Difference between revisions of "Paper - Contributions to the embryology of the marsupialia 4"

From Embryology
(9 intermediate revisions by the same user not shown)
Line 1: Line 1:
{{Hill1910 header}}  
=The Early Development of the Marsupialia, with Special Reference to the Native Cat (''Dasyurus Viverrinus'')=
[[File:James Peter Hill.jpg|thumb|alt=James Peter Hill|link=Embryology History - James Hill|James Peter Hill]]
=The Early Development of the Marsupialia, with Special Reference to the Native Cat (Dasyurus Viverrinus)=
(Contributions to the Embryology of the Marsupialia, IV.)  
(Contributions to the Embryology of the Marsupialia, IV.)  
Line 17: Line 15:
Table op Contents.
==Table of Contents==
Introduction . . . . . .2
[[Paper - Contributions to the embryology of the marsupialia 4-1|Chapter I.  -  Critical Review of Previous Observations ON THE Early Development op Marsupialia]]
[[Paper - Contributions to the embryology of the marsupialia 4-2|Chapter II.  -  The Ovum of Dasyurus]]
1. Structure of Ovarian Ovum . . .11
2. Maturation and Ovulation . . .21
3. Secondary Egg-membranes . . .23
4. Uterine Ovum . . . .25
[[Paper - Contributions to the embryology of the marsupialia 4-3|Chapter III.  -  Cleavage and Formation of Blastocyst]]
1. Cleavage . . . . .28
2. Formation of Blastocyst . . .37
[[Paper - Contributions to the embryology of the marsupialia 4-4|Chapter IV.  -  Growth of Blastocyst and Differentiation OP THE Embryonal Ectoderm and the Entoderm]]
1. Growth of Blastocyst . . .43
2. Differentiation of the Embryonal Ectoderm and
the Entoderm . . . .52
3. Establishment of the Definitive Embryonal Area 65
4. Summary . . . .72
[[Paper - Contributions to the embryology of the marsupialia 4-5|Chapter V.  -  Some Early Stages of Perameles and Macropus]]
[[Paper - Contributions to the embryology of the marsupialia 4-6|Chapter VI.  -  General Summary and Conclusions]]
[[Paper - Contributions to the embryology of the marsupialia 4-7|Chapter VII. -  The Early Ontogeny of the Mammalia in the Light of the Foregoing Observations
1 . The Early Development of the Prototheria . 86
2. The Early Development of the Metatheria and Eutheria . . . .96
3. The Entypic Condition of the Eutherian Blastocyst .... Ill
Addendum ...... 121
List op References ..... 122
[[Paper - Contributions to the embryology of the marsupialia 4-Plates|Explanation of Plates]]
“ In mammalian embiyology very many sm-prises are yet in store for
ns ” (Hnhrecht, ’08).
The present contribution contains an account of the nrincipal results and conclusions at which I have arrived after a
somewhat protracted and much interrupted study of an
extensive collection of early developmental stages of Marsupials, ranging from the fertilised egg to the blastocyst in which
the two primary germ layers are definitely established. I
believe I nm now able to give for the first time an account of
early Marsupial ontogeny, based on the examination of an
adequate material, and both consistent in itself and with Avhat
we know of the early development in the other two Mammalian sub-classes. The material at my disposal was obtained
during my tenure of office in the University of Sydney, and
with the aid of grants from the Royal Society and of a George
Heriot Research Fellowship. It represents the proceeds of
some eight years’ collecting, and comprises a fairly complete
series of stages of the native cat (Dasyurus viverrinus),
together with a few early stages of other Marsupials, notably
Perameles and Macropus.
Dasyurus proved in many ways a convenient subject for
embryological purposes. It can readily be trapped in many
WELLCOME NeAv South Wales; it lives and breeds fairly well
LIBF’^R’in captb ity, and though always somewhat intractable, it can,
its size, be easily handled, and so may be subjected
if iiecBSsary to daily exarainationd But it lias this great disadvantage, which it apparently shares with other Marsupials,
that a very variable period intervenes between coitus and
ovulation. As a consequence, the obtaining of any desired
cleavage or early blastocyst stage is largely a matter of
chance.^ It is true that the changes which take place in the
pouch, in correlation with ovulation and the events connected
therewith, do afford in the case of late pregnant females some
indication of the stage of development likely to be met with,
but these changes are at first of too indefinite a character to
be of much service beyond indicating that ovulation may have
taken place.
Dasyurus breeds but once a year, the breeding season
extending over the winter months  -  May to August. One
remarkable featui'e in the reproduction of Dasyurus, to which
I have directed attention in a previous paper (Hill, ’00), may
be again referred to here, and that is the fact that there is no
correlation between the number of ova shed during ovulation
and the accommodation available in the pouch. The normal
number of teats present in the latter is six, though the
pi-esence of one or two supernumerary teats is not uncommon;
the number of ova shed at one period is, as a rule, far in
excess of the teat number. I have, for example, several
records of the occurrence of from twenty to twenty-five eggs,
two of twenty-eight, one of thirty, and one of as many as
thirty-five! (twenty-three normal blastocysts and twelve
^ Perameles, on the other hand, though quite common in many parts
of tlie State, is hy no means such a convenient type. It is much less
easily trapped than Dasyurus, does not live nearly so well in captivity,
and is pai'ticularly difficiilt to handle. I have to thank Mr. D. G. Stead,
now of the Department of Fisheries, Sydney, for first directing my
attention to the breeding habits of Dasyurus, and also for providing
me with the first female from which I obtained segmenting eggs.
^ For example, I obtained unsegmented ova from the uteri, four, five,
six, seven and eight days after coitus, 2-celled eggs six and seven
days after, 4-celled eggs eleven and eighteen days after. In one case
the young were bom eight days after the last observed act of coitus,
in another sixteen days after, and in yet another twenty days after.
abnormal), there can be little doubt that Dasyurus, like
various other Marsupials (e.g. Perameles, Macropus, etc.),
has suffered a progressive reduction in the number of young
reared, but even making due allowance for that, the excess
in production of ova over requii’eraents would still be remarkable enough. Whether this over-production is to be correlated
in any way with the occurrence of abnormalities during early
development or not, the fact remains that cleavage abnormalities are quite frequently met with in Dasyurus.
Technique.  -  As fixatives, I have employed for ovaries
the fluids of Hermann, Flemming, Ohlmacher, and Zenker;
for ova and early blastocysts, Hermann, Flemming, Perenyi,
and especially picro-nitro-osmic acid (picro-nitric acid [Mayer]
96 C.C., 1 per cent, osmic acid 2 c.c., glac. acetic acid 2 c.c.) ;
for later blastocysts, the last-named fluid especiall}'^ also
picro-corrosive-acetic aud corrosive-acetic.
To facilitate the handling of ova and early blastocysts
during embedding, I found it convenient to attach each
specimen separately to a small square of pig’s foetal membrane
by means of a dilute solution of photoxylin (1 to 2 per cent.),
Orientation of the specimen was then easily effected during
final embedding, under the low power of the microscope. The
larger blastocysts were double-embedded in photoxylin and
paraffin, the cavity of the blastocyst being tensely filled with
the photoxylin solution by means of a hypodermic syringe
fitted with a fine needle.
For the staining of sections, Heidenhain’s iron-htematoxylin method proved the most satisfactory, and was almost
exclusively employed. Entire portions of the blastocyst wall
were stained either with Ehrlich’s or Delafield’s haematoxylin.
I am much indebted to Mr. L. Scbaeffer, of the Anatomical
Department of the University of Sydney, and to Mr. F.
Pittock, of the Zoological Department, University College,
for invaluable assistance in the preparation of the photomicrographs reproduced on Plates 1 - 5, and also to Mr. A.
Cronin, of Sydney, and Miss M. Rhodes, for the drawings
fi’om their respective pencils reproduced on Plates 6 aud 7.
To Miss V. Sheffield I am indebted for tlie original of fig. 63.
To my friend Dr. F. P. Sandes, Sydney, I am indebted for
kind help in the revision of certain parts of the manuscript.
Chapter I.  -  Critical Eeview op*Peevious Observations on
THE Early Development of the Marsupialia.
Apart from the very brief abstract of a short paper on the
development of Dasyurus, which I read before Section D of
the British Association in 1908 (included in Dr. Ashworth’s
Report, ^Natui'e,’ vol. Ixxviii), our knowledge of the processes
of cleavage and germ-layer formation in the Marsupialia is
based (1) on the well-known observations of the late Emil
Selenka (’86) on the development of the Virginian opossum
(Didelphys mars upialis), published in 1886 as Heft 4 of
his classical 'Studien’; and (2) on those of W. H. Caldwell
(’87) on the uterine ovum, and cleavage process in the native
bear (Phascolarct us cinereus).
Selenka’s account of the mode of origin of the germ-layers
in Didelphys differs widely, us the sequel will show, from my
description of the same in Dasyurus. Now Didelphys and
Dasyurus are two marsupials, admittedly allied by the closest
structural ties, and we should therefore not expect a priori
that they Avould differ fundamentally in the details of their
eai’ly ontogeny, however much they might diverge in respect
of the details of their embryonal nutritional arrangements.
Furthermore, we might reasonably hope, in view of the
generally admitted relationships of the Marsupialia, that a
knowledge of their early development would aid us in the
interpretation of that of Eutheria, or, at least, that their
early developmental phenomena would be readily comparable
with those of Eutheria. It cannot be said that Seleuka’s
observations realise either of these expectations. Whichever view is taken of Seleuka’s description of the opossum,”
writes Assheton (’98, p. 254), “ many obvious difficulties
remain for the solution of which no satisfactory suggestion
can as yet be offered.”
As concerns niy own observations, I venture to think it is
possible to bring* them into line with what we know of the
eai’ly ontogeny in the other two mammalian sub-classes, and
I have attempted to do so in the concluding chapter of this
paper, with what success the«‘eader can judge, whilst as regai'ds
the divergence between Selenka’s results and my own, I am
perfectly convinced that the explanation thereof is to be
found in the fact that the whole of Selenka’s early material
was derived from but two pregnant females, and that much
of it consequently consisted of eggs which had failed to
develop normallj". From the one female, killed 5 days
after coition, he obtained one egg in the 2-celled stage,
one with about twenty cells and nine unfertilised ova. From
the second, killed 5 days 8 hours after coition, he obtained
“ ausser zwei tauben, 14 befruchtete Eier niimlich je ein Ei mit
4, 8, 42, 68 Zellen, eine junge und eine iiltere Gastrnla mit
noch dicker Eiweisschicht und endlich acht auch gleichen
Entwickelungsstufe stehende Aveit grossere Keimblasen, deren
AVand noch grosstentheils einschichtig war” (’86, p. 112).
Selenka recognised that the last-mentioned blastocyst “die
normale Entwickelungsphase reprasentiren,” since he found
as a rule that all the embryos from one uterus Avere in the same
developmental stage. Nevertheless he proceeded to describe
the segmenting eggs and the two “ gastrulas ” Avhich lagged so
far behind the blastocysts, as if they Avere perfectly normal
developmental stages. He does, indeed, question Avhether or
not the 42-celled stage is normal, but decides in the affirmative, “ denn Avenn ich von zwei Zweifelhaften Fallen absehe,
so habe ich niemals Eier aus den ersten Tag aufgefunden,
Avelche auf irgend Avelche Anomalie der EutAvickeluiig
hiiiAviesen.” This, hoAvever, can hardly be accepted as a
satisfactory reason for his conclusion, since apart from the
other eggs of the same batch, he had but the two eggs from
the other female for comparison, viz. the 2-celled egg (and
even that is, in my vieAV, not quite normal), and the 20-celled
egg, Avhich is stated to have suffered in preparation. AVith
the exception of the tAvo eggs just mentioned, all the crucial
early stages (ranging from the 4-celled stage to the completed
blastocyst), on whose examination Selenka based his account
of germ-layer formation in Didelphys, would thus appear to
have been derived from a single female.^ No wonder it is
impossible to reconcile his description either with what we
know of germ-layer formation in the Prototheria and Eutheria
or with my account of the same in Dasyurus.
My own experience with the latter has shown me that no
reliance whatever is to be placed on segmenting eggs or
blastocysts which exhibit marked retardation in tbeir stage
of development as compared with others from the same
uterus, and also that batches of eggs or blastocysts in which
there is marked variation in the stage of development attained
should likewise be rejected. Abnormalities in the process of
cleavage and of blastocyst formation are by no means uncommon in Dasyurus, and during the earlier stages of my
own work I spent much time and labour on the investigation
of just such abnormal material as that on which Selenka, no
doubt unwittingly, but I feel bound to add, with an utter
disregard for caution, based his account of the early development of Didelphys.
I propose now, before passing to my own observations, to
give a short critical account of Selenka’s observations, my
comments being enclosed in square brackets.
The uterine ovum of Didelphys is enclosed by (1) a relatively thin “ granulosamembran,” formed by the transformation of a layer of follicular cells [really the shell-membrane,
first coiTectly interpreted by Caldwell C87) and formed in the
Fallopian tube] ; (2) a laminated hiyer of albumen, semitransparent ; (3) a zona radiata, not always recognisable [in my
experience invariably distinct] .
Cleavage begins in the uterus, is holoblastic, and at first
equal. A 2-celled stage is figured (Taf. xvii, fig. 3) [not
quite normal as regards the relations of the blastomeres], and
also a 4-celled stage [normal in appearance except for the
* The collection of my own early material of Dasyurus has involved
the slaughter of over seven dozen females.
enormous tliickness of the albumen layer], iu which the four
equal-sized blastomeres are radially arranged round a cleavage
cavity and are conical in form, their upper ends being more
pointed, their lower ends thicker and richer in yolk-material.
The nucleus of each is excentric, being situated nearer the
upper pole. [This description is applicable word for word to
the4-celled stage of Dasyurus.]
An 8-celled stage (fig. 6) is next described, seven of the
blastomeres being equal in size and one being smaller. They
are arranged somewhat irregularly iu two circles. [This stage
I regard as abnormal both in respect of the arrangement of
the blastomeres and the occurrence of irregularity amongst
them.] Selenka (p. 119) thought it probable that the third
cleavage planes cut the first two at right angles and divided
each of the first four blastomeres into a smaller ectodermal
cell and a larger more granular entodermal, but states that
he was unable to establish this owing to the opacity of the
albumen-layer. [My observations show that it is the fourth
cleavage in Dasyurus, not the third, which is equatorial, unequal, and qualitative, and that even then the cells formed are
not ectodermal and entodermal insignificance. The albumen
is normally never opaque.]
A 20-celled stage is mentioned, but not described, since it
suffered in preparation. It is said to have a large entoderm
cell in the cleavage cavity. [A statement of very doubtful
value, since the blastomeres were admittedly pressed together
and probably displaced by the shrunken egg-membranes.]
The next stage described is a spherical “gastrula^’ (Taf. xvii,
figs. 7, 8), composed of fox’ty-two cells with an open "blastopore^’ at the vegetative pole, a smaller opening at the animal
pole, and a large “ ur-entoderm ” cell in the cleavage-cavity,
just inside the "blastopore.” The wall of the "gastrula”
consists of cells graduated in size ; those in the region of the
blastopore are the largest and richest in deutoplasm, those at
the opposite pole are the smallest and most transparent. [This
is a very characteristic stage in the formation of the blastocyst,
with which I am quite familar in Dasyurus. Selenka’s speci
men, judging from Dasyurus, is normal as regards the constitution of its wall and the occurrence of an opening at each pole.
The lower opening, however, has no blastoporic significance,
but, like the upper, owes its presence to the mode of formation
of the blastocyst-wall by the spreading of the blastomeres
towards the poles of the sphere formed by the egg-envelopes.
Selenka’s blastopore simply marks the last point of closure.
This specimen I hold to be abnormal from the presence of
the so-called “ urentoderm ” cell in its interior.- I figure
(PI. 3, fig. 37) a section of a fairly comparable and undoubtedly abnormal blastocyst of Dasyurus in which there is
also present in the blastocyst cavity a large free cell. Here
this latter is unquestionably a blastomere of the lower hemisphere, which, having failed to divide, has become enclosed
by the spreading of its neighbours. Selenka^s “ urentodermzelle’'’ I regard as a similarly displaced blastomere.]
A 68-celled “gastrula^’ (figs. 9 and 10) is next described.
It is essentially similar to the preceding, only the “blastopore ” has closed.
The succeeding stage (fig. 11) is a somewhat older
“ gastrula,” in which gastrulation is said to be still in
progress, since over the lower pole, in the ]-egion of the now
closed blastopore, it is no longer possible to say which cells
belong to the ectoderm, which to the entoderm. The latter
layer is described as being several cells thick in the blastoporic
region, and as in course of spreading round inside the ectodermal wall of the“gastrula” to war ds the u pper o r ani mal
pole. [This specimen is undoubtedly abnormal, at all events
there is no comparable stage in Dasyurus. It is difficult to
obtain a clear idea of Seleuka’s conception of the mode of origin
of the germ-layers, but he evidently held (cf . pp. 116 and 119)
that the large yolk-rich cells of the lower (“ blastoporic ”)
pole constitute the anlage of the entoderm, and that they
become inturned at the “ blastopore ” and pi’oliferate to form
the definitive entoderm, which then gradually extends round to
the animal pole, in contact with the inner surface of the wall
of the gastrula., that wall forming the ectoderm. He appa
rently did not regard the “ urentodermzelle ” as the sole
progenitor of the entoderm, but simply as an entoderm-cell
precociously inturned from the blastoporic ” margin.
This view of Selenka, however, lands us in the predicament
of having to regal’d the embryonal area as differentiating over
the vegetative hemisphere, since in the next stage the
‘'blastopore^’ is described as being situated excentrically in
that ai’ea. Either Selenka’s detei’mination of the poles in the
42-celled blastocyst is wrong, or the entoderm does not
originate as he describes it. My own observations force me
to accept the latter alternative. In his paper Selenka gets
over the difficulty very easily by altering the orientation of
his figures. On Taf. xvii, the figures of sections of blastocysts are so placed that the “blastopore ” is below, next the
bottom of the plate. These figures I hold to be correctly
orientated. On Taf. xviii, the figui’es are inverted, so that
the “blastopore” is above; as the result the animal pole of
fig. 11, Taf. xvii, becomes the' vegetative pole of the stage
next described (fig. 2, Taf. xviii).]
The stage just referred to, described as an “eiformige
gastrula,” is represented in a drawing made from the fresh
specimen as lying quite free in a large perivitelline space
enclosed by a very thick layer of albumen, outside which is
the “ granulosa-membran.” In section (fig. 2) a mass of
entoderm is seen to reach the sm-face at one pole (marked hi.)
uppermost in the figure, whilst other entodermal cells are
shown spreading from this towards the lower pole. The
ectoderm of the wall is represented as composed of definitely
cubical cells. [The presence of a large perivitelline space,
by itself stamps this specimen as not normal. The sectional
figure must be schematic.]
The last of Selenka’s early stages to which reference need
be made here is formed by eight “ gastrulas” (blastocysts),
reckoned as ten hours after the commencement of cleavage
[a reckoning I consider of no value] (Taf. xviii, figs. 3 and 4).
The embryonal area is now distinguishable by the larger size
of its ectodermal cells. The entoderm is unilamiuar, and has
extended beyond the limits of the embryonal area. The
position of the “ blastopore ” is said to be marked in all by a
mass of coag’ulum attached to the wall, and in three by a
definite opening' as well. It is situated excentrically in the
embryonal area. [Except for the ‘‘ blastopore and the
presence of a thick layer of albumen, this blastocyst stage is
quite comparable with the corresponding one in Dasyurus;
the latter, however, is considerably larger. Of Selenka’s
early material, I think it is these blastocysts alone which had
any chance of giving origin to normal embryos.]
W. H. Caldwell, who, as Balfour student, visited Australia
in 1883-4, obtained a very rich collection of early marsupial
Tnaterial, of which, unfortunately, no adequate account has
ever been published. He gave, however, in his introductory
paper on the ‘ Embryology of the Monotremata and Marsupialia^ (^87), an account of the structure of the ovum, both
ovarian and uterine, in Phascolarctus, and he showed that
the ovum during its passage down the Fallopian tube becomes
enclosed outside the albumen layer in “a. thin transparent
membrane, ’0015 mm. thick,” which he homologised with the
shell-membrane of the monotreme egg. This impoi*tant discovery of the existence of a shell-membrane in the Marsupialia I can fully confirm. I am, however, unable to accept
his interpretation of the internal structure of the ovum
of Phascolarctus, or his remarkable statement that cleavage
in that form is of the meroblastic type. Cleavage is not
described in detail, nor is any account given of the mode
of origin of the germ-layers.
Chaptee JI.  -  The Ovum op Dasyueus.
1. Structure of the Ovarian Ovum.
The full-grown ovarian ovum of Dasyurus (PI. 1, fig. 1)
appears as a rounded, or more usually, ovalish cell, the
diameter of which varies in section in ten eggs measured
from '28 X ‘126 mm. to '27 x '26 mm. (average, ’24 mm.),
and is therefore large relatively to the ova of Eutheria. It
is enclosed by a tliin, but very definite I’efractive membrane
or zona (vitelline membrane of Caldwell) of an approximate
thickness of ’002 mm. (fig. 1, z-p-), on which the cells of the
discus proligerus (fig. I, d.p.) directly abut, a differentiated
corona radiata and syncytial layer being absent. It appears
to be identical in its relations and optical characters with the
membrane investing the monotreme ovum, and never shows
in section any trace of radial striations (though I believe I
have detected an extremely faint appearance of such in the
fresh zona), or of the extension into it of protoplasmic
processes from the adjacent cells of the discus ‘proligerus,
such as Caldwell figures in the case of the ovum of Phascolarctus (cf. his PI. 29, fig. 5). Within the zona the
peripheral cytoplasm of the ovum is differentiated to form an
exceedingly thin but distinct bounding layer or egg-membrane
(vitelline membrane, sensu stricto).
The cytoplasmic body of the ovum exhibits a very obvious
and striking differentiation into two regions in correspondence
with the presence in it of two definitely localised varieties of
deutoplasmic material, respectively granular and fluid. Peripherally it consists of a relatively narrow cytoplasmic zone of
practically uniform width, dense and finely granular in
appearance owing to the presence in it of numei’ous particles
of deutoplasmic nature. This we may distinguish as the
foi'mative zone (fig. 1, f.z.). In it lies embedded the large
vesicular nucleus (about ‘06 x ‘03 mm. in diam.). Centrally
and forming the main bulk of the ovum is a mass of greatly
vacuolated cytoplasm presenting the appearance of a clear
wide-meshed reticulum. Its framework is coarser peripherally
where it passes over without definite limit into the formative
zone, with which it is structurallv identical, but much finer
and wider-meshed centrally, so fine, indeed, that it almost
invariably breaks down under the action of fixatives, and
appears in sections as an irregular space, perhaps crossed by
a few fine interlacing strands (fig. 1, d.z.). The meshes of
this reticulum are occupied by a clear fluid which must be
held to constitute the central deutoplasm of the egg. We
may accoi'dingly designate this central reticular area as the
deutoplasmic zone.
If we pass noAv fi-om the full-grown to the ripe ovarian
ovum (PI. 1, figs. 2 and 3), i. e. an ovum in which either thefirst polar spindle has appeared or the first polar body hasalready been separated off, it at once becomes evident that
important changes have occurred in the disposition and
relative proportions of the two constituent regions of the eggcytoplasm. The full-grown ovum is of the centrolecithal
type, the central deutoplasmic zone forming its main bulk
and being completely surrounded by the thin formative zone.
The ripe ovum, on the other hand, exhibits an obvious and
unmistakable polarity, and is of the telolecithal type, as the
following facts show. The cytoplasmic body evidently consists of the same two regions as form that of the full-gi-own
ovum, but here the dense formative region now forms its
main bulk, and no longer surrounds the clear deutoplasmic
region as a uniform periphei*al layer. It has not only
increased considerably in amount as compared with that of
the full-grown egg, and at the expense apparently of the moreperipheral coarser portion of the deutoplasmic zone, but it
has undergone polar segregation, with the result that it now
occupies rather more than one hemisphere of the egg as a
dense finely granular mass, with vacuoles of varying sizesparsely scattered through it (figs. 2 and 3, /.z.). It
accordingly defines one of the ovular poles. The opposite
pole is just as markedly chai’acterised by the presence immediately below it of a more or less rounded clear mass,
eccentrically situated, and composed of an extremely finecytoplasmic reticulum with wide fluid-filled meshes. It is
completely surrounded by formative cytoplasm (though over
the polar region the enclosing layer is so extremely thin that
it here almost reaches the sui'face), and its cytoplasmic framework is perfectly continuous with the same, the line of
junction of the two being abrupt and well defined. So
delicate, however, is this framewoi’k that it breaks down
more or less completely under the action of fixatives of such
excellence even as the fluids of Flemming and Hermann,
and thus in sections usually all that represent it ai’e a few
irregular cytoplasmic strands crossing a large, sharply defined
clear space (figs. 2 and 3, d.z.). The mass in question has
thus all the characters of the deutoplasmic zone of the fullgrown ovum, and it must undoubtedly be held to represent
the central portion of that which has not been utilised in the
upbuilding of the formative cytoplasm, and which has been
forced to take up an excentric position immediately below
the polar region of one hemisphere, owing to the increase of
the formative cytoplasm and its segregation in the other
The ripe ovum of Dasyurus thus possesses a polarity which
in its way is equally as striking as that of the Monotreme
egg. Towards the one pole the main mass of the ovum is
composed of dense, slightly vacuolated formative cytoplasm,
in which the polar spindle is situated peripherally, but nearer
the equator than the formative pole. Toward the opposite
pole and practically reaching the surface is a rounded mass
of greatly vacuolated deutoplasmic cytoplasm. Eoughly,
the formative cytoplasm constitutes about two-thirds of the
bulk of the ripe egg, the deutoplasmic the remaining third.
Such being the structure of the ripe ovarian egg, if we
classify it at all, we must place it, it seems to me, with eggs
of the telolecithal type. My view of the significance of this
marked polar differentiation of the constituent materials of
the ripe ovum of Dasyurus I shall presently indicate. Meantime I would lay special emphasis on the fact that the
eccentric mass of deutoplasmic cytoplasm represents matei’ial,
surplus deutoplasmic material which has not been utilised in
the upbuilding of the formative cytoplasm.
The fact of the occurrence in the Eutherian ovum of a
polar differentiation of its constituent materials is now
definitely established, thanks especially to the valuable
researches of Prof. 0. Van der Stricht and his pupils  -  H.
Lams and the late J. Doorme. In this connection I wish to
refer here in some detail to the extremely interesting obser
vations of Van der Stricht [’03, ’05] on the structure and
polarity of tlie ovum of the bat (Vesperugo noctula), since
these observations are in essential agreement with my own
on the ovum of Dasyurus, and enable me to affirm that the
polar differentiation herein recorded for the first time for the
Marsupial ovum is attained as the result of vitellogenetic
processes, which essentially correspond with those of the ovum
of the bat. Van der Stricht, as is well known, has made a
special study of the process of vitellogenesis in the Eutherian
ovum, and is, indeed, at the present time the foremost
authority on this particular subject, so that his views are
worthy of all respect.
Study of the oocyte of Vesperugo during the period of
growth shows, according to Van der Stricht, that “ a un
moment donne du developpement du jeune oeuf, les boyaux
et amas vitellogenes [derived, according to him, from ‘ une
couche vitellogene, mitochoudriale,’ present in the young
oocyte in the first stage of growth] disparaissent au profit du
vitellus, dont la structure pseudo - alveolaire s’accentue
graduellement.” The full-grown oocyte at the stage just
prior to the appearance of the first polar spindle is characterised by the presence of this “ pseudo-alveolar structui’e ”
throughout the extent of its cytoplasmic body. The alveoli
or vacuoles are of variable size, are filled by a clear liquid,
and “ correspondent incontestablement au deutoplasma de
I’oeuf. A ce stade du developpement de I’oocyte, ce vitellus
nutritif, auquel s’ajoutent bientot des granulations graisseuses,
est repandu uniformement dans toutes les profondeurs du
cytoplasme. Nulle part on ne constate une zone deutoplasmique distincte d’une zone de vitellus plastique.” In
Dasyurus the stage in vitellogenesis which almost exactly
coiTesponds with that of the full-grown oocyte of Vesperugo
just described is seen in oocytes not quite full-gi*own. In
fig. 4 is shown an oocyte of Dasyurus (•26 x ‘20 inm. in
diameter), in which the same pseudo-alveolar sti’ucture as
described by Van der Stricht for the Vesperugo oocyte is
perfectly distinct. Here, however, fatty particles are not
apparent, and the peripheral portion of the cytoplasm tends
to be free from vacuoles. In Dasyurus the formation of these
deutoplasmic vacuoles begins in oocytes about ’2 mm. or less
in diameter. This characteristic pseudo-alveolar ” stage is
followed in both Vesperugo and Dasyurus by one in which
there is recognisable in the cytoplasmic body of the ovum a
differentiation into a dense peripheral zone and a central
vacuolated area. In Vesperugo this stage is attained about
the time of appearance of the first polar spindle, whilst in
Dasyurus it is attained somewhat earlier, always prior to the
formation of the latter. So close is the agreement between
the two forms that Van der Stricht’s desci’iption of the bat’s
egg at the time of appearance of the first polar spindle might
equally well be applied to the full-grown ovum of Dasyurus.
He writes [’03, p. 43] : “Vers I’epoque de I’apparition du
premier fuseau de maturation, le vitellus prend un autre
aspect. La partie centrale deutoplasmique conserve une
structure pseudo-alveolaire, mais dans le voisinage immediat
du premier fuseau et daus toute I’etendue de la couche
peripherique du protoplasme, apparait une mince zone de
vitellus compact et dense, plus ou moins homogene ou les
vesicules claii’es font defaut. . . . A ce moment, on
distingue daus I’odcyte de V. noctula une zone centi’ale
tres etendue, riche en deutoplasme et une zone corticale
tres mince, riche en vitellus plastique.” This centrolecithal phase, as we may term it, is followed in Vesperugo
during fertilisation and the separation of the second polar
body by a telolecithal phase characterised by a distinct
polarity. “ La zone de vitellus plastique s’epaissit encore,
mais surtout a un pole de I’oeuf, a celui oppose au pole ou se
detachent les deux globules polaires. Ce pole, ou s’accumule
graduellement le vitellus formateur, merite le nom de pole
animal. II est oppose au pole d’expulsion des globules
polaires, vers lequel est refoule le deutoplasme, et qui se
comporte desormais comme le pole vegetatif. Pendant que
les deux pronucleus male et femelle se ferment, le vitellus
plastique augmente graduellement en abundance au pole
animal, tandis qu’il diminue an p61e vegefcafcif, et le deutoplasme, parseme d’un plus grand nombre de boules graiss6uses,
constitue une masse spberique excenfcriqne, voisine des deux
globules polaires” (Van der Stricbfc, ’03, pp. 44 - 45). It is
evident, then, that the fertilised ovum of Yesperugo exhibits a
polarity comparable with that of the ripe ovarian ovum of
Dasyurus, and that the vitellogenetic processes in the ova of
these two widely separated forms proceed along lines almost
identical, at all events so far as their broad outlines are concerned. In both we find during growth a progressive
vacuolisation of the egg-cytoplasm consequent on tlie elaboration of a deutoplasmic fiuid. In both, the '^pseudo-alveolar ”
condition so engendei-ed is followed by one in which there
is recognisable a differentiation into a peripheral "formative ” zone rich in deutoplasmic granules, and a central
" deutoplasmic ” zone rich in fluid yolk, and finally in both
there occurs a segregation of the granular "formative” and
fiuid yolk-constituents to opposite I’egions of the egg, with
resulting attainment of a definite polarity. In view of the
close general agreement in the vitellogenetic processes, and in
the constitution of the ova in Vesperugo and Dasyurus, it
might be expected that the poles would accurately correspond,
but such is not the case if Van der Stricht’s determination of
the poles in the ovum of Vesperugo is correct. In the latter,
according to Van der Stricht, the deutoplasm is located at
that pole from which the polar bodies are given off; at the
opposite pole the "plastic” vitellus accumulates, and close to
it the two pronuclei unite and the first cleavage spindle is
formed. Accordingly Van der Stricht concludes that "le
premier pole correspond au pole vegetatif, le second au p61e
animal des oeufs a deutoplasme polaire (0. Hertwig).” In
Dasyurus, on the other hand, I am perfectly convinced (and
adequate reason for my conviction will be forthcoming in the
course of my description of the processes of cleavage and
germ-layer formation) that the pole of the ripe ovum in
relation to the mass of deutoplasmic cytoplasm is not the
vegetative pole, but represents morphologically the upper or
von. 56, PAUT 1.  -  Nicw series. 2
animal pole of the egg, the opposite pole in relation to which
the formative cytoplasm is situated being' the lower or
vegetative. The deutoplasmic cytoplasm thus lies in the
upper hemisphere, whilst the formative cytoplasm occupies
the lower. If Van der Stricht’s determination of the poles of
the ovum of Vesperugo be accepted, then we must conclude
that the poles of the Dasyurus ovum are exactly reversed as
compared with those of the bat^s egg. In this connection it
may be recalled that Lams and Doorme [’07] have demonstrated the occurrence in Cavia of an actual reversal of the
original polarity of the ovum, pi’ior to the beginning of
cleavage. These facts may well give us pause before we
proceed to attach other than a purely secondary significance
to the exact location of the formative and deutoplasmic constituents in the Metatherian and Eutherian ovum. But
besides this apparent difference in the location of the deutoplasmic constituents of the ova of Dasyurus and Vesperugo,
there exists yet another which concerns the fate of these constituents in the respective eggs. In Vesperugo, Van der
Stricht shows that the “ deutoplasm ” remains an integral
part of the egg, and I’etains its polar distribution in the
blastomeres up to at least the 4-celled stage. ^ In Dasyurus,
on the other hand, the fate of the deutoplasmic mass is a very
different, and, indeed, a very remarkable one. It does not
remain an integral part of the segmenting egg as in Vesperugo,
but prior to the completion of the first cleavage furrow it
becomes bodily separated off, apparently by a process of
abstriction, from the formative cytoplasm as a clear rounded
mass which takes no further direct part in the developmental
processes. As soon as its elimination is effected, the remainder
of the cytoplasmic body of the ovum, formed of the formative
cytoplasm alone, divides into the first two equal-sized blastomeres, the first cleavage plaue being coincident with the polar
diameter and at right angles to the plane of separation of the
deutoplasmic mass, or “yolk-body ” as we may term it (PI. 2,
figs. 14-16, 19, ij.b.), so that it is this formative zone of the
* Vide, however, “Addendum” (p. 121).
ovutn which is filone concerned in tlie production of the
embryo and its foetal membranes.
We have but to recall the conclusion already reached that
the clear vacuolated zone at the upper pole of the ripe ovum
of Dasyurus consists of surplus material, mainly in the form of
fluid of deutoplasmic nature which has not been utilised in
the upbuilding of the formative cytoplasm, and the signiflcance of this remarkable and, so far as the Mammalian ovum
is concerned, absolutely unique occurrence becomes at once
manifest.’^ We have to do here with an actual elimination of
surplus deutoplasmic material by the Marsupial ovum  -  a phenomenon only paralleled elsewhere, so far a,s I am awai’e, and
even then but distantly, by the curious temporary separation
of the so-called yolk-lobe which occurs during the cleavage
of the yolk-laden eggs of certain Molluscs (Nassa, Ilyanassa,
Modiolaria, Aplysia, Dentalium) and Annelids (Myzostoma,
Chjetopterus) . In these forms cleavage of the ovum into the
first two blastomeres is accompanied by the sepai’ation of a
portion of the ovular substance in the form of a non-nucleated
mass or so-called yolk-lobe. This latter, which has been shown
to be connected Avith the formation of determined organanlagen, reunites Avith one of the tAvo blastomeres, and then
the same process of abstriction and reunion recurs at the
second cleavage.^ We have here evidently a purely adaptive
phenomenon, the object of which no doubt is to permit of the
total cleavage of the yolk-laden ovum on Avhat are presumably
the old ancestral lines, and I believe a comparable explanation
Avill be found applicable to the elimination of surplus yolkmaterial by the Marsupial ovum.
As regards the significance of the occurrence of the deutoplasmic zone in the ovum of Dasyurus, holding the views that
I do as to the phylogeny of the Marsupialia (viz. that the
Metatheria and Eutheria are the divergent branches of a
* Yide “Addendum” (p. 121), in Avliich reference is made to the discovery by Prof. Van der Stricht of the elimination of deutoplasm in
the ovum of Yesperugo.
® Of. Korschelt u. Heider, ‘Lehrbuch d. vergl. Entwicklungsgeschichte,’ Lief. 3, p. 107, 1909.
common stock, itself of Prototlierian derivation), and bearing
in mind the occurrence of an undoubted repi’esentative of the
shell round the Marsupial ovum, I venture to see in the fluidmaterial of the deutoplasmic zone the partial and vestigial
equivalent of the yolk-mass of the monotreme egg. In other
words, I would regard the deutoplasmic fluid as the product
of an abortive attempt at the formation of such a solid yolkmass. The objection will no doubt be forthcoming that this
interpretation cannot possibly be correct since the supposed
equivalent of the yolk-mass in the Dasyure ovum is located,
on my own showing, at the wrong pole  -  at the upper instead
of at the lower. But its precise location does not seem to me
to be a matter to which we need attach any great importance,
since it has doubtless been adaptively determined in correlation with the special character of the cleavage process.
The belief that the minute yolk-poor ovum of the Eutheria
is no pure primarily holoblastic one, but that it has only
secondarily arrived at the total type of cleavage as the result
of the all but complete loss of the yolk ancestrally present in
it, consequent on the substitution of the intra-uterine mode of
development for the old oviparous habit, is now widely held
amongst Mammalian embryologists. Hubi’echt, however, is
an exception, wedded as he is to a belief in the direct derivation of the Eutheria from Protetrapodous ancestors with yolkpoor, holoblastic eggs. Whether the interpretation I have put
forward, viz. that the non-formative or deutoplasmic zone of
the Dasyure ovum is the reduced and partial equivalent of the
yolk-mass of the Monotreme egg, be accepted or not, I venture
to think that my discovery of an actual elimination of deutoplasmic material by the Marsupial ovum affords a striking
confirmation of the truth of the prevailing conception as to
the phylogeny of the Eutherian ovum, and I further venture
to think that the facts I have brought forward in the preceding
])ages justify us in' regarding the ripe ovarian ovum of
Dasyurus as being potentially of the yolk-laden, telolecithal
type, and the uterine ovum, by bodily casting out the superfluous part of its deutoplasm, as becoming at the same time
secondarily lioniolecithal and secondarily lioloblastic. Ihe
Marsupial ovum presents itself to my mind as tlie victim of
tendencies conditioned by its ancestry, and in particular it
appears as if its inherited tendency to elaborate yolk had not
yet been brought into accurate correlation with the other
changes (reduction in size, intra-uterine development), which
it has undergone in the course of phylogeny. As the consequence it manufactures more yolk than it can utilise, and so
finds itself under the necessity of getting rid of the surplus.
Whether or not a comparable elimination of deutoplasmic
material occurs in the ova of other Marsupials, future investigation must decide. I should be quite prepared to find variation
in this regard, correlated perhaps with the size of the egg.
In the case of Phascolarctus, Caldwell gives the diameter of
the ovum as '17 mm., and his figure of a (horizontal ?) section
of the uterine ovum (here produced as text-fig. 1, p. 27) shows
a differentiation of the cytoplasmic body of that into vacuolated and granular zones quite comparable with that of the
Dasyure ovum. From the few measurements of ova of other
marsupials that I have been able to make, it would appear
that the ovum of Trichosurus approximates in size to that of
Dasyurus, whilst that of Perameles and pi-obably also that
of Macropus are smaller. From Selenka’s figure I have
calculated that the ovum of Didelphys measures about 'IS x
•12 mm. in diameter. In the smaller ova it is quite likely
that yolk-formation may not proceed so far as in the relatively
large ovum of Dasyurus.
2. Maturation and Ovulation.
The details of the maturation process have not been fully
worked out, owing to lack of material. As in the Eutheria
(Sobotta, Van der Stricht, Lams and Doorme, and others),
the first polar body is separated off in the ovary, the second
apparently in the upper part of the- Fallopian tube where
entrance of the sperm takes place. The first polar figure
(late anaphase observed, fig. 5) lies in the formative cyto
plasm, close below and afc right angles to the zona. Its exact
site is subject to some slight variation, and is best described
as adjacent to the equatorial region of the egg, sometimes
nearer the lower pole, more usually, perhaps, nearer the
upper. Centrosomes and polar radiations were not observed.
The heterotypical chromosomes (gemini) have the form of
somewhat irregular, more or less angular granules. I have
not been able to determine their number. The figure is
barrel-shaped, and almost as broad as long, measuring
*015 X ‘OIS mm. The first polar body (fig. 6, is small
relatively to the size of the egg, its diameters varying round
•03 X '01 mm., and its shape is that of a flattened bi-convex
disc. In uterine eggs thex'e is some evidence pointing to the
probability of its having undergone division.
The second polar spindle (figs. 3 and 7) lies immediately
subjacent to the first polar body in the fully ripe ovarian
ovum. It is shorter than the first, measuring '013 mm., and
xnuch narrower. The second polar body measures about
’015 X '01 mm. in diameter, and is thus smaller than the first.
I have only seen the second polar body in uterine ova, and
therefore can only presume that it is sepai’ated off in the
upper part of the Fallopian tube, subsequently to the penetration of the sperm, as in Eutheria.
Ovulation takes place irrespective of whether copulation has
occurred or not, and it is a fact Avorbhy of record that, even
if the ova be not fertilised, the pouch and mammary glands
undergo the same series of growth changes as are characteristic of, at all events, the earlier stages of normal
The follicular cells of the discus proligerus investing the
ovum are already in the ripe follicle in a state of disruption,
and I believe they separate completely from the ovum at the
moment of dehiscence, so that, except for the zona, the ova
are quite naked Avhen they enter the tube. I have no evidence
of the existence outside the zona of a layer of proalbumen
such as Caldwell describes round the ovum of Phascolarctus.
Apparently the ova are shed almost simultaneously, and they
must pass with considerable rapidity down the tubes to the
uteri wliere cleavage begins, for I have only once found a
tubal ovum, and that one had evidently been retarded for
some reason, and was polyspermic.
3. The Secondary Egg-membranes: Albumen and
Shell- membrane.
During the passage of the ovum down the tube it is
fertilised, and becomes enclosed externally to the zona by
two secondary layers formed as secretions by the cells of the
oviducal lining. First of all, the ovum becomes surrounded by
a transparent to semi-transparent laminated layer of albumen,
•015 to ■02’2 mm. in thickness, composed of numerous very
delicate concentric lamellae, and having, normally, numbers
of sperms imbedded in it (figs. 8-11, alb., sp.). Then outside
the albumen layer there is laid down a definite, but at first
very thin, double-contoured membrane (figs. 8 and 10, s.m.),
which, following Caldwell, I have no hesitation in homologising with the shell-membrane of the Monotreme egg.
Caldwell in 1887 described and figured a definite membrane
enclosing the uterine ovum of Phascolarctus, externally to,
and quite distinct from the albumen, which he interpreted
as the representative of the shell-membrane of the Monotremata, but owing apparently to the fact that Selenka
altogether failed to recognise its true nature in Didelphys,
since he regarded it as a derivative of the follicular epithelium, and termed it the “ granulosa-membran,” this highly
significant discovery of Caldwell has been largely ignox’ed.
Such a membrane is constantly present and easily recognisable
in all the Marsupials (Dasyurus, Perameles, Trichosurus,
Macropus, Petrogale, Phascologale, Acrobates, Phascolarctus,
Bettongia), of which I have had the opportunity of studying
early developmental stages. It is laid down in the Fallopian
tube, is perfectly distinct from the albumen, and increases in
thickness in the uterus, and if it has not the significance
which Caldwell has suggested, then I must leave it to those
who decline to accept Caldwell’s interpretation to put forward
an alternative one, since I am unable to do so.
The shell-membrane of Dasyurus (PI. 1, figs. 8-11; PI. 2,
figs. 17, 18, s.m.) is a ti*ansparent, perfectly homogeneous
layer, highly refractive in character and of a faint yellowish
tint. When fully formed it possesses firm, resistant properties,
recalling those of chitin, and is doubtless composed of a
keratin base. It is distinguishable at once from the albumen
by its optical characters and staining reactions, so that there
is not the slightest justification for the supposition that it
may represent simply the specially differentiated outermost
portion of that layer. In ova which have just passed into the
uterus (fig. 10) the shell-membrane is extremely delicate, its
thickness being only about '0016 mm., but even before cleavage
begins it has increased to ”002 mm. (fig. 12) ; in the 2-celled
stage (fig. 18) it has reached ‘005 mm., in the 4-celled stage
(fig. 22) ’0072 mm., whilst in the 16-celled stage (figs. 24-26)
it has practically attained its maximum thickness, viz., ‘0075'008 mm. Caldwell’s measurements in the case of Phascolarctus agree closely with the above (shell of unsegmented
ovum from the uterus, '0015 mm. thick, that of the '3 mm.
‘‘ovum,” '01 mm.). Its presence renders the thorough
penetration of ova and early blastocysts with pai'affin a
capricious and frequently troublesome operation, and its
resistant shell-like nature becomes only too obvious in the
process of section-cutting, since it cracks with the utmost
readiness (cf. PI. 3, figs. 32, 37).
The occurrence of a shell-membrane round the Marsupial
ovum is a feature of considerable phyletic significance, as I
need hardly point out. It shows us that the ancestors of the
Metatheria must have been oviparous, or must themselves
have come from an oviparous stock, which there is no valid
reason for supposing was other than Prototherian in its
characters. It also renders untenable the views of Hubrecht to
the effect that the Metatheria are the descendants of Eutheria,
whilst the Eutheria themselves have been directly derived
from some presumed viviparous group of hypothetical Prote
trapods, unless we are to suppose that the Metatheria are
even novv on the way to acquire secondarily the oviparous
habit, much in the same way as the Mouotremes, according
to Hubi-echt, have long since succeeded in doing.
q’be occurrence of a shell-membrane round the Marsupial
ovum has also aai important ontogenetic significance in relation to the mode of formation of the blastocyst, as I shall
endeavour presently to show.
4. The Uterine Ovum.
The unsegmented ovum from the uterus (figs. 8-13)
consists of the following parts :
(1) The shell-membrane externally, ’OOlfi -  ’002 mm. in
(2) The laminated layer of albumen, '015 -  "022 mm. or
more in thickness.
(3) The zona, about -0016 mm. in thickness.
(4) The perivitelline space, between the zona and the
ovum, occupied by a clear fluid which coagulates under the
action of certain fixatives, e. g. Hermann’s fluid (fig. ll,p.s.),
and which has diffused in from the uterus. The minute
polar bodies lie in this space, usually nearer the upper pole
than the lower.
(5) The ovum proper.
The entire egg is spherical in form, and varies in diameter
in the fresh state from about ’3 mm. to ’36 mm. (average
about ‘32 mm.).
The ovum itself is ovoidal, its polar diameter always slightly
exceeding the equatorial. Its average diametrical measurements in the fresh state run about ’25 x '24 mm., though I
have records of ova measuring as much as ‘3 x ‘29 mm., and
I find that there is an undoubted slight variation iu the size
of the ova of even one and the same batch, as well as iu those
from different females.
The uterine ovum exhibits the same marked polarity as
cliaracterises the ripe ovarian ovum (the upper pole being
marked by the vacuolated deutoplasmic zone (figs. 8-11, d.z.),
and so far as its cytoplasmic body is concerned it shows no
essential difference from that.
Examined fresh in normal salt solution, the formative cytoplasm forming the bulk of the ovum appears dense, finely
granular, and of a very faint lightish-brown tint, its opacity
being such that the two pronuclei situated in its central region '
can just be made out. In section, this central region is distinguishable from the peripheral zone by its uniform, more
finely granular character and by the absence of the fluid-filled
vacuolar spaces which are generally present in the latter
figs. 10 and 12). The deutoplasmic zone at the upper pole,
which is only partially visible in the entire egg owing to the way
in which it is enclosed by the formative cytoplasm (figs. 8, 9,
d.z.), presents a characteristically clear or semi-transparent
vacuolated appearance in the fresh state, but may have einbedded in it a small dense mass (fig. 8, cf. also figs. 11 and 14),
evidently formed by the transformation of a portion of its fluid
constitutent into the solid state, and so to be regai’ded as compai'able with a bit of formative cytoplasm.
In most of the unsegmented uterine ova at my disposal the
male and female pronuclei have attained approximately tlie
same size and lie in proximity in the central more homogeneous region of the formative cytoplasm (figs. 10-12). The
transformation of the sperm-head into the male pronucleus
probably takes place during the passage of the ovum down
the tube, and was not observed, and I am as yet uncertain
whether the pronuclei unite to form a single cleavage nucleus
or give origin directly to the chi’omosomes of the first cleavage
Caldwell figures (’87, PI. 30, fig. 5) a section through the
uterine ovum of Phascolarctus which I reproduce here as
Text-fig. l,in order to facilitate comparison with my figs. 11 and
12, with which it shows an essential agreement, apart from
the pi’esence of follicular cells in the albumen which I have
never observed in Dasyurus, and making allowance for the
difference in sectional plane. The figure is stated to represent
“the seventeenth section of a vertical longitudinal seines of
thirty-five sections through the segmenting ovum, containing
two nuclei, taken from the uterus and measuring T7 mm. in
diameter.” Caldwell has, I think, fallen into several errors
in his interpretation of the structural features seen in this
Text-pig. 1.
Section of uterine ovnni of Phascolarctus cinereus.
(After Caldwell.)
figuie. In the first place, the sectional plane appears to me
not to be vertical as in my own figs. 11 and 12, but horizontal,
and to have passed through the lower portion of the deutoplasmic zone, shown in the figure as a central markedly
vacuolated area. Then there is no evidence to be derived
from the figure in support of the description of the ovum as
segmenting. The part inside the zona {vm.) labelled and
described as protoplasm with finest jolk-granules,” I would
interpret simply as coagulum in the perivitelline space, whilst
the so-called “ segmentation nuclei ” Ua) situated in it are
probably the polar bodies or their derivatives. The part
labelled y«, and designated “ white yolk,” I would regard as
the ovum itself. It exhibits an obvious differentiation into a
central vacuolated area and a peripheral, dense, granular
zone with scattered vacuoles, and I thiuk there can be little
doubt but that the former corresponds to the deutoplasmic
zone of the Dasyure ovum, the latter to the formative zone.
It is these errors of intei-pretatiou apparently which misled
Caldwell into making the statement, now widely quoted in
the text-books, that cleavage in Phascolarctus is of the
meroblastic type.
Chapter III.  -  Cleavage and Formation op the Blastocyst.
1. Cleavage.
Cleavage begins in the uterus as in Didelphys, Phascolarctus, and no doubt Marsupials in general. The first externally visible step towards it consists, as already described,
in the elimination by abstriction of the deutoplasmic zone at
the upper pole. The yolk-body so formed appears as a
definitely limited, clear, rounded mass which lies in contact
with the slightly concave upper surface of the formative
remainder of the ovum. It is quite colourless and transparent
except for the frequent occurrence in it of a small, more or
less irregular opaque mass, representing probably a condensation product of its fluid material (cf. PI. , figs. 8, 14, y.h.).
Consisting as it does of a very delicate cytoplasmic reticulum
with fluid-filled meshes it is extremely fragile, and is seen to
advantage only in fresh material (figs. 14 and 19, y.h.). It
takes no direct part in the later developmental processes>
though during the formation of the blastocyst it becomes
enclosed in the blastocyst cavity and finally undergoes disintegration therein, its substance becoming added to the fluid
which Alls the same, so that it may be said, in this indirect
way, to fulfil, after all, its original nutritional destiny. Separa
tion of the yolk-body is rapidly followed by the completion
of the division of the formative remainder of the ovum into
the first two blastomeres, the plane of division being coincident with the polar diameter or egg-axis and at right
angles to the plane of separation of the yolk-body (PL 2, fig.
14). I obtained relatively little material between the stage
of the unsegmented ovum with two equal-sized pronuclei seen
in fig. 12 and the 2-celled stage (fig. 14), both of which
are well represented in my material, so that it would appear
that the separation of the yolk-body and the division of the
formative remainder of the ovum are effected with considerable
rapidity. Fig. 13 shows, however, a section of an unsegmented ovum in which the chromosomes of the metaphase
of the first cleavage figure are visible in the central region
of the formative cytoplasm, but situated, it is worthy of note,
rearer the future upper pole than the lower pole. The deutoplasmic zone {d.z.) still forms an integral part of the egg, and
there is no sign of commencing abstriction. I have also
sections of ova in a still more advanced stage of the first
cleavage, in which the daughter-nuclei have but recently been
constituted and are still quite minute, and the cleavage furrow
is well marked on the surface of the egg. In these ova the
yolk-body is already separated, so that we may conclude with
a fair degree of certainty that its elimination about coincides
with the first appearance of the cleavage furrow.
Figs. 14-16 show the 2-celled stage, respectively in side,
lower polar, and end views. The blastomeres are of approximately equal size and otherwise quite similar. Selenka also
found the same to be the case in Didelphys, though in the
single specimen of the 2-celled stage he had for examination
(Taf. xvii, fig. 3) the blastomeres are displaced and somewhat
shrunken. Each blastomere has much the shape of a hemisphere from which a wedge-shaped segment has been sliced
off, a form readily accounted for when we take account of the
effect of the elimination of the deutoplasmic zone. After
that event, the formative remainder of the ovum has the
form of a sphere from which a somewhat bi-convex lens
shaped piece lias been gouged out at the upper pole.
Consequently, when it divides along its polar diameter, the
resulting blastomeres will have the form of hemispheres with
obliquely truncated upper surfaces or ends, which will be proportionately thicker than the lower ends. In correlation
therewith we find the nucleus of each blastomere situated
slightly excentrically, rather nearer the upper than the lower
pole (fig. 18). The rounded yolk-body lies partly enclosed
betAveen the upper truncated surfaces of the blastomeres.
Two-celled eggs ai’e shown in vertical section in figs. 17
and 18. The cytoplasm of the blastomeres exhibits a wellmarked differentiation into two zones corresponding to that
ah’eady seen in the formative cytoplasm of the uusegmented
egg, only much more accentuated, viz. a dense, fine-grained
perinuclear zone, and a less dense, more vacuolated peripheral
zone, in which there is present a coarse, irregular network of
deeply staining strands, recalling the frameAvox’k of mitochondrial origin described by Van der Stricht (’04, ’05) in
the human ovum and that of Vesperugo. We have here in
this differentiation of the cytoplasm, evidence of the occurrence of an intense metabolic activity Avhich has resulted in a
marked increase in the amount of deutoplasmic material
present in the blastomeres as compared Avith that found in
the ovarian egg or even in the unsegmented uterine egg.
The blastomeres consequently present a someAvhat dense
opaque appearance Avhen examined in the fresh state, their
nuclei being partially obscured from view. Amongst the
Eutheria, various observers (Sobotta, Van der Stricht, Lams
and Doorme) have described a similar inci’ease in the deutoplasmic contents of the egg after its passage into the
Fallopian tube or uterus.
The second cleavage plane is also vertical and at right
angles to the first. The resulting four equal-sized blastomeres
vieAved from the side (PI. 2, fig. 19) are seen to be ovalish in
outline, their loAver ends being slightly narroAver and more
pointed than their upper ends, Avhich diverge someAvhat to
enclose the lower part of the yolk-body. Seen from one of
tli 0 poles, in optical section (figs. 20, 21), they appeal triangular 'with rounded corners and centrally directed apices.
The space occupying the polar diameter, which they enclose
is the cleavage cavity. The blastomeres are now somewhat
less opaque than those of the 2-celled stage, so that their
nuclei, excentrically situated nearer their upper ends and
enclosed in the central granular zone of the cytoplasm, can
now be fairly distinctly made out in the fresh egg.
The arrangement of the blastomeres at this stage is
exceedingly characteristic, and is identical with that of the
blastomeres in the corresponding stage of Amphioxus or the
frog, but is quite different from that normal for the 4-celled
stage of the Eutheria. They lie disposed radially or meridionally around the polar diameter, occupied by the cleavage
cavity, their thicker upper ends partially surrounding the
yolk-body. Selenka figures a precisely similar arrangement
in his 4-celled stage of Didelphys, so that we may conclude
it holds good for the Marsupials in general.
Whilst, then, in Marsupials the first two cleavage planes are
vertical or meridional, and at right angles to each other, and
the first four blastomeres are arranged radially around the
polar diameter (radial type of cleavage), in the Eutheria
such is never the case, at all events normally, so far as is
known. In the Eutheria the first four blastomeres form, or
tend to form, a definite cross-shaped group, as the result
apparently of the independent division of thefii’st two blastomeres in two different planes at right angles to each other,
the division planes being meridional in the one, equatorial
in the other. ^ This pronounced diffei’ence in the spatial
relations of the fij’st four blastomeres in the Metatheria and
Eutheria is a feature of the very greatest interest and importance, since it is correlated with and in part conditions
the marked dissimilarity which we meet with in the later
developmental occurrences in the two groups, in particular
in the mode of formation of the blastocyst in the two.
* Compare in this connection Assheton’s remarks (’09, pp. 232-233),
which have api>eared since this chapter was written.
Moreover, so far as the Eutheria are concerned, it affords us,
I believe, a striking and hitherto unrecognised example of a
phenomenon to which Lillie (’99) has directed attention, viz.
adaptation in cleavage.
Fig. 22 shows a horizontal section through the 4-celled
stage, and fig. 23 a vertical section of the same. The blastomeres in their cytoplasmic characters essentially resemble
those oE the 2-celled stage, but the peripheral deutoplasmic
network is here more strongly developed, and it is especially
worthy of note that it is more marked towards the lower
poles of the blastomeres (fig. 23), as also appears to be the
case in the 2-celled stage. The shell-membi’ane measures in
thickness ’0072 mm.
The next succeeding (third) cleavages are again meridional, each of the four blastomeres becoming subdivided
vertically into two, not necessarily synchronously. Fig. 53.
PI. 6, shows a side view, and fig. 54 a view from the lower
pole of a 6-celled egg, two of the blastomeres of the 4-celled
stage having divided before the other two. The blastomeres have moved apart, and now form an' open ring
approximately equatorial in position, and surrounding the
central cleavage space, the upper opening of which is
occupied by the yolk-body. I have failed to obtain a
perfectly nonnal 8-celled stage, nevertheless the evidence
clearly shows that the first three cleavage generations in
Dasyurus are meridional and equal, and that the resulting
eight equal-sized blastomeres form an equatorial ring in
contact with the inner surface of the sphere formed by the
zona and shell-membrane.
Whilst, then, the first three cleavage generations are
meridional and equal, the succeeding divisions (fourth cleavage
generation), on the contrary, are equatorial and unequal, each
of the eight blastomeres becoming divided into a smaller,
more transparent upper cell, with relatively little deutoplasm,
and a larger, more opaque lower cell with more abundant
deutoplasmic contents. In this way there is formed an
exceedingly characteristic 1 6-celled stage, consisting of two
superimposed rings, each of eight cells. The upper ring of^
smaller and clearer cells partially encloses the yolk body, and
is situated entirely in the upper hemisphere of the sphere
formed by. the egg-envelopes. The lower ring of larger, more
opaque cells lies approximately in the equatorial region of the
said sphere. This 16-cell ed stage is.figm’ed in fig. 55, PI. 6,
as seen from the side, and in fi'g. 56 as seen from the upper ^
pole, both figures being taken from a spirit egg ’37 mm. in
diameter. The marked' differences in the cells of the two
rings are well brought out in the micro-photographs reproduced
as figs. 24, 25, and 26, PI. â–  2. Pigs. 24 and 25 represent
horizontal sections of an egg '38 mm. in diameter, the former
showing the eight cells of the lower ring, and the latter the
eight cells of the upper ring. Pig. 26 shows a vertical section
through an egg also of a diameter of "38 mm., but with
seventeen cells, one of the original eight cells of the upper ring
having divided and one beinginprocessof division. Thesection
passes through the yolk-body [y.h.), which is seen as a faintly
outlined structure lying in contact with the zona between the
two cells of the upper ring (/.c.).
The shell-membrane in eggs of this 16-celled stage has
attained a thickness of -0075 mm., and the albumen layer has
been almost completely absorbed, so that the zona now lies
practically in apposition with the shell-membrane, the two
together forming a firm resistant sphere, to the inner surface
of which the blastomeres are closely applied. The separation
between the zona and shell-membrane seen in the figures is
largely, if not wholly, artificial. ,
The average^ measui’ements of the cells of the two rings ip
the '38 mm. egg, :^ured in figs. 24 and 25, are as follows :
Ppper ring cells. Lower ring cells.'
Diameter . : -06 x -058 mm. . -OOx-Ofidinm.^,'
Vertical height , -OSSimm. . -llS mm.
Nucleus . j . ‘0165 mm. ■ ■ • . •02 mm.
These measurements demonstrate at a glance the distinct
difference in. size which exists between the cells of the two
rings, whilst the cytoplasmic differences between them. ai-e
VOL. 56, PART 1.  -  NEW SERIES. ""S
equally evident from an inspection of the micro-photographs,
figs. 24-26. In the larger cells of the lower ring (fig. 24,
tr.ect.) the nucleus (rich in chromatin and nucleolated)
is surrounded by a perinuclear zone of clearer, coarsely
vacuolar cytoplasm, outside of which is a densely granular
deutoplasmic zone, which extends to within a short distance
of the periphery of the cell-body. In the smaller cells of the
upper ring (fig. 25, /.c.) the cytoplasm is coarsely reticular,
with a tendency to compactness round the nucleus, and its contained deutoplasmic material is spare in amount as compared
with that of the lower cells, being mainly located in a quite
narrow peripheral zone. The upper cells thus appear relatively
clear as compared with the dense, opaque-looking lower cells
(fig. 26).
It becomes evident, then, that we have to do here, in this
fourth cleavage generation, with an unequal qualitative
division of the cytoplasm of the blastomeres of the 8-celled
stage. Just such a division as this we should expect if the
deutoplasmic material were mainly aggregated towards the
lower poles of the dividing cells. The evidence shows that
this is actually the case. In the 2-celled and especially in the
4-celled eggs we have already seen that the deutoplasmic
network is already most strongly developed towards the lower
poles of the blastomeres. This polar concentration of the
deutoplasm reaches its maximum in blastomeres of the 8celled stage, and confers on these an obvious polarity.
Although I failed to obtain normal examples of the latter stage,
I have fortunately been able to observe the characters of the
blastomeres in sections of eggs with twelve, thirteen, and
fourteen cells respectively.
In the 12-celled egg (PI. 6, fig. 57), measuring '38 mm.
in diameter, four of the eight original blastomeres are still
undivided ; the remaining four have undergone division
unequally and qualitatively, one but recently, so that 4 +
(4 X 2) = 12. The undivided blastomeres are large (average
diameter, '11 x '076 mm.) and ovoidal in form, their lower
ends being thicker than their upper, and they exhibit a well
marked polarity. The nucleus lies excentrically in the
upper half of the cell, just above the equator, and is surrounded by a finely granular zone of cytoplasm, outside which
is a thin irregular ring of deutoplasmic material. The cytoplasm of the apical part of the cell is clear and relatively free
from deutoplasm ; that of the lower half, on the other hand,
is so rich in deutoplasm as to appear quite dense and opaque.
The conclusion is therefore justified that the blastomeres of
the 8-celled stage possess a definite polarity, which has beenj
acquired as the result of the progressive concentration of
deutoplasmic material at their vegetative poles during the>
cleavage process. Division, in the equatorial plane, of cells
so constituted must necessarily be unequal and qualitative, so
far at least as the cytoplasm is concerned.
In the 13-celled stage three of the original eight blastomeres are in process of division, and five have already divided
unequally and qualitatively, so that 3 -f- (5 x 2) = 13, and in
the 14-celled stage two of the original blastomeres are in
division and sis have already divided : 2 -t- (6 x 2) = 14.
The significance to be attached to this characteristic unequal
and qualitative division of the blastomeres of the 8-celled stage
to form two superimposed cell-rings, markedly differentiated
from each other, we shall presently consider. Meantime I
may categorically state the conclusions I have reached in
regard thereto. The wall of the blastocyst in Dasyurus is at
its first origin, and for some considerable time thereafter,
unilaminar throughout its entire extent, and I regard the
upper cell-ring of the 16-celled stage as giving origin to<
the formative or embryonal region of the unilaminar wall,
the lower cell-ring ae furnishing the extra-embryonal or nonformative remainder of the same. I shall therefore refer to
the upper cell-ring and its derivatives as formative or
embryonal, and to the lower, cell-ring and its derivatives
as non-formative or extra-embryonal.
The formative or embryonal region furnishes the embryonal
ectoderm and the entire entoderm of the vesicle, and I accordingly conclude that it is the homologue of the embryonal knot
or inner cell-mass of the EutHerian blastocyst. The nonformative or extra-embryonal region directly gives origin to
the outer extra-embryonal layer of the bilaminar blastocyst
wall, i.e. to that layer which in the Sauropsida and Prototheria is ordinarily termed the extra-embryonal ectoderm. I
regard it as such, and as the homologue of the so-called
trophoblast (or as I prefer to term it, the “ trophoblastic
ectoderm” or “ tropho-ectoderm ”) of the Eutherian blastocyst.
A word or two here before concluding this section by way
of summary, as to the condition of the enclosing egg-envelopes.
During the sojourn of the egg in the uterus the albumen is
gradually resorbed, and by about the 16-cell stage it has all
but completely disappeared, thus permitting the zona to come
into direct apposition with the inner surface of the shellmembrane. The shell-membrane itself increases very considerably in thickness during cleavage, and by the 16-celled
stage had practibklly reached its maximum, viz. '0075•008 mm., i.e. it is nearly five times thicker than that of the
ovum which has just entered the uterus. The thickened
shell-membrane by itself is firm and resistant, and it becomes-,
still more so by the application-of the zona to its inner surface,
the two together forming a spherical supporting case round
the segmenting egg, to the inner surface of which the blastomeres become closely applied. .
The existence of such a firm supporting envelope round
the Marsupial egg is, in my view, a feature of very great
ontogenetic significance, and one which must be taken into
account in any comparison of the early developmental occurrences in the Metatheria and Eutheria. As the sequel will
show, the mode of formation of the blastocyst in these two
sub-claisses is fundamentally different, and in my opinion
the explanation of this difference is to be found in the
retention by the Metatheria of a relatively thick resistant;
shell-membrane, and its complete disappearance amongst the
Eutheria. r j
2, Formation of the Blastocyst.
It is characteristic of the Marsupial that the cleavage-cells
proceed directly to form the wall of the blastocyst, without
the iuterventioii of a morula stage, as in’ the Eutheria.
The fil'tli cleavages are meridional, each of the eight cells
of the two rings of the 16-celled stage becoming subdivided
vertically into two, so that there results a 32^celled stage
consisting of two rings, each composed of sixteen cells. As
might be expected, the smaller less- yolk-rich cells of the
upper ring tend to divide more rapidly than the larger yolkladen cells of the lower ring, but the difference in the rate of
division of the two is only slight. I have, for example,
sections of a 17-celled stage (that already referred to, hg. 26)
consisting of nine formative cells (= 6 + [1 x 2] + 1 in
division) and eight non-formative cells, and also of a 31-celled
stage (PI. 6, tig. 59, seen from lower pole; cf. also tig. 60,
showing a side view of another 31-celled egg, both eggs
"375 mm. in diameter), consisting of sixteen formative and
fifteen non-formative cells, of -which one is in process of
division. But I have also preparations of 32-celled eggs with
ail equal number of formative and non-formative cells,
showing that the latter may make up their leeway, the former
resting meantime. On the other hand, the cells of the two
rings may divide more irregularly, as evidenced by a stage
of about forty-two cells, consisting approximately of twentythree formative cells ( = 9 -f- [7 x 2J ) and nineteen nonformative (= 13 -f [3 X 2]). Whatever the rate of division,
the important point is that the division planes are always
radial to the surface, so that all the resulting blastomeres retain a superticial position in contact with the inner
surface of the supporting sphere formed by the zona and
shell-membrane. In apposition with the continuous surface
afforded by that, the blastomeres, continuing to divide,
gradually spread round towards the poles, the descendants of
the upper or formative cell-ring gradually extending towards
the upper pole marked by the yolk-body, whilst those of the
lower or uou-formative cell-riug similarly spread towards the
lower pole. As the blastomeres divide and spread they
become smaller and more flattened, and gradually cohere
together, and so in this way they eventually give origin to a
complete unilaminar layer lining the inner surface of the
sphere formed by the egg-envelopes. It is this unilaminar
layer which constitutes the wall of the blastocyst.
The just completed blastocyst of Dasyurus is a spherical
fluid-hlled vesicle measuidng about '4 mm. in diameter (PI. 3,
flgs. 27-29, PI. 6, figs. 61, 62), and invested externally by
the thin zona and the shell-membrane {â– 0075-0078 mm. in
thickness). The albumen layer has completely disappeared,
and the shell-membrane, zona, and cellular wall are from
without inwards in intimate apposition. The smallest complete vesicles which I have examined measure ‘39 mm. in
diameter (figs. 27, 61), and in one of these I find the cellular
wall consists approximately of about 108 cells. In four other
eggs of the same diameter and from the same female the wall
of the blastocyst is as yet incomplete at the lower pole (fig.
31, l.'p.), and in these, rough counts of the cells yielded the
following respective numbers  -  89, 93, 121, 128. In another
also incomplete blastocyst of the same batch, ’41 mm. in
diameter (fig. 32), the cellular wall consists of about 130 cells.
'I’he largest complete blastocyst in this same batch measured
"49 mm. in diameter, so that we have a range of variation in size
of the just completed blastocyst extending from ‘39 to '49 mm.
The unilaminar wall of the blastocyst consists of a continuous layer of more or less flattened polygonal cells (figs.
27-29, 61, 62) lying in intimate contact with the zona, itself
closely applied to the shell-membrane. Over the lower hemisphere the non-formative cells are on the whole larger and
plumper than the formative cells of the upper hemisphere,
and in sui-face examination they appear somewhat denser
owing to the fact that they possess much more marked perinuclear zones of dense cytoplasm than do the formative cells
(cf. fig. 63, representing a ‘6 mm. vesicle). In sections,
however, this latter difference is much less obvious, indeed.
is hardly, if at all, detectable, so that one has to depend
partly on the relative thickness of the cells, partly, and,
indeed, mainly, on the yolk-body in determining which
hemisphere is which.
The blastocyst cavity is tensely filled by a coagulable fluid
derived from that poured into the uterine lumen through the
secretory activity of the uterine glands. Also situated in the
blastocyst cavity, in contact with the inner surface qf the
wall in the region of the upper pole, is the spherical yolkbody (fig. 29, y.h.). It becomes overgrown and enclosed in
the blastocyst cavity as the result of the completion of the
cellular wall over the upper polar region, much in the same
sort of way as the yolk in the meroblastic egg becomes
enclosed by the peripheral growth of the blastoderm. In the
majority of my sections of early blastocysts the yolk-body
has been dragged away from contact with the formative cells
through the coagulation of the albuminous blastocystic fluid,
and lies more or less remote from the wall enclosed by the
coagulum, except on the side next the upper hemisphere (fig.
31, y.h., c.g.). In two instances, one of which is shown in
fig. 32, 1 find the yolk-body had become so firmly attached to
one of the formative cells that the coag-ulum formed during
fixation failed to detach it, and only succeeded in drawing it
out to a pear-shape.
The yolk-body, it may here be mentioned, persists for a
considerable time in the blastocyst cavity; I have found it
shrunken indeed, but still recognisable, in relation to the
embryonal area in vesicles 4'5-6 mm. in diameter. And
there may even appear within it peripherally, irregular strands
which stain deeply wit^ iron-h<ematoxylin and which recall
those forming the periphei’al deutoplasmic network of the early
blastomeres. Eventually, however, it seems to disappear, its
substance passing into the blastocystic fluid, so that, as already
remarked, it fulfils in this indirect way its original destiny.
Normally the cavity of the just completed blastocyst contains no cellular elements whatever. In one otherwise
perfectly normal blastocyst (’39 mm. diam.) I find present.
j. p. hilL. ‘
however, a small spheroidal body *028 mm. in diameter,
composed of glasSy-looking cytoplasm enclosing a central
deeply staining granule. This I interpret as a cell or cellfragment which has been accidentally separated off from the
wall, and which has undergone degeneration. In later
blastocysts such cellular bodies exhibiting more or less
evident signs of degeneration are of fairly common occurrence. They are of no morphological significance.
Selenka^s Blastopo.re.” -  r-N'ormally the wall of the
blastocyst is first completed over the upper hemisphere, in
correspondence with the fact-that the formative cells not
only divide somewhat more rapidly than the non-formative
but have a smaller extent of surface to cover, since the upper
cell-ring from which they are derived lies about midway
between the upper pole of the sphere formed by the eggenvelopes and the equator of the same, whilst the lower cellring from which the noil-formative cells arise is approximately
equatorial in position. We thus meet with stages in the
formation of the blastocystic wall such as are represented in
surface view on PI. 3, fig. 30, and in section in figs. 31 and
32, in which the blastocystic cavity, prior to the completion
of the cellular walTbver the lower polar region, is more or less
widely open below. There can be no doubt, I think, but that
this opening corresponds to that observed by Selenka in his
42-celled “gastrula^’ of Didelphys and regarded by him as
the blastopore, since he believed the entoderm arose from its
lips. My observations conclusively show that it has no
connection whatever â–  with the entoderm, this layer arising
from the formative region of the upper hemisphere, and that
it is a mere temporary opening of no morphological significance, blastoporic or other. Pi’ior to the completion of
the wall at the upper pole a corresponding opening is temporarily present there also. Both owe their existence to the
characteristic way in which the blastocyst wall is formed by
the spreading of the products of division of the two cell-rings
of the 16-celled stage towards opposite poles in contact with
the surface provided by the enclosing egg-envelopes.
I have met with one specimen, an incomplete blastocyst
•39 mm. in diameter (belonging to the same batch as the
other blastocysts referred to in this section^), in which the
lower hemisphei’e would appear to have been completed before
the upper, for the yolk-body lies in contact with the zona in
the region where the cellular wall is as yet absent, and that
the yolk-body has not been secondarily displaced is proved by
a micro-photograph of the specimen in my possession (taken
immediately after its transference to the fixing solution), in
which the yolk-body is seen to lie at the unclosed pole in
exactly the same position as in the sections.
In connection with this exceptional specimen, it may be
recalled that Selenka, in his 68-celled “ gastrula” of Didelphys
(fig, 10, Taf. xvii), figures the wall as complete at the lower
pole, the “blastopore” having alx’eady closed, but as still incomplete at the upper pole, there being present a small opening
leading into the blastocyst cavity. In the 42-celIed “gastrula”
(fig. 8, Taf. xvii) this same opening and the “blastopore” as
well are present. The occurrence of these openings at
opposite poles, and the general agreement in the constitution
of the blastocyst wall (larger, more yolk-rich cells at lower
pole, smaller, less yolk-rich cells at upper), in the corresponding stages in Didelphys, and Dasyurus justify the conclusion that the blastocyst of the former develops in the same
way as does that of the latter. It is worthy of remark,
however, that the just completed blastocyst of Didelphys
appears to be considerably smaller than that of Dasyurus.
Selenka unfortunately gives no measurements of his early
stages, but I have calculated from the figure, the magnificatiou
of which is given, that the ^8-celled blastocyst has a diameter
of about ‘IS? mm. The corresponding stage of Dasyurus
measures about '39 mm., and is therefore nearly three times as
• This batch, from female 2 b, 16 . vii . ’01, comprised altogether
twenty-eight eggs, of which some eighteen were normal complete and
incomplete blastocysts mm. in diameter) and ten abnormal, four
of these being unsegmented ova.
.1. P. HILL.
Selenka’s Urentoderinzelle.  -  Whilst the 42- and 68celled blastocjsts described by Selenka may be regarded
as normal so far as the occurrence of polar openings and
the constitution of their wall ai‘e concerned, I hold them to be
abnormal in respect of the presence in each of a single large
yolk-laden cell, regarded by Selenka as entodermal in significance. It is well to point out that Selenka was not able
actually to determine the fate of this cellj he merely presumed
that it took part in the formation of the definitive entoderm.
No such cell occurs in normal blastocysts of Dasyurus at any
stage of development, and in my opinion Selenka’s “ urentodermzelle” is none other than a retarded and displaced
blastomere, i.e, a blastomere which has failed for some
reason to divide, and which has become secondarily enclosed
by the products of division ot its fellows, and I am
strengthened in this interpretation by the occurrence in
an abnormal blastocyst of Dasyurus of just such a large
cell as that observed by Selenka. The vesicle in question
is one of the batch already referred to, and measured ’397 mm.
in diameter. The cellular wall (fig. 37) is apparently normal,
but is incomplete' at one spot, and the gap so left is occupied
by a large binucleated cell, i-ich in deutoplasm and measuring
•12 X ’072 mm. (fig. 37, abn.). This cell corresponds in its
size and cytoplasmic characters with a non-formative blastomere of about the 16-celled stage, and I l egard it simply as
a blastomere which has failed to undergo normal division.
In another abnormal blastocyst (‘39 mm. diam.) from the
same batch, the cellular wall appears complete and normal,
but the blastocyst cavity contains a group of about sixteen
spherical cells averaging about ‘032 mm. in diameter, and in
yet another abnormal egg ot’ the same diameter and batch
there is present an incomplete layer of flattened cells over
one hemisphere, and towards the opposite pole of the eggsphere there occurs a group of spherical cells of variable size
and some of them multinucleate. In this abnormal egg it
appears as if the formative cells had divided in fairly normal
fashion, whilst the nou-formative cells had failed to do so.
Chaptee IV.  -  Growth oe the Blastocyst and Differentiation OF THE Embryonal Ectoderm and the Entoderm.
1. Growth of the Blastocyst.
In the preceding chapter we have seen that the cleavage
process in Dasyurus results in the formation of a small
spherical vesicle, about '4 mm. in diameter, Avhich consists,
internally to the investment formed by the apposed zona and
shell-membrane, simply of a cellular wall, unilaminar throughout its entire extent, and enclosing a fluid-filled cavity
normally devoid of any cellular elements. The stage of the
just completed blastocyst is followed by a period of active
growth of the same, and it is a noteworthy featui’e in the
development of Dasyurus that during this time the blastocyst
undergoes no essential structural change, but remains unilaminar until it has reached a diameter of from 4'5 to 5'5 mm.
Even during cleavage, the egg of Dasyurus increases in
diameter, partly owing to the thickening of the shell membrane, partly, and, indeed, mainly, as the result of the accumulation of uterine fluid under pressure within the egg-envelopes,
but the increase due to these causes combined is relatively
insignificant, being only about '1 mm. As soon, however, as
the cellular wall of the blastocyst is completed, rapid growth
sets in, under the influence of the hydrostatic pressure of the
fluid, which tensely fills the blastocyst cavity, with the result
that the small relatively thick-walled blastocyst becomes
convei’ted into a large extremely thin-walled vesicle, but
beyond becoming very attenuated, the cellular wall during
this period of actjve growth uudei'goes no essential change,
and retains its unilaminar character until the blastocyst, as
already mentioned, has reached a diameter of from 4'5 to 5‘5
mm. In vesicles of about this size there become differentiated
from the formative cells of the upper hemisphei-e the embryonal ectoderm and the entoderm, and this latter layer then
gradually spreads round inside the non-formative (extraembiwonal ectodermal) layer of the lower hemisphere so as to
form a complete lining' to tlie blastocyst, whicli thereby
becomes bilaminar. Sucli a marked enlargement of the blastocyst prior to the differentiation of the embryonal ectoderm
and entoderm as is here described for Dasyurus does not
apparently occur, so far as known, in other Marsupials : in
Perameles, for example, the embryonal ectoderm and the
entoderm are in process of differentiation in vesicles a little
over 1 mm. in diameter (v. p. 77), in Macropus these two layers
are already fully established in a vesicle only *8 mm. in
diameter (v. p. 79), and much the same holds good for Ti'ichosurus and Petrogale. It is pai’alleled by the marked growth
which in the Monotremes follows the completion of the blastocyst and which precedes the appearance of embryonal diffei-entiatiou. It must be remembered, however, that the growing
blastocyst in the Monotreme is bilaminar and not unilaminar
as in Dasyurus, owing to the fact that the entoderm is established as a complete layer at a very much earlier period than
is the case in the latter. I am nevertheless inclined to regard
the attainment by the Dasyurus blastocyst of a large size,
prior to the differentiation of the embi’yonal ectoderm and the
entoderm, as a more primitive condition than that found in
other Marsupials. The pronounced hypertrophy which the
uteri of Dasyurus undergo during the early stages of gestation, an hypertrophy which appears to be proportionately
greater than that met with in other forms,^ is no doubt to be
correlated with the presence in them of such a considerable
number of actively growing blastocysts.
Selenka states (Heft 5, p. 180) that he examined seven
blastocysts of Dasyurus “-f mm.” in diameter, taken from a
female fifteen days after copulation. He describes their
structure as follows : “ Man unterscheidet (1) eine sehr
zarte aussere, homogene Haut (Granulosamembran), (2)
' Por example, the uteri of a female (5, 18 . vii . '01) from which 1
obtained twenty-one normal vesicles, 4'5-6 mm. in diameter, with the
embryonal area definitely established, measured as follows : Left uterus,
4'5 X 4'7 X 1'4 cm. (fourteen vesicles) ; right uterus, 4'5 X 4'2 x 1‘45 cm.
(seven vesicles and one shrivelled).
darmiter ein Lagei’ von Ektodermzellen, welche im Gebiete
des Embryonalschildes prismafcich, am gegeniiberliegenden
Pole nahezu kubisch, im iibrigen abgeplattet erscbeinen, (3)
ein inneres zusammenliangeudes Lager von abgeflacbten Entodermzellen.” This description, apart from the reference to
the thin shell-membrane, is entii’ely inapplicable to blastocysts
of Dasyurus of the mentioned size which I have studied.
I have examined a practically complete series of vesicles of
Dasyinms ranging from ’4 mm. to 4 mm. in diameter and all of
them without exception are unilaminar.
Of vesicles under 1 mm. diameter I possess serial sections
of more than two dozen, I’anging from ’5 mm. to '8 mm. in
diameter, and obtained from three different females. These
differ structurally in no essential respect from the just completed blastocysts. A surface view of a blastocyst '6 mm. in
diameter is shown in fig. 63, PI. 6; in this the difference in
the cytoplasmic chai’acters of the cells of opposite hemispheres
is clearly brought out, the non-formative cells of the lower
hemisphere having much more marked perinuclear zones of
dense cytoplasm (deutoplasm) than the formative cells of the
upper hemisphere ; moreover, the former cells tend to be of
larger superficial extent than the latter. Pig. 34, PI. 3,
represents a section of a blastocyst '57 mm. in diameter, and
fig. 35 a section of one ’73 mm. in diameter. These blastocysts differ in no essential way from the ’43 mm. blastocyst
represented in fig. 33. As in the latter, the cellular wall is
unilaminar throughout, but both it and the shell-membrane
have undergone considerable attenuation. Moreover in these
blastocysts, apart from the clue afforded by the shrivelled
yolk^body, it is practically impossible to determine from the
sections which is morphologically the upper hemisphere and
which the lower. In fig. 36, from a '6 mm. blastocyst, on the
other hand, the cells of the hemisphere opposite the yolk-body
(y.b.) are larger than those of the hemisphei'e adjacent to
which that body is situated. In the '57 mm. blastocyst the
shell-membrane has a thickness of ‘0052 mm., in the -73 mm.
blastocyst it measures -0045 rnrh., and in a -84 mm. blastocyst
•0026 mm. The zona is now no longer recognisable as an
independent membrane. In blastocysts of this stage of
growth a variable number of small spherical cells or cellfragments are frequently met with in the blastocyst cavity,
usually lying in contact Avith the inner aspect of the cellular
wall (fig. 34, i.c.). In some blastocysts such structures are
absent, in others one or two may be present, in yet others
numbers of them may occur. They raa,y be definitely nucleated,
but this is exceptional; more usually they contain one or more
deeply staining granules (of chromatin ?), or are devoid of
such. They ai’e of no morphological importance, and I think
thei*e can be no doubt that they represent cells or fragments
of cells which have been separated off from the cellular wall
during the process of active growth. They are of common
occurrence in later blastocysts, and it is possible the so-called
“ yolk-balls ” observed by Selenka in Didelphys are of the
same nature.
If we pass now to vesicles from 1 to 3 or 3'5 mm. in
diameter, we find the wall still unilaminar, but considerably
more attenuated than it is in the blastocysts last referred to.
In a vesicle with a diameter of 1’24 mm. the shell-membrane
has a thickness of about '0015 mm., whilst the cellular wall
has a thickness of only ’0045 mm. In a 3’5 mm. vesicle the
shell-membi’ane measures about •0012 mm., Avhilstthe cellular
wall ranges from •OOlS to •OOIS mm. in thickness. A small
portion of the wall of a vesicle, 2^4 mm. in diameter, is shown
in PI. 6, fig. 64. In these later vesicles I have failed to detect,
either in surface examination of the vesicles in to to or in
sections, any regional differences between the cells indicative
of a differentiation of the wall into upper or formative, and
lower or non-formative, hemispheres. Everywhere the wall
is composed of flattened, exti’emely attenuated cells, polygonal
in surface view, and all apparently of the same character. It
might therefore be supposed that the polarity, which is recognisable in early blastocysts, and which is dependent on the
pronounced differences existent between the cells of the
upper and lower rings of the 16-celled stage, is of no funda
mental importance, since it apparently becomes lost at an
early period during the growth of the blastocyst. Such an
assumption, however, would be very wide of the maxk, as I
hope to demonstrate in the next section of this paper, and,
indeed, in view of the facts already set forth, is an altogether
improbable one.
Reappearance of Polar Differentiation in the
Blastocyst Wall.  -  Following on the period of what may
be termed the preliminary growth of the blastocyst, in the
course of which the original polar differentiation in the
blastocyst wall apparently becomes obliterated, is an
extremely interesting one, during which that differentiation
again becomes manifest. In view of the fact (1) that the
fourth cleavage in Dasyurus is of the nature of a qualitative
cytoplasmic division, and (2) that approximately one half or
rather less of the unilaminar vosicle wall is formed from the
eight smaller and less yolk-rich cells of the upper ring of the
16-celled stage, and its remainder from the eight larger
more yolk-rich cells of the lower ring, it thus becomes a
question of the first importance to determine if we can the
significance of that differentiation.
Amongst the Eutheria, it has been conclusively shown by
various observers (Van Beneden, Heape, Hubrecht, Assheton,
and others) that there occurs during cleavage an early
separation of the blastomeres into two more or less distinctly
differentiated groups, one of which eventually, by a process
of overgrowth, completely encloses the other. The peripheral
cell-group or layer forms the outer extra-embryonal layer of
the wall of the later blastocyst (the trophoblast of Hubrecht,
or trophoblastic ectgderm as I prefer to term it). It therefore
takes no direct part in the formation of the embryo, and may
be distinguished as non-formative. The enclosed cell-group,
termed the inner cell-mass or embryonal knot, gives rise, on
the other hand, to the embryonal ectoderm as well as to the
entire entoderm of the vesicle, and may accordingly be distinguished as formative. May it not be, then, that we have
here at the fourth cleavage in Dasyurus a separation of the
blastomeres into two determinate cell-groups, respectively
foi’mative and non-formative in significance, entirely compar-,
able with, and, indeed, even more distinct than that which
occurs during cleavage in the Eutheria ? I venture to think
that the evidence brought forward in this paper conclusively
justifies an answer in the affirmative to that question.
If we assume that the upper cell-ring of the 16-celled stage
in Dasyurus is formative in destiny and the lower cell-ring
non-formative, then we might naturally expect to find in the
unilaminar wall of the later blastocyst some differentiation
indicative of its origin from two distinct cell-groups, and
indicative at the same time of the future embryonal and
extra-embryonal regions. Now just such a differentiation,
does, as a matter of fact, become evident in vesicles 3'5 to
4’5 mm. in diameter. We have already seen that the wall in
early blastocysts ’4 to '8 mm. in diameter exhibits a wellmarked polar differentiation in correspondence with its mode
of oi’igin from the diffei-entiated cell-rings of the 16-celled
stage, its upper hemisphere or thereabouts consisting of
smaller cells, poor in deutoplasm, its remainder of larger
cells, rich in deutoplasm. .In later blastocysts, 1-3 mm. or,
more in diameter, it is no longer possible to recognise this
distinction  -  at all events I have failed to observe i't  -  but if
we pass to blastocysts 4-5 mm. in diameter, in which the wall
is still unilaminar, we find on careful examination of the
entire vesicle under a low power that there is now present a
definite continuous line^ which encircles the vesicle in theequatorial region so as to divide its wall into two hemi-,
spherical areas (PI. 4, fig. S8,j.L). If we remove and stain;
a portion of the wall of such a vesicle, including this line,)
and examine it microscopically (figs. ,42-46.), it becomes
apparent at once, from the .disposition of the cells on either
side of the bns, that we have to do with a sutural line or line
of junction produced by the meeting of twp sets- of cells,'
which are pursuing their .own, independent courses of growth
and division. ; The, cells never cross the demax’cation line
from the one side tn the other, but remain strictly confined
to their own territory, so that we are justified in regarding
the vesicle wall as composed of two independently growing
zones. Now tlj^ existence of two such independent zones in
the unilaminar wall is, to my mind, only intelligible on the
view that they are the products of two originally distinct,
predetermined cell'groups, and if this be admitted, then I
think we are justified in concluding, in view of the facts
already set forth, that tlie two zones in question are derived,
the one from the upper cell-ring of the 16-celled stage, the
other from the lower ring ; that, in other words, they represent respectively the upper and lower hemispheres of the
early blastocysts.
If, now, we find that the embryonal ectoderm and the entoderm arise from one of these two regions of the unilaminar
wall, whilst the other directly forms the outer extra-embryonal
layer of the later (bilaminar) vesicle, then we must designate
the former region as the upper or formative, and the latter as
the lower or non-formative. Further, bearing in mind the
characters of the cells of the two rings of the 16-celled stage,
T think we are justified in holding that the formative region
is derived from the ring of smaller, less yolk-rich cells, and
the non-formative region from the ring of larger, more yolkrich cells, even if it is impossible to demonstrate an actual
genetic continuity between the constituent cells of these two
rings and those forming the independently growing areas of
the later blastocyst. I have recently re-examined a series of
vesicles, measuring 1’5-1'8 mm. in diameter, obtained from a
female killed in 1906, and I have so far found it impossible,
either in the entire vesicle or in portions of the wall stained
and mounted on the flat, to distinguish between the cells over
opposite hemispheres. Thus the only actual guide Ave have
for the determination of the poles in such vesicles is the
yolk-body, and though the latter is liable to- displacement, it
is Avorthy of record that I have several times found it in
relation to the formative area in vesicles 4‘5-6 mm. in
diameter, but never in relation to the non-formative region.
This evidence is, therefore, so far as it goes, confirmatory of
VOL. 56, PART 1.  -  NEAV SERIES. 4
the conclusion reached above, viz, that the formative hemisphere is derived from the smaller-celled ring of the 16-celled
stage. On that conclusion is based my interpretation of the
poles in the unsegmented ovum, and of the two cell-rings o£
the 16-celled stage as respectively upper and lower.
Of vesicles ov'er 1 mm. in diameter, the smallest in which I
have been able to detect the sutural line above referred to
measure 3'25 mm. in diameter. In three lots of vesicles, 3'5
mm. in diameter from three different females, I have failed to
X’ecognise it, whilst in two other lots, respectively 3'75 mm.
(average) and 4 mm. in diameter, the line appears to be in
course of differentiation as in the 3’25 mm. vesicles. A
portion of the wall of one of the 3'5 mm. vesicles just referred
to is shown in PL 4, fig. 41, and a portion of the wall of the
3'25 mm. stage, including the sutural line, in fig. 42. Both
vesicles were fixed in the same fluid, viz. picro-nitro-osmic
acid. Comparison of the two figures reveals the existence, quite
apart from the presence of the junctional line in fig. 42, and its
absence in fig. 41, of certain more or less obvious differences
between them. In fig. 41 the cells are larger, and their cytoplasmic bodies are inconspicuous, being fairly homogeneous
and lightly staining. In fig. 42, on the contrary, the cellbodies are strongly marked, the cytoplasm being distinguishable into a lighter-staining peripheral zone, and a much more
deeply staining perinuclear zone, showing evidence of intense
metabolic activity. This latter zone is more or less vacuolated,
and contains, besides larger lightly staining granules, numerous
smaller ones of varying size, stained brown by the osmic acid
of the fixative. In the 4 ram. vesicles the cells show pi-ecisely
the same characters; in the 3'75 mm. vesicles, which were
fixed in a picro-corrosive-acetic fluid, the granules ax’e absent
from the cytoplasm, otherwise the cells are similar to those
of the other two. Mitotic figures are common. The sutural
line is recognisable in all three sets of vesicles (3'25, 3’75, and
4 mm.) (fig. but I cannot be certain that it runs con
tinuously round, and it appears to have a rather more sinuous
course than in later blastocysts. The cells of the two regions
of the bliistocyst wall, separated by the sutural line, differ
somewhat in tlieir characters. On one side of the line (fig.
42, tr.ect.) the cells appear to be on the whole slightly larger,
and of more uniform size than they are on the other, and they
also stain somewhat more deeply. Comparison with later
blastocysts shows that the region of more uniform • cells is
non-formative, that of less uniform, formative. At^this stage,
however, the differences between the cells of the two regions
are as yet so little pi’onounced that it is practically impossible
in the absence of the sutural line to say to which hemisphere
an isolated piece of the wall should be referred.
I am inclined to regard the sutui'al line in these vesicles as
being in course of differentiation, and judging from the disposition of the cells on either side of it, I think its appearance
is to be correlated with the marked increase in the mitotic
activity of the cells of the two hemispheres which sets in in
vesicles of 3-4 mm. diameter. The preliminary increase in
size of the blastocyst up to about the 3 mm. stage might be
described as of a passive character, i.e. it does not appear
to be effected as the result of the very active division of the
wall-cells, but is characterised rather by a minimum of mitotic
division and a maximum of increase in surface extent of the
cells, due to excessive stretching consequent on the rapid
imbibition of uterine fluid. Once, however, the requisite size
has been attained, the cells of the unilaminar wall commence
to divide activel}', and doubtless as the outcome of that
wave of activity, the sutural line makes its appeai-ance
between the two groups of independently growing cells.
On the inner surface of the blastocyst wall, especially in
the region of the formative hemisphere, there are present
in these vesicles numbers of small deeply staining cells of
spherical form, and containing osmicated granules similar
to those in the wall-cells. They may occur singly or in groups,
and appear to me to be of the same nature as the inteimal cells
of the earlier blastocyst. In addition to these cells, there are
present clusters of cytoplasmic spheres, staining similarly to
the spherical cells, and apparently of the nature of fragmeiita
tion products formed either directly from the â– \vall-cells or
from these internal cells.
2. Differentiation of the Embryonal Ectoderm and
the Entoderm.
After the preliminary growth in size of the blastocyst is
completed, the next most important step in the progressive
development of the latter is that just dealt with, involving
the appearance of the sutural line, with resulting re-establishment of polar differentiation in the blastocyst wall. Following
on that, we have the extremely important period during
which the embryonal ectoderm and the entoderm become
definitely established.
For the investigation of the earlier phases of this critical
period I have had at my disposal a large number of
unilaminar blastocysts derived from three females, distinguished in my notebooks as (3, 25 . vii . ’01, with fifteen
vesicles of a maximum diameter of 4‘5 mm. ; 8 . vii . ’99, with
twelve vesicles, 4‘6 .mm. in diameter ; and 6 . vii . ’04, with
twfenty-two vesicles, 4‘5 and 5 mm. in diameter. These three
lots of vesicles may for descriptive purposes be designated
as ’01, ’99, and ’04 respectively.
The ’01 vesicles are distinctly less advanced than the
other two. The sutural line is now, at all events, definitely
continuous, and can readily be made out in the intact vesicle
with the aid of a low-power lens (PI. 4, fig. 38, j.L), but
the differences between the cellular constituents of the two
hemispheres which it separates are much less obvious than
they are in the ’99 and ’04 vesicles. Here, again, one
hemisphere forming half or perhaps rather more of the entire
vesicle is distinguished from the other by the greater uniformity and the slightly deeper staining character of its
constituent cells (figs. 43 and 44, tr. ecL). This hemisphere,
subsequent stages show, is the lower or non-formativ^
hemisphere. It is characterised especially by the striking
'uniformity in the size of its cells. Over the opposite hemisphere, the upper or formative one (figs. 43 and 44, the
cells are more variable iu size, the nuclei thus appearing less
uniformly and less closely arranged, and they stain,. on the
whole, somewhat less deeply than those of the lower hemisphere. The non-formative cells are on the average smaller
than the largest of the formative cells, but they are more
uniform iu size, and their nuclei thus lie at more regular
distances apart, and appear more closely packed. They are
also richer in deutoplasmic material, and so stain rather more
deeply than the formative cells. Sections show that the
cellular wall is unilaminar throughout its extent, and that,
whilst it is somewhat thicker than that of 3‘5 mm. vesicles,
it is still very attenuated, its thickness, including the shellmembrane, ranging-from ‘004 to '008 mm. I have examined
a number of series of sections taken through portions of the
wall known to include the sutural line, and find it quite
impossible to locate the position of the- latter; indeed, I
cannot certainly distinguish between the formative and nonformative regions.
In the blastocyst cavity, lying in contact with the inner
surface of the wall, and most abundant in the region of the
formative hemisphere, there are present numbers of deeply
staining spherical cells with relatively small nuclei similar to
those described in connection with the 3'25 mm. vesicles.
They occur singly or in groups, and may appear quite normal
or may show more or less evident signs of degeneration.
Their nuclei may stain deeply and homogeneously, or may be
represented by one or two deeply staining granules, vacuoles
may occur in their cytoplasm, and spherical cytoplasmic masses
of very variable size, with or without deeply staining granules
of chromatin) may occur along with them. In sections and
preparations of the wall of these, and other 4*5 mm. vesicles
there are to be found, in both the formative and non-formative
hemispheres, small localised areas from which such spherical
cells are being proliferated off in numbers together. PI. 5,
fig. 47, from the formative hemisphere of an ^04 vesicle shows
One of the most marked examples of such proliferative. activity
that I have encountered. A similar but smaller proliferative
J. ?. HILL.
area occurs on the non-formative hemisphere of the same
These spherical cells are, I am convinced, of no morphological importance, and are destined sooner or later to degenerate. They have certainly nothing to do with the
entoderm, the parent-cells of that layer arising exclusively
from the formative hemisphei'e and not from cells such as
these, which are budded ofE from both hemispheres. The fact
that they are, in unilaminar vesicles, more numerous over the
formative hemisphere may perhaps be taken as an indication
of the greater mitotic activity of the formative as compared
with the non-formative cells.
The Primitive Entodermal Cells.  -  Following closely
on the stage represented by these ’01 blastocysts is the extremely important one constituted by the ’99 and ’04 vesicles
before referred to. This stage is the crucial one in primary
germ-layer formation, and marks the transition from the unilaminar to the bilaminar condition, since in it the entodermal
cells are not only distinctly recognisable as constituents of the
formative region, but are to be seen both in actual process of
separation from the latter and as definitely internal cells, frequently provided with, and even connected together by,
pseudopodial-like processes of their cell-bodies. Such cells
are already present in the ’01 vesicles (fig. 71), and probably
also in the blastocysts in which the sutural line first makes
its appearance, but are much less conspicuous than in these
older blastocysts.
The ’99 blastocysts are distinctly more advanced than the
’01 batch and are just a little earlier than the ’04 lot. The
former measui’ed, as already mentioned, 4’5 mm. in diameter,
the latter 4’5 and 5 mm. (the majority being of the latter
size). In my notes, on the intact ’99 vesicles I find it stated
that one hemisphere, forming rather less than half of the
entire extent of the vesicle wall, appeared somewhat denser
than the other, the sutural line marking the division between
the two. I naturally inferred at the time that the denser
hemisphere corresponded to the embryonal region of the
Eutherian blastocyst and the less dense to the extra-embryonal region of the same, but just the reverse proves to hold
true for the ’04 vesicles, the formative hemisphere in these
appearing less dense than the non-formative. I cannot now
test my former inference by direct observation since I do not
appear to have any of the ’99 vesicles left intact, but amongst
my in toto preparations of the vesicle wall I find one
labelled as from the “ lower pole ” which unmistakably
belongs to the formative hemisphere, hence I conclude that
the denser and slightly smaller region which I originally
regarded as formative is really non-formative, a conclusion
which brings the ’99 vesicles into agreement with the ’04
In these latter vesicles the sutural line and the two regions
of the wall can be quite readily made out on careful examination under a low power with transmitted light. The one
region appears slightly denser (darker) and has more closely
arranged nuclei (i. e. is composed of smaller cells) than the
other. On the average this denser region appeal’s to be
rather the less extensive of the two ; the two regions may be
about equal ; on the other hand the denser may be the smaller.
Examination of stained preparations of the wall demonstrates
that the darker hemisphere is non-formative, the lighter,
formative. It would therefore seem that in certain of these
’04 vesicles the formative region has grown more rapidly than
the non-formative.
In stained preparations of the wall both of the ’99 and ’04
vesicles, the differences between the two hemispheres are now
so well marked that there is no diflBculty in referring even an
isolated fragment to its proper region. The non-formative
hemisphere differs in no essential way from that of the ’01
vesicles, and as in these, is readily distinguishable from the
formative by the much greater uniformity in the size and
staining properties of its cells (fig. 45), as well as by the fact
that there are no primitive entodermal cells such as occur in
relation to the formative hemisphere, in connection with it.
Its constituent cells are on the average distinctly smaller than
the largest of the formative ; their nuclei lie nearer each other,
with the result that in surface examination of the blastocyst
the non-formative region appears rather denser than the
formative. In in toto preparations of the wall the former
usually stains darker than the latter (fig. 45), but this is not
always the case ; in fig. 46, from an ’04 vesicle, there is
practically no difference in this respect between the two
regions ; in yet others of my preparations of ’99 vesicles the
formative region has stained more deeply than the nonformative.
The formative hemisphere in the earlier blastocysts of this
particular developmental stage was described (ante, p. 51) as
differing from the non-formative in that its constituent cells
were much less uniform in chai*acter than those of the latter.
This same feature, but in much enhanced degree, characterises
the formative region of the vesicles under consideration, for it
can now be definitely stated that the latter I'egion is constituted by cells of two distinct varieties, viz. (1) moi*e lightly
staining cells which form the chief constituent of the formative region, its basis so to speak, and which are on the
average larger than those of the other variety, and (2), a less
numerous series of cells, distinctly smaller than the largest
cells of the former variety, and with denser, more granular and
more deeply staining cytoplasm, and frequently met with in
mitotic division (cf. PI. 6, fig. 65). The two varieties of cells are
intermingled promiscuously, the smaller cells occurring singly
and in groups but in a quite irregular fashion, so that here
and there we meet with patches of the wall composed exclusively of the larger cells.
The evidence presently to be adduced shows that the larger
cells furnish the embryonal ectoderm, and that the smaller
cells give origin to the primitive entodermal cells from which
the definitive entoderm arises. The smaller cells may therefore be regarded as entodermal mothei’-cells. Whether these
latter cells are progressively formed from the larger cells
simply by division, or whether the two vaifieties become
definitely differentiated from each other at a particular stage in
development, must for the present be left an open question. Of
the actual existence in tlie unilaminar formative region of these
’99 aud ’04 blastocysts of two varieties of cells, respectively
ectodermal and entodermal in significance, there can be no
doubt. In preparations of the formative region, however,
whilst one can without hesitation identify certain cells as
being in all probability of ectodermal significance and others
as prospectively entodermal (cf. figs. 65, 66), it must be
admitted that one is often in doubt as to whether one is
dealing with small ectodermal cells or with genuine entodermal mother-cells. It is, therefore, hardly to be wondered
at that I have not yet been able to satisfactorily determine
at what precise period the entodermal mother-cells first
become differentiated, though judging from the facts that
in the eai-liest vesicles in which the sutural line is recognisable one region of the wall already differs from the other in
the less uniform size of its constituent cells, and that internally
situated entodermal cells are already present in small numbers
in the ’01 vesicles (fig. 71), I incline to the belief that it
will probably be found to about coincide with the first
appearance of the sutural line. To this question I may
perhaps be able to return at some future time.
In addition to the presence of these entodei’mal mothercells, which enter directly into its constitution, the formative
region of the ’99 and ’04 blastocysts is. characterised by the
occurrence on its inner surface of definitely iuteimal cells,
which generally agree with the former cells as regards size
and staining properties and are evidently related to them. It
is these internally situated cells which directly give origin to
the definitive entoderm of the later blastocysts, and one need,
therefore, have no hesitation in applying to them the designation of primitive entodermal cells. They are exclusively found
in relation to the formative hemisphere, and appear in in toto
prepai’atious as flattened, darkly staining cells closely applied
to the inner surface of the unilaminar wall, and disposed quite
irregularly, singly, and in groups. They vary greatly in
number in blastocysts of even the same batch, but on the
J. r. HILL.
wholo are most abundant in the ^04 series, and they also
exhibit a remarkable range of variation in shape. They may
have a perfectly distinct oval or rounded outline (figs. 67, 71,
72), or, as is more frequently the case, they may lack a
determinate form and appear quite like amoeboid cells owing
to their possession of cytoplasmic processes of markedly
pseudopodial-like character (fig. 69). Frequently, indeed,
the cells are connected together by the anastomosing of these
processes, so that we have formed in this way the beginnings
at least, of a cellular reticulum (figs. 68, 69,70).
The question now arises. How do these primitive entodermal cells originate from the small, darkly staining cells of
the unilaminar formative region designated in the foregoing
as the entodermal mother-cells ? I can find no evidence that
the primitive entodermal cells are formed by the division of
the mother-cells in planes ta.ngential to the surface ; on the
contrary, all the evidence shows that we have to do here with
an actual inward migration of the mother-cells, with or without previous mitotic division, such inward migration being
the outcome of the assumption by the mother-cells, or their
division products, of amoeboid properties ; in other words, the
evidence shows that the formation of the entodei’m is effected
here not by simple delamination (using that term in the sense in
which it was originally employed by Lankester), but by a process involving the inward migration, with or without previous
division, of certain cells (entodermal mother-cells) of the unilaminar parent layer, a process comparable with that found in
certain Invertebrates (Hydroids) and distinguished by Metschnikoff as '^gemischte Delaminatiou.”
In this connection it has to be remembered that the cells of
the unilaminar wall of the blastocyst are under considei’able
hydrostatic pressure, and, in correlation therewith, tend to
be tangentially flattened, though the flattening in this stage
is much less than in the earlier blastocysts. From a series of
measurements made from an '04 vosicle, I find that over the
formative region the ratio of the breadth to the thickness of
the cells varies Horn 6 : 1 to 2 : 1, and even to 3 : 2. On the
whole cells of the type indicated by the ratio 6 : 1 predominate,
and we should hardly expect to find such cells dividing tangentially. In fact, the only undoubted examples of such division I
have met with occur in the single abnormal vesicle present in
the ’04 batch. In this particular vesicle, which had a diameter
of 3 mm. and was thus smaller than the others, thei’e was
present on what appeared to correspond to the formative
hemisphere of the normal blastocyst a well-defined and conspicuous ovalish patch, 1’23 x ’99 mm. in diameter.^ Sections
show that over this area the cells of the unilaminar wall are
much enlarged and , more or less cubical in form, their thickness varying from ‘012 to ‘019 mm. These cubical cells
exhibit distinct evidence of tangential division, both past and
in progress. But in normal vesicles, whilst mitotic figures are
quite commonly met with in the cells of the formative region
(in which, indeed, they are more numerous than in those of
the non-formative region), I have failed to find in my sections
after long-continued searching even a single spindle disposed
directly at right angles to the shell-membrane ; the mitotic
spindles lie disposed either tangentially to the surface or
obliquely thereto.
For the determination of the mode of origin of the
primitive entodermal cells, it is absolutely necessary to
study both in to to preparations of the formative region,
i.e. small portions of the unilaminar wall stained and
mounted on the flat, and sections of the same. Sections alone
are, on the whole, distinctly disappointing so far as the
question under discussion is concerned, and, indeed, give one
an altogether inadequate idea of the primitive entodeimial cells
themselves, seeing that practically all one can make out is that
1 Curiously enough, amongst the ’99 vesicles there also occurred
a single small one, likewise 3 mm. in diameter, and with a thickened
patch 1-28 X 1 mm. in diameter, quite similar in its character to that
described in the text. I am as yet uncertain whether the thickened
area in these two vesicles represents the whole of the formative hemisphere of normal blastocysts or only a hypertrophied part of the same,
or whether, indeed, it may not represent the I'etarded non-formative
there are present, in close apposition with the inner surface of
the unilaininar wall, small, darkly staining cells, apparently
quite isolated from each other and usually of flattened form
(figs. 73, 74, 76, ent.). One has only to glance at a wellstained in to to preparation of the formative region (cf.
fig. 70) to realise how inadequate such a description of the
primitive entoderm cells really is.
Sections nevertheless do yield valuable information on
certain points. Besides affording the negative evidence of
the absence of tangential divisions and the positive evidence
that the primitive entodermal cells are actually internal (figs.
73, 74, 76), they show that growth of the wall iu thickness
has already set in, and that it is most marked over the
formative region, though the thickness attained by the cells
is as yet very unequal (figs. 73-76). Measurements takeu
from an ’04 vesicle show that over the non-formative region
(fig. 77) the cells vary in thickness from *006 to '009 mm.,
whilst over the formative region the range of variation is
greater, viz. from ‘006 to ‘013 mm., so that we may conclude
that the latter region is on the average thicker than the
former (cf. figs. 73-76, with fig. 77 depicting a small portion
of the non-formative region). It is still impossible to determine the position of the sutural line, even in sections of
fragments of the wall known to contain it.
The entodermal mother-cells are not very readily recognisable in sections. In fig. 75, however, which is drawn
from an accurately transverse section through the formative
region of an ’04 vesicle, there is depicted what is undoubtedly
an entodermal mother-cell {ent.). The interesting point
about this particular cell is that its cell-body, whilst still
intercalated between the adjoining cells of the unilaminar
wall, has extended inwards so as to directly underlie one of
the wall-cells. ' Division of such a cell as this would necessarily result in the production of an internally situated cell
with all the relations of one of the primitive entodermal type.
The inwardly projecting spheroidal cell situated immediately
to the left (in the figure) of the one just refeiTed to, I also
regard as an entodermal mother-cell. Cells of this type are
not infrequently met with in sections; they nsually stain
somewhat deeply, and are often found in mitosis.
The evidence obtainable from the study of in to to preparations conclusively proves that some at all events of the
primitive entodermal cells are actually derived from the entodermal mother-cells, much in the-way suggested above, whilst
others of the primitive entodermal cells are directly formed
from mother-cells which bodily migrate inwards.
Fig. 65, PI. 6, represents a small portion of the formative
region of an ’04 vesicle viewed fPom the inner surface. In
the centre of the figure, surrounded by the larger, lighter
staining (ectodermal) cells of the wall, is a smaller cell in the
telophases of division, the cytoplasm of which is granular and
stains deeply. That cell unmistakably forms a constituent of
the unilaminar wall. I regard it as an entodermal mothercell. Fig. 66 shows another cell of the same character in the
anaphases of division, which likewise forms a constituent of
the unilaminar wall, but which differs from the corresponding
cell in fig. 65 in that its cytoplasmic body has extended out
on one side (lower in the figure), so as to directly underlie
part of an adjacent ectodermal cell. In other words we have
here a surface view of the condition represented in section in
fig. 75, only the entodermal mother-cell depicted therein is not
actually in process of division. Fig. 67, taken from the same
preparation as fig. 65, shows what I take to be the end result
of the division of such a cell as is i-epresented in the two
preceding figures. Here we see two small deeply staining
cells towards the centre of the figure, which from their disposition and agreement in size and cytological characters
are manifestly sister-cells, and the products of division of
just such an entodermal mother-cell as is represented in fig.
65, or, better, fig. 66. The one cell (upper in the figure) is
more angular in form and manifestly still lies in the unilaminar wall; the other (lower in the figure) is ovalish in form
and is no longer a constituent of the unilaminar wall, but is
on the contrary a free cell, definitely internal both to the
J. P. HILL. â–  . â–  :
latter and to its sister-cell. It is, in fact, a primitive entodermal cell, as comparison with fig. 68 proves, and that it has
been formed by the division of a mother-cell situated in the
unilaminar wall can hardly, I think, be doubted. Its sistercell, which is still a constituent of the wall, would presumably
have migrated inwai-ds some time later.
It is to be noted that the primitive entodermal cell referred
to above and those depicted in figs. 71 and 72 are definitely
contoured, ovalish and I’ounded cells, entirely devoid of processes. In these respects they differ markedly from the entodermal cells shown in fig^. 68-70, which are very variable in
form owing to their possession of more or less elongated
pseudopodial-like processes. It might thex’efore be inferred
that the formation of these processes only takes place after
the entodermal cells have become definitely internal. Such
an inference, however, would be incorrect, for I have abundant
evidence showing that such processes may be given off from
the entodermal mother- cells whilst they are still constituents
of the wall. In in toto preparations, it is often difficult to
determine with certainty whether a particular entodermal cell
still enters into the constitution of the unilaminar wall or not.
In the portion of the formative region of a ’04 vesicle depicted
in fig. 70, however, I am satisfied that all the entodermal
cells therein shown (they are readily distinguishable by their
smaller size and more deeply staining character) are, with the
possible exception of the one on the extreme right, at least
partially intercalated between the larger ectodermal cells of
the wall. Some of them are entirely situated in the wall ;
others have extended inwards in varying degree so as to
partially underlie the ectodermal cells. It is these latter
entodermal cells in particular which exhibit the cytoplasmic
processes above referred to. As the figure shows, these processes have all the characters of pseudopodia,; they vary in
size, form, and number from cell to cell, individual processes
may be reticulate and their finer prolongations may anastomose with those of others, and they are formed of cytoplasm,
less dense and rather less, deeply staining than that of the
cell-bodies from which they arise. Attention may be specially
directed to the cell towards the left of the hgure (mai'ked ent.).
Here we have an entodermal cell whose cytoplasmic body is
evidently still partially intercalated between the cells of the wall,
but which is, at the same time, prolonged inwards (towards
the left) so as to underlie the adjoining ectodermal cell.
From this inward prolongation there are given off two slender
processes, one short and tapering, the other very much
longer ; this latter, after becoming vei’y attenuated, gradually
widens to form an irregular fan-shaped expansion, suckerlike in appearance, and produced into several slender
threads, which is situated adjacent to the nucleus of
the ectodermal cell on the extreme left. Then from the
right side of the same cell there is given off a small inwardly
projecting bulbous lobe which may well be the start of just
such another process as arises from the left side. Processes
of the peculiar sucker-like type just described, formed of a
slender elongated stem and a distal expanded extremity from
which delicate filamentous prolongations are given off, are
abundantly met with in preparations, and strikingly recall the
pseudopodia of various Ehizopoda. They are seen in connection with other entodermal cells in fig. 70, and with many
of those in fig. 68. I regard them as veritable pseudopodia.
Towards the right side of fig. 70 the two entodermal cells
there situated stand in direct protoplasmic continuity by
means of two slender connecting threads, whilst the upper of
these two cells is again joined by a very fine process to the
irregular pseudopodial expansion which arises from one of
the two entodermal cells situated nearer the middle of the
figure, and that same expansion is directly connected with the
second of the two entodermal cells just mentioned, so that we
have here established the beginning of a cell-network, prior
to the complete emancipation of its constituent entodermal
elements from the unilaminar wall. We have, then, clear
evidence that the entodermal elements in Dasyurus, prior to
their separation from the unilaminar formative region ai*e
capable of exhibiting amoeboid activity, since not only may
J. r. HILL.
they send lobose prolongations of their cytoplasmic bodies
inwards below the adjacent ectodermal cells, but they may
emit more or less elongated processes of indubitable pseudopodial character, which similarly lie in contact with the inner
surface of the wall-cells. Furthermore, we have evidence
that these pseudopodial processes may anastomose with each
â– other so as to initiate the formation of an entodermal reticulum,
whilst the cells from which they arise are still constituents of
the unilaminar wall  -  an especially noteworthy phenomenon.
Certain of the primitive entodermal cells, as we have seen,
are at first devoid of such processes, but since they all
eventually form part of a continuous reticulum, it is evident
that the entodermal elements are capable of emitting pseudopodial processes as well after as before their separation from
the formative region.
Finally, in view of the fact that the entodermal mothei'-cells
depicted in fig. 70 are not actually in process of division, and
therein differ from those of figs. 65 and 66, we may conclude
that the formation of the primitive entodermal cells is effected
either with or without the pi*evious division of the mother-cells.
If Ave admit, as I think on the evidence we must admit,
that the entodermal cells in Dasyurus are endowed with
amceboid properties, then Ave are relieved of any further
difficulty in regard to the mechanism of their inAvard migration
from the unilaminar Avail. Doubtless, in the case of those
entodermal mother-cells Avhich do not undergo division, the
precocious formation of the above-described pseudopodial
processes which spread out from the cells like so many
suckers considerably facilitates their direct detachment from
amongst the cells of the Avail. In the case of those primitive
entodermal cells Avhich originate as the direct products of
division of the mother-cells, it no doubt depends on a variety
of circumstances (e.g. actual form of the dividing cell,
direction of the spindle, etc.) whether they exhibit amoeboid
activity precociously (i.e. before their actual i separation), or
only at a later period.
The entoderm varies considerably in its degree of diffe
rentiation iu different vesicles of this stage, and even in
different parts of the formative region of one and the same
vesicle. In some vesicles there are relatively few primitive
entodermal cells, in othei*s they are much more abundant.
Fig. 68, from the formative region of an ^04 vesicle, shows a
typical patch of them and illustrates very well the highest
stage of differentiation which they attain in these vesicles. The
entodermal cells therein depicted all appear to be definitely
internal, and it is especially worthy of note that the portion
of the unilaminar wall in relation to them is composed exclusively of the larger, lighter staining cells. It is these cells
which directly form the embryonal ectoderm of the blastocysts
next to be described. The entodermal cells are obviously
amoeboid in character (obsei’ve especially the cells near the
middle of the figure), and are in active process of linking
themselves together into a cellular reticulum. In fig. 69 is
shown a small portion of the formative region of another ^04
vesicle. A single entodermal mother-cell in process of
division occurs in position in the unilaminar Avail, which is
otherwise composed of ectodermal cells, whilst internally there
are present three entodermal cells, already linked together by
their pseudopodial processes. ^Jfiie two lowermost cells afford
especially striking examples of amoeboid activity, the elongated
pseudopodial process of the cell on the left terminating iu a
well-marked reticulation in definite continuity Avith the corresponding, but shorter and thicker process of the cell on the
3. Establishment of the Definitive Embryonal
FolloAving directly on the stage represented by the ’04
blastocysts described in the preceding section is one designated in my list as 5, 18 . vii . 01 and referred to here as 5, ’01.
It comprises twenty-two blastocysts obtained from a female
killed fifteen days after coition and all normal, Avith the
exception of one Avhich Avas shrivelled, and all in precisely
VOL. 56, PAllT 1. NEW SERIES. 5
.T. P. HILL.
the same stag-e of development. They measured from 4‘5 to
6 mm. in diameter.
In this stage the formative region of the preceding blastocysts has become transformed into the definitive embryonal
area (embryonic shield, Hubrecht) as the result of the completion of that process of inward migration of the entodermal
mother-cells which we saw in pi-ogress in the vesicles last
described, and the consequent establishment of the entoderm
as a continuous cell-layer undeidying and independent of, the
embryonal ectoderm constituted by the larger passive cells of
the original unilaminar formative layer.
In the entii*e blastocyst (PI. 4, fig. 39) the embryonal area
is quite obvious to the naked eye as the more opaque, hemispherical region, forming rather less than half the entire
extent of the vesicle wall ; the larger remainder of the same
is formed by the much more transpai-ent, non-formative or
extra-embryonal region. Sections of the entire blastocyst
show (1) that the embryonal area is bilaminar over its entii-e
extent, its outer layer consisting of embryonal ectoderm,
already somewhat thickened, its much thinner inner layer
consisting of entoderm, partly still in the form of a cellular
reticulum, and (2) that the extra-embryonal region is still
unilaminar throughout and composed of a relatively thin
layer of flattened cells (extra-embryonal or trophoblastic ectoderm, trophoblast [Hubrecht])^ (PI. 8, fig. 78). The entoderm
is co-extensive at this stage with the embryonal ectoderm,
and terminates in a wavy, irregularly thickened, free, edge
(PI. 5, fig. 49), which over most of its extent either directly
underlies or extends very slightly beyond the line of junction
between the embryonal and extra-embryonal ectoderm. The
junctional line is thus not very easily seen. In fig. 48, however,
’ In consonance witli my conviction that this layer is homologous
both Avith the so-called trophoblast of Eutheria and the exti-a-embryonal
ectoderm of Prototheria, and in view of the theoretical signification
which Hubrecht now insists should be attached to the term “ trophoblast.” and which I am wholly unable to accept, I venture to suggest as
an alteiTiative name for this layer that of “ tropho-ectoderni.
a small portion oP the line shows with sufficient distinctness, I
think, to demonstrate its identity with that of the preceding
The vesicle wall in all my sections of this stage appears
to be somewhat thinner than that of the ’04 blastocysts, but
apart from this apparently variational difference the present
blastocysts are almost exactly intermediate between the latter
and those next to be described.
The embryonal ectoderm (fig. 78, emb. ect.) appears in
section fairly uniformly thickened, though its cells are still of
the flattened type. In surface view in in toto preparations
(cf . fig. 48), they exhibit the same polygonal form and lightly
staining qualities as the larger cells of the formative region
of the ’04 blastocysts, which we have already identified as
prospective embryonal ectodermal cells. The junctional line
between the embryonal ectoderm and the extra-embryonal is
now for the first time readily distinguishable in sections
(fig. 78). The extra-embryonal ectoderm (tropho-ectoderm)
(PI. 5, figs. 48 and 49, PI. 8, fig. 78, tr. ect.) differs in no
essential respect from the corresponding layer in the ’04
The entoderm in these blastocysts is exceedingly closely
adherent to the inner surface of the embryonal ectoderm and
cannot be removed therefrom by artificial means. It varies
slightly in its character in different vesicles and in different
parts of its extent in the same vesicle. Mostly it appears as
a continuous thin cell-layer (figs. 49 and 78, ent.), but here and
there patches occur in which the cells form a reticulum quite
similar to that shown in fig. 68 of the preceding stage.
The next stage (designated in my list as 8 . vi . 01), and the
last of Dasyurus that need be described in the present communication, comprises eleven vesicles (5-5'5 mm. in diameter),
in which the embryonal area is conspicuous and distinctly in
advance of that of the preceding vesicles, but is still devoid
of any trace of embryonal differentiation (PI. 4, fig. 40;
PI. 8, fig. 79).
The embryonal area is hemispherical in form (its greatest
diameter varying’ from 3'5 to 4 mm.) in all except two of the
blastocysts, in which it is elongate, with longer and shorter
diameters. It occupies about a third or less of the entire
extent of the vesicle wall, and thus appears relatively smaller
than that of the preceding (.5, ’01) vesicles. The entoderm now
extends for a distance of about 1 mm. beyond the limits of
the area, so that in the entire vesicle (fig. 40) three zones
differing in opacity are distinguishable, viz. the dense hemispherical zone at the upper pole, constituted by the embryonal
area; below that, a less dense, narrow annular zone, formed of
extra-embryonal ectoderm and the underlying peripheral
extension of the entoderm ; and finally, the still less dense
hemispherical area, forming the lower hemisphere of the
blastocyst and constituted, solely by extra-embryonal ectoderm. Thus approximately the upper half of the blastocyst
is bilaminar, the lower half unilaminar. Sections show that
the embryonal ectoderm (fig. 79, emh. ect.) is now a quite
thick layer of approximately cubical cells, whilst the extraembryonal ectoderm {tr. ect.) is formed of relatively thin
flattened cells. The line of junction between the two is perfectly obvious, both in sections (fig. 79) and in surface view
(PI. 5, fig. 50). The embryonal ectodermal cells, though
much thicker than the extra-embryonal, are of less superficial
extent; their nuclei therefore lie closer together than those
of the latter, moreover they are larger, stain more deeply, and
are more frequently found in division, all of which facts
testify to the much greater growth-activity of the embryonal
as compai’ed with the exti-a-embryonal ectoderm at this stage
of development (cf . fig. 50, emh. ect. and tr. ect.-, in the preparation from which this micro-photograph was made the entoderm
underlying the embryonal ectoderm has been removed, whilst
it is still partially present over the extra-embryonal ectoderm).
The entoderm (fig. 79, ent.) over the region of the embrvonal area is readily separable as a quite thin membrane,
and is then seen to consist of squamous cells, polygonal in
outline, and either in direct apposition by their edges or connected together by minute cytoplasmic processes. Beyond the
embryonal area, liowever, its peripliei'al extension below the
extra-embryonal ectoderm is much less easily separable in the
intact condition (cF. fig. 50), because oF its greater delicacy
due to the fact that it has here largely the form of a cellular
reticulum. In this extra-embryonal region the entodermal
cells are frequently found in mitosis. Ic would appear, then,
that the entoderm is first laid down in the region of the embi'yonal area as a cellular reticulum, which later becomes
ii’ansformed into a continuous cell-membrane, and that its
peripheral extension over the inner surface of the extraembryonal ectoderm is the result of the growth and activity
oF its own constituent cells.
This peripheral growth continues until there is formed
eventually a complete entodermal lining to the blastocyst
cavity. The rate of growth appears to be somewhat variable.
In a series of primitive streak vesicles (6-6'75 mm. in diameter)
the lower third oF the wall is, I find, still unilaminar. In
another series of vesicles of the same developmental stage
(4'5-6 mm. in diameter) a unilaminar area is present at the
lower pole, varying from I x ‘5 mm. in diameter to as much
as 4 mm. Even in vesicles 7-7'5 mm. in diameter a unilaminar patch may still occur at the lower pole, but in vesiqles
8'5 mm. in diameter (stage of fiat embryo) the entodermal
lining appears always to be complete.
The Origin of the Entoderm in Eutheria.  -  The
remai’kable facts relative to the origin of the entoderm in
Dasyurus which I have been able to place on record in the
jireceding pages, thanks to the large size attained by the
blastocyst prior to the differentiation of the formative germlayers and to the circumstance that the formative cells are
not arranged, as they are in Eutheria, in the form of a more
or less compact cell-mass, but constitute a thin unilaminar
cell-layer of relatively great extent which can easily be cut
up with scissors, and which, after staining and mounting on
the fiat can be examined under the highest powers, throw, it
seems to me, a new and unexpected light on the mammalian
entoderm, and at the same time help to fill the considerable
gap whicli has hitherto existed in our knowledge of its early
ontogenesis. Although the mode of origin of the entoderm
in Dasyurus would appear, in the present state of our knowledge, to find its closest parallel, not amongst vertebrates, but
in certain invertebrates (cf . the mode of origin of the entodermal cells from the wall of the blastula in Hydra as
described by Brauer^), the observations of Assheton (’94)
on the early history of the entoderm in the rabbit, when
viewed in the light of the foregoing, seem to me to afford
ground for the belief that phenomena comparable with those
hei’e recorded for Dasyui’us will eventually be recognised as
occurring also in Eutheria.
Hubrecht (’08), in his recent treatise on early Mammalian
ontogeny, deals very briefly with the question of the origin
of the entoderm in the latter group, merely stating that
“ from the inner cell-mass arises by delamination a separate
lower layer which we designate as the entoderm of the
embryo. These entoderm cells wander in radial direction
along the inner surface of the trophoblast, which in many
cases is thus soon transformed into a didermic structure.
. . . When the entoderm has separated off by delamina
tion from the embryonic knob, the remaining cells of the
latter form the ' embryonic ectodei'm,’ which is thus situated
between the entoderm and the trophoblast.”
Assheton, in the paper just referred to, has given a careful
account of the first appearance of the entodermal cells in the
rabbit, and of what he believes to be the mode of their
peripheral extension below the trophoblastic wall of the
blastocyst. He shows that the inner cell-mass, at first
spherical, gradually, as the blastocyst enlarges, fiattens out
below the “ covering layer ” of the trophoblast until it forms
an approximately circular plate “ nowhere more than two
cells thick.” During the process of flattening, cells are seen
to jut out from the periphery of the mass; these eventually
separate, and appear as rounded cells scattered irregularly
over the inner surface of the trophoblast and ‘^extending
' ‘ Zeitschi’. f. wiss. Zool.,’ Bd. Hi, 1891.
over an arc of about 60° from the upper pole in all directions.”
These “ straggling” cells, as Assheton terms them, as well as
the innermost cells of the now flattened inner cell-mass, are
regarded as hypoblastic and the outermost cells of the same
as epiblastic (embryonic epiblast). “The hypoblast, as a
perfectly definite layer, is formed by the time the blastodermic vesicle measures '5 mm. in diameter, that is, about the
102nd hour after coition. It is not, however, as yet by any
means a continuous membrane ; it is a network or fenestrated
membrane. For this reason, in section it appears to be
represented by isolated cells lying beneath the embryonic
disc (v. fig. 29, Sy.)” (cf. Dasyurus). In considering the
question how the peripherally situated (“ straggling ”) entodermal cells, which are undoubtedly derived from the inner
cell-mass, “apparently Avander round the inside of the blastodermic vesicle,” he I’eaches the conclusion that this is not the
result of amoeboid activity or growth “in the sense of migration ” on the part of these cells, but “ is only an apparent
growth round produced by the more rapid growth of a
zone of the [trophoblastic] wall of the vesicle immediately
surrounding the embryonic disc, in which zone the marginal
cells of the inner mass lie.” He is unable to find any
evidence of the production of pseudopodial processes by
these pei’ipheral entodermal cells, the majority of them
appearing at first to be quite isolated from each other and
approximately spherical. “Certain of the cells here and
there are connected by threads of protoplasm, but this, I
think, is not a sign of pseudopodic activity, but merely
indicates the final stage in division betAveen the tAvo cells.”
By the sixth day the hypoblast of the embryonic disc has
assumed the lorm of a continuous membrane, composed of
completely flattened cells, Avhilst the peripheral hypoblast
cells have become more numerous, and “many of them,
possibly all of them, are noAV undoubtedly connected by more
or less fine protoplasmic threads.” Such, in brief, is
Assheton’s account of the early history of the entodenn in
the rabbit; it presents obvious points of agreement with my
own for Dasyunis, and I ventui’o to think the agreement is
even greater than would appear from Assheton’s conclusions.
In adopting- the view that the more active growth of the
region of the blastocyst wall immediately surrounding the
inner cell-mass is the sole causal agent in effecting the separation and peripheral spreading of the entodermal cells, I cannot
but feel, in view of his own description and figures and of my
own results, that he has attributed a much too exclusive importance to that phenomenon and a much too passive role to the
entodermal cells themselves. In Dasyurus the inward migration and the later peripheral spreading of the entodermal
cells is effected without any such marked unequal growth of
tlie blastocyst wall as occurs, according to Assheton, in the
rabbit, as the direct outcome of their owu inherent activity,
and I believe the possession of a like activity characterises
the entodermal cells of the rabbit. The evidence of Assheton’s
own fig. 40, which shows in surface view a portion of the
vesicle wall with the peripheral entodermal cells in relation
thereto, and which should be compared with my figs. 68 and
69, conclusively demonstrates, to my mind, the possession by
these cells of amoeboid properties, and thus support is
afforded for the belief that the separation of the entodermal
cells from the formative cell group (inner cell-mass) is here
also the expression of an actual migration. Whether or not
the strands of protoplasm which Assheton (’08, '09) describes
as present in the sheep, pig, ferret, and goat, connecting the
inner lining of the inner mass to the wall of the blastocyst,
and which he interprets as tending “ to show that the inner
lining of the inner mass is of common origin with the wall of
the blastocyst,” are of any significance in the present connection, I cannot certainly determine.
4. Summary.
The results and conclusions set forth in the preceding
pages of this chapter may be summarised as follows;
(1) The unilaminar wall of the blastocyst of Dasyurus con
sists of two regions distinct in origin and in destiny, viz. an
upper or formative region, derived from the upper cell-ring
of the 16-celled stage, and destined to furnish the embryonal ectoderm and the entoderm and a lower or nonformative region derived fi-om the lower cell-ring of the
mentioned stage, and destined to form directly the extraembryonal or trophoblastic ectoderm (tropho-ectoderrn) of the
bilaminar vesicle.
(2) The formative region, unlike the non-formative, is
constituted by cells of two varieties, viz. : (i) a more
numerous series of larger, lighter-staining' cells destined
to form the embryonal ectoderm, and (ii) a less numerous
series of smaller, more granular, and more deeply staining
cells, destined to give origin to the entoderm and hence
distinguishable as the entodermal mother-cells.
(3) The entodermal mother-cells, either without or subsequently to division, bodily migrate inwards from amongst the
larger cells of the unilamiuar wall and so come to lie in
contact with the inner surface of the latter. Tkey thus give
origin to the primitive entodermal cells from which the
deKnitive entoderm arises. The larger passive cells, which
alone form the unilamiuar wall after the inward migration of
the entodermal cells is completed, constitute the embryonal
(4) The entodermal cells as well before as after their
migration from the unilamiuar wall are capable of exhibiting
amoeboid activity and of emitting pseudopodial processes, by
tlie anastomosing of which there is eventually formed a
cellular entodermal reticulum underlying, and at first coextensive with, the embryonal ectoderm.
(5) The definitive entoderm thus owes its character as a
connected cell-layer primarily to the formation of secondaiy
anastomoses between the pseudopodial processes emitted by
the primitive entodermal cells (or entodermal mothercells).
(6) The assumption by the entodermal cells of amoeboid
j^roperties whilst they are still constituents of the unilamiuar
wall affords an intelligible explanation of the mechaiiisin of
their inwai'd migration.
(/) The entoderm is first laid down below the formative or
embryonal region of the blastocyst; thence it extends gradually by its own growth round the inner surface of the uuilaramar non-forrnative region so as to form eventually a
complete entodermal lining to the blastocyst cavity. In this
way the blastocyst wall becomes bilaminar throughout.
(8) The bilaminar blastocyst consists of two reguous, respectively embryonal and extra-embryonal. The embryonal
region (embryonal area) is constituted by an outer layer of
embryonal ectoderm and the underlying portion of the entoderm, and the extra-embryonal, of the extra-embrvonal or
trophoblastic ectoderm (tropho-ectoderm), which is separated
from the embryonal by a well-marked junctional line, together
with the underlying portion of the entoderm, which is perfectly continuous with that below the embryonal ectoderm.
(9) The formative or embryonal region of the blastocyst
in Dasyurus is from the first freely exposed, and at no time
daring the developmental period dealt with in this paper
does there exist any cellular layer externally to it, i. e. a
covering layer of trophoblast (Deckschicht, Kauber’s layer)
is absent and there is no entypy of the primary germ-layers
(cf. p. 111).
Chapter V.  -  Some Early Stages op Perameles and
The early material of Perameles and Macropus at my
disposal comprises only a small number of stages, but is of
special importance, since it enables me to demonstrate that
so far as these particular stages are concerned, the early
developmental phenomena in these forms are essentially the
same as in Dasyurus, and thus affords ground for the belief
that there is one common type of early development throughout the series of the Marsupialia. Moreover, it is of interest
since it reveals the. existence of what might be termed
specific differences in the early development of these Marsupials, especially in regard to the time of appearance of the
entoderm. In Dasyurus, it will be remembered, the primitive
entoderm cells first become definitely recognisable as internally situated cells in vesicles 4‘5 mm. in diameter. In
Perameles they occur in vesicles just over 1 rnm. in diameter,
while in Macropus they are already present in a blastocyst
only -35 mm. in diameter, so that it would appear that the
entoderm is differentiated much earlier in the higher, more
specialised types than in the more generalised forms. This
difference in time of appearance of the entoderm is perhaps
to be correlated with a difference in size of the ovarian ova
in the three genera mentioned.
1. Perameles.
The earliest material of Perameles I possess consists of two
eggs of P. obesula, which I owe to the skill and enthusiasm
of my friend Mr. S. J. M. Moreau, of Sydney. Egg -Ameasures '23 mm. in diameter, and egg B, ‘24 x ‘23 mm.
The former consists of thirty-two cells, the latter of thirty. In
both the shell-membrane has partially collapsed, but the general
plan of arrangement of the blastomeres can still fairly readily
be made out. Fig. 51, PI. 3, represents a micro-photograph
of a section of egg B, the better of the two. It shows the
.shell-membrane (nearly '005 mm. thick) externally, considerable remains of the albumen between that and the
deeply stained zona, and then, closely applied to the inner
surface of the latter, the blastomeres arranged in the form of
an inverted D, so as to enclose a central space, open below
as the figure stands. This latter opening extends through
the series, and it seems probable that there was a corx*esponding one opposite to it in the intact egg. Evidently we
have hei’e a stage in the formation of the blastocyst, in which
the blastomeres are in course of spreading towards one or
both of the poles of the sphere formed by the egg-envelopes.
J. r, HILL,
â– just as liappeus in the corresponding' stage of Dasyurus (cf.
fig. 51 with fig. 31j though the latter represents a somewhat
older stage in Dasyurus). The blastocyst-wall here appears
relatively more extensive than in the 32-celled stage of
Dasyurus, an apparent difference which may perhaps be accounted for by the difference in size of the respective eggs
(•24 mm. as compared with '36 mm.) . The blastomeres situated
adjacent to the opening and those on the right side of the
figure tend to be more flattened and of greater superficial extent than the remainder, but I can recognise no
difference in the cytological characters of the cells. The
space or cleavage cavity enclosed by the blastomeres is partly
occupied by a granular coagulum, and towards the opening
there is present a lightly staining reticular mass, which
i*ecalls the yolk-body of Dasyurus, though I am not prepared
to affirm that it is of that significance. The fixation of the
specimen is not quite perfect.
My next stage of Perameles is constituted by a blastocyst
of P. nasuta, for which I am again indebted to Mr. Moreau
measuring in the preserved condition ’29 x '26 mm. Pig. 52,
PI. 3, shows a section of this blastocyst. Structurally,
it corresponds in all essential respects with the '43 mm.
blastocyst of Dasyurus, figured on the same plate (fig. 33).
The blastocyst Avail is complete and unilamiuar throughout.
It is distinguishable into tAvo regions, a more extensive region
over Avhich the cells are large and flattened and a less extensive,
composed of smaller but thicker cells (left side of fig. 52).
In the early blastocysts of Dasyurus, it may be recalled, the
evidence showed that the region of more flattened cells is
formative in significance, that of more bulky cells, non-formative. It is possible the same holds good for this Perameles
blastocyst. On the other hand, the structural condition of
the stage next to be described rather supports the vieAv that
the smaller region, composed of plumper cells, is in this case
formative. That view seems to me the more probable of the
two, but there is a considerable difference in size betAveen the
present blastocyst and those next available, so that it is
impossible to decide this point witli certciinty. The blastocyst cavity is partly occupied by coagulnm. There are no
cells present in it, but the question of the presence of a yolkbody must remain open. The shell-membrane (‘0045 mm. in
thickness) and zona are in close apposition.
Following this early blastocyst, I have three vesicles of
P. nasuta, two of them measuring 1‘3 mm. in diameter,
the other PI mm. In their stage of development they
agree pretty closely with the 4'5-5 mm. vesicles of Dasyurus,
referred to in the preceding pages under the designation
6, '04, the entoderm being in process of differentiation. The
formative region was readily distinguishable in the intact
vesicles as a darker patch occupying about three eighths of
the surface extent of the wall. In section (PI. 8, figs. 80, 81)
it is characterised by its greater thickness as compared with
the non-formative or trophoblastic region, and by the
presence below it of numbers of primitive entodernial cells.
Compared with the corresponding stage in Dasyurus, the
chief difference consists in the relatively much greater thickness of the cells of the formative region in the Perameles
vesicle. The latter cells are here already more or less definitely cubical in shape, their thickness varying from '09
mm. to ’015 mm., and altogether they form a layer of a much
more uniformly thickened character than that of the 6, '04
vesicles of Dasyurus. The trophoblastic ectoderm (figs. 80,
81, tr. ect.) is composed of somewhat flattened cells, .varying
in thickness from ‘005 to ’008 mm.
The primitive entodermal cells (figs. 80, 81, ent.) are
present below the formative region in fair abundance, more
especially around the periphery of the same, which may thus
appear somewhat thickened (fig. 81). 4’he cells vary in size
from ‘01 X '007 mm. to ’024 x '009 mm., and they stain on the
whole somewhat more deeply than the formative cells, to
whose under-surface they are closely applied. They occur
groups. Mitotic figures are frequently met
with in the cells of the formative ai’ea (observe the obliquely
disposed figure in one of the formative cells in fig. 81), and
J. r. HILL.
■they also occur in the primitive entodermal cells. Examination of the sections leaves no doubt in one’s mind as to the
source of the entodermal cells. They are undoubtedly derived
from the formative region of the vesicle wall. The 'shellmembrane has a thickness of about '0027 mm.
2. Macro pus.
Of Macropus the earliest stage I have examined is a blastocyst of M. ruficollis, -25 x *21 mm. in diameter. It is not
in a quite perfect state of preservation, but is in a sufficiently
good condition to enable me to say that the wall is complete
and unilaminar throughout, just as in the ‘29 x "26 mm.
blastocyst of Perameles. The shell-membrane has a thickness
of about -005 mm., and there are still remains of the albumen
between it and the zona.
My next stage (figs. 82-85) is a blastocyst of the same
species, *35 mm. in diameter. It unfortunately suffered in
preparation, but practically the whole of the formative area
of the blastocyst wall and part of the trophoblastic ectoderm
are comprised in the sections (PI. 9, fig. 82), so that it is still
possible to make out its chief structural features. In its stage
of development this blastocyst closely agrees with the last
described blastocysts of Perameles. The formative area of
the wall is perfectly distinct in the sections because of its
greater thickness and the presence below it of the primitive
entodermal cells. It attains its gi-eatest thickness (*027 mm.)
peripherally, whilst it is thinnest centrally (*006 mm.), so that,
taken as a whole, it is not quite such a uniformly thickened
layer as is that of the Perameles blastocysts. Primitive entodermal cells are present below it, but not in great abundance
(figs. 82, 84, 85, ent.). In fig. 83, a formative cell is seen in
division, the axis of the spindle being oblique to the surface.
The trophoblastic ectoderm (figs. 82, 83, tr. ect.) is composed
of the usual flattened cells, and varies in thickness from
*005 to *0067 mm.
In the blastocyst cavity, adjacent to the trophoblastic
ectoderm on the left side of fig. 82, there is visible a small
spherical cell similar to the degenerate cells met with in
blastocysts of Dasyurus.
My last stage of M. ruficollis comprises an excellently
preserved blastocyst, measuring '8 mm. in diameter, in which
the embryonal ectoderm and the entoderm are definitely
established. It thus corresponds to the 8, ’01 stage of
Dasyurus (blastocysts o -  5'5 mm. diameter). The embryonal
area is circular and measures '468 mm. in diameter. Its
constituent cells are cubical and from ’008 to ‘OlS mm. in
thickness, Avhilst the trophoblastic ectoderm is formed of
flattened cells, -006 ram. in thickness. The entoderm is
present as a continuous layer of attenuated cells below the
embryonal ectoderm, and it probably also forms a continuous
layer below the trophoblastic ectoderm. Entodermal cells are
certainly pi*esent over the lower polar region of the vesicle,
but it is difficult to be certain from the sections whether or not
they form a perfectly continuous layer. The shell membrane
has a thickness of •0026 mm.
I have a corresponding blastocyst of Petrogale penicillata •915 mm. in diameter, with an oval, embryonal area
•525 X ^45 mm. in diameter, and a later blastocyst of M.
ruficollis P46 mm. in diameter, with a circular embryonal
area ’57 mm. in diameter.
Chapter VI.  -  General Summary and Conclusions.
The observations recoi'ded in the pi’eceding pages and the
conclusions deducible therefrom may be summarised as
follows ;
(a) Ovum.  -  The uterine ovum of Dasyurus is characterised
(1) by its large size relatively to those of Eutheria; (2) by
the presence externally to the zona of a layer of albumen and
a shell-membrane, both laid down in the Fallopian tube and
homologous with the corresponding structures in the Mouotreme ovum, the shell-membrane, like the shell of the latter,
inci’easing in thickness in the uterus; (3) by its marked
J. r. HILL.
polarity, its lower two thirds consisting of formative cytoplasm, dense and finely granular in appearance, owing to the
presence of fairly uniformly distributed deutoplasmic material,
and containing the two pronuclei, its upper third being
relatively clear and transparent, consisting as it does of a
delicate reticulum of non-formative cytoplasm, the meshes of
which are occupied by a clear deutoplasmic fluid. Study of
the pi'ocess of vitellogenesis in ovarian ova demonstrates that
this fluid represents surplus deutoplasmic material which has
not been utilised in the upbuilding of the formative region of
the ovum.
The fate of the clear non-formative portion of the ovum is
a very remarkable one. Prior to the completion of the first
cleavage, it is separated off from the formative remainder of
the ovum as a spherical mass or yolk-body, Avhich takes no
direct part in development, though it becomes enclosed iu the
blastocyst cavity on completion of the blastocyst wall at the
upper pole. Its contained deutoplasmic fluid is to be regarded
as the product of an abortive attempt at the formation of a
solid yolk-mass, such as is found in the Monotreme ovum.
By its elimination the potentially yolk-laden telolecithal ovum
becomes converted into a secondarily homolecithal, holoblastic
one. All the evidence is held to support the conclusion that
the Marsupials are descended from oviparous ancestors with
ineroblastic ova.
(b) Cleavage.  -  Cleavage begins in the uterus, is total, and
at first equal and of the radial type. The first two cleavage
planes are meridional and at right angles to each other.
The resulting four equal-sized blastomeres lie disposed radially
around the polar diameter like those of the Monotreme (not
in pairs at right angles to each other as in Eutheria), and
enclose a segmentation cavity open above and below, their
upper ends partially surrounding the yolk-body. The third
cleavage planes are again meridional, each of the four blastomeres becoming subdivided equally into two. The resulting
eight cells form an equatorial ring in contact with the inner
surface of the sphere formed by the egg-envelopes. They
contain deutoplasmic material, which is, however, located
mainly in their lower halves. The ensuing fourth cleavages
are equatorial, and in correlation with the just-mentioned
disposition of the deutoplasm, are unequal and qualitative,
each of the eight blastoraeres becoming subdivided into an
upper smaller and clearer cell, with relatively little deutoplasm fairly uniformly dispersed through the cytoplasm, and
a lower larger, more opaque cell Avith much deutoplasm,
mainly located in a broad zone in the outer portion of the
cell-body. A 16-celled stage is thus produced in which the
blastomeres are characteristically arranged in two superimposed rings, each of eight cells, an upper of smaller, clearer
cells next the yolk-body, and a lower of larger, denser cells.
The former is destined to give origin to the formative or
embryonal region of the blastocyst wall, the latter to the
non-formative or extra-embryonal region of the same.
(c) Formation of the Blastocyst.  -  There is in the
Marsupial no morula stage as in Butheria, the blastomeres
proceeding directly to form the wall of the blastocyst. The
cells of the two rings of the 16-celled stage divide at first
meridionally and then also equatorially, the division planes
being always vertical to the surface. The daughter-blastomeres so produced, continuing to divide in the same fashion,
gradually spread towards opposite poles in contact with
the inner surface of the fii-m sphere formed by the zona and
the thickened shell-membrane. Eventually they form a complete cellular lining to the said sphere and it is this which
constitutes the wall of the blastocyst. The latter is accordingly unilaminar at its first origin, and it remains so in
Dasyurus until it has attained, as the result of active gi’owth
accompanied by the imbibition of fiuid from the uterus, a
diameter of 4-5 mm. It consists of two parts or regions,
distinct in origin and in destiny, and clearly marked off from
each other in later blastocysts by a definite junctional line
approximately equatorial in position, viz. an upper, embryonal
or formative region derived from the upper cell-ring of the
16-celled stage, and a lower, extra-embryonal or nonVOL. 56, PART 1. NEW SERIES. 6
formative region derived from tlie lower cell-ring of the same
(d) Later History of the Two Region s of the Blastocyst Wall (for details see pp. 72-74). -  From the embryonal
region are derived the embryonal ectoderm and the entire
entoderm of the vesicle. I conclude^ therefore, that it is tlie
homologue of the inner cell-mass or embryonal knot of the
Eutherian blastocyst. The extra-embryonal region directly
furnishes the outer extra-embryonal layer of the vesicle wall,
i. e. the outer layer of the omphalopleure and chorion of later
stages. Assuming, as the facts of comparative anatomy and
palaeontology entirely justify us in doing, that the Mammals
are monophyletic and of reptilian origin, and further assuming
that the foetal membranes are homologous structures throughout the Amniotan series (also in my view a perfectly
justifiable assumption)^, then the homologies of this extraembryonal region of the Marsupial blastocyt are not far to
seek. It is clearly the homologue of the extra-embryonal
ectoderm of the Sauropsidan and Monotreme egg, and the
homologue also of the outer enveloping layer of the Eutlierian
blastocyst, to which Hubi’echt has given the special name of
“ trophoblast.” In my view the trophoblast is none other
than extra-embryonal ectoderm which in the viviparous
mammals, in correlation with the intra-uterine mode of
development, has acquired a special significance for the
nutrition of the embryo.
These, then, are my conclusions, and to me they seem on
general grounds perfectly obvious, viz. : (1) that the embryonal or formative region of the unilaminar Marsupial
blastocyst is the homologue of the inner cell-mass or
* How Assheton can maintain f 09, p. 266) “ that the amnion of the
rabbit is not more homologous to the amnion of the Sauropsidan than
the homy teeth of Ornithorhynchns ai-e homologous to the true teeth
of the mammal or reptile, which they have supplanted,” how he can
hold this view and yet proceed to utilise the presence of the amnion as
one of the leading charactei-s distinguishing the Amniota from the
Anamnia, I fail to comprehend. Surely the presence of a series of
purely analogous structures in a group is of no classificatory value.
imposed cell-viiigs, respectively and non-formative cell-rings of
formative (emlDryonal) and non- the Metatherian.
formative (extra-embryonal) in
O CJ r
^ ct CO <D o
‘7i ce O r
,0 c 2 «w
^ >
' W ^ ^*“1 ^
> -p «f-i ^ O ^
c! ^ 03 . d ^
CD 03 d So,
2^.2 «
>> o o
S -M
^ 2
^ 2
O • -  ' pO
^ g-r 5 s
d ^ «4-j
L-r 03
'O ^-g_fe^.S
rj 03 4 ^ C -P
•S ^ S:^ aj'5 ^
_ -g U3 ri 2< § ®
(D cu 5 i
d d 3 ^ 2 ■‘^ S
= a CO 4 j
o o ^ a ^ a
-iO Q)
O *H O 03 03
•P * 03
>% O -P
o d 3
" d ci_i
m o
(D «
-d 'p
-p c ,
/-< -U r -  •
03  - 
dJ > P ^
. "2 =2
S (T 2 a
d .!ll _d
•a a3 ^ q2
ri H) (d „
s 3 3“
•g o 0 13
ar ^ o
a ®
'§ 2
o . .
d 03 03
^ p
d I -  '
d ^ O
O 03 ^
^ 5S d
3-2 o
3 a’
0) 03
03 03 -r
r -  -U O
CO 'a
• I-H CD
a a 2
^ ?3 ^
“ ^ Jh
d o d
03 «M
> O
«w d
5-t ^
-P 03
8 g
ra •- a
3 a a>
3 aaj
® o o
^ ^ s
2 ^
§ § 2
O â– 
-5 ’’"-p
03 03 O "d
t> K d d
• d 03 O
f- ~*z o^
Q d ^ o <x>
«+H d CD , j3
2 S as 3 ^
^=2^' 5^2
O 3 Td S
CO a =2
d 03
2 !^
d CO
^ 03
03 â– 
d d ^
d d «4H
0 - O '^-l
ft i<1N
5 c
* CO
03 -P
pd M
-P 03
ni ^ ^
d ’d P
d F ^ 03
of 03 O
03 5
3 -p^
an" 2
03 ^ 03
pd «w pd .;-p O -p -p
O rc5
2 ^
,2 -p 'a
^ o
*P 5?
(H 2-2
> ^ d
to O o
ni5 ^
d ep 03
03 O
' CD
(1) <SM
60 O
S 03
- O
S=^ ;
03 '
^-2 is
(>,’3 .2
-P Ph
“ 2=m
£ 00
-a 03 -S
2 3-5
P ^ P
o ° a
0 ) g
"oj “
0 ^
1 0)
2 s
-p -P
«iM *>
o ^
CD o
60 d
•d d
:5 §
03 rj
t O .d
-2 8 .S
d -p
^ CD P
• d d i> ,
-P p -  ! -P
^ • I
CD d d
d d
03 ? r
2 o o
' «4-(
03 d
d= o
CD '
60 2
03 rd
‘r^ -P»
t>, d
03 •
o CD -d
> d
$ -p ^ ^
•L O r ■
â– 9 a
2 g a .2 i ° 'd
^ 9 ~S ^ d  - 
<D tc a S ra p £
fcc a a a a -S
^ 03 ?H d ud ^
O d
03 o
2 o
a o g 73
a a a (u
o =p o
a'a 2'^
aj a Ej P
S ’5b o
2 £ . P
d p -  ^ d
;-t 03 d * 1 ^
•Soli's a
03 3 pa
-P «+H
1 -^ P-- • -P
CO d O 03
?-• d d
-a g^J’a
a a is
an (D g a
2 ci’a"!
^ ^ g a
^ ns 3
a H a
'  -  a ^
>, a •- ra
^ 5 2 2
a ^ -a 03
£=« -^ a
.a o=w ^
na o
-a â– 5 a -2
2 & .2 a
a 0 bD p
p p a ^
O 6D P cai
^ -a
fl' _
a -d -a
a 03 a
£-5 .a
ai a
' £ -p o
a =f-i
ra a
a a
>-> a -p
a o a
o ^a
g^'o J § S
izi p a-^ o ^3
r3 O «4-j
ro ^
*“ 03
=p .3 ®
o a 03
- -p -p
a -p
.2 03 ^
§ 3 “ 2 I
a p a
• 43 =p a
d ^
d ^ o
?-• 5 o
o fe a
; a 2 2
a p -p
o o_
-p =4-1 'a
a . '
io .2 a
a a d
-a p 'â– 
03 J-i
<44 >
^•■0 a
a i °
p r 03
Cm -P
° S Bo a 'g
«W I y
a P a -p
g 2
^ 03 2
o CO i pd
d n
1=^ ;
1 p
- '’^ a
° 00 a a
03 a -a
_, bD ^ o
1i5 6“o
d a -a a
(D ^ CO P •
O ^ ^ d ,
03 P 60 o d
pd S d d 03 o
-p -4J P ^
CD CD ;-j
s 2 p
a a-j: d
*03 9 d ^ d
f[^ >d C5 03 fl-)
S .-p .-g
p ’> p
a g d
a 03 a
p "d
a â– 
6E ‘
o q a
03 o
03 03
' O
a ,B .
p^ a -fi
o a a
.2 a "p < 1 ^
a bn 2 ""
d o a -4
q 2 ^ a
a pH a
a a
^ ^-2
d-d d CO
4 . O
, o^
q -p p
Pj P. a
o -a
‘^’a 2
^ p a a
a g-2 p.^
d ^
'd d
cw M
o d
03 03
-tS 2
03 -
a bD
O g 3 -M .P
p,.d d 'w 4J Vp
t,.- 03 -p «4-l X
03 di d o 03
d CD
CD '
d 2 o
^ 04
d ^ ^
03 CD s. •
H CO ^
d 03 p :
040 rt
o E ®
.a d s
.p S a
a a
=H =p
o ;a
I * 03
q a ^
3 ^ -4J
0 gs CO
a q (H
” o p a
p= a a
B = 4 H a
"2 'a
d -p P« o
S q 3 -B
a I 3 a
â– p d a a
p=-S;g a
o S<| '•' q
3S a ®
CO d ^
,2 • d p:
-£“ a
a-d'S £
4-2 p 4 be a
d 03
embryonal knob of tlie Eutberiaii blastocyst ; and (2) that the
extra-embryonal or noii-formative region of the same is the
homologne of the extra-embryonal ectoderm of the Sauropsida and Monotremata and of the trophoblast of the
As regards conclusion (1) there is not likely to be much
difference of opinion, but as regards (2), whilst perhaps the
majority of embryologists support the obvious, not to say
common-place view which I here advocate, it seems certain
that it will prove neither obvious nor acceptable to those
mammalian embryologists (I refer specifically to my friends
Professor A. A. W. Hubrecht and Mr. R. Asshetou) who, with
only Selenka^s account of eaidy Marsupial ontogeny before
them, have formulated other and quite divergent views as to
the morphological nature of the outer enveloping layer of the
Eutherian blastocyst. It is therefore necessary to discuss
this question further, though I would fain express my conviction that had the observations recorded in this paper been
earlier available, much vain speculation as to the phytogeny
of the trophoblast might possibly have been avoided.
Chapter VII.  -  The Early Ontogeny op the Mammalia in
THE Light op the Foregoing Observations.
In entering on a discussion of the bearings of the results
of my study of the early development of Marsupials on
current interpretations of early Mammalian ontogeny, and
especially of the homologies of the germ-layers, I desire at
the outset to emphasise my conviction that, specialised
though the Marsupials undoubtedly are in certain features of
their anatomy, e. g. their dentition, genital ducts, and mammary apparatus, the observations recorded in the preceding
pages of this paper afford not the slightest ground for the
supposition that their early ontogeny is also of an aberrant
type, devoid of signiffcance from the point of view of that of
other mammals. On the contrary, I hope to demonstrate
that the Marsupial type of early development not only readily
falls into line with that of Eutheria, and with what we know
of the early development of the Prototheria, but furnishes
ns with the key to the correct interpretation of that extraordinarily specialised developmental stage, the Eutherian
blastocyst. In particular I hope to show that the description
which I have been able to give of the mode of formation of
the' Marsupial blastocyst, bridges in the most satisfactory
fashion the great gap which has till now existed in our
knowledge of the way in which the transition from the
Monotrematous to the Eutherian type of development has
been effected.
1. The Early Development of the Monotremata.
Our knowledge of the early development of the oviparous
mammals is admittedly still far from complete. Nevertheless
it is not so absolutely fragmentary that it can be passed over
in any general discussion of early mammalian ontogeny, and
I certainly cannot agree with the opinion of Assheton (’08,
p. 227) that from it “we gain very little help towards the
elucidation of Eutherian development.” On the contrary, I
think that the combined observations of Semon (’94), and
Wilson and Hill (’07) shed most valuable light on the early
ontogenetic phenomena in both the Metatheria and Eutheria.
I propose therefore to give here a very brief resume of the
chief results of these observers,^ and at the same time to
indicate how the knowledge of early Monotreme ontogeny
we possess, limited though it be, does help us to a better
understanding of the phenomena to which I have just
The ovum, as is well known from the observations of
Caldwell. (’87), is Reptilian in its character in all but size.
It is yolk-laden and telolecithal, the yolk consisting of
discrete yolk-spheres, and it is enclosed outside the zona
(vitelline membrane) by a layer of albumen and a definite shell.
* In so doinff I have largely utilised the phraseology of Wilson and
Hill’s paper (’07).
At the moment of entering the oviduct it has a diameter of
3‘5-4 mm. (2‘5-3 mm. according to Caldwell), and is therefore
small relatively to that of a reptile of the same size as the
adult Monotreme, but large relatively to those of other
mammals, being about twelve times larger than that of
Dasyurus, and about eighteen times larger than that of the
Cleavage is meroblastic. The first two cleavage planes are
at right angles to each other, as iii the Marsupial, and divide
the germinal disc into four approximately equal-sized cells
(Semon, Taf. ix, fig. 30). Each of these then becomes subdivided by a meridional furrow into two, so that an 8-celled
stage is produced, the blastomeres being arranged symmetrically, or almost symmetrically, on either side of a median line,
perhaps corresponding to the primary furrow (Wilson and Hill,
p. 37, text-figs. 1 and 2). Imagine the yolk removed and the
blastomeres arranged radially, and we have at once the open
ring-shaped 8-celled stage of Dasyurus. The details of the
succeeding cleavages are unknown. Semon has described a
stage of about twenty-four cells (Semon, Taf. ix, fig.31),inwhich
the latter formed a one-layered circular plate with no evidence
of bilateral symmetry, and this is succeeded by a stage also
figured by Semon (figs. 32 and 33, cf. also Wilson and Hill,
PI. 2, fig. 2), in which the blastoderm has become sevei’al
cells thick, though it has not yet increased in surface extent.
It is bi-convex lens-shaped in section, its lower surface being
sharply limited from the underlying white yolk. No nuclei
are recognisable in the latter, either in this or any subsequent
stage, nor is there ever any trace of a syncytial germ-wall,
features in which the Monotreme egg differs from the
The next available stage, represented by an egg of Ornithorhynchus, described by Wilson and Hill ('07, p. 38, PI. 2, fig.
4), and by an egg of Echidna, described by Semon (’94, p. 69,
figs. 22 and 33), is separated by a considerable gap from the
preceding, and most unfortunately so, since it belongs to the
period of commencing formation of the germ-layers. The
J. r. HILL.
cellular lens-shaped blastoderm of the preceding stage has
now extended in the peripheral direction so as to enclose
about the upper half of the yolk-mass, and in so doing it has
assumed the form, almost exclusively, of a unilaminar thin
cell-membrane, composed of flattened cells and closely applied
to the inner surface of the zona. At the embryonic pole,
however, in the region of the white yolk-bed, there are
present in the Ornithorhynchus egg a few plump cells,
immediately subjacent to the unilaminar blastoderm, but
separate and distinct from it, whilst in the Echidna egg
Semon’s figure (fig. 33), which is perhaps somewhat schematic,
shows a group of scattered cells, similar to those in the
Ornithorhynchus egg but placed considei’ably deeper in the
white yolk-bed. Unfortunately we have no definite evidence
as to the significance of these internally situated cells. One
of two possible interpretations may be assigned to them.
Either they represent the last remaining deeply placed cells
of the blastodisc of the preceding stage, which have not yet
become intercalated in the unilaminar blastodermic membrane
believed by Semon to be the condition attained in eggs of
about this stage of development, or they are cells which have
been proliferated off from this unilaminar blastoderm, to
constitute the parent cells of the future yolk-entoderm. As
regards Echidna, Semon expresses a definite enough opinion ;
he holds that these deeply placed cells actually arise by a
somewhat diffuse proliferation or ingrowth from a localised
depressed area of the blastoderm at the embryonic pole, and
that they give origin to yolk-entoderm. This interpretation
of Semon seems probable enough in view of the mode of origin
of the entoderm in the Metatheria and Eutheria. Moreover
in the next available stage, an egg of Ornithorhynchus, just
â– over 6 mm. in diameter, described by Wilson and Hill, the
blastoderm is already bilaminar throughout its extent, so that
we .might veiy Avell expect to find the beginnings of the entoderm in the somewhat younger eggs.
In the 6 mm. egg just referred to, the peripheral portion of
the utjilaminar blastoderm of the preceding stage has grown
SO as to enclose the entive yolk-mass in a complete ectodeimal
envelope, whilst iiiteimally to that a complete lining of yolkentoderm has become established. As tlie result of these
changes, and of the imbibition of fluid from the uterus, the
solid yolk-laden egg has become converted into a relatively
thin-walled vesicle or blastocyst, possessed of a bilaminar
wall surrounding the partly fluid vitelline contents of the egg.
Throughout the greater part of its extent the structure of the
vesicle wall is very simple. It consists externally of an
extremely attenuated ectodermal cell - membrane closely
adherent to the deep surface of the vitelline membrane
(zona), and within that of a layer of yolk-entoderm, composed
of large swollen cells, containing each a vesicular nucleus,
and a number of yolk-spheres of varying size. Over a small
area, overlying the white yolk-bed, however, the ectodermal
layer of the wall presents a different character to that
described above. Its constituent cells are here not flattened
and attenuated, but irregnlai'ly cuboidal in form and much
more closely packed together; moreover they stand in proliferative continuity with a subjacent mass of cells, also in
process of division. The irregular superficial layer and this
latter mass together form a thickened lenticular cake, "5 mm.
in greatest diameter, projecting towards the white yolk-bed
but separated from it by the yolk-entoderm, which retains
its character as a continuous cell-membrane. This differentiated, thickened area of the wall, situated as it is at the upper
pole of the egg, as marked by the white yolk-bed, must be
held to represent a part of the future embryonal region.
Wilson and Hill incline to regard it as in some degree the
equivalent of the “primitive plate” of Eeptiles and as the
initial stage in the formation of the primitive knot of latex;
eggs. This question, however, does ixot closely concern us
here : the point I wish to emphasise is the relative inactivity
of the cells composing the embryonal region of the blastoderm
in the Monotreme as compared with the marked activity displayed by those constituting the peripheral (extra-embryonal)
region of the same. It is these latter cells which by their
rapid growth complete the envelopment of the yolk-mass and
so constitute the lower hemisphere of the blastocyst.
Ihe bilaminar blastocyst of the Monotreme, foi'nied in the
manner indicated above, is entirely comparable with the
Marsupial blastocyst of the same developmental stage. There
are differences in detail certainly (e.g. in the characters,
time of formation, and rate of spreading of the entoderm,
in the mode of formation of the blastocyst cavity and in its
contents, in the apparent absence in the Monotreme of any
well-marked line of division between the embryonal aud extraembryonal regions of the ectoderm, in the relatively earlier
appearance of differentiation in the embryonal region in the
Monotreme as compared with the Marsupial), but the agreements are obvious and fundamental ; in particular, I would
emphasise the fact that in both the embryonal region is
superficial and freely exposed, and forms part of the blastocyst wall just as that of the reptile forms part of the general
blastoderm. Moreover, should future observations confi^rm
the view of Semon that the primitive entodermal cells of the
Monotreme are proliferated off from the embryonal region of
the unilaminar blastoderm, then we should be justified in
directly comparing the latter with the unilaminar wall of the
Marsupial blastocyst, and in regarding it also as consisting
of two differentiated regions, viz. a formative or embryonal
region, overlying the white yolk-bed, and giving origin to
the embryonal ectoderm and the yolk-entoderm, and a nonformative region which rapidly overgrows the yolk-mass so
as to eventually completely enclose it, just as does the less
rapidly growing extra-embryonal ectoderm of the Sauropsidan blastoderm.^ Meantime I see no reason for doubting
that this rapidly growing peripheral portion of the unilaminar
blastoderm of the Monotreme is anything else than extraembryonal ectoderm homogenous with that of the reptile.
Indeed, I am not aware that any embryologist except Hubrecht
thinks otherwise. Even Asshetou is, I believe, content to
* We should further he justified in concluding that the entoderm is
similar in its mode of origin in all three mammalian sub-classes.
regard the outer layer of the Monotrerae blastocyst ns
ectodermal. Hubrecht’s view is that the primitive eiitodermal
cells of Semon give origin, not to yolk-entoderm, but to the
equivalent of the embryonal knot of Eutheria, whilst the
uuilaminar blastodermic membrane itself is a larval layer
-  the trophoblast  -  that portion of it overlying the internally
situated cells representing the covering layer (Rauber’s layer)
of the Eutherian blastocyst. ‘'For this view,” remarks
Assheton [^09, p. 283), “1 can see no reason derivable from
actual specimens described and figured by those four authors”
(Caldwell, Semon, Wilson and Hill), with which criticism I
am in entire agreement, as also with the following statement,
which, so far as the Metatheria are concerned, is based on
my own results: “Neither in the Prototheria [n ] or the
Metatheria is there really any tangible evidence of a trophoblast occui*ring as a covering layer over the definitive epiblast
as in Eutheria” (p. 234).
In connection with the peripheral growth of the unilaminar
blastoderm in the Monotreme, it is of interest to observe that
this takes place, not apparently in intimate contact with the
surface of the solid yolk, as is the case with the growing
margin of the extra-embryonal ectoderm in the Saui’opsidan
egg, but rather in contact with the inner surface of the
thickened zona, perhaps as the result of the accumulation in
the perivitelline space of tiuid which has diffused into the latter
from the uterus. In other words, the peripheral growth of
the extra-embryonal ectoderm to enclose the yolk-mass appears
to take place here in precisely the same way as the spreading
of the non-formative cells in Dasyurus to complete the lower
pole of the blastocyst. In my view the latter phenomenon
is none other than a recapitulation of the former ; on the
other hand, I regard the spreading of the formative cells in
Dasyurus towards the upper pole as a purely secondary
feature, conditioned by the loss of the yolk-mass and the
attainment of the holoblastic type of cleavage.
If it be admitted that the outer extra-embryonal layer of
the Monotreme blastocyst is homogenous with the extra
embryonal ectoderm of the Keptile, then it seems to me there
is no escape from tlie conclusion that these layers are also
homogenous with the non-formative region of the unilaminar
Marsupial blastocyst. I need only point out here that the
chief destiny of each of the mentioned layers, and I might
also add that of the outer enveloping layer of the Eutherian
blastocyst (the so-called trophoblast), is one and the same,
viz. to form the outer layer of the chorion (false amnion,
serous membrane) and omphalopleure (unsplit yolk-sac wall.
Hill [’97]),^ and that to deny their homogeny to each other
implies the nou-homogeny of these membranes and the amnion
in the Amniotan series, and consequently renders the group
name Amniota void of all moi’phological meaning.
The rapidity Avith which the enclosure of the yolk-mass
is effected, and the relative tardiness of differentiation in the
embryonal region are features Avhich sharply distinguish the
early ontogeny of the Monotremes from that of the Sauropsida,
and which, in my view, are of the very greatest importance,
since they afford the key to a correct understanding of the
peculiar coenogeuetic modifications observable in the early
ontogeny of the Metatheria and Eutheria. To appreciate the
significance of these featui-es it is necessary to take account
of the great difference which exists between the Sauropsidan
and Monotreme ovum in regard to size, as Avell as of the very
different conditions under Avhich the early development goes
on in the two groups. The Sauropsidan egg is large enough
to contain Avithin its OAvn confines the amount of yolk necessary for the production of a young one complete in all its
parts and capable of leading an independent existence
immediately it leaves the shell. Furthermore, it is also large
’ In certain Ainniotes the layers in question appear also to participate
in the formation of the inner lining of the amnion (amniotic ectoderm)
(cf . Assheton [’09], pp. 248-9), but this does not affect the statement in
the text. In the Saxu'opsida and Monotremata I think I am coia-ect in
saying that no sharp distinction is recognisable between the embi'yonal
and extra-embryonal regions of the ectoderm, hence it is difficult, if not
imj)ossible, to determine with certainty their relative participation in
the formation of the amniotic ectoderm.
enough to provide room for tlie development of an embryo
without any secondary growth in size after it leaves the ovary.
Moreover we have to remember that after it has become
enclosed in the shelly it remains but a short time in the oviduct
and receives little or no additional nutrient material from the
oviducal walls. The yolk-mass in any case retains its solid
character; there is no necessity for its rapid enclosure, and
so enclosure is effected slowly, contemporaneously with the
differentiation of the embryo.
In the Monotreme the conditions are altogether different.
The ripe ovarian ovum when it enters the oviduct has a
diameter of about 3-5 to 4 mm., and is thns considerably
smaller than that of a Eeptile of the same size as the adult
Monotreme. The amount of yolk which it is capable of containing is not anything like sufficient to last the embryo
throughout the developmental period, and, moreover, it does
not provide the space essential for the development of an
embryo on the ancestral Reptilian lines. As Assheton (’98,
p. 251) has pointed out, “ the difference in size between
the fertilised ovum of a reptile or bird or of a mammal
is very great ; but the difference in size between the
embryo of, say, a bird with one pair of mesoblastic
somites and of a mammal of the same age is comparatively
small. This means that nearly the same space is required
for the production of the mammalian embryo as of the
Sauropsidan, and has to be provided.” In the Monotreme
not only is additional room necessary, but also additional
nutrient material, sufficient with that already present in the
egg to last the embryo throughout the period of incubation.
Both are acquired contemporaneously during the sojourn of
the egg in the uterine portion of the oviduct, wherein the egg
increases greatly in size. When it enters the uterus, the
Monotreme egg has a diameter, inclusive of its membranes, of
about 4-5 mm. ; when it is laid, it measures in Ornithorhynchus, in its greatest diameter, 16-19 mm., and somewhat
less in the case of Echidna. Prior to the enclosure of the yolk
the increase in diameter, due to the accumulation of fluid in
the perivitelliue space and between the zona and shell, is but
slight. But as soon as the yolk becomes suiTonnded by a
complete cellular membrane, i.e. as soon as the egg has
become converted into a thin-walled blastocyst, rapid growth
sets in, accompanied by the active imbibition of the nutrient
fluid, which is poured into the uterine lumen as the result of
the secretory activity of the abundantly developed uterine
glands. The fluid absorbed not only keeps the blastocyst
turgid, but it brings about the more or less complete disintegration of the yolk-mass, its constituent spherules
becoming disseminated in the fluid contents of the blastocyst
cavity. Although a distinct and continuous subgerminal
cavity, such as appears beneath the embryonal region of the
Sauropsidan blastoderm, does not occur in the Monotreme
egg, vacuolar spaces filled with fluid develop in the white
yolk-bed underlying the site of the germinal disc and appear
to represent it. As Wilson and Hill remark (’03, p. 317),
“ one can, without hesitation, homologise the interior of the
vesicle with the subgerminal cavity of a Saui’opsidan egg,
extended so as to include by liquefaction the whole of the
yolk itself.” In the Marsupial the blastocyst cavity has a quite
different origin, since it represents the persistent segmentation
cavity, whilst in the Eutheria the same cavity is secondarily
formed by the confluence of intra- or intei*-cellular vacuolar
spaces, but no one, so far as I know, has ever v^entured to
assert that, because of this difference in mode of origin, the
blastocyst cavity in the series of the Mammalia is a nonhomogenous formation.
To return to the matter under discussion, it appeal’s to me
that the necessity which has arisen, consequent on the I’eduction in size of the ovum, for rapid growth of the same in
order to provide room for the development of an embryo and
for the storage of nutrient material furnished by the maternal
uterus, affords a satisfactory explanation of the much more
marked activity of the extra-embryonal I'egion of the blastoderm as compared with the embryonal, Avhich is such a striking
feature in the early ontogeny of the Monotremes, and not
only of them, but, as Assheton has pointed out (’98, p. 251),
of the higher mammals as well (cf. the process of epiboly and
the inertness at first displayed by the formative cells of
the embryonal knot as compared with the activity of the nonformative or tropho-ectodermal cells), an activity which
results in the rapid completion of that characteristically
mammalian developmental stage  -  the blastocyst or blastodermic vesicle.
The necessity for the early formation of such a stage,
capable of rapidly growing in a nutrient fluid medium
provided by the mother, has profoundly influenced the early
ontogeny in all three mammalian subclasses, and natui*ally
most of all that of the Eutheria, in which reduction of the
ovum, both as regards size and secondary envelopes, has
reached the maximum. And I think there can be little
doubt but that it is this necessity which has induced that
early separation of the blastomeres into two categories,
respectively formative and non-formative in significance,
which has long been recognised as occurring in Eutheria, and
which I have shown also occurs amongst the Metatheria.
This early separation of the blastomeres into two distinct
groups is not recognisable in the Sauropsida, and the idea
that it is in some way connected with the loss of yolk which
the mammalian ovum has suffered in the course ofphylogeny,
was first put forward, I believe, by Jenkinson. In his paper
on the germinal layers of Vertebrata (’06, p. 51) he writes:
“ Segmentation therefore is followed in the Placentalia by
the separation of the elements of the trophoblast from those
destined to give rise to the embryo and the remainder of its
foetal membranes, and this ^precocious segregation’
seems to have occurred phylogenetically during
the gradual loss of yolk which the egg of these
mammals has undergone.” Whether or not such a
precocious segregation ” has already become fixed in the
Monotremes,future investigation must decide (cf . ante, p.90).
Ihe loss of yolk, with resulting reduction in size which the
Monotreme ovum has suffered in the course of phylogeny, we
must assume to have taken place gi-adually and in correlation
with the longer retention of the egg in the oviduct, the
elaboration of the uterine portion of the same as an actively
secretory organ, and the evolution of the mammary apparatus.
The Monotremes thus render concrete to us one of the first
great steps in mammalian evolution so far as developmental
processes are concerned, viz. the substitution for intra-ovular
yolk of nutrient material furnished directly by the mother to
the developing egg or embryo. We see in them the beginnings of that process of substitution of uterine for ovarian
nutriment which reaches its culmination in the Eutheria with
their microscopic yolk-poor ova and long intra-uterine period
of development. The Marsupials show us in Dasyurus an
interesting intervening stage so far as the ovum is concerned,
in that this, though greatly reduced as compared with that
of the Monotreme, still retains somewhat of its old tendencies
and elaborates more yolk-material than it can conveniently
utilise, with the result that it has to eliminate the surplus
before cleavage begins. But as coucerns their utilisation of
intra-uterine nutriment, they have specialised along their
own lines, and instead of exhausting the possibilities implied
by the presence of that, they have extensively elaborated
the mammary apparatus for the nutrition of the young, born
in a relatively immature state, after a short period of intrauterine life (cf. Wilson and Hill [T7, p. 580]).
In view of the fact that the young Monotreme enjoys three
developmental periods, viz. intra-uterine, incubatory, and
lactatory, the question might be worthy of consideration
whether it may not be that the Marsupial has merged the
incubatory period in the lactatory, the Eutherian the same in
the intra-uterine.
2. The Early Development of the Metatheria and
It will have become evident Horn the foregoing that the
Metatherian mode of early development is to be regarded as
but a slightly modified version of the Prototherian, such
differences as exist between them being interpretable as coenogeuetic modifications, induced in the Metatherian by the
practically complete substitution of uterine nutriment for
intra-ovular yolk, a substitution which has resulted in the
attainment by the marsupial ovum of the holoblastic type of
cleavage. In tlie present section I hope to demonstrate how
the early ontogeny of the Metatlieria enables us to interpret
that of the Eutheria in terms of that of the Prototheria.
If we proceed to compare the early development in the
Metatlieria and Eutheria, we encounter, from the 4-celled
stage onwards, such obvious and profound differences in the
mode of formation of the blastocyst, and in the relations of
its constituent parts, that the differences seem at first sight
to far outweigh the resemblances. Nevertheless, apart from
their common possession of the same holoblastic mode of
cleavage, there exists one most striking and fundamental
agreement between the two in the fact that in both there
occurs, sooner or later during the cleavage process, a separation of the blastomeres into two distinct, pre-determined cellgroups, whose individual destinies are very different, but
apparently identical in the two subclasses. In tlie Marsupial,
as typified by Dasyurus, the fourth cleavages are, as we have
seen, unequal and qualitative, and result in the separation of
two differentiated groups of blastomeres, arranged in two
superimposed rings, viz. an upper ring of eight smaller, less
yolk-rich cells, and a lower of eight larger, more yolk-iuch
cells. The evidence justifies the conclusion that the former
gives origin directly to the formative or embryonal region of
the vesicle wall, the latter to tlie non-formative or extraembryonal region.
Amongst the Eutheria the evidence is no less clear. It has
been conclusively shown by various observers (Van Beneden,
Duval, Assheton, Hubrecht, Heape, and others) that, sooner
or later, there occui's a separation of the blastomeres into two
distinct groups, one of which eventually encloses the other
completely. The two groups may be clearly distinguishable
t.C.TTV. (€cj
emJb. ect.
Diagrams illustrating the mode of formation of the blastocyst
in Metatheria (a-d) and Eutheria (1-3). b.c. Blastocyst cavity.
i.c.m. Inner cell-mass, 'pr.amn.c. Primitive amniotic cavity.
r.l. Rauber’s layer. s.c. Segmentation cavity. For other
reference letters see explanation of plates (p. 125).
in eai'lv cleavage stages, owing to diffecences in the characters
and staining reactions of their cells, and in such cases there
is definite evidence of the occurrence of a process of overgrowth
or epiboly, whereby one group gradually grows round and
completely envelops the other, so that in the completed
morula a distinction may be drawn between a central cellmass and a peripheral or enveloping layer (rabbit. Van
Beneden; sheep, Assheton). In other cases, where it has
been impossible to recognise the existence of these two
distinct cell-groups in the cleavage stages, we nevertheless
find, either in the completed moimla or in the blastocyst, that
a more or less sharp distinction may be drawn between an
enveloping layer of cells and an internally situated cell-mass
(inner cell-mass).
E. van Beneden, in his classical paper on the development
of the rabbit, published in 1875, was the first to recognise
definitely the existence of two categories of cells in the
segmenting egg of the Eutherian mammal. In this form he
showed how in the morula stage a cap of lighter blastomeres
gradually grows round and envelops a mass of more opaque
cells by a process of overgrowth or epiboly. In his more
recent and extremely valuable paper on the development of
Yespertilio (’99), he again demonstrated the existence of two
groups of blastomeres as well in the segmenting egg as in the
completed morula, but failed to find evidence of epiboly in all
cases. Nevertheless he holds fast to the opinion which he
expressed in 1875 : “ Que la segmentation s’accompagne, chez
les Mammiferes placentaires, d’un enveloppement progressif
d’une partie des blastomeres par une couche cellulaire, qui
commence a se differencier des le debut du developpement,”
and states that “dans tons les oeufs arrives a la fin de la
segmentation et dans ceux qui moutraient le debut de la
cavite Blastodermique j’ai constamment rencontre une couche
peripherique complete, eutourant de toutes parts un amas
cellulaire interne, bien separe de la couche enveloppante.”
The latter layer he regards as corresponding to the extraembryonal ectoderm of the Sauropsida, and points out that
“ chez tons les Choi'des les premiers blastomeres qui se
differencient et qui avoisinent le pole animal de I’oeuf sont
des elements epiblastiqnes. C’est par la couolie cellulaire qui
resulte de la segmentation ulterieure de ces premiers blastomeres epiblastiqnes que se fait, cbez les Sauropsides, benveloppement du vitellus. Dans Toeuf reduit a n’etre plus
qu’une sphere microscopiquej bepibolie a pu s’achever des la
fin de la segmentation, voire meme avant bachevement de ce
phenomene.” The “ amas cellulaire interne ” (embryonal
knot, inner cell mass). Van Beneden shows, differentiates
secondarily into “ un lecithophore et un bouton embryonnaire.’’ The former is the entoderm of other authors, the
latter the formative or embryonal ectoderm. Hubrecht, in
the forms studied by him (Sorex, 'I’upaia, Tarsius^) finds
a corresponding differentiation. In Tupaia he describes the
morula stage as consisting of a single central lightly staining
cell, which he regards as the parent cell of the inner cell-mass
of later stages, and of a more darkly staining peripheral layer
which forms the unilaminar wall of the blastocyst. Here,
then, the parent cells of the two cell-groups would appear to
be separated at the first cleavage. Hubrecht, like Van
Beneden, holds that the inner cell-mass furnishes the
embryonal ectoderm and the entire entoderm of the blastocyst.
The peripheral layer he has termed the trophoblast (’88, p.
511), and in his paper on the placentation of the hedgehog
(’89, p. 298) he defines the term as follows: “I propose to
confer this name to the epiblast of the blastocyst as far as it
has a dix’ect nutritive significance, as indicated by proliferating
processes, by immediate contact with maternal tissue, maternal
blood, or secreted material. The epiblast of the germinal
ai-ea  -  the formative epiblast  -  aud that which will take part
in the formation of the inner lining of the amnion cavity is,
ipso facto, excluded from the definition.” Thus the name
* In Erinacens the entoderm, from Hubrecht’s observations, appears
to be precociously differentiated, prior to the separation of the embryonal
ectoderm fi'om the overlying trophoblast, but the details of the early
development in this form are as yet only incompletely known.
trophoblasb was originally employed by Hubrecbt as a convenient term designatory of what he at the time regarded as
the extra-embryonal ectoderm of the mammalian blastocyst.
In the course of his speculations on the oingin of this layer,
however, he has reached the conclusion that it is really of the
nature oP'a larval envelope, an Embryonalhiille (^08, p. 15),
inherited by the mammals, not from the reptiles (which have
no direct phylogenetic I'elationship to the latter), but from
their remote invertebrate ancestors ('Vermiform pi’edecessors
of coelenterate pedigree, provided with an ectodermal larval
investment [Laiwenhiille] ”).
Assheton, again, although he was unable to convince himself (’94) of the correctness of van Beneden’s account of the
occurrence of a process of epiboly in the segmenting eggs of
the rabbit, finds in the sheep (’98) that a differentiation into
two groups of cells is recognisable “ perhaps as early as the
eight segment stage,” and that one of the groups gradually
envelops the other. “Let it be noted,” he writes (’98, p. 227),
“ that we have now to face the fact, based on actual sections,
that there is in certain mammals a clear separation of
segments at an early stage into two groups, one of which
eventually completely surrounds the other,” and instances
Van Beneden’s observations on the rabbit (of the correctness
of which he, however, failed to satisfy himself, as noted above),
Duval’s observations on the bat, Hubrecht’s on Tupaia, and
his own on the sheep. Assheton thinks this phenomenon
“ must surely have some most profound significance,”
but finds himself unable to accept the interpretations of
either Van Beueden or Hubrecht, and puts forward yet
another view, “ based on the appearance of some segmenting
eggs of the sheep ” (’08, p. 233), “that in cases where this
differentiation does clearly occur, it is a division into epiblast
and hypoblast, the latter being the external layer” (’98, p. 227).
Assheton thus differs from all other observers in holding that
the inner cell-mass or embryonal knot of the Eutherian
blastocyst gives origin solely to the formative or embryonal
ectoderm, and I believe 1 am correct in stating that he also
J. p. mi,L.
differs from all other observers in holding that the outer
enveloping layer of the same is entodermald
The fact, then, of the occurrence amongst Eutheria of a
“precocious segregation ” of the blastomeres into two distinct
groups, one of which eventually surrounds the other completely, is not in dispute, though authorities differ widely in
the intei'pretation they place upon it. In the Eutherian
blastocyst stage, the enveloping layer forms the outer unilaminar wall of the vesicle, and encloses the blastocyst cavity
as well as the other internally situated group. This latter
typically appears as a rounded cell-mass, attached ac one spot
to the inner surface of tlie enveloping layer, but more or less
distinctly marked off from it. It is genei-ally termed the
inner cell-mass or embryonal knoc (“ amas cellulaire interne ”
of Van Beneden). For the enveloping layer Ilubrecht’s name
of “ trophoblast ” is now generally employed, even by those
who refuse to adopt the speculative views with which its
originator has most unfortunately, as I think, enshrouded this
convenient term.
I have demonstrated the occurrence of an apparently comparable “precocious segregation^^ of the blastomeres into
two distinct groups in one member of the Metatheria which
there is no reason to regard as an abeirant type, and I have
shown beyond all shadow of doubt that from the one group,
which constitutes what I have termed the formative region
of the unilaminar vesicle-wall, there arise the embi*youal
ectoderm and the entire entoderm of the vesicle, both embryonal and extra-embryonal, and that the other group, which
constitutes the non-formative region of the vesicle-wall,
directly furnishes the extra-embryonal ectoderm, i.e. the
ectoderm of the omphalopleui’e and chorion."
* Assheton states ('08, p. 233, cf. also ’98, p. 220) that his interpretation “ owes ranch also to the theoretical conclusions of Minot and
Robinson.” However that may be, both Minot and Robinson in their
most recent writings continue to speak of the chorionic ectoderm.
^ Whether or not it participates in the formation of the ainniotic
ectoderm future investigation must decide.
As resrards Eutheria, we have seen that Van Beneden and
Hubrecht, though their views in otlier respects are widely
divero-ent, both ag'ree that the inner cell-mass of the blastocyst furnishes the embryonal ectoderm (as well as the amniotic
ectoderm wholly or in part) and the entire entoderm of the
vesicle. That, in fact, is the view of Mammalian embryologists
generally (Duval and Assheton excepted),^ and if we may
assume it to be correct, then it would appear that the later
history of the formative region of the Marsupial blastocyst
and that of the inner cell-mass of the Eutherian are identical.
That being so, and bearing in mind that both have been
shown, at all events in certain Mammals, to have an identical
origin as a group of precociously segregated blastotneres,^ I
can come to no other conclusion than that they are homogenous formations. If that be accepted, then this fact by itself
renders highly probable the view that the so-called trophoblast of the Eutherian blastocyst is homogenous with the
non-formative region of the Metatherian vesicle, and v?hen
we reflect that both have precisely the same structural and
topographical (not to mention functional) relations in later
stages, inasmuch as they constitute the ectoderm of the chorion
and omphalopleure (with or without participation in the
formation of the amniotic ectoderm;, and that both have a
similar origin in those Mammals in which a precocious segregation of the blastomeres has been recognised, their exact
* The view of Duval [’95], based on the study of Vespertilio, that the
inner cell-mass gives rise solely to entoderm, and that the enveloping
layer furnishes not only the extra-embryonal but also the embryonal
ectoderm, is shown by Van Beneden’s observations on the same form to
be devoid of any basis of fact. Assheton’s views are referred to below
(p. 110).
- The fact that the phenomenon of the “ precocious segregation” of
the blastomeres into two groups with deteiminate destinies has already
become fixed in tlie Marsupial lends additional weight to the view of
Van Beneden that such a segregation will eventually be recognised as
occurring in all Eutheria without exception. Without it, it is difficult
to understand how the entypic condition, characteristic of the blastocysts of Ml known Eutheria, is attained, imless by differentiation in
situ, which .seems to me highly improbable.
J. r. HILL.
homology need no longer be doubted. In the preceding section
of this paper (ante, pp. 91, 92) I have shown reason for the
conclusion that the non-formative region of the Marsupial
blastocyst is the homologue of the extra-embryonal ectoderm
of the Monotreme and Reptile, and if that conclusion be
accepted it follows that the outer enveloping layer of the
Eutherian blastocyst, the so-called trophoblast of Hubrecht,
is none other than extra-embryonal ectoderm, as maintained
by Van Beneden, Keibel, Bonnet, Jenkinson, Lee, MacBride
and others, the homologue of that of Reptilia.
I am therefore wholly unable to accept the highly speculative conclusions of Hubrecht, set forth with such brilliancy
in a comparatively recent number of this Journal (’08), as
to the significance and phylogeny of this layer. These conclusions, on the basis of which he has proceeded to formulate
such far-reaching and, indeed, revolutionary ideas not only
on questions embryological, but on those pertaining to the
phylogeny and classification of vertebrates, have already
been critically considered by Assheton (’09) and MacBride
(’09), also in the pages of this Journal, and found wanting,
and they are, to my mind, quite irreconcilable with the facts
I have brought to light in regard to the early development
of Marsupials. I yield to no one in my admiration for the
epoch-making work of Hubrecht on the early ontogeny and
placentation of the Mammalia, and I heartily associate
myself with the eulogium thereanent so admirably expressed
by Assheton in the cx’itique just referred to (p. 274), but
I am bound to confess that as concerns his views on the
phylogeny of this layer, which he has termed the “ trophoblast,” he seems to me to have forsaken the fertile field of
legitimate hypothesis for the barren waste of unprofitable
speculation, and to have erected therein an imposing edifice on
the very slenderest of foundations.
Before I proceed to justify this, my estimate of Hubrecht’s
views on the phylogeny of the trophoblast, let me first set
forth his conception so far as I understand it. He starts
with the assumption that the vertebrates (with the exception
of Ainpliioxus, the CyclostoineSj and the Elasraobi'anclif!) are
descended from “vermiform predecessors of coelenterate
pedigree” possessed of free-swimming larvte, in which there
was present a complete larval membi'ane of ectodermal derivation, and of the same order of differentiation “as the outer
larval layer which in certain Nemertines, Gephyreans, and other
worms often serves as a temporaiy envelope that is stripped
off when the animal attains to a certain stage of development.”
When, for oviparity and larval development, viviparity and
embryonic development became established in the Protetrapodous successors of the ancestral vermiform stock, the
larval membrane did not disappear. On the contrary, it is
assumed that it merely changed “its protective or locomotor
function into an adhesive one,” and so, development now
taking place in utero, it is quite easy to understand how tlie
larval membrane could gradually become transformed into
a trophic vesicle, containing the embryo as before, and
functional in the reception of nutriment from the walls of
the maternal uterus. The final stages in the evolution of
this trophic vesicle constituted by the old larval membrane
are met with amongst the mammals, since in them it
became vascularised so as to constitute a “yet more
thorough system of nourishment at the expense of the
maternal circulatory system.” Such, then, is the phylogeny
of the trophoblast according to Hubrecht. The Eutheriau
mammals, which it is held trace their descent straight back to
some very early Protetrapodous stock, viviparous in habit and
with small yolk-poor, holoblastic eggs, exhibit the trophoblast in its most perfect condition. Hubrecht therefore starts
with them, and attempts to demonsti’ate the existence of a
larval membrane, or remnants of such, externally to the
embryonal ectoderm in all vertebrates with the exceptions
already mentioned. There is no question of its existence in
the Meta- and Eutherian mammals. “We may,” writes
Hubrecht (’08, p. 12), . . . “insist upon the fact that
. . . all Didelphia and Monodelphia hitherto investi
gated show at a very early moment the didermic stage out of
which the embryo will be built up enclosed in a cellular
vesicle (the troplioblast), of which no pai‘t ever enters into
the embryonic organisation.” The common possession by the
Metatheria and Eutheria of a larval membi'ane is after all
only what might be expected, “since after Hill’s (’97)
investigations, we must assume that the didelphian mammals
are not descended from Ornithodelphia but from monodelphian
placental ancestors.” As concerns the Prototheria, although
they cannot in any sense be regarded as directly ancestral to
the other mammals, we nevertheless find the trophoblastic
vesicle “ compax’atively distinct.” “In many reptiles and
birds,” however, it is “.distinguished with great diflSculty