If the cells of the inner mass do not multiply quickly enough, and if they are not connected together firmly, and if they adhere by some means or other to the outer layer, then, as the outer layer expands, there must be a tendency for the general separation of the cells of this inner mass, and this will be most apparent at the edges. That is to say, the centre, originally thicker than the edges, will become thinner, and at the periphery the edges will be drawn out to an irregular or regular fringe according to less or more regular local growth and expansion of the outer epiblastic layer.

If the spread of these is due to this cause, and if the growth is equal over the whole surface of the sphere, then, in a section taken through the centre of the sphere and the centre of the embryonic disc, the arc along which these isolated cells are found should subtend an angle equal to that subtended by the compact embryonic disc up till now; that is to say, an angle of about 80°. But on measurement at a rather later stage, 102 hours after coition, represented in fig. 39, the angle is found to be very considerably wider than 80°, in fact somewhere between 110° and 130° (the limit of these cells in question being very irregular), and by the eighth day an angle of 200°.

Hence it would seem that this does not account for the apparent growth round. If, however, we have any reason to suppose that the cells of the outer layer multiply more rapidly and thus allow the expansion of that part of the sphere to take place quicker around a zone bounded roughly by the edges of the compact embryonic area on the one hand and some other parallel line about the original equator on the other hand, we could still consider that the suggested cause is the correct one. I think we have, for these reasons:

The blastodermic vesicle of 96 hours is a sphere. The blastodermic vesicle of 120 hours is no longer a sphere, but is a body of such a nature that a horizontal plane taken through its longest diameter will not pass through the equator, but will be nearer the upper pole than the lower.

The circumference of this plane and of all planes taken parallel to it are, I think, as yet circles. There is nothing so far to indicate which will be anterior or posterior end of the embryo. The blastodermic vesicle of about the 140th hour is in shape of the same character, but more markedly so.

In the blastodermic vesicle of the 175th hour sections taken parallel to the equator are now no longer circles, but are ovoidal. The fact that there are now a long and a short axis may be entirely due to the pressure upon the vesicle exerted by the walls of the uterus; but I want to point out more especially that the horizontal sections are not true ellipses, but have one end larger than the other. The large end corresponds approximately, but by no means always exactly, with the future posterior end of the embryo.

A median longitudinal vertical section of this stage is very instructive, for the asymmetry is very marked. One end, the future posterior end, is much more bulky than the anterior end. The embryonic area is always placed nearer to the anterior than to the posterior end.

Fig. 42 shows four stages, namely, at the 100th hour, the 125th, 140th, and 175th. A. is the more anterior, P. the more posterior end. The thick black line in each case is the outer layer of cells or epiblast; the thin black line represents those cells of the inner mass which become separated to form the hypoblast. It is quite clear that a very great change in shape as well as size has occurred between the 96th and 168th hours. How has this been produced?

I have argued that there is a hydrostatic pressure within the blastodermic vesicle causing it to expand. The pressure is sufficient to cause the stretching and flattening out of the cells of the wall of the vesicle, and also to cause the stretching and expansion of the very tough albumen layer. On the other hand, we know that the pressure is not so great as to rupture either the vesicle wall or the albumen layer; in other words, the vesicle wall and albumen layer are sufficiently strong to resist the hydrostatic pressure for, at any rate, some considerable time. The albumen layer becomes continually more

and more stretched, and quite late it does, as a matter of fact, rupture, but at a time (the ninth day) that does not concern the present question. No addition is ever made to the thickness of the albumen layer after the embryo leaves the Fallopian tube; it is an inanimate structure. The cellular wall of the blastodermic vesicle is, on the contrary, living, and capable of adding to itself at any part of its area. Like the albumen layer, it becomes greatly stretched and becomes very thin (v. figs. 21-29), and unless it received additional matter (i.e. multiplication of the cellular units) it would rapidly thin out altogether. After about the 100th hour the cellular wall ceases to get any thinner. Up to this moment we must suppose that the rate of increase of hydrostatic pressure has been in excess of the rate of addition of material to the cellular wall of the vesicle and so has stretched it, but from now there is no appreciable thinning out of this cellular wall (v. figs. 27, 28, 29, and 34).

This means, I believe, that the increase of cellular tissue just balances the increase of hydrostatic pressure, and so the vesicle grows in size, the thickness of the cellular wall remaining unaltered. Now there is no reason, as far as I can see, to doubt the albumen layer being equally tough on all sides of the embryo. The albumen layer is not secreted by the embryo, but is applied by the Fallopian tube.

Although preserved specimens and sections are not well adapted for this purpose, still my figures show that there is no regularity at all in these preserved specimens, such as a thicker part of the albumen layer being present over any one part of the embryo in stages up to the 96th hour (v. figs. 16, 20, 21, 22, 23, and 24), and in the fresh specimens the outer and inner limits of the albumen layer present in optical section true concentric circles. Therefore I do not think we can attribute any subsequent difference in thickness of the albumen layer to a difference in texture acquired by that albumen layer during its deposition. At a later stage we do find a difference in thickness occurring as a constant character.

The difference I find is as follows:

The most marked difference concerns the part of the albumen layer adjoining the embryonic disc. This is in the later stages very much thicker than elsewhere. Figs. 30 and 31 are portions of the upper and lower poles of an embryo taken from a rabbit of the one hundred and forty-fourth hour.

The albumen layer is at the upper pole about twice as thick as it is at the lower pole. This is a perfectly constant feature, and becomes more and more marked the later the stage.

Hydrostatic pressure exercises its influence equally in all

directions. If it encounters less resistance in one direction than in another it will cause greater effects in that one direction. I have argued above that at any one moment (between one hundredth and one hundred and sixty-eighth hours) the hydrostatic pressure within the vesicle on the one hand, and the living cellular wall of the vesicle together with the nonliving albumen layer on the other hand, are in a state of equilibrium. The blastodermic vesicle is always taut, but does not rupture. But the next moment the hydrostatic pressure has increased on the one hand, and the albumen layer has become thinner and the cellular wall has increased its material -and still the equilibrium is maintained. It seems to me to be clear that the degree of rapidity of increase of the cellular wall must be a factor in the resistance afforded to the hydrostatic pressure by the two walls of the vesicle.

If this is so, then it follows that at any area where the increase of cellular tissue is greatest, there the hydrostatic pressure will exert its greatest influence, and there the albumen layer will, as a consequence, be thinnest. This is, of course, dependent on the close attachment of the cellular layer to the albumen layer. I am bound to confess I do not find it easy to prove that there is no sliding of the albumen layer over the cellular layer; but, on the other hand, I see no evidence to suggest that there is such a sliding.

Accordingly, I take it that the great thickness of the albumen layer adjoining the embryonic disc over that adjoining the lower pole shows that there has been less stretching and

therefore less growth at the embryonic pole than at the lower pole.

It may be said that this is due to the fact of the cellular wall at this point being practically several layers thick. But it must be remembered that the cells of the inner cell-mass are very loosely arranged (figs. 28, 30, 35) at this time, and certainly do not give me the impression of being under great tension, as are the outer cells.

As far as the tension is concerned, I believe the outer layer has to bear it very nearly all, even in the embryonic disc. Sections show how very thin it is here.

Granted that they are only very equally stretched, it follows, I think, since the albumen shows not the same amount of stretching as elsewhere, that the multiplication of cells must have been going on more slowly in the embryonic region. Now from measurements of the albumen layer it will be seen that the thinnest place is not usually at the lower pole, but somewhere between the equator and the upper pole. This, then, marks out as an area a zone of more rapid multiplication of cells, that area which lies just outside of the embryonic disc.

Now we know that upon the eighth and many subsequent days there is a very great activity evinced by the cells of a zone immediately surrounding the embryonic disc; it is the zone called by Duval the ectoplacental area. If it is an area of very great activity upon the tenth day, of great activity upon the ninth and eighth days, when does it begin to be an area of activity? Why not upon the seventh or sixth, or even on the fifth day?

I believe that it does begin as early as the fifth day, and that it is to the presence of this area of more active cell division round the embryonic disc that the shape of the blastodermic vesicle is as I have described it to be (v. fig. 42), and that it is to this that the apparent growth round of the hypoblast cells is due.

The ectoplacental area is, as is well known, much more strongly developed all round the posterior end of the embryo