256 10 



easily be used for a rapid determination of the species. As mentioned above 

 there is found an abundance of forms, which can be united naturally into smaller 

 limited "groups of forms", which term I prefer to use here for what I have named 

 and treated as "species" in the following pages. Many of these probably include 

 several real species, which are now in process of developing into separate species, 

 but owing to the lack of sufficient material cannot with certainty be recognized 

 as. such. Many of these groups of forms are well defined and easily recognizable, 

 while others are only faintly characterized and pass gradually into others. Still 

 I think I am right in separating groups of forms under a special name, when 

 at least a couple of characters, constantly occurring together and by which the 

 forms differ from related ones, can be pointed out, as I have done for instance in 

 separating D. pilosiila and D. argentina from /). oHgocarpa. 



The characters, which are of particular importance for the distinguishing of 

 the species, are: (1) rhizome, (2) shape of lamina, (3) size, (4) number of veins, 

 (5) pubescence, (6) texture, (7) position and shape of the sori, and (8) the sporangia. 

 Of these I have above more explicitly mentioned the two first and the fourth, and 

 it is evident that their importance for the recognition of species is essential. 



The size of the leaf is, of course, very variable within the same species, 

 especially as to its length, while the width or, in other words, the length of the 

 middle pinnæ is somewhat more constant. Still the breadth of the pinnæ is of 

 greater importance, as it rarely varies beyond '-centim.; in connection with this the 

 number of veins is very nearly constant in segments of normal size. The length 

 of the basal pair of segments, compared to the othei-, is also of considerable 

 importance. In a great number of species the basal segments are of equal size, in 

 D. opposita and its nearest allies they are either both or only one distinctly longer, 

 and in some species (D. rudis and its relatives) they are, especially in the larger 

 pinnæ, smaller, often even very small. It is remarkable that the length of the 

 basal segments is in correlation to the reduction downwards of the leaf; the leaves 

 of the types I and III (see schematic figures aliove) have nearly always equal-sized 

 basal segments, leaves of type II nearly always enlarged, of tj'pe IV very often re- 

 duced basal segments — still exceptions from this general rule may be found. 



As to the pubescence of the species, or in a wider sense, its covering 

 of trichomes: hairs, scales and glands, there may be found within the species a 

 great variation in the density of such coating, its kind seems, however, to be a 

 very constant character. Most variable is the coating with glands on the under- 

 surface of the leaf, many species occurring as densely or sparsely or even not 

 glandulose forms. In some species, as D. pacliyrachis and D. Sprengelii, the cha- 

 racteristic large, sessile glands are however an excellent diagnostic mark. The most 

 constant is the presence or the lack of scales on stem and rachis. In the group 

 taken as a whole scales occur only sparsely, still most species bear at the apex 

 of the rhizome and at the base of the stem some few, often finely pubescent, 

 deciduous scales. A thick cluster of scales at the base of the stem, as is seen in 



