35 
The first reference to the var. Gmelini made in North American litera- 
ture is probably one by Vasey who, describing Agropyron Scribneri in Bui. 
Torr. Bot. Cl. X, p. 128, 1883, says that this species “is perhaps A. caninum 
var. Gmelini , Led./’ a combination, by the way, which, as Pease and Moore 
point out (13, page 66), was not made by Ledebour. 
Later, in U.S. Dept. Agr., Div. Agrost., Bull. No. 4, pages 30-31 
(1897), Scribner and Smith describe A. gmelini as a species occurring from 
Washington to western Nebraska, stating that “this plant agrees with the 
figure and description of Triticum caninum gmelini Griseb. in Ledeb. Icon. 
FI. Ross, t. 248.” 
On the strength of this assertion Pease and Moore (13, page 75) 
use the name Gmelini (Ledeb.) Pease & Moore for one of the varieties 
of A. caninum, thus admitting Ledebour’s var. Gmelini as a plant occurring 
in North America. 
Recently, however, LIulten (4, page 151) has questioned the 
identity of the Asiatic var. Gmelini and Scribner and Smith's A. gmelini, 
pointing out that the description of the latter does not fit the Asiatic plant. 
He, therefore, hesitates to identify the American plant with the Asiatic, 
and considers the latter a species of its own, calling it A. Turczaninowii 
Drobov. 
An examination of specimens of A. Turczaninowii cited by Hulten 
and kindly placed at the writer's disposal by Dr. Samuelsson, Stockholm, 
has revealed that the Asiatic plant certainly is very different from what 
Scribner and Smith named A. Gmelini. In the first place, the Kamchatka 
plant has pronounced rootstocks. One of the specimens, No. 2946a, has a 
rootstock with two internodes, one of which is about 1 cm. long and the 
other 2 cm. long. This characteristic at once separates A. Turczaninowii 
Drobov from A, gmelini of Scribn. and Smith, which latter the 
authors, in their analytical key to the species of Agropyron (16, pages 
25-26), describe as caespitose with intravaginal innovations. There are 
also differences in some characters of the spikelets, which made it quite 
clear that the two plants belong to different species. Thus, for instance, 
A. Turczaninowii has the empty glumes as well as the lemma very strongly 
scabrous, a character beautifully illustrated on Ledebour's tab. 248, a 
photograph of which has been procured through the kind assistance of 
Prof. Fernald, whereas in Scribner and Smith's A. Gmelini the lemma is 
smooth or finely scabrous, as in the plant commonly known as A. Richard- 
soni Schrad. In fact, most of the specimens cited by Scribner and Smith 
represent typical A. Richardsoni or forms thereof, as the writer has ascer- 
tained through an examination of the material in the United States National 
Herbarium, kindly loaned by Prof. Hitchcock. 
A . caninum var. Gmelini f. Pringlei (Scribn. & Sm.) Pease & 
Moore was originally placed by Scribner and Smith as a variety of their 
A. gmelini, which, in turn, as mentioned above, was based on Triticum 
caninum var. Gmelini Griseb. apud Ledeb. Even a most cursory com- 
parison between Ledebour’s tab. 248 of var. Gmelini and Scribner and 
Smith’s type of A. gmelini var. Pringlei makes it abundantly clear that the 
two belong to different species. The latter is a high-alpine grass of low 
stature with a tufted habit and the culms geniculate at base. The short 
leaves are smooth underneath and the lemma is glabrous except for the 
minutely scabrous tip. 
