Simply stated, the difference between these two examples does not consist in the existence of a relationship in the one case and the absence of a relationship in the other; it consists in the fact that in the case of the flowers the relationship occurs now and then but is not permanently knowable. It is possible that there is a natural law controlling the relation between color and odor, and if that law were known there would be no question of accident or of analogy, but of law. Our ignorance of such a law, in spite of the multiplicity of instances, lies in the fact that we are concerned only with the converse relationships and not with the common cause of perfume and color. Suppose I see on the street a large number of people with winter over-coats and a large number of people with skates in their hands, I would hardly ask whether the coats are conditioned or brought out by the skates or the skates by the coats. If I do not conclude that the cold weather is the condition both of the need of over-coats and the utility of skates, I will suppose that there is some unintelligible reflexive relation between over-coats and skates. If I observe that on a certain day every week there regularly appear many well-dressed people and no workingmen on the street, if I am ignorant of the fact that Sunday is the cause of the appearance of the one and the disappearance of the other, I shall try in vain to find out how it happens that the working people are crowded out by the well-dressed ones or conversely.