KEYS TO THE FERNS OF BORNEO 
293 
them. • The old artificial tribes, the Aspidieae, Asplenieae, 
Davallieae, Pterideae, and Vittarieae and Polypodieae, not 
to speak of the Acrosticheae as a tribe in the older sense, 
have, therefore, to be recast or abandoned. As to the 
family conception, it seems to me clear that Dryopteris, 
and even Athyrium, as well as Per cinema,, Diacalpe, 
Monachosorum and Aerophones, and probably all of the 
Woodsieae, must fall in one family with Cyathea ; while 
the descendants of Dennstaedtia, including possibly the 
Polypodieae and Vittarieae, and certainly all of the real 
Davallieae and Pterideae, form another family; and while 
the descendants of the Matoniaceae fall in that family— 
or else we must include in the Polypodiaceae, as com¬ 
pletely as we know how to do so, the nearest common 
ancestry of all of these distinct phylogenetic lines. If we 
adopt the former alternative, the definition of the 
resulting families by diagnosis will become a pure impos¬ 
sibility. There seems to me, therefore, no alternative 
but to include Cyatheae, Matoniaceae and Polypodiaceae 
in one family, retaining, for the resulting great family, 
the name by which the overwhelmingly largest part of it 
is already known. 
In every attempt at a natural classification there are 
two inherent difficulties, and a third very great one which 
it is our business to outgrow. The first of these is the 
fact that our arrangement is necessarily a linear one, in 
which at the completion of one phylogenetic series we 
move back to an earlier, more primitive point to begin 
the treatment of another series. The second difficulty is 
the fact that in the evolution of series of plants, nature 
has not had regard for our convenience of definition, and 
that it is very unusual to find a large natural series which 
can be easily distinguished from other natural groups by 
diagnosis. In practice, this difficulty is passed around 
by making artificial keys, by which those who are not yet 
