We generally think of interstellar space as excessively cold. It is quite true that any thermometer placed there would show a temperature only about 3° above the absolute zero--if it were capable of registering so low a reading. Compact matter such as a thermometer, or even matter which from the ordinary standpoint is regarded as highly diffuse, falls to this low temperature. But the rule does not apply to matter as rarefied as the interstellar cloud. Its temperature is governed by other considerations, and it will probably be not much below the surface-temperature of the hottest stars, say 15,000°. Interstellar space is at the same time excessively cold and decidedly hot.[21] _The Sun’s Chromosphere_ Once again we shift the scene, and now we are back in the outer parts of the sun. Fig. 10[22] shows one of the huge prominence flames which from time to time shoot out of the sun. The flame in this picture was about 120,000 miles high. It went through great changes of form and disappeared in not much more than twenty-four hours. This was rather an exceptional specimen. Smaller flames occur commonly enough; it seems that the curious black marks in Fig. 1, often looking like rifts, are really prominences seen in projection against the still brighter background of the sun. The flames consist of calcium, hydrogen, and several other elements. We are concerned not so much with the prominences as with the layer from which they spring. The ordinary atmosphere of the sun terminates rather abruptly, but above it there is a deep though very rarefied layer called the chromosphere consisting of a few selected elements which are able to float--float, not on the top of the sun’s atmosphere, but on the _sunbeams_. The art of riding a sunbeam is evidently rather difficult, because only a few of the elements have the necessary skill. The most expert is calcium. The light and nimble hydrogen atom is fairly good at it, but the ponderous calcium atom does it best. The layer of calcium suspended on the sunlight is at least 5,000 miles thick. We can observe it best when the main part of the sun is hidden by the moon in an eclipse; but the spectroheliograph enables us to study it to some extent without an eclipse. [Illustration: Fig. 10. SOLAR PROMINENCE] [Illustration: Fig. 11. STAR CLUSTER ω CENTAURI] On the whole it is steady and quiescent, although, as the prominence flames show, it is liable to be blown sky-high by violent outbursts. The conclusions about the calcium chromosphere that I am going to describe rest on a series of remarkable researches by Professor Milne.