98 PITTONIA. 
they are flattened to the degree last indicated. ^ Moreover, to 
carry the generalization a little way in another direction; 
subulate-dilated filaments distinguish the principal group of 
our Californian wild gooseberries, but no one has thought of 
removing them from Ribes on that account, nor, I may add, 
on account of their peculiar cucullate-involute petals, which 
are a good deal more unlike those of East American and Old 
World phases of their genus than Horkelia petals are differ- 
ent from those of ordinary Potentilla. 
If, then, we are to have in Saxifraga, as we do, the fila- 
ments ranging from filiform to deltoid, may we not allow a 
less range of diversity (an equal range is not required) in’ 
Potentilla? There is, it must again be said, no difference of 
habit to be noted between the original Horkelia and Potentilla 
glandulosa, which latter is the commonest western representa- 
tive of the old type of the genus. But there is a new species 
herein to be described, so like H. Californica in appearance 
that in collecting it I made no duplicates, supposing it to be 
only a form of that species with leaves thinner and more dis- 
sected, but which I now find to combine a campanulate calyx, 
and definite (ten) stamens of Horkelia with filaments that are 
filiform. So complete an invalidation of Chamisso's genus 
was neither expected nor required. 
Distinetions of the breadth and depth of the calyx failing 
to commend themselves as of generic value in the Rosaceo- 
Saxifrageous alliance, and few and definite stamens occurring 
in even the strictest type of original Potentilla, Ivesia has no 
better foundation than that of the paucity of its pistils. 
Several of the species have, pretty constantly, two only, and 
one of them only one. But this, which is certainly an extreme 
condition of things in a genus normally, and on the whole so 
largely polygynous, is attended with no change in the nature 
of the earpel itself. That is essentially the same in all; 
moreover, the transition in Ivesia from polygynous to mo- 
nogynous species is as gradual as can be, some having almost 
always three pistils, others four or five; and upon the whole, 
these all, as regards their fruits, have the same relation to 
