26 BULLETIN: MUSEUM OF COMPAEATIVE ZOOLOGY. 



distance from that structure, and having described an arc of 180°, take 

 up their position in the cytoplasm between the nucleus and the cell 

 membrane. Other investigators upon chilopods — Meves und von Korff 

 (:Ol) for Lithobius forficatus ; P. Bouin (:Ol) for Lithobius forficatus ; 

 and P. Bouin et R. Collin (:Ol) for Geophilus linearis — have arrived 

 at similar results. That this never occurs in the first spermatocyte 

 division of Scolopendra heros, I can state with certainty after having 

 observed several hundred cells in the prophase and in other stages of 

 mitosis. There can be no doubt, however, that such phenomena do 

 occur in other genera of Chilopoda (Lithobius, Scutigera, Geophilus), 

 where I have myself observed them. In the second spermatocyte of 

 Scolopendra there is present a condition similar in some respects to 

 that described by Carnoy, and it is possible that these cells have been 

 mistaken by him for first spermatocytes. 



As the centrosomes continue to diverge upon the nuclear membrane, 

 the astral rays surrounding them become more and more marked. 

 Whereas at first these were of a granular nature, now they become 

 more clear cut and approach more nearly the fibrillar structure char- 

 acteristic of them during the succeeding metaphase. Before the cen- 

 trosomes have separated very far, each one becomes elongated in a 

 plane tangent to the nuclear membrane and exhibits signs of a second 

 division (Figs. 21, 22). This division, however, is not fully accom- 

 plished until the following metaphase. Often during this period of the 

 prophase the nuclear membrane is distorted in the vicinity of the cen- 

 trosomes, giving it the appearance of undergoing amoeboid movement. 

 In some cases this seems to be due to an attractive influence exerted by 

 the centrosomes, as is shown by the existence of protuberances (Fig. 22), 

 but in other cases the membrane is affected in exactly the opposite way 

 (Fig. 21), which suggests a repellent force on the part of the centrosome. 

 This distortion of the nuclear vesicle is most marked at the stage im- 

 mediately preceding the disintegration of its membrane (Figs. 29, 30). 

 At this time, when the two compound centrosomes are at opposite poles 

 •of the nucleus, the diameter of the vesicle connecting the two cen- 

 trosomes is considerably shorter than those at right angles to it (Figs. 

 29, 30). 



At this stage the centrosomes are never immediately in contact with 

 the membrane, but are at a distance from it equal to perhaps half of 

 their own diameter. That part of the nuclear membrane which is near 

 the centrosome is invariably bent inward, forming a cup-like depression, 

 in which the double centrosome lies. Under high magnification it is 



