be able to discern here and there a stream of molten iron issuing from a blast furnace, or perhaps to witness the operation of the forging of an anchor under the steam-hammer, to watch the rolling of the plates for an armour-clad, or to see the more humble operations of the blacksmith or nail-maker. But he would surely form an entirely erroneous impression as to the quantity of iron on this earth, or as to the extent in which it was employed in the varied purposes of the arts, if he concluded that there was no iron on our globe at all except that which happened at the moment to be in that particular incandescent state in which alone it was visible to him. If he were gifted with reasoning powers he would say, "It is quite true that I can only see the iron while it is red-hot, but I know that for iron to be red-hot on the earth's surface is an exceptional and abnormal condition of a very temporary or intermittent character. No doubt, every piece of iron may have to be red-hot once, or more than once, but the total duration of such phases of incandescence is quite insignificant under ordinary circumstances when compared with the periods in which the iron is cold and invisible. I, therefore, cannot refuse to believe that there must be an amount of iron on the earth which I do not see, but which bears a proportion to that which I do see in the ratio of thousands or millions to one."
Precisely similar is the way in which the astronomer who is properly familiar with the theory of probabilities will approach the study of the stars. He will reflect that each mass of matter must be cold and invisible for by far the greater part of the period of its existence; he will reflect that on rare occasions, separated by intervals
