15(1 3o. II. •;.] CLASSIKrCATIOX nj' Ko-SIL I'LANTS. 375 



to be delimited, in order that the nauit's may not he used 

 difterently l)y difterent writers. I have tried to chdimit Pfi- 

 lophylhtm by aceeptini^ Feistmantki/s definition which makes 

 the downward trend of both pinna-edges at the base the ge- 

 neric character. This is thus no new idea, the genus having 

 been accepted in this sense also by other palaeobotanists af- 

 ter the days of Feistmantel. This character, it must be ad- 

 mitted, is not a very opportune one, since it is often only 

 seen with difficult}-, and it does not mark off the genus very 

 distinctly from all forms of Zamifrs, yet it seems to l^e the 

 better one under the circumstances. As typical members of the 

 genus Za mites may be regarded the species with large fronds 

 of which the pinna-bases are distinctly rounded and characte- 

 rized by a more or less marked callosity, Z. gigas being the 

 best known representative of this group. The callosity can- 

 not always be observed, however, and its occurrence is gener- 

 ally not considered as a conditio sine qua non for a reference 

 to Zcunites. To the same genus are often referred smaller 

 forms without callosity and with the pinna-basis very slightly 

 rounded. As well-known representatives of this group of forms 

 may be regarded the species of Zamitcs described by Heer* 

 from Greenland; and it was because of the resemblance of the 

 Antarctic fronds to these Grreenland forms that I referred the 

 former to Zamites. 



The difference between these two groups is in many cases 

 obvious, and I have considered it in the paper quoted by 

 distinguishing the two types as different sections of the ge- 

 nus, resp. Kuzaiiiitcs and Snhzaniitis. It may be that the 

 latter section would ])e better designated by a new generic 

 name, but I did not judge such a course sufficiently war- 

 ranted. If, on the other hand, the choice is between includ- 

 ing these forms in Zdiuitts and referring them to PtilojtlniJ- 

 turn, it would appear that the former genus is decidedly t(> be 

 preferred. 



* Flora I'ossilis arrtica. Vol. 3, li>73, p. •;:', :ni.l lull . VU 11 ami 1.'.. 



