66 Rhodora [APRIL 
Wheeler. New Mexico: Fort Wingate, Dr. Matthews.  SASKATCH- 
EWAN: 1857-58, E. Bourgeau. British CoLuMBIA: Fraser River, 
1858, Lyall; Chilliwack Valley, 1901, J. M. Macoun, no. 26486. 
Var. foliosum (Fernald) comb. nov. E. purpureum var. foliosum 
Fernald, Rnopona x. 86 (1908).—Leaves large, ovate-oval, with a 
tendency toward a 3-ribbed base, thin, nearly or quite glabrous: 
inflorescence small, surrounded and usually much surpassed by the 
large lance-ovate upper leaves and bracts.—Newfoundland and 
Saguenay County, Quebec to northern New England and possibly 
Lake Superior. NEWFOUNDLAND: Manuel's River, 1894, Robinson & 
Schrenk; Brigus Junction, 1911, Fernald & Wiegand, no. 6275; Grand 
Falls, 1911, Fernald & Wiegand, no. 6276; Bluff Head, Bay of Islands, 
1898, A. C. Waghorne, no. 18; Harry's River, 1910, Fernald & Wie- 
gand, no. 4133. QuEBEc: Dartmouth River, 1904, Collins, Fer- 
nald & Pease; Mingan, 1915, H. St. John, no. 90749. Prince Ep- 
WARD ÍsrAND: Bloomfield, 1912, Fernald, Long & St. John, no. 8095. 
New Brunswick: Kent County, 1870, J. Fowler. Maine: Van 
Buren, “the common form,” 1900, M. L. Fernald; Hamlin, 1901, 
Fernald; Fort Fairfield, 1901, E. F. Williams; Boundary Lake, 1902, 
Eggleston & Fernald; Fairfield, 1916, Fernald & Long, no. 14648; 
Bridgton, 1912, L. R. Martin. New Hampsurre: Ice Gulch, Ran- 
dolph, 1908, A. S. Pease, no. 11498; Pittsburgh, Fernald & Pease, 
no. 17068; Jefferson, 1916, Pease, no. 16837. MicHiGAN: moist 
thickets, Keweenaw County, 1887, O. A. Farwell, no. 441 (doubtful). 
E. maculatum is a northern plant of the richer, more calcareous 
soils, and is therefore nearly or quite absent from the sandy coastal 
plain from New England southward. In Central New York it is 
the characteristic joe-pye weed, occurring in great abundance every- 
where in low grounds except perhaps in acid bogs. Locally it is 
somewhat variable as affected by shade and moisture, but geograph- 
ically it is much more so. In the northeast the var. foliosum seems 
distinct enough to warrant separate treatment, but the other varia- 
tions pass one into the other so gradually that no practical dividing 
lines can be drawn. Thus the leaves of the northeastern specimens 
are only slightly hairy beneath, but as one travels westward they 
gradually become more hairy until some plants on the Great Plains 
have the leaves distinctly whitened on the under surface. Locally, 
too, the degree of pubescence may fluctuate with the exposure. The 
more canescent western plants have been separated by Robinson 
as E. purpureum var. Bruneri (Gray) Robins. (l. c.). Several sheets 
of var. foliosum from Gaspé have the upper portion of the stem, and 
to a certain extent the under surface of the leaves, villous instead of 
