Page:Experimental researches in chemistry and.djvu/496

1855.] analogies; to present the fundamental idea in every form, proportion, and condition; to clothe it with suppositions and probabilities, —that all cases may pass in review, and be touched, if needful, by the Ithuriel spear of experiment. But all this must be under government, and the result must not be given to society until the judgment, educated by the process itself, has been exercised upon it. Let us construct our hypotheses for an hour, or a day, or for year; they are of the utmost value in the elimination of truth, 'which is evolved more freely from error than from confusion' but, above all things, let us not cease to be aware of the temptation they offer; or, because they gradually become familiar to us, accept them as established. We could not reason about electricity without thinking of it as a fluid, or a vibration, or some other existent state or form. We should give up half our advantage in the consideration of heat if we refused to consider it as a principle, or a state of motion. We could scarcely touch such subjects by experiment, and we should make no progress in their practical application without hypothesis; still it is absolutely necessary that We should learn to doubt the conditions we assume, and acknowledge we are uncertain, whether heat and electricity are vibrations or substances, or either.

When the different data required are in our possession, and we have- succeeded in forming a clear idea of each, the mind should be instructed to balance them one against another, and not suffered carelessly to hasten to a conclusion. This reserve is most essential; and it is especially needful that the reasons which are adverse to our expectations or our desires should be carefully attended to. We often receive truth from unpleasant sources; we often have reason to accept unpalatable truths. We are never freely willing to admit information having this unpleasant character, and it requires much' self-control in this respect, to preserve us even in a moderate degree from errors. I suppose there is scarcely one investigator in original research who has not felt the temptation to disregard the reasons and results which are against his views. I acknowledge that I have experienced it very often, and will not pretend to say that I have yet learned on all occasions to avoid the error. When a bar of bismuth or phosphorus is placed between the poles of a powerful magnet, it is drawn into a position across the line