PHASES OF POTENTILLA. 99 
their polygynous congeners that species of Rubus with very 
few and definite drupelets bear to others in whose fruits they 
are indefinitely numerous. 
Too much account has been made, in times past, of the 
delicate, almost capillary inflorescence of T. sanlotinoides. Tt 
was this peculiarity which, in the main, influenced M. Baillon 
when he proposed to make of it a separate section of Potentilla 
under the name Stellariopsis. | Yet this author did not, like 
Messrs. Brewer and Watson, erroneously characterize the in- 
florescence. The latter have called it a “diffuse panicle,” * 
whereas it is truly cymose, however diffuse, and, being cymose, 
is essentially that of all species. of Potentilla whose flowers 
are not solitary. 
I have seemed called upon to produce this explicit and 
somewhat lengthy statement of the case in hand, because, 
while neither Horkelia nor Ivesia found recognition either 
by Bentham or Baillon in their great treatises upon genera, 
in following those authors I am doubtless dissenting from the 
opinion of the most eminent and experienced of American 
botanists ; for Professor Gray has, in recent years, here and 
there expressed his mind as for the retention of the first, if 
not for the second of the genera. It is not easy for an Ameri- 
can who is not an authority, to take a step like this, knowing 
as he does beforehand that American authority will be against 
him. But unless I have culpably neglected the opportunities 
of the last ten years, I really ought to possess considerably 
more knowledge of the plants in question, than has been en- 
joyed by any other who has written upon them ; and, feeling 
obliged to take the ground here taken, I have endeavored to 
give the reasons. 
In assigning to the species their names under Potentilla, 
little effort will be made to group them. Even as subgenera 
Horkelia and Ivesia can not be limited otherwise than arbi- 
trarily. 
* Bot. Cal. Geol. Surv. i, 183. 
