366 Prof. Wood and Mr. Moore on the Fluorescence 



absorption spectrum above, fluorescent below). The fluor- 

 escent spectrum in the green-blue region may be divided 

 into three groups of bands: — (1) Those from 5340 to 5080, 

 consisting of six broad hazy bands which correspond in position 

 to the fluted bands of the absorption spectrum ; higher dis- 

 persion would doubtless resolve them into lines. (2) Those 

 from 5080 to 4865, a group of irregularly fluted bands, the 

 heads of which point towards the red, i. e. they are strongest 

 on the less refrangible side, and shade oif on the side of the 

 shorter wave-lengths. (3) Those from 4865 to 4600, w^hich 

 appear under the dispersion used as uniform narrow bands. 

 The spectrum is the exact complement of the absorption 

 spectrum taken wdth the same instrument^ and further work 

 with the concave grating of 14 feet radius will show^ w^hether 

 all of the very minute and narrow lines in the flutings of the 

 latter spectrum are present in the former. We feel sure that 

 the spectrum can be photographed with the large grating 

 without difficulty. 



Visual observations have shown that the red fluorescence 

 is also fluted, though the flutings are only to be seen when 

 the vapour is very dense and the illumination very intense. 

 It has not yet been definitely proven whether this fluting 

 belongs properly to the fluorescent spectrum or is the result 

 of absorption. A new" form of tube has been designed wliich 

 it is believed wdll give a much denser vapour and make the 

 settlement of this question possible. 



The complementary nature of the photographs of the 

 fluorescent and absorption spectra might lead one to suppose 

 that the wave-length absorbed was re-emitted without change 

 of wave-length. To test whether or not this was the case it 

 w-as necessary to illuminate the vapour with monochromatic 



The Fuess monochromatic illuminator, which enables one 

 to cut out a region of any width from a very intense solar 

 spectrum and focus this light at a definite point in sj^ace, 

 was arranged so as to send a cone of violet light down the 

 tube, the rays meeting just above the mouth of the crucible. 

 No fluorescence was detected. On gradually increasing the 

 wave-length of the light by turning the graduated screw 

 wdiich rotated the prisms of the instrument it was found that 

 the fluorescence appeared when wave-length 4600 was reached. 

 On looking into the end of the large tube a very beautiful 

 phenomenon was seen. The cone of deep blue monochro- 

 matic light was distinctly A-isible owing to traces of oxide 

 floating about, while at the point where the rays met in the 

 dense sodium vapour there appeared a brilliant spot of ^/^//oiczV/i- 

 green fluorescent light. As the w^ave-length was further 



