| 1917] | Weatherby,— Color Forms of Impatiens biflora 117 
form. According to collectors’ notes and what I have seen myself, 
they vary somewhat in color, from crimson to pink and brownish red, | 
and very much in number. In some individuals they are few and 
scattered and in others so numerous as to coalesce in an irregular 
crimson spot on the lower spreading petal. But in all these forms 
they exist without essential change of hue, quite independently of the 
variations of the body-color of the perianth. In a fifth form, however, 
the perianth is orange, as in the typical form, but the spots are en- 
tirely absent. Rough tests of the pigment of these spotless flowers 
gave different results from the same tests applied to flowers of two 
spotted forms, of which fresh material was available at the time. 
This form may be called 
Forma immaculata, f. nov. Perianthiis aurantiacis, sine maculis. 
Perianth orange, without spots. Marne: Springy woods, St. Francis, 
Aug. 16, 1893, Fernald, no. 21a (TYPE, in Herb. Gray); Bar Harbor, 
Aug. 31 (no year given), Kate Furbish. Vermont: Mt. Mansfield, 
Aug., 1877, Faron. Munnesora: Lac qui Parle (no date or collector), 
“without spots.” 
Certain facts about these forms may be worth noting. As might 
be expected in a species producing cleistogamous flowers, they show 
abundant ability to maintain themselves. Forma Peasei was col- 
lected at the same station in Jefferson, New Hampshire, in 1901 and 
1907. Forma albiflora at the Billerica station, where conditions are 
unfavorable for its spreading, was present in about the same quan- 
tity in 1914 as when I first saw it there in 1911. Forma citrina at 
Thompson, where conditions were more favorable, increased consid- 
erably between 1908 and 1914, though the station is now likely to be 
exterminated by the building of a state road. That is, these forms 
are apparently not recurrent but unstable variations, like, for in- 
stance, Viola pedata, f. rosea,! but, once established, tend to remain as 
fixed and genetically constant lines. 
At the Billerica and Thompson stations the forms occur in pure 
colonies, associated with, even mingled with, the typical form, but 
not grading into it nor into any of the other variant forms. That is, 
at Billerica, all the variant plants are f. albiflora, at Thompson, 
practically all f. citrina. At the latter place, out of scores of plants, 
two exceptions were noted in which the color was somewhat inter- 
mediate between f. citrina and typical I. biflora. These, however, 
1 See Ruopora, xiv. 22 (1912). 
