DISTINCTIONS IN CULTIVATED BARLEYS. 
11 
Table I. — Influence of geographical location on the length of the culm in IS 
representative selections of barley grown at four ividely separated points, the 
selectio-ns being arranged in the order of their height at each point. 
St. Paul, Minn. 
Williston, N. Dak. 
Moccasin, Mont. 
Chico, Cal. 
Hordeum vulgare 
Oderbrucker 
Servian 
S. P. I. No. 20375. 
Odessa 
Hordeum vulgare 
Oderbrucker. 
Hordeum vulgare 
Smyrna 
Abyssinian. 
Princess 
Oderbrucker 
Smyrna. 
Manchuria 
Summit 
S. P.I. No. 20375 
Kitzing, 6-rowed 
Servian 
S. P. I. No. 20375 
Surprise 
Odessa. 
Kitzing, 6-rowed 
S. P. I. No. 20375 
Princess 
Kitzing, 6-rowed. 
Princess. 
Kitzing, 6-rowed 
Smyrna 
Abyssinian 
Kitzing, 2-rowed. 
Surprise. 
Hordeum vulgare. 
Kitzing, 2-rowed 
Odessa 
Kitzing, 2-rowed 
The great variation evidenced by these few selections is sufficient 
to show that the length of culm can not be of much taxonomic value. 
There are varieties which are persistently below average height, and 
others that are as persistently above, but beyond that it is difficult to 
make an unqualified statement. Locally, this measurement is of 
more significance and can often be used to advantage in the study of 
nursery selections. The differences it reveals are important in 
breeding, no matter to what cause they may be due. 
DIAMETER OF THE CULMS. 
Measurements have not been found very useful in revealing small 
differences in the diameter of the culm. *The experimental error is 
large, owing to the fact that the diameter varies on the same plant 
with the culm selected, on the same culm with the internode chosen, 
and on the same internode with the distance from the node. A part 
of this variation was avoided by measuring the greatest diameter of 
the first elongated internode, but even then the results were unsatis- 
factory. There are varietal differences, but they must be great 
enough to be seen optically before the error of measurement is re- 
duced to the point where it becomes negligible. As a group, the 
nutans has smaller culms than the Manchuria, but among the Man- 
churia strains there is little difference. Only once in these investi- 
gations has this character been used to isolate a type. This type has 
proved to be stable, and perhaps the effort of measuring hundreds of 
selections is rewarded by the one strain obtained, as it is very prom- 
ising. 
THICKNESS OF CULM WALLS. 
A large number of determinations were made of the thickness of 
the walls of the culm, with even less satisfaction than in those of the 
diameter. Measurements finer than one-tenth of a millimeter are 
impracticable, owing to the variation within the plant and culm. 
This does not give range enough to disperse the varieties. For in- 
