Appendix B

THE most interesting event in astrophysics since the first edition of this book was published is the identification of nebulium by I. S. Bowen in the autumn of 1927. We said (p. 55): 'Nebulium is not a new element. It is some quite familiar element which we cannot identify because it has lost several of its electrons.' It turns out that nebulium is oxygen which has lost two electrons. Nevertheless there are unforeseen circumstances connected with this identification. It is not difficult in the laboratory to remove two electrons from an oxygen atom and so obtain a nebulium atom ; but in the laboratory this does not emit nebulium light. Our failure to imitate nebulium light artificially was not due to inability to batter atoms sufficiently powerfully; it was because we were unable to leave them sufficiently at peace.

In the lectures reference was made to a race between the experimental and the theoretical physicists to solve the mystery of nebulium; success was actually reached by a happy combination of the work of both. To understand this we must first remark that when the positions of a few lines in a spectrum have been measured experimentally the positions of some more lines can be calculated by a simple and exact rule. Consider for example three states of an atom; the rule is that the frequency of the spectral line emitted in a transition from state 3 to state 1 is the sum of the frequencies of the lines emitted in the transitions from state 3 to state 2 and from state 2 to state 1. Similarly if there are 10 states of the atom it is sufficient to measure experimentally the frequencies of 9 spectral lines; we can then, by repeated applications of the theoretical rule, compute the whole spectrum which consists of 45 lines corresponding to the 45 possible pairs of states between which a transition might occur. It is the ambition of theoretical physics to compute the spectrum purely from a knowledge of the electronic constitution of the atom without any help from observation, but up to the present this has only been done for the very simplest type of atom; less ambitiously, however, theoretical physics can take a spectrum which has been partially observed and compute the remainder of it. The light which comes to us from oxygen ions in the remote nebulae does not belong to any part of the oxygen spectrum known experimentally ; nevertheless its source is certainly oxygen because it is part of the oxygen spectrum as completed in the way we have indicated.

You may take a horse to the water but you cannot make it drink. In the laboratory we can produce doubly ionized oxygen atoms and bring them into state 2, but we cannot persuade them to drop back into state 1. By theory we know precisely the kind of light they would emit in the transition; and now we recognize that the atoms are freely giving this light in the nebulae although they will not do it at our command. With us the atoms will not jump either way between state 1 and state 2; they always go round via states 3 or 4. There are always a number of lines of the completed spectrum which are missing experimentally because the atom refuses the corresponding jump; there is in fact a selection rule which enables us to predict the missing or 'forbidden lines'. It appears, however, that the prohibition has reference to the circumstances of terrestrial experiment and does not apply universally.

When an atom is excited, we may picture its electron as a guest in an upper story of an old-fashioned hotel with many alternative and interlacing staircases; he has to make his way down to the lounge -- the normal unexcited condition. There are many different ways down, but there is not always a direct way from one landing to another. Sometimes the guest comes to a cul-de-sac from which there is no way leading downwards, and the only thing to do is to ascend to a higher landing and try a new descent. The cul-de-sac corresponds to a state of the atom which is not the lowest, but from which there is no unforbidden transition to a lower state. Such a state is called metastable. There are plenty of unforbidden passages upwards, but to take any of these the atom would need to be supplied with energy from external sources; spontaneously it can only go downwards, and from a metastable state all downward passages are forbidden. There are three ways in which it may be released from the cul-de-sac. Firstly, by absorbing light it may obtain the energy needed to ascend to a higher state from which it can try another way down with perhaps better luck next time. Secondly, when collision with another atom or free electron occurs the usual rules are suspended; the guest, so to speak, finds himself transported to the lounge by an earthquake. But in this case the light corresponding to the transition is not emitted, the energy being got rid of in another way. Finally, if after waiting a long time release does not come by the first two methods, the electron will venture on the forbidden passage and the forbidden line of the spectrum will be emitted. 'Emergency passage' would perhaps be a better description than 'forbidden passage'.