CAMBRIDGE UNIVERSITY PRESS

C. F. CLAY, MANAGER

itoiltrott: FETTER LANE, E.G.

(Elltttimrgf) : 100 PRINCES STREET

H. K. LEWIS AND CO. LTD., 136 GOWER STREET, W.C.
fieto gorh: G. P. PUTNAM'S SONS
anli Calcutta : MACMILLAN AND CO., LTD.
Toronto: J. M. DENT AND SONS, LTD.

THE MARUZEN-KABUSHIKI-KAISHA

All rights reserved

A

GROWTH IN LENGTH

EMBRYOLOGICAL ESSAYS

BY

RICHARD ASSHETON, M.A., Sc.D., F.R.S.

Trinity College, Cambridge

Lecturer in Biology at Guy's Hospital in the University of London

Lecturer in Embryology at the Imperial College

Lecturer in Animal Embryology in the

University of Cambridge

With 42 Illustrations

Cambridge :

at the University Press

1916

PREFACE

ri THE three Lectures forming the first part of this book were
given as one of the "Advanced Courses in Zoology" in the
University of London. They were delivered at Guy's Hospital
on January 23rd, January 30th and February 6th, 1913, under
the title 'The Growth in Length of the Vertebrate Embryo."
They are now published in accordance with the known intention
of the author, and in response to the expressed wish of some of those
who heard them.

Although my husband had destined this work for publication,
he had not actually prepared it for the press ; but the Lectures
were typewritten and some corrections and notes had been added
to them. For help in arranging them for print I am greatly
indebted to Professor Stanley Gardiner, F.R.S., and to Professor
J. P. Hill, F.R.S., who have most kindly read the MSS; and to
the latter especially for valuable suggestions as to some altera-
tions in the original wording, which while adapted for lecturing
purposes might not always have been clear to the reader.

The figures have been drawn by Miss D. Thursby-Pelham from
the sketches given in the MSS referring to lantern slides and
diagrams displayed during the lectures.

The cost of publication is partly defrayed by a grant from the
Royal Society.

The second part of the book consists of a reprint of a paper
on the Mechanics of Gastrulation which appeared in the Archivfilr
Entwicklungsmechanik der Organismen in 1910. It is here included
as the subject is kindred to that of the lectures and also because
it is often in demand and further copies are unobtainable.

vi Preface

As these "Growth Problems" occupied my husband's studies
and interests perhaps to a greater extent than did any other
of the varied Biological work undertaken by him, this small
memorial volume may be of special interest to those who knew
him as researcher and teacher. It is now published in the hope
that it may fulfil his constant wish to inspire others to delve
into this, to him, fascinating field of research. It represents the
summation of work and thought carried on for more than twenty
years.

FRANCES A. E. ASSHETON.

GRANTCHESTER,

April, 1916.

CONTENTS

PAGE

PUBLICATIONS BY THE SAME AUTHOR .... viii
THE GROWTH IN LENGTH or THE VERTEBRATE EMBRYO

LECTURE I 1

LECTURE II 24

LECTURE III 47

THE GEOMETRICAL RELATION or THE NUCLEI IN AN IN-
VAGINATING GASTRULA (e.g. AMPHIOXUS) CONSIDERED
IN CONNECTION WITH CELL RHYTHM, AND DRIESCH's
CONCEPTION OF ENTELECHY 65

LITERATURE REFERRED TO . 99

PUBLICATIONS BY THE SAME AUTHOR

On the Development of the Optic Nerve of Vertebrates and the
Choroidal Fissure of Embryonic Life."
Quarterly Journal Microscopical Science. Vol. xxxiv, 1892.

A Re-investigation into the Early Stages of the Development of
the Rabbit."
Quarterly Journal Microscopical Science. Vol. xxxvn, 1894.

;' On the Phenomenon of the Fusion of the Epiblastic Layers in the
Rabbit and in the Frog."
Quarterly Journal Microscopical Science. Vol. xxxvn, 1894.

" On the Causes which lead to the Attachment of the Mammalian
Embryo to the wall of the Uterus."
Quarterly Journal Microscopical Science. Vol. xxxvn, 1894.

"The Primitive Streak of the Rabbit, the causes which may
determine its Shape, and the part of the Embryo formed by
its activity."
Quarterly Journal Microscopical Science. Vol. xxxvn, 1894.

"On the Growth in Length of the Frog Embryo."

Quarterly Journal Microscopical Science. Vol. xxxvn, 1894.

"Notes on the Ciliation of the Ectoderm of the Amphibian

Embryo."
Quarterly Journal Microscopical Science. Vol. xxxvm, 1896.

An Experimental Examination into the growth of the Blastoderm
of the Chick."

Proceedings of the Royal Society. Vol. LX, 1896.

Publications by the same author

''' An account of a Blastodermic Vesicle of the Sheep of the seventh
day with twin germinal areas."

Journal of Anatomy and Physiology, April, 1898.

"The Development of the Pig during the first ten days."

Quarterly Journal Microscopical Science. Vol. XLI, 1898.

" The Segmentation of the Ovum of the Sheep, with observations
on the Hypothesis of a Hypoblastic Origin for the Tropho-
blast."

Quarterly Journal Microscopical Science. Vol. XLI, 1898.

"

:<0n Growth Centres in Vertebrate Embryos.1

Anatomischer Anzeiger. Vol. xxvn, 1905.

'The Morphology of the Ungulate Placenta, particularly of that
organ in the Sheep, and notes upon the Placenta of the
Elephant and Hyrax."

Proceedings of the Royal Society. Vol. LXXVI, 1905.

;'0n the Foetus and Placenta of the Spiny mouse, Acomys
cahirinus."

Proceedings of the Royal Society. Vol. LXXVI, 1905.

;' Certain Features characteristic of Teleostean Development."

Guy's Hospital Reports. Vol. LXI, 1907.

'The Development of Gymnarchus niloticus."

Budgett Memorial Volume, University Press, Cambridge.

'' Report on sundry Teleostean Eggs and Larvae from the Gambia
River."

Budgett Memorial Volume, University Press, Cambridge.

"The Blastocyst of Capra, with some Remarks upon the Homologies
of the Germinal Layers of Mammals."

Guy's Hospital Reports. Vol. LXII, 1908.

:'A new species of Dolichoglossus."

Zoologischer Anzeiger. Vol. xxxm, 1908.

Publications by the same author xi

" Professor Hubrecht's paper on the Early Ontogenetic Phenomena
in Mammals."

Quarterly Journal Microscopical Science. Vol. LIV, 1909.

" Tropidonotus and the Archenteric Knot of Ornithorhynchus."
Quarterly Journal Microscopical Science. Vol. LIV, 1909.

:

The geometrical relation of the Nuclei in an Invaginating Gastrula
(e.g. Amphioxus) considered in connection with Cell Rhythm
and Driesch's conception of Entelechy."
Archiv fur Ent. Mech. der Organismen. Vol. xxix, 1910.

"Variation and Mendel."

Guy's Hospital Reports. Vol. LXIV, 1910.

'Loxosoma loxalina and Loxosoma saltans, a new species."

Quarterly Journal Microscopical Science. Vol. LVIII, 1912.

" Gastrulation in Birds."

Quarterly Journal Microscopical Science. Vol. LVIII, 1912.

" A review of Dr Beard's book on the Enzyme Treatment of Cancer
and its scientific basis."

Guy's Hospital Gazette, February, 1912.

"Fission of the Embryonal Area in Mammals."

IXth International Zoological Congress. Monaco. 1913.

(With Dr A. Robinson.)
"Formation and Fate of the Primitive Streak."

Quarterly Journal Microscopical Science. Vol. xxxn, 1891.

(With Dr T. G. Stevens.)
"The Elephant's Placenta."

Quarterly Journal Microscopical Science. Vol. XLIX, 1905.

THE GROWTH IN LENGTH OF THE
VERTEBRATE EMBRYO

LECTURE I

The mode of growth of the main axis of the Vertebrate Embryo
has been for a good many years a subject of controversy among
Comparative Anatomists and Embryologists. The problem is not
such an easy one as it seems at first sight to be.

From a zoological point of view it is a very important question ;
because according to the way in which we are disposed to answer
it we shall form our ideas as to the relation which the main axis
of the vertebrate bodv bears to the main axis of the bodies of

V

animals of other phyla. Also upon this problem depends to a
large extent the view we take about metamerism. An Annelid is
a distinctly metamerically segmented animal — so is a Vertebrate.
Is this similarity a case of homology or of convergence? And
what is the relation that the metamerism of the one bears to that
of the other ?

The problem of the growth in length of the Vertebrate Embryo
may be said to have originated with the essay by Prof. Wilhelm His
published in the year 1874 on the formation of the trout embryo
("Ueber die Bildung des Lachs Embryo/' Sitzungsberichte der
naturforschenden Gesellschaft zu Leipzig, I. Jahrgang (1874), S. 30);
the theory of the mode of growth in length of the embryo
enunciated in this paper was expanded in subsequent papers,
among which may be mentioned his well-known work Unsere
Korperform und das physiologische Problem ihrer Entstehung,
published in the same year.

The theory which His then put forward has been termed the
concrescence theory. It is with respect to embryos of Teleosteans
and Elasmobranchs that one is most inclined at first sight to
adopt this theory of the growth in length of the Vertebrate.

A. v. E. 1

2 Growth in length of the Vertebrate Embryo

Oddly enough it is also to embryos of this type that those
who oppose this view can best turn for their experimental evidence
of a destructive character against the theory.

The theory of concrescence is as we shall see more difficult to
combat with reference to the Amniota, for in them the theorv

V

confronts us in a more subtle form.

The eggs of most of the members of these two classes — the
Elasmobranchs and Teleosteans — are large and are heavily laden
with food yolk.

After the egg of an Elasmobranch has undergone a partial or
meroblastic segmentation,, a little disc or cap of cells is formed
upon one pole. This we may call the upper pole, because in most
cases, though not in all, the egg floats with the pole bearing this
cap uppermost.

The cap of cells soon begins to expand and to become differen-
tiated into distinct layers, an outer or upper one which is a
continuous membrane, and is called the ectoderm or epiblast;
and an inner tissue composed of loose cells and a yolk mass which
together may be called the endoderm or hypoblast. This inner
layer is continuous with the outer one peripherally, but at the
posterior end it is continued inwards for a short distance and then
becomes continuous with the yolk mass. The whole disc derived
from the segmented region of the egg is called the blastoderm or
blastodisc, the rest of the egg is called the yolk mass. This is the
gastrula stage, and the narrow shallow cleft, which in the Elasmo-
branch extends under the edge of the blastoderm, is really the
primitive gut cavity, and it may be called the archenteron. The
edge of the blastoderm, where the ectoderm and. endoderm are
continuous with each other is therefore the blastopore lip; and
the great yolk mass must be regarded as a piece of the floor of the
gut projecting out of the gastrula mouth.

The diagram, Fig. 1, is really a compromise between the
Elasmobranch and the Teleostean condition. In the Elasmobranch
there is a much greater difference in size between the blastoderm
and the yolk mass than in the Teleostean.

The Teleostean furthermore is complicated by the presence of
an outer envelope which is split off from the ectoderm layer of
cells. This layer is at first continuous with the yolk, but at a

Formation of Blastoderm 3

later stage it is also split off from the superficial layer of yolk.
This outer investment therefore prevents the existence of an
actual passage from the cleft-like gastrula cavity to the exterior,
so that the archenteron is not very obvious in this class of fishes
at this stage.

B

Fig. 1. Diagram to show the growth of the blastoderm over

the yolk.

A, B, C, the egg as seen from above; D, E, F, the same stages seen from the side;
bl blastoderm ; m merocytes ; the yolk is represented by dots.

1—2

Growth in length of the Vertebrate Embryo

As development proceeds, the blastoderm in each case becomes
expanded and it gradually surrounds the whole yolk; but long
before it does so there appears a thickening of the outer layer just
inside the blastopore rim, and the rim itself also is thickened
about this region. The inner edge of this thickened knob is the
anterior end of the future neural tube, and the rim is therefore
the most posterior part of the embryo. As the blastoderm
expands this thickening is apparently extended inwards from the
rim. The process that is really taking place is a growth of the
thickened area — which is usually called the " embryo " --extending
backwards and keeping pace with the expanding blastopore rim.
Accordingly this rim is often called the germ ring.

In the Teleostean, which gives a much truer idea of the growth

processes than does the Elasmo-
branch, the germ ring soon reaches
the equator of the egg. By this
time the :' embryo '' has also
increased a corresponding amount
in length, and its anterior end is
definitely differentiated into the
anterior portion of the fish, Fig. 2.
After the equator has been
attained the germ ring quickly
Fig. 2. Diagram to show the pro- spreads on to the lower pole and

gress of the germ ring as it travels . . 111

over the yolk at four successive m a short time the whole egg is
stages- enveloped.

6 blastopore ; gr germ ring ; ^i £ ,

l, 2, 3, 4, successive stages. The part of the germ ring

opposite to the part where the

thickening occurs, travels much more quickly over the yolk than
the thickened part.

Thus it is seen that the blastopore becomes more and more
reduced until it actually closes. By this time the greater part
of the "embryo"' has been laid down, in fact all of it except the
tail; and we have watched the growth in length of the embryo
from the time when the head region first appeared until just
before the tail begins to grow. The question is how has this
growth taken place ?

When once an embryo is formed it may increase either in

Concrescence Theory 5

general size, or in length only ; it may increase as a whole, or in
part ; in any case the increase comes from a general multiplication
of its cells, that is to say, it is the result of a general interstitial
growth.

Or again there may be certain definite centres which continue
for some time as centres of indifferent cell proliferation, and which
give rise to growth in length of parts, as for instance, in the
case of the limbs, or of the body as a whole.

Now in the particular case which we are considering, namely
the growth in length of the fish embryo, it might be possible
theoretically to ascribe the change from the condition seen in
the earlier stage A, Fig. 3, to that of the later stage B, Fig. 3,
as being due to one of three causes.

B C

Fig. 3. Concrescence theory, after His.

The arrows indicate the direction in which the edges are approaching
each other ; the dotted area represents yolk.

1. It might be due to a general interstitial growth whereby
a — b becomes a'--b'. In this case the foundation of the whole
embryo would be considered as having been definitely laid down
between these two points from the beginning.

This explanation however is probably not correct, because,
as is perfectly well known, in all vertebrate embryos the anterior
part of the embryo develops long before the hindermost. In Fig. 3
the earlier stage shows only the rudiment of the fore part of
the brain. The rest of the brain and the spinal cord and the
protovertebrae are added on subsequently, in regular succession.

Experimental evidence is also against this hypothesis.

2. Or, we may regard the advancing rim as being an area of
special cell proliferation, and as giving rise to the more posterior
part of the embryo as it passes backwards over the yolk.

6 Growth in length of the Vertebrate Embryo

We shall find that in the Anamnia the experimental evidence
supports this view.

3. Then again it is possible theoretically to hold the view that
the main axis of the embryo is really formed by the coalescence of
the lips of the rim — each half side of the rim representing one-half
of the embryo. This view is known as the concrescence theory.

Since the rim is regarded as the lip of the blastopore (the
uncovered yolk as a plug filling up the blastopore) then, according
to the concrescence theory of vertebrate development, the blasto-
pore is supposed to fuse by the concrescence of its lateral lips,
each lip giving rise to the corresponding half of the embryo,
Fig. 3 C, and the blastopore finally closing.

According to this theory the relation of the main axis of a
Vertebrate to the blastopore is very different from that which it
has if we adopt alternative No. 2. According to the concrescence
theory the relation of the main axis to the old blastopore of
coelenterate days — the gastraea stage — is the same in Vertebrates
as it is in Annelida and Arthropoda. If we adopt No. 2, the
main axis is in a direction at right angles to what it is in Annelida
and Arthropoda.

This theory of concrescence started by His in 1874 has been
brought forward time after time by embryologists in newer and
more subtle forms. It has received support from many authorities,
and has been vigorously opposed by others.

GROWTH CENTRES.

Quite apart from the question of concrescence — or the heresy of
concrescence, as I would prefer to call it — there is another question
of considerable importance, and of especial interest with respect
to the difficult question of gastrulation and the formation of the
germ layers, the ectoderm and endoderm.

If we look at an early stage of the trout embryo we see a
perfectly circular disc of cells. Later the rim thickens along the
border which will ultimately be the posterior and dorsal part of
the embryo and rim. But what about the thickened area just
within this rim? Fig. 1, C. Is that derived from the rim, or
is it formed so to speak in situ on the blastoderm ? And what

Protogenesis and Deuterogenesis 7

about the fate of the blastoderm just in front of this knob?
Does not that give rise to part of the embryo? There can be
no doubt about the answer.

All parts of the blastoderm of a Teleostean or an Elasmobranch
become parts of the embryo.

Thus we have a part of the embryo formed from the layer of
cells which exists as a result of segmentation, and another part
which conies into being by the activity of the germ rim whether
by concrescence or not.

That is to say we have two definite and in a sense independent
centres of growth, one of these comes into being as the result of
the fertilisation of the egg, — and the other comes into being
subsequently.

These may be termed the primary growth centre and the
secondary growth centre — or to use terms which I suggested some
years ago, the protogenetic and deuterogenetic growth centres. The
activities of these centres may be called protogenesis and
deuterogenesis respectively.

In order to understand the action of these centres one must
have a clear idea of the formation of the germ layers and of the
process and meaning of gastrulation.

I propose therefore to describe in some detail the formation
of the germ layers in the Amphibia.

Rana temporaria.

An egg of Rana temporaria is pigmented to a greater or less
extent. There is much variation, in some the lower hemisphere
is almost quite devoid of pigment, in others it is pigmented but
is slightly lighter near the lower pole.

The whiter pole is the more heavily laden with yolk and it
floats downwards in water. To all appearances the egg is radially
symmetrical. Soon after fertilisation a curious crescentic whiten-
ing of part of the pigmented area can be noticed, this is called the
"grey crescent."

It is said

(1) That this is always immediately opposite to the point
at which the spermatozoon has entered;

(2) That the future sagittal plane of the animal coincides

8

Growth in length of the Vertebrate Embryo

approximately with the centre of this crescent and that of
the egg.

Segmentation as a rule follows the radial type.

A pit begins on the upper pole of the black hemisphere, and
passes round as a furrow dividing the egg into two equal segments,
Fig. 4 A, B and E, F. Very soon, after about two hours or less
according to the temperature, two furrows start from the centre
of this furrow in the black hemisphere and gradually extend so
as to divide each segment again into two, so that we have four
segments of equal size all lying in the same plane, as in Amphioxus.

F G

Fig. 4. Egg of Eana temporaria.

H

A, B, C, D, diagrammatic views of segmentation as seen from above; E, F, G, H,

corresponding stages seen from the side.

If we compare several views of the lower pole we shall find
that there are several varieties of mode of meeting of the furrows
at the lower pole. Also we shall notice that the spot where the
furrows meet is not in the centre of the white hemisphere.

The next set of furrows is horizontal. These may be seen
commencing from the edges of the two previous furrows and
eventually meeting one another, so cutting the four segments into
eight.

In comparing this stage with Amphioxus we note that it
corresponds to the third division plane in the radial type of
Amphioxus, but is much nearer to the upper pole. There is thus
a greater difference between the micromeres and macromeres.

Segmentation

9

The radial character is continued in some cases by the division
in vertical planes, of each of the four micromeres, and at a slower
rate, of the four macromeres, Fig. 5 A. But sometimes a dis-
tinctly bilateral phase is entered upon, Fig. 5 B.

A B

Fig. 5 A, Radial division; J5, Bilateral division.

After this stage the segmentation proceeds less regularly and
the bilateral symmetry if established is soon lost. As early as
the eight segment stage a space is apparent between the segments
very much nearer the upper pole than the lower. This is the
first sign of the blastocoel or segmentation cavity. At about the
sixteen segment stage this cavity is quite enclosed, and we have
a condition comparable to that of Amphioxus, the chief difference
being that the segments are more uneven in size, fewer in number
and correspondingly larger, so that the blastocoel is correspondingly
smaller, Fig. 6 B. Segments are now formed by tangential
planes of division, so that the wall becomes first double and then
many-layered, and at the close of segmentation we have a blastula
very characteristic of Amphibians in which there is an upper pole
of very small black cells and a lower pole of considerably larger
and whiter cells, bounding between them the eccentrically placed
segmentation cavity, Fig. 6 D.

There is a stage during the earlier phases to which I would
draw special attention. In Fig. 6 C it will be noticed that there
are very few division furrows extending to the lower pole, and no
horizontal ones. This is of course due to the fact that the
horizontal planes always result in the production of a small and
a large segment, owing to the greater amount of yolk in the lower
part.

In Rana we have a good example of an egg which is holoblastic.

10

Growth in length of the Vertebrate Embryo

but in which gastrulatiori can be said to be produced by invagina-
tion in a very small degree only.

Gastrulation is produced by a splitting among the segments,
caused by a differentiation of an upper layer of small segments
above a layer of larger segments. We have seen that in the
advanced stage of segmentation the blastula of Rana has a many-
layered wall of which the upper part consists of smaller cells and

B

c D

Fig. 6. Frog's egg. Segmentation.

A, 8 cell stage; B, 16 cell stage; C, 32 cell stage; D, stage with segmentation

cavity (sg) fully developed.

the lower of larger cells, Fig. 6 D, with a zone of more or less
abrupt passage between the two ; this will be seen in sections to
vary slightly according to the radius along which the sections are
cut.

At this stage, on the surface, one sees that the upper pole is
formed by small black cells and the lower pole of whiter and larger
cells. There is no sharp line between the two, and yet there is
sufficient difference to enable one to speak of an advance of the

Gastrulation

11

small black cells into the region previously occupied by the large
white cells. This advance is an advance of a zone of differentia-
tion ; it is certainly not an actual movement of black cells over
the white, in Rana, though it may be so in some other Amphibia.
Thus there is a tendency during segmentation for the black area
to increase at the expense of the white, for pigment is deposited
as the larger cells segment into smaller ones. Gastrulation is
first rendered evident by the appearance of a slight groove running
parallel to the equator of the egg but some distance below it,
Fig. 7.

A

dl vl

Fig. 7. Diagram to illustrate the external views and corresponding
internal changes during gastrulation in the frog.

A, B, C, D, external views; E, F, G, H, median sections; arch archenteron;
bl blastopore ; dl dorsal lip of blastopore ; vl ventral lip of blastopore ; sg seg-
mentation cavity ; in H the segmentation cavity has almost disappeared.

We will now very briefly look at the general characters as one
sees them in surface view and sections, without at first considering
the mechanism.

If we examine the groove in section it will be noticed that
there are signs of pressure, the cells are compressed and there
is more pigment deposited between the segments at this point
than elsewhere.

If we watch the surface we find, Fig. 7 A—D, that the groove

12 Growth in length of the Vertebrate Embryo

which at first appeared almost straight and transverse, soon
lengthens and becomes crescentic, and finally its horns meet and
we have a white patch of larger segments surrounded by a dark
sharply defined rim. The white patch is formed by a plug of
yolk cells which almost completely blocks the blastopore, but all
round the plug the archenteric cavity is visible as a chink between
the plug and the lip rim. This rim is the blastopore lip.

Examination of vertical sections, Fig. 7 E — H, shows that a
narrow slit, continues inward from the groove parallel with the
surface and some half-a-dozen cells below it. The slit gradually
lengthens inwards from the groove and eventually expands into
a cavity as the blastopore forms and closes. This cavity is the
future gut cavity, and it may for the moment be called archen-
teron, though as we shall see it corresponds to rather more than
what we shall eventually term archenteron.

By this act of splitting the gut cavity is initiated, and the
formation of the two primary layers is established, firstly in the
region of what will be the future dorsal surface, then in the lateral
and lastly in the ventral region.

The process of gastrulation is complicated in the Amphibia
by the fact that the gut cavity is formed dorsally a considerable
time before it is formed ventrally — or perhaps it would be more
correct to say that it is formed anteriorly sooner than posteriorly.
It is still further complicated by the origin of the process of growth
in length which also begins first on the dorsal side.

Experimental observations show quite unmistakeably that as
soon as the blastopore lip is formed it begins to grow over the
yolk, and does not cease to grow in the dorsal region until the full
complement of segments have been formed. The blastopore lip con-
stitutes a ring-like area of proliferating tissue which is continuously
adding on new material to all previously existing material with
which it is in contact. It forms the deuterogenetic centre.

The series of sagittal sections, shown in Fig. 7, indicate the
character of the changes which take place internally and are
accompanied by surface views.

The slit seen in F running in from the groove shown in surface
view in B widens out in G and H to form a capacious chamber
which is the archenteron — the first formed part of the gut cavity.

Formation of the Segmentation Cavity 13

The segmentation cavity which is so large and conspicuous an
object in E becomes smaller and appears to be obliterated by
gradual compression owing to the advance of cells caused by the
expansion of the gastrulation slit.

At the same time it must be noted that there is some uncertainty

•/

as to whether the whole of the segmentation cavity is obliterated
in this way, or whether some of it may not remain and become
merged with the new archenteric cavity derived from the slit, and
thus in fact help to form the future gut cavity.

There is no doubt that the wall between the original segmenta-
tion cavity and the slit cavity becomes very thin. According to
Hertwig this wall breaks down in the frog. According to Brachet
it breaks down sometimes in Rana temporaria though not always,
but in Siredon it always does break down, as according to Brauer
it does in Hypogeophis. Hence we may say that in Amphibia
the segmentation cavity either disappears by compression of its
walls or it may become confluent with the anterior region of the
archenteron.

To compare this condition with Amphioxus we may refer to
the diagrams of sections drawn in corresponding positions. It
will be seen that a certain amount of invagination occurs. A
mark placed in the yolk plug at X should appear within the gut
cavity at a later period, Fig. 8.

There are no essential differences with respect to gastrulation
in the many genera of Anura and Urodela which have now been
studied.

In the case of the Gymnophiona however we have an egg
which is distinctly nearer to the meroblastic type than is that of
either of the other two Amphibian orders. It is almost a centro-
lecithal egg.

In the egg of Hypogeophis, which has been very well studied
and described by Brauer, the segmentation is partial; but there
is a very rapid superficial differentiation so that a circular blasto-
pore is formed very much as in the other Amphibia. The
segmentation at one time shows a cap of small cells upon a mass
of still unsegmented yolk, with here and there a nucleus near to
the segmented pole. There is no very well marked segmentation
cavity, but small cavities appear here and there among the small

14

Growth in length of the Vertebrate Embryo

cells, Fig. 26 A. Brauer seems to draw a sharp distinction
between "animal cells" and "vegetative cells." I cannot see the
importance of so doing. All his figures show that the superficial
layer of the whole egg very quickly becomes segmented or at
any rate it becomes covered with a layer of cells. Possibly this
may be to a certain extent due to a sliding of the animal cells
over the surface, but it is far more probably due to a progressive
differentiation,, as it is in the frog.

G

Fig. 8.

A, B, C, D, gastrulation stages in Amphioxus. After MacBride, 1910. E, F, G,
corresponding stages in Rana temporaria ; the white parts represent proto-
genetic ectoderm ; the black protogenetic endoderm, and the thick and thin
lines deuterogenetic endoderm and ectoderm respectively; sg segmentation
cavity.

However that may be, a stage is eventually reached when
there is a circular blastopore established, with lips continuous
all round and with a superficial layer of cells on the outside and
an inner cell mass, Fig. 27, p. 49.

In this stage the cavities which represent the segmentation
cavity, become confluent with a space caused by splitting between
the cells that radiate inwards from the blastopore lip. Thus in
Hypogeophis we have at this stage the future gut cavity well
defined into two regions : that which was in existence actually
before or at the time of the beginning of the blastoporic ingrowth,
and that which has been produced in connection with the growth
of that edge.

Formation of the Gut Cavity 15

Thus in Rana, Siredon, and Hypogeophis one sees very plainly
from anatomical observations that the gut cavity is due to two
processes.

A point to be noted is that when as in Amphioxus and Rana
there is a real segmentation cavity, namely a cavity produced
by the geometrical arrangement of the early segments, then this
cavity, as in Amphioxus, is roofed over by cells which will
eventually become ectoderm cells, and its floor is formed by cells
which will eventually become endoderm cells.

Before it becomes converted into gut cavity there is a movement
of segments so as to line the ectodermic roof. This movement
in part is no doubt due to pressure caused by the expansion of
the gut cavity, though partly perhaps to the pressure produced
by the growth backwards of the blastopore lips. It can be taken
as representing a modified form of invagination.

Invagination such as one sees in a yolkless egg (Amphioxus)
is impossible — therefore gastrulation is affected by some other
means.

It is doubtful to what extent the gut cavity is due to the
splitting referred to above.

This is rather an important matter; because all the part of
the gut cavity which is produced either by splitting or by any
modification of invagination, or by conversion of the segmentation
cavity (so catted) clearly differs from that which is produced by
the GROWTH BACK OF THE BLASTOPORE LIPS. The former is in all
cases protogenetic, THE LATTER is DEUTEROGENETIC.

The distinction between the two parts of the gut cavity is
quite distinctly seen in Axolotl and Hypogeophis, as Brachet's
figures of the former and Brauer's of the latter will demonstrate.

In Rana it is a little more uncertain. I have often tried to
test this point by experiment in Rana temporaria and I think it
can be done but it is not easy. It might be done by inserting
fine hairs into the yolk cells just behind the dorsal lip of the
blastopore and then tracing them after growth in length has
continued for some time.

It seems quite possible that the greater part of the gut cavity
of protogenetic origin is derived from the conversion of the
segmentation cavity into gastrula cavity by the movement of

16 Growth in length of the Vertebrate Embryo

those cells to which we alluded in the preceding paragraph. But
in any case there is undoubtedly some splitting, even if it is only
such as is sufficient to make the cavity derived from the segmenta-
tion cavity continuous with that derived from the overgrowth of
the blastopore lip.

It might be well to restrict the term archenteron to the proto-
genetic part of the gut cavity and to call the deuterogenetic part
metenleron ("neo-enteron" of de Lange), Fig. 9.

dl

Fig. 9. Rana. Diagram to show protogenetic and deutero-
genetic tissues with a dividing line A, B.

ecf en protogenetic ectoderm and endoderm ; ec'. en' deuterogenetic ectoderm and
endoderm; dl dorsal lip of blastopore ; vl ventral lip of blastopore; a archen-
teron ; m metenteron.

At first sight a difficulty arises because the cavity in question
seems to be roofed in by deuterogenetic, while it is paved by
protogenetic tissue. Really however the portion of the floor
which corresponds to the deuterogenetic roof is the little piece
near the letter v, Fig. 10. The yolk plug is the hypertrophied
floor of the protogenetic area or true archenteron, but it is pro-
jected out temporarily through the biastopore mouth. At a later
stage the yolk plug is actually withdrawn — this in itself may
perhaps be an invagination process.

The reason why the roof of the metenteron is so much longer
than the floor, is that the dorsal lip by which the roof is formed
comes into being so much sooner in the frog than does the ventral
lip.

For a time the whole annular area of deuterogenetic activity

Development of the Tail 17

continues, with the result that the metenteron is increased in
length and new tissue is added on to the back, sides, and ventral
walls of the animal, and to all organs which at this time are present,
and thus the whole embryo, gut, nervous system, notochord,
mesoderm, increases in length — by the addition of new material
posterior to all previously existing tissues.

The Tail

After a certain time however and again as a result of the
activity of the deuterogenetic centre, the lips of the blastopore
approach and coalesce, and this is followed almost at once by the
dying out of the ventral and ventro-lateral parts of that area of
cell production.

The dorsal region and dorso-lateral region however continue
their growth for a considerable time longer, but the chief parts
with which they are now continuous are the neural tube, the
notochord, and the dorsal and dorso-lateral parts of the ectoderm
and mesoderm.

The metenteron is on the whole more ventral, so that there
is no great tendency for the gut cavity to be continued backwards.
Also none of the ventral or ventro-lateral part of the body wall
is continued backwards. Growth in length ceases with regard to
these parts.

In some Amphibians (some Urodela) the blastopore partly
closes by the coalescence of its lateral lips, leaving the extreme
ventral and the extreme dorsal parts open. The ventral part
remains open as the anus after the dying out of the ventral region
of the deuterogenetic centre. The dorsal part becomes involved
in the folding up of the neural folds which adjoin its anterior and
antero-lateral margins and it forms in that way a temporary
communication between metenteron and neural tube. This is
called the Neurenteric canal.

In Kana and other' Anura the ventral part of the blastopore
closes along with the middle part; but the anus forms in the
closed area and, as the deuterogenetic area dies out almost
immediately, no growth in length occurs here.

Thus the dorsal part alone continues to grow and to add on
more and more segments, so producing the chordate tail, which

A. v. E. 2

18

Growth in length of the Vertebrate Embryo

released from the encumbrance and discomforts of a digestive
system has led to the wonderful success achieved by the vertebrate
phylum.

It is necessary to realize that the metenteron has a wide roof
and narrow floor, because otherwise it is difficult to interpret
sections taken vertically and transversely through the egg, e.g.
along the line A — B, Figs. 9 and 13.

mes

Fig. 10 alter de Lange, 1912. Diagram of embryo of Megalobatrachus.

an anus ; A archenteron ; M metenteron ; ec protogenetic ectoderm ; en proto-
genetic endoderm ; dl dorsal lip ; v ventral lip of blastopore ; mes protogenetic
mesoderm; mes' deuterogenetic mesoderm; nc neurenterjc canal; ec' deutero-
genetic ectoderm ; en' deuterogenetic endoderm.

The deuterogenetic tissues have in the Frog a more compact
and less yolky character than the protogenetic tissues. The two
regions of course merge one with the other along the lateral walls.

Thus we get the appearance of a special dorsal plate forming
the nerve plate and notochord, Fig. 10. It was this appearance
which caused Lwoff in 1894 to describe what he termed the
ectoblasto-genetic plate as a turning in of the ectoderm to form
the notochord and the roof of the gut cavity.

Experiments on Gastrulating Eggs 19

De Lange has found evidence of these two growth centres in
Megalobatrachus japonicus which quite agrees with what we find
in Rana and Hypogeophis.

De Lange suggests a further term tritogenesis to describe the
outer growth of the tail. But clearly the growth of the tail is
a simple continuation of the deuterogenetic growth, and it certainly
is not anything like so fundamental a change as that from proto-
genesis to deuterogenesis.

Experimental Observations.

The recognition of these two growth centres in the frog is due
to experimental work. It is possible by marking parts of the
surface of the gastrulating egg to realize movements which other-
wise escape one.

Observations by marking the surface can be made in various
ways. It is possible to cauterise a few cells and the scar thus
produced can be traced during a shorter or longer period of time
and thus the relative movements of parts determined.

Or bristles may be inserted, at any rate in the egg of Rana
temporaria, with many satisfactory results. A bristle makes an
unmistakeable landmark, which cannot be said of the scar made
by cautery. In cauterisation it is possible for the cauterised cells
to slough off: and to get carried away from the spot which they
are intended to mark. The bristle method it may be noted
cannot be used with Triton eggs. Then again stains such as Nile
blue sulphate may be used, as shown by Goodale, who stained
little groups of cells round the equator of the egg of Spelerpes
bilineatus shortly before gastrulation. The effect of this treat-
ment is to stain the yolk grains so that the cells go on dividing
and thus the individual descendants can be followed to a certain
extent. Apparently all eggs are not equally suitable for this
staining method. But experiments on Spelerpes and Amblystoma
were successful.

It may be said at once that the results of all these methods
of experimental research are contrary to the theory of concrescence,
whether they are made upon Elasmobranch, Teleostean or
Amphibian eggs.

By the bristle method it is possible in the frog's egg to make

2—2

20

Growth in length of the Vertebrate Embryo

out with considerable accuracy the area over which the main axis
of the embryo extends. It has been made quite certain that
there is a part of the embryo which is developed, so to speak,
in situ on the egg, and a part later in origin which is formed by
the activity of the blastopore lip.

These are the protogenetic and deuterogenetic regions respec-
tively.

c'

Fig. 11. Rana temporaria.

A, B, C, Eggs to show position of bristle; A', B', C', Longitudinal sections of the
same embryos after some daj^s; br bristle; bl blastopore; dl dorsal lip.
Drawings of eggs and tadpoles in which bristles have been inserted to deter-
mine the parts of the embryo derived from the primary and secondary growth
centres. In A the bristle was inserted 5 mm. anterior to the dorsal lip of the
blastopore at its first appearance. In B and C the bristle was inserted about
the same distances in front of the dorsal lip but at later periods.

By carefully marking various parts at different stages of
development we are able to arrive at the following conclusion.
The protogenetic area includes :

Forebrain, eyes, mouth, heart, nasal organs, pharynx,
and probably the front end of the notochord and perhaps the
mid-brain.

Experiments on Trout Eggs 21

The deuterogenetic area includes :

Hindbrain, spinal cord, renal organs, muscles of the trunk,
and the whole tail ; but probably not much of the gut except
the rectum and anus.

The accompanying Fig. 11 will show the kind of evidence
obtained by the method of using bristles as a means of marking
regions on the egg. It is difficult to obtain the exact boundary
between the influence of the two activities, because bristles
cannot be placed nearer to the lip of the blastopore than about
0-5 mm. without causing serious malformations. The bristles
were actually placed round the region of Brachet's blastopore
virtuel (see pp. 25, 26).

Experimental Evidence as to Concrescence.

The first to test this theory by experiment was Kastschenko,
who in 1888 operated upon the embryos of an Elasmobranch.
He destroyed one part of the rim and found that such destruction
did not prevent the formation of the main axis of the embryo.

The most complete investigation that has been made is that
of F. Kopsch who performed a large number of experiments upon
the developing embryo of the trout. I give here some figures
showing the nature of the operation and the results.

In Fig. 12 A the mark op represents in A the point in the rim
of the disc — i.e. lip of the blastopore, at which an injury was
effected. Fig. 12 B shows the position of the injury at a later
stage when the embryo had eight or nine pairs of protovertebrae.

According to the concrescence theory, the germinal disc rim
growing round should have coalesced with the corresponding part
on the other side, and formed the mid-dorsal part of the embryo,
namely the neural tube, notochord, and, perhaps the proto-
vertebrae. But, as the examination of sections taken through
B — b showed, the formation of these parts is quite normal, and
the injury lies far to the side — in the region of the latero- ventral
part of the future body wall. The injury has brought about a
failure on the part of the rim to progress over the yolk, and a
corresponding defect at the side — which is just what one would
expect if the embryo is formed in the way described.

Fig. 12 C is a figure of a slightly more advanced embryo

22

Growth in length of the Vertebrate Embryo

which had been operated upon in the same way as B. Here the
operation scar certainly appears to have moved towards the
mid-dorsal line. But if we come to examine sections, we
find that the mid-dorsal and dorso-lateral parts of the embryo
are quite unaffected. It is the part of the embryo ventral to
that which will be derived from the protovertebrae which is
inj ured.

B

Fig. 12 after Kopsch. Egg of trout. Diagram to show the
result of injury to the rim of the blastoderm.

A blastoderm; op injury; B the embryo some days later; C another specimen.

The notochord, neural tube, and even the protovertebrae are
normally developed, but a little of the tissue beyond — that is to
say, ventral to this point — is lost. This means that it is the side
or lateral body wall of the future animal which will show a
deficiency, and this is only what one would expect upon the non-
concrescence theory.

An injury made in the mid-dorsal line completely prevents
any addition to the mid-dorsal region of the embryo. If the
embryo were formed by concrescence this would not be so.

Experiments on Trout Eggs 23

An injury inflicted at a point between that of Fig. 12 A and
the tail bud brings about a defect in the dorso-lateral region,
leaving intact the median structures, such as the notochord and
the floor of the neural tube.

These experiments, to my mind, are quite conclusive against
the theory of concrescence or confluence in any form, but it is
only fair to say that this view is not held by all.

Even Kopsch, from whose work I have taken these figures,
believes that his experiments indicate a certain amount of con-
crescence for the hinder end. He shows how, for instance, as in
Fig. 12 C, that there is a movement of the point of injury nearer
to the median line. It is true there is a change of position with
reference to the yolk mass, but where is there any evidence that
the dorsal or dorso-lateral region is affected by such an injury?
It is in the much distended latero-ventral region of the young
embryo that the injury has been made, and it is in that region of
the fully- formed embryo that the injury persists. So long as the
notochord and the floor of the neural tube are uninjured by such
experiments as these, so long is there a complete absence of
evidence for concrescence or confluence.

Goodale stained groups of cells in the equatorial region just
prior to gastrulation. These groups lengthened out meridionally
and converged upon the blastopore.

This showed that as deuterogenesis effected its work, the lips
of the blastopore progressed leaving trails behind them which
showed not a trace of concrescence.

LECTURE II

We have been able to come to a fairly decisive conclusion with
respect to the growth in length of embryos of the Anamniate
Vertebrates. We have arrived at this conclusion chiefly through
anatomical observations and also from evidence obtained by actual
experiment upon certain fishes and Amphibia.

We have seen that the production of undifferentiated cells
which go to build up the general form of the embryo occurs around
two centres of activity. The first may be said to be around the
protoplasmic centre of the fertilised egg, and to begin with
segmentation and to end (so to speak) with the accomplishment
of gastrulation — this is the primary or protogenetic growth centre.
The second begins with the formation of the blastopore lip at
the commencement of gastrulation, and may be said to be
around a centre situated in the middle of the blastopore, for
the form which this centre of growth takes in the Anamnia is
annular.

This is the secondary or deuterogenetic growth centre. It is
placed eccentrically to the primary growth centre and the effect
of its action is the growth in length of the embryo, i.e. deutero-
genesis. New tissue is added on to all previously existing tissues
with which it is in contact.

Hence we are able to distinguish between protogenetic ectoderm
and deuterogenetic ectoderm, protogenetic part of the nerve tube,
and deuterogenetic part of the nerve tube, protogenetic part of
notochord, deuterogenetic part of notochord, and so on.

The clear appreciation of this fact renders intelligible the differ-
ences said to occur in the method of formation of the notochord and
explains how it can arise either from endoderm or mesoderm or
ectoderm. It does away also altogether with the idea of gastru-
lation occurring in two phases as Hubrecht, Keibel and some

Formation of Archenteron and Metenteron 25

other embryologists have asserted, because the two phases are
quite distinct, the first being the real gastrulation, the effect of
protogenesis, while the second so-called phase of gastrulation is
a totally different phenomenon, namely the growth in length of
the already formed gastrula by accession of material from behind,
so that a spherical form is converted into a cylinder.

In all the Anamnia the rule is that the gastrula cavity or
archenteron is formed with an opening to the exterior — the
blastopore. In all the Anamnia the form of the deuterogenetic
proliferating area tends to be annular and the blastopore is the
hole in this ring.

Conditions in Amniota.

In the Amniota we find a very different state of affairs.

It is extremely doubtful if the part of the gut cavity which
I have termed archenteron (Fig. 13), as distinct from metenteron
ever forms in connection with a blastopore.

An opening between it and the exterior may occur later but
that is of a different character and it is really not an archenteric,
but a metenteric opening. These two openings are what Brachet
calls blastopore virtuel and blastopore reel, Fig. 13.

The metenteric opening or notopore of some authors (blastopore
reel) (of which the more dorsal part becomes neurenteric canal for
a while) subsequently disappears, and the ventral part becomes
the anus.

In the Amniota the archenteron always arises as a split either
between cells, or else as vacuoles among cells which, will be
endoderm cells; and in most cases it does not arise so much as
the result of the geometrical effect of segmentation (as in
Amphioxus or the Frog), Fig. 6, as by internal secretion of fluid,
Fig. 15 C.

There is only one exception to this, and this a very obvious
exception indeed, namely, the segmentation of the Marsupial
mammals. But the process in this case is very different from
that of any other vertebrate, though at the same time it may be
said that it can be derived fairly easily from a meroblastic type
of egg similar to that of Prototheria or Sauropsida.

26

Growth in length of the Vertebrate Embryo

In Anamnia the archenteron is formed at the same time that
the deuterogenetic centre begins to come into existence.

In Amniota the archenteron is formed before the deuterogenetic
centre has come into existence.

In Anamnia the archenteron is formed either by invagination
of one wall of the blastula within the other, or by a process of
splitting or delamination, producing a split which is always open
to the exterior, or by a combination of both processes. In

a

Fig. 13. Diagram to show the opening of the blastopore in Anamnia — the line
a — b divides protogenetic tissue (left) from deuterogenetic (right).

Dots and fine lines protogenetic endoderm and ectoderm respectively; broken
lines = deuterogenetic endoderm; thick lines = deuterogenetic ectoderm; the
bracket on the right marks the metenteric opening, or notopore (Brachet's
"blastopore reel"), the bracket to the left marks the archenteric opening
('* blastopore virtuel ").

Amniota the archenteron is formed by internal secretion of a

«,

cavity among or within the yolk cells.

These differences are in correlation with a physiological process
which differentiates the mode of development very sharply
indeed from that of the Anamnia. This is the fact that in an
early stage of development, a fluid is secreted into the two great
cavities of the embryo, namely the archenteron and the coelom.

This condition of oedema may be said to be the physiological
essential cause of the formation of the amnion and allantois,
which are the most distinguishing features of the Amniota as
compared with other vertebrates. If this is so we can see why

Formation of Coelomic Cavity

27

there should not be a blastopore during the time of the formation
of the archenteron. If there were a blastopore at this time, clearly
there could be no cavity produced by swelling.

The amnion is essentially an organ of the coelom. This view has
been taken by Hubrecht in his paper upon the " Early development
of the Hedgehog and its significance for Phylogeny of the Amnion."

But it has also probably arisen in correlation with a distended
archenteron.

In Elasmobranchs there is a very heavily yolked egg, but there
is no amnion. Nevertheless it is very interesting to note in the
Elasmobranchs a condition which might very easily lead to the

am

A B

Fig. 14. A, Scyllium; B, Amniote.

Diagrammatic section. Black = endoderm ; dotted line = mesoderm ; line barred
with white = ectoderm ; coe coelom ; am amnion ; arch archenteron.

formation of an amnion of the amniotan type. In Dogfish
embryos of about half an inch in length, at a time just after
the closure of the blastopore, there will be found two large
coelomic cavities upon the yolk sac, that is to say cavities of the
so-called exembryonic coelom, which are morphologically and
physiologically in the position occupied by the lateral folds of the
amnion of an amniote (Fig. 14).

In Scyllium there is no accumulation of fluid within the gut
cavity, and the rather long yolk stalk keeps the so-called embryo
part well away from the yolk sac.

But a very slight alteration of conditions could convert this
into an amniote as Fig. 14 illustrates.

28 Growth in length of the Vertebrate Embryo

There is one other point to be insisted upon in considering
this question of growth in length in Amniota — not only is there
this fluid enclosed in the archenteron, but there is an increasing
amount of pressure, so that we have here to deal not only with
a hydrostatical condition but also with a hydrodynamical con-
dition, and this must have its effect upon the mechanics of
development.

The development of the Rabbit.

We will now take the formation of growth centres in the
embryo of the Kabbit in more detail. This animal is chosen
because of all vertebrates none other has been so far described
that shows in so distinct a manner the relations of the primary
and secondary growth centres.

We know that the blastocyst of the Rabbit attains a large size
from causes within itself before it becomes influenced by the
walls of the uterus, for this reason it is more suitable than are the
blastocysts of most mammals, and because of the absence of yolk
it is more suitable than those of birds or reptiles, for study of
growth centres.

In the accompanying Fig. 15 the early stages in an ordinary
Eutherian are shown — such as that of the Pig, which we may take
as typical of the sub-class — in which there is some, but very little,
indication of what has been called inversion (entypy).

The chief points are that the egg has but little yolk though
some is present in the form of fat drops. The segmentation is
holoblastic, the four segment stage is criss-cross, Fig. 15 A, and
the result of segmentation is a solid sphere of cells all much
of the same size— this stage is termed the morula, Fig. 15 B.
A cavity arises by the formation of vacuoles within some of the
more internal cells, Fig. 15 C. The vacuoles run together, the
fluid increases and the cavity gets larger and larger till it occupies
a position that leaves only a heap of cells at one pole, and a layer
reduced to a single cell in thickness elsewhere, Fig. 15 D.

The heap of cells becomes flattened, and three layers are
distinguishable at this point ; the outer is termed trophoblast, the
middle ectoderm, the inner endoderm, Fig. 15 E.

Early Stages of the Pig

29

The middle layer or ectoderm is distinguishable earlier, and in
some mammals very much earlier, than the other two are dis-

B

Fig. 15. Pig.

A, 4 segment stage; B, morula stage; C, blastocyst showing vacuoles; D, 4£ day
blastocyst; E, 7 day blastocyst; al albumen layer; ec ectoderm; en endo-
derm ; / oil drops ; vac vacuoles ; tr trophoblast ; im inner mass.

30 Growth in length of the Vertebrate Embryo

tinguishable from each other. In some mammals it may be
different, as in Manis and Galeopithecus, where it is said that the
trophoblast is to be made out before the ectoderm and endoderm
are distinguishable from each other.

Then the circular ectoderm plate becomes bent inwards and by
its growth causes the rupture of the trophoblast above it, thus
allowing the plate to appear on the surface; pieces of trophoblast
upon the surface degenerate and disappear.

All this time the symmetry has been markedly radial or
spherical. The cavity of the blastocyst becomes the gut, or at
any rate the fore gut is derived from it. It is the protogenetic
enteron or true archenteron. There is no blastopore. Up to now
everything that has happened has been the result of protogenesis
and the result is on the whole spherical or radial.

In the Babbit the history up to this point is much the same,
the chief points of difference being:

(i) The egg contains protein not fat.

(ii) There is a very tough albuminous layer laid on during
the passage of the ovum down the Fallopian tube, and this acts
like the leather coat of a football in keeping the more distensible
inner vesicle taut, and also it prevents the embryo (blastocyst)
from coming in contact with the walls of the uterus for some
considerable time.

(iii) There is no trace of inversion. The rupture of the
overlying trophoblast is general.

(iv) The endoderm does not completely surround the inner
surface of the trophoblast.

(v) The archenteron arises more as a split between cells than
as vacuolations within cells.

The failure of the endoderm to line the ventral part of the
blastocyst looks as though there were no ventral wall to the gut
cavity if — as some embryologists think — the trophoblast is
ectodermic. Of course if it is endodermic or homologous with
yolk cells, then there is no such difficulty.

We can now consider the next stage in development. Proto-
genesis has come to the end of its solitary reign. The radial
symmetry to which it gave rise now becomes profoundly modified
by the origin of a secondary growth centre resulting in growth

Early Stages of the Rabbit

31

in length. The conditions in the Rabbit are such that the plane
of the deuterogenetic area coincides at first with that of the
protogenetic area, instead of being at right angles to it as in the
Anamnia.

This is (i) partly on account of the great distension of the
archenteric space, and (ii) partly because the deuterogenetic
centre first shows itself as a circular spot instead of a ring.

In the Rabbit we have the stages of development of the two
centres shown in quite a diagrammatic form in Fig. 16.

During the seventh day the circular embryonic area is marked
out very distinctly in front, but less distinctly behind.

Fig. 16. Rabbit seventh and eighth day. Diagram
to show the secondary growth centre.

The white areas are protogenetic, the shaded areas are deuterogenetic, the doubly
shaded portions represent the primitive streak.

The embrvonal area is coterminous as vet with the ectoderm.

u •/

The surrounding superficial cells are all trophoblastic. Internally
the endoderm lines both the trophoblast and ectoderm. About
the time that the deuterogenetic centre begins to appear, there is
a slight thickening of the endoderm just under the anterior and
lateral margins of the ectoderm. This is the first sign of any
mesodermal formation, but no cells are actually budded off here
until the deuterogenetic tissues have become apparent. This
semicircle is however the first sign of the mesoderm from which
the heart and pericardium will be developed, but we will confine
ourselves at present to the external characters, that is to say
to the general form. The deuterogenetic centre begins as a

32 Growth in length of the Vertebrate Embryo

proliferation of cells of the outer layer of the hind part, probably
of the actual edge of the embryonal area.

The proliferation is from the under surface. But we must
presume also an addition of cellular units to all tissues in contact
with the proliferating area. The cells which are proliferated
within give rise chiefly to mesoderm and anteriorly also to noto-
chord, that is to say deuterogenetic notochord. We will return to
this point later.

There can be no doubt that here is an area of intense cell
divisional activity.

This centre of activity has apparently arisen suddenly. It is
close to but not coincident with the centre of the primary centre,
which we may take as being practically in the centre of the circular
embryonal area.

Now what must be the effect of the counter action of these
two centres? We must consider them for the moment apart
altogether from any phylogenetic point of view. Here we have
two actively productive centres of tissue, one of which is especially
active. They lie on the same plane. They are subjected from
within to an ever increasing hydrostatical and therefore a hydro-
dynamical pressure. Surely the result must be that they tend to
elongate along the lines joining their respective centres.

If one area is more yielding than the other then that area will
be elongated the most.

What are the conditions here? Inside there is an increasing
hydrodynamic agent. This is shown by the way in which the
albumen layer is kept taut and becomes thinner, and later it is
seen by the expansion of the uterine wall. Therefore the cells
are stretched also. But where there is also a tendency to expansion
due to cell multiplication — there will be an additional tendency
for the cells to stretch.

Again if the cell multiplication is greater in one area than in
the other, the area in which the cell multiplication is greatest will
give way more easily than the other.

The deuterogenetic area seems to be an area in which cell
multiplication is very intense, and therefore we find that this part
expands more markedly than the protogenetic area.

The result is that the posterior half of the embryonal area is

Development of Rabbit

33

drawn out,, but the anterior half of the deuterogenetic area is
drawn out to a still greater extent,, the latter appearing as a conical

HKCS^^^S^uaE/
en

^"^J^^^kB^

t''~**v3$'i$f

&.---&^y^^&£5zz^^

Fig. 17. Rabbit embryonal areas.

D is a transverse section through A, 1 72 hours ; # through 5, 180 hours ; F through
C, 192 hours; in A the primitive streak is elongated and slightly grooved;
in B the primitive streak is at the maximum length and the groove at maxi-
mum depth ; in C the groove has disappeared, en endoderm ; ec ectoderm ;
mesmesoderm; ps primitive streak; prg primitive groove.

A. V. E.

34 Growth in length of the Vertebrate Embryo

or wedge-like strand of thickened proliferating tissue running into
the also drawn out but less conspicuous hinder region of the
protogenetic embryonal area.

This proliferating area is what we term the "primitive streak."

^^HM^^^^ta

Along it a groove soon appears.

Why should it be grooved ?

Let it be noted that in the first place it is not grooved until
it has attained a considerable length. There is no groove at first.

Secondly the groove disappears long before the streak dis-
appears.

The difference in the characters at these different times is
seen in sections, Fig. 17.

It is at least possible that the grooving may be very largely
due to mechanical causes.

As the deuterogenetic mesoderm passes to either side by being
crowded out in the median line, it will tend to be drawn away
from the median line by the general expansion of the blastodermic
vesicle which is still going on. This must place some tension
upon the under surface of the proliferating area, if there is any
continuity, so that each side will be drawn away. The result is
that a groove arises along the middle of the streak, except at the
anterior end where there is less lateral tension, and where there
is also an opposing compression in front, due to the antagonistic
effects of the protogenetic and deuterogenetic centres.

Posteriorly also there is no groove or if there were any it would
tend to be transverse, as indeed it is actually found in the Sparrow.

The part of the deuterogenetic proliferating area which will
be most expanded on this hypothesis is the latero-anterior, or,
speaking morphologically, the dorso-lateral part, that is to say
the part equivalent to the dorso-lateral part of the blastopore lip
of the Frog, Fig. 18.

Thus we see that the part which is most expanded is the part
which in the Frog gives rise to the neural plate and mesodermic
somites.

If we look at figures of the Rabbit's blastoderm from the time
the primitive streak first appears until the time of the formation
of the first 7-8 proto vertebrae — we can see throughout the following
features very well defined.

Comparison of Rabbit and Frog

35

(i) The general radial character of the protogenetic area,

(ii) After a certain moment the primitive streak becomes
contracted — shortened and thickened,

(iii) An additional part of the embryo has been intercalated
between the primitive streak and the radial protogenetic area.

This intercalated region is approximately all that part of the
embryo which includes the protovertebrae. Now we find that in
the Amphibian (Frog) it is just that part which is formed from the
lips of the blastopore.

If we can assume that this intercalated region has been added
by, and at the expense of, the primitive streak, we arrive at the

v

A B

Fig. 18. A, Frog; B, Rabbit.

Corresponding deuterogenetic proliferating areas; vl ventral;
dl dorsal lip of blastopore.

conclusion that physiologically the primitive streak is identical
with the blastopore lip region of the Anamnia.

To prove this by experiment is probably impossible in
Mammals, but in Birds the supposition has been put to a practical
test.

If we can assume that the primitive streak of the Bird is the
same as that of the Rabbit, then we can say that experiment
proves that the primitive streak is converted into the meta-
merically segmented part of the animal, but that the anterior
part of the animal is developed quite independently of the
primitive streak.

3—2

36 Growth in length of the Vertebrate Embryo

Thus again we find that in the Mammal and Bird, there is a
protogenetic and a deuterogenetic region; and that they corre-
spond to one another.

The protogenetic region gives rise to the fore brain, the mouth,
the pharynx, the heart and pericardium, and the deuterogenetic
region to all the rest of the animal in both cases.

Fate of the Primitive Streak.

We have so far considered the question of the fate of the
primitive streak generally. We must now take it more in detail.

A very interesting paper has recently been published post-
humously which was written by E. van Beneden in 1889 on
the development of the Rabbit. The work was begun in 1881-2,
when the figures were drawn, and was published in 1912.

The view van Beneden takes is that the first formed didermic
condition of the blastocyst of the Amniota is not comparable to
the didermic condition of lower vertebrates.

The inner layer which is usually called the endoderm or
hypoblast is, according to van Beneden, not homologous to the
endoderm of Amphioxus. It is something else and he calls it
LecithopTiore.

As I will shew presently, what van Beneden calls hypoblast is
what I should call deuterogenetic endoderm. His lecithophore is
what I term protogenetic endoderm and this, according to my way
of looking at things, is very clearly comparable to the true endo-
derm of the lower chordates.

Further, van Beneden takes a very different view of the
development of the Babbit from that outlined above. He takes
no heed of internal hydrodynamic pressure, or tensions, or any-
thing of that kind. To him the cells at the lower pole of an
expanding blastodermic vesicle of the Rabbit, are squamous not
because of the pressure from within to which they are subjected —
but he thinks on the contrarv that the blastodermic vesicle

\r

expands because the cells stretch themselves out.

He does not regard the cavity of the vesicle as archenteron,
nor the inner layer of cells as endoderm and says that "Gastru-
lation': does not occur until after the primitive streak has

van Beneden on Rabbit

37

been formed, and also that the canal of Lieberkiihn is the
invagination cavity and is therefore archenteron. In his view
the blastocyst of a six-day Rabbit before the primitive streak
begins to form is not a gastrula, but a blastula, Fig. 19. The

Fig. 19. Diagrams to illustrate van Beneden's interpretation.

A, Mammal; B, Elasmobranch. bl hlastodisc; m merocytes; the lined area Is

the lecithophore.

inner layer of cells together with the fluid van Beneden calls
lecithophore, and he says they are homologous to the unsegmented
yolk of an Elasmobranch with its merocytes.

Fig. 20 after van Beneden. Rabbit eight days. Plate XIV, Section 14.

Transverse section through embryonal area, end (endoderm) archenteric
plate of van Beneden; lee lecithophore.

Clearly if Riickert is right in ascribing the origin of the
merocytes to the accessory spermatozoa, this homology will not
hold good, but it might apply to the Teleostean egg.

Gastrulation according to van Beneden occurs by invagination,
producing the groove along the primitive streak and the canal of

38 Growth in length of the Vertebrate Embryo

Lieberkithn. This area which " invaginates " is said to be endo-
dermal and since it clearly does not invaginate at the posterior
end, for there is no groove there — it constitutes the equivalent of
a yolk plug. Therefore after invagination we have the relation
as shown in Fig. 20.

Van Beneden tries to show that the whole of the notochord
and gut lining is derived from this archenteric rudiment and he
brings forward some very remarkable evidence from the corre-
sponding stages in the Bat in support of this. But even with
sections which favour his views there still remains a considerable
amount of gut lining which must be derived from the "lecitho-
phore."

Thus Fig. 21 which is taken from Fig. 15 of his Plate XXI.

Fig. 21 after van Beneden. Bat.

Sagittal section through the anterior end of an embryo with six protovertebrae.
ar pi remains of archenteric plate projecting into the foregut: lee lecithophore.

It is however not at all certain that the remnant of archenteric
plate has not shifted its position. It is degenerating tissue and
may quite well move, as I have seen degenerating trophoblast
move in the Pig.

Again the pericardium is very clearly indicated at quite an
early stage by the thickening of the inner layer (lecithophore of
van Beneden) and this layer certainly occupies the position
marked lee in the above figure.

van Beneden on Bat 39

Van Beneden regards the whole of the notochord, the whole
of the gut lining and the whole of the mesoderm, with the possible
exception of the endothelial lining of blood vessels, as being derived
from the primitive streak, i.e. by invagination. The primitive
streak is converted into the embryo by concrescence of its lips.
Even if the primitive streak does become converted into " embryo ':
by concrescence, this does not necessarily prove that the original
blastopore existed along the dorsal surface of the ancestral
chordate.

Because :

(i) The primitive streak appears in the first instance as an
already closed blastopore.

(ii) The grooving is I believe due to mechanical causes.

I would interpret the bat condition (Fig. 21) as follows.

All van Beneden's archenteric plate I regard as deuterogenetic.
His discarded " lecithophore " is protogenetic. His figures of bats
certainly do not support his contention that no notochord is
formed from the lecithophore, (v. Plate XVIII, fig. 7, 8). There
is evident thickening along the middle line due to proliferation
of nuclei at right angles to the surface.

Gastrulation and Growth in Length in Birds.

The segmented part of the ovum at the end of seven hours
after fertilisation consists of a little cap of cells occupying the
centre of the upper pole. These cells have been produced by the
segmentation of the germinal disc. They may be said to be in
cellular continuity with the so-called unsegmented yolk mass around
and below them. We may as well term this cap of cells the blasto-
derm. It is important to remember that this cap of cells is
continually being added to peripherally and perhaps beneath as
well. It should be noted that whenever a new segment is cut off
from the peripheral yolk mass the formation of such a new cell is
preceded by the division of a nucleus, and while one daughter
nucleus is contained in the new cell thus cut off and added to the
blastoderm, the other daughter nucleus remains behind in the
yolk mass, and after awhile it again divides with the result that
another cell is added to the blastoderm, and so on. Therefore

40

Growth in length of the Vertebrate Embryo

ao

end th

morphologically we must look upon the "yolk mass" as being
merely the more ventral segment of a fully segmented ovum.

If we examine the egg of a sparrow just after it has been laid
it has in many cases very much the appearance just described.
There is however usually a slight cavity below the lowest layer of
small cells and the general yolk mass. In the case of the new
laid egg of Gallus the cavity is much more evident.

The blastoderm consists of a definite outer membrane of cells,
of practically a single layer in thickness in its central part though
many layered round the peripheral area. A distinct cavity,
Fig. 23 sg, filled with fluid and containing many more loosely
arranged cells underlies the thin central part of the outer membrane.
In surface view this central part appears dark owing to the trans-

lucency of the membrane itself
and the cavity into which we are
looking; just as a window ap-
pears dark when looked into
from outside. The thickened peri-
pheral part, as it is opaque
and rests on a solid yolk, looks
lighter in a surface view, Fig. 22.
These are the areas of the blasto-
derm known as area pellucida
and area opaca respectively. In
section one would always see in
addition to the cellular membrane, the non-cellular vitelline
membrane extending over blastoderm and yolk alike.

The cavity containing the fluid lying under the area pellucida,
is usually called the subgerminal cavity, and it continually in-
creases in size by further accumulation of fluid as development
proceeds. It is in fact the future gut cavity, and ought more
strictly to be called the archenteron. Until recently it has
generally been assumed that it arises in the bird's egg just as
the corresponding cavity arises in the reptile's egg, by the
gradual accumulating of fluid between the small cells of the
segmented blastoderm and the large cells forming the unseg-
mented yolk.

Then as the result of the gradual multiplication of the cells

Fig. 22 after Schauinsland. Sparrow
egg new-laid.

ap area pellucida ; ao area opaca ;
end th thickening of endoderm.

Germinal Layers

41

which lie between the outer layer and the yolk, and their arrange-
ment into a definite membrane, an inner sheet of tissue is pro-
duced, forming eventually the gut epithelium.

This, at any rate, is the condition attained after a few hours
of incubation and as soon as this has been accomplished we can
recognise the presence of what are known as the primary germinal
layers, and the process of gastrulation has been finished. Gastrula-
tion is simply the formation of a cavity which is the future gut
cavity, and the lining of this cavity by a definite layer of cells,
the future epithelium of the true gut or mesenteron.

There is as yet, Fig. 23, no opening to the exterior, no mouth
nor anus.

ec-

ABC

Fig. 23. A, Bird in oviduct; B, bird laid; (7, bird after short incubation.
ec ectoderm; end endoderm; sg segmentation cavity.

The outer membrane about this period becomes dissociated
from the underlying layer, at any rate at its margin, and it
extends beyond it and creeps slowly over the yolk.

The names of these several layers and parts are as follows.

The outer layer is called the ectoderm, the inner the endoderm.

The endoderm is obviously continuous at its margin with the
general yolk mass, and is greatly thickened here.

The thickened margin is known as the germinal wall, the
cavity as the subgerminal cavity, or archenteron.

If we take the account just given of the origin of the gut
cavity in the bird's egg as correct, it will fit in very well with the
origin of the gut cavity in both the other great classes of Amniotes,
the Reptiles and Mammals.

The condition of a solid morula is converted into the condition
represented in Fig. 24, by accumulation of fluid in spaces among

42 Growth in length of the Vertebrate Embryo

the lower layers of the segmented germinal disc, or between them
and the largest yolk segments forming the basal part of the ovum.

An entirely different account of gastrulation in the Pigeon's
egg has been given by Patterson, which if confirmed would
necessitate an entirely different interpretation of the facts so
far described. But it is extremely difficult to reconcile his
account of the early development of the bird's egg with the facts
of Mammalian and Reptilian embryology, and I am very doubtful
as to whether Patterson's conclusions can be accepted (v. my
paper on " Gastrulation in Birds," Q. J. M. S. Aug. 1912).

The chief point in dispute concerns the exact mode of formation
of the lower layer or endoderm. That we must leave as uncertain,
only noting that the anterior edge of the germinal wall is reached
by the advancing edge of endoderm at about the time of the
laying of the egg.

All investigators however are agreed upon the structure of
the blastoderm after laying and during the first hour of incuba-
tion, and we may now proceed to consider its later history.

The freshly laid egg of Gallus if it has been fertilised, shows the
blastoderm lying on the upper pole of the yolk. In this the
following parts as already noted are discernible :

Area pellucida (through which the nucleus of Pander can
often be seen) ;
Area opaca.

Surrounding the margin of the area opaca a slight groove
usually runs separating the blastoderm from the surrounding yolk.

Sections show that this blastoderm consists of a segmented
disc lying loosely on and connected with the sub-lying yolk, and
having an outer well-defined layer of cells, and under this a large
number of loosely arranged cells, which form an incomplete roof
to a distinct cavity between them and the yolk. This is the
subgerminal cavity. There is a good deal of variation in the
state of the egg when laid, sometimes it is further advanced than
this. Also some blastoderms may be broader, measuring quite
a millimetre more than others. In such cases there is a greater
attenuation of the cellular layers, and the subgerminal cavity is
much larger.

The looser inner layer of cells when the section is taken along

Development of Chick 43

the direction of the future main axis, is thicker towards its posterior
end.

Of course after laying there is an almost complete cessation of
the developmental processes as the egg cools, so the egg remains
in a dormant state, though in hot summer weather development
makes some slight progress. Then as soon as incubation begins
the cells again become active and the changes effected are rapid.
During the first few hours the chief change is in connection with
the margin of the blastoderm.

The groove seen in the unincubated egg quite disappears and
the outermost layer of cells extends over the yolk beyond the
zone where nuclei lie scattered within the yolk. This layer is now
quite distinct from the underlying cells throughout the whole area

ect

lUiilK

4rogoo^°0°Vs

Fig. 24 after Duval. Section through blastoderm of chick

incubated ten hours.

A , anterior ; P, posterior, ect ectoderm ; end endoderm ; mes mesoderm ;
6V7 segmentation cavity; y yolk (subgerminal).

and it constitutes the ectoderm. Its progress in the future over
the yolk is by its own interstitial growth. It receives no additional
cells or nuclei from the volk.

*/

The loose cells below have by the 10th hour of incubation,
Fig. 24, formed a continuous sheet of tissue, this is the endoderm
and it is connected all round its margin with the yolk, which there
contains many nuclei, and it is called the germinal wall. According
to some authors, Duval, Marshall and Balfour, some of the cells
of this loose mass are left between the sheet of endoderm and the
ectoderm, and they are said to be more numerous in the anterior
part of the area pellucida. On the other hand other authors
(e.g. Lillie) deny the presence of these cells.

There can be no doubt about their presence in the Sparrow,
where they form a many layered mass circular in outline in the

44 Growth in length of the Vertebrate Embryo

anterior part of the blastoderm. These are the first traces of the
middle layer of cells or mesoderm.

The ectoderm even at the time of laying is thicker near the
centre of the area pellucida than elsewhere. In Gallus after some
5-10 hours of incubation this thickening is more marked still over
a wide area occupying the more anterior part of the area pellucida,
and a little later still a much more decided thickening can be
seen in the more posterior part of the area pellucida extending
from about the centre of the area backwards nearly to the margin
of the posterior part of the area opaca. This thickening is the
first sign of a structure which has long had "the name of primitive
streak, but it should be recognised at once as the growing point
for the increase of length of the embryo, while the less well-defined
thickening of the ectoderm in front is the beginning of the plate
of tissue which will ultimately form the anterior part of the
central nervous system — the fore and perhaps mid brains.

Such is the condition at the llth hour of incubation, Fig. 25.
Transverse sections through these two regions reveal certain well-
marked differences. A transverse section through the anterior
region along A-B shows us a decided thickening of the ectoderm,
but it is a wide diffuse thickening, having a quite sharp inner
surface; a few scattered cells are present between the layers
which are the mesoderm cells already referred to : while a section
along C-D shows a much more restricted area of thickening
which has a ragged inner surface from which mesoderm cells
are being actively proliferated.

Up till now in surface view the whole blastoderm has appeared
circular, as also the area opaca and the area pellucida. But now
the outline of the area pellucida becomes elongated and it can be
clearly seen that this elongation of the area pellucida is an
elongation backwards, and that this elongation involves the
elongation of the primitive streak. By the 18th hour the primitive
streak has stretched backwards and it now reaches from the
centre of the original circular area almost to the area opaca. It
is a more or less straight line, and is deeply grooved along its
whole length except at the two ends. The groove is deepest at
the anterior end, and it ceases abruptly against a rather well-
marked mass of cells which is sometimes called "Hensen's knot"

Development of Chick

45

or the primitive knot, but posteriorly it gradually shallows out
into a rather expanded thickening which forms the end of the
streak.

In front of the primitive streak the wide thickened area of the
ectoderm is now more sharply marked out from the rest of the
ectoderm. It constitutes the neural plate and is traversed longi-

end

. -•••..>-•.-. •/ •••»"**»*dLJ"ilV J

** »°»- ^/- — "^ ^^ ^ ^ fci* *• ** ^

«° •-•'*< **^° --- •

^

O o ° ••'••**• ••• ..... o

B

Fig. 25 after Duval. Chick at llth hour of incubation.

(7, surface view; A and 5, sections taken along the lines A — B and C — D
respectively, ect ectoderm; mes mesoderm; end endoderm; sg subgerminal
cavity; y yolk; ps primitive streak; ap area pellucida; ao area opaca.

tudinally by a rather wide shallow groove, the neural groove.
This groove is absent just in front so that posteriorly the neural
plate is partly divided by the shallow neural groove, but in front
the two halves are united.

The study of a series of transverse sections makes the features
peculiar to all these several parts clear and also explains their
relations to the whole egg and to each other.

46 Growth in length of the Vertebrate Embryo

The most important feature is perhaps one which it is difficult
to see in the chick in surface view.

This is the fact that the primitive streak is clearly an area of
intense cell proliferation and that cells are being continually
budded off from its whole length into the space between the
ectoderm and endoderm ; and that these are spreading out to the
sides and behind and in front as well. This is also mesoderm
but mesoderm having a different origin to that which was
mentioned before, as occurring in the anterior region of the
blastoderm, cf. Fig. 25^4.

Experimental proof.

It is possible to apply just the same tests experimentally to
the eggs of Birds as have been already described for the eggs
of Frogs.

The Bird's egg may be opened by removal of a small piece
of the shell, a mark made on the blastoderm, and the shell
replaced.

Or the whole egg, yolk and albumen, may be turned out into
a suitably sized vessel, and the yolk kept down below the surface
of the albumen by means of a glass ring ; the bristle or hair can
then be inserted and the vessel, after being covered with parch-
ment or even with a glass lid, can be placed in an incubator,
where development will proceed for two or three days almost
normally.

By this means it will be found that part of the animal is
formed in front of the primitive streak, that is to say by proto-
genesis, and part by the activity of the primitive streak, that is
to say by deuterogenesis, and that the two areas correspond to
the two parts derived from those centres in the Amphibian.

LECTURE III

The conditions in the Reptilia are in some respects intermediate
between those of the Amphibia, Birds and Mammals. In each
case there is no early formed blastopore and the archenteron is
formed by the accumulation or secretion of fluid among the yolk
cells, that is to say among the endoderm cells.

This is very clearly shown in Will's figures of Platydactylus.
There are the same primary and secondary growth centres at
work, but the conditions under which they work are quite different.
The difference in the conditions which determine the difference
in effects seems to be (i) the less abundant, and less rapid accumula-
tion of fluid or perhaps it would be more accurate to say the less
early accumulation or secretion of the fluid among the endoderm
or yolk cells; (ii) the form of the secondary "centre" is discoid
instead of annular. The condition is intermediate between that
of an Amphibian like Hypogeophis, and a Mammal, as Fig. 26
will illustrate.

There would seem to be an accumulation of fluid among the
endoderm cells in Hypogeophis, but this never causes a swelling
because the space wherein it accumulates opens at once to the
exterior by the "blastopore" shown in Fig. 27.

In the Bird and the Mammal, as we have seen, the fluid
accumulates very rapidly and there is no way for its escape, so
the cavity enlarges, becoming the subgerminal cavity in the one
case, and the blastocyst cavity in the other. In both cases it
eventually forms the anterior part of the gut cavity. It can
therefore be called the archenteron.

In the Reptile there is an undoubted accumulation of fluid as
in the Mammals and Birds and there is at this time no blastopore

48

Growth in length of the Vertebrate Embryo

opening — but there is a difference and this is a difference of
degree. In the Keptile the accumulation is slower.

If we compare the stages of Mammal, Bird and Reptile at the
moment when the deuterogenetic centre becomes evident we find

fi

1

Fig. 26. 1. Hypogeophis after Brauer, 1907. 2. Platydactylus after Will,
1892, showing little fluid. 3. Vespertilio after van Beneden, 1912, showing
much fluid, fl fluid.

very marked differences. In the Mammal the blastocyst cavity
is already large and is rapidly growing by continued increase of
hydrodynamical pressure producing tension in the layers above.
In the Bird the same condition is almost as evident. It is only
masked by the presence of the yolk mass, Fig. 28.

Reptile compared with Mammal and Bird

49

But in the Reptile at this stage there has been very little
accumulation and therefore presumably very little tension.

The accumulation of fluid does occur later, but not until the
deuterogenetic centre has become well established, and the result

bl

Fig. 27 after Brauer, 1907. Longitudinal section
through the egg of Hypogeophis.

bl blastopore; ec ectoderm; en endoderm; gc future gut cavity.

is therefore a mass of cells in the form of a heap instead of a long
drawn out streak, as in the case of Mammals and Birds.

When the accumulation does come, that is to say when the
stretching is produced which results in the rupture of the posterior

A B

Fig. 28. A, Mammal; B, Bird.

Diagram to compare corresponding stages in the formation of the segmentation
cavity, be blastocyst cavity; sg segmentation cavity.

part of the roof of the subgerminal cavity or archenteron, there
occurs at the same time a perforation within the deuterogenetic
tissue which corresponds in position with the metenteric opening
referred to in a previous paragraph.

A. v. E. 4

50

Growth in length of the Vertebrate Embryo

At the same time the deuterogenetic centre has been acting
just as it does in an Amphibian; that is to say dorsally it has
been adding on new tissue to the previously existing tissue, new
ectoderm to the primarily formed ectoderm,, new neural plate to
that previously formed and so on, producing growth in length.

This dorsal region which is homologous to the dorsal lip of
the blastopore therefore grows backwards, just as the dorsal lip
of the blastopore of the Frog grows back, and in some cases there
results a condition very similar to that found in the Frog. But in
spite of the similarity in appearance there is an all-important
difference.

The dorsal and dorso-lateral lips in the Frog grow back over
the yolk, and an endodermal yolk plug eventually fills the opening
of the blastopore and projects out through it for a time.

ABC

Fig, 29. A, Amphibian; B, Reptile; C, Bird or Mammal.
Small dots, protogenetic tissue ; large dots, yolk ; lines, deuterogenetic tissue.

In the Reptile, the dorsal lip, and the dorso-lateral lips are
growing back over the morphologically coalesced ventro-lateral
and ventral lips of the blastopore — not over yolk or endoderm at
all — that is to say, not over protogenetic tissue but over deutero-
genetic tissue.

There is no blastopore in the Reptile, and therefore there can
be no yolk plug.

The canal formed in this way is neurenteric canal as Kupffer
correctly named it. It has also been termed mesodermal sac by
Hertwig. It is wholly deuterogenetic.

The accompanying diagram, Fig. 29, should make this clear
as seen in surface view. The small dots, large dots and shaded
lines represent the protogenetic tissues, yolk and deuterogenetic
tissues respectively.

Theories with respect to Amphioxm 51

Growth in length occurs by the activity of the deuterogenetic
mass, which as in the Anamnia is most active dorsally, giving rise
to the bulky dorsal region of the body.

The Lower Chordates.

Among the lower chordates the question of growth in length
has been considered with respect to Amphioxus by a number
of investigators. Hatschek originally described growth in length
as being due chiefly to the activity of two large cells which he
termed polar cells, and from which he supposed all the hinder
mesodermal segments were formed. Others, Wilson, Morgan, etc.,
have shown that these polar cells were mythical and they described
growth in length as due to a general mass of proliferating tissue
as in the higher chordates.

Lwoff, Cerfontaine and others have also worked at the subject.

Hatschek and Cerfontaine have maintained that growth in
length is due to concrescence of the dorsal lips of the blastopore
in precisely the way that we have shown to be quite untenable
for the Teleostean and Elasmobranch fishes.

Lwoff described the dorsal wall of the hinder part of the embryo
as forming from the dorsal lip by a process of inrolling of the
ectoderm, this giving rise to what he called an ecto-blastogenetic
plate.

MacBride has more recently gone into the matter and has
come to the conclusion that there is no such thing as a concrescence
of the dorsal lips of the blastopore, but that increase in length
is due to a growing point exactly comparable to that which I
have already described as a deuterogenetic centre, due to the
general activity of the lips of the blastopore. MacBride however
considers that the closing of the blastopore is very largely due
to the growth upwards of the ventral lip of the blastopore; a
condition which does not fall very well into line with the method
by which the Frog's blastopore closes. Here, as in the anamniate
craniates, the more active part of the blastopore rim is the dorsal.

This fact according to MacBride becomes evident very early,
even while gastrulation is in progress (Fig. 8).

In the same way as in the higher chordates the ventral part

4—2

52 Growth in length of the Vertebrate Embryo

of the deuterogenetic centre dies out after a time, and the dorsal
and dorso-lateral part continues to grow, giving rise to the tail.

In the Ascidians the conditions are much the same, as shown
by Conklin in the development of Cynthia. Here the activity of
the growth of the ventral lip is perhaps more marked than in
Amphioxus.

In the Hemichordata the question has not been considered
very thoroughly.

It is well known that no part of their body can be compared
with the tail of the higher chordates.

If we are to compare their condition with that of the higher
chordates, we must suppose either that

(i) The deuterogenetic centre dies out wholly almost as soon
as formed, and that all growth is interstitial and protogenetic ; or,

(ii) The deuterogenetic centre continues active all round the
blastopore, which however closes, with the result that the anus is
at the extreme end of the body, and that there is no tail.

The egg of the Balanoglossidae forms a hollow blastula, which
invaginates as in Amphioxus and forms a gastrula. The gastrula
elongates. The blastopore closes but the anus forms at the spot
where the blastopore has closed. An anterior unpaired pouch
and two pairs of lateral pouches are given off from the gut which
form mesodermic sacs, the five cavities which are known as the
proboscis or head cavity, the collar cavities and the trunk cavities.
There is no segmentation of the trunk cavities as in Amphioxus
although the trunk grows out to a far greater extent.

This latter is an important point for it is not unreasonable to
suppose that segmentation may be due to growth being from a
definite terminal growing point.

If growth in length is produced like the tail of a vertebrate
is produced, or like a stem of a tree is produced, it is clear that
although median organs may be indefinitely prolonged by this
means, and yet retain their original characters, lateral organs
cannot be indefinitely prolonged and at the same time continue
to act as lateral organs.

A notochord, or a spinal cord, can be produced indefinitely,
and these organs do not show segmentation.

But a lateral blood-vessel, or a lateral nerve, or a branch of

Dolichoglossus 53

a tree or a leaf could not be produced in the same way and yet
retain their original character. They can however be repeated-
and indeed we find in organisms in which growth in length occurs
in this way that lateral organs are repeated, and not drawn out
-and this fact may be at the bottom of metameric segmentation.
If so it follows that metameric segmentation may have arisen
independently in many different phyla and that it is a result of
growth in length from a definite growing point ; and as the growing
points may have had different origins, there need be no homology
between the metamerism of one phylum, for instance the Annelida,
with that of another phylum, for instance the Vertebrata.

Fig. 30. Dolichoglossus serpentinus.
df dorsal furrow of proboscis ; br gill clefts ; g gonads ; v bright vermilion spots.

As regards Balanoglossus it seems probable that there is no
deuterogenesis, or if it occurs it soon dies out altogether. The
great increase in length of the trunk is due to general interstitial
growth. This seems probable from the fact that the bright spots
which occur on the trunk of Dolichoglossus become gradually
more and more separated from one another as the organism
elongates, Fig. 30.

The bright vermilion spots are the remnants of pigment formed
in the gonads, and probably of the nature of excretory products.
The gonads, however, are closely packed against one another,
whereas the spots towards the end of the trunk are perhaps an

54 Growth in length of the Vertebrate Embryo

inch or more apart. This seems to prove that their separation
and hence the growth in length is due to a general interstitial
growth and not to a terminal area of cell proliferation.

If deuterogenesis occurs it occurs all round the anus (or
blastopore) and there is no tail. The absence of tail is a very
important fact, for it is in the Hemichordata alone among
chordates that there is no tail.

This is so important that it would justify the classification of
the Chordata into Chordata caudata, and Chordata ecaudata.

Conclusion.

To sum up we may say, both from anatomical observation
and from experiment, that growth in length of the embryo in all
anamniate chordates must be considered to be due to the origin
of a special area of cell production round the lips of the blastopore,
which converts the spherical form of the gastrula into the cylin-
drical form of the later embryo. Since this area of necessity comes
into being only after the gastrula is formed, we may recognise
two regions in the later embryo. One of these regions is the
direct result of the segmentation of the ovum culminating in the
gastrula, and having the general character of a radially symmetrical
form, and this on the whole is to be identified with the coelenterate
phase of evolution. The region of the body so arising has been
named the protogenetic region. The other region is that of later
origin produced by the proliferation of the lips of the gastrula
mouth. This has been called the deuterogenetic region. The
part formed from the protogenetic region includes the fore-
brain, probably also the mid-brain, the mouth, and possibly the
hind-brain as far as the origins of the 5th and 8th nerves, the
branchial region and heart and probably much of the gut. The
part formed from the deuterogenetic region comprises the re-
mainder of the hind- brain and spinal cord and tail, the whole of
the metamerically segmented mesoderm, and in the craniates the
renal organs. As regards the reproductive organs there is much
evidence to show that in the craniate chordates the actual germ cells
are, as one would expect, protogenetic in origin, but that they
migrate during development into the deuterogenetic region and
here undergo their maturation, and eventually find their way to

Origin of tailed Chordate Stage

55

the exterior by means of the deuterogenetic channels of the
coelom or renal apparatus.

The same relations between the two regions probably hold
good for the amniotes, though in them experimental evidence is
obtained less easily.

Now it is conceivable that having attained the stage of a
coelenterate the gastrula may take very different courses of
development. For instance, the whole deuterogenetic centre
might continue active, and thus produce a long cylindrical addition
to the spherical gastrula, Fig. 31 A. Or it might cease almost

Fig. 31. Diagrams of sagittal sections of stages in the development of a craniate

chordate, showing the growth centres.

A, Gastraea stage; B, Balanoglossus stage; G, tailed chordate stage.

as soon as formed, in which case there would be no great alteration
from the radially symmetrical form, Fig. 31 B. The organism
might retain more or less perfectly its radially symmetrical form.
This would be due to the dying out altogether and at once of the
deuterogenetic centre of activity. Or a portion of the deutero-
genetic region might remain for a longer time active while the
rest dies out earlier, Fig. 31 C. The portion that remains active
might be that forming one border of the blastopore, while all
round the rest of the blastopore the activity dies out. The
latter part therefore would be left stationary while the active
part grew back.

It may be suggested that deuterogenesis in the embryo
is a geometrical consequence following gastrulation (which may

56 Growth in length of the Vertebrate Embryo

perhaps also be regarded as a necessary culmination of proto-
genesis). If so these consequences of embryonic development
must have had their effect upon evolution. For instance, one
of the first effects of deuterogenesis must be in many cases a
tendency to close the blastopore.

Now clearly in evolution to close the only entrance to the
digestive tract would be fatal. Therefore there must always have
been and must always be the counteracting circumstances of the
'' necessity to live'1 opposing the geometrical tendency, just as
there always are and always have been the counteracting
influences of the force of gravity which would bring us to the
earth, and the necessities of living that keep us erect. The
continual warring of these tendencies at each repetition of the
life cycle, the one to close, the other to keep open the gastrula
mouth, may well have given the impetus in evolution which has
led to the very varying fates of the blastopore in the different
groups of animals, and the consequent relation of main axis to
the plane of the blastopore.

The two phenomena, the difference in the fate of the blasto-
pore, and the difference in the fate of the deuterogenetic centre,
are of fundamental importance in our attempt to understand the
relationship of the chief phyla of the animal kingdom. Let us
take a very brief survey of what we know of the actual fate of
the blastopore in individual species. In Coelenterates the fate is
simple. The blastopore remains open, acting as chief inhalent
and exhalent aperture. The deuterogenetic centre, if there ever
is one, dies out wholly and at once. In Echinodermata the blasto-
pore may be said to remain, but as anus only. A new opening
is formed, which is the mouth. The deuterogenetic centre probably
dies out wholly and at once. In the Chordata the blastopore re-
mains in part as anus or else the anus opens where the blastopore
has closed. We may regard this as a re-opening of a part — the
ventral part — of the blastopore. To be accurate it is the more
ventral part of the blastopore which is anus, the dorsal part, after
becoming enclosed by the neural folds forms the neurenteric canal,
which ultimately closes. The mouth is never formed from any
part of the blastopore; it is always a new opening. The whole
deuterogenetic area continues active for a short time, then the

Origin of Mouth and Anus 57

dorsal portion alone continues, with the result that a tail is formed.
This I have endeavoured to show is the only conclusion to which
we can come as the result of experimental and anatomical
observation in the chordate embryos. Therefore the relation of
the main axis of the Vertebrata is, as I have indicated, at right
angles to the plane of the blastopore in Fig. 31 C.

This conclusion is diametrically opposed to the conclusion of
His, Hubrecht, Dohrn, Semper, Sedgwick and others, who derive
the dorsal surface of the chordate from the coalescence of an
elongated blastopore surface.

So even MacBride, who after fully accepting as true the views
of growth in length as I have indicated them, comes to the con-
clusion, that "the mouth and anus in all animals in which two
such openings are found, owe their origin to the division of a long
slit-like coelenterate mouth," and suggests that the seam which
originally connected mouth and anus was situated " on the ventral
and not on the dorsal surface."

This supposition has not a trace of embryological evidence in
its favour.

Goodale's experiments already alluded to show that there is
no suggestion of concrescence either along the ventral or dorsal
border connecting mouth and anus.

MacBride suggests that the growth upwards of the ventral lip
of the blastopore in Amphioxus is reminiscent of this concrescence.
It is true that a similar and more obvious coalescence of the
ventral part of the blastopore lip occurs in the Amphibia, but the
most ventral part either remains as anus or re-opens as anus —
never is there any sign of connection with the mouth. We might
legitimately regard the Amphibian condition as representing an
ancestral coelenterate or actinian phase in which the neurenteric
end or opening represents the actinian inhalent channel; whilst
the anal end or opening represents the exhalent channel.

One may argue thus. When the chordate central nervous
system began to become bulky and in forming a tube involved
the inhalent channel (as neurenteric channel), then one or more
of many other openings which in no way owed their origin to the
blastopore became an inhalent aperture or mouth. There is no
difficulty in assuming the adoption of a new opening for the mouth

58

Growth in length of the Vertebrate Embryo

either in the Chordate phylum or in the Echinoderm phylum.
In both these phyla openings from the gut or its pouches are
common — as they are in the Coelenterate. In the coelenterate we
find numerous apertures in the tentacles and in the body wall, as
for instance the cinclides of Anemones. In Chordates apertures of
this nature are numerous, e.g. the gill clefts of any class ; or the
intestinal pores of certain Hemichordata.

In the Echinoderm also the mouth is the new opening. The
hydropore is also a new opening into an archenteric diverticulum.
It may be objected that gill clefts are essentially exhalent and not
inhalent apertures. No one, however, would deny the homology

II

Fig. 32. Diagram to show the closure of the blastopore, and to show the part
that remains open as anus, an. I, Urodela; II, Anura.

of the spiracles of the Dog-fish with the branchial clefts. It is
almost certain that they must at one time have been exhalent
apertures. Now they are inhalent. There is no reason why the
present mouth should not have arisen from either an inhalent or
exhalent pore, provided it occurred in a suitable place.

In both the Chordata and Echinodermata we find strong
embryological evidence for the assumption that the blastopore
became anus only and that the present mouth is a new pore. In
every case described the anus is always either derived directly
from the blastopore or else it opens at the spot where the blastopore
has closed. Thus the Chordata and Echinodermata agree in their
main axial relation to the plane of the blastopore, for, in each case
it is at right angles to this plane. They also agree in various other

Fate of Blastopore in Annelida 59

respects. But the two phyla differ in respect to deuterogenesis.
In the Echinoderms it died out — such elongation as there is is
probably interstitial, as possibly in Balanoglossus. In the caudate
Chordates it continued active dorsally, and as already described
produced the tail, see Fig. 31.

Now it has probably been very different with regard to the
other phyla. Unfortunately at present there is no experimental
evidence to help us, but a good deal of accurate anatomical
evidence is available.

What is the fate of the blastopore in Annelida ? The evidence
suggests that there is an elongation of the original gastrula mouth
in the plane of the blastopore, which elongated mouth becomes
reduced by a coalescence of the middle lips into two terminal
openings (as in the Urodele Amphibians) forming mouth and
anus respectively.

And further it would appear that the deuterogenetic centre
remains active between the two openings, but more particularly
near the anal opening producing growth in length. The direction
of which elongation is therefore at right angles to the direction of
the Chordate elongation.

Thus the fate of the blastopore in Hydroides uncinatus (Eupo-
matus) is described thus by Shearer: "The blastopore at first
lies exactly in the middle of the ventral plate... when fully formed
it is an elongated slit, somewhat enlarged at its anterior end.
This end never completely closes, but... becomes the mouth.
The posterior portion closes completely, the anus breaking through
almost immediately where the last portion of this part of the
blastopore disappears." So also in Polygordius. Conn says of
serpula : " the blastopore becomes an elongated slit, the lips of
which meet in the middle and close, forming the rudiment of the
future gut. For a short time the digestive tract remains attached
to the ectoderm throughout the length of the blastopore, but
after a little it only retains this connection at either end. The
digestive tract becomes hollow and acquires two openings to the
exterior at the two points of its previous connection with the
ectoderm."

In Amphitrite ornata according to Mead the blastopore closes
by concrescence from behind forwards but the extreme anterior

60 Growth in length of the Vertebrate Embryo

end remains open as the mouth, and the anus breaks through at
the posterior end of the seam.

Balfour says that Stossich and also Willemoes-Suhm state that
in one species of Serpula the part of the blastopore remaining
open is the anus.

In Oligochaeta the blastopore becomes the mouth.

This can all be easily rendered intelligible by assuming that
in evolution the blastopore elongated and then by closing centrally
gave rise to the mouth and anus.

Growth in length in this case has been due to the persistence
of part of the blastopore rim as the deuterogenetic centre. But
which part? Clearly not the mid-dorsal as in Chordates. It is
probably caused by the retention of a part of the coalesced lip.
This is the part which has been sometimes named the teloblast.
It gives rise to the trunk and the so-called coelomesoblast ; and as
it is situated between the mouth and anus the effect of its activity
is of course to separate these two openings more and more.

In Hydroides uncinatus (Eupomatus), Fig. 33, the cells which
give rise to the coelomesoblast from which the greater part of
the trunk mesoblast is derived have their origin in the lips of the
blastopore, as Shearer has shown, or as Treadwell has shown in
Podarke, near to the future posterior end.

In many Annelids (and Molluscs) there are other mesodermal
cells which are not derived from these two blastopore lip cells,
but either from the ectoderm or other part of the endoderm.
These form the so-called larval mesoblast or ectomesoblast.

If we use the same terminology as that which we have used
with regard to the Chordates, then clearlv the ectomesoblast is

V

protogenetic mesoderm, and the coelomesoblast in these forms is
deuterogenetic. The growth in length is due to deuterogenesis but
it is due to the persistence of a different part of the deuterogenetic
ring, and the result is a lengthening in the plane of the blastopore
in Annelids, and not at right angles to it as in Chordates.

Moreover in Annelids both mouth and anus have had their
origin in the gastrula coelenterate mouth, whereas in Chordates
only the anus has been derived therefrom, Fig. 31.

The group most closely allied to the Annelida is undoubtedly
the Arthropoda. In these there is plenty of indirect evidence

Conditions in the Arthropoda

61

that the plane of the gastrula mouth is the plane of the future
ventral surface. For instance in the Copepod Cetochilus the
blastopore is a longitudinal slit on the future ventral surface and
it closes from before backwards, the part which closes last being

Fig. 33. A, B, Podarke, after Treadwell; C, D, Hydroides

uncinatus, after Shearer.

Semi-diagrammatic views to show the derivation of mouth and anus openings,
an anus; mes ectomesoblast ; ewendoderm; M e coelomesoblast ; storastomo-
daeum; m mouth; hk head kidney; st stomach; av anal vesicle.

near the site of the future proctodaeal opening. In Limulus
and the scorpion or in insects such as Doryphora there is a median
ventral furrow which represents the blastopore; and above all
there is the evidence of the intermediate form Peripatus, where
the blastopore actually elongates into a long narrow slit which

62 Growth in length of the Vertebrate Embryo

closes in the middle region but the ends remain permanently open
as mouth and anus.

So in the Mollusca in some, as for instance, Ostrea, the blasto-
pore remains open as the mouth, in others, Pisidium, it remains
as the anus, in others, Teredo, it closes altogether. Such incon-
sistencies taken in conjunction with what has just been said in
regard to the Annelida and Arthropoda may well be interpreted,
as Balfour suggested long ago, as indicating that in the Molluscs
also the gastrula mouth has given rise in the course of evolution
to both mouth and anus.

In the Mollusca the deuterogenetic activity would seem to
occupy a position corresponding to that of the Annelid and
Arthropodan group, and we can recognise protogenetic and
deuterogenetic mesoderm produced in a very similar manner, but
the deuterogenetic activity dies out much earlier and growth in
length is limited.

The conditions of development are thus totally different to
those of the Chordata and Echinodermata in which the gastrula
mouth gave rise to anus only.

In the Nemertean phylum the gastrula mouth would appear
to have become mouth only, and the anus is a new formation.

It may be asked what about metameric segmentation? Is
the metamerism of the Annelid in no way homologous to that of
theChordate?

To this I would reply, no more so than the metameric segmenta-
tion of the Annelid is homologous to that of the fir tree.

The similarity is one of analogy. The deuterogenetic regions
of the Annelid and of the Chordate show metameric segmentation
because in each case growth in length is due to the same physio-
logical cause, which is quite analogous to that which produces
growth of the stem of a fir tree. Axial organs can be indefinitely
produced in length without interfering necessarily with their
utility. Thus the intestine of an earth-worm, the notochord or
spinal cord of a vertebrate, or the pith of a fir tree, can be
indefinitely extended and yet retain their original functions.

But we cannot imagine lateral organs like lateral blood vessels,
lateral nerves, lateral muscles, limbs, leaves on a stem, being

Conclusion 63

expanded indefinitely without altogether losing their utility, so
such organs must be repeated and not expanded when the great
growth in length occurs. Hence it is not necessary to regard
metamerism as in itself indicating homology.

The object of this short course of lectures has been to show
what the real relation of the main axis of the body of a Chordate

V'

is to the plane of the gastrula or coelenterate mouth, and to
indicate that it is probably quite different to that of the Annelid,
Arthropodan and Molluscan groups. Such a conclusion if correct
is quite destructive to the hypothesis of His, Semper, Dohrn,
Patten, Gaskell and others who would derive chordates from
various highly organised invertebrate groups. It agrees with
the views of Sedgwick, van Beneden and Hubrecht in seeking the
origin of chordates as far back as the Coelenterata or their
immediate descendants, but is equally destructive of the view that
the main axis of the chordate body is parallel with the plane of
the coelenterate mouth — on the contrary it shows it to be at
right angles thereto.

THE GEOMETRICAL RELATION OF THE
NUCLEI IN AN INVAGINATING GASTRULA
(e.g. AMPHIOXUS) CONSIDERED IN CON-
NECTION WITH CELL RHYTHM, AND
DRIESCH'S CONCEPTION OF ENTELECHY

A. V. E.

PART I

'No kind of causality based upon the constellations of single
physical and chemical acts can account for organic individual
development.... Life, at least morphogenesis, is not a specialised
arrangement of inorganic events ; biology, therefore, is not applied
physics and chemistry : life is something apart, and biology is an
independent science."

:The Science and Philosophy of the Organism'1 by Hans
Driesch, being the Gifford Lectures for 1907, p. 142.

Although one may subscribe to the opinion expressed in the
above quotation from Driesch, nevertheless on reading the whole
book it strikes one that Driesch's "Entelechy" is something much
more than the minimum contained in that statement, and is
indeed a somewhat mystical conception.

To fill in the sentences omitted in the above quotation of
Driesch, the omission of which was indicated by the asterisks:
'This development is not to be explained by any hypothesis
about configuration of physical and chemical agents. Therefore
there must be something else which is to be regarded as the
sufficient reason of individual form-production. We now have
got the answer to our question, what our constant E consists in.
It is not the resulting action of a constellation. It is not only
a short expression for a more complicated state of affairs, it
expresses a true element of nature"

Then again later "our vitalistic or autonomous factor E con-
cerned in morphogenesis" is named "Entelechy," p. 144. But
Driesch's Entelechy does seem to me to be a more complex con-
ception than that of a " true element of nature" or a form of energy
in any way comparable to the single physical forces. It is a

5—2

68 Geometrical relation of Nuclei

coordinating and regulating influence with very diverse and
mysterious ways and means of application, not unlike the re-
capitulatory ;' constraint" which twenty years ago was held to
cause organisms to develop along a certain ancestral course, and
is amenable with difficulty, if at all, to exact computation.

While welcoming the conception because it involves the
postulation of some form of energy not present in inanimate
bodies, one may ask whether for the present it would not be more
profitable to go no further in the search for the unknown vitalistic
factor than the recognition and investigation of a special form of
energy not so wholly different from certain known forms of energy.

Such a form of energy might be conceived of as acting from a
centre like gravitation, statical electricity or magnetism ; as causing
movements of attraction or repulsion according to the conditions
under which it is acting; as being a constant attribute of living
matter and exerting no influence on non-living material; and as
possessing the peculiarity of automatic and rhythmic alternation
between unipolar and bipolar states.

May not Driesch's "Entelechy511 prove to be the more complex
action of some such simpler form of energy in combination with
other forces ?

May not Entelechy bear to this simpler form of energy some
such relation as that, for example, which the gyroscope bears to
gravity, and the real vital factor be something rather less complex
and also easier to investigate by exact methods of mensuration
and computation?

In 1894 Roux made the disco verv that the blastomeres of cells

•/

of Rana fusca towards the end of segmentation, if isolated and
floated in a suitable medium, will show distinctly attractive pro-
perties inter se, so that the sides of each cell will become drawn
out towards a closely neighbouring cell, and such cells, eventually
moving towards each other and coming actually into contact
become pressed and flattened up against one another. This
phenomenon Roux termed cytotropism, which he has since
changed to cytotaxis, a term also adopted more recently by
Przibram.

If this alleged attraction is a reality — and if it is a universal
law, then the discovery of this fact must be regarded as one of

Cytotaxis 69

great importance and of far-reaching consequence. Indeed this
property has already been recognised and called upon as an
explanation of certain developmental phenomena as for instance
by Hjalmar Theel who observed that apparent attractions and
repulsions occur among the blastomeres of the segmenting egg,
Echinocyamus pusillus.

Zur Strassen especially, has endeavoured to apply the principle
of chemotaxis and cytotaxis to the formation of simple epithelia
in general and the blastula epithelium of the developing egg of
Ascaris in particular.

He suggests that a simple, single-layered epithelium of cubical
or columnar cells, supported by a sub-lying tissue or membrane,
may be the result of a chemotactic attraction between the base-
ment membrane and the several cells producing a result such as
would be produced by placing a pile of shot on a table and then
shaking until the shot all settled down into a uniform unilaminar
layer.

For the production of more complicated cases he supposes a
much more complicated series of attraction zones, so that one
part of a cell attracts its neighbour more strongly than the next
part — the parts beyond again still more strongly and so on, and
thus he accounts for the curves seen in various "free" epithelia.

When some years ago I originally came to believe in the
phenomenon now known as cytotropism or cytotaxis (though
perhaps cytokinesis would be a still more appropriate term),
I arrived at it as the result of experiments made with a view to
determine the cause of invagination such as occurs during the
process of gastrulation in Amphioxus.

The mechanics of the invagination process as seen in Amphioxus
during gastrulation have been often discussed, e.g. Hatschek,
Goette, zur Strassen, Rhumbler, etc.

Rhumbler in his recent analysis of the factors involved in the
process of gastrulation by invagination distinguishes between the
factors which increase invagination when once initiated and those
which are the actual cause of it. In his model made of flexible
steel rings he perceived that there would have to be an actual
force which acts as for instance one's finger acts in producing an
invagination in an indiarubber ball, either a pushing from without

70 Geometrical relation of Nuclei

or a pulling from within which initiates the invagination. Once
started, then lateral pressure caused by the multiplication of
cell units will increase it; but without forces other than those
which he supposes to exist, lateral pressure cannot initiate
invagination.

It has been suggested that invagination of a gastrula such as
that of Amphioxus may be brought about by the absorption of
the fluid within the blastocoel, causing the collapse of one side.
Hatschek suggested this as a possible cause in Amphioxus ; but
apart from the unlikelihood of the thicker side being the invagi-
nated side as is actually the case in Amphioxus, if this were the
cause, it has been completely put out of question by the fact
noted by Morgan and Hazen that invagination will occur when
a small aperture exists in the blastula wall.

Rhumbler like Goette points out that the shape assumed by
cells forming the wall of a blastula is conical, with the smaller end
pointing inwards, but that the cells of the invaginated side of a
gastrula are conical with the smaller ends pointing towards the
original outside, and Rhumbler suggests that a change in shape of
the cell has been the cause of the invagination. Surely this is con-
fusion of effect and cause. The living cell in an undifferentiated
state is an almost perfectly fluid body and the rectangular or
hexagonal shape assumed in early undifferentiated tissues is much
more easily to be explained as the result of environmental pressure
than as being due to its own intrinsic properties. And does not
a cell isolated from blastula or gastrula tend to assume a spherical
form?

He argues, however, that the change in shape is due to the
absorption of fluid by the inner portions of the cells causing
thereby a bulging of their inner surfaces. This, occurring in many
or most of the cells of one part, would, he says, produce the change
in shape, and together with the lateral pressure due to multiplica-
tion of the ectoderm units would bring about invagination.

Przibram commenting on this and other suggestions offered
concludes " Blastula tion and gastrulation depend on chemotactic
effects started by processes of assimilation, which not only cause
passive mechanical displacements, but also active migrations of
cells."

Experiments with Rubber Balls 71

But beside the objection raised above that the conical shape of
the cells would seem to be more the effect than the cause, and the
subsequent change of shape during invagination likewise an effect
rather than a cause, it is very doubtful whether, if the changes in
assimilation and surface tensions suggested do occur, they could
produce a swelling of the inner part of the cell and corresponding
reduction of the outer, because we know that the cell when isolated
tends to assume a spherical form ; that is to say cells do not retain
a conical shape after the mutual pressure has been removed, thus
showing that they have no intrinsic tendency to be conical (v.
Herbst's figures), and so can hardly be so prone to become conical
in the obverse direction as to be able to exert the pressure necessary
to convert a blastula into a gastrula.

In my original experiments on the process of gastrulation in
Amphioxus, I experimented with indiarubber balls, which were
supposed to represent isolated cells, and which like cells, when
pressed together become flattened. In most respects an india-
rubber ball behaves as an isolated blastomere does when at rest.
I tried by various means representing increase in number illus-
trating, that is to say, more rapid growth at various places, or
by variation in the size or shape of ball, to reproduce gastrulation
such as occurs in a typical way in Amphioxus, but by no means
could I do this until I used indiarubber cord applied so as to
represent a mutual attraction between cell and cell.

By means of a circle of indiarubber balls strung together by
indiarubber cord I could reproduce exactly the process of con-
version of a blastula into a gastrula, or to be more accurate, the
conversion of a circle into a double crescent, one part of the circle
becoming " invaginated " into the other.

The necessary conditions are: (i) that the balls shall be
sufficiently numerous to cause tension in the elastic cord — the
balls themselves then becoming flattened against each other, the
free inner and outer surfaces remaining convex.

(ii) The cord must pass nearer to the outward surfaces of
a certain number of the balls along one part of the circle than
elsewhere. That is to say, if the cord pass through the centre
of the balls for two-thirds of the circle, and pass, say half way
between the centre and the outer surfaces of the balls forming

72

Geometrical relation of Nuclei

the remaining third, then that remaining third will be invaginated,
if the third and last condition is fulfilled, namely,

(iii) That the whole circle is sufficiently large; because the
invagination process can be inhibited if the circle is so small that
the pressure between the neighbouring balls at the point of contact

Fig. 34. Diagram representing a circle of spheres in which the relative position of
the centres of attractions (b) to the centres of the mass (g) necessary to cause
invagination of one part of the circle within the other part are shown.

is greater than the attractive force between the attractive centres
of ball and ball, (which ex hypotliesi are eccentric to the whole
mass of the ball over a certain small area of the sphere or arc of
the circle,) in this case no invagination will take place. But
increase the number of balls and then the invagination may occur.
The accompanying diagrams show how this must be so.

Forces exerted between Cells

73

Fig. 34 represents a series of spheres arranged in a circle. Each
sphere has its actual centre indicated by the black spots g, </....
And each sphere is supposed to be made partly at any rate of living
material and to have a centre from which it exerts an attractive
force upon its neighbours. This centre, b, b... is shown not to
correspond with the actual centre, but to lie near the internal pole
of the spheres forming the upper part of the diagram, and nearer
to the outer margin in the spheres forming the lower part of the
diagram.

Under these circumstances the form taken by the ring of
spheres would not be a circle, but one side — the lower side would
be invaginated within the upper — provided that the attractive
force between cell and cell (contiguous cells) is strong enough.

Fig. 35. Parallelogram of forces concerned in the invagination of
a circle of spheres such as shown in Fig. 34.

Fig. 35 illustrates in detail the action of the forces exerted
between cell and cell of the lower part of the diagram.

Instead of taking the whole lower part into consideration,
I have taken only three spheres, in the middle one (Y) of which
the centre of attraction (A) is shown to be near to the outer
margin of the circle while in the two side ones (X and Z) it is co-
incident with the centre of the spheres at B and C.

The parallelogram ACDB will be the parallelogram of forces
exerted upon A by the attractive powers of the spheres X and Y,
and Z and Y. The lines AB, AC, may be taken to represent the
direction of these forces and the line AD the resultant.

If the tissues are perfectly compressible the sphere Y must
move in the direction of D.

74 Geometrical relation of Nuclei

But if the tissues are only imperfectly compressible there will
then be a counteracting force exerted on Y by its neighbours at
the points of contact, tending to prevent the movement of Y
towards D. The strength of this counteracting force will depend
inter alia on the position of these points of contact. The nearer
they are to the point D, the greater will be their influence, the
further removed from D, the less will be their influence and the
less will be the tendency to prevent the movement of Y towards
B. Now clearly with a sphere (or ring of balls as in the diagram)
any enlargement of the sphere (or ring) will move the points of
contact further from D and diminish the counteracting force due
to physical pressure and will therefore favour the movement of
Y towards D. Thus the invagination of the lower pole of the
diagram depends upon (1) the presence of an attractive force acting
between sphere and sphere, (2) the difference, relative to the centre
of the mass, of the position of the centre from which the force acts
in the upper and lower parts of the ring, (3) the relative strengths
of the supposed inter-spheric attractions and the counteracting
physical pressure at the points of contact of sphere and
sphere.

Let us see to what extent it is possible to find these conditions
in the developing blastula and gastrula in a typical case of invagina-
tion such as we get during the gastrulatiori of Amphioxus. In the
first place we must regard the phenomenon of intercellular attrac-
tion as an established fact, and believe that an attraction exists
between cell and cell as an attribute of a living cell, as much as
gravity is an attribute of all matter.

Like gravity we must consider that it can be described as
acting from a centre. Just as there is a centre of gravity for
every mass, and just as the centre of gravity may be not the
actual centre of the mass, so there must always be an attraction
centre in a cell, and this attraction centre need not correspond to
the actual centre of the cell mass, nor to its centre of gravity.
Thus in a bird's ovum for instance, which we may consider to be
a sphere, the centre of gravity is certainly below the actual centre
of the sphere (i.e., nearer the "vegetative" pole) while its attraction
centre will probably be very much above the actual centre of the
sphere.

Nucleus as centre of attractive Force 75

Given these conditions, then the mechanics of the rubber ball
model may be reproduced by the living blastula and gastrula.

The next question to be considered is, can we locate the
attraction centre?

There can be little doubt that if it exists, it is coincident with
either the nucleus (Hertwig's law 'The nucleus tends to take up
a position in the centre of its sphere of influence, i.e., of the
protoplasmic mass in which it lies," Przibram), or more probably
still, the centrosome of the resting cell. The centrosome is, how-
ever, not always easy to see and at any rate it is not often drawn
in figures of sections of developing embryos, but the nucleus is
usually given, and as the nucleus is never far removed from the
centrosome we may for the purpose of the argument take the
centre of the nucleus as indicating approximately or perhaps
actually the position of the centre of the attractive force. When
I first endeavoured to work this out, the only figures available
were those of Kowalewski and Hatschek, and those of Hatschek
were regarded as authoritative. Now unfortunately for my
hypothesis the position shown by Hatschek for the nuclei did
not support in any way the contention that the supposed attraction
centre of the invaginating cells was more towards the outer surface
than the inner surface as was necessary to fulfil the required con-
ditions. On the contrary just the opposite was shown, e.g., v.
Hatschek, Fig. 21, though not in all cases, e.g., Fig. 20, 2 b.

But although Hatschek's figures are unsurpassed for clearness
and general accuracy, yet as regards the position of nuclei, and of
cell walls and in some other respects, they must be regarded as
diagrammatic, as recent work on Amphioxus has shown.

During the last fifteen years sections of the gastrulating Amphi-
oxus have been drawn by several authors which are probably more
accurate in this respect, because they agree with one another,
though disagreeing with Hatschek. They all agree in showing
the nuclei of the cells about to invaginate and the invaginating
cells as lying close up to the outer surface of the cells, and what
is more to the point in most cases distinctly more eccentric than
the nuclei of the cells which do not participate in the actual
invagination. I allude to the drawings of Sobotta, Samassa,
Morgan and Hazen, Wilson, MacBride, Cerfontaine and Legros.

76

Geometrical relation of Nuclei

I reproduce here figures from Sobotta, Samassa and Cerfontaine
which illustrate my point as well as possible (v. Fig. 36).

In Samassa' s figure A, for instance, the difference of the
position of the nuclei as regards its eccentricity is represented
by the figures 1 : 1«73 for the fourteen uppermost cells compared
with 1 : 2- 40 for the twelve lower larger cells, the figure 1 repre-
senting the average distances of the centre of the nucleus from
the outer margin of the cell, the figures 1*73 and 240 the average

Fig. 36. Drawings of sections of the gastrulating blastula of Amphioxus, showing
how the nuclei of the cells occupy the position within their cells required by
the hypothesis explained in the text. A after Samassa, B after Sobotta,
C after Cerfontaine.

distance of the centre of the nucleus from the inner margin of the
uppermost 14 and lowermost 12 respectively.

It is a very remarkable fact that in all these figures the nuclei
of the invaginating cells occupy the position required by the hypo-
thesis. Moreover a similar state occurs in many other gastrulating
blastula, as, for instance, may be seen in the figures of Synapta by
Selenka, Balanoglossus by Bateson, etc.

The condition that determines the position of the attraction
centre is probably the degree of concentration of the protoplasm.

Position of " centre" of Attraction

77

If there is a form of energy which is an attribute of living cells
and which under certain conditions is manifest as an attractive
force acting from a centre, the "centre" must be the centre of the
living substance, and if the protoplasm is vacuolated through part
of its mass by dead material, for instance, food yolk, then the
centre will be driven away from the yolk mass just as the centre
of gravity is driven away from the centre of a mass by vacuolation
such as one may get in part of a lump of slag. Further it must

*

be supposed that the divisions of the blastomeres take place in
such a way that the more yolk-laden parts of the cells are directed

Fig. 37. Diagrams illustrating how a mass of inert material within an ovum
causing eccentricity of the centre of attraction may be divided in the vertical
plane without disturbing the original proportion of its distribution between
upper and lower segments of the segmented ovum, but how a horizontal plane
of division upsets this arrangement.

inwards, and that therefore the disproportion between the segments
set up by the first horizontal furrow, that is to say, at the third
generation of blastomeres, is retained till the end of segmentation.

The material which vacuolates the protoplasm and causes the
eccentricity of the centre of attraction in the case of Amphioxus is
no doubt the food yolk, small though the amount of it may be.

It is clear that on the above hypothesis the distribution and
quantity of yolk relative to the upper and lower poles, is an all
important factor in the process of invagination of the blastula, and
anything which would interfere with the normal distribution of yolk
would probably render the invagination imperfect or impossible.

78 Geometrical relation of Nuclei

This is very prettily shown by Wilson's experiments on the
blastomeres of Amphioxus. Wilson found that by shaking apart
the blastomeres formed by the first cleavage furrow, each separated
blastomere developed normally and became a perfect blastula,
gastrula, and metameric embryo, but half the size of a normal one.
So also each of the first four segments became perfect gastrulae
and free swimming embryos. Now in both these cases the plane
of division is vertical and median and divides the egg in such a
way that the relative distribution of the protoplasm and yolk
between upper and lower poles is undisturbed.

But a very different result was obtained by the segments of the
eight cell stage. In this case the eight cells are derived from the
four cells by an equatorial or horizontal division, which entirely
upsets the proportion of yolk to protoplasm. Each of these eight
segments isolated by Wilson continued to divide, but since their
composition was utterly unlike the composition of the whole ovum
of Amphioxus, each became an embryo unlike an Amphioxus
embryo, and although a few became blastulae not one ever became
a gastrula. Wilson says: "None of the J embryos, as I believe,
are capable of full development. I have isolated a considerable
number of the J blastomeres, and of the later embryos of the
various types (i.e., '(a) perfectly flat plates of cells, (6) more or
less curved plates and (c) blastulas one-eighth the normal size,
either closed or with a pore at one side'), and have observed
hundreds of all stages without once obtaining a gastrula."

Morgan a few years later repeated these experiments and he
gives on the whole a similar, but not so dogmatic an account as
Wilson's. Thus he says some of the J- blastomeres fail to gastrulate
while a few of the J blastomeres form blastulae which "partially
gastrulate." "Many of the J blastomeres develop into hollow,
swimming blastulae, which swim around for several hours after
the normal blastula has gastrulated. Some of these blastulae
flatten at one pole as though about to gastrulate, but do not seem
to be able to develop further."

The shaking of the eggs however seems to cause abnormalities
even in the whole ova.

I think we may safely conclude that whereas the rule is that
J and J blastomeres develop as the whole ovum normally does,

Experiments on Amphioxus Eggs 79

the J blastomeres will not invaginate. Again, Wilson did find
some very small gastrulae "even less than J the normal size," but
he says " these gastrulas however did not arise from 8-celled stages
but from 2- and 4-celled, and I can only explain their origin by
supposing either that the J or -| blastomeres underwent preliminary
fission before beginning their progressive development, or that they
were mechanically broken into smaller fragments by the operation
of shaking as often happens in the case of entire undivided ova."

Thus a gastrula J or y1^ normal size may be obtained, but
these are derived from \ half embryos or J half embryos, in all
of which cases the reduction is due to vertical partitions which
do not disturb the proportions of upper and lower poles as a
horizontal partition does. On the hypothesis which I suggest
these observations receive a perfectly comprehensible explanation.

Granted that there is such a form of energy we must inquire
whether the attractive property is constant or whether it is only
temporary or whether it is varying in intensity. For instance, it
might be more intense after the complete formation of each cell
or blastomere, and gradually diminish with the age of the cell,
while passing gradually into the bipolar state to become again
more evident as an attractive force after the recurrence of cell
division.

Anyone who is familiar with the fresh segmenting ova of
mammals will remember how when first found — presumably
while still alive — the segments appear pressed together so that
the sides in contact with other cells are flattened, but that after
a time, when presumably the cells have died, this flattening
against one another disappears, each segment at any rate in the
2- and 4-celled stages becomes spherical instead of oval and
touches its neighbouring segments at points only.

Fig. 38 shows the 2-segment stage of a rabbit. Each segment
is flattened against the other, but this is not due to insufficiency of
space within the zona radiata because the cells do not touch the
zona on all sides — and I do not see how it can be explained by
reference to capillarity or surface tension or cements, because such
causes ought to have the same effect after death as before. It
may be objected that the change from the oval closely pressed
together condition is due to a lower temperature bringing about a

80

Geometrical relation of Nuclei

solidification or change in consistency of the fluids of the cells of
a hot blooded animal. I am not clear how far this is due to loss of
heat or how far it is due to loss of vitality. By use of a hot stage
I have seen the spherical segments regain their oval and mutually
compressed condition. But I have not as yet succeeded in ob-

250p.nt.

335pm.

B

Fig. 38. A, ovum of rabbit 2 segment stage after removal from the Fallopian tube
and examination in aqueous humour (rabbit) at the ordinary room temperature.
Drawn with camera lucida. B, the same under the same conditions 45 minutes
later. The segments are more spherical. C, the same after another 35 minutes
during which the heat of the stage had been raised approximately to blood heat.
The segments had regained their oval condition and more closely pressed
together though internally there was a distinction between an inner granular
and outer clearer part not noticeable in the egg when first seen in the Fallopian
tube. The polar bodies showed "amoeboid movements." D, the same two
hours later during which time the high temperature had been maintained.
The outer clearer ectoplasm has diminished in extent. E, the same three
hours twenty minutes later; the general appearance is now quite normal.
F, the same some twelve hours later during which time the temperature had
sunk to that of the room.

serving the division of a segment so cannot say whether the change
is due to physical condition change, or to a re-attainment of
vitality.

Przibram, when writing of the significance of the processes of
cell division, refers to the analogy between the mitotic and certain

Nature of the attractive force 81

other figures thus. " The mitotic forms have often been compared
(Fol, Ziegler, Gallardo,) to lines of force in an electric or magnetic
field, but this has not furnished any deeper analogy; in contrast
to the electric and magnetic figures with dissimilar poles we are
dealing in the case of mitosis with similar poles; this may be
recognised from the migration of the chromosomes in opposite
directions, and still more clearly from the association of several
poles to form figures which do not in any way disturb the spindle
and aster figures. The crossing of the rays has been urged as an
objection to this comparison."

Przibram clearly favours a solution of the problem along the
line of differential hydrostatic states, the accumulation and with-
drawal of water from various parts of the alveolar protoplasm and
he suggests as "a provisional solution of the difficult problem
of mitotic cell division" the following statement: 'The common
cause of the mitotic phenomena lies in a localised secretion
('condensation') of a more liquid substance, the Enchylemma,
and in the transformation, caused by the redistribution of the
liquid, of a monocentric system of surface tension into a dicentric
system."

Przibram does not allude to the paper by Marcus Hartog in the
Proceedings of the Royal Society, 1905, Vol. LXXVI, where that
author meets at any rate some of the objections alluded to above,
such as the anastomosing and crossing of the rays, which he finds
actually were reproduced in his magnetic model.

Hartog also obtained, though not very perfectly, triaster and
tetraster conditions, although working with unlike poles of the
magnet and a 'third of zero sign — in the models the core of a
magnet without a coil."

Still, these experiments, although they tend to make the
analogy between this force of the dividing cells [termed "karyo-
kinetic force," or ;'mito-kinetism" (Hartog)] closer, leave the
analogy an analogy only, indicating that the effect upon cell
material is in its action from a centre similar to that of a force
like magnetism or electricity on substances variably permeable to
those forces acting in a field from a centre.

Whatever the nature of the force may be, if it act under those
conditions from a centre, analogous effects may be produced quite

A. v. E. 6

82 Geometrical relation of Nuclei

apart from any exact or near similarity existing between the sup-
posed forces.

The explanation of invagination during gastrulation in Amphi-
oxus is put forward as a little further evidence in favour of the
existence of some force which acts like gravitation or magnetism
or statical electricity, and can be described as acting from a centre,
but which is probably of a different nature and with different laws,
and is an attribute of living matter alone.

For it has been made clear, I hope, that Amphioxus in its
process of gastrulation fulfils all the conditions required by the
model of indiarubber balls, provided it can be conceded that
(1) there is an attractive force between cell and cell, "acting from
a centre"; (2) that the nucleus approximately indicates the
position of the centre from which this force may be described as
acting; (3) invagination can only take place after a certain size
relative to the size of the unsegmented ovum, the distribution of
yolk, and degree of force of attraction has been attained by the
blastula.

By reason of the original structure of the ovum the centres of
attraction take up their required position whereby invagination
must ultimately occur. So long as the essential structure is not
interfered with, the ovum may be subdivided without destroying
the power to invaginate.

A gastrula j1^ the normal size may be obtained so long as the
subdivision of the ovum has been always by a vertical plane. But
the gastrula stage is not reached if the division has been made
horizontally.

This fact receives satisfactory explanation on the hypothesis
enunciated above.

The attraction of daughter cells immediately following separa-
tion is of course very marvellous ; because whatever may be the
details and however complicated may be the mechanism of cell
division, yet the process of cell division reduced to its simplest
expression must imply the origin of bipolarity or a bicentred con-
dition involving an antagonism or repulsion between two centres
within a body which at one time had been single centred. That
is to say, the incipient daughter nuclei (and centrosome of course)
while still forming parts of their disappearing parent seem to repel

Protogenesis and Deuterogenesis 83

*

one another, and yet immediately on separation (which separation
is by no means always nor even perhaps ever immediately absolute)
appear to attract one another strongly.

But such a state of things is not unknown in other analogous
phenomena in which attraction and repulsion form the visible
effects of the internal force. In a bar magnet there are strong
force "centres" of the two poles, while midway between the poles
there is but little attractive force. Break the magnet and im-
mediately strong attraction centres occur near the point where
before there was none or but little, even if the broken ends are in
loose contact. I do not mean to compare the supposed form of
energy in any way with magnetism except by analogy; for
magnetism is always bipolar while the form of energy under
consideration appears to be automatically unipolar and bipolar
alternately.

PART II

If the account given here of the gastrulation of Amphioxus is
correct, it will be realised that the formation of the gut cavity and
the definitive ectode'rm and endoderm at any rate of the anterior
part of the future animal is the direct result of and indeed part of
segmentation. The product is on the whole a radially symmetrical
animal. It is what I called in earlier papers on the Kabbit the
result of the primary centre of growth (and later by paraphrase
protogenesis) . It is to be sharply distinguished from growth in
length, which by the origin of a secondary growth- centre system
(deuterogenesis) round the lips of the blastopore, converts the
radially symmetrical organism into a bilaterally symmetrical one
by adding on a metamerically segmented trunk and tail. In a
similar way I attempted to show that the archenteron and
definitive ectoderm and endoderm are formed as a direct result of
protogenesis in the Amphibian, v. " The Growth in Length of the
Frog Embryo."

The whole process of formation of the gastrula with its archen-
teron is regarded as being the simple working of the driving power,
so to speak, of this supposed form of energy guided by the

6—2

84 Geometrical relation of Nuclei

extrinsic factors, in this case chiefly food yolk granules which are
distributed in larger or smaller grains in greater or lesser quantities
within the meshes of the living protoplasm in the different regions
of the egg.

In the paper alluded to before, I have attempted to show
how the process may work in an Amphibian egg, but a mathe-
matical demonstration alone would prove or disprove its correct-
ness.

In the developing egg of Rana temporaries we see a gradual
conversion of large segments into small segments, there being
always a rather sudden transition from the smaller, usually darker
cells, into the larger, usually less pigmented cells; this not very
well denned transition area forms what we might call a differentia-
tion zone gradually passing from the upper towards the lower
pole. There is no migration of cells, it is merely a wave like
progression of a zone of differentiation. After passing the equator
one arc of the zone instead of progressing over the surface dips
inwards, giving rise to a deep groove and then a cleft — the first
sign of the blastopore.

A section taken vertically through the egg and this groove
shows the same rather sudden mergence of the smaller dark cells
into the larger light cells at the bottom of this groove just as it
appears on the surface at the opposite side of the zone.

The part of the zone still on the surface sweeps on but adjoining
the spot where the groove is, more and more of it turns inwards
until at last the whole zone of mergence disappears from the surface
but may be found beneath, where it is probably still extending
and giving rise by differentiation to gut lining cells and causing
a cavity by a split, which widens later into the true archenteron.
This departure inwards — which does not involve any movement of
cells — is shown in the accompanying figure (Fig. 39), and con-
stitutes protogenesis, or as I called it in the paper alluded to, the
result of the primary centre of cell activity. The process is some-
what masked by another phenomenon, the growth backwards of
the rim of the blastopore Bl as soon as and wherever formed, which
ultimately produces the growth in length of the animal, and meta-
merically segmented region, i.e. deuterogenesis (v. arrow in
drawing). The ventral part dies out as an active proliferation

Mathematical Problems

85

area very soon, and the dorsal part continues and gives rise to
the tail. Could not the mathematician tell us exactly, given

a

d'

Bl

Fig. 39. An attempt to explain graphically the origin of the true archenteron in
the Anuran egg by cleavage as the direct action of the same forces which
produce the true archenteron in Amphioxus by invagination. a-af a line
which represents a zone of mergence of small and large segments (the effect
of cell division in the horizontal plane) at an early stage of segmentation [cf.
Morgan and Ume Tsuda, Q. J. M. 8. Vol. xxxv, PL 24, Fig. 111]. 6-6' the
zone at a later stage, c-c'-c" the zone at a still later stage after gastrulation
has begun. The zone has progressed superficially until after passing the
equator; then at one point it has left the surface and passed inwards and
backwards c". d-d'-d" the zone still later. It is now about to leave the
surface at all points and to complete the ventral lip of the blastopore. The
split caused by its ingression below the surface becomes archenteron — as soon
as the sides of the cleft separate. The "lips" formed by the ingression,
become ipso facto the secondary growth centre, because, whereas before the
turning in, any increase in bulk of the superficial layer (i.e. the blastula wall)
leads to mere increase in the size of the radially symmetrical form, after the
ingression any increase in size must on account of the folding or doubling of
the superficial layer on itself bring about a conversion of the radial to the
cylindrical form and thus initiate growth in length (deuterogenesis). This
must be true of all gastrulae.

a force such as postulated, what the distribution of the inert
or impermeable material must be to produce this form? The

86 Geometrical relation of Nuclei

problem is necessarily further complicated by the interaction of
the numerous force centres, though probably this effect is less
where yolk is abundant than where it is absent.

Some deny the occurrence of any form of imagination in Rana.
With them I would agree, except in so far as the withdrawal of
the yolk plug inwards from the blastopore (anus of Rusconi) to
form the gut floor at a later period may be called an invagination.
The cases of the Urodela and the Cyclostomata appear to be rather
different. Here it looks as though there were at a rather earlier
period a distinct inrolling of the lower pole to form the ventral
wall of the archenteron — really very much like the withdrawal of
the yolk plug just mentioned. Could not the mathematician
decide whether this can be or must of mechanical necessity be so ?
Could not the mathematician decide whether, on the hypothesis
of universal cell attraction put forward above, the combined effect
of attraction between cell and cell could account for the alleged
inrolling of the yolk mass in certain cases ?

To return for a moment to the differentiation zone alluded to
in the last paragraph but one.

After this has disappeared from the surface, that is to say, after
the complete formation on the surface of the circular blastopore or
anus of Rusconi, what is the subsequent history of this differen-
tiation zone? Does the differentiation cease? Or does it con-
tinue till it dies out by coalescence of all parts of the zone, during
which time a complete layer is differentiated as gut epithelium?
That it continues for a while, causing a split as s.uggested in my
paper, 1894, I am confident, but it is a moot point whether it is
not in some cases brought to an end by the confluence of the split
produced thereby with the original segmentation cavity as main-
tained by Hertwig, etc.

Brachet's careful work on Siredon and Rana temporaria makes
it probable that confluence is the rule in the former as it seems to
be in the Gymnophiona (Brauer) and may or may not occur in the
latter. But in either case the true archenteron is of protogenetic
origin and is the immediate result of the simple working of forces
above postulated, guided by extrinsic factors. Probably this could
be demonstrated mathematically.

So also in Amphioxus the gut cavity is brought into being by

Evidence from Rabbit Embryo 87

no other means than the working of the same universal force but
guided by differently arranged extrinsic factors (chiefly yolk), and
by the interaction of the differently arranged intrinsic factors — the
individual systems of the supposed form of energy.

There is no mysterious determinant, id or engram which in the
one case causes the gut cavity to be formed by splitting, in the
other by invagination, in a third case, e.g., mammal, by infiltration
of fluid combined with splitting.

The relationship of the special area of cell activity, which in
the vertebrate embryo gives rise to the growth in length (deutero-
genesis), to the primary area (protogenesis) again most probably is
capable of almost exact elucidation by mathematical means. In
my papers on the Kabbit there will be found figures in which
this relationship is roughly indicated, but they lack conviction,
because, if for no other reason, they lack mathematical proof.

There is certainly no mammal, probably no animal, better
suited for the study of this particular point than the Rabbit,
because the Rabbit embryo develops from the first to the eighth
day entirely free from the influence of the uterus, and is almost
devoid of yolk, so that we are there dealing almost exclusively
with the living matter and its properties. The zona radiata is in
this case an important extrinsic factor, but this is wholly outside
the living embryo.

If we get the relationship of the deuterogenetic to the proto-
genetic part of the organism clearly and mathematically defined,
should we have an explanation of the origin of the former? Or
should we have to fall back upon some supposed determinant, id
or engram ?

Research might give it by showing how it is in reality present
from the first, though not apparent in the earlier stages (e.g., in
Rabbit 1-5 days though obvious during the 7th day). If so, then
there is no need for a special id or other determinant which induces
at the right moment the growth in length.

A little reflection makes one realise that, with reference to those
animals in which an archenteron is produced by invagination or by
a process such as that characteristic of the Anura whereby a blasto-
pore and "blastopore lips" are formed, deuterogenesis is initiated
by the very fact of gastrulation, so that in those cases one can

88

Geometrical relation of Nuclei

speak of deuterogenesis being the direct and inevitable consequence
of the general structure of the unsegmented egg. It depends upon
the doubling of the expanding area upon itself.

This may be recognised more clearly by a glance at the diagram,
Fig. 40.

Let A represent a section of a sphere composed of living cells,

Fig. 40. A, a diagram representing the blastula stage. B, the gastrula. (7, an
embryo after growth in length has begun, the effect of which is indicated by
the dotted lines. Theoretically closure of the blastopores should precede
accumulation. Actually the two processes go on together as in the diagram.

which are actively dividing. The divisions occur only, or at any
rate chiefly, in the radial plane and, we will imagine, with equal
rapidity at all points. If this occurs in conjunction with the
growth of the divided cells to their "normal" size, the result
must be general increase in the size of the ring, or sphere, of which
the ring is supposed to be a section. The radial symmetry is in
no wise necessarily disturbed. For, given any point therein, the

Origin of Deuterogenesis 89

net result of growth must be expansion in all directions from that
point as indicated by the arrows. For simplicity I represent three
such areas of expansion by the three pairs of arrows. Now let it
be supposed that the original arrangement of the forces involved
be such that when a certain size has been attained (as in the
description of the invaginating gastrula of AmpJiioxus discussed
in the earlier part of this essay) invagination must occur, then as
a result, the proliferating areas are doubled upon themselves,
Pig. 40 B, and at the point of doubling the resultant effect must
be a radial accumulation of cells and the consequent destruction of
the radial symmetry, and the origin of growth in length (accom-
panied by at any rate partial closure of the blastopore), which will
be continued so long as the margins of folding retain their power
of production of undifferentiated cells. This is shown to be so in
the diagram by the direction of growth which before invagination
has a counterbalancing effect and so the status quo or radial
symmetry is retained, but, as soon as an antagonism is converted
into a co-operation and the balance is upset, the blastopore must
tend to close and on the accumulation of radially arranged material
growth in length ensues. That is to say, deuterogenesis is as in-
evitable a consequence of the doubling of the proliferating area as
the doubling is, on my hypothesis, an inevitable result of the original
general structure of the egg and it requires no special determinant
to call it into being.

I confess that with regard to the origin of the deuterogenetic
centre in Amniota it is not so simple a matter. In Amniota, as
I believe, there is no true blastopore. In a Rabbit for instance the
deuterogenetic centre seems to arise "of its own accord." For the
moment the problem must remain unsolved, but further observa-
tion and reflection may show it to be as inevitably a consequence
of what has gone before, as it is in the case of gastrulas with
typical blastopore.

Clearly this explanation of the origin of the deuterogenetic
centre, is, if correct, applicable to all gastrulae formed by in-
vagination, or indeed, all that have a distinct blastoporic lip. If
an organism retains a radial symmetry after gastrulation, it must
do so by the dying out of the deuterogenetic centre of cell pro-
liferation. By a retention of one part, and suppression of another

90 Geometrical relation of Nuclei

different types of bilaterally symmetrical forms will be attained.
It is characteristic of the Chordata (excepting probably the
Enteropneusts) that the ventral part of the deuterogenetic centre
dies out early, so that growth in length in that phylum is chiefly
attributable to the dorsal part of the original blastopore lip. This
dorsal part is the part of the lip, which is earliest formed in
Amphibians, and as it begins its growth at once, before the whole
archenteron is completed, the correct understanding of the pro-
cess has been a matter of some difficulty, and has been arrived at
only after much research and not a little controversy.

No doubt the denial of ids and engrams during development
only puts back the difficulty to the previous generation, to the
building up of the germ cells. But there is this difference. The
potter turns out jar after jar exactly alike not because the clay
contains an id or engram that causes the special turn of the lip or
width of neck to appear at the right time, but because the mould
is the same in each case.

The development of the individual is the building up of the
mould which forms the next generation (i.e., the germ cells), which
in most cases is all that is concerned, but in mammals at any rate
the moulding is continued to a much later period of the succeeding
generation's span of existence.

The attraction and repulsion observed between cell and cell
are certain of the manifestations of this supposed form of energy
— but probably not by any means all; just as attraction and
repulsion are manifestations of electrical energy under certain
conditions, but are not by any means the only manifestations.
In nerve impulses we may, for instance, really be experiencing
manifestations in another way of the same form of energy which
under other conditions produces the attractions and repulsions
and the figures of strain in the dividing cells, and the actual cell
division.

Herbst made the remarkable observation that the segments of
developing Echinus eggs and others when allowed to develop in
sea water which has been deprived of calcium salts, separate from
one another, instead of adhering ; and that although they live so
long as to develop cilia as in a fully formed blastula, nevertheless
they do not form a completely closed vesicle. This observation

Experiments on Echinus' eggs 91

perhaps tends to show how important this attraction property is
for the normal development of embryos. The fact that under
these peculiar conditions the living blastomere does not show
attractive properties does not necessarily upset the hypothesis
that attraction is a general truth, any more than the fact that
although a glass vesicle sinks in ordinary water, yet, if a certain
percentage of some salt be added to the water it exhibits repulsion
rather than attraction to the centre of the earth, upsets the
universality of the force of gravitation. In the latter case a
better knowledge of the laws of gravitation enables one to under-
stand the "abnormality," whereas the abnormal behaviour of
segments developing in sea water without calcium salt remains
for the present a mystery. It has been suggested in this particular
case that the action of the calcium salt is to help in the formation

N.

of a cementing material which in its absence is not formed. This
however seems to me hardly a sufficient explanation, for the
presence of a cement would not of itself account for the flattening
of the segments against one another, nor for the fact that on adding
again the necessary calcium, adherence follows.

The idea of Entelechy developed by Driesch in his Gifford
Lecture is of the most intense interest; but it must be allowed
that the conception is almost mystical.

Now, although the development of the egg up to the formation
of the archenteron is, as compared with the later stages of ontogeny
or still more as compared with regenerative processes, almost in-
finitely simple, yet nevertheless it involves stages of the greatest
importance, namely establishment of the gut cavity, and differen-
tiation of the tissue into the groundwork of the ectodermal, and
endodermal tissues, that is to say, the formation of layers whose
subsequent fate can be foretold.

If Entelechy is the ruling influence of life, ought it not to be
the ruling influence in such a process as the gastrulation of the
Amphioxus by invagination of a blastula? If it can be shown
that this process is explicable by a general application of a simple
force in combination with other well-known factors, the probability
of which is claimed to have been shown in the foregoing pages and
elsewhere not only in Amphioxus, but in Lepus and Rana, may we
not doubt whether, if it were possible to analyse the almost

92 Geometrical relation of Nuclei

infinitely more complex processes of later development, we should
find so mystical an explanation as entelechy necessary ? In other
words, may not this simple force be the sole factor in the develop-
ment of organisms which is to be regarded as "vitalistic," that is
to say peculiar to living matter — exhibiting the essential character
which distinguishes living from non-living substances.

For by this supposed form of energy, I do not mean a mysterious
metaphysical influence, but a form of energy comparable to gravity,
electricity, or magnetism — in some respects similar to these but in
other respects differing from each, and a form which could be in-
vestigated by the ordinary methods of mensuration and computa-
tion available to the mathematician.

Professor Bateson, who was so good as to read the foregoing
pages, pointed out to me (among other kind and valuable criticisms,
for which I desire to express my thanks) with reference to the
hypothesis of an intercellular attraction as a possible cause of
invagination, that, given the fact that the pull between cell and
cell is eccentric at one part of a ring, or over one area of a hollow
sphere, invagination would occur if the mutual pull were due to
a contraction of protoplasm just as much as it would on the
hypothesis of intercellular attraction.

No doubt that is so; and it must be admitted that whereas
contractility is recognised as an attribute of protoplasm, the
attraction between cell and cell as indicative of a special vitalistic
property is "not proven."

If contraction is the cause of invagination it is necessary to
assume either:

(a) that there is an early differentiation into contractile tissue
for the special purpose of causing gastrulation by- invagination, or

(b) that there is a relationship between more contractile and
less contractile protoplasm so as to exert a greater pull eccentrically
over a certain area as regards the cellular units corresponding in
effect to the result of the forces postulated by the attraction hypo-
thesis.

It must be admitted, that there is nothing visibly present which
suggests a differentiation into more and less contractile protoplasm.

The contraction hypothesis necessitates a cement, or other firm
means of attachment between cell and cell, or the units would be

Experiments on Frogs' eggs 93

pulled asunder. So far as I know invagination to form a gastrula is
never spasmodic. There are never any perceptible contractile
movements. On the other hand, the attraction hypothesis requires
no special teleological developments but presupposes a constant
and universal principle which, guided by other factors, in this case
largely inanimate, leads inevitably under normal conditions to a
certain definite result.

With a view to determining the nature of the alleged attractions
between cell and cell such as have been stated to occur between
the isolated blastomeres of gastrulating frogs' eggs, I repeated
some of Roux's experiments by isolating the blastomeres of Rana
temporaries and Bufo vulgaris and observing any changes in position
or shape which take place when such cells are floating in ordinary
water, or in a mixture of salt solution and albumen as used by
Roux. On the whole I found the attraction to be less noticeable
than I had expected.

In many cases, that is to say between many pairs of cells which
were closely situated to each other, there was no movement at all.
This Roux also found. But on the other hand certain cells origin-
ally a short distance apart from one another — perhaps half a cell
diameter — certainly approached one another and came actually in
contact; but I did not see any that became pressed together so
as to be flattened, indicating any considerable strength of attrac-
tion force.

I experimented on frogs' eggs which were laid on the 27th March,
and on toads' eggs a fortnight or so later in April.

They had been some time in dishes on my balcony, and were
in the gastrulating stage, the lower lip of the blastopore was just
about to form. I broke the eggs and gently teased the cells apart,
some in the water in which they were and some in a mixture of
| per cent, salt solution and filtered white of egg after Roux's
instructions.

With the former I was much more successful in observing
attraction movements, with the latter I was more successful in
seeing pseudopodial movements.

After treatment of an egg as described a large number of the
cells are isolated, though many pieces of gastrula wall remain
unteased.

94 Geometrical relation of Nuclei

Many also are broken; but floating among the debris are
hundreds of isolated cells as well as pairs and aggregates of three
and more. Many of the cells are quite spherical as soon as one
gets the preparation under observation. Other cells are oblong,
or compressed in some way or other, and are seen to assume
slowly a more and more spherical condition. Some of the pairs
are so placed that the individuals touch one another at a point
only, others are flattened against one another. The latter are
probably cells in the act of division.

I have seen isolated cells which are quite close to one another
come in contact, but I have not seen them flatten up against each
other.

Isolated cells which are farther apart sometimes show less rapid
and less regular attraction movements. For instance : there were
two cells separated from each other by 6'5 divisions of the micro-
meter. The one on the left, the larger of the two, moved only
slightly; the one on the right, the smaller, moved considerably,
so that eventually they were only 2-0 divisions apart. The move-
ments were of two kinds, slow sliding movement and sharp jerks.

But many of such movements, obviously, need have nothing to
do with attraction and are amoeboid.

There are undoubted changes of shape in the isolated cells.
The first tendency on separation is towards the assumption of a
spherical form. This is followed by more or less marked changes
in shape, sometimes so great as to amount to the protrusion of
pseudopodia. So that it is probable that the slower and less certain
movements are due to amoeboid activity, and the approximation of
two cells may be fortuitous. And if the protrusion of pseudopodia
is evidence of the contractility of some part of the cytoplasm of the
cell, then clearly the cells must at this time be capable of contractile
movements, which if applied in a particular way would bring
about the invagination of the blastula as Professor Bateson
suggests.

Thus in Fig. 41, two isolated blastomeres from a toad's egg in
the gastrulating stage are shown as they appeared at intervals
between 5.30 p.m. and 7.30 p.m. on April 15. In this case the
slight reduction in the distance between the two cells must almost
certainly have been due to the amoeboid movements exhibited

Experiments on Toads' eggs

95

by each, and in respect of one, these movements were so great as
to result in well marked pseudopodia.

In Fig. 42, which is a series of drawings of three isolated cells
from gastrulating toads' eggs similar amoeboid changes are seen
and a slight approximation between two of the cells, a and 6, has
been effected.

Now these changes in shape of the isolated cell indicate in
all probability contractility, at least that is one explanation of
amoeboid movement, and so show that the power to exert a pull
by contraction is not out of the question.

5.30

o o

556

o o

O o

6.9

o

6.13

o

o

611k

o

o

o

7.30

Fig. 41. Seven drawings of two isolated cells of a Toad's egg in the gastrulating
stage, drawn with a camera lucida at intervals as indicated.

On the other hand there are movements, seen especially when
the preparation is first made, which cannot be ascribed to any
form of amoeboid activity, and as far as I can see not open to
explanation by contractility.

Again, it may be that the attraction between cell and cell is
greater immediately after division than at a later period. One
reason why such movements occur more obviously immediately
after the preparation is made than later is probably to be sought
in the tendency of the isolated cells to adhere to the slide, which

96

Geometrical relation of Nuclei

friction may be sufficient to prevent movement due to intercellular
attraction.

If there is an attraction between cell and cell as Roux and
others have suggested, that intercellular attraction is a factor
which must be taken into account in the study of the mechanics
of development. It is probably only in the more delicate tissues
of the early stages that the attraction force can have any mechani-
cal effect. In later stages and in the more bulky tissues its effect

Z4i

0 ©
©

£54

© ©
0

3.5

0 0
©

0 &'"

0

2.56

0 0

0

3.9

& &
0

249

'© 0

0

0 0
"0

3.13

Q ©
0

Fig. 42. Nine drawings of three cells isolated from a Toad's egg in the gastrulating
stage drawn under a camera lucida at intervals of about four minutes as
indicated by the figures.

would be masked by many other factors. It is quite conceivable
that contraction, at least in its primitive condition as an attribute
of protoplasm and attraction may be phenomena due to the same
cause as, for instance, contraction of an envelope like the earth's
atmosphere is due to the same cause as attraction of the earth's
satellite, namely to gravitation.

Whether invagination of the Amphioxus gastrula and of other
gastrulas of a similar type is due to the existence of an intercellular

Conclusion 97

attraction or to an intracellular contraction (if not to some other
cause altogether) the geometrical relationship of the nuclei con-
sidered in either connection is a matter of some interest.

While strongly inclined to adhere to the attraction hypothesis
because it is simpler and more fundamental in its application I fully
admit the force of Professor Bateson's criticism.

I admit quite frankly that I am advocating a vitalistic explana-
tion for the prime biological phenomenon, namely cell division.
By this I mean a form of energy evolved by and peculiar to the
complex nature of the molecule of protoplasm — or of protoplasms
-which exhibits so long as it retains the power of persistence an
unceasing recurrence of a bipolar state out of a unipolar state
which since the material is fluid induces the separation of the
material aggregated round each pole, thereby multiplying the
units and tending towards the establishment of more and more
complex systems.

At so early a stage of inquiry it is impossible to determine how
the bipolar state is to be derived from the unipolar state. Possibly
it may be always bipolar, but when the energy is at its lowest ebb,
namely when the cell is, as we say, in its resting stage, the poles
are so close to one another as to be indistinguishable and to act
in respect of external bodies as a single pole. But when the evolu-
tion of the energy grows in intensity as the result of metabolic
activity the poles are driven further and further apart owing to the
fluid nature of protoplasm until their separation occurs, whereupon
each mass becomes instantly bipolar; but now the storm is over
and the two poles in each daughter cell or nucleus lie so close
to one another as to be indistinguishable. It is interesting to
remember that in some cases the centrosome (which may when
present mark the position of the pole or poles) appears to be
doubled almost as soon as the daughter nucleus is formed.

Perhaps, however, this may be altogether too fanciful ; but my
point is, if we are to have a vitalistic theory of biology I submit
that it must be sought for somewhat along the lines here indicated
rather than in the more mystical form presented by Driesch's con-
ception of Entelechy, though Entelechy might be a coordinating
complex of this and other forces.

It is with the hope of gaining the attention of those who are

A. V. E. 7

98 Conclusion

mathematicians as well as biologists that I venture to offer this
essay for publication.

My sincere thanks are herewith tendered to Professor Bateson,
Mr Blackman, and Dr Jenkinson who were so good as to read the
draft of the above essay, for their trouble in doing so, and for their
criticisms.

LITERATURE REFERRED TO

ASSHETON, R. (1894). "The Primitive Streak of the Rabbit, the
causes which may determine its Shape, and the part of the
Embryo formed by its activity."

Quart. Journ. Micr. Sci. Vol. xxxvu.

55

(1894). ;'0n the Growth in Length of the Frog Embryo.

Quart. Journ. Micr. Sci. Vol. xxxvu.

(1894). ''A. Re-investigation into the early stages of the
Development of the Rabbit."

Quart. Journ. Micr. Sci. Vol. xxxvu.

(1896). "An Experimental Examination into the growth
of the Blastoderm of the Chick."

Proc. Roy. Soc. Vol. LX.

- (1898). "The Development of the Pig during the First Ten
Days." Quart. Journ. Micr. Sci. Vol. XLI.

(1898). " The Segmentation of the Ovum of the Sheep, with
observations on the Hypothesis of a Hypoblastic Origin for
the Trophoblast." Quart. Journ. Micr. Sci. Vol. XLI.

(1905). "On Growth Centres in Vertebrate Embryos."

Anat. Anz. Vol. xxvu.

(1907). "Certain Features Characteristic of Teleostean
Development." Guy's Hospital Reports. Vol. LXI.

— (1912). ;; Gastrulation in Birds."

Quart. Journ. Micr. Sci. Vol. LVIII.

BALFOUR, F. M. (1885). "Vertebrate Embryology." Macmillan.

100 Literature

BATESON, W. (1884). "The Development of Balanoglossus."

Quart. Journ. Micr. Sci. Vol. xxiv.

VAN BENEDEN, E. (1912). "Recherches sur 1'Embryologie des
Mammiferes." Arch, de Biol. Vol. xxvu.

BRACKET, A. (1903). ;' Recherches sur Fontogenese des Amphibiens,
Urodeles et Anoures (Siredon pisciformis — Rana temporaria)."

Arch, de Biol. Vol. xix.

BRAUER, A. (1897). "Beitrage zur Entwickelungsgeschichte der
Gymnophiona." Zool. Jahrbuch.

CERFONTAINE, P. (1907). " Recherches sur le developpement de
FAmphioxus." Arch, de Biol. Vol. xxn.

CONKLIN, E. G. (1905). "The Orientation and Cell Lineage of the
Ascidian Egg (Cynthia partita)."

Journ. Acad. Sci. Philadelphia. Ser. 2. Vol. xm.

CONN, H. W. (1884). "Development of Serpula."

Zool. Anz. Vol. vii.

DOHRN, A. (1875). " Der Ursprung der Wirbeltiere und das Princip
der Functionswechsels." Leipzig.

DRIESCH, H. (1908). "The Science and Philosophy of the
organism." (The GifTord Lectures. 1907, 1908.) London.

DUVAL, M. (1889). "Atlas d'Embryologie." Paris.

GASKELL, W. H. (1908). "The Origin of Vertebrates." London.

GOETTE, A. (1884). "Abhandlungen zur Entwicklungsgeschichte
der Thiere." Heft 2. Hamburg u. Leipzig.

(1886). " Untersuchungen zur Entwicklungsgeschichte von
Spongilla fluviatilis." Loc. cit. Heft 3.

GOODALE, H. D. (1911). "The Early Development of Spelerpes
bilineatus." Amer. Journ. Anat. Philadelphia.

HARTOG, M. (1905). "The dual force of the dividing cell. Pt 1.
The Achromatic Spiral Figure illustrated by Magnetic Chains
of Force."

Proceedings of the Royal Society. Series B. Vol. LXXVI.

Literature 101

HATSCHEK, B. (1881). "Studien iiber die Entwicklung des
Amphioxus." Arch. Zool. Inst. Wien. Bd. iv.

(1909). :; Studien zur Segmenttheorie des Wirbeltiere."

Morph. Jahr.

HERBST, C. (1900). 'Ueber das Auseinandergehen von Furchungs-
und Gewebezellen in kalkfreiem Medium."

Archiv f. Entw.-Mech. Bd. ix.

HERTWIG, 0. (1906). "Handbuch der Vergleichenden und
Experimentellen Entwickelungslehre der Wirbeltiere." Vol. i.
Jena.

His, W. (1874). 'Unsere Korperform und das Physiologische
Problem ihrer Entstehung." Leipzig.

(1874). "Ueber die Bildung des Lachs-Embryo." Leipzig.

HUBRECHT, A. A. W. (1905). " Gastrulation of Vertebrates."

Quart. Journ. Micr. Sci. Vol. XLIX.

(1912). "Friihe Entwickelungsstadien des Igels und ihre
Bedeutung fiir die Vorgeschichte (Phylogenese) des Amnions."

Zool. Jahr. Bd. in.

JENKINSON, J. W. (1913). "Vertebrate Embryology." Oxford.

KASTSCHENKO, N. (1888). "Zur Entwickelungsgeschichte des
Selachierembryos." Anat. Anz. m.

KEIBEL, F. (1901). "Die Gastrulation und die Keimblattbildung
der Wirbeltiere." Erg. d. Anat. und Entw.-Gesch. Vol. x.

KLAATSCH, H. (1897). "Bemerkungen iiber die Gastrula des
Amphioxus." Morph. Jahrb. Bd. xxv.

KOPSCH, F. (1896). "Experimentellen Untersuchungen iiber
den Keimhautrand der Salmoniden."

Verh. u. Anat.-Gesel. Berlin.

KOWALEVSKI, K. (1867). " Entwicklungsgeschichte des Amphioxus
lanceolatus." Mem. Acad. Imp. de St Petersbourg. T. xi.

102 Literature

DE LANGE, D. (1912). " Mitteilungen zur Entwickelungsgeschichte
des Eiesensalamanders (Megalobatrachus maximus Schlegel)."

Anat. Anz. Vol. XLII.

LEGROS, B. (1907). "Sur quelques cas d'asyntaxie blastoporiale
chez FAmphioxus." Mitt. Zool. Stat. Neapel. Bd. 18.

LILLIE, F. (1908). "Development of the Chick." New York.

LWOFF, B. (1894). "Die Bildung der primaren Keimblatter und
die Entstehung der Chorda und des Mesoderms bei den
Wirbeltieren." Bull, de la Soc. Imp. des Nat. de Moscou.

MACBRLDE, E. W. (1898). ' : The early Development of Amphioxus."

Quart. Journ. Micr. Sci. Vol. XL.

(1910). "The Formation of the Layers in Amphioxus and
its Bearing on the Interpretation of the Early Ontogenetic
Problems in other Vertebrates."

Quart. Journ. Micr. Sci. Vol. LIV.

MARSHALL, A. M. (1893). "Vertebrate Embryology." London.

MORGAN, T. H. (1896). "The number of Cells in Larvae from
Isolated Blastomeres of Amphioxus."

Arch. f. Entw.-Mech. Bd. in.

(1907). "Experimental Embryology." Macmillan.

and HAZEN, A. P. (1900). ' The Gastrulation of Amphioxus."

The Journal of Morphology. Vol. xvi.

PATTEN, W. (1912). "The Evolution of Vertebrates and their
Kin." Philadelphia.

PATTERSON, J. T. (1909). "Gastrulation in the Pigeon's Egg.
A Morphological and Experimental study."

Journ. of Morph. Vol. xx.

PRZIBRAM, H. (1908). " Embryogeny." University Press, Cam-
bridge.

BEINKE, F. (1900). "Zum Beweis der trajectoriellen Natur der
Plasmastrahlungen." Archiv f. Entw.-Mech. Bd. ix.

Literature 103

RHUMBLER, L. (1902). ;' Zur Mechanik des Gastrulationsvorganges,
insbesondere der Invagination."

Arcliiv f. Entw.-Mech. Bd. xiv.

Roux, W. (1894). "Ueber den ' Cytotropismus ' der Furchungs-
zellen des Grasf rosettes (Rana fusca)."

Archiv /. Entw.-Mech. Bd. I.

- (1896). "Ueber die Selbstordnung (Cytotaxis) sich 'beriih-
render' Furchungszellen der Froscheier."

Archiv f. Entw.-Mech. Bd. in.

RUCKERT, J. (1899). "Die Erste Entwickhmg des Eies der
Elasmobranchier." Jena.

SAMASSA, P. (1898). "Studien iiber den Einflusz des Betters auf
die Gastrulation und die Bildung der primaren Keimblatter
der Wirbeltiere. IV. Amphioxus."

Archiv f. Entw.-Mech. Bd. vii.

SEDGWICK, A. (1884). ;' On the Origin of Metameric Segmentation
and some other Morphological Problems."

Quart. Journ. Micr. Sci. Vol. xxiv.

(1886). " On the Fertilised Ovum and the Formation of the

Layers of the S. African Peripatus."

Quart. Journ. Micr. Sci. Vol. xxvi.

SEMPER, C. (1874). "Ueber die Stammesverwandtschaft der
Wirbeltiere und Anneliden." Cent. f. Med. Wiss.

SHEARER, C. (1911). "On the Development and Structure of the
Trochopore of Hydroides Uncinatus (Eupomatus)."

Quart. Journ. Micr. Sci. Vol. LVI.

SOBOTTA, J. (1897). " Beobachtungen iiber den Gastrulations-
vorgang beim Amphioxus."

Verh. d. Physikal.-Mediz. Ges. zu Wilrzburg. Bd. xxvi.

STOSSICH (1878). "Beitrage zur Entwickelung der Chaetopoden."

Sitz. d. k. k. Akad. Wiss. B. 77- .

104 Literature

THEEL, HJALMAR (1892). "The Development of Echinocyamus
pusillus."

N. Acta Reg. Soc. Sc. Upsala. Na. Ser. Abstr. in Nature.
Vol. XLVIII. p. 330.

TREADWELL, A. L. (1901). "The Cytogeny of Podarke obscura."

Journ. of Morph. Vol. xvn.

WILL, L. (1892). "Beitrage zur Entwickelungsgeschichte des
Reptilien." Zool. Jahr. Vol. vi.

WiLLEMOES-SuHM, E. V. (1871). " Biologische Beobachtungen
iiber niedrige Meeresthiere." Zeit. f. Wiss. Zool. Vol. xxi.

WILSON, G. B. (1892). :'0n multiple and partial Development in
Amphioxus." Anat. Anz. Vol. vn.

(1893). "Amphioxus and the Mosaic Theory of Develop-
ment." Journ. of Morph. Vol. vin.

ZUR STRASSEN, 0. (1903). "Ueber die Mechanik der Epithelbil-
dung." VerJi. d. deutsch. zool. Ges. S. 91-112.

CAMBRIDGE: PRINTED BY j. B. PEACE, M.A., AT THE UNIVERSITY PRESS

• 0

10

t>

H

OQ:

W;

<

cai

S

CO

w

o

•«-*:

bl)
O

* £

«! *r«
O

•B

University of Toronto
Library

DO NOT

REMOVE

THE

CARD

FROM

THIS

POCKET

Acme Library Card Pocket

Under Pat. "Ref. Index File"
Made by LIBRARY BUREAU