CHYTRIDIALES 61 



myxamoebac divide independently. Finally, the protoplasm 

 splits up into uninucleate bits, and each of these rounds up, 

 assumes a wall, and becomes a spore. The spores formed by 

 each myxamoeba remain attached to one another, the whole 

 being in the form of a hollow sphere, which, according to Schroter 

 (1899) and Schwartz (1911) is enveloped in a very thin soral 

 membrane. These spore balls are later freed by the disintegra- 

 tion of the host, and each spore germinates b}^ a single zoospore. 



Two other species, S. junci Schwartz (1910) and S. graminis 

 Schwartz (1911), were later added to the genus. These forms 

 correspond to *S'. veronicae Schrot. in all the essential features 

 of the life cycle, though the spore balls are much less regular 

 in form, the spores sometimes adhering merely in chains. The 

 host is not stimulated to form galls, the attacked roots being 

 even thinner than the normal ones. When Maire and Tison 

 (1911) founded their new germs Lignier a, to include all the species 

 of the family which do not cause hypertrophy of the host, they 

 transferred to it S. junci; and the following year S. graminis, 

 which had meanwhile been pubhshed, was placed in Ligniera by 

 Winge (1912). Subsequently, Cook (1926) has shown by inocu- 

 lation experiments that these two species are identical, and that 

 several other species (L. bellidis Schwartz, L. mcnthae Schwartz, 

 L. alismatis Schwartz, L. pilorum Fron & Gaillat) more recently 

 described were also based on this one species. The genus 

 Ligniera contains several other species (L. radicalis Maire & 

 Tison, L. verrucosa Maire & Tison, L. isoetes Palm). Cook 

 (1926) considers it highly probable that L. radicalis is also 

 identical with *S. junci. On the other hand, L. verrucosa Maire 

 & Tison is clearly a distinct species, marked by its roughened 

 spores and thc'r less evident tendency to cling together in definite 

 masses; and L. isoetes Palm, if in fact a member of this group, 

 is certainly distinct. 



As the genus Ligniera Maire & Tison was based merely on 

 a host reaction the writer is wholly disinclined to accept it as 

 valid. It is clear that two hosts may react quite differently 

 to a given parasite, one being stimulated to gall formation, while 

 the other is not. It seems best for the present to include all of 

 the species in Sorosphaera, even though in some the tendency 

 to form definite spore balls is much less evident than in the type 

 species S. veronicae Schroter. The genus as thus reconstituted 

 includes besides the type also *S. junci Schwartz, and tentatively 



