more and more anteriorin each ontogeny, until its front border reached the nasal sac.Secondly, that the antero-posterior width of the commissuredecreased and its posterior edge came to lie very much fartherforwards, in such a way that the fusion was now between thetrabecula and the pterygoid process only, rather than, as inA s c a p h u s , between the trabecula and both the pterygoidprocess and the body of the quadrate (see Pusey, 1938, fig. 8).As a result of these two changes an increasing vacuity wasopened up behind the commissure. Ascaphus has remainedat the stage where the evolutionary processes are only justbeginning, and where the subocular vacuity is insignificant andthe whole of the commissure is broad and lies far back in theskull.

The importance of this whole argument is that it stresses thepoint that t he commissu re of the m o d e r n - t y p e t ad -pole is a p a r t of t he ba sa l process of the q u a d r a t ewhich p r i m i t i v e 1 y belongs to a p a r t of the jawsys tem ly ing far back in the head . The commis-


sure is t hus not a neomorph and is not a s t r u c t u r ep r i m i t i v e l y be longing to the e thmoid region ofthe s n o u t . This contention is but a slight modification ofthat set out on pp. 528 and 529 of a previous paper (Pusey, 1938),but it is in sharp contrast to the views held by Gaupp (1906),Edgeworth (1925), Save-Soderbergh (1936), and Watson (1940),all of whom look on the commissure as essentially an ethmoidalstructure. Gaupp's and Edgeworth's theories have been criti-cized in the paper mentioned, but Save-Soderbergh's andWatson's views call for remark here.

(/) A Cri t ic ism of Save -Soderbe rgh ' s Theoryof the Na tu re of the Commissura.

In interpreting the adult structures of the snout of L y r o -cep~halus eur i , a triassic Stegocephalian, Save-Soderberghderived assistance from the l a rva l cartilages of Kana tem-p o r a r i a, one of the most specialized of present-day frogs. Asa result, he marks a postulated ethmoidal cartilage bridgebetween the front end of the quadrate pterygoid process and theethmoidal region of the skull of his animal, as a commissuraquadrato-cranialis anterior (his figs. 7 and 8). Had he beenable to use the then unknown structures of the more primitiveAscaphus for comparison, this interpretation would hardlyhave been possible. To use a pure ly l a rva l specializationof an admittedly specialized modern animal to interpret thestructures in an -a d u 11 of an ancient stock, particularly whenthe structure of the modern form is known to be lost at meta-morphosis, seems to me but to court disaster. Watson, in hisinterpretation of the skull of B r a n c h i o s a u r u s , has acceptedSave-Soderbergh's lead (see Watson, 1940, fig. 23, and p. 226).

It may be well to collect together the evidence against thistheory and to try to show finally that the commissure cannot beconsidered to be an ethmoidal structure.

(1) In Ascaphus the commissure is a mid-cranial and notan ethmoidal structure and, though its front border admittedlyreaches up to the unusually posteriorly-placed internal nostril,its hind border is in the auditory region (Text-fig. 7 D).

(2) In L y r o c e p h a l u s and Branch io sau rus the

162 H. K. PUSEY

connexion in question is between the tip of the pterygoid processof the quadrate and the ethmoidal part of the trabecula. Thewhole of the pterygoid process and of the pterygoid bone liesbehind it. In A s c a p h u s, on the other hand, the commissureconnects (1) the whole upper surface of the pterygoid process(and not just its tip) and (2) the greater part of the inner borderof the body of the quadrate, to the trabecula, along their re-spective lengths, so that no part of the pterygoid process, orof the pterygoid bone rudiment underlying it, lies behind thecommissure; in fact these structures lie rather in front of thecommissure (Text-fig. 7 D).

(8) In modern-type frog larvae the tip of the future adultpterygoid process lies in front of the anterior border of the com-missure, as the processus quadrato-ethmoidalis (see Pusey, 1938,fig. 8, pqe, and Text-fig. 7 E of this paper) laterally to the internalnostril and not postero-medially as Save-Soderbergh's theoryrequires.

(4) The pterygoid bone rudiment lies yet farther forwards inmodern-type frog larvae and not behind the processus quadrato-ethmoidalis.

None of this evidence, however, is finally conclusive, butthere are two further points.

(5) I am not satisfied that the interpretation of the impres-sions in the palatine bone, on which the theory of Save-Soder-bergh and Watson turns, requires the presence of a commissureas well as of an antorbital process, with its anterior and posteriormaxillary processes. It seems to me to be possible that thegroove which Watson attributes to the presence of a com-missure in B r a n c h i o s a u r u s (see his fig. 23) may have beendue solely to the presence of the thickened posterior edge of theantorbital process, whilst the groove attributed to a p o s t e r i o rmaxillary process may, in fact, have been due to the a n t e r i o rmaxillary process. In such a case the commissure would havebeen entirely absent. The slight distortion which these antor-bital cartilages had undergone, in terms of the Eanid (adult)condition, would be due to the 'great forward extension of theHub-temporal I'OBI-KI' in B r a n e h i o s a u r u s , which Watsonhimself mentions on p. 220. The same argument could apply to


Save-Soderbergh's findings in the case of L y r o c e p h a l u s .Thus the whole system, which he shows as the antorbital processwith its anterior and posterior processes in his figs. 7 and 8, etc,.could well be the forwardly-arched roof of the nasal capsulestanding out from the thickened edge of the actual antorbitalprocess which he marks as the commissure. This roofingcartilage is apparently pierced by a hole due to the presenceof a dorsal process of the dermopalatine bone, and it is this holewhich gives rise to the idea of a separate antorbital process infront of it. Now there is a similar hole in the cartilaginous nasalroofing of Di scog lossus (post-metamorphic stage) throughwhich the lateral branch of the profundus nerve passes from theorbit into the capsule. If this hole were larger, Discog lossuswould then parallel L y r o c e p h a l u s in this connexion, with-out, of course, retaining a commissure in this region.

It may be noted that neither Save-Soderbergh nor Watsonmakes use of the evidence which Edgeworth (1925) has collectedof an anterior connexion of the tip of the pterygoid process tothe base of the antorbital process in the Urodeles. This com-parison is, of course, ruled out for Save-Soderbergh owing to hisideas of the separate origin of the Urodeles from the Fish.Edgeworth himself considered this Urodele connexion to be thehomologue of the anuran commissure, but this homology issubject also to the very criticisms set out above and was, infact, so criticized in my previous paper.

(6) Finally, to stress the primitively posterior origin of thecommissure, one last point may be discussed, which seems to bethe most important in this connexion. Gaupp (1893 and 1906)and Pusey (1938) have both drawn attention to the fact that atmetamorphosis, in the modern-type frog larvae, a great partof the cartilage of the trabecula is destroyed by erosion. Thiserosion extends from under the ascending process behind, to thefront border of the commissure in front. It certainly results inthe breakdown of the attachment of the commissure to theskull and thus allows the whole quadrate bar to move freelybackwards (Gaupp and Pusey). But it has always been a puzzleas to why so extensive a length of cartilage is destroyed, fornot only is the outer part stripped off the whole central length

164 H. K. PUSEY

of the trabecula, but the entire bar is destroyed in places, sothat the optic foramen becomes vertically confluent with thecranio-palatine foramen, and the oculomotor foramen with theprimary carotid foramen. Now if fig. 17, PI. 12, fig. 19, PI. 13,and Text-fig. 7 D and E are studied, it will be seen that thecommissure of Ascaphus is attached to the trabecular baralong just this very region which is destroyed in the modern-type frogs. De Villier's work (1934) shows that the commissuremust be wholly detached from the trabecula at metamorphosisin Ascaphus , and this detachment is presumably achieved,as in E a n a , by the erosion of the cartilages concerned. Itwould appear, therefore, that once this band of trabecularcartilage had become sensitive in the ancestor to respond to thestimulus which brings about the erosion, this sensitivity,together with the activating stimulus, has been retained by thedescendants, so that the destruction has continued, though theneed for it posteriorly has been superseded by the forwardm i g r a t i o n of the commissure . The phenomenonis thu's an e v o l u t i o n a r y , phys io log ica l rel ic whichis important in that it supports the view that the commissurewas primitively farther back in the head, as it is to-day inAscaphus , and was not always an ethmoidal structure, as itis in modern-type tadpoles.

It is probable that a conclusive answer can finally be givento this question by a study of the larval stages of the otherLiopelmid frog, L iope lma . If this frog, with its own peculiarterrestrial development, is like Ascaphus in possessing acommissure lying far back in the temporal region of the head,it would leave little doubt that this, and not the ethmoidalregion, is the primitive site of this controversial structure.

(g) Changes in the A u d i t o r y Regions ofAnuran Skul l s .

Watson's researches (1940) make it clear that, whilst thecommissure was moving forwards in phylogeny, the basitra-becular process was moving backwards, so that it finally came tounderlie the front part of the auditory, capsule, with which itbecame fused. This absorption of the process into the capsule


has had far-reaching effects on the anatomy of this part of theanuran skull. It probably led (1) to the suppression of the rootof the process in L i o p e l m a and all modern-type frogs, butnot yet in A s c a p h u s (Text-fig. 7 D and B) ; (2) to the totalsuppression of the true cartilage floor of the capsule inA s c a p h u s . Then as the basitrabecular process became moreand more reduced in development—only its outer end remainingas the pseudobasal process—(3) the prefacial commissure, pre-sent in the Liopelmidae and the Discoglossidae, was lost in allother modern-type frogs; and (4) the auditory capsule regaineda true floor and side wall of its own above and beside the nowconjoined facial and trigeminal ganglia, thus refilling the gapsmade by the absence of the basitrabecular root and the prefacialcommissure. This capsular floor is already appearing late in thelarval development of the Discoglossid frogs; in the moremodern-type frogs (e.g. Kan a) it makes its appearance muchearlier in ontogeny.

In the early phylogenetic stages of the incorporation of thebasitrabecular process into the auditory capsule, the posteriorcorner of the palatoquadrate basal process must have keptcontact with the outer end of the process (Text-fig. 7 B, C, and D),as it does to-day in A s c a p h u s with its complete basitrabecularprocess fused to a 'posterior basal process'. I n D i s c o g l o s s u sa small remnant of this posterior basal process is still present(Text-fig. 6, bpr), but it is without a basitrabecular partnerduring larval life. At metamorphosis, however, it articulates,for a day or two only, with the young cartilage of the developingpseudobasal process (Text-fig. 7 B) (= the outer end of thebasitrabecular process, Pusey, 1938) and then is later totallydestroyed by erosion. This posterior basal process fails todevelop in any other modern-type frog so far described.

Attention should be drawn to the extreme similarity, evento details, of the anterior part of the auditory capsule and thebasal-basitrabecular articulation in Ascaphus and in theUrodeles. De Villiers (1934) has drawn attention to thissimilarity in the adult skull of A s c a p h u s . A comparison ofjg. 10, PI. 9, and figs. 11, 12, PI. 10 with the sections of Sala-m a n d r a figured by Goodrich (1930, fig. 453 c and D) brings

166 H. K. PUSBY

out the similarity; also fig. 17, PI. 12, may be compared withGoodrich's fig. 261. Probably this common condition is arrivedat by parallel evolution in the two orders. Such parallel evolu-tion is much more likely to have gone on in stocks which wereclosely related genetically than in stocks independently derivedfrom separate groups of Fish, in the manner suggested by theSwedish workers, Holmgren and Save-Soderbergh. There isthis difference, however, that in the Urodeles the entire basalprocess must have passed backwards along with the basi-trabecular process, without the anterior part having separatedoff to become a commissura quadrato-cranialis anterior. Thearticular region of the quadrate also remained far back in thisorder, so that the larva remained like the adult and no extensivemetamorphosis of these parts was necessary.

(h) The A r t i c u l a r Region and t he Body ofthe Q u a d r a t e .

In R a n a and other modern-type frogs the articular regionof the tadpole quadrate lies under the olfactory foramen; inA s c a p h u s it lies just behind the optic foramen. A comparisonof fig. 8, PL 36, of Pusey, 1938, with fig. 15, PI. 11, and fig. 19,PI. 13, shows how the change has been brought about. From theevidence of Craniates as a whole we know that primitivelythe body of the quadrate was a mass of cartilage lying under theauditory capsule. In Ascaphus this block has been elongatedanteriorly so that the quadrate is a fairly long band when seenfrom above and its articular region has moved in phylogenyalong the under edge of the pterygoid process (Text-fig. 7 D). Inthe change from the Ascaphus to the Ranid condition, thequadrate bar, as a whole, has moved forwards, carrying thecommissure and the pterygoid process with it, so that the frontlimits of these two structures have advanced from a posi-tion midway between the optic and olfactory foramina, inA s c a p h u s , to a position under the olfactory foramen inR a n a ; Text-fig, 7 E shows a midway stage in this evolution.But further, the articular region has moved forwards muchm o r e r a p i d 1 y than the main body of the quadrate, passingfrom the level of the optic foramen to a position in front of


the olfactory foramen (Text-fig. 7 D and E). It has done this bydeveloping, in successive ontogenies, further and further for-wards along the under side of the pterygoid bar, obscuring, asit did so, all but the quadrato-ethmoidal tip of this bar in thesubstance of the commissure and the body of the quadrate.During this general advance the cartilage of both the quadratebar and the commissure, as represented in A s c a p h u s , hasbeen lightened posteriorly by the formation in them of thesubocular vacuity. As suggested in the previous paper, the sub-ocular shelf (? of the trabecula) and Gaupp's processus pseudo-pterygoideus probably represent remnants of this suppressedcartilage in the modern-type tadpoles.

(i) The Otic P r o c e s s e s .Whilst the main body of the quadrate was moving forwards

in phylogeny, its outer, upper border—the otic process—wassimilarly being elongated in a forward direction; all the time,however, it retained its connexion with the wall of the auditorycapsule behind in the form of the 'tadpole otic process'. InA s c a p h u s the front border of the ancestral otic process hasreached forwards to the level of the oculomotor foramen, where-as in E a n a it has reached to the olfactory foramen (where it isrepresented by the front border of the quadrate muscular pro-cess; see Pusey, 1938, p. 523). Thus in all frogs thewhole l a t e r a l bo rde r of t he long q u a d r a t e barr e p r e s e n t s the much n a r r o w e r o t ic process of thea n c e s t o r and , as u s u a l , Ascaphus has r ema inedat t he ha l f -way c o n d i t i o n .

An examination of the figures of Ascaphus shows that there aretwo regions in which the posterior part of the quadrate bar is fusedto the wall of the auditory capsule. These regions are parted bya foramen (fvcl) through which a vein passes downwards to jointhe vena capitis lateralis, after draining blood from the overlyingadductor jaw muscles. An examination of E a n a t emp or a r iatadpoles in sections shows that a vein of precisely similar rela-tions runs backwards from the more forwardly-lying jaw muscles,1 i pass downwards over the hind edge of the quadrate bar,in f ront of the t a d p o l e o t i c p rocess , to join the

168 H. K. PUSBY

lateral head vein. For this reason I have marked the posteriorfusion in Ascaphus as the tadpole otic process. The anteriorfusion seems to be an a d d i t i o n a l s t r e n g t h e n i n g of theautostyly of the jaws achieved by a fusion of the tip of theauditory capsule to the hinder border of the ascending process.Unfortunately this critical vein seems to be wholly absent inDiscoglossus and in Bombina , where, too, the tadpoleotic fusion is also absent. In Discoglossus , however, theback of the quadrate bar is cupped and fits very closely againstthe rounded end and side of the auditory capsule, as a ball andsocket joint. There is, in fact, no connexion between the two,although this could easily be achieved and the condition pointsto the way in which this fusion has been brought about in theadmittedly specialized Ascaphus . The evidence could, ofcourse, be read in the reverse way: the additional fusion couldbe thought of as the primitive condition in the ancestral froglarva and a,s only recently lost in Discoglossus ; but it seemsmore probable that it is one of the secondary features inAscaphus .

(j) The Muscular P rocess .

The elongated otic process of Ascaphus gives origin tocertain mandibular and hyoid muscles. The two hyoid musclesmainly concerned are the suspensorio-angularis muscle (sam) tothe jaw and the orbitohyoideus muscle (ohm) to the ceratohyal.They take origin all along the under side and outer edge of thepalatoquadrate bar from the roof of the quadrate tunnel in frontto the tadpole otic process behind. There is thus no specializedand sharply upraised muscular process which, in modern-typefrogs, is such a characteristic of the larval palatoquadrate. Nonethe less, the front part of the quadrate bar over the tunnel isslightly raised above the general level, and I have consequentlymarked it as a muscular process in the figures, although it isbut an incipient one.

The fact that the muscular origins are widespread and notlocalized in Ascaphus upholds the view of an elongated oticprocess in this genus. We now, however, must face the questionof why there is a well developed and localized muscular process


in the modern-type tadpoles. The answer turns on three points.(1) The ceratohyal has moved to a more forward position in themodern-type frogs; (2) it has shortened and its posterior cornerhas been brought forwards relative to the bar as a whole;(3) as a result the primitive origins and insertions, particularlyof the orbitohyoideus muscle, would have been brought muchcloser together unless a muscular process was developed to carrythe origin up and away from the insertion. It is a well knownfact that a vertebrate muscle fibre can only reduce its length bycontraction to a certain fixed fraction of its resting length.Consequently, if a moving part of the skeleton is to be activatedto the same extent in two animals with different anatomicalrelations of the skeletal parts, the length of the muscle fibresconcerned must be maintained above a certain minimum value,and this can best be achieved by keeping the origin and insertionat a suitable distance apart. A comparison of fig. 19, PL 13,of this paper with fig. 1, PI. 33, of my previous paper makes theposition clear. In A s c a p h u s the fibres of the orbitohyoideusmuscle run a sufficiently long course to the tip of the ceratohyalto allow of the necessary contraction without the presence of atall muscular process. If, however, in Eana and the modern-type frogs there was no upraised muscular process the musclefibres would be very much shortened and would therefore beless effective. Consequen t ly , a muscu l a r process isdeve loped in the m o d e r n - t y p e frogs to keep theor ig ins and i n s e r t i o n s of the muscles s u i t a b l yp a r t e d from one a n o t h e r .

The change in orientation of the ceratohyal in advanced frogsis obviously a part of the general tendency to forward migrationof all splanchnic structures. The reduction in length of the barobviously prevents its posterior and now outer end from pro-jecting far out beyond the overlying quadrate in a way thatwould upset the streamlining of the larval head. When theceratohyal thus came to underlie one part only of the quadratebar in the modern-type frogs, instead of lying beneath thegreater part of the bar as in A s c a p h u s , the origins of theorbitohy )ideus and other muscles also came to be localizedabove it around the edges of the muscular process, instead of

NOS. 334-5 N

170 H. K. PUSEY

being spread out all along the quadrate otic process. Thus ,t h e r e can be no doub t t h a t the presence of amuscu l a r process in the m o d e r n - t y p e frog t ad -pole is but a necessa ry resu l t of the genera lforward movement of the s t r u c t u r e s u n d e r l y i n gthe sku l l . The actual muscular process is made by a modi-fication of the front border of the elongated ancestral oticprocess, as was suggested in a previous paper (Pusey, 1938,p. 523).

(fc) The Origins of the Adduc to r JawMuscles .

The same argument, as set out above, can be applied to thelength of the muscles fibres and to the variations in the origins ofthe mandibular muscles of A s c a p h u s on the one hand and ofthe advanced frogs on the other. The larval A s c a p h u s is uniqueamong frogs in that the superficial and deep posterior levatormuscles of the jaw take origin far back on the roof of the auditorycapsule, as well as on the quadrate bar. In all other frog tad-poles these muscles arise only from the hinder edge of thequadrate bar and its ascending process. Correlated with thisdifference is the different position of the lower jaw in the twoopposing types of organization. In Ascaphus the lower jawlies below and just in front of the level of the optic foramen(see fig. 19, PI. 13), whereas in other frogs it lies under the regionof the olfactory foramen (see fig. 1, PI. 33, Pusey, 1938). Thedistance between these two positions (as measured on either ofthe types concerned) is approximately equal to the distancebetween the sites of origin of the muscles in the two forms. Thus,since the jaw lies more posteriorly in Ascaphus , the muscleorigins must also lie farther back in order to give the muscles asufficient length to carry out effective contraction. But when,as in the modern-type frogs, the jaw moved forwards, carryingwith it the muscular insertions, the muscular origins were ableto descend from the capsular roof to a more forward position,without affecting the degree to which the muscles could shorten.Thus th i s difference in muscu la r or igins is also


r e l a t e d to the forward m i g r a t i o n of t he sp lanch-nic s t r u c t u r e s .

(I) The Ascend ing P rocess of t h eP a l a t o q u a d r a t e .

The ascending process is unusually broad and strong inAscaphus and is attached to the pila antotica and the orbitalcartilage high up along the side of the skull. This primitivehigh-level attachment is retained in the Discoglossidae; in otherfrogs, however, the attachment is rather to the trabecula, orthe base of the pila. The process appears to be very short inAscaphus because its hind border is completely fused to thefront face of the auditory capsule almost as far laterally as thelarval otic process. The normal long chink between the processand the capsule is only represented medially by the dorsaldivision of the prootic foramen (ftd) and laterally by the smallvenous foramen (foci). The presence of this additional fusionbears out the view put forward in a previous paper (Pusey,1938, p. 534) that the 'bony bridge' described by de Villiers(1934) in the adult skull of Ascaphus is but the root of theascending process which is retained through the period of meta-morphosis without the destruction which is the rule in modern-type frogs. It will be interesting to observe what bone ossifiesthis bridge after metamorphosis. Will it prove that Ascaphusadds to its other primitive characters the retention of anindependent epipterygoid bone, or will it be found that theprocess is ossified from the prootic centre? Watson (1940) hasfound a separate epipterygoid bone in M i o b a t r a c h u s, butthis region of the skull of P r o t o b a t r a c h u s , with itsforamina, is unfortunately not well known.

(TO) The L a r v a l Otic P rocess of X e n o p u s .

Before finally leaving the above topic it may be well to makea suggestion about the otic attachments of the larval palato-quadrate of Xenopus laevis to the auditory capsule. Dr. Patter-son (1939) gives excellent reconstructions in her recent paper on<rT1he Head of Xenopus laevis', and her figures show conditionsat least superficially similar to those of A s c a p h u s . That is

172 H. K. PUSBY

to say, that in X e n o p u s there are two fusions of the back ofthe palatoquadrate bar to the auditory capsule, one to theanterior tip of the capsule and one more postero-laterally to itsside wall, the two being parted by a foramen over the thymusgland; see her fig. 14, PI. 12. Her fig. 23, PL 13, makes it clearthat neither fusion is present in the early (10 mm.) larva, but thateach is already well formed in the 28 mm. stage (fig. 27 a to c,PI. 15) and is further consolidated in the 60 mm. stage (fig. 24,PI. 14). She marks the antero-medial of these connexions as thelarval otic process and in this she follows the lead of Kotthaus(1933), Edgeworth (1935), and de Beer (1937). It seems to memore probable, however, that this title should be reserved forthe postero-lateral connexion, for this latter seems to be madewith a more usual part of the capsular wall and bears moretypical relations to the thymus gland, the lateral head vein, thebranches of the IX nerve, and the ramus connecting it to theVII nerve; there is also now the evidence from A s c a p h u s .Unfortunately, the vein which I have used as the critical testfor the otic process in A s c a p h u s and R a n a is apparentlyabsent in X e n o p u s , judging from Miss Patterson's carefuldrawings of selected sections (figs. 6 to 17, Pis. 11 and 12). Theadductor jaw muscles, which this vein drains in other frogs, arepoorly developed and have very anteriorly placed origins whichfail to reach back to anywhere near the capsule in this genus(cf. her figs. 8 and 9). I raised this point in a letter to MissPatterson and, in reply, she was kind enough to make and sendto me a wax-plate reconstruction of the auditory region of her60 mm. stage, for which I would again thank her here. She still,however, maintained the view published in her paper; yet astudy of her model leaves me convinced that it is none the lessthe postero-lateral connexion which is the larval otic processin X e n o p u s .

I would also like to correct a statement which Miss Pattersonmakes on p. 179. What she there claims to be the 'quadrato-ethmoidal cartilage' is surely a chondrified trabecular-quadrateligament and not a chondrified quadrato-ethmoidal ligament(see particularly her figs. 6, 23, 24, and 27 a). A comparison ofher fig. 27 c with my fig 19, PI. 13, taken in conjunction with


remarks made about this ligament in other primitive frogs, onp. 112 of this, paper, suggests, as at least a possibility, that thetentacular cartilage of Xenopus is the homologue of thelateral cartilage wing (clsl) of the supra-rostral system ofAscaphus .

(n) The Eoof of t h e ' Q u a d r a t e T u n n e l ' .

Before leaving the palatoquadrate I would draw attentionto the exact similarity of anatomical relations between thecartilage forming the roof of the quadrate tunnel (at) inAscaphus (see p. 121) and the small, sickle-shaped cartilagein, say, E a n a or Discoglossus , which is the first rudi-mentary trace of the annulus tympanicus. It is worth remember-ing that Gaupp (1893 and 1906) describes the formation, inE a n a , of a mesenchyme cloud from the base of the muscularprocess at its front edge, in early stages. This cloud remainsthroughout the larval life surrounding a branch of the carotidartery and touching the quadrate above and below it. Itbecomes denser as metamorphosis approaches and the annularcartilage condenses only from its outer part, laterally to theartery. If, however, the whole cloud were to chondrify, it wouldcomplete a tunnel as in Ascaphus . Now Ascaphus hasno middle ear apparatus nor tympanic ring in the adult (deVilliers, 1934), a condition which must certainly be one ofsecondary loss, at least as far as the columella system is con-cerned, so that it would be unwise to draw any far-reachingconclusions from this otherwise primitive frog as to the ancestralorigin of the annulus, but the condition may perhaps be taken asfurther evidence to support Gaupp's view that the annulus,which is peculiar to the Anura, is a separated part of thequadrate bar.

(o) The S u p r a - r o s t r a l Sys tem.Without further embryological evidence from earlier stages

little can be said of the homologies of the parts of this system tothe structures in other Vertebrates than the Anura. I havesought to compare the structures with parts of the nasalapparatus of Urodeles and other types, but have failed. There

174 H. K. PUSEY

need be no surprise that the system is divided into one medianand two lateral parts, for this condition is foreshadowed by theevidence collected together from other frogs by van Seters(1922) and shown in his fig. 10. I have found that in Dis-cog lossus the system is also deeply cleft into three sectionswhich are, however, fused to one another (Text-fig. 7 E). Itappears, therefore, that the supra-rostral system of the ancestorwas divided up into three pieces (or if the medial piece is reallyof dual origin, possibly into four pieces). It is rather (1) theextreme posterior position of the whole system and (2) the rigidattachment of its median piece to the trabeculae, which are ofinterest in A s c a p h u s . The posterior position is to be ex-pected to some extent from what has been said on pp. 146 to 148,but in A s c a p h u s this posterior position seems almost to beexaggerated in the way expected if the criticism given on p. 148et seq. were valid. That is to say, that the whole apparatus seemsto be further back than it would be if it were just a slightlymodified arrangement of the tips of the trabeculae. Further,the fusion of the medial piece to the under sides of the trabeculaeseems also to be very much exaggerated and to be carried alsosomewhat far back. These backward displacements are probablysecondary characters related to the peculiar behaviour of thetadpole. They are part of a general tendency, discernible alsoin other structures, to supply the mouth and pharynx cavitywith the maximum skeletal support on all sides and particularlyunderneath, so that the sucker has a rigid roof which can belifted up as a wThole and can withstand the atmospheric pressure.If fig. 15, PI. 11 is superimposed on fig. 21, PI. 14, so that thejaws in the two coincide, it will be seen that the mouth andpharynx cavities are largely ringed round with cartilage strutswhich leave few membraneous spaces between them. (1) Theposterior jaw cartilages are broad plates; (2) the space betweenthe anterior jaw cartilages and the ceratohyals is supported bythe unique posterior spurs of the jaw cartilages; (3) the cerato-hyals are broad plates and the medial part of the hyobranchialapparatus is also much consolidated; and (4) the gill bars,particularly the anterior three pairs, are also widened antero-posteriorly. Thus the peculiar box-like structure of the supra-


rostral system, in its medial part, supplies the anterior sectionof this general basket work, which as a whole is presumably asecondary character related to the sucker mechanism.

8. THE CHARACTERS WHICH ASCAPHUS SHARES WITH THEURODELES.

(1) Short trabecular horns.(2) The nasal sacs lie laterally to the olfactory foramina and

not largely in front of them.(3) The anatomy of the auditory capsule at its anterior end,

e.g. a c o m p l e t e b a s i t r a b e c u l a r p rocess acts as theeffective anterior floor and is intimately fused to the capsularwalls; the true capsular floor is present only as membrane; ajoint anterior acustic foramen and exit for the facial nervewhich appears to run through the capsular cavity to separateforamina for its hyomandibular and palatine branches.

(4) A prefacial commissure is present.(5) Absence of a taenia tecti medialis and a t. t. transversalis

in both larva and adult.(6) Parachordal cartilage underlies the notochord (as well as

overlying it in the usual anuran manner, so that the notoehordis encased on all sides in A s c a p h u s ) .

(7) The relatively posterior position of the palatoquadrateand its parts, and of many other splanchnic structures (seep. 146 et seq.).

(8) The pterygoid process is unobscured below in the body ofthe quadrate.

(9) The posterior position and the extent of the pterygoidbone-rudiment (= ligamentum quadrato-ethmoidale).

(10) The high-level attachment of the ascending process of thequadrate to the pila antotica and to the orbital cartilage and theretention of its root throughout adult life.

(11) The articulation between the quadrate basal process(in Ascaphus , only its posterior part) and a complete basi-trabecular process of the cranial floor.

(12) The absence of a well developed muscular process fromthe otic process of the quadrate.

(13) The posterior position of the whole hyobranchial

176 H. K. PUSEY

apparatus and the more antero-posterior, rather than medio-lateral, orientation of the gill bars.

(14) The absence of a processus branchialis between the basesof the Ilnd and Illrd gill bars, and the freedom of the IVtharch from the hypobranchial plate.

(15) The presence of forked ' urobranchial' prongs on theunder side of the basibranchial copula (copula II).

(16) The mandibular muscles are not greatly subdivided;thus there are five pairs only ( A s c a p h u s ) instead of ten pairs(modern-type frogs); there are five pairs in Urodeles.

(17) The intermandibularis muscle partly underlies the inter-hyoideus muscle.

(18) Levator mandibulae anterior (pterygoid) fibres, thoughfused withL. m. posterior profundus fibres, arise high up on the or-bital cartilage in front of the quadrate ascending process and theexit of nerves V 2 and 3 (and not from the front face of the auditorycapsule below the ascending process as in modern-type frogs).

(19) The hyoid muscles are not greatly subdivided; thus thereare five pairs in A s c a p h u s instead of seven pairs as in themodern-type frogs; the Urodeles have from four to six pairs.

(20) Of the five pairs of hyoid muscles mentioned above, onepair—the branchio-hyoideus externus—is shared with theUrodeles, but is absent from all modern-type frogs except theDiscoglossidae.

(21) The posterior origins of the Levatores arcuum bran-chialium muscles on the posterior wall of the auditory capsule(I and II, and III in part) and on the sheaths of the neckmuscles (II and III in part, and IV).

(22) Subarcualis rectus IV passes from arch IV to arch I(but not to a branchial process as in other frogs).

(23) A s c a p h u s is more primitive than the Urodeles inpossessing Subarcuales obliqui muscles arising on arches '? Vand IV, and is like them in having similar muscles arising onarches III and II. Modern-type frogs have only one of thesemuscles arising from arch II (transversus ventralis II, of Edge-worth). A s c a p h u s is like the Urodeles in that these musclesare inserted on the forked urobranchial prongs of the basi-branchial copula.


(24) The persistence of a Transversus ventralis IV muscle inold larvae of A s c a p h u s . This muscle develops, but is soonlost, in the ontogeny of modern-type frogs.

(25) See p. 145 for a ? abnormality of the geniohyoideus musclein this specimen of A s c a p h u s .

(26) The attachment of the Eectus cervicis muscle to theurobranchial prongs of the basibranchial copula.

9. THE CHARACTERS WHICH ASCAPHUS SHARES WITH

D I S C O G L O S S U S .

(1) The supra-rostral system is cleft into three parts, butthese are not separate from one another in D i s c o g 1 o s s-u s asthey are in A s c a p h u s . .

(2) The attachment of the trabecular-quadrate ligament tothe lateral wing of the supra-rostral and not to the tip of thetrabecular horn.

(3) Certain auditory capsular arrangements, e.g. the commonentrance to the anterior acustic foramen and the facial nervetunnel; the facial nerve is free in the apparent capsular cavityin young larvae of D i s c o g l o s s u s ; separate foramina for VIIpalatine and hyomandibular.

(4) A prefacial commissure is present.(5) Absence of a taenia tecti medialis in the cartilage skull roof.(6) The high-level attachment of the quadrate ascending

process to the pila antotica and to the orbital cartilage.(7) The close association of the anterior tip of the auditory

capsule with the posterior border of the ascending process andthe body of the quadrate as a ball and socket joint; comparethe actual fusion in A s c a p h u s .

(8) The presence of a small 'posterior basal process' on thequadrate (see Text-fig. 6, p. 155).

(9) The absence of a processus branchialis joining the Ilndand Illrd gill bars. There is, however, an incipient process pro-jecting forwards from the Illrd bar in Di scog lossus .

(10) A dorso-ventrally deep, but laterally narrow urobranchialkeel on the basibranchial copula of D i seog lossus , not,however forked at its posterior end as in A s c a p h u s .

(11) Presence of a branchio-hyoideus externus muscle.

178 H. K. PUSEY

(12) Absence of a hyoangularis muscle.(13) Posterior origins of the suspensorio-angularis and suspen-

sorio-hyoideus muscles.(14) The quadrato-angularis muscle arises medially to the

ceratohyal-quadrate articulation, not laterally to it.(15) The Subarcualis IV muscle passes from arch IV to arch I

and not to a processus branchialis.(16) The simple form of the Subarcualis I muscle.

10. SUMMARY.

This paper gives the first account of the larval cranial anatomyof either of the genera of Liopelmid frogs.

A single, partly grown larva o fAscaphus t r u e i , Stejneger,has been studied in transverse sections and in two-dimensionalreconstructions. Its chondrocranium, jaws, gill arches, and headmuscles are described and figured. Comparisons are madethroughout with similar structures of Urodeles and certain otherfrogs, particularly D i scog lo s sus p i c t u s and K a n a t e m -p o r a r i a . A summary of the characters which A s c a p h u sshares with the Urodeles is given on pp. 175-7 and with Dis-cog lossus on pp. 177-8. The reader is referred to these listsas an important part of this summary.

Noble (1931, &c.) considers A s c a p h u s (with Liopelma) tobe one of the two most primitive living frogs. The findings ofthis paper are in full agreement with this view. Thus larvalA s c a p h u s is shown to be a persistently primitive 'link-animal' whose cranial structures, in almost every case, differfrom those of other frogs—often radically—and throw muchlight on the evolution of the modern-type frog tadpole from theunknown (larval) ancestor.

A s c a p h u s is shown to have more characters in commonwith the Urodeles than any other frog larva yet described. Mostof these are probably a simple retention of an ancestral Amphi-bian plan which led on to the frogs and Urodeles (contrastthe writings of Holmgren and Save-Soderbergh). Others seemto link these two orders even more closely together. Suchare: (1) The presence of 'urobranchial' prongs on the basi-branchial copula and the attachment to them of Subarcuales


obliqui and Eecti cervicis muscles; (2) the presence of a pairof Branchio-hyoideus extemus muscles and other similaritiesof the musculature. The relationship of Ascaphus to theGymnophiona is far less marked.

A s c a p h u s , however, has remained more primitive than thepresent-day Urodeles by retaining: (1) a ? Vth gill bar, with itsSubarcualis rectus and S. obliquus muscles, and (2) four pairsof S. obliqui muscles instead of two. In these points it is, infact, the most primitive living tetrapod.

A s c a p h u s is, however, somewhat specialized in relation to asucker mechanism and to a peculiar method of larval progres-sion which it employs. These have led to an exaggerated auto-styly of the palatoquadrate which has developed an additionalfusion to the anterior tip of the auditory capsule; to a rigidfusion of the central part of the supra-rostral system to the skull;to the general heavy build of the head cartilages and to the greatsize of several of the mandibular and hyoid muscles; to a generalconsolidation and widening of parts of the hyobranchial ap-paratus ; to a widening of the posterior jaw cartilages and thedevelopment from them of unique posterior spurs. A pre-oralmouth cavity and a long 'posterior narial tube' to the innernostril are also parts of this specialization.

Among the frogs A s c a p h u s is shown to be most nearlyrelated to the Discoglossidae, which appear to have been derivedfrom an ancestor with many Ascaphus-like characters. This isin further agreement with Noble's classification and is particu-larly true of Discog lossus p i c t u s .

A s c a p h u s is unique among frogs in the posterior positionof its splanchnic head structures; see the list on p. 147.

A forecast is made of the probable evolution of the modern-type tadpole's jaw system from that of the unknown ancestorand this is diagrammatically summed up in Text-fig. 7, pp. 158-9.

Evidence is collected to show that the ' anterior basal process'(= commissura quadrato-cranialis anterior) is not an ethmoidalstructure by origin, as has been held up to now, and conse-quently Save-Soderbergh's use of it to explain an ethmoidalstructure in a Stegocephalian Amphibian is criticized.

An account of the muscles has been given in summary form

180 H. K. PUSEY

on pp. 126 to 146 and cannot be further condensed here. But itmay be noted that Edgeworth's theories (1935) of the primitivemuscular content of a single branchial segment break downwhen applied to A s c a p h u s. It is now probable that a singlesegment could simultaneously contain a Subarcualis rectus, aS. obliquus, and a Transversus ventralis muscle. Further, Edge-worth's term 'Transversus ventralis I I ' must be changed to' S. obliquus II' in the frogs and his ' S. rectus IV' must probablybe changed to 'S. recti IV, III, and I I ' in the Urodeles.

11. LITERATURE CITED.

Archey, G., 1922.—"Habitat and Life History of Liopelma hochstetteri",'Reo. Canterbury (N.Z.) Mus.', 2.

Beer, G. R. de, 1937.—'Development of the Vertebrate Skull.' Oxford.Denburgh, van, 1912.—'Proo. Calif. Acad. Soi.', 3.Eaton, T. H., 1936.—"Myology of Salamanders; I, Muscles of the Head",

' Journ. Morph.', 60.Edgeworth; F. H., 1925.—"Autostylismof Dipnoi and Amphibia", 'Journ.

Anat., Lond.', 59.• 1935.—'Cranial Muscles of Vertebrates.' Cambridge.Frazier, M., 1924.—"Contribution to the Anatomy of Amphibian Larynx",

'Journ. Morph. Physiol.', 39.GSaige, H. T., 1920.—"Observations upon Habits of Ascaphus truei", 'Occ.

Papers Mus. Zool. Univ. Mich.', 84.Gaupp, E., 1893.-—"Primordialcranium u. Kieferbogen v. Rana", 'Morph.

Arb.', 2.1904.—"Das Hyobranchialskelet", 'Anat. Hefte, Ergeb.', 14.1906.—"Entwicklung des Kopfskelettes", in Hertwig's 'Handbuch

der vergl. u. exper. Entwick. der Wirb.', Jena.Goodrich, E. S., 1930.—'Studies on the Structure and Development of

Vertebrates', London.Holmgren, N., 1934.—" Origin of the Tetrapod Limb " , ' Acta Zool.', 14.Kotthaus, A., 1933.—"Entwicklung des Primordialcraniums von Xenopus

laevis", 'Z. wiss. Zool.', 144.Litzelmann, E., 1923.—"Visceralapparat der Amphibien", 'Z. Anat.

EntwGesch.', 67.Luther, A., 1914.—"Vl. d. vom N. trigeminus versorgte Musculatur der

Amphibien", 'Acta Soc. Sci. fenn.', 44.Noble, G. K., 1924.—"New Spadefoot Toad from Oligocene of Mongolia",

'Amer. Mus. Novit.', 132..—— 1927.—-"Value of Life History Data in study of Evolution of Amphi-

bia", 'Ann. N.Y. Acad. Sci.', 30.—-— 1931.—•' Biology of the Amphibia.' New York.


Parker, W. K., 1876.—"Structure and Development of Skull in Batrachia;Part I I " , 'Philos. Trans.', 166.

1881.—"Structure and Development of Skull in Batrachia; PartI I I " , ibid., 172.

Patterson, N. F., 1939.—"Head of Xenopus laevis", 'Quart. Journ. MicrSci.', 81.

Piveteau, J., 1937.—"Un Amphibien du Trias inferieur", 'Ann. Paleontol.'26.

Pusey, H. K., 1938.—"Structural Changes in the Anuran Mandibular Archduring Metamorphosis, Rana temporaria " , ' Quart. Journ. Micr. Sci.', 80.

1939.—"Methods of Reconstruction from Microscopic Sections",'Journ. Roy. Micr. Soc.', 59.

Save-Soderbergh, G., 1934.—"Evolution of Vertebrates, and Classifica-tion", 'Ark. Zool.', 26, A, no. 17.

1935.— "Dermal Bones of Head of Labyrinthodont Stegocephaliansand Primitive Reptilia", 'Medd. Greenland', 98, no. 3.

1936.—"Morphology of Triassic Stegocephalians from Spitsbergen",'Svenska Akad. Handl.', 16.

Schiilze, F. E., 1892.—"U. d. inneren Kiemen der Batrachierlarven",' Abh. Konigl. Preuss. Akad. Wiss., Berlin.'

Seters, W. H. van, 1922.—"Developt. du Chondrocrane d'Alytes obstetri-cans avant la Metamorphose", 'Arch. Biol.', 32.

Trewavas, E., 1933.—"Hyoid and Larynx of Anura", 'Philos. Trans.',222, B.

Villiers, C. G. S. de, 1934.—"Cranial Anatomy of Ascaphus truei", 'Bull.Mus. Comp. Zool., Harv.', 77.

Wagner, D. S., 1934i—1. "Structure of Inner Ear in Liopelmidae", 2."Cranial Characters of Liopelma hochstetteri",' Anat. Anz.', 79.

Watson, D. M. S., 1940.—" Origin of Frogs", 'Trans. Roy. Soc. Edin.', 60.

12. A B B R E V I A T I O N S U S E D IN THE F I G U R E S .

a, artery; ac, auditory capsule; ajc, anterior jaw cartilage (= cartilagolabialis inferior, Gaupp); ap, anterior pit in cranial floor; asc, anteriorsemicircular canal; at, arterial tunnel through quadrate; b, brain; ba I-IV,branchial arches I-IV; ? ba V, probable Vth branchial arch ( = spiculumIV); bbc, basibranchial copula; bhc, basihyal copula; bhem, branchio-hyoideus externus muscle; bpr, basal process of quadrate (posterior part);bt, basitrabecular process; bv, blood-vessel; c, carotid artery; cb, cartilagebridge; cbm I to III, 1st to I l lrd constrictores branchiales muscles; cbr,branch of carotid artery; cf, cranial floor; ell, ceratohyal; el, cartilage ledgebelow nasal sac; clsl, lateral wing of supra-rostral cartilage (part of cartilagolabialis superior, Gaupp); clsm, median part of supra-rostral cartilage (part ofcartilago labialis superior, Gaupp); con, conus; cq, cranio-quadrate pas-sage ; cqa, commissura quadrato-cranialis anterior (anterior part of quadrate

182 H. K. PUSBY

basal process); ct, trabecular horn; dbm, diaphragmato-branchialis muscle;de, deep epidermis; dm, dense mesenchyme; dpba I, dorsal process of 1stbranchial arch; eft, epithelial band; en, external nostril; erm, externalrectus muscle; / , region of fusion of quadrate with auditory capsule; faa,foramen acusticum anterius; fam, median acustic foramina; fap, foramenacusticum posterius; fc, foramen cranio-palatinum; fcp, foramen caroticumprimarium; fen, foramen endolymphaticum; ff, facial foramen (as yetundivided); fhm, foramen for hyomandibular branch of facial nerve; fj,foramen jugulare; fo, fenestra ovalis; focn, foramen oculomotorii; fol,foramen olfactorium; fon, foramen opticum; fpi, foramen perilymphaticuminferius; fps, foramen perilymphaticum superius; ft, trigeminal foramen;ftd, trigeminal foramen (dorsal division for V2 and V3); ftv, trigeminalforamen (ventral division for V J ; fu, fusion of processus posterior hyaliswith 1st branchial arch; fvcl, foramen for branch of vena capitis lateralis;g, glottis; glim, geniohyoideus muscle; hbp, hypobranchial plate; htp, hornytooth-plate of upper ' l ip ' ; ifrm, inferior rectus muscle; ihm, interhyoideusmuscle; in, internal nostril; iom, inferior oblique muscle; ipm, inter-mandibularis posterior muscle; irm, internal rectus muscle; I, ligament;la, left auricle; labm I and II, 1st and Ilnd levatores arcuum branchialiummuscles; labm IV, levator ( +constrictor) IV muscle; lac, 'cartilage ledgeof auditory capsule', de Villiers ( = outer edge of basitrabecular process andpost-palatine commissure); Ic, lateral semicircular canal; Us, lower(posterior) ' lip', forming sucker; Imam, levator mandibulae anterior articu-laris muscle; Impm, levator mandibulae posterior profundus muscle,? + levator mandibulae anterior muscle; Imsm, levator mandibulaeposterior superficialis muscle; Iq, quadrato-ethmoidal ligament ( = rudi-ment of pterygoid bone); Is, lymph space; 1st, tendon of levator mandi-bulae posterior profundus muscle; It, tendon of levator mandibulaeposterior superficialis muscle; mfc, membraneous true floor of auditorycapsule; mo, mouth opening; msr, membraneous skull roof; mtr, membrane-ous temporal roof; n, notochord; nc, notch in quadrate for articulation ofceratohyal; nm, musculature of neck; runt, notch for 'posterior narial tube';ns, nasal sac; nsr, V-shaped notcli in supra-rostral cartilage; oa, occipitalarch; oac, outline of auditory capsule; oc, orbital cartilage; occ, occipitalcondyle; oec, open end of cranium; ohm, orbito-hyoideus muscle; op.operculum; pa, ascending process of quadrate; paq, articular region oquadrate; pbc, palatine branch of carotid artery; pcqa, posterior border oicommissura quadrato-cranialis anterior; pf, foramen for palatine branchof facial nerve; pfc, prefacial commissure; php, pouch of pharynx; pia, pilaantotica; pit, pituitary gland; pitf, pituitary fossa; pjc, posterior jawcartilage (cartilago Meckelii, Gaupp); pm, preoral buccal cavity; pmq,muscular process of quadrate; pnt, 'posterior narial tube'; pot, processusoticus quadrati (larval); ppc, post-palatine commissure; pqe, processusquadrato-ethmoidalis; pre, pars reuniens; psp, pseudo-basal process(== detached outer end of basitrabecular process); psq, posterior spur oiquadrate; pt, profundus tunnel; ptc, pterygoid process of quadrate; q,


quadrate; qdm, quadrato-angularis muscle; ra, right auricle; ram, rectusabdominis muscle; rent, rectus cervicis muscle; rcsm, rectus cervicis super-ficialis muscle; s, postero-ventral spur of posterior jaw cartilage; sam,suspensorio-angularis muscle; saq, surface of ceratohyal articulating withquadrate; saom II to V, Ilnd to Vth subarcuales obliqui muscles; sarm I,IV, and V, 1st, IVth, and Vth subarcuales recti muscles; shm, suspen-sorio-hyoideus muscle fibres; soc, subopercular cavity; som, superioroblique muscle; sov, subocular vacuity; sp II and III, spicula over Ilndand Illrd branchial arches; srm, superior rectus muscle; t, trabeculacranii; toe, tip of auditory capsule; tc I-II, and III-IV, terminal com-missures between branchial arches I and II and III and IV; tg, thymusgland; tJig, thyroid gland; til, horny teeth of lower ' lip'; tm, taenia tectimarginalis; tql, trabecular-quadrate ligament; ts, tectum synoticum (in-complete); tul, horny teeth of upper ' lip'; tvm IV, transversus ventralisIV muscle; ul, upper (anterior) 'lip'; uls, upper anterior 'lip', formingsucker; upbc, urobranchial prong of basibranchial copula; v, ventricle;va, ventral aorta; vcl, vena capitis lateralis; vclb, branch of vena capitislateralis; vn, valve of external nostril.

I, olfactory nerve; / b, branch of olfactory nerve; / / , optic nerve; / / / ,oculomotor nerve; IV, abducens nerve; Fx, profundus branch of trigeminalnerve; Vg, trigeminal ganglion; Vr, root of trigeminal nerve; Vx b, branchof profundus nerve; V2 b, branch of maxillary branch of trigeminal nerve;Vs b, branch of mandibular branch of trigeminal nerve; VII b, branch offacial nerve; VII g, facial ganglion; VII hm, hyomandibular branch offacial nerve; VII p, palatine branch of facial nerve; VII r, root of facialnerve; VII—VIIIr, joint root of facial and auditory nerves; VIII b,branch of auditory nerve; VIII g, auditory ganglion; IX b, branch ofglossopharyngeal nerve; IX re, ramus communicans from glossopharyn-geal to facial nerve.

13. DESCRIPTION OF PLATES 6-14.

All figures are drawn or reconstructed from a series of trans-verse sections of a single larval specimen of A s c a p h u s t r u e i :overall length 28 mm., tail length c. 18 mm., no hind legs present,having completed about one half to two-thirds of its larval life.

(The sections were mounted in rows of nine, with four rowsto each slide.)

Eelative to the sections shown in figs. 1, 2, 3, PL 6; figs. 4, 5,PI. 7; figs. 6, 7, PL 8; figs. 8, 9, 10, PL 9; figs. 11, 12, 13, 14,PL 10, and in Text-figs. 1 to 5, all reconstructions are reversed,so that the animal's right side has become the apparent left.

184 H. K. PUSEY

PLATE 6.

Fig. 1.—Slide 4, row 1, section 2; through line AA of the reconstructions:fig. 15, PI. 11, fig. 17, PI. 12, and fig. 19, PL 13.

Fig. 2.—Section 4-3-4; through line BB of the reconstructions.Fig. 3.—Section 5-3-2; through line CC of the reconstructions.

PLATE 7.

Fig. 4.—Section 6-1-7; through line DD of the reconstructions, in-cluding fig. 21, PL 14.

Fig. 5.—Section 6-4-7; through line EE of the reconstructions.

PLATE 8.

Fig. 6.—Section 7-3-6; through line FF of the reconstructions.Fig. 7.—Section 8-3-1 (in part); through line GG of the reconstructions.

PLATE 9.

Fig. 8/—Section 9-1-1; through line HH of the reconstructions.Fig. 9.—Section 9-3-6 (in part); through line II of the reconstructions.Fig. 10.—Section 10-1-7 (in part); through line J J of the reconstructions.

PLATE 10.

Fig. 11.—Section 10-2-2 (in part); through line KK of the reconstruc-tions.

Fig. 12.—Section 10-3-1 (in part); through line LL of the reconstructions.Fig. 13.'—Section 11-1-5 (auditory region); through line MM of the

reconstructions.Fig. 14.'—Section 11-2-4 (auditory region); through line NN of the re-

constructions.PLATE 11.

Fig. 15.—Reconstruction of the chondrocranium and jaws: seen dorsallyand 23° from the left side. (Fig. 15, PL 11 may be fitted to fig. 21, PL 14.)

Fig. 16.—Reconstruction of the supra-rostral system and anterior endof the cranium: seen from in front. Fig. 16 may be fitted to Fig. 20,PL 13.

PLATE 12.

Fig. 17.—Reconstruction of the chondrocranium and upper jaw systems;the lower jaw has been removed: seen ventrally and 10° from the animal'sloft side.

Fig. 18.—As Fig. 16, but seen from behind.

PLATE 13.

Fig. 19.'—Reconstruction of the chondrocranium, jaws, and ceratohyal;lateral view.

Fig. 2O.<—Reconstruction of the chondrocranium and lower jaw: seenfrom in front. The anterior end of the cranium and the supra-rostral system

HEAD OP ASCAPHUS 1 8 5

PLATE 14.(see figs. 16 and 18) have been cut away. Fig. 16, PI. 11 may be fitted toFig. 20.

Fig. 21.—Reconstruction of the lower jaw and hyobranchial cartilages:seen dorsally and 23° from the left side. (Fig. 21 may be fitted to fig. 15,PI. 11.)

Fig. 22.—Reconstruction of the auditory capsule: seen from in front.The front of the cranium and the palatoquadrate have been cut away.

Fig. 23.—As fig. 22, PI. 14, but with parts of the nerves V, VII, VIII,and IX added.

NOS. 334-5

Quart. Journ, Micr. Sci. Vol. 84, N.8., 1 L 6"

v. -..

//. K. Pusey del.

Quart. Journ. Micr. Sci. Vol. 84, N.S., PI. 7

vcLb

FIG. 4

erm pbc! 11/ l.

mtr,VI I>Lmsm '•-. '

r,'"lv r " msr V l l p y i b ycLb,r . • ' * ! I '• .'

FIG. 5

//. K. Puscy del.

Quart. Journ. Micr. Sbi. Vol. 84, N.8., PI 8

vcLb

LmsmLmpm

mtr.

Vllp—

Lmem

Lmsm

V,b

FIG. 6

oc V2-3

IV- .vcLb.Lmsm

.Lmpm

,mtr

sam

' S a m

Fio. 7

/C. Pusey del.


VII bLmsm

vcLb

FIG. 8

(mm ,tm

msr. ?

thrrr':

Lpm bbc ghm

I mmnc I s ghm ch |Xb cbr '. • ohm

11. K. Pvsey del.

vclb .^ cm a i ,

• V Lmpm I!' Lmsm

<cLb

Quart. Journ. Micr. Sci. Vol. 84, N.8., PI. JO

ts>. vclb tm

vf- V l l l g - ^Lmsm J ^-5f7

v -vii g

msr

Vr

/ a VII b

g'hm Lhm

•/"; ••cbr upbc-''

-.' • : v 'vcLb i,hm'

'ohmghm ba sarml

msr Fo a Lmpm

c l ) ( rc. ' I A 0 • ^cLb \cbr sam Ls

Lmpm

/ c IXbrrvtLff'^"-"phpI X r c dpbal ' U b m l

.'/. /v. Piisey del.


AI mm

cLsm

cLsL

pbc -.., ^

.Li.-T:'~'"- / .-.it''' .-

16

cLet- nsr

II. K. Fusty del.

Quart. Journ. Micr. Sci. Vol. 84, N.8., FL 12

A

oacFo

occ1mm Vj 17

.-- oc

;J. K Pusey del.

Quart. Journ. Micr. Sci. ' Vol. 84, N.S., PL 13

/A /B /C /D /E /F /S /H /I J//K/L /M/N

pa Pbd b s

P / oac

cb (> tinb. • \

FoLocc

18

/A /B /C /D /E /F /G /H /I

Fcp pLbF t s

3C

pob

pacj

i l . 1C. Pitsey del.

Quart. Journ. Micr. Sci. Vol. 84, N.8., PI. 1,

btoc

FtdV2-3

V I I 3 2 3

H. K. Pusey del.

On the Head of the Liopelmid Prog, Ascaphus truei. I. The ......On the Head of the Liopelmid Prog,...

Documents

Transcript of On the Head of the Liopelmid Prog, Ascaphus truei. I. The ......On the Head of the Liopelmid Prog,...