The spermatogenesis of Peripatus (Peripatopsis) balfouri. 327 



the centrosomes, could not be determined. These converging linin 

 fil)rcs are the first appearance of the mantle fibres of the spindle, 

 which converge towards and then connect with the centrosomes. A 

 careful study was made of a number of cells in this stage (which is 

 not at all abundant), and the result of this examination was that no 

 evidence was obtained to show that mantle fibres grow from the 

 centrosomes to the chromosomes, while all the appearances suggest 

 that the mantle fibres are formed by the distal linin fibres of the 

 iiuchuis first directing themselves towards, then attaching themselves 

 to, the centrosomes. My observations show that these distal linin 

 fibres have no orderly arrangement in the nucleus, until the nuclear 

 membrane begins to disappear, when they all direct themselves towards 

 the centrosomes. 



This observation stands in full accord with our present conception 

 of the rôle of centrosomes and chromosomes; the centrosomes form 

 attractive centres which attract to themselves (by a force yet un- 

 explained) the various fibrous structures of the cell; whereas the 

 chromosomes show no phenomena suggestive of an attraction for other 

 structures. And after making this observation I was agreeably sur- 

 prised to find that it suffices to fully explain why in Peripatus^ and 

 why also in probably most other objects, the distal ends of the mantle 

 fibres are attached to certain definite points on the surface of the 

 chromosomes. When all the chromosomes lie in the equatorial plate 

 it can be clearly determined that all the chromosomes are so placed, 

 that the distal ends of the mantle fibres are connected to the distal 

 ends of the univalent halves of the bivalent chromosomes (Fig. 177, 

 179—182, 184—186). There is apparently absolute regularity in this 

 arrangement in all the cases studied by me. Now if for these cells 

 in Perij>ahis the origin of the mantle fibres had not been determined, 

 yet their regular attachment to the distal ends of the chromosomes 

 would render it veiy likely that these fibres had not grown out from 

 the centrosomes to the chromosomes; for on the assumption of such 

 a mode of formation, how could it be explained that the distal ends 

 of such outgrowing fibres attach themselves with the greatest regularity 

 to the morphological distal ends of the chromosomes, when the chromo- 

 somes vary much in form, and when the chiomosomes exert apparently 

 no attractive influence upon other structures? On the other hand, all 

 the subsequent observations here detailed will be found to corroboiate 

 the observation, that the mantle fibres grow out, under the attractive 

 agency of the centrosomes, from definite points on the chromosomes 



