CYTOKINESIS. Ill 



planes (Sachs), or to the principle of smallest surfaces (Plateau, Berthold). Almost 

 all investigators agree that the direction of division is due to the position of the 

 mitotic figure and the orientation of the figure is usually supposed to be actively pro- 

 duced by the figure itself, or by some part of it. According to Roux ('95) there is 

 immanent in the nucleus, a direction of division which may be independent of the 

 chief dimensions of the protoplasmic body. Rauber ('83) holds that the position of 

 thespindle is the result of the mutual attractions of neighboring asters. Heiden- 

 hain (94) refers the direction of division to a definite angle of rotation of the 

 centrosomes. 



In the segmenting eggs of the gasteropods which I have studied the direction 

 of division is not primarily due to any of the factors named, though several of these 

 principles are well illustrated in these cells. The spindle usually, though not in- 

 variably, lies in the longest axis of the protoplasmic mass and it probably always 

 lies in the direction of least resistance, though this in itself is no explanation of the 

 direction of division. As has already been pointed out (p. 105), the angle of diver- 

 gence of the centrosomes and the initial position of the mitotic figure may not cor- 

 respond with its final position, while at the same time the lobing of the cytoplasm 

 may indicate the final position of the spindle and the direction of the coming divi- 

 sion ; in fact the form of the cell and the movements of the cell contents may pro- 

 claim where the next division will occur long before the spindle is formed. 



The alternation in the direction of successive cleavages is not due to the mere 

 divergence of the centrosomes in planes successively at right angles to one another, 

 but rather to regular alternations in the rotations of the cell contents; the lack of 

 alternation is associated (at least in one generation of cells, see p. 88) with the lack 

 of rotation of the cell contents during the preceding telophase. 



When for any reason the mitotic figure is prevented from assuming its normal 

 position, the cytoplasm may divide in the normal place and manner, thus giving rise 

 to a cell which is normal in appearance except that it contains no part of the 

 nucleus or spindle. 1 From these facts I conclude that the position of the spindle is 

 the result of movements and stresses in the cytoplasm. In normal cell division the 

 spindle takes a position of equilibrium between the two portions of the dividing- 

 cell, so that the equatorial constriction cuts through the middle of the spindle ; if, 

 however, the spindle is prevented from assuming this position of equilibrium it may 

 be cut through nearer one end than the other or may be left entirely to one side of 

 the new cell wall. Therefore the position of the spindle and the direction of divi- 

 sion are functions of the cytoplasm, rather than of the nucleus, centrosome or 

 spindle. 



(c) Size of Daughter Cells. — The inequality of cell division leads to some of 

 the most characteristic and important features of differential cleavage, while the 

 varying sizes of blastomeres have a definite prospective significance in development, 

 as Lillie ('95, '99) and Conklin ('97, '98) have pointed out. 



1 I have seen several such cases, but a more detailed account of them must be postponed to 

 another paper. 



