General Remarks. 



The 280 species, dealt with in this work, are grouped into ten subgenera. In 

 a recent j)reliminary paper') I have characterized the subgenera and pointed out 

 the principles, upon which my classification is based. Referring to that paper I 

 shall here remark only, that the kind of the trichomes, the most important cha- 

 racter, is not absolutely constant within each subgenus. That character is, like all 

 other characters, not sufficient for basing a natural classification upon it alone, but 

 it is, no doubt, the best and most constant character, by which groups of related 

 species can be distinguished from each other. Against my classification that objec- 

 tion may be made that I, when using a single character as distinguishing mark, 

 cannot omit errors similar to those so evident in all earlier classifications based 

 upon a single character. To this I want to say that my classification is not based 

 upon a single character, the structure of the trichomes. By examining again and 

 again and comparing the thousands of specimens I rather quickly succeeded in 

 being able to group the different forms thus, that the species referred to each 

 group were no doubt intimately related. The affinity is not shown by a single 

 common character, but by a certain conformity in the structure, colour and general 

 habit of the related species. That conformity is easily seen by the trained eye, 

 but it is very difficult to describe. Having sorted my material after that method 

 I then found, that the structure of the trichomes of species referred to a certain 

 group was remarkably alike and on further examination I found, that no other 

 character is so constant as that mentioned. The find was of much value to me, 

 for now it became an easy work to characterize my new subgenera in such a 

 manner, that each pteridologist can refer with appioximate accuracy a given form 

 to its right systematic position. As a matter of fact I shall point out that all 

 the 280 species, four or five perhaps excepted, could be determined to subgenus 

 from an examination of the scales and hairs alone. 



Qualitatively the ten adopted subgenera are not of exactlj' the same value. If 



') Gn a natural classification of the species of Dryopteris. Biologiske Arbejder, tilegnede Eug. 

 Warming 3. Nov. I'JU, pp. 73 85. 



