LYCOPERDACEAE 103 



epigean, and this is borne out by the plant just mentioned, as the outer layer has not 

 been stripped off and is furthermore free of trash. 



In attempting to find a natural classification for the species of Geaster, one meets 

 with difficulties and with differences of opinion among authors. What are the char- 

 acters expressing the most fundamental relationships; and those of lesser degrees of 

 importance? All arrangements heretofore proposed have in certain sections obviously 

 violated, it seems to us, the true relationships. De Toni (Rev. Mycol. 9: 64. 1887. 

 See also Sacc. Syll. 7 1 : 70, and Hollos, I.e., p. 51) divides the true Geasters first into a 

 fornicate and a cupulate group (Fornicati and Cupulati) and the remaining species which 

 are subdivided into the sulcate-striate group (Striati) and the fimbriate group (Fim- 

 briati). As a result of the first two groupings there are thrown together the very dis- 

 similar fomicatus and radicans, while the near relatives fomicatus and rufescens, and 

 coroiiaius and minimus are separated. The third group is a natural one, but the last 

 should be rearranged. 



Morgan's two main groups, DepeUiti and Pelliculosi, are in part natural, but are 

 unnatural in having species with sulcate-plicate peristomes in each. His two sub- 

 groups under DepeUiti are unnatural for the same reason. He does not make the 

 mistake of making the hygroscopic character a leading one. In Lloyd's last outline of 

 the Geasters (Myc. Notes, p. 317), as well as in The Geastrae, he first divides them into 

 the Rigidae and the Xon-rigidae, the latter with six groups. This throws together such 

 distantly related plants as mammosus and Drummondii and separates species with 

 sulcate peristomes. 



Of the six groups of Lloyd's Non-rigidae, the first and last are natural in their in- 

 clusions, though Hariotii is excluded from the first. The other three groups are in our 

 opinion unnatural to a greater or less degree. That the hygroscopic character is not the 

 most fundamental one seems obvious. Both hygroscopic and non-hygroscopic species 

 are found in the very natural group with sulcate peristomes, as well as in the non- 

 sulcate group. In related species and even in the same species at times the intensity of 

 this character is very variable and in individuals may practically disappear. For 

 example, in arenarius the rays while hygroscopic are usually pliable and thin, while in 

 the very closely related floriformis they are much more rigid. In asper individuals 

 appear rarely without any hygroscopic tendency at all, though the species is markedly 

 hygroscopic. 



Within certain limits the structure of the mouth area is the most dependable 

 character. The sulcate group is a natural one and there are no confusing intermediates 

 connecting it with the non-sulcate group. So-called intermediates like Morganii are 

 not troublesome, as their crumpled mouths rarely bear any real resemblance to a truly 

 sulcate peristome. In the non-sulcate group the distinction between the definite 

 and indefinite peristome, while adequate enough as a rule, is in some cases obscured by 

 variation. This is particularly true in floriformis and arenarius. In constructing the 

 key we have used the characters in the order of their importance as it seems to us, and 

 the key is therefore as natural a grouping of the species as we can arrange at present. 



Literature 



Cunningham. Species of the Genus Geaster. Proc. Linn. Soc. New South Wales 51: 72, pis. 2-7. 



1926. 

 Cunningham. The Development of Geaster velutinus. Trans. Brit. Myc. Soc. 12: 12, 9 text figs. 



1927. 



