August 24, 1900.] 



SCIENCE. 



295 



as soon as the experimental study was ex- 

 tended that no such general law could be 

 laid down. For the same region in the 

 spectrum of an element the separation may 

 vary from apparently nothing up to an 

 Angstrom unit or more for a moderately 

 intense field. 



The high value of ejin indicated by 

 this phenomenon is significant. It is 

 of about the same order as that found 

 for the cathode ray particles and the ions 

 caused by uranium and Rontgen rays and 

 ultraviolet light ; but the value for electro- 

 lytic ions is only about 400. 



Zeeman's experiments were soon repeated 

 by other investigators, including Lodge, 

 Michelson, Preston and Cornu; and it was 

 not long before magnetic fields were used 

 of strength sufficient to fully resolve the 

 several components. It then became known 

 that the phenomenon was not nearly so 

 simple as the first observations would in- 

 dicate. Lodge* first noticed indications of 

 a quadruplet in the case of the i)-lines in- 

 stead of a triplet, and later Prestonf and 

 Cornu I observed unmistakable quadruplets 

 both in the case of D^ and in the spectra of 

 cadmium and magnesium. A little later 

 Becquerel and Deslandres§ discovered in 

 the iron spectrum a new type of triplet in 

 which the states of polarization of the inner 

 and outer components were interchanged. 

 In February, 1898, Michelson published in 

 the Astrophysieal Journal a paper giving re- 

 sults obtained with the interferometer, 

 some of which are not in accord with those 

 obtained before and since then with the 

 grating. Among other things he said that 

 all lines are divided into what may pro- 

 visionally be called triplets of approxi- 

 mately the same width, each member of a 

 so-called 'triplet,' however, being itself 



*The Electrician, June 18, 1897. 

 fProc. Soy. Soc, vol. 63, p. 26. 

 tComptes Bendus, vol. 126, p. 181. 

 ICompies Bendus, April, 4, 1898, p. 997. 



complex, making the whole magnetic group 

 formed from a single natural line quite 

 complicated. As has already been stated, 

 other investigators have found the degree 

 of separation to vary quite strikingly even 

 for neighboring lines, and although re- 

 searches with the grating have discovered 

 many complicated lines, the number of 

 these is very small compared to those that 

 appear as simple triplets. Professor Mich- 

 elson maintains in defense of his methods 

 that the resolving power of a grating is not 

 sufficient to reveal the finer structure of 

 the line as indicated by the interferometer. 

 This last is no doubt true, but on the other 

 . hand the interferometer method is exceed- 

 ingly indirect, and one hesitates before 

 accepting conclusions drawn from an esti- 

 mated visibility-curve as to the distribu- 

 tion of intensity in such a complicated 

 source as he advocates. In any case, the 

 assumption is involved that the source is 

 symmetrical, and this certainly is not al- 

 ways true. An example is found in the 

 cadmium group 4678, 4800, 5086, and the 

 similar group in the spectrum of zinc. 

 Each of these lines in the spark-spectrum 

 shows a decided shading on the red side, 

 which is retained by each component when 

 they are separated by the magnetic field, 

 making the whole group quite asymmetri- 

 cal. It has also been shown* that many 

 triplets and quadruplets are asymmetrical 

 in separation. It is quite possible that 

 such cases may account for some of Pro- 

 fessor Michelson's results. 



It cannot be denied, however, that, al- 

 though most lines become simple triplets 

 in the field, many are more complicated 

 than the simple theory would indicate, 

 many being fourfold and some at least six- 

 fold, while some seem not to be affected by 

 the field. Several theories have been de- 



*Zeemaii, Proc. Boy. Amst. Acad. Sei., Dee. 30, 

 1899. Eeese, J. H. TJ. Circulars, June, 1899 ; June, 

 1900. 



