56 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. 



or have even disappeared. As the constriction proceeds, the filaments 

 are forced inward in the form of a bundle, the density of which con- 

 tinually increases as the diameter becomes less and less. The free end 

 of this bundle has a tendency to retain its old position in the cytoplasm, 

 so that in the later stages some of the fibres are curved outward toward 

 the cell membrane, the ends being much farther apart than, the central 

 portion (Plate 5, Fig. 70). It can readily be understood that this 

 arrangement of the interzonal filaments obscures very much the astral 

 fibres, which are still active. The majority of these, at the time the 

 cleavage furrow has nearly reached the centre of the cell, extend in a 

 straight line from the bottom of this furrow to the centrosomes, but the 

 number of filaments is so great that this arrangement is considerably 

 obscured. The " Zwischenkorper " thus formed does not persist very 

 long, but soon disappears, probably as a result of the rotation of the cells 

 upon each other, which seems always to occur during the telophase. 

 In Figure 70, where the constriction is as yet hardly complete, this 

 rotation has apparently already begun. 



These conclusions concerning cell division, while by no means new, 

 contain several new features. Every point is based upon actual ob- 

 servations made upon different stages in the process of mitosis, and I 

 believe the conclusions drawn from these observed facts are correct. 

 As regards the source from which the power requisite for cell division 

 proceeds, it is necessary to conclude that it proceeds either from the 

 centrosome, from the archoplasm, or from the two working together 

 and in harmony. It is necessary to say that during these phenomena 

 the centrosome acts at least as the directive centre. During the 

 earlier stages of actual division (metaphase and anaphase) the separa- 

 tion of the chromatin masses and elongation of the cell are accom- 

 panied by the moving apart of the centrosomes. This divergence must 

 be accomplished, I believe, either by the activity of the centrosomes 

 themselves, or through some force which is exerted upon these organs 

 by the archoplasmic fibres. The only part of the archoplasm which 

 could accomplish this is the central spindle. To be sure, the cell mem- 

 brane and centrosome could be brought together by the contraction of 

 the astral rays in the polar region. But it is readily seen that this 

 would not have the observed effect. The contraction of these structures 

 would cause the membrane to approach the centrosome and the cell 

 would be elongated in the plane of the equatorial plate. Then, in addi- 

 tion, these polar rays are less developed than any others of the cell 

 The lengthening of the cell occurs at right angles to the equatorial plane, 



