"Assuming that the number of things in the universe is infinite, and assuming all observed A's are B's, there will be an infinite number of classes in which all A's are B's as well as of classes in which not all A's are B's. It would seem to follow that, however many A's have been observed to be B's and none have been observed to be not B's, it is even chances that all A's are B's. This does not invalidate the conclusion that induction by simple enumeration is invalid."