doubt, with greater effort and constraint, but still not altogether impracticable. But the artificial and mutilative character of such experimentation is felt here in a way different from its pressure in other cases. And what is more important, to institute an experiment or set on foot a scientific observation (and to observe means to _watch_ a definitely restricted natural process with a view to answer some question about it), presupposes--as we have already seen--a tolerably definite provisional theory as to the general lie of the country to be investigated. Only when the country has been reasonably well mapped out in provinces and provided with some system of roads, can these problems of detail--questions to be answered Yes or No--be profitably put. And it is--in some parts of the historical sciences at least--somewhat premature to put questions requiring a categorical reply. There is only the vague _malaise_ of felt difficulty to guide us. We do not, in many cases, know what it is that we want to know; for, it demands a good deal of wisdom and trained art to put the proper or reasonable question,--so much so, indeed, that to succeed in formulating your question fully is equivalent or nearly equivalent to being able to answer it. The value of observations and experiments--which are ways of putting nature to the question and it may be to the torture--depends entirely upon the knowledge and the command of general ideas possessed by the observer and experimenter. And the same may be said of the reduced and tabulated conspectuses of the results of many observations and experiments which are called Statistics. Their value depends on the truth and breadth of view which presided at their collection and arrangement[5].