29 
appressed. In the corresponding North American forms no movement, 
or at most a very slight one, took place. The spikelets remained closely 
appressed to the axis both during and after anthesis. 
It was also noticed that in the European A. caninum the anthers are 
about 3 mm. long, whereas, in the North American corresponding forms 
and in A. tenerum they are at most 2 mm. 
The above observations led to the suspicion that what so far has 
passed as A. caninum in North America might not be conspecific with the 
true A. caninum of Europe. At the time that the biological differences 
between the European and the North American plants were discovered, 
sufficient European material was not available to determine conclusively 
whether the North American forms, morphologically, were sufficiently 
different to deserve specific distinction. Later, however, through the 
kindness of Dr. G. Samuelsson, of the State Museum (Riksmuseet), Stock- 
holm, Sweden, a much appreciated opportunity has been given to examine 
a large and representative collection of A. caninum and allied species from 
various parts of Europe. After comparing the European collection with 
so-called A. caninum from North America and its allies, in collections of the 
United States National Herbarium, Washington, D.C., the Gray Herb- 
arium, Cambridge, Mass., the Herbarium of the New York Botanical 
Garden, New York — to the curators of which the writer herewith expresses 
his sincere thanks — as well as in the National Herbarium of Canada, 
Ottawa, a task that entailed an examination of nearly three thousand 
sheets, the conclusion was reached that the true A. caninum (L.) Beauv. of 
Europe does not exist on the North American continent, except perhaps as 
an accidental and rare introduction. 1 
The most conspicuous difference between the European A. caninum 
and its North American relatives is that the former has the empty glumes 
coarsely 3-nerved, whereas in the North American forms, including the 
so-called A. tenerum , the empty glumes, with exceedingly few exceptions, 
have at least 4 or 5 nerves. In the North American forms the nerves are 
less prominent than in A. caninum of Europe, a fact that, to a certain 
extent, at least, may be accounted for by the former having thick, firm 
glumes, whereas in the latter they are thin and membranous. In the 
European A. caninum, furthermore, the empty glumes have a rather broad, 
scarious margin which, on the external side of the glumes, is widened into a 
projecting edge immediately below the apex. In the North American 
forms the scarious margin is much narrower and in many cases obsolete. 
It gradually peters out upwards and does not extend below the apex. 
As pointed out, the North American forms have the empty glumes 
normally at least 4- or 5-nerved. Among the odd three thousand specimens 
examined, a few — less than half a dozen — from eastern Canada and north- 
eastern United States, have most of the empty glumes 3-nerved as in A. 
caninum of Europe, but in these cases the scarious margin is not dilated 
below the apex. These, at first somewhat puzzling, specimens represent 
starved forms, occurring on Scrub pine land, in Thuja swamps, and in 
similar, poor habitats. That they do not belong to the European A. 
caninum is furthermore evident from the size of' the anthers which are 
about 1 mm. long when emptied. 
‘Cfr. Britton and Brown: “Illustrated Flora of the Northern United States, Canada, and the British Posses- 
sions,” vol.I, p. 285 (1913). 
