But, then, in any given case, whether we ought to yield our belief in anything where we have only circumstantial evidence to guide us, there are certain rules to be observed. The first of these rules is, that every fact in a collection of proofs from which we are to draw a certain inference must be proved independently by direct evidence, and must not be itself a deduction from some other fact. This is the first step in the process of arranging a chain of moral evidence. There is a maxim in this branch of the law of evidence that you can not draw an inference from an inference. In other words, you can not infer a fact from some other fact, and then unite the former with two or more independent facts to make a chain of proofs. Every link in the chain must have its separate existence, and its existence must be established by the same kind and degree of evidence as if it were the only thing to be proved. The next rule is to place the several facts, when so proved, in their proper relation to each other in the group from which the inference is to be drawn. In circumstantial evidence a fact may be established by the most direct and satisfactory proofs, and yet it may have no relation to other facts with which you attempt to associate it. For example, suppose it to be proved that A on a certain occasion bought a certain poison, and that soon after B died of that kind of poison; but it does not appear that A and B were ever seen together, or stood in any relation to each other. The fact that A bought poison would have no proper relation to the other fact that B died of that kind of poison. But introduce by independent evidence the third fact, that A knew B intimately, and then add the fourth fact, that A had a special motive for wishing B's death, you have some ground for believing that A poisoned B, although no human eye ever saw the poison administered. From this correlation of all the facts in a body of circumstantial evidence, there follows a third rule, namely, that the whole collection of facts, in order to justify the inference sought to be drawn from them, must be consistent with that inference. Thus, the four facts above supposed are entirely consistent with the hypothesis that A poisoned B. But leave out the two intermediate facts, or leave out the last one, and B might as well have been poisoned by C as by A. Hence there is a fourth rule: that the collection of facts from which an inference is to be drawn must not only be consistent with the probable truth of that inference, but they must exclude the probable truth of any other inference. Thus, not only must it be shown that A bought poison, that B died of poison, that A was intimate with B and had a motive for wishing B's death, but, to justify a belief in A's guilt, the motive ought to be shown to have been so strong as to exclude the moral probability that B was poisoned by some one else, or poisoned himself. It is in the application of these rules that in