280 



mentioned on p. 210, this genus is characterised by the fact that the cells in the 

 upright filaments, of which the frond (the perithallium) is composed, are connected 

 by transverse pits, the origin of which must be of a secondary nature. I have not, 

 however, been able to follow their development, and particularly did not succeed 

 in ascertaining the co-operation of the nuclei in their formation. In the remaining 

 members of the family of Corallinacese, on the other hand, there is a different method 

 by which the cells in various filaments may enter into direct communication one 

 with another, to wit, by dissolution of the separating wall, whereby an open con- 

 nection is established between the cells. This feature has already been referred to 

 above (p. 210) where it was also pointed out that more than two cells may fuse 

 together, and that the cell-fusions may involve fusion of the nuclei (cf. figs. 136, 139, 

 156 and many others). Only in two of the Danish Corallinacese have the fusions 

 hitherto not been shown {Melobesia minutula and Choreonema Thuretii). 



Entirely similar cell-fusions were demonstrated in various Squamariacese, viz, 

 Cmoriopsis danica (p. 185 fig. 107), Cruoriopsis gracilis (p. 188 fig. Ill), Rhododermis 

 elegans (p. 198 fig. 118) and Rhododermis Georgii (p. 199 fig. 119). In Hildenbrandia, 

 on the other hand, they were not found. 



That cell-fusions are important as facilitating connection between cells and 

 cell-filaments not directly in communication by plasma-continuity can hardly be 

 doubted. We find them also particularly numerous in the "roof" above the con- 

 ceptacle 'in the Corallinaceaj, i. e. between cells whose indirect connection below 

 has been interrupted by the formation of the conceptacle. Comparison with Hilden- 

 brandia, which lacks cell fusions, supports this view, as the roof of a conceptacle, 

 which grows in extent through the continued sporangia formation, consists of dead 

 and more or less disorganised cells, save at the margin, undoubtedly owing to the 

 fact that the connections below have been interrupted, and those to the sides are 

 wanting (cf. p. 204 and figs. 125, 126.). 



3. Alternation of generations and alternation of nuclear phases. As we know, 

 there has in several Floridese been shown to exist a regular alternation between a 

 haplophase, consisting of the sexual generation, and a diplophase, consisting of two 

 generations, viz; the cystocarp or gonimoblast, and the tetraspore-bearing plant ^). 

 A like course of development must be presumed to take place in all Floridese with 

 normal fertilisation, and having tetrasporangia. Svedelius has called these Floridese 

 diplobiontic, in contrast to the haplobiontic, which lack tetraspores, and in which 

 the chromosome reduction takes place by division of the zygote nucleus^). Here 

 then, we have but two generations, the sexed plant and the cystocarpium, both 



Comp. H. Kylin, Die Entwick. u. syst. Stell. von Bonnemaisonia asparagoides etc. Zeitschr. f. 

 Botanik, 8. Jahrg., 1916, p. 570. — J. Buder, Zur Frage des Generationsvvechsels im Pflanzenreiche. Ber. 

 deut. bot. Ges. Bd. 34. 1916, Heft 8. — O. Renner, Zur Terminologie des pflanzlichen Generationswech- 

 sels. Biolog. Ceiitralblatt. Bd. 36, 1916, p. 337. 



'-) N. Svedelius, Zytolog.-entwickelungsgesch. Stud, iiber Scinaia furcellata. N. Acta reg. soc. sc. 

 Upsal. Ser. IV. vol. 4 no. 4. Upsala 1915, p. 42. 



