Relations are a class of general ideas familiar to savage thought, often appreciated with great subtlety and especially prominent in some magical operations. Relations are not only thought of by savages but compared, and likenesses discovered between them that may often surprise us. A gardener of New Guinea, having planted taro, ensured the growth of the crop by saying: “A muræna, left on the shore by the tide, was exhausted and on the point of expiring, when the tide returned, and it revived and swam away.” And he struck the ground with a branch three times.[98] He saw that renewal of life in the muræna and in the taro were the same thing; so that to describe one must strengthen the other—such is the “force” of a spell! The most important relation involved in knowledge and in its practical applications is causation. Savages who have no word for causation and have never thought of it in the abstract, must always act as if they assume it. This is apt to be misunderstood. It is written: “The natives [of Australia] have no idea of cause and effect. They notice that two things occur one after the other, and at once jump to the conclusion that one is cause and the other effect.” They have, then, some idea of the relation, though ill-discriminated. Thus, having noticed that the plover often cries before rain, they imitate the cry when performing rites to bring on rain. We rather suppose that an atmospheric change, preceding the approaching rain, excites the plovers. But that does not occur to the natives: that “one after the other” is the same as effect after cause, or (as Logicians say) _post hoc, ergo propter hoc_, lies at the bottom of innumerable superstitions. The subconscious control exercised by this latent form of thought is very imperfect. The maxims of indirect Magic, that a likeness may be substituted for the thing itself, or a part of it for the whole, are merely formulæ of causation assumed under an erroneous belief as to what can be a cause. There are cases in which the like may be a substitute (one man for another) and a part may nearly serve for the whole (part of a broken knife). Hence these principles were stated and avowed by physicians and alchemists of the Middle Ages, and observed in their practice.