42 SECTIONAL ADDRESSES. 



can be regarded as complete, and many which have not been explored. 

 at all. Naturally, the spectra in which the regularities have not yet been 

 traced are those which present special experimental difficulties. The 

 spectra of some of the rare earths, for example, consist of a vast number 

 of lines, and the determination of their structure will be extremely 

 laborious, demanding in the first instance great accuracy in wave-length 

 measurements. On the other hand, boron presents a difficulty because 

 its lines in the ordinary region of observation are too few for complete 

 classification. Very little is known also of the structure of the spectra 

 of the halogen elements, and much work remains to be done on the 

 classification of the lines of some of the inert gases, in continuation of 

 Paschen's masterly analysis of the spectrum of neon. 



Furthermore, predictions which can be made remain to be tested ; 

 for example, the spectrum of doubly-ionised scandium, which, in opposition. 

 to earlier expectations, should not resemble that of neutral potassium, is. 

 as yet unknown. Results of great importance to theoretical progress with 

 respect to atom building may confidently be expected also from investiga- 

 tions of the spectra of other elements at successively higher stages of 

 ionisation, as witness the results already obtained for the spectra of 

 numerous elements in which the outermost shells of the atoms have been 

 reduced to a single valency electron. *i 



The present resources of experimental spectroscopy would appear ta 

 be adequate for the elucidation of the majority of the outstanding 

 problems. For most elements the conditions of excitation can be sa 

 modified that the spectrum is well under control, so that all the lines, or 

 only a selection of them, can be produced at will. In the discussion of 

 regularities it is, in fact, often required to excite the complete spectrum, 

 including the fainter lines, for the completion of multiplet groups. On 

 the other hand, as an aid to the determination of ground terms, one 

 desires to produce the smallest number of lines that an element can be 

 induced to give. 



The old, well-tried methods of exciting substances to luminosity — the 

 flame, arc, spark, and vacuum tubes — have by no means been superseded. 

 They provide the observer with a wide range of exciting energies, and 

 seem likely to continue long in use as standard methods applicable to 

 most of the elements. While the flame yields only the lines representing 

 combinations with the deepest terms of the spectrum, the spark with some 

 elements, such as silicon, is capable of exciting even trebly-ionised atoms, 

 and it is usually possible to sort out the lines associated with atoms at 

 different stages of ionisation by merely observing the extensions of the 

 lines from the tips of the electrodes. Experiments on the absorption 

 spectra of metallic vapours will no doubt also continue to be of effective 

 service in the identification of lines which originate in the normal atomic 

 energy levels, or in the verification of deductions as to the normal states 

 based upon analyses of the more complex emission spectra. 



The older methods of observation, however, have been supplemented 

 by numerous other experimental arrangements. Some of these, like the 



" A. Fowler, Phil. Trans., A, vol. 225, p. 1 (1925). MilUkan and Bowen, Phys. 

 Rev., Sept. 1924 ; Phil. Mag., May 1925. J. A. Carroll, Phil. Tram., A, vol. 225, 

 p. 357 (1925). D. R. Hartree, Roy. Soc. Proc, A, vol. 106, p. 552 (1924). 



