WOOD. — SPECTRA OF SODIUM VAPOR. 251 



was found impossible to illuminate the vapor with light from but one 

 of them, and have at the same time sufficient illumination to excite 

 much fluorescence. I have not yet found much evidence of regularity 

 in the distribution of the lines in this case, though there is undoubted 

 evidence of two series in the immediate vicinity of the exciting lines. 

 I have indicated with the letters A and B the lines which appear to 

 belong together (compare with helium excitation). We have an 

 enormous number of lines in the yellow-green region, since we have a 

 double stimulation at the opposite end of the spectrum. There seems 

 to be some regularity here, but it is difficult to say which lines belong 

 together. 



Bismuth Excitation. 



The light of the bismuth arc makes a beautiful stimulus for the 

 fluorescence, since it contains but a single operative line, the strong 

 one at 4724. It gives rise to a very regular series in the blue-violet 

 region, the lines appearing to fall midway between the lines of the 

 third and fourth magnetic series (Plate 2, e, and chart). Though the 



o 



wave-length of the exciting line is only two Angstrom units longer than 

 that of the zinc line 4722, the spacing of the series in the two cases is 

 quite different. The same thing has been noticed in the case of the 

 shortest cadmium and zinc lines, which makes it seem possible that in- 

 teresting results may be obtained by altering the wave-length of the 

 exciting line, either by pressure or a powerful magnetic field. Exper- 

 iments in this direction will be made next winter. 



In both of these cases, in each of which we have excitation by lines 

 of nearly the same wave-length, the wider-spaced series is produced 

 when the stimulation is by the longer wave-length. It remains to be 

 determined whether we take hold of different absorption bands and 

 excite entirely different series, or whether we stimulate the same vi- 

 brator in each case, the spacing of the resulting lines depending upon 

 how nearly we approach its natural period in our exciting vibrations. 



In addition to the regularly spaced lines in the violet, we have a 

 complex assortment of lines in the yellow-green region, the intervening 

 portion being totally devoid of lines. One of these lines, A. = 5.300, has 

 a broad diff"used wing, and it is perhaps worthy of remark that in the 

 spectrum excited by the two zinc radiations we have a hazy doublet at 

 this point, in the spectrum excited by zinc 4811, a single line, and in 

 the spectrum excited by cadmium 480, two faint lines. Some of the 

 other lines have wings, as will be seen from the chart, and at wave- 

 length 546 we find a broad hazy band. All of these peculiarities com- 



