366 bulletin: museum of comparative zoology. 



not appear until much later, and from the beginning are usually 

 " erythrophile." 



In general, the character of the nucleoli in Gonionemus supports the 

 foregoing generalizations, but the chief nucleolus in this animal shows 

 one marked deviation from the usual condition, in that it retains through- 

 out its history the affinity for basic dyes characteristic of its youngest 

 stages. The chemical evidence, then, suggests that this structure retains 

 its primitive composition of chromatin. But there could hardly be 

 stronger morphological evidence that this is not the case than is afforded 

 in its history. To begin with, we have seen that this nucleolus corre- 

 sponds exactly with the plasmosome of somatic cells in origin, since it is 

 merely the persistent central portion of a nucleolus of which the periph- 

 eral shell has taken part in the formation of the chromatin net of the 

 pseudoprophase. Furthermore, it does not contribute in any way to the 

 chromatin net of the germinative vesicle, but, on the contrary, appears 

 to be wholly inert. Nor does there appear to be any good evidence that 

 the deeply staining granules and fibres which occasionally occur in its 

 ground substance are stored chromatin, as suggested by Carnoy ('85), 

 Meunier ('86), Moll ('93), and Hartmaun (:02) ; but it is more likely 

 that these appearances in the vacuolated nucleolus are in reality caused by 

 branching channels in the vacuolar substance, as Montgomery ('98 b ) be- 

 lieves to be the case in Polydora. If this view be correct, the staining 

 reactions of the chief nucleolus of Gonionemus are most readily explained 

 on the assumption that by-products of chromatin may sometimes be 

 acid instead of basic. 



Those authors who have traced the ultimate fate of the chief nucleo- 

 lus in coelenterates have found that it either disintegrates (Brauer, 

 91 a , '91 b ; Maas, '99 ; Morgenstein, : 01 ; Harm, : 02 ; Wulfert, : 02), or 

 is cast out into the cytoplasm. All are agreed that it does not contribute 

 in any way to the formation of the chromosomes. A diametrically op- 

 posite view, namely, that it does so contribute, is maintained by Hart- 

 mann (: 02) and Guenther (: 03 b ) for certain echinoderms. But their 

 results have been criticized by Dublin (: 05), who has shown from his 

 own studies that observations based, as were theirs, on haematoxylin 

 staining are of little value in this question. In Pedicellina he found 

 stages where this method seemed to show that chromatin strands were 

 being detached from the chief nucleolus, just as described by Guenther 

 (: 03 a ). However, Auerbach preparations proved that this was not the 

 case, for, in Dublin's own words, " The nucleolus, at this point completely 

 vacuolated, stains intensely red ; and in most striking contrast, all the 



