M.—AGRICULTURE. 261 
Although the statistical investigation is only recently begun, mathematical 
expression has already been given to the relationship between rainfall 
and yield of wheat and barley under different fertiliser treatments, and 
precision has been given to some of the ideas that have hitherto been only 
general impressions. If on an average of years a farmer is liable to a 
certain distribution of rainfall, it is becoming possible to advise as to 
fertiliser treatment which enables the plant to make the best of this rainfall. 
Unfortunately, few other Experimental Stations possess such complete 
masses of data as Rothamsted. Methods are now being devised, however, 
both by Fisher and by the able English investigator who modestly conceals 
his identity under the pseudonym ‘Student,’ for the study of smaller 
numbers of data, and it is hoped that these or others equally effective 
will be applied to the results of field experiments accumulated at various 
Experimental Stations throughout the world. A massed attack by a 
competent band of statisticians on the whole of the data of the best Experi- 
mental Stations, dealing with yields of crops under different conditions of 
nutrient supply, temperature, rainfall and other factors that go to make up 
_ the aggregate called season, would yield information of extraordinary value. 
_ Investigations of this kind, however, are necessarily slow, and they do 
not themselves afford complete information ; their value lies in the fact 
that they reduce a very complex problem to a set of single-factor problems 
of the type with which the scientific investigator is already familiar. 
In the meantime, while this work is proceeding, much is being done 
by observational methods. At Rothamsted the field plots are under 
continual observation by a group of three workers, a physiologist, an 
ecologist, and an agriculturist, who study such factors as rate or habit 
of growth, earliness of starting or maturing, degree of resistance to 
insect or fungus attack ; their observations are fully recorded and brought 
before the chemical, physical and botanical departments at regular and 
frequent intervals. Certain of the experiments are repeated at other 
centres on closely similar lines for purposes of comparison. In consequence 
our old field plots which have been studied for the past eighty years by 
Lawes, Gilbert, Warington, and Hall, and might have been supposed to 
have no further tales to tell, are found to be still yielding results of great 
interest in agricultural science and practice. 
—_— 
The Results Obtained : Alterations in the Plant. 
We shall begin with the results obtained by effecting alterations in the 
plant. Reference has already been made to the changes brought about 
by the plant breeder, and we need not stop to argue whether the great 
improvements in crops made in pre-Mendelian days by the Suttons and 
Findlay in potatoes, by Chevalier in barley, by the Gartons in oats, Vilmorin 
in sugar beet, and others, should be labelled empirical or scientific. There 
are certain other changes in plants, however, of a purely temporary nature, 
hia have been induced by changes in conditions. Itis a commonplace 
_ among farmers that certain soil conditions influence not only the yield 
te also the quality of crops. The leaf and root are more easily affected 
than the seed. Thecase of mangolds has beeninvestigated at Rothamsted ; 
- the sugar content of the root, an important factor in determining feeding 
value, was increased by increasing the supply of potassium to the crop. 
_ Middleton at Cockle Park showed “that grass increased in feeding value— 
: 
B 
