courts of justice the minds of jurymen often become perplexed with doubts which they can not account for, or else they yield a too easy credence to the guilt of the accused when the question of guilt depends upon circumstantial evidence. I shall not spend much time in contending that these rules of evidence must be applied to scientific investigations which are to affect our belief in such a proposition as the descent of man from a common ancestor with the monkey. This is not only an hypothesis depending upon circumstantial evidence, but it is professedly a deduction from a great range of facts and from a very complex state of facts. In reasoning upon such subjects, when the facts which constitute the chain of circumstantial evidence are very numerous, we are apt to regard their greater comparative number as if it dispensed with a rigid application of the rules of determination. Every one can see, in the illustration above employed, borrowed from criminal jurisprudence, that the facts which constitute the chain of circumstantial evidence ought to be rigidly tested by the rules of determination before the guilt of the accused can be safely drawn as a deduction from the facts. But, in reasoning from physical facts to any given physical hypothesis where the facts are very numerous, there is a strong tendency to relax the rules of evidence, because, the greater the accumulation of supposed facts becomes, the greater is the danger of placing in the chain of evidence something that is not proved, and thus of vitiating the whole process. To this tendency, which I have observed to be very frequent among scientists, I should apply, without meaning any disrespect, the term invention. A great accumulation of facts is made, following one another in a certain order; all those which precede a certain intermediate link are perhaps duly and independently proved, and the same may be the case with those which follow that link. But there is no proof of the fact that constitutes the link and makes a complete chain of evidence. This vacuity of proof, if one may use such an expression, is constantly occurring in the writings of naturalists, and is often candidly admitted. It is gotten over by reasoning from the antecedent and the subsequent facts that the intermediate facts must have existed; and then the reasoning goes on to draw the inference of the principal hypothesis from a chain of proof in which a necessary intermediate link is itself a mere inference from facts which may be just as consistent with the non-existence as with the existence of the supposed intermediate link. In such cases we are often told very frankly that no one has yet discovered that the intermediate link ever actually existed; that the researches of science have not yet reached demonstrative proof of the existence of a certain intermediate animal or vegetable organization; that geological exploration has not yet revealed to us all the specimens of the animal or vegetable kingdoms