The spermatogenesis of Peripatus (Peripatopsis) balfouri. 329 



valent chromosomes is directed towards one pole of the spindle, the 

 otlier univalent chromosome towards the opposite pole. 



Accordingly, in the completed monaster stage all the bivalent 

 chromosomes are so arranged, that their transverse constrictions (i. e. 

 the points where the central bands of linin join the central ends 

 of the univalent chromosomes) lie in the equatorial plane. In the 

 equatorial plate, on lateral views, the true outlines of the chromo- 

 somes can be best determined for those which lie nearest the peri- 

 pliery : there it can be seen that in the most frequent type of chromo- 

 some, the bent dumbbells, the convexity of the chromosome is usually 

 directed away from the central axis of the spindle. The mantle fibres 

 are very conspicuous by deep staining with iron haematoxyline or the 

 triple stain of Hermann, and the distal end of each univalent chromo- 

 some seems to be connected with one of the pairs of centrosomes by 

 two mantle fibres. It is, however, difficult to determine whether there 

 are two fibres separate for their entire length; it seems rather, that 

 tlie mantle fibre is unsplit at the point of attachment to the centro- 

 some, and it is certain that it is split at its point of attachment on 

 the chromosome. At each pole of the spindle there are two centro- 

 somes: what I could not satisfactorily determine by observation, was 

 whether from each centrosome there passes one fibre to one end of 

 the chromosome (i. e. whether the originally single distal linin fibre 

 splits into two each attached to a centrosome), or whether only one 

 of the centrosomes of each pair is connected by a mantle fibre with 

 a chromosome (and then this mantle fibre splits at its distal end). The 

 difficulty in determining this point lies in the fact, that the line 

 joining the two centrosomes of a pair is not perpendicular to, but 

 makes an angle with, the axis of the spindle. But from the relations 

 to be observed in later stages, the former alternative may be assumed 

 to be the correct one, namely that the originally single mantle fibre 

 has split through its entire length into two, and that the proximal 

 end of the one fibre is attached to the one, the end of the other 

 fibre to the other, centrosome of the pair. 



When the nuclear membrane disappears, the pole fibres are seen 

 to have diujinished greatly in volume and number (Figs. 175, 176, 

 179—182). This is in striking contrast to their great development 

 during the time of migration of the centrosome pairs (Figs. 162—169). 



No traces of nucleoli are to be seen at the completed monaster 

 stage. When the chromosomes are arranged in the equator of the 



