WOOD. — SPECTRA OF SODIUM VAPOR. 257 



They have some connection with the D lines, I feel sure, for they are sym- 

 metrically arranged on each side of them. If the photograph had been 

 made with a grating, we should of course call them ghosts. It may be 

 that they are analogous to satellite lines, but if they are, we are cer- 

 tainly dealing with the phenomenon on a grand scale, for the fourth 

 one is not far from the sodium doublet at 5688 ! All of these points 

 will be more fully investigated during the coming year. 



Other Possible Excitations. 



It has occurred to me during the preparation of this paper that 

 very interesting results would be obtained by exciting the fluorescence 

 with the light selectively rotated by the vapor in a magnetic field, i. e. 

 by the magnetic bright-line spectrum. This light is fairly intense, 

 and it would be interesting to see whether the intensity distribution 

 among the excited fluorescence lines was the same as in the magnetic 

 spectrum. 



What I most need, however, is a set of screens which will enable 

 me to separate lines such as those of copper without resorting to the 

 systems of prisms and lenses. A good collection of solutions of the 

 rare earths would probably be very useful in the work. Erbium, 

 praseodymium, and neodymium I have, but I should feel very 

 grateful for the loan of any others which might prove serviceable, or for 

 any suggestions regarding other possible sources of monochromatic 

 light. As I have said before, the instrument most needed is a light 



siren ! 



Composite Excitation. 



At the top of the chart just below the magnetic series will be found 

 a spectrum containing about two hundred lines. This is a composite 

 drawing made by superposing all of the drawings made of the simple 

 spectra excited by monochromatic stimulation. It contains many lines 

 not found in the complex spectrum excited by white light. In the 

 latter, between wave-lengths 5000 and 5100, we find but ten or a 

 dozen lines, while in the composite spectrum there are at least twenty. 

 This circumstance is of interest in connection with the periodic dark 

 regions of the complex spectrum, which give it a fluted appearance. 

 The formation of these flutings requires further study, as their position 

 shifts as we alter the wave-length of the exciting light, which in this 

 case is a rather broad, isolated band from the continuous spectrum 

 obtained with the monochromatic illuminator. The phenomenon was 

 more fully described in the earlier paper, but requires further study. 



VOL. XLIl. — 17 



