PHASES OF POTENTILLA. 97 
blages of plants not distantly related to this one, such as have 
been already cited; for some Saxifrageous genera are ap- 
parently allied to Potentilla more closely in nature than is 
represented by their placing in our books. In Saxifraga the 
organ under consideration ranges from rotate to campanulate, 
and from polysepalous in some species to gamosepalous in 
others. In Ribes it even takes a wider range of forms, run- 
ning all the way from the cyathiform to the long tubular ; 
and that phase of the genus which exhibits the elongated 
tubular ealyces appertains to the same phytographieal dis- 
trict which gives us our Horkelia or species of Potentilla 
with calyx modified, but less so, in the same direction. 
With reference to the few and definite stamens of Horkelia 
as a mark of distinction between it and ordinary Potentilla, 
suffice it to say that among the latter, stamens when as 
many as twenty, twenty-five or thirty to a flower, are often 
pretty definitely numerated, and are apt to be arranged more 
or less manifestly in five sets of four, five or six in each set, 
and that there is an Atlantic American species, P. pentandra, 
perfectly true to the old type of the genus, in which the 
stamens are definitely five only ! 
But the most essential character of Horkelia, according to 
the latest plea which has been made for the genus, is that of 
the petaloid dilatation of its filaments! The omission of this 
feature from the generic character as framed in the first place 
by Chamisso, and as repeated by Endlicher, can not have 
been an oversight ; it is too conspicuous a feature. But either 
of those celebrated authors may be presumed to have gener- 
alized upon this subject of dilated filaments, far enough to 
see that, in this particular alliance at least, it is a circumstance 
which will not bear weight as a generic character. 
Within Saxifraga transitions are made from filiform fila- 
ments to such as are upwardly dilated, to subulate, and ETOR 
to triangular-petaloid, and we have yet no Horkelia in which 
1 «This last and most distinguishing feature was omitted from the 
original generic character, and also from that of Endlicher.” A. Gray. 
Proc. Am. Acad. vi. 528 (1865). i 
