VOL. XI, NO. 10.—BOTANICAL GAZETTE.—OCT., 1886. 
Memoranda of a revision of the North American Violets.’ I. 
ASA GRAY. 
It seems most natural to throw all the Candollian groups into 
one, except the section Melanium, which includes the pansies ; in 
this, following the late M. Boissier ; and to arrange our violets in 
SIX primary sections, upon characters of vegetation taken along 
with differences in the stigma. 
Grovr I. Strictly acaulescent; the dissected leaves and scapes all directly 
from an erect and short thick caudex rather than rootstock, never stoloni- 
ferous: corolla beardless: large antrorse-terminal stigma wholly beakless and 
naked, 
V. pedata L., with var. bicolor Pursh, fide Raf. 
Group II. Acaulescent; the leaves and scapes springing directly from 
the summit of a rootstock, or later more or less from runners: style with in- 
flexed or truncate and beardless summit and an antrorsely beaked or short 
pointed small stigma. 
* Rootstocks thick and short, multicipital, ascending or little creeping, 
never filiform nor stoloniferous, often fleshy-dentate: corolla only sacca 
spurred, blue or violet, occasionally varying to white; at least lateral petals 
bearded. Species connected by transitions. 
V. pedatifida Don. Syst. i. 320 (1831). V. delphinifolia Nutt. 
in Torr. ray, Fl. i. 136 (1838). This earlier name 
clearly belongs here and must be adopted. It is the V. pinnata 
of Richardson (not the Linnean species, which has longer and 
narrower spurs), the V. pedata of Hooker’s Flora as to the plant 
of Saskatchewan, ete. It has often been confounded with that 
species; but its affinities are with V. palmata, indeed is probably 
variety of that species, with all the 
t might take the much earlier name of 
1Read before the A. A. A. S., Buffalo meeting, 1886. 
