of the Fishery Board for Scotland. 



37 



Kupffer concluded that in the pike the second furrow is an equatorial 

 one, and inferred that the same is the case in the herring. Hoffmann, 

 on the other hand, has described the first furrow as equatorial in direction, 

 dividing the germinal protoplasm at the outset into two layers — the 

 archiblast and parablast. 



With these exceptions, an equatorial furrow has not hitherto been 

 described at so early a stage in the segmentation of fish ova, though it is 

 possible, from the nature of the furrows, that it has been overlooked. 

 The point, if of general application, has considerable phylogenetic in- 

 terest. An analogy can thus be instituted between the segmentation of 

 meroblastic ova such as those of fishes, and the holoblastic ova of Amphibia 

 and allied forms. In both cases the first equatorial division of the ovum 

 takes place with the formation of the third furrow, and the ovum then 

 becomes divided into an animal and a vegetative pole. The animal pole 

 is represented in the herring by the archiblast, which goes on segment- 

 ing. The vegetative pole consists of only one large cell containing the 

 yolk in the centre, which is surrounded by a thin layer of protoplasm — 

 the parablast. I have endeavoured to explain the cause of this unicel- 

 lular vegetative pole in a paper which was read before the Eoyal Physical 

 Society of Edinburgh in January last, so that the subject need not be 

 discussed further here. 



The archiblast goes on segmenting in the usual way, while the para- 

 blast does not for the time being undergo division, but simply increases 

 in bulk by a further assimilation of yolk. When at first formed the 

 parablast was thickest under the archiblast, but gradually the protoplasm 

 sinks from that area, and collects in a comparatively thick layer around 

 the yolk. 



The segmentation in the archiblast takes place to a great extent by a 

 process of vacuolation, and the straight planes of cleavage usually found 

 in karyokenetic cell division seldom occur in the earlier stages. In the 

 future plane of cleavage a number of vacuoles arise, which increase in 

 size and run together, forming larger vacuoles, and in this manner the 

 two adjacent masses of protoplasm become ultimately separated. Before 

 the process is complete the two resulting daughter cells are connected 

 together by bridge-like strands of protoplasm, the substance of which is 

 gradually withdrawn into the daughter cells. I have described the 

 process in detail in a paper ' On the Nature of the Segmentation Process 

 in Fish Ova' {Proc. Roy. Phys. Sac. Session 1885-86). 



The primary segmentation phase in the archiblast may be said to ter- 

 minate at a stage corresponding to that shown in fig. 8, An egg at this 

 stage measured 1*4288 mm. in the direction of its long axis, and 1*1656 

 mm, in the axis at right angles to it. The disc in the same egg was 

 P0904 mm. in diameter, and had a greatest thickness (as seen in optical 

 section) of -3948 mm. The disc extends over 110 degrees of the circum- 

 ference of the yolk. 



The cells which form the morula mass in this stage have hitherto been 

 considered as derivatives of the archiblast only. Such, however, is not 

 the case so far as the herring is concerned, and I have reason to suppose 

 that the herring is not alone in this respect. In a case of this sort a 

 study of the living egg is of little value, and it is absolutely necessary 

 to study a series of consecutive sections in order to understand what takes 

 place. In the interval between the stages shown in figures 7 and 8 the 

 parablast has increased considerably in bulk, and a portion of it is seen 

 in sections to form a comparatively thick layer between the archiblast 

 and the yolk. Nuclei appear in this portion at comparatively regular 

 intervals by the process which has been termed ' free cell formation.' I 



