FINAL CONSIDEEATIONS 309 



comparatively minor order, which were overlooked or not 

 suspected when fertiliser actions first began to be studied, 

 because in most cases the agent in the process is that part of 

 the substance which possesses no value as a fertiliser. For 

 example, sulphate of ammonia was considered as a source of 

 nitrogen only, the sulphuric acid it contained was entirely 

 ignored and regarded as of no account. Similarly with nitrate 

 of soda, the nitrogen is the important part upon which its value 

 as a fertiliser depends ; the mistake came in supposing that 

 the soda was entirely without effect. The same state of affairs 

 has occurred over and over again in the history of science ; the 

 broad conclusions reached by early generations of investigators, 

 which become the staple of the text-books and the dogma of 

 the lecture rooms, and in the process always grow cruder and 

 more hard and fast in statement than is justified by the original 

 researches, prove eventually to be no more than first approxima- 

 tions to the truth. To complete the story, a second, sometimes 

 even a third, term requires to be introduced, the course of 

 events in nature being always much more complex than the 

 nice water-tight statements which our minds like to evolve 

 under the guise of laws. These second approximations, which 

 may become large enough to override the main truth, often 

 make themselves evident to the practical man, who delights in 

 them as proofs that theory and practice do not square, though 

 as theory can never be more than a method of explaining and, 

 in its turn, predicting the practice, any want of agreement 

 between the two must only mean that the practical man is 

 dealing with an imperfect theory. 



However, it is the duty of the scientific man to recognise 

 that practical affairs will always be stretching the range of 

 actions upon which he founded his general statements into 

 regions where they will no longer fit the facts, and by picking 

 up the hints of such discrepancies as the practical man can 

 often supply, the theory may be founded on a more accurate 

 basis. 



