CYTOKINESIS. 109 



The immediate cause of these telokinetic movement, as also of those in the 

 unsegmented egg, may be found at least in part in the movement of the spheres to 

 a free surface of the cell, but the orientation of these movements cannot at present 

 be further explained than to refer them to the structure of the protoplasm.^ 



2. Dzfferenlial Cell Division. — As has been emphasized alread}- (p. 6), 

 cell division is typical and non-difterential when it occurrs at regular intervals 

 or at the same time in cells of the same generation (rhythmical), when successive 

 divisions are at right angles (alternating), when the daughter cells are of similar size 

 (equal) and are composed of similar materials (homogeneous). Divisions are differ- 

 ential when they depart from these typical conditions in one or more respects, thus 

 becoming non-rhythmical, non-alternating, unequal or heterogeneous. 



{a) Rhythm of Division. — A definite order of cleavage is highly character- 

 istic of gasteropods. In Crepidula one of the first two blastomeres usually divides 

 slightly earlier than the other, and in the formation of the first, second, third and 

 fourth quartettes one cell of each quadrant forms earlier than the other three. 

 These differences in the time of formation of the different cells of a quartette are least 

 in the first quartette and greatest in the fourth, where the cell 4d is formed at the 

 25-cell stage while the other cells of this quartette are not formed until the 52-celI 

 stage. In the subdivision of the different quartettes this same lack of rhythm is 

 obsei'ved, the cells which formed first being usually the earliest to divide. To this 

 rule, however, there are several notable exceptions ; for example, the trochoblasts 

 which are formed at the first division of the first quartette do not again divide until 

 more than one hundred cells are present. In this case the lack of rhythm in the 

 divisions leads to important differentiations, since the large trochoblasts give rise to 

 certain of the large cells of the velum and head vesicle. 



The old view (Balfour '80) that the rate of division depends upon the presence 

 or absence of yolk, cells with yolk always lagging behind those without it, is unten- 

 nable since this lack of rh3'thm frequently concerns purely protoplasmic cells. For 

 example, none of the ectomeres of Crepidula contain j^olk, yet they divide at very 

 different rates, while on the other haiid many yolk-laden cells divide more frequently 

 than those without yolk (see Wilson, 1900, p. 366). 



The rhythms of division of centrosome, nucleus and cell body go on more or 

 less independently of one another. Boveri ('97) has shown that in echinoderm eggs 

 the centrosomes preserve their rhythm of division even when separated from their 

 nuclei, and I have observed the same thing in enucleated blastomeres of Crepidula. 

 Furthermore, the rhythm of cell and nuclear division are more or less independent 

 of each other ; in certain abnormal eggs of Crepidula I have observed that normal 

 and characteristic cell division may occur in enucleated blastomeres, and on the 

 other hand nuclear division may go on in regular manner and at regular intervals 

 in the absence of cell division. There is therefore no absolutely necessary connec- 

 tion between the division of nucleus, centrosome and cell body. 



Driesch ('98) has shown that in cross-fertilized eggs of echinids the rhythm of 



' See similar conclusions readied b}' Lillie (1901) in the case of Vnio. 



