DISTINCTIONS IN CULTIVATED BARLEYS. 17 
vulgare. This is easily demonstrated by weighing kernels from side 
and central spikelets. In the Horcleum sativum vulgare the lateral 
kernels, compared with the central ones, are actually greater in rela- 
tive weight than is the case in the Horcleum sativum hexastichum. 
Differences other than density are likely to be due to the nature 
of the attachment of the lateral spikelets. Systematists describe the 
barlejr spikelets as sessile. This is true in most cases, but it ap- 
proaches an exception in Horcleum sativum hexastichum. In this 
group the central spikelets are sessile as usual, but the lateral ones 
either possess an elongation of the base of the flowering glumes or 
else are pedicellate. Among the barleys collected by the writer is a 
Greek form in which the lateral spikelets are elevated upon a pedicel 
that is over one-half as long as the length of the rachis internode 
itself. This pedicel is jointed both at its attachment to the rachis 
and at its attachment to the floret. It is the longer attachment of 
the lateral spikelets that allows the characteristic radial arrangement 
of Hordeum sativum hexastichum. Density is, however, a parallel 
factor. The compactness of the spike forces the kernels to assume 
certain relations. Both in Hordeum sativum hexastichum and in 
Horcleum sativum erectum, the kernels are placed at a much wider 
angle with reference to the rachis than in Horcleum sativum vulgare 
and Hordeum sativum nutans. The Swedish Plant-Breeding Associ- 
ation at Svalof has considered the angle of the inclination of the 
kernels as one of the more important of their notes. It is the opinion 
of the writer, however, that, with rare exceptions, it will vary di- 
rectly with the density, and is therefore superfluous if the latter 
measurements be taken. 
In breeding, density has not been utilized as fully as its value 
seems to warrant. Yoss (25), Kornicke (15), and Atterberg (2), 
have used it in group classification, and Atterberg, Blaringhem (8), 
and the breeders at the Svalof station have used it in studies of 
variation and purity, but in the opinion of the writer its possibilities 
in the isolation of types and in the identification of strains have been 
far from exhausted. 
In the years from 1909 to 1913 a close study of density was made, 
both upon general farms and in experiment-station nurseries. In 
this study, 100 spikes of' each variety were taken without other 
choice than that they were not diseased or dwarfed. On each of 
these spikes 10 internocles of the rachis were measured; that is, the 
distance was between six spikelets on one side of the rachis. From 
these measurements the number of internodes per decimeter was 
computed and this number taken as the unit of density. The for- 
mula was then D=l,000-^-L, where L was the length in millimeters 
of 10 internodes of the rachis. 
52783°— Bull. 137—3 
