The Spermatogenesis in Pentatoma up to the Formation of the Spermatid. 47 



except by assuming that here the central spindle had become narrower. 

 The five cases in regard to which there can be no doubt (Figs. 101 

 —103, 119, 121), however, show that when the centrosomes first 

 wander apart a central spindle connects them. There are no pole 

 radiations at this stage. I have been unable to determine with surety 

 centrosomes and central spindles from this stage, until the centrosomes 

 have reached opposites sides of the nucleus (Figs. 150 — 159). The 

 nuclear membrane is still intact, but thinner than previously. Around 

 each ceutrosome, when they have attained this position, is a well 

 marked pole radiation, the rays of which insert themselves on the 

 centrosomes, and gradually extend outwards until they reach the cell 

 membrane (Figs. 150, 153 — 159). At first the pole rays appear micro- 

 somal, but subsequently become homogeneous, though this difference 

 has not been shown in my figures; this agrees with the observations 

 of Drüner ('94) on cells of Salamandra. Those pole rays are more 

 numerous and more pronounced which pass towards the equator of 

 the cell, than those which pass to the cell membrane (a fact noted 

 also by Drüner, Meves and others), and they never produce any 

 pulling in of the cell membrane above the centrosomes (such as has 

 been described by Van Beneden, '83, for Ascaris). Further, it may 

 be noted that the achromatic spindle does not occupy any particular 

 position with regard to the line of the greatest amount of cytoplasm. 

 The early stages of the central spindle have been described and 

 figured {c.s.p.l, Figs. 101—103, 119, 121, and possibly 137, 138). 

 But when the two centrosomes are situated at opposite poles of the 

 nucleus (Figs. 150 — 159) the nuclear membrane is still present, and 

 no fibres connecting the centrosomes are to be seen in the nuclear 

 cavity. What then has become of the central spindle? Unfortunately 

 I was unable to find any centrosomes in the stage immediately preceding 

 the latter period, and so cannot answer this question positively. Two. possi- 

 bilities suggest themselves: either the central spindle has disappeared, 

 or else it still persists, and in the latter case is so closely apposed 

 to the outer surface of the nuclear membrane that it cannot be 

 distinguished. In the latter case, the nucleus being round or ovoid 

 and the centrosomes at opposite poles of it, the central spindle must 

 be curved around the surface of the nucleus. I give a figure (c.s.p. 2?, 

 Fig. 150) of a case showing two centrosomes lying on the surface of 

 a Qucleus, in which, by careful focussing of the microscope, two fibres 

 could be made out which seemed to connect the centrosomes. This 

 was the only case where I could distinguish such apparent connecting 



