directly applicable to practice. It is not to be wondered at, then, that in a time when agricultural experimentation is attracting attention as never before, field experiments should be multiplied on every hand, and that the public should regard the making of them as one of the chief ends, if not the chief end, of experiment stations and experimental farms. Both in Europe and America a vast deal of time and money has been devoted to their execution, and not only have organizations undertaken them, but private farmers have been urged to experiment on their own farms, both with a view to obtaining a better knowledge of the needs and capacities of their soils, and in the hope of advancing the science of agriculture.
The literature of the subject is voluminous, particularly in relation to the use of manures and fertilizers, and it might be expected that by this time our knowledge of these matters would be tolerably complete.
When, however, we come to look for the results of all this work, we find them surprisingly meager in comparison with the expenditure of time and labor which they have cost. That many valuable results have been reached goes without saying; but relatively their number is small, while the number of uncertainties and contradictions is remarkably great. The weakest portion of agricultural chemistry is that relating to fertilizers and manures; that is, precisely that part which we should expect to find well developed.
This state of affairs could not fail to impress thoughtful students of agricultural science, and cause them to seek out the reason why a method, which is apparently based upon a correct principle, and which has been executed with so much labor and care, has yielded, on the whole, such unsatisfactory returns.
Quite recently two German investigators, Professor Paul Wagner in Darmstadt, and Professor G. Drechsler in Göttingen, have given especial attention to this question, and have reached some interesting and important results, a brief account of which may not be uninstructive at a time when such general attention is being given to agricultural experimentation. These two experimenters have worked quite independently of each other, and their substantial agreement is strong evidence of the correctness of their conclusions.
We have said that the fundamental idea of the field experiment is essentially scientific; but, while this is true, a more critical examination shows that the way in which this idea has been carried into execution has been far from scientific. The scientific method of experiment requires two things: 1. All the conditions of the experiment must be identical, with the exception of the one whose action is to be tested, and that must vary to a known extent. 2. The limits of error of the methods of weighing, measuring, etc., used, must be known, to the end that we may know whether any difference which may be observed is accidental or significant.
