BELL. — ULTRAVIOLET COMPONENT IN ARTIFICIAL LIGHT. 23 



important. The very high ultra violet output reached by the carbon 

 arc is as has already been pointed out largely due to the very intensive 

 cyanogen bands about in the middle of the ultra violet spectrum 

 and the output of wave length below 300 /x^t is materially less than it 

 is in the quartz lamp operated without its globe. 



At the other end of the list stand the G. E. M. lamp and the ordinary 

 Cooper-Hewitt tube, the former showing a very low ultra violet 

 output by reason of its relatively low temperature and the latter 

 by reason of the fact that the extreme ultra violet is entirely cut off 

 ^by the tube, and the middle ultra violet being very weak in the 

 mercury spectrum, the main body of the energy is of wave length 

 greater than 365 ijl/jl. In fact since the spectrum of the G. E. M. 

 lamp runs down nearly to wave length 300 /x/x, and is strong only 

 between say 360 and the ^■isible, the energy distribution of the spectra 

 of these two illuminants is singularly similar, considering their wide 

 difference in character. 



The Nernst and tungsten lamps produce rather more total ultra 

 violet than the Cooper-Hewitt tul)e, most of the output being toward 

 the visible spectrum. The Nernst lamp operated without its globe 

 gives a spectrum relatively stronger in the further ultra violet, reach- 

 ing wave length 300 mm with a considerable degree of strength and 

 stretching beyond it. All the lamps running with glass globes show 

 a weak spectrum in that region. For this reason the quartz lamp 

 with its regular diffusing globe shows an ultra violet output per watt 

 almost as low as the G. E. M. lamp, the cut off of the globe in the 

 ultra violet region being very striking. The magnetite arc both with 

 and without its globe gives a considerable ultra violet output. The 

 globe cuts off much less ultra violet than in the case of the quartz 

 lamp, the latter being relati\ely rich in the rays which the glass most 

 effectively absorbs. 



Table H shows the percentage of energy cut off by the Euphos glass 

 in each of the illuminants investigated as compared with the total 

 energy which was transmitted by the quartz water cell, and also the 

 relative horizontal c. p. of the sources dealt with. The percentage 

 ratios of ultra violet are therefore numerically higher than they would 

 be in the case of admitting the whole infra red to the thermopile. 

 The relative composition of the various sources, however, is well 

 expressed by the data. 



