with; we are practically dealing with a cooling globe of which the exterior absorbing layers consist of hydrogen, iron, magnesium, and sodium. And now perhaps it will be obvious why I was anxious in this general statement to begin as near as I could at the beginning of things. It is only by going back in that way that it is possible to explain this enormous development of hydrogen in the hottest stars. We saw that first one or perhaps two unknown substances—together with hydrogen, carbon, magnesium, manganese, lead, and iron—wrote their record in the spectrum, and that finally hydrogen was present in excess in the hottest stars. By the phenomena of comets it has been demonstrated that the radiant energy of our sun, and therefore the radiant energy of all other masses of equal temperature to our sun, drives, in all probability, everything of the nature of a permanent gas, like hydrogen or carbon compounds, away from the center of the system. Thus we may possibly explain the absence of oxygen and carbon from the sun; but hydrogen is present. The unknown substance or substances are concerned in most of the actions which take place in the hottest parts of the sun, and they are always associated with hydrogen. In the atmospheres of the hottest stars, again, hydrogen is enormously developed. Now that hydrogen, we have reason to believe, can not have passed the cordon to which I referred. The only supposition is that it and the unknown substances have as such been produced by the dissociation of the chemical elements of which the meteoritic particles which have formed the star in the manner I have indicated are composed. Here, then, we have a series of facts which add very great probability to the idea which has been arrived at on other grounds, that the chemical elements themselves are forms of hydrogen, or have a common origin.
On the right-hand part of the temperature curve the hottest state of things is represented at the top and the coolest at the bottom, and we pass through groups IV, V, and VI. As the temperature runs down, the hydrogen gradually disappears; as this happens in a mass of gas, the temperature of which is gradually but constantly reduced, we can only suppose that it is used to form something else. We get association due to reduced temperature in the same way that we get dissociation due to increasing temperature. The sun is a star just about half-way down the descending side of the curve; we know on other grounds that the sun is cooling.
The next part of the story is this: with decreasing hydrogen we get gradually associated an increasing quantity of the metallic elements (group V), and subsequently of carbon; but now the carbon vapors are absorbing, they are not radiating—in other words, the spectrum includes dark bands instead of bright ones, as they