388 COMPARATIVE MORPHOLOGY OF FUNGI 



number of terminal cells which cut off basal segments successively. Later 

 they usually divide once, resulting in the association of antheridial cells in 

 pairs (Fig. 261, 9 to 10). The opposite series, arching somewhat, become 

 separated by a space which forms the central chamber into which the 

 spermatia are discharged. In T. africanus the trichogyne develops at 

 the same time with the antheridial branchlets and is ordinarily in close 

 contact with them, perhaps in actual contact with the cavity containing 

 the spermatia (Fig. 261, 2). Fertilization might thus occur directly 

 from this cavity, making an actual discharge unessential. This possibil- 

 ity is further suggested by the fact that the exposed portions of the 

 trichogynes are relatively thick-walled, without the usual thin-walled 

 receptive region. 



Where no individualized antheridium is produced, the appendage 

 becomes more or less indeterminate in growth, producing a variable 

 number of sterile branches, as in T. acuminatus (Fig. 256, 6). In T. 

 chaetophilus, (Fig. 261, 7), which produces no visible antheridial cells, 

 there has been seen a trichogyne, sometimes found between the 

 perithecium and the base of the appendage. 



Since there is such complete disagreement in the interpretation of 

 many phenomena in this group between Gaumann and Thaxter, in the 

 following discussion, I have attempted to present accurately Gaumann's 

 views and Thaxter's unpublished criticism of them. 



In the Laboulbeniales the asterinoid habit has become most strongly developed; 

 but it has undergone extensive morphological degeneration owing to extreme xerophyt- 

 ism. Biologically, they seem to have undergone a development from endo- to com- 

 plete ectoparasitism, like that we have met in the Perisporiales and Sphaeriales. 

 A few species, as Trenomyces histophthorus on Mallophaga, are endoparasitic, like 

 Leveillula taurica of the Erysiphaceae and Lanomyces tjibodensis of the Perisporiaceae, 

 and with their well-developed haustoria penetrate the whole fatty tissue of the 

 host. Other forms, as Dimeromyces rhizosporus, Ceraiomyces Dahlii and Arthro- 

 rhynchus Cyclopodiae, correspond to the Phyllactinia type of the Erysiphaceae; 

 they form richly branched, intramatrical haustoria which spread only in the 

 immediate vicinity of the point of infection. Other forms, as Rhizomyces ctenophorus 

 and Moschomyces insignis, penetrate the chitinous integment of the host and form 

 a lobate haustorium on its interior (Fig. 251, 2). In most species the intramatrical 

 portion of the thallus is reduced to an ungulate haustorium, the black foot, which is 

 let into the integument of the host, goes no further into the interior of the body and 

 apparently takes its nourishment through the host integument (Fig. 251, 3). Analogies 

 to this relationship are rarely known in fungi; one was earlier described for Harpochy- 

 trium Hyalothecae (Fig. 23, 1) ; another may be seen in Fig. 262, 1. On the basis of the 

 observation that the insects suffer no serious injury and the assumption that a suffi- 

 cient intake of nourishment through the integument is impossible, the idea has been 

 expressed that the purely ectoparasitic Laboulbeniales are only saprophytes like 

 the epiphytic Sphaeriales of the Limacinia-Teichospora type. This idea is hardly 

 correct, since many species are adapted to definite genera of insects. 



It seems rather a question of end forms of an asterinoid developmental series 

 in which the step from ectoparasitism to epiphytism is small, as in the Erysiphaceae. 



