Page:Makers of British botany.djvu/186

 148 species: and natural habit is often a safer guide than minute microscopic characters." Thus we see that in his method convenience of diagnosis is put before the use of important structural characters. I have recently found reason to uphold the opinion of Mettenius on this point, and to confirm Plagiogyria as a substantive genus.

Similarly, the genera Lophosoria and Metaxya will have to be detached from Alsophila: Prantl removed Microlepia from Davallia into his new family of the Dennstaedtiinae, where they are related with Patania (Dennstaedtia), which Hooker had merged into Dicksonia. Goebel also has detached Hecistopteris which Hooker had placed in Gymnogramme, and has placed it with the Vittarieae. These are all examples of the way in which further study is tending to reverse the excessive merging of genera, which Hooker carried out in the interest of diagnostic convenience.

The general conclusion which we draw from contemplating Sir William Hooker's work on the systematic treatment of ferns is that it was carried out consistently to the end under the influence of the current belief in the Constancy of Species. The methods were not phylogenetic, as they have since become under the influence of evolutionary belief. The problem seems to have been to depict and describe with the utmost accuracy the multitudinous representatives of the Filicales, and to arrange them so that with the least possible difficulty and loss of time any given specimen could be located and named. But the result is not to dispose them in any genetic order. Even the arrangement of the larger genera according to the complexity of branching of the leaves appears as a method of convenience rather than of genesis, and subsequent inquiry is tending to show that so far as such series really exist, they will require to be read in converse. Goebel, in his paper on Hecistopteris, remarks that "the systematic grouping of the Leptosporangiate Ferns, as it is at present, e.g. in the Synopsis Filicum, is artificial throughout; it is adequate for the diagnosis of Ferns, but it does not give any satisfactory conclusion as to the affinity of the several forms." He proceeds to say that "a thorough investigation, taking into account the general characters of form of both the