To deal with molecules in a book on the behavior of the chemical elements seems a little bit awkward, since the information on molecules seems apparently unsuitable to add to our knowledge of the behavior of the elements. There are, however, several good reasons for dealing with molecules. The first is that, for some elements, we only know their existence through the molecules in which they participate. This is for instance the case for fluorine. In the second place, molecules permit easy study of the different isotopes of an element. As an example consider C, whose isotopes were first studied through the molecular bands of C12/C12 and C12/C13 (Sanford 1940). The study of isotopes is an extremely fruitful subject, which is by no means exhausted. In the third place, molecules are very sensitive indicators of spectral peculiarities. For instance the first indications of the non-uniform composition of stellar atmospheres came through the study of molecules in late type stars. It was found that the division of these stars into two families, one with carbonated molecules (then called R- and N-type stars) and the other group with metallic oxides (the M- and S-type stars) could not be explained by a uniform chemical composition. That such an abundance effect is more easily observed in molecules rather than in atoms or ions, is due to the fact that molecules are very sensitive to differences in chemical composition.

These three arguments largely justify the existence of a separate section on molecules. Another independent reason is that we have been unable to find an updated summary on molecules in stars.