218 Rhodora [NOVEMBER 
setts specimen is referred to this variety by Schiffner himself and agrees 
closely with most of the specimens from Connecticut. Instead of 
growing in quiet pools, where typical forms of the species flourish, the 
variety grows on rocks and stones in brooks. Schiffner states that no 
similar form has been found in Europe and suggests that the American 
plant ought perhaps to be raised to specific rank. Aside from the 
slightly smaller leaf-cells, however, it seems to show no good differ- 
tial characters, and it is probable that the differences in the size of the 
cells are associated in some way with the differences in habitat. In 
any case the writer is inclined for the present to include the var. 
Sullivantii among the forms of Ch. fragilis. 
Although there is usually little difficulty in recognizing CA. fragilis, 
the diagnostic characters, as indicated above, are drawn from differ- 
ences in general appearance, in habit, and in size, rather than from 
morphological or structural differences. Even the size of the leaf-cells, 
which Schiffner emphasizes especially, is subject to considerable 
variation, and the differences in the form of the leaves are too slight 
and too inconstant to be considered seriously. According to Müller 
the characteristic features of Ch. fragilis are due entirely to environ- 
mental factors and disappear when the plant establishes itself in drier 
localities. Under these circumstances, in his opinion, the plant 
becomes quite indistinguishable from Ch. polyanthus. Schiffner shows 
pretty conclusively that this is not the case. He describes a variety 
subterrestris of Ch. fragilis, which grew in a locality alternately wet 
and dry, and shows that this is distinct from Ch. polyanthus. He 
admits also a variety submersus Loeske of Ch. polyanthus, which has 
slightly larger cells than the typical form of the species but which is 
evidently distinct from Ch. fragilis. According to Macvicar, who 
retains Ch. fragilis as a variety under Ch. polyanthus, the plant is even 
closer to Ch. pallescens and cannot always be distinguished from it. 
He emphasizes the more opaque and rarely emarginate leaves, those 
of Ch. pallescens being translucent and frequently emarginate. 
Aside from the New England stations listed above, Ch. fragilis may 
be recorded from the following North American localities: Torbay, 
Newfoundland (Howe & Lang, 1358); near Montreal, Quebec (Du- 
pret); Port Renfrew, Vancouver Island (Miss Gibbs); Tacoma, 
Washington (Flett); Yosemite Valley, California (Cooke); Tate Mine, 
near Marysvale, Utah (Jones). 'The last named specimen belongs 
to the variety calcareus Schiffn. 
