First, let us bear in mind that anthropologists are apt to have their favourite theories—conclusions, that is, which are the legitimate result of reasoning inductively on the class of facts which they have more particularly studied. Thus Mannhardt had his theory of the Vegetation-spirit, Robertson Smith that of the sacramental meal, Usener that of the Sondergötter, Dr. Frazer that of divine Kingship; all of which are perfectly sound conclusions based on facts which no one disputes. They have been of the greatest value to anthropological research; but when they are applied to the explanation of Roman practices we should be instantly on our guard, ready indeed to welcome any glint of light that we may get from them, but most carefully critical and even suspicious of their application to other phenomena than those which originally suggested them. It is in the nature of man as a researcher, when he has found a key, to hasten to apply it to all the doors he can find, and sometimes, it must be said, to use violence in the application; and though the greatest masters of the science will rarely try to force the lock, they will use so much gentle persuasion as sometimes to make us fancy that they have unfastened it. All such attempts have their value, but it behoves us to be cautious in accepting them. The application by Mannhardt of the theory of the Vegetation-spirit to certain Roman problems, e.g. to that of the Lupercalia,19 and the October horse,20 must be allowed, fascinating as it was, to have failed in the main. The application by Dr. Frazer of the theory of divine Kingship to the early religious history of Rome, is still sub judice, and calls for most careful and discriminating criticism.21