215 
shorter, are of the same pale color, and are evidently striate— 
striate-angled, however, rather than furrowed. The seeds of majus 
prove to differ from those of Caxadense in respect of size, color and 
Surface character, much as those of mutilum differ from those of 
boreaie. The broad-based leaves of majus and mutilum here comes 
_ to mind and suggests a sort of parallelism between Canadense and 
majus on the one hand and éoreale and mutilum on the other. 
It may be noted that sayus has the smallest and narrowest 
seeds of the entire series; doreade, the smallest plant of all, having 
the largest seeds. 
The relationship between Canadense and doreale needs no ex- 
tented comment. It is true that from the leaves alone some 
forms of each plant would scarcely be regarded as different, but 
the character of the inflorescence of each is perfectly destinctive : 
Canadense will be found to differ constantly in its more strict and 
compound, minutely bracted cymes, with more or less tapering cap- 
sules and calyx lobes. 
It has just been said that forms of Canadense cannot by the 
leaves alone be separated from forms of doreale. This is equally 
true of majus. All three species assume in some of their states a 
narrowly. oblong form of leaf which is practically the same in all. 
In specimens of majus, however, which exhibit this form of feaf, 
the other distinctive characters remain unimpaired: it maintains 
the large pod, long tapering sepals, and characteristic seeds. Not 
So, however, Canadense. In the oblong-leaved form collected at 
York Harbor, there is a general difference also in the inflor- 
€scence and capsules; the cyme is less strict and more delicately 
branched, even becoming somewhat diffuse, and the short-pedi- 
celled capsules are more numerous and crowded, and only about 
half the usual size, 1-2” instead of 22" long; the seeds also 
take on a darker color, and seem to approach those of mutilum 
Indeed, some examples of this form have much similarity = 
less branched forms of mutilum, and actually seem to be impli- 
cated somehow with that species. In fact, in the case of some 
immature specimens, it can aaeoy be said to which species they 
belong. 
Compared with fully typical Canadense, this form reer . 
striking contrast, and in studying the group I have found it con- 
venient to label it var. parvicarpum. 
