TRANSACTIONS OF SECTION D. 743 



Weismauii. The file of granules next becomes double, probably owing' (o their 

 bipartition, and the nuclear segments widen out, as a preparation for splitting 

 between the two lines of granules. The nuclear wall now disappears ; the nuclear 

 segments become grouped about the centre of the cell, constituting the 'equatorial 

 plate ' of Strasburger, the ' monastroid ' of Flemming, and the cytoplasm shows a 

 bipolar arrangement. Each segment splits into two secondary segments, one of 

 which goes to either pole of the cell, and the assemblage of these daughter seg- 

 ments receiving a wall constitutes the nucleus of the daughter cell. On their way 

 to the pole the daughter segments are usually /\-shaped wilh the apex in front, 

 and they frequently retain this shape for some time in the young nucleus, giving 

 it a polar or monaxlal structure. Kabl, who found traces of this an-angement 

 persist in the ' resting nucleus ' of the larval salamander, formed the generalisation 

 that the threads formed in an old nucleus on its 7»v/j>a>-«fi'ows for division are the 

 same as those which constituted it on its original formation ; and that where, this 

 arrangement is seeminf/ly lost in the 'resting nucleus^ definite paths exist along 

 which the errant granules return to their original stations to re-form identical 

 segments. This we may term Rabl's ' doctrine of the personality of the nuclear 

 segments.' On the assumption of the truth of this doctrine Weismann has based 

 a complex theory of heredity, fully and finally developed in his essay on Amphi- 

 mixis ; but despite its ingenuity it must fall if Rabls doctrine is not proved. But 

 Rabl's doctrine has never commanded the universal assent of histologists, and Ihe 

 author's researches have yielded additional evidence against it. He adduces the 

 following arguments : — 



1. The resting nucleus, in the vast majority of cases, shows no trace of polar 

 structure, but a network of protoplasm which is either impregnated witb 

 chromatin, or bears on its threads chromatin granules of irregular size ; and this 

 network, granules and all, is probably in constant motion. Such a nucleus, 

 characteristic of tissue cells and most brood-mother cells, such as ovarian ova, &c., 

 is sometimes found even between successive brcod divisions. Hence we infer that 

 Rabl's polar type of nucleus is characteristic not of the resting state but of the 

 absence of it, the preparations for a new division supervening before the nucleus 

 has as it were settled down from its constitution. 



2. In the pollen mother cells of Tradescantia spp., on their first division, the 

 segments undergo changes (now first described by the author) during the recon- 

 stitution of the nucleus, which give no support to Rabl's doctrine. 



o. In the resting nucleus of the ordinary type there is no indication of a path 

 by which the granules could return to their primitive position. 



4. The chromatin granules themselves should have a personality if Rabl's 

 doctrine be correct ; but in the ordinary resting nucleus they are subject to increase 

 and decrease of size, probably appearance and disappearance — phenomena which 

 make their personality doubtful. 



5. In certain cases of brood-cell formation, not all the segments of the equa- 

 torial plate are divided between two equivalent daughter nuclei. In Hemerocallis 

 fulva some segments pass laterally to form one or more lateral nuclei (Stras- 

 burger). In Tradescantia some segments may be retarded, and stay in the cyto- 

 plasm (as now shown by the author). On Rabl's doctrine the personality of the 

 pollen grains should be defective, with a corresponding result in the offspring; on 

 Weismann's theory these groups should evince increased variability, but neither 

 anomaly has been observed. 



6. The nucleus of the brood-mother cell of most gametes on its first division 

 presents a number of segments which is half of the number that constituted the 

 nucleus on its formation. As the mitosis is often in all other respects normal, 

 this process of ' reduction ' is irreconcilable with Rabl's doctrine, and Weismann's 

 explanations based on purely hypothetical differences from ordinary mitosis are 

 inadmissible. 



7. In some cases a cell on its mitosis presents a greater number of segments 

 than those which constituted its nucleus on formation. Thus in the embryo-sac 

 oi Lilium the first mitosis shows twelve nuclear segments; but the second mitosis 

 shows sixteen in the lower cell, while the upper repeats twelve segments. Such 



