4 BULLETIlsr 622, U. S. DEPARTMENT OF AGRICULTUEE. 
detail, his conception of varieties was such as to establish most of 
them upon a firm basis. His treatment of the larger groups, on the 
other hand, was less satisfactory. He recognized hexastichan^ 
tetrastichum^ intermedium^ and distichum as his primary groups. 
These are more fully discussed later. A third paper appeared in 
1895, and in 1908, after his death, his son, M. Kornicke, published 
a fourth treatment. Xeither of these latter possesses the clearness 
of the earlier papers, and they include many doubtful varieties. 
Many of the additions were of hybrid origin and were unfixed at 
the time of publication. Kornicke's failure in many instances was 
due to the general lack of genetic information at that time. Eecent 
discoveries have made the purification of a type rather simple. For 
instance, his suhcomutum was selected from a mating of trifurcatwn 
and steudelu and was far from fixed. A constant variety of this 
type is readily obtained from a cross of cornittum, X pallidum, or of 
horsfoi'dianuTTi X haxtonl. Many varieties attributed to Kornicke in 
the ke}^ which follows were heterozygous forms with him which he 
Avas never able to fix. Most of these have been recognized in this 
l^ublication, because fixed forms corresponding to the descriptions 
have since been produced. 
In 1885, the j^earln which the '* Handbuch des Getreid^baues ** was 
published, a paper by Yoss appeared. Voss follows Jessen (18.55) 
and places all cultivated barleys under the single species Hordeum 
sativum. Although his handling of the varieties is inferior to that 
of Kornicke, his treatment of the larger groups is decidedly better. 
He uses polystichon, distkhon^ and defciens as his subspecies. He 
indicates that deficiens is subject to the same variations as distichon 
by subdividing it into dense and lax divisions. He substitutes m- 
equale for the term tetrastichum of Kornicke, pointing out that there 
is no such thing as a 1-rowed barley. 
Following Yoss came Atterberg, who published various papers 
between 1889 and 1899. In the latter year he advanced a new basis 
of classification. He made four subspecies under Hordeum, sativum 
Jess., namely, comwAine^ Tnacralepis^ furcatumy^ and i/mrme. As 
both th€ character of the outer glumes and the appendages to the 
lemma were used, the result was rather involved and seems to offer 
little compensation for so abrupt a departure. Atterberg's concep- 
tion of the regular occurrence of the subfactors through each of 
the groups gives evidence of a familiarity with a larger number of 
varieties than had been accessible to previous taxonomists. His sys- 
tem of uniform recurrence of names in his subgroups seems inad- 
visable and likely to lead to confusion, even though, as he points 
out, he uses only 33 terms to designate 188 forms, while Kornicke 
used 78 terms to designate 74 varieties. 
