XIII.] EXPERIMENTAL WORK B V FARMERS 375 



of the plants as regards disease, insect attacks, etc., and 

 in the texture of the soil, from having due weight. In 

 pot work the artificiality of the conditions is increased ; 

 pot experiments are only of value to the investigator in 

 clearing up the earlier stages of an enquiry before the 

 applications to practice begin to be considered. 

 Deductions from pot experiments with regard to field 

 work must be drawn with great caution and always 

 regarded with suspicion. 



It will follow from what has been said about the care 

 with which field experiments are to be conducted and 

 the large margin of error inherent in their results even 

 under favourable conditions, that they are hardly to be 

 lightly entered upon by the ordinary busy farmer, and 

 that the advice so often given to him to work out by 

 experiment the manures best suited to his own farm 

 would really involve a disproportionate amount of work. 

 Single-handed it will take too long and cost too much 

 to arrive at accurate results ; the farmer's experimenting 

 should be done as part of a large co-operative local trial 

 designed to establish the characteristics of the soil on 

 which he is working, in regard to its customary cropping. 

 Even the usual habit of testing a fertiliser by leaving 

 one breadth of the field unmanured may often deceive ; 

 concentrated nitrogenous manures easily show up under 

 such conditions, because they affect the colour and vege- 

 tative appearances of the crop so markedly ; basic slag, 

 again, effects an extraordinary change in the appearance 

 of some pastures ; but, speaking generally, the effects of 

 the mineral and some mixed manures are only to be 

 seen in the yield. Experience shows that it is difficult 

 to detect by eye the difference between two adjoin- 

 ing plots when the yield of one is 20 per cent, above 

 the other; differences like 10 per cent, will always go 

 unperceived. In estimating the yield of root crops the 



