— 6 5 — 
var. cristulata as a possible subspecies, on account of the fact that its distinctive 
features seem to be hereditary. Miiller 1 accords varietal rank to the var. cristu- 
lata but reduces the var. gracillima to a form and groups the vars. turfosa, sub- 
aquatica and inundata under the forma elatior Gottsche, as mere synonyms. 
According to Loeske 2 the var. gracillima , although apparently only a biological 
“Kleinform” of Eucalyx crenulata (as he calls it), adapted to hard and bare foot- 
paths, occurs nevertheless so independently in the Harz Mountains and other 
localities that he gives it specific rank under the name E. gracillima. He sug- 
gests, moreover, that his E. gracillima may represent the original species, from 
which E. crenulata has been derived. Warnstorf seems to be even surer that the 
var. turfosa is not an aquatic form of N. crenulata. According to his last-expressed 
opinion 3 it is a distinct species, which he designates Haplozia turfosa Warnst. 
Of Schiffner’s six varieties the only one which is certainly known from North 
America is the var. gracillima. This sometimes occurs in pure mats, just as it 
does in Europe, and sometimes in admixture with the typical form of the species. 
Under the latter circumstances it presents the appearance of being a juvenile 
state of the typical form, and this may well be its true explanation. Even if 
this is the case, however, there may be tw;o closely related races or “small” spe- 
cies, one which advances beyond the juvenile state and one which does not. 
This would account for the fact that the var. gracillima evidently occurs by 
itself in certain regions, to the exclusion of typical N. crenulata. Unfortunately 
very little is known at the present time about the effects produced on the Hepat- 
icae by modifications of the environment, so that the various problems raised 
by var. gracillima and by the other varieties of N. crenulata must still await 
solution. 
As Howe has pointed out the Californian forms of N. rubra resemble the 
var. gracillima rather than typical N. crenulata. The leaves almost always 
lack enlarged marginal cells, and the leaf-cells throughout are essentially the 
same in structure. This condition tends to persist, even in the most robust 
forms, but the marginal cells then acquire characteristics unlike those found 
in robust forms of N. crenulata. In order , to make these points clear the struc- 
ture of the leaves in the two species will be considered at some length, even at 
the risk of repeating well-known facts. 
In iV. crenulata the leaf-cells and their walls have been variously described 
by writers. According to Schiffner the walls, both in the typical form and in 
the var. cristulata , are almost uniformly thickened throughout, and he uses 
this feature in separating critical forms of the species from N. hyalina (Lyell.) 
Carringt., where trigones are invariably present. According to Muller the 
walls are thin and trigones are either slightly developed or absent altogether. 
According to Hesselbo , 4 who discusses Icelandic forms, plants which are “low” 
1 Rabenhorst’s Kryptogamen-Flora 6: 542. 1909. 
2 Hedwigia 49 : 9. 1910. 
3 Kryptogamenfl. der Mark Brandenburg 2 : in 3. 1906. 
4 Rosenvinge & Warming, Bot. Iceland 1 : 410. 1918. 
