5 GO TRANSACTIONS AND PROCEEDINGS OF THE 



which in the older parts bore the numerous " cystocarps " 

 were always characterised by the presence of numerous 

 sporangia with tetraspores, arranged in pairs in the outer 

 cortex of the younger branches, while in E. cervicornis 

 sporangia and cystocarps were always borne on different 

 individuals. A detailed examination of the development 

 brought the cause of this difference to light. 



I first found that in E. cervicornis the cystocarps 

 always proceeded from well-formed procarps. In the 

 younger branches of the thallus these procarps are in great 

 numbers on the margin between the closely coherent 

 tissue of the wall and the interwoven rhizoids filling the 

 internal cavity. They grow out towards the exterior of the 

 thallus from the lower cells of the wall-forming filaments, 

 and are represented when mature by small short umbel-like 

 fascicles of filaments, the lateral branchlets of which fre- 

 quently develop a terminal carpogonium, while one of the 

 upper cells of the central main branch is converted into an 

 auxiliary cell rich in protoplasm.^ These brush-like pro- 

 carps thus appear sometimes on the inner side of the 

 thallus wall, surrounded by its interstitial mucilage, while 

 in other cases they lie in the interior of the internal cavity 

 of the shoot closely pressed against the inner surface of 

 the wall. 



From each fertilised procarp there arises a single cysto- 

 carp. After fertilisation of one of the numerous carpogones, 

 the auxiliary cell becomes fertilised by a second act of 

 fecundation, it then swells up and the gonimoblast proceeds 

 from it towards the outside of the thallus in the form of 

 a small compact tuft of closely coherent filaments. This 

 gonimoblast during its growth presses the superincumbent 

 wall slightly outwards, but is itself at the same time forced 

 by its enlargement increasingly further into the inner 

 cavity of the shoot, w^ithin which it appears embedded 

 when mature. In consequence of the extremely close 

 union of its congenitally coherent branched filaments it 



^ Thus several (or even numerous) carpogone filaments, with terminal 

 carpogones, which ripen in succession, and a single auxiliary cell 

 filament have always a share in tlie formation of each procarp. It 

 nevertheless appeared to me in isolated cases as if in speciallv thick 

 copiously branched procarp fascicles, there were occasionally formed 

 two or three auxiliary cells on distinct branchlets of the fascicle. 



