510 SECTIONAL TRANSACTIONS.—M. 
ultimate problem would be to discover the relative value of different inherited plant 
characters in terms of value of Milk or of Meat. It is fairly obvious that theories of 
chromosomal effects on these relative values are far too complex for determination. 
There are too many genes involved. We are reduced to empirical methods aided by 
statistical analysis. 
The ultimate object of the cultivation of barley is the production of the maximum 
quantity of grain on unit area of the quality required either for feeding to animals, or 
for conversion into beer; the latter being its more profitable use. 
The breeder of any hybrid cereal requires to adopt methods which will give 
indications of value at an early stage in the multiplication of his individual plants. 
Nothing in the way of plant selection can be done until the F3 generation is reached 
and by this time there will be some hundreds of plants, any one of which may be a 
more desirable parent of a new race than any other. It is impossible to distinguish 
effectively between inherited and fluctuating differences due to environment at this 
stage. At least two thousand plants of any single plant culture must be raised and 
compared with a similar number of other single plant cultures before this distinction 
can begin to be established and comparisons made of genetic characters. To bring 
hundreds of cultures to the stage of field trials would require very extensive areas, 
expenditure and time. 
A Table is exhibited showing for two races of barley ‘factors of productivity ° in 
nine years. In each of these nine years eight races of barley were grown chequerboard- 
wise, with either sixteen or twenty replications for each race. The table shows for 
each year the means of determinations made on two of the eight races and the ratios 
of these values on the 164 pairs of plots. The two races selected for this sample of the 
available data are Spratt-Archer and Plumage-Archer, both of which are widely 
cultivated and together occupy more than half the total barley area of Great Britain. 
Produce of grain on unit area is given sufficiently, closely by nx 7’x H when n is 
the number of plants, T the average number of tillers surviving to produce ears, and 
E the average weight per ear. The same yield (Y) is also given by nx PX M when n 
is as before the number of plants, P the average weight of the entire plants, and M 
the ‘ Migration Co-efficient,’ i.e. the ratio of weight of ears to weight of entire plant. 
This is therefore TE/P. It follows of course that PM=TE and is the mean weight of 
ears per plant. It also follows that the products of the ratios, PA/SA, for each 
determination give the ratio of the yields. 
The data show that for these two races a prediction made in seven of the nine 
years from the factors T and M only, in the form PA/SA (TM—-03), would have given 
a somewhat better estimate of the mean relative yields for the nine years than is 
given by the actual yields in the same years. It cannot be claimed that this does more 
than point the way to further research for similar equations which might be of general 
application. It is, however, only to be expected that a result of this order should be 
obtained owing to the fact that under the climatic conditions of this country the 
critical periods for cereal crops are at the tillering and grain-forming stages. 
Differences of habit in races at these stages are well shown to be hereditary, whilst 
differences in plant population which greatly affect total produce are mainly due to 
incidental and environmental conditions. Moreover, both 'T and M can be determined 
Sete in any year on small plots with a much lower probable error than can gross 
yields. 
Concerning quality of produce, Dr. L. R. Bishop, working at Rothamsted for the 
Institute of Brewing, has calculated from a great mass of data a prediction equation 
based mainly on nitrogen content of grain which gives with a very small probable 
error the amount of brewers’ extract obtainable. Differences in nitrogen content are 
clearly shown to be hereditary and, applying Dr. Bishop’s equation to the above 
prediction equation for relative yields, we get a prediction for the mean relative 
quantity per acre of matter available for brewing given by the two races when grown 
under parallel conditions for nine years. The validity of this prediction is confirmed 
by results obtained from trials of the same races made in 1925/6/7/8 by the National 
Institute of Agricultural Botany at five stationsin England. There were ten replicated 
field plots of each race in nineteen trials, giving 190 pairs of plots. The produce of 
each of the nineteen trials was malted and the relative values of brewery extract was 
determined. 
The final ratio obtained by combining the TM equation, given above, with Dr. 
Bishop’s equation gives for the 164 pairs of Warminster plots PA/SA=:955. The 
a ee 
