7 253 



Phegopteris and Leptogramma have indeed no value as characters of groups of reallj' 

 related species within Eudrifopteris. still it becomes necessary to look for others, 

 according to which it may be possible to group related forms in a natural manner. 

 As we have nol any safe base for a natural classification in the generative organs, 

 we must seek in the vegetative organs especially the different architecture and 

 structure of the leaf for available characters. It is, however, by no means easv 

 to find such, as there is hardly a character which constantly reappears in even 

 closely related species. By studying a long series of forms, however, it be- 

 comes possible to arrange the species in natural groups, the peculiarities of which 

 are not in a single character but in a sum of common features, which is very 

 difficult to express in words alone. That, for instance, D. opposita and D. Spren- 

 gelii each by itself represents a natural group is perfectly clear to me without its 

 being possible to explain in a few words the characteristic difference between them. 

 As, however, from practical reasons it would be desirable to give these groups 

 distinctive marks, I shall here brielly mention two characters which are the 

 basis for the main divisions in the following key; still one must always bear 

 in mind that these two characters like all others are rather vague, though in 

 most cases they will contribute to facilitating the right determination. The two 

 characters are: (1) the number of veins in the segment, and (2) the reduction of 

 the leaf towards its base. 



Although the number of veins in a segment is of course partly depen- 

 dent on its length, yet it is subject to other laws and must be a character 

 inherited by the species, which is not subject to any great variation. It is parti- 

 cularly the distance between the veins themselves which is of importance, and one 

 might possibly by minute measurements of these distances, determine how much 

 thej' vary within the same species and thus point out an absolute specific character 

 in contrast lo other s])ecies; but such measurements, where we would have to 

 count with fractions of millimeters, will in reality be difficult to carry out. It 

 may be sufficient to state the length of the segment and the number of its veins, 

 in which manner one can much more easily tell the density of the veins. As 

 D. tenerrima, the segments of which are of about the same length as found in D. 

 Sprengelii, has only 10 veins to a side, while D. Sprengelii has 16—18, and as the 

 forms closely related to these species all show the same difference, it is natural to 

 suppose, that the two species mentioned belong to two different groups, the separa- 

 tion of which is of older date than the separation of the single species of each 

 group. This seems justify us in seeing a character in the number of veins, which 

 may be used when dividing the species into larger natural groups. The distinguishing 

 number 10 as used in the key, is, of course, somewhat arbitrary, and not all species 

 of the two groups have more (or fewer) veins than 10". 



' After having written tlie above I liavc seen that Cope land ("The comparative ecology of San 

 Kamon Polypodiaceæ", Philippine .Journ. of Science Section C. Botany, 2. Jan. I'.t(l7, pag 47) has recently 

 measured the distances between the veins of a number of Philippine ferns, and he has arrived at the 



