170 ANNUAL REPORT SMITHSONIAN INSTITUTION, 1^)28 



of a gas. But the hypothesis has proved disappointing, and a much 

 more probable hypothesis is, I think, that the atoms are not stripped 

 quite bare, but that in most stars they retain a few rings of electrons 

 which give the atoms so much size that they jostle one another about 

 like the molecules of a liquid. This hypothesis explains beautifully 

 the otherwise puzzling fact that stars of large mass fall into distinct 

 groups of what may almost be described as " standardized " sizes. 

 On the " liquid-star '' hypothesis, these different sizes correspond to 

 the different sizes possible for the stellar atoms which may have 0, 

 1, 2, or 3 rings of electrons left, but can not have fractional numbers. 

 The largest stars of all, such as Betelgeuse, have three rings left, 

 while minute stars, such as Van Maanen's star, consist of atoms most 

 of which are stripped quite bare, so that there is almost no limit to 

 the closeness with which they can be packed together. An average 

 handful of the matter of which this star is composed would contain 

 about 10 tons. 



Thus the observed sizes of the stars proclaim the secret of the 

 structure of the atom. The sizes of the stars are discontinuous 

 because the sizes of atoms broken down to different stages are dis- 

 continuous. These discontinuities can be traced in turn to the discon- 

 tinuities which form the central feature of the new quantum dy- 

 namics. Thus the distinguishing characteristic of the laws which 

 govern the most minute processes in nature is transmitted directly 

 into the large-scale phenomena of astronomy and governs the dis- 

 tribution of the huge masses of the stars. The infinitely great is 

 never very far from the infinitely small in science, but it would be 

 hard to find a more sensational illustration of the unity of science 

 than that I have just giA^en. 



On this hypothesis, not only do the observed sizes of the stars dis- 

 close the general structure of the atom, which is old knoAvledge, but 

 they also reveal the detailed structure of the particular atoms of 

 which the stars are composed, and this is new knowledge. To be 

 precise, the observed sizes of the stars disclose the atomic weights of 

 the stellar atoms; they indicate that the stellar atoms are probably 

 rather heavier than the heaviest atom, uranium, known on earth. 

 The atoms which reveal their presence in stellar spectra are, of course, 

 atoms of the ordinary terrestrial elements — hydrogen, iron, calcium, 

 and the like. These, being the lightest atoms in the star, must natu- 

 rally float up to its surface, and as the earth was originally formed 

 out of the surface of the sun the earth is necessarily composed of 

 them. But it now appears likely that down in the depths of the 

 stars are other unknown and heavier atoms. We may almost say 

 that it must be so, for no terrestrial atoms, not even radium or 

 uranium, can produce anything like the amount of energy which these 

 stellar atoms are observed to produce. 



