THE SOIL AND ITS EFFECT ON PLANT GROWTH. 21 



order that their experiments may furnish information which will 

 stand the fire of criticism, and he can further, with mathematical 

 accuracy, detect in their results any flaws or inconsistencies in 

 the methods of experiment adopted. The Statistical Department 

 at Rothamsted has not long been in existence, and this account 

 of its work will therefore be mainly an indication of the general 

 lines of investigation being followed, and of the objects which 

 are more immediately in view. In the first place, a comprehen- 

 sive study is being made of the causes of variation in the yield 

 of crops. Every farmer knows that considerable variations are 

 due to season and weather conditions, but nobody has yet made 

 any attempt to analyse exactly to what extent these and other 

 influences operate. The yield of crops is affected by a number 

 of factors, as, for example, manuring, weather, variety, soil 

 population, weeds, physical condition of soil (tilth), diseases, 

 and so on. The relative importance of these factors is not known, 

 nor is it certain that other unknown factors are not of importance. 

 In the investigation of the subject a beginning has been made on 

 the records of crop yields collected at Rothamsted. This is by 

 far the longest and most complete body of material in existence, 

 and from its study it is expected that certain general principles will 

 emerge, in the light of which the shorter series collected at other 

 experimental centres may be interpreted. The study of the results 

 obtained during the past 70 years on the Broadbalk Wheat plots 

 has shown that in most cases the weather is responsible for any- 

 thing up to 75 per cent, of the variation in jdeld from year to 

 year. But weather is itself a combination of many factors, and a 

 detailed study of the effect of the various factors which contribute 

 to weather is at present in progress, based, in the first instance, 

 on the daily records of rainfall, which date back to February, 1853. 

 As a result of this investigation, it is expected that provisional 

 answers will be found to such questions as the following : — How 

 can the season, in so far as its effect upon crop yield is concerned, 

 be deduced from weather records as taken at the present time ? 

 What is the best season from the farmer's point of view, and what 

 are the critical periods of the year ? With what accuracy can 

 the crop be predicted from weather records alone ? How does 

 manuring affect the influence of weather upon the crop ? How 

 can a farmer make the most of a good season or mitigate the 

 adverse effects of a bad one ? How should manurial treatment 

 suited to one type of climate be modified to suit another ? 



The Physical Properties of the Soil. 

 We turn now to consider the investigations in progress bearing 

 on the second main problem which the Institute at Rothamsted 



