2 8 2 The Botanical Gazette. [October, 



On the structure and dimorphism of Hypocrea 



tuberiformis. 



<;EO. F. ATKINSON 



(WITH PLATE XXV.) 



Hypocrea tuberiformis was described as follows by Berkeley 

 in Grevillea IV, p. 13, from specimens collected by Ravenel 

 in South Carolina: "Magna tuberiformis mycelio radiata 

 alboaffixa. On stems of Arundinaria. Car. Inf. Ravenel. 

 No. 1220. Forming either a large mass | of an inch across, or 

 two or three distinct subglobose individuals, fixed to the 

 stem by a radiating white rugose mycelium; at first yellowish, 

 then black." 



^ It was also distributed in Rav. F. Am. n. 733, and in Rav. 

 Fung. Car. n. 52. Most of the specimens collected by Rav- 

 enel were probably sterile, since there is, to my knowledge, 

 no published description of the forms of this dimorphic fungus. 

 Saccardo (Sylloge Fungorum II, p. 534) repeats Berkeley's 

 description. Ellis and Kverhart (Journal of Mycology II, 

 p. 68) place it in the group of Hypocreacae, which constitutes 

 Saccardo's genus Hypocrclla, and add the following note, 

 probably taken from specimens distributed by Ravenel: 

 "Apparently the original specimens were imperfect and, as 

 those in Rav. F. Am. are either young or sterile, we can only 

 say that the perithecia are subcylindrical and stand on the 



millim. high." 



'/ 



Patouillard (Bull, de la Societe Mvcologique de France, VI, 

 2e fasc. pp. 107-9, 1890) describes a new genus of Hypo- 

 creace.e under the name Dussiella. The perfect condition of 

 the fungus he examined was deposited in the Berlin Museum 

 and collected on stems of Arundinaria at Caracas, the coni- 

 dial stage from specimens collected by Duss in Martinique 

 He considers these forms to be the conidial and ascosporous 

 stages of Hypocrea tuberiformis B. & R., but unquestionably 

 it is a very distinct fungus from the one described by Berke- 

 ley, from Ravenel's collection. 



The perithecia are entirely immersed in the stroma, para- 

 ges are present, and while, as he states, the linear aW 

 pores .how its relationship to Epichloe and Hvpoarlla, the 



