106 Nichols and Franklin — A Spectro-photometric 



During these measurements the current, potential and length 

 of arc were very nearly constant. We found, however, that 



they might be varied through 

 the widest limits compatible 

 with the maintenance of the 

 arc, without appreciably 

 changing the character of 

 those portions of the spectrum 

 which were under observa- 

 tion. The length of arc was 

 reduced from 3-5 mm to l ,mm , 

 the shortest arc which the 

 lamp in question could be 

 made to sustain, without in- 

 troducing any change in the 

 distribution of intensities ; 

 and a " short-arc " lamp using 

 the same carbons, with 21 

 amperes, 27 ,6 volts and an arc 

 too short to be measured, gave 

 almost precisely the same 

 values throughout the spec- 

 trum as those obtained from 

 the long arc lamp. (See 



Curves of relative intensities in the spec- Table II.) 

 trum of the electric arc Curve III is from A stu dy of the spectrum 

 measurements below the horizontal plane, ~ , v u+ tT" "U '• 



curve IV from measurements above the oi tne arc-llgnt wnicn m- 

 horizontal plane (16°), curve V is from eluded Only the regions lying 

 Vogel's measurements of the Poucault betweenA= 7530 and A=4500, 

 regulator. i i i i • 



would, however, be incom- 

 plete and misleading. Our measurements of that part of the 

 spectrum show, as will be seen by an inspection of Table II and 

 the curves already referred to, that the distribution of energy 

 between the above mentioned limits is very similar to that in 

 the spectrum of the lime-light. The latter is slightly stronger 

 in the red and weaker in the violet than the arc-light, but the 

 difference is not sufficient to account for the much greater 

 brilliancy of color in the electric-light, nor for its higher actinic 

 value. The source of the difference in character, and also of 

 the well known excess of violet light in the " long-arc," as com- 

 pared with the " short-arc " light, is to be found, however, for 

 the most part among still shorter wave-lengths. The electric 

 arc possesses an extremely brilliant band, really of complex 

 structure, but apparently continuous under the conditions of 

 observation with which we have to deal here, having its maxi- 

 mum in the neighborhood of wave-length 4280 and entirely 

 embraced within the region lying between A=4320 and A=4250. 



