No. 3 was also in two segments, and was examined, measured, and drawn in aqueous humour. Result of measurement :

No. 4, the nearest specimen to the uterus, was as yet unsegmented, possibly an immature or unfertilised one, as the ovum completely filled the space within the zona pellucida. I could not find any trace of polar bodies.

All three specimens from this rabbit, which were in two segments, I examined most carefully, to find, if possible, evidence of the marked difference in appearance described by van Beneden. I treated all with 2 per cent. osmic acid and used weak staining solutions of picro-carmine and aniline blueblack, of picro-carmine, of Beale's carmine and aniline blue-black respectively, but to no purpose. They were then mounted in glycerine jelly, and now, after the lapse of a year, I still fail to find any difference between the two segments.

From another rabbit I obtained two specimens in the twosegment stage, but was unfortunately only able to make a very cursory examination. One of these I thought I distinguished as having a larger darker segment and a smaller clearer one, but I could only examine it while lying on the folds of the Fallopian tube, the unevenness of which might easily have caused one segment to appear darker. So I cannot advance even this instance in support of there being a marked difference in appearance between the two segments.

Another rabbit, with ova aged 25 hours, gave me four, of which two were in the two-segment stage, and two showed four segments. Here again there was no perceptible difference. I did not measure these while fresh. One I have since measured,

and find that the longitudinal and transverse axes of the one measure exactly the same as those of the other.

A tabular view of the measurements of these two first cleavage segments may be of interest, showing how the size of one specimen may vary with that of another, and the variation in size of the two segments of one and the same specimen.

Each segment of each specimen when examined immediately after the death of the animal showed a denser, more granular inner portion, and a clearer, almost hyaline outer layer, the nucleus being situated in the denser inner portion (v. fig. 2). This difference is more evident in the early stages of segmentation, up to the time that there are twelve to sixteen segments, than later.

Stage with Four Segments.

The second plane of cleavage seems to be at right angles to that of the first. It appears that the two segments divide about, if not exactly, at the same time. This occurs about twenty-five or twenty-six hours after coition. The time between the segmentation of one into two and between two into four appears to be nearly two hours. Since there may be some difference in size between the two primary segments, it follows that there is also very frequently a difference in size between

the cells of the four-segment stage, and, as one would expect, it may appear that two segments are small and two larger, as in fig. 6.

Development between the 26th and 72nd Hours.

Stages of Segmentation between Four-Segment Stage and Commencement of Cavity of the Blastodermic Vesicle.

The further cleavage of the four segments does not occur in each segment at the same moment. A stage of five segments or even seven may often be found.

In a rabbit killed 27 hours after coitus (Rabbit 36) I found in the right Fallopian tube two ova in the usual locality for this age, which is about 40 mm. above the uterus; one of them was in five segments, the other in seven. The fivesegment one is shown in fig. 7.

In this specimen it will be seen that two (L2.) spheres are almost exactly the same size, and that these are considerably larger than two (S3.) of the remaining three, and slightly larger than the third (S.). This specimen I surmise may have been one in which the two primary segments were unequal, and that each of these divided into the approximately equal spheres, and that at the moment of examination one of the daughter cells of the primary smaller (?) one had divided again into two very nearly equal spheres (S3.).

It should be noted that in van Beneden's account of the process the larger primary segment gives rise to the more. rapidly dividing daughter cells forming his so-called epiblast; while in this case it seems to be the descendant spheres of the smaller primary segment which appear to be the more ready to undergo division.

Fig. 14 is a camera lucida drawing of the other embryo of the same Fallopian tube, seen as a transparent object. In this specimen one segment (L2.) is larger than the others. The four marked S3. were approximately equal, and slightly larger than the two (L.). May this be interpreted as follows? The ovum which gave rise to this embryo divided into two

slightly unequal spheres. The smaller of these two divided into two, each of which has divided into two, the result being four approximately equal spheres (S3., S., S., S3.). The larger primary spheres divided into two, one of which is still undivided (L.), the other having divided into two (marked L3.). Here again there seems to be a tendency for the descendant spheres of the smaller primary to undergo division first.

In the Fallopian tube of the left side of the same rabbit I found six embryos, of which some were in seven segments, others in eight. I had not time, unfortunately, to measure all these or draw them carefully. The embryos of this rabbit seem to be unusually far advanced for their age.

The stage with eight segments is a very common one to find between the 29th and 44th hours. This may be accounted for by there being a rather long resting stage after the production of the eight segments.

One rabbit (Rabbit No. 19) presented a very curious condition of its Fallopian tube, a condition, I believe, that has been noticed before, but I cannot remember by whom. The rabbit was a very large, healthy English doe. Both Fallopian tubes were almost filled with ova. On examining the Fallopian tubes with a lens before opening I imagined I had come across a marvellous find of embryos-in all about fourteen "ova" were to be seen shining through the wall of the left Fallopian tube. Of these only three turned out to be embryos, and were mingled with at least twelve apparently disintegrating unfertilised ova. The right Fallopian tube likewise contained a number of disintegrating ova, as well as four perfectly normal embryos.

Some of these unfertile ova were more thickly coated with albumen than is usual for normal eggs, possibly due to their having been a long time in the Fallopian tube. It is a matter of curiosity why these unfertilised or pathological ova had not passed down the Fallopian tube, but had allowed the fertile ova to pass them, as one at least had succeeded in passing the whole twelve bad ova. One can hardly believe that all these ova had left the ovary at the same time and together with

the ova which became fertilised and developed into normal embryos.

The corpora lutea are often very difficult to make out at this early stage, but I am pretty certain that there were certainly not more than five corpora lutea in one ovary and six in the other.

Fig. 9 is drawn from one of the fertile ova of this rabbit. I have reproduced it here to show the appearance of the denser inner and clearer outer layer of protoplasm of each sphere, described above as occurring in the earlier stages.

Possibly this appearance is constant in later stages, but as the cells get smaller the difference becomes less easy to distinguish.

I have avoided the use of the terms macrosphere and microsphere, because a morphological significance is sometimes given to these terms which I think is not advisable, at any rate in the case of the rabbit.

At the same time it is necessary, in discussing van Beneden's description and deduction, to pay especial attention to the question of the fate of the segments derived from the two first spheres.

It is hardly necessary to point out the very great difficulty in determining the fate of the descendant segments of the two primary segments when size is the only character on which we can rely, and when even in size there may be no difference. Again, if the ovum may sometimes divide into two segments of different size, may not the two primary segments also segment unevenly?

I can see no way of determining the question except by watching the division of one specimen, and this is, as far as I have tried, impossible. Even if it were accomplished, it must be under conditions which can hardly be called normal, and therefore not a very safe ground on which to base either facts or theory.

Certainly, in cases where the primary division is so unequal as in fig. 3, the subsequently formed segments could, on the whole, probably be classed as larger ones derived from the