82 



situated directly on tlie basal disc or at the end of short unbranched erect filaments 

 (fig. 42). An equally extreme reduction occurs in Ch. reducta, the frond of which 

 consists of creeping filaments bearing sessile or short-stalked sporangia (fig. 49). 



In certain species the thallus is partly endophytic. In Ch. cyiophaga the 

 development begins as in Ch. polyblasta, and it is only when the plant has become 

 multicellular that some of the cells in the creeping filaments produce short fila- 

 ments from their underside, which penetrate into the cells of the host plant (Por- 

 phyra umbilicalis) , pushing aside the protoplasm and taking without doubt nutri- 

 ment from it. This plant is thus a true parasite. The intracellular filaments or 

 haustoria do not seem to penetrate from one cell into another but they may make 

 their way again to the surface of the host plant (figs. 50, 51). In Ch. Dumontice the 

 development begins in the same way but the endophytic filaments are intercellular 

 and become much longer (fig. 52). These intercellular filaments are still more de- 

 veloped in Ch. Nemalionis, where they form a widely extended system of branched 

 threads, giving off free filaments at many points Ihrough the surface of the host 

 (Nemalion), while creeping epiphytic filaments are wanting (figs. 53, 54). The ger- 

 mination has not been observed in this species. Finally, there is a group of spe- 

 cies the filaments of which are entirely endophytic. Ch. endozoica Darb. forms a 

 transition to this group, the (endozoic) filaments sending out through the surface 

 of the Alcyonidium it inhabits numerous short slightly branched sporangia-bearing 

 filaments. In Ch. emergens only the solitary short-stalked sporangia are free (fig. 55), 

 and in Ch. imniersa and Ch. Polyidis the solitary sporangia are even more or less 

 sunk in the host plant (figs. 56, 58, 60) ^ 



Most of the Chantransia are usually epiphytic and then not bound to parti- 

 cular host plants ; several species also occur on Hydroids and Bryozoa , further on 

 Mollusc-shells, Ch. efflorescens even on stones. Probably other species may also 

 sometimes grow on stones but have not been detected there on account of their 

 small size. On the other hand, the endophytic species appear to occur only in 

 one particular species of Algse, or several nearly related. Thus, Ch. Dumontice has 

 been found growing only in Dumontia filiformis, Ch. cytophaga only in Porphyra 

 umbihcalis , Ch. corymbifera only in Helminthocladia , Ch. Nemalionis in Nemalion lu- 

 bricum and multifidum, Ch. immersa in Polysiphonia nigrescens and violacea and 

 in Rhodomela subfusca. The endozoic Ch. endozoica occurs only in Alcyonidium 

 gelatinosum. 



The form of the chromatophore is of great systematic value as pointed out 

 by Kylin (1906, p. 122). In the vast majority of Danish species the cells contain 

 only one chromatophore, but these may again be divided into two groups. In a 

 fairly large number of species the chromatophore has a central body lying in the 



> Rhodochorton Biebneri Batters (Journ. of Botany 1897 p. 437 and 1900 Tab. 414 fig. 17), which 

 is endophj'tic in Gloiosiphonia capillaris, is evident!}' a Chantransia belonging to this group, to judge 

 from tlie mode of growth, the hairs and the chromatophore; its name must therefore be Chantransia 

 Brebneri (Batt.) Rosenv. Tiie genus Colaconema Batters (see page 71 note) seems also to comprise spe- 

 cies referable to the group of the endopliytic Chantransice. 



