This may be a glorious though a remote prospect. But surely it is above all things needful that our science should be scientific. She must not blink facts, merely because they do not fit into her scheme or hypothesis of the nature of things, or of religion. She really must give as much prominence to the evidence which contradicts as to that which supports her theory in each instance. Not only must she not shut her eyes to this evidence, but she must diligently search for it, must seek for what Bacon calls _instantice contradictorim_, since, if these exist, the theory which ignores them is useless. If she advances an hypothesis, it must not be contradictory of the whole mass of human experience. If science finds that her hypothesis contradicts experience, she must seek for an hypothesis which is in accordance with experience, and, if that cannot be found, she must wait till it is found. Again, science must not pile one unverified hypothesis upon another unverified hypothesis till her edifice rivals the Tower of Babel. She must not make a conjecture on p. 35, and on p. 210 treat the conjecture as a fact. Because, if one story in the card-castle is destroyed by being proved impossible, all the other stories will 'come tumbling after.' It seems hardly necessary, but it is not superfluous, to add that, in her castle of hypotheses, one must not contradict, and therefore destroy, another. We must not be asked to believe that an event occurred at one date, and also that it occurred at another; or that an institution was both borrowed by a people at one period, and was also possessed, unborrowed, by the same people, at an earlier period. We cannot permit science to assure us that a certain fact was well known, and that the knowledge produced important consequences; while we are no less solemnly told that the fact was wholly unknown, whence it would seem that the results alleged to spring from the knowledge could not be produced. This kind of reasoning, with its inferring of inferences from other inferences, themselves inferred from conjectures as to the existence of facts of which no proof is adduced, must be called superstitious rather than scientific. The results may be interesting, but they are the reverse of science. It is perhaps chiefly in the nascent science of the anthropological study of institutions, and above all of religion, that this kind of reasoning prevails. The topic attracts ingenious and curious minds. System after system has been constructed, unstinted in material, elegant in aspect, has been launched, and has been wrecked, or been drifted by the careless winds to the forlorn shore where Bryant's ark, with all its crew, divine or human, lies in decay. No mortal student believes in the arkite system of Bryant, though his ark, on the match-boxes of Messrs. Bryant and May, perhaps denotes loyalty to the ancestral idea.