COMPLETE AND INCOMPLETE CLEAVAGE. 105 



(Marshall, No. 37; SoUas, No. 51), ia A.lcyonarians (Kowa- 

 levsky and Marioa, No. 33) and most Arthropoda that we 

 shall have to seek the nearest erabryological counterpart of the 

 pi'ocess by which the Metazoa arose from the Protozoa. 

 If this is so, how are we to account for the frequency of the 

 cases in which the furrows, dividing the ovum, completely 

 separate the segments from each other ? In the first place, I 

 would ask, are the cases so numerous as is supposed? It seems 

 to me extremely probable that it will be found, on renewed 

 investigation, that incomplete cleavage takes place in many 

 forms in which it has been assumed that complete cleavage is 

 the rule. The complete cleavage of small ova is such a striking 

 phenomenon, and so readily lends itself to speculative sugges- 

 tions, and has in this form so dominated the views of mor- 

 phologists {vide especially Flemming's remarks above referred 

 to), that I cannot help feeling that it may, in some cases in 

 which perhaps the observation was difficult, have been assumed 

 to occur on insufficient evidence. And this feeling is rather 

 confirmed by the well-known prevalence of the habit of as- 

 suming cell boundaries when they cannot be seen. Almost 

 every embryological memoir bears on its plates numerous 

 examples of this habit. 



In the second place, it seems possible that the complete 

 cleavage, found so conspicuously in small ova, may be sus- 

 ceptible of a mechanical explanation. The clean rounded form 

 of the spheres at the moment of division is unlike anything 

 else in the animal kingdom, and is suggestive rather of an 

 intensely active force in the centre of the cell, which compels 

 for the moment the assumption of this form in the protoplasm 

 over which it has dominion, than of a tendency inherited from 

 an adult ancestor. I would refer in this connection to Brook's 

 observations on the total segmentation of Lucifer (No. 22). 

 He describes how, at the moment of activity, the segments 

 round themselves oflF, touching only at one point, while in the 

 intervals of rest they flatten out against one another, and 

 possibly become partly fused. The same phenomenon is found 

 in other Crustacea {vide Balfour, No. 17, p. 112), and it 



