blackman: the spermatogenesis of scolopendra. 109 



in the large spermatocytes, but is gradually broken down, as the astral 

 rays become more and more marked. Tlie first change in this transfor- 

 mation is noticed in the region immediately surrounding the centrosome. 

 Here the arehoplasm becomes less granular, more transparent, and astral 

 rays, faint and short, appear. As the number and length of these in- 

 creases, more and more arehoplasm is dissolved and beeomes transformed, 

 until, at the opening of the metaphase, none remains in the granular con- 

 dition (Plate 5, Figs. 57-51); Plate 9, Figs. 154, 155). 



The phenomena in the early stages of the second spermatocyte are 

 similar to those occurring in the first spermatocyte of the large type. In 

 another connection is mentioned evidence furnished by the spermatids, 

 showing that in these also the arehoplasm undergoes similar changes. 

 Here the origin of the axial filament is seen to be similar to that of the 

 astral rays at other stages. Both arise in connection with the centrosome, 

 and both are of arohoplasmic origin. 



I think there should be no hesitation in concluding that in Scolopen- 

 dra the arehoplasm is a distinct substance, and that it is derived, in part 

 at least, from the arehoplasm of the parent cell. This is undoubtedly 

 true of all generations of the cell after the spermatogonia, for the pres- 

 ence of this modified cytoplasm can be detected at all times. However, 

 it certainly undergoes very marked changes in its morphological, and 

 perhaps in its chemical, nature ; first appearing in the granular form, 

 then in the diffuse or dissolved condition, and finally in the fibrillar. 



Thus, the conditions in Scolopendra would seem to realize Boveri's 

 contention that the arehoplasm is a modification of the cytoplasm and 

 that the astral rays are different from the cytoplasmic reticulum. In 

 all stages of mitosis the cytoplasmic reticulum in Scolopendra remains 

 entirely distinct from the astral radiations, except in the region occupied 

 by the spindle, where no reticulum occurs, although the structure of the 

 reticulum is often obscured by the astral rays. 



This view of the persistence of the arehoplasm as opposed to the older 

 hypothesis (Biltschli, '76) of fibrillar persistence is supported by the ob- 

 servations of many of the later investigators. Griffin ('99) and Coe 

 ('99) find that the new asters as they arise around the divided centro- 

 some are entirely independent of the old ones, which still converge 

 toward the old position, even though the centrosome and new aster have 

 moved away. Similar results have been obtained by MacFarland ('97). 

 Lillie ('98), Smallwood (:04), and others. The fact that the astral rays 

 pei-sist in the absence of the centrosome also shows, I think, that they 

 are definite structures and not due merely to diffusion currents. As the 



