We have seen how the stars shew as great a range of candle-power as there is between a glow-worm and a searchlight; while their range of size is as that between a speck of dust and a motor-car. The range in their weights is much smaller, but still it is about equal to that between a feather and a football. And in every respect the sun is somewhere about average. It could hardly be expected to strike the exact happy mean in every way, but it never misses it badly. To put the same thing in another and less complimentary way, the sun is totally undistinguished in all respects—in weight, in size, in temperature and in candle-power.

Clearly, however, we get very little knowledge of the general nature of the stars from a mere mention of extremes and of one average star. We should not know much about the English population if we had only been told the heights and weights of the shortest dwarf and the tallest man, and that a particular man 5 feet 9 inches high was a good average Englishman in all respects. We want a more detailed knowledge as to the classification of the stars by size, candle-power and weight.

Suppose that all the entrants to a dog-show broke loose and ate their labels, and had to be reclassified.