502 THE POPULAR SCIENCE MONTHLY. 



such a case could not arise, the water-supply being uniform and in ex- 

 cess of the needs of the plants. The same considerations apply to 

 other conditions, of course, though less markedly than to the water- 

 supply. In carefully conducted pot experiments it is possible to have 

 practically all the conditions controllable, while duplicate trials will 

 show the degree of accuracy obtained. 



"What, now, is the value of this method, as compared with properly 

 conducted field experiments, in the study of agricultural questions ? 

 Can it replace them either partially or wholly ? An intelligent reply 

 to these questions must distinguish between the various kinds of prob- 

 lems which present themselves for solution. 



In the first place, many purely scientific problems demand attention. 

 These are of the first importance, for^ until we can master them, all 

 attempts to apply science to practice will have but partial and uncer- 

 tain success. Such problems are, for example, the most suitable form 

 in which certain fertilizing substances may be applied (sulphate or 

 chloride of potassium, nitrates or ammonium salts, soluble or reverted 

 phosphoric acid, etc.), the effect of differing degrees of fineness, or of 

 a more or less uniform distribution at different depths in the soil, the 

 effect of different manurings upon the chemical composition and feed- 

 ing value of the plants produced, the specific needs of different plants 

 as regards fertilizers, etc., etc. 



Such problems as these can be solved only by scientific methods of 

 experiment, in which all the conditions are under control. Just as the 

 question, what substances are essential to plant-growth, was not solved 

 by field experiments, but by the method of water-culture, in which no 

 soil at all is used, so questions such as were just mentioned seem likely 

 to reach their solution by a method almost equally removed from the 

 conditions of practice. But while the method of pot experiments ap- 

 pears well adapted to resolve scientific questions, and while its results 

 (if reached legitimately, and tested carefully) are true independently 

 of any extraneous considerations, those results need to be tested under 

 actual working conditions ; not as to their truth — that is settled — but 

 as to their applicability to practice. It is true, as a scientific fact, that 

 certain varieties of feldspar contain several per cent of potash, and it 

 is also true that potash is an indispensable element of plant-food ; but 

 he who should therefore try to supply potash to his crops by means of 

 ground feldspar, would be likely to meet with very indifferent suc- 

 cess. He would not thereby disprove the fact that feldspar contains 

 potash, or that potash is indispensable to plants. He would simply 

 show that to these two facts there must be added some information as 

 to the availability of feldspathic potash as plant- food, and so his field 

 experiment would be the starting-point of a new series of scientific in- 

 vestigations, which should show whether the first-named facts were 

 capable of any useful application. 



The method of exact field experiments, then, as developed by 



