108 
W. C. Etheridge 
have not used them for the differentiation of classes. In the present 
classification the dimensions of grains are seldom used for any but a 
descriptive purpose. There are very few varieties the grains of which 
may distinctly be characterized by dimensions. In nearly all varieties 
the measurements of grains cannot be classified; they are transitional 
between types and between individual varieties. 
DOUBLE-GRAINS 
The so-called " double-grain" in cultivated oats is a condition of the 
spikelet in which the second grain is either partly or wholly inclosed by the 
lemma of the defective first grain (Plate XXI, 3) . It is found only in spike- 
lets that have two grains, and has been considered by other writers both 
as a mere abnormality and as a varietal characteristic. Atterberg (1891) 
believed double-grains to be due to unfavorable conditions of weather 
at the time of ripening, and therefore of no importance in classification; 
but he also noted their more numerous occurrence in certain varieties 
than in others. Fruwirth (1907) speaks of normally developed double- 
grains, and observed a varietal tendency to produce them. Nilsson-Ehle 
(1906) apparently regarded double-grains as a character of little impor- 
tance in oat breeding, since under the environmental conditions existing 
in Sweden they composed but a small proportion of the total number of 
grains in the panicle. He found, however, among different varieties a de- 
cided range in the occurrence of double-grains, in respect of which 
there was a varietal stability under somewhat different environments. 
The investigation of Bohmer (1908-09) shows a greater tendency by 
some varieties than by others to produce double-grains, but the variation 
in their production was greater as between seasons than as between vari- 
eties. Krogmann's data (1908) show a considerable range among varieties 
with respect to the kernel content of double-grains. 
From the foregoing views the status of the double-grained spikelet may 
be defined as an abnormality resulting from incomplete development, but 
toward which there is a varietal tendency. This conclusion has been fully 
confirmed in the present work, but in the production of double-grains a 
varietal tendency much greater than that reported by others has been 
observed. For example, in the widely different varieties Storm King 
and Canadian, the typical spikelet is double-grained, although the data of 
Nilsson-Ehle and those of Bohmer show, in varieties of Swedish and 
