724 Proceedings of the Royal Irish Academy. 



number of tlie diromosomes of tlie nuclear plate of cases of normal 

 karyokinesis in tlie endosperm very probably often arises in these 

 multipolar divisions ; for it is easily seen that in this way nuclei con- 

 taining abnormally large and abnormally small numbers of chromosomes 

 would be formed. Doubtless nuclear fusion is also often responsible 

 for the nuclei with large numbers, as Strasburger has pointed out. A 

 curious case of this latter process is figured in fig. 1 4, where two of 

 the daughter nuclei formed by the fission of two neighbouring nuclei 

 have united together immediately after division. 



In the multipolar nuclear divisions the same transition forms 

 between direct and indirect nuclear division have been observed 

 (fig. 12). _ 



It not infrequently happens that apparently without any formation 

 of a thii'd pole a few chromosomes fail to attach themselves to either of 

 the daughter nuclei of a division, but take up a position at the edge 

 of the rudiment of the cell-plate, and towards them the fibres of the 

 achromatic spindle appear drawn (fig. 15). 



The portions of the nuclear thread which in the abnormal divisions 

 protrude from the daughter nuclei as well as the fragments of the 

 thread which remain external to them after the completion of division 

 become vacuolated so as to remind one of Yan Beneden's and Neyt's 

 figures of the reconstruction of the daughter nuclei in Ascaris megalo- 

 cepliala by the vacuolation of the chromosomes of the diaster (jN'ouvelles 

 recherches sur la fecondation et la division mitosique chez I'ascaride 

 megalocephale. Leipzig, 1887). 



The ultimate fate of the daughter nuclei formed by direct division 

 I have not been able to ascertain. ^ It seems probable, since these 

 divisions are found in such early stages of the growth of the endosperm, 

 e.g. before the formation of the cell-partitions, that these nuclei con- 

 tinue to divide either by direct division or by karyokinesis. It is, 

 however, quite possible that neither they nor their descendants ever 

 divide by karyokinesis unless they imite with other nuclei. 



In conclusion it may be noticed that the abnormal divisions 



1 Ziegler (Archiv f. Mikrosk. ATiatomie, 30), coneliules that direct nuclear 

 fragmentation only occurs wlien nuclei take on a specialized function and then 

 disintegrate. Again the same author (Biol. Centralb. 11) notices that nuclei 

 wldeh divide by amitosis are exceptionally large ; that dii'ect fission occurs chieflj' 

 (perhaps exclusively) in the nuclei of cells in which the secretory and assimilatory 

 processes are very intense. 



Von Eath (Zool. Anzeiger, 1891) believes it probable that nuclei -which 

 have once divided directly uever divide by mitosis ; that dii-ect division cannot be 

 repeated indefinitely ; that regenerative growth never takes place by amitosis. 



