Contributions to the Embryology of theMarsupialia.

By

•las. F. Hill, B.Sc.(Edin.), F.L.S.,Demonstrator of Biology in the University of Sydney, N.S.W.

With Plates 1 and 2.

II. On a Further Stage in the Flacentation of Ferameles.

IN the present paper I deal with a stage in the placenta-tion of Perameles somewhat later than my previous Stage D(1). Beyond the demonstration of the persistence of the entireyolk-sac wall, no strikingly new facts have been brought tolight in the investigation, but I have been enabled to essen-tially confirm my previous account, and also to add somefuller information on certain points of detail, especially asregards the yolk-sac vessels. In addition, I give an account,in the sequel, of the foetal membranes of a species ofM&cropus, for comparison with those of Perameles, and Ialso take this opportunity of offering some further remarkson the phenomena of parturition, and of correcting a mis-interpretation contained in my previous account of thatprocess, since an extended investigation1 into the morpho-logy of the female urogenital organs of Perameles, withspecial reference to the phenomena of parturition, has con-

1 Communicated to the Linnean Society of New South Wales, March,1899.

VOL. 4 3 , PAET 1. NEW SERIES. A

2 JAS. P. HILL.

vinced me that " the common uterine canal" of my formerpaper is in reality formed by the union posteriorly of twomedian vaginal cul-de-sacs, which constitute a medianvaginal apparatus homologous to that existing in otherMarsupials. If for the above misleading expression the readerwill substitute common median vagina, the main facts in myprevious short account of the process of parturition remainsubstantially correct.

It is with much satisfaction, in view of my interpretationof the allantoic placenta of Perameles as a primitive featureof its organisation, that I am thus enabled to bring its appa-rently peculiar parturition phenomena into line with thoseoccurring in certain other Marsupials, which, like Perameles,give birth to the young through a direct median passage.Further, the investigation above referred to has led me tothe conclusion that the urogenital organs of Perameles are ina condition which can only be described as persistentlyembryonic, and which is without doubt much more primitivethan that existing in any other Marsupial hitherto described,a conclusion which I believe serves materially to strengthenthe position I have taken up as regards the significance ofthe occurrence of an allantoic placenta in the genus.

My work for the past year has been carried out with theaid of a grant from the Royal Society, to whose Committee Idesire to express my grateful thanks.

STAGE D.—P. OBESULA.

The female, a specimen of the short-nosed Bandicoot, P.obesula, upon the examination of whose genital organs theaccount of the present stage is based, came into my hands afew hours after death. Post-mortem change, however, wasnot at all apparent, and the genital organs were at once cutout and preserved in picro-nitric acid.

Both uteri were much enlarged; the right measured 22 mm.in length by 11 mm. in breadth, the left 21 mm. by 15 mm.Each contained a single embryo (PI. 1, fig. 3), measuring in

EMBRYOLOGY OF THE MARSUPIALIA. 3

greatest length 12-5 mm. and in head length 6"5 mm. Itsstructural characteristics are given in the Appendix.

Both uteri were submitted to macroscopical examination,while the right uterus and its embryo were alone examinedmicroscopically.

FCETAI MEMBRANES.—In fig. 1, PL 1, the embryo is

shown in situ in the right uterus, the ventral wall of whichhas been removed. It was noticed at once that the yolk-sacWall was not even slightly adherent to the surface of thetnucosa. Whether this non-adherence is due to post-mortemchange or to natural causes I am unable to decide. It isquite possibly due to the latter, since the yolk-sac placenta isnow in a largely degenerated condition.

In the figure, the vasculur omphalopleure and the yolk-sacsplanchnopleure, which were cut through along their line ofattachment to the margin of the allantoic placental area, havebeen reflected outwards. The embryo is seen lying on itsright side, with its head end next the anterior end of theuterus, and its long axis in the antero-posterior axis of thesame. In the left uterus the head of the embryo lay nextthe posterior end of the uterus, so that its position is aninconstant one. The embryo is situated, of course, in theextra embryonic splanchnocoele, which has been opened intoby the reflection of the yolk-sac splanchnopleure, and isinvested by the amnion. The latter is reflected from thedistal end of the thick stalk, which arises from the mid-regionof the ventral abdominal wall. This stalk—the umbilicalstalk (fig. 1, u. s.)—is a tubular prolongation of the abdo-minal wall, which encloses a portion of the coelom, and theupper ends of the allantoic and yolk-stalks (fig. 4, PI. 1).Prom the yolk-stalk the yolk-sac splanchnopleure is reflectedclosely round the embryo. It becomes continuous with thevascular omphalopleure, along the line of union of the latterwith the marginal zone of chorion which is present round theallantoic placental area.

The allantoic stalk (fig. 1, all. s.), on leaving the umbilicalstalk, passes outwards and forwards, and then, curving

4 JAS. P. HILL.

mesially, forms a single loop upon itself, hence it passes overthe right side of the embryo to reach the centre of the allan-toic placental area (fig. 1, pi. a.), a portion of which is visiblein the figure, just behind the back of the embryo. In fig. 2,PI. 1, the embryo has been removed to expose the placentalarea. In both uteri the placental area is roughly oval inshape, as in Stage D, and situated on the mesial side of theuterus. It is bound by a continuous ridge-like thickening ofthe mucosa, so that the area itself appears depressed belowthe general level of the inner surface of the uterus. Throughthe thin walls of the vesicular portion of the allantois theirregularly ridged surface of the placental syncytium couldbe made out. The inner surface of the mucosa of the rest ofthe uterus also presents a ridged appearance.

The placental area of the right uterus measures approxi-mately 8 mm. in length by 6 mm. in breadth; that of the leftuterus was somewhat larger, measuring 6"5 mm. by 9 mm.The allantoic stalk has now a length of about 16 mm., i. e. justdouble the length of the stalk in the preceding Stage D, andit is now also just about double as thick as in that stage.

As in previous stages, the outer or placental surface of thevesicular part of the allantois is found not to occupy theentire extent of the original chorionic area, but there is leftoutside its margin, i. e. outside the margin of the placentalarea, a narrow annular zone of persistent pure chorion, whichsurrounds the placental area, and which is attached on theone side around the margin of that area, and on the otherpasses into the vascular omphalopleure, along the line ofjunction of the latter with the yolk-sac splanchnopleure(PI. 2, fig. 6,ch.).

The placental area of the present stage as seen in surfaceview is essentially similar to that of Stage D. The allantoicstalk (PI. 1, fig. 2, all. s.) is seen to join the inner surface ofthe flattened vesicular portion of the allantois near its centre.In the stalk run the two allantoic arteries and the single vein.The latter is situated above the two arteries, and is hence notvisible in the figure. At the distal end of the stalk the

EMBRYOLOGY OF THE MAESUP1ALIA. 5

vessels branch out over the inner (ccelomic) wall of thevesicular part of theallantois. The two arteries—one runningforwards to supply the anterior portion of the placental area,the other passing back to similarly supply its posteriorportion—eventually divide, each into two branches. Thesepass outwards, dividing as they go, to the margin of theallantoic vesicle, round which they and their branches passto reach its outer or placental surface, where they break upinto capillaries in the allanto-chorionic mesenchyme.

The allantoic vein is formed by the union of two mainfactors (fig. 2, all. v.) which accompany the correspondingarteries. The main arterial branches are each accompaniedby a corresponding venous trunk, but the smaller arterialbranches may or may not be so accompanied. The vitellineartery (fig. 2, vit. a.), emerging from under the distalmargin of the umbilical stalk, passes in the yolk-sacsplanchnopleure, at first almost directly backwards, thencurves mesially and slightly forwards to reach the margin ofthe chorionic zone, where it passes over into the vascularomphalopleure. There it at once divides into its twobranches, which, diverging in opposite directions, form thesitius terminalis (figs. 1 and 2, $. t.), the two branches beingunited by a delicate anastomosis.

The vascular omphalopleure (figs. 1 and 2, vase, omph.) isreadily distinguishable macroscopically from the bilaminaromphalopleure (figs. 1 and 2, bil. omph.) lying outside thesinus, by its more opaque appearance and yellowish colour.

Owing to the fact that the vitelline artery at once divideson reaching the vascular omphalopleure, the vascular areahas at this point no existence, but as the two branches of theartery pass out, they at the same time diverge away from themargin of the placental area, and the vascular area thusgradually increases in width until a maximum of 5 mm. isreached about opposite the point of bifurcation of thevitelline artery. The vascular area has a maximum diameterof 12*5 mm. It will be shown later that the extent of thevascular area of the Perameles embryo of this stage is

6 JAS. P. HILL.

considerably less than that of a Macropod embryo at asomewhat earlier stage of development. What, however,especially characterises the vascular area of this stage, aswell as that of my preceding Stage D, is its relativelyextremely poorly developed capillary system. Branchesare, indeed, given off from the sinus into the vascular ai'ea,but these are small and by no means numerous, and ap-parently undergo very little secondary branching.

I describe below the foetal membranes of an embryo ofMac r op us parma, and from that account and the accom-panying figure (PI. 2, fig. 9) it will readily be seen thatthe branches given off from the sinus terminalis into thevascular area are not only numerous but richly branched;they form a highly developed capillary system, almost entirelyunrepresented in Perameles, in correspondence with the factthat in Macropus the embryonal nutrition is carried onsolely by means of the yolk-sac vessels of the vascular area,while in the Perameles embryo of this stage the allantoiccirculation is now playing the dominant part in the nutritionof the embryo, only an insignificant proportion of the nourish-ment necessary for the growth of the latter being conveyedthrough the yolk-sac vessels. In my previous paper I putforward the suggestion that probably the entire yolk-sac walldisappeared before the end of intra-uterine life. The presentstage, however, shows that, at all events up to this periodof development, the entire yolk-sac wall remains intact,although greatly reduced in functional importance, as shownby the poor development of its vessels.

The yolk-sac splanchnopleure is supplied with blood bynumerous fine wavy vessels (fig. 1, a. v.), which arise directlyfrom the vitelline artery—especially from its first part—as itcourses over that membrane. In fig. 2 one main branch isseen passing from under the free margin of the umbilicalstalk. I t divides up into numerous small wavy brancheswhich pass up parallel with the vitelline veins. Behind thisnumerous branches arise separately from both sides of theartery (fig. 2, vit. a.). In the terminal part of the course of

EMBRYOLOGY OF THE MABSDPIALIA. 7

the latter fewer and shorter branches are given off. Theblood from these fine vessels apparently passes over into theVascular area at the line of junction of the splanchnopleurewith the vascular omphalopleure.

Semon (2) has also described the occurrence of similar finebranches of the vitelline artery in the two forms examinedby him ( j iEpyprymnus ru fe scens and P h a s c o l a v c t u sc inereus ) , and such also occur in M a c r o p u s parma.

The blood from the vascular area is returned by twovitelline veins (figs. 1 and 2, vit. u.),each of which is formedby the union of two main factors, shortly after they entertjie yolk-sac splanchnopleure. As in Stage D, the twovitelline veins run back in the splanchnopleure over the leftside of the head of the embryo (fig. 1), enclosing betweenthem a narrow triangular area altogether devoid of vessels.The two veins gradually approximate as they pass back andeventually run side by side. In this region the amnion isfound connected with the yolk splanchnopleure by a narrowwedge-shaped band of mesoderm which separates the twoveins (PI. 1, fig. 5, mes.). This narrow connection is allthat is now left of the much larger pro-amniotic remnantseen in Stage D. Not only is the area now less in extent,but the mesoderm has penetrated completely across it. Theectoderm of the amnion over the remnant is thickened (fig.5, ect), the cells being club-shaped in form, with the nucleussituated in the outer projecting part of the cell. The ento-derm of the yolk-sac splanchnopleure over the remnant islikewise thickened, but irregularly (fig. 5, ent.).

As the vitelline veins are traced back towards the umbilicalstalk they gradually become smaller, and at the same timepass over from the yolk splanchnopleure towards the amnion,in the now greatly thinned bridge of mesoderm of the pro-aniniotic remnant, which posteriorly becomes continuous withthe thin, attenuated yolk-stalk carrying the vitelline artery.At the anterior margin of the umbilical stalk the two veinsunite into a single trunk. This runs up in a thin fold ofmesoderm continuous with that of the pro-amniotic remnant,

8 JAS. P. HILL.

and which connects the yolk-stalk with the inner surface ofthe tubular umbilical stalk (fig. 4). In the upper part of theumbilical stalk the vitelline vein becomes considerablyreduced in size, and is a much smaller vessel than theallantoic vein. In this stage the vitelline vein is very re-markably reduced as compared with the same vessel inStages C and D. The vitelline artery in the upper part ofthe umbilical stalk is also a quite insignificant trunk, abouthalf the size of one of the allantoic arteries, although itsdiameter is considerably greater as it courses over the yolk-sac splanchnopleure. These facts are sufficient to show thatat this stage the yolk-sac vessels play a quite insignificantpart in the nutrition of the embryo as compared with theirfunctional importance in Stages 0 and D.

The bilaminar omphalopleure calls for no special mentionhere, beyond the remark that it composes the larger half ofthe entire surface of the globular embryonic formation.

STRUCTURAL DETAILS.

The following account of the structural features of theuterus and foetal membranes is made very brief, in view ofthe close agreement of this stage with my previous Stage D.In writing up the present paper I have had, after an intervalof a year, to carefully re-examine many of my old prepara-tions, and compare them with those of this stage, with theresult that I am able to substantially confirm my previousaccount.

I. UTERUS.—The serosa, muscularis, and corium presentessentially the same appearances as in Stage D. The uterineglands are considerably enlarged, and lined by a low cubicalepithelium; and very numerous blood-vessels occur in themuch attenuated interglandular connective tissue.

(a) A l l an to ic P lacen ta l Syncytium.—This has thesame average thickness as in Stage D, viz. "12 mm., and showsno differences from that stage worthy of remark beyond thefact that its capillary system is somewhat better developed.

KMBRYOLOGY OF THE MAESUPIALIA. 9

the capillaries being on the whole larger. As in that stage,owing to the projection of the superficial capillaries, thesyncytial surface presents an irregularly ridged appearance(PI. 2, fig. 7, syn. c).

(b) Syncyt ium beyond Al lan to ic P l a c e n t a l Area.—The portion in contact with the vascular omphalopleure isalso of the same average thickness as in Stage D, viz. '09 mm.In other respects also the same features are visible. Thesuperficial capillaries are fairly abundant, but on the wholevery much smaller than those of the placental area, and notquite so well developed as in Stage D.

The syncytium in contact with the bilaminar omphalopleureis a thinner and much less regular layer, and less vascular.

II. FCETAL MEMBRANES.— (a) Chorionic Ectoderm.—Over the placental area usually single, much degenerated, anddeeply staining chorionic ectoderm cells are still to be found,but are not numerous. Of the marginal zone of chorionicectoderm, seen persistent in Stage D, only traces now remain,in the form of isolated and often much altered cells. Occa-sionally such cells are found forming the transition to theectoderm of the zone of pure chorion (PI. 2, fig. 6, ch.) whichexists between the outer margin of the allantois and thevascular omphalopleure (PI. 2, fig. 6, ch. ect.). The ectodermcells of this chorionic zone are usually enlarged close to themargin of the placental area, but rapidly diminish in size asthey pass outwards.

(b) Al lan to i s and Al lantoic Placenta.—The allantoicstalk is more or less rounded in section, and where it liesenclosed by the umbilical stalk has a diameter of "5 mm.The allantoic canal (fig. 4, all. cl.) is no longer perfectly con-tinuous, and its entodermal lining is only recognisable inplaces. Curiously enough, the allantoic vessels appear to besmaller than in Stage D, but the difference in size is not sodisproportionate as in the case of the vitelline vessels in thetwo stages. As regards the vesicular portion of the allantoisand the allantoic placenta, I have little of importance to addto' the description already given of the same parts in Stage

10 JAS. P. HILL.

D. It may be mentioned that small numbers of maternalleucocytes occur in the allantoic cavity.

A fairly typical section of the allantoic placental area isshown in PI. 2, fig. 7, from which it will be seen that, exceptperhaps for an increase in size of both foetal and maternalcapillaries, the description already given of the allantoicplacenta in Stage D holds good certainly up to this stage ofgestation, and without doubt also for the whole period ofrntra-uterine life, a conclusion already arrived at from the.conditions presented by the post-partum uterus described inmy previous paper.

(c) Vascular Omphalopleure and Yolk-sac Placenta .—The entodermal cells of the vascular omphalopleure, some ofwhich wei'e found in Stage D to be undergoing enlargement,are now found to have all increased in size, and to form auniformly much thicker layer than in that stage. The cellsare mostly somewhat cubical in form, with rounded projectingends (PI. 2, fig. 8, ent.). Their protoplasm and nuclei staindeeply. Over the sinus terminalis (fig. 8, s. t.) the entodermcells are also enlarged and, as in previous stages, club-shapedin form, with the nucleus situated in the outer projectingpart of the cell.

It seems highly probable that this hypertrophy of theentodermal cells of the vascular omphalopleure, commencingas it does about the time the allantoic placenta begins itsfunctional activity, is to be associated with the functionalretrogression of the yolk-sac placenta, consequent upon theestablishment of the allantoic one. For, whether the ecto-derm of the vascular omphalopleure be adherent to thesurface of the syncytium or not, the condition of the yolk-sacvessels shows us that the yolk-sac placenta, if we can anylonger speak of such, is now functionally the mere shadow ofits former self. The ectoderm of the vascular omphalopleure,though still a comparatively thin layer, is slightly thickerthan in preceding stages, the cell bodies being richer inprotoplasm (PI. 2, fig. 8.).

(d) Bilaminar Omphalopleure.—The ectoderm of the.

EMBRYOLOGY OF THE MAESUPIALIA. 11

bilaminar oinphalopleure (PI. 2, fig. 8, ect.) is just aboutdouble as thick as that of Stage 0. The cells are large andSomewhat irregular in shape, with projecting outer ends.Their protoplasm is considerably vacuolated. In Didelphys,according to Selenka (3, pp. 137 and 138), the ectodermal cellsof both the vascular and bilaminar omphalopleure undergoenlargement during the last two days of intra-uterine life.Selenka regards the enlargement of these cells as associatedWith their functional activity in transmitting the secretion ofthe uterine glands into the yolk-sac, but here again I aminclined to look upon the hypertrophy as a sign of degenera-tion.

The entoderm of the bilaminar omphalopleure is a verymuch thinner layer than the ectoderm, and consists of flat-tened attenuated cells (PI. 2, fig. 8).

III . GENITAL ORGANS.—As I have already given elsewherel

an extended account of the structure of the urogenital organsand of the main phenomena connected with the act of parturi-tion, a brief r e s u m e of some of the results of that portion ofmy investigation will here suffice.

The female of the present stage proves to have been inher first pregnancyj as shown by the condition of the genitalOrgans.

Each uterus passes back and opens into a short blindlyending canal, the median vaginal cul-de-sac, which isexternally not marked off from the uterus, while internallythe os of the latter is exceedingly ill-defined, a conditionwhich led me in my previous paper to misinterpret theVaginal cul-de-sacs as posterior prolongations of the. uteri.The two vaginal cul-de-sacs, separated by a common partitionwall, lie embedded in the connective tissue of the anterior endOf the urogenital strand between the lateral vaginal canals,here bent slightly outwards. From the anterior ventral endOf each cul-de-sac, just below the opening of the uterus, therepasses forwards in the connective tissue underlying theposterior portions of the uteri (uterine necks) the anterior

1 'Proc. Linn. Soc. N.S.W.,' March, 1899.

12 JAS. P. HILT,.

forwardly directed portion of the lateral vaginal canal, toopen into the base of the corresponding vaginal caecum.The latter, functioning as a receptaculum seminis, is simplyan outgrowth of the lateral vaginal canal at the point atwhich it bends round to pass posteriorly. Throughout theirentire extent the lateral vaginal canals are embedded in thetissue of the urogenital strand, which represents nothing elsethan the persistent genital cord of the pouch young.

Sections through the urogenital strand of this stage showthat the connective tissue separating the lateral vaginalcanals, in which in uni- or multi-parous females the pseudo-vaginal passage is found to occur, is quite uniform incharacter, and only differs from that of the virgin in beingvery vascular.

In females, then, which have not given birth to young,the two median vaginal cul-de-sacs do not open into eachother, and, as I pointed out before, no trace of a pseudo-vaginal passage is present. But after parturition has oncebeen effected, the two median vagina are found to open intoeach other posteriorly to form a short common median vagina,and from this the pseudo-vaginal passage leads away. Theyoung in Perameles thus reach the exterior by way of amedian passage constituted in front by a short epitheliallylined tube a few millimetres in length—the commonmedian vagina—and behind by a relatively extremely longcleft-like space, 3 to 4 cm. in length, entirely destitute of aDepithelial lining—the median pseudo-vag ina l passage—situated in the connective tissue between, the lateral vaginalcanals.

In the paper above referred to the phenomena of parturi-tion are described at length, and the conclusion is arrived at" that Perameles, in respect to the phenomena connectedwith that process, stands in no way alone amongst Marsupialsas an aberrant and specialised type, but, quite on the contrary,exhibits more primitive features in the mode of birth of theyoung than are shown by any other Marsupial hithertodescribed as possessing a direct median passage." In

EMBRYOLOGY OF THE MARSUPIALIA. 1 3

concluding the description of this stage I may remark that,since the completion of my previous paper on the placenta-tion of Perameles, I have discovered that the native cat,Dasyurus v ive r r i na s , is also provided with a placentalConnection of a somewhat complex nature and probably ofyolk-sac origin, but whether or not there is also present anftllantoic placental connection I am unable to state. Myseries of stages is as yet very incomplete, and does notpermit me to enter into details, but it is certain that oversome portions of the inner surface of the uterus there ispresent a syncytial layer formed by the fusion of the ectodermof the bilaminar omphalopleure with the uterine epithelium,the nuclei of which form nests, vividly recalling the similarnuclear nests of the syncytium of Perameles, and this conjointlayer is vascularised by maternal capillaries. But whereasin Perameles the syncytium clothing the inner surface of theuterus is entirely of maternal origin, being derived from theUterine epithelium, in Dasyurus it is partly of foetal origin(foetal ectoderm) and partly of maternal origin (uterineepithelium).

The foetal membranes are left behind in the uterus afterparturition, and, as in Perameles, the foetal and maternalportions of the placenta are absorbed in s i t u through theagency of maternal leucocytes. In other words, the placentaof Dasyurus, like that of Perameles, is contra-deciduate incharacter. "We have, then, in any discussion bearing on thegeneral question of the significance of the occurrence of auallantoic placenta in Perameles, to face the fact that a placentaof the contra-deciduate type has been found to exist in twodistinct genera of Polyprotodont Marsupials, and that theuterine epithelium in both forms undergoes an essentiallysimilar transformation into a syncytium vascularised bymaternal capillaries.

14 . JAS. P. HILL.

APPENDIX.

Organisation of Embryo, P. obesula.(Plate 1, fig. 3.)

Measurements.—Greatest length, occiput to roundedhinder end of body, 125 mm. Head length 6*5 mm.

Form of Body.—Head raised, but. not so much as innew-born, and forming an acute angle with the trunk. Pro-minent snout. Slight neck protuberance. Lips fused late-rally to form " Saugmund," the line of junction stillrecognisable. Mouth lozenge-shaped. Tongue projectsprominently from mouth, and is grooved dorsally. Byesdistinct, roughly oval in shape, and not covered by epi-trichium. Eyelids not yet united. Position of lachrymalduct recognisable. Position of external ear just recognis-able, covered by epitrichium. In front of cloacal opening isa small projecting "genital Hooker."

Limbs.—Much as in new-born, except that the claws areless developed. Fore-limb flexed at elbow, and palmar sur-face directed backwards. The first and fifth digits quitesmall tubercles j second, third, and fourth provided withshort, blunt, recurved claws; the third digit the largest.Hind limb paddle-like, with plantar surface directed me-sially. Digits all indicated, but not free from each other;the fourth the largest.

Notochord and Vertebral Column.—Notochord intra-vertebrally constricted. Marked cartilaginous centra withtransverse processes and neural arches, the latter not yetunited above the spinal cord.

Nervous System.—Marked fissura arcuata along mesialhemisphere wall. Coi'pus striatum not yet prominent. Bud-ding from hypophysis anteriorly, but not very marked.

Eye.—Outer wall of optic cup about half pigmented.Ovalish lens cavity. Optic stalk solid in its middle portion.Eyelids not yet united. Eye muscles developing.

EMBRYOLOGY OP THE MARSUPIAL1A. 15

Ear.—Semicircular canals formed. Cochlea bent. Perioticcapsule partly pro-cartilage, partly cartilage. External audi-tory meatus plugged by epidermal cells.

Nose.—Much as in new-born. Ghoanss formed; turbinalprojections arising, and Jacobson's organ formed and itscartilage present. Solid lachrymal ducts, which reach thenasal epithelium.

Mouth.—Palate formed, and invests glottis posteriorly.Olub-like tooth germs, but no enamel organs. Tonguegrooved and with distinct taste-buds. Submaxillary salivaryglands open into mouth.

A l imen ta ry Canal, etc.—Lateral thyroid Anlagen onsides of trachea. Thymus Anlagen approximated in theirmiddle regions, but free posteriorly. Extensive pancreas,its duct opening into bile-duct. Lungs with numerous simplealveoli. Cartilaginous rings round trachea. Diaphragmcomplete and muscle-fibres developing in it. Intestinal loopno longer projects at navel.

H e a r t and Vessels.—Ventricular septum not yet com-plete. Auricular septum complete and a foramen ovalepresent.. Sinus venosus still extensive. Truncus aortsoalmost completely divided. Left ductus Botalli present.Single dorsal aorta. Renal portal circulation in connectionwith Wolffian bodies. Allantoic circulation greatly pre-dominates over yolk-sac circulation.

U r o g e n i t a l System.—Mesonephros of large size. Peri-toneal funnels of Mullerian ducts present. Commencingr^nal tubules. Ureters open into urogenital sinus mesiallyto Wolffian ducts. Genital Leisten prominent. Sex ?

Sk in and Skeleton.—Thin nucleated epitrichial layeroh epidermis. Hair Anlagen well advanced on head. No ossi-fication in cartilaginous skeleton. Pre-maxillary, maxillary,and palatine ossifications. Rudimentary clavicle ossified.1

1 Tliis is also present in the 8-75 mm. enibryo of Stage D.

16 ' JAS. P. HILL.

I I I . On the Foetal Membranes of Macropus parma.

THE following account of the foetal membranes of thissmall wallaby is based on the examination of a single em-bryo from the left uterus, for which I am much indebted tomy friend Mr. J. J. Fletcher, M.A., B.Sc. -

HISTORICAL.

In 1834, Owen (4) gave an account of the foetal membranesof an embryo of M. major, 7 lines in greatest length; butalthough no allantois was found in this embryo, he stated,from the presence in young mammary foetuses of "the Kan-garoo, Phalangista, and Petaurus," of " the remains ofa urachus and umbilical vessels," that " i t would appear thatan allautois and umbilical vessels are developed at a laterperiod of gestation, but probably not to a greater extentthan to serve as a receptacle of urine" (4, p. 342). Theaccuracy of this conclusion he definitely established in 1837,through the examination of a uterine embryo of M. major,10 lines in length. He says (5, p. 83), "Below the neck ofthis [vitelline] sac there is extended a second pyriform sac,about one sixth the size of the vitelline sac, having numerousramifications of the umbilical vessels and constituting atrue allantois. This sac was suspended freely from the endof the umbilical cord; it had no connection at any part ofits circumference with the chorion, and consequently wasequally free from attachment to the parietes of the uterus inwhich the foetus was developed."1

Between 1837 and 1881 nothing of importance seems tohave been published on the embryogeny of Macropus. Inthe latter year Chapman again described (7) the foetal mem-branes of M. major, " six eighths of an inch in length fromthe mouth to the root of the tail." His account essentially

1 Owen states that lie was anticipated in the description of the allantois ofthis embryo by Coste (6), with whom he had examined the foetal membranes.

EMBRYOLOGY OJ? THE MARSUPIALIA. 17

confirms Owen's observations, and adds nothing of import-ance thereto. In 1884, Caldwell (8), in a paper on thearrangement of the foetal membranes in Marsupials, firstclearly pointed out that the embryo lay in the splanchnocoele,surrounded by the invaginated upper portion of the yolk-sac wall. He described the general arrangement of thefoetal membrane in Phascolarctus cinereus, and showedthat the allantois in this form fuses with the discoidal area oftrue chorion. He also referred to the membranes of M.rufficollis, but in such a way as to lead the reader tosuppose that in this form also the allantois fuses with thechorion.

Semon (2), in 1894, described the foetal membranes ofPhascolarctus cinereus (confirming Caldwell's discoveryof the fusion of the allantois with the chorion), as well asjEpyprymnus rufescens. In this latter form he insistedon the rudimentary character of the allantois and its non-fusion with the chorion, and expressed the opinion that sucha condition would be found to hold good for the Macropodida3in general. In 1895 I shortly stated (9) that such wascertajnly the state for four species of the genus, viz. M.parma, M. ruficollis, M. robustus, and M. major.

The present communication deals with M. parma alone.

External Characters of Embryo.

The embryo, at least so far as external characters go,appears to be in a somewhat earlier stage of developmentthan the previously described embryo of P. obesula. Un-fortunately it is not sufficiently well preserved for sectionis-ing. It has a greatest length of 10 mm., and a head lengthof about 6 mm., and is characterised as follows :—Head bentalmost parallel with trunk. Prominent neck protuberance.Lips not fused. Eye rounded. Lachrymal groove present.Rounded external nares. Snout little prominent. Wideexternal auditary meatus, with triangular ear-pinna. Fore-limb not flexed, digits all present, and with claw Anlagen,

VOL, 4 3 , PART 1. NEW SERIES. B

18 ,TAS. P. HILL.

palmar surface directed mesially. Hind limb paddle-like,digits not indicated, plantar surface directed mesially.

Foetal Membranes .

The foetal membranes in M. parma have the same generalarrangement as in i E p y p r y m n u s rufescens , as describedby Semon (2, p. 25), i. e. the embryo enclosed in its amnionlies in the extra-embryonic splanchnoccele surrounded by theupper part of the yolk-sac—the yolk-sac splanchnopleure—which is invaginated into the cavity of the latter. Thesplanchnocoele is closed externally by a discoidal area ofchorion, round the margin of which the yolk-sac splanchno-pleure becomes continuous with the vascular omphalopleure.The allantois is a comparatively small short-stalked vesicle,which never reaches and fuses with the chorion, but lies per-manently buried in the splanchnocoele.

The entire globular embryonic formation, of this stage ofM. pa rma has a diameter of 15 mm.

In Plate 2, fig. 9, the cavity of the yolk-sac has beenopened by the removal of the greater part of the bilaminaromphalopleure, exposing the embryo in its amnion, closelysurrounded by the invaginated yolk splanchnopleure (y. spl.),together with the inner surface of the vascular area (vase,ompli.).

The discoidal area of chorion, concealed in the figure bythe embryo, is readily distinguishable externally from thevascular omphalopleure by its more transparent and smootherappearance. Compared with the extent of the choi-ionic areain Perameles and in Phascolai*ctus, where, according toCaldwell (8), the chorion has a diameter of 12 mm., the areais here a small one, measuring in greatest diameter 4-5 mm.

The surface of the vascular area (vascular omphalopleure)presents a roughened reticulate appearance in correspondencewith the ridged surface of the uterine mucosa. It is distin-guished from the bilaminar omphalopleure by its denser anddarker appearance. The latter (Plate 2, fig. 9, bil. omph.)

EMBRYOLOGY OF THE 1IARSUPIALIA. 19

occupies rather less than half the surface of the entireembryonic formation. Its surface is also irregularly ridged.

The vitelline artery, leaving the yolk-stalk on the left sideof the embryo, courses in the yolk splanclmopleure backwardsover its dorsal surface to reach the junction of the lattermembrane with the vascular omphalopleure (Plate 2, fig. 9,vit. a.). Here it passes over into the latter, does not at oncedivide as in Perameles, but runs on for a distance of about4 turn, before dividing into its two branches, which form thesinus terminalis (s.t,). The sinus is completed by a narrowanastomosis, as in Perameles. It measures in diameter 14mm. by 12 mm. The vascular area has a greatest width of8 mm. opposite the point of bifurcation of the vitelline artery,and a least width of 4 mm. at that point. It is thus evidentfrom these measurements that in mere surface extent thevascular area of this Macropod embryo very considerablyexceeds that of the Perameles embryo described in the pre-ceding pages. And not only is the area greater in extent,but its capillary system is also much more richly developedthan in that form, in correlation with its high functionalimportance in the nutrition of the embryo. From the sinusthere pass off into the vascular area numerous branches,1

which divide up into much finer branches to form a richcapillary system, from which the factors of the vitelline veinstake their origin. The numerous fine long wavy branches ofth$ vitelline artery which run in the yolk splanchnopleureare especially well developed (Plate 2, fig. 9, a. v.). Above,these branches come off directly from the vitelline artery,while below they arise from an independent branch ofth$ artery which runs parallel with the same. The bloodcirculating in these vessels must, as in Perameles, pass overinto the vascular area at the junction of the yolk splanclmo-pleure with the vascular omphalopleure. After the vitellineartery has passed over into the vascular omphalopleure, itstill continues to give off small branches from both sidesalmost up to its point of bifurcation.

i In the figure these branches have been represented too short.

20 JAS. P. HILL.

The vitelline veins (Plate 2, fig. 9, vit. v.) show the usualrelations, and include between them a very large area ofyolk-sac splanchnopleure altogether devoid of vessels. Theanterior vein is formed by the union in the yolk splanchno-pleure of a larger anterior and a smaller posterior factor,which pass over from the vascular area. There each isformed by the union of lesser factors coming from the capil-lary net of the same. Similarly, the posterior vein is formedby the union of two factors, a smaller anterior and a muchlarger posterior factor, returning the blood from the poste-rior portion of the vascular area. The anterior vein passesback in the splanchnopleure over the head of the embryobehind the eye, while the posterior one passes in transverselyon a level with the snout. The two unite at the yolk-stalkto form the single vitelline vein.

The allantois is a small, globular, thin-walled vesicle pro-vided with a very short stalk, and measuring in diameter 5"5mm. It is situated mainly on the embryo's right side, lyingbetween the snout and the rounded hinder end of the body.In the figure (Plate 2, fig. 9, all.) it is seen through theyolk splanchnopleure, in contact with which it lies. Itsvessels are poorly developed. As Owen long ago pointedout for M. major, and as Semon has recently described for^SDpyprymnuSj so here the allantois is a quite rudimentarystructure which lies remote from the chorionic area, andnever comes into contact, much less fusion, with the same.Such a specialised condition of the allantois is probablygeneral for the entire family of the Macropodidae. That italso exists outside the limits of that family we know fromthe observations of Osborn (10) and Selenka (3) on Didelphys,and probably the condition is also general for the familyPhalaugeridas—it certainly holds for P e t a u r u s sc iurensaccording to Semon, and from my own observations I can addfor Tr i chosurus vulpecula . But whether the conditionmay turn out to be the general one for Marsupials or not, itis admitted on all hands to be an essentially modified one.

UNIVEHSITV OP SiTDNEY, N.S.W. ;March 16th, 1899.

EMBRYOLOGY OP THE MARSUPIALIA. 21

LIST OP PAPERS REFERRED TO.

1. HILL, J. P.—"The Placentation of Perameles," 'Quart. Journ. Micr.Sci.,' vol. xl, p. 385.

2. SEMON, R.— "Die Embryonallmllen der Monotiemen und Marsiipialer,"' Zool. Forsch. in Australien u. dem Malayisclien Archipel.,' BJ. ii.

3. SELENKA, B.—' Studien iiher Entwickelungsgeschichte der Tliiere : ' IV(1 and 2), Das Opossum (Didelphys virginiana), Wiesbaden,1886.

4. OWEN, R.—"On the Generation of Marsupial Animals, etc.," 'Phil.Trans.,' 1834.

5. OWEN, R.—"On the Existence of an Allantois in a Foetal Kangaroo,"'Proc. Zool. Soc.,' pt. v, 1837, p. 82. Cf. also ' Comp. Anat. ofVertebr.,' vol. iii, p. 569; and Todd's 'Cyclop, of Anat. and Pliys./vol. iii, p. 324, fig. 141.

6. COSTE, M.— "Vesicule allantolde observed dans l'ceuf des Kangaroo,"' Compt. Rend. Acad. Sci.,' Paris, torn, v, 1837.

7. CHAPMAN, H. C.—"On a Foetal Kangaroo and its Membranes," 'Proc.Acad. Nat. Sei. Philadelphia,' pt. iii, 1881, p. 468.

8. CALDWELL, W. II.— "On the Arrangement of the Embryonic Mem-branes in Marsupial Animals," ' Quart. Journ. Micr. Sci.,' vol. xxiv,1884.

9. HILL, J. P.—"Preliminary Note on the Occurrence of a Placental Con-nection in JPerameles obesula and on the Foetal Membranes ofCertain Macropods," 'Proc. Linn. Soc. N.S.W.,' vol. x (2nd ser.),pt. iv, 1895.

10. OSBORN, H. F.—"Observations upon the Foetal Membranes of theOpossum and other Marsupials," 'Quart. Journ. Micr. Sci.,' vol. xxiii,1883. "The Foetal Membranes of the Marsupials: the Yolk-sacPlacenta in Didelphys," 'Journ. of Morph.,' vol. i, 1888.

22 JAS. P. HILL.

EXPLANATION OP PLATES 1 & 2,

Illustrating Mr. Jas. P. Hill's paper, " Contributions to theEmbryology of the Marsupialia" (II and III).

All sections drawn were outlined by means of Zeiss's camera lucida.

L I S T O F C O M M O N R E F E R E N C E L E T T E R S .

Amn. Amnion. a. v. Branches of vitelline artery on yolk splanclmopleure.all. a. Allantoic artery, all. c. Allantoic cavity, all. cap. Allantoic capil-lary, all. cl. Allantoic canal, all. enl. Allantoic entoderm. all. vies. Al-lanto-chorionic mesenchyme. all. v. Allantoic vein. bil. omph. Bilaminaromphalopleure. eh. Cliorion. ch. eel. Chorionic ectoderm, cm. Cnelom.cos. to. Ccelomic wall of allantois. eel. Ectoderm, enl. Entoderm. ex. syn.Syncytium beyond allantoic placental area. mes. Mesoderm of pro-amnioticremnant, pi. a. Allantoic placental area. pi. st/n. Syncytium of allantoicplacental area. Som. Somatic mesoderm of cliorion. Sp. Space in syncytium(lymph space ?). s. t. Sinus terminalis. syn. c. Capillary of. syncytium.w. s. Umbilical stalk, ut. w. Uterine wall. vase. omph. Vascular omphalo-pleure. vit. a. Vitelline artery, vil. v. Vitelline vein. y. spl. Yolk-sacsplanclmopleure.

F I G . 1.—Right uterus, P . o b e s u l a , with ventral half of its wall removedand portions of the foetal membranes reflected, to expose the embryo. Xnearly 7.

F I G . 2.—The same, afler removal of the embryo, to show the placentalarea {pi. a.), x G^\j.

P I G . 3.—Embryo, P . o b e s u l a . X 6^ j .

I"TG. 4.—Transverse section through the umbilical stalk, som. Somato-

pleure of body.wall. X 90.

T I G . 5.—Transverse section through pro-amniotic remnant, x 90.

T I G . 6.—Section through margin of allantoic placental area. X 120.

F I G . 7.—Section through allantoic placeutal area, showing the relation ofthe fetal and maternal capillaries. X 3S0.

FIG. 8.—Section through the sinus terminalis (s. I.), and the adjacent por-tions of the vascular {vase, omph.) and bilaminar omphalopleure {bil. omph.).X 150.

F I G . 9.—Uterine-embryo of M. p a r m a in its fetal membranes, seen afterremoval of the greater portion of the bilaminar omphalopleure. X 7i .

%twr6.c9oxm, ' 7& 43. IsM /.

Fie. I.

y.spo.- •

vu, a . •-•

bil.oniok — •; \-t—'r- •

- ait

A C. f .: R K I"!

Fiif. S

r - i ; • • - / ; / > .

J P H & r CV; F Huth. L;thr 2d:-