720 



Dr. R. A. Houstoun on the Relative 



observer's two "weak" curves was taken, and the reciprocal 

 of slit-width plotted against s, the drum reading. This gave 

 two prismatic luminosity curves for each observer. These 

 curves were integrated and then replotted, with their ordi- 

 nates shortened or lengthened, so as to make their areas the 

 same, This was done so that they would all contribute in 

 the same degree to the average. The average of all the 

 strong intensity curves and of all the weak intensity 

 curves was then taken, and they are shown in the following 

 diagram (fig. 2) : — 



Fia:. 2. 



A-SO^m. 



soo 



sso 



The full curve is for the illumination of § and the broken curve for 

 the illumination of g-^ metre-candle. 

 The diameter of the pupil of the eyepiece was 7 mm. 



The extreme red is not given, as it was found in the course 

 of the tests that the readings on the drum were not reliable 

 for that part of the spectrum, owing to the sprino- for 

 rotating the prism becoming slack there. The maxima of 

 the curves are situated at 576 and 560 jxjx. 



The sixth column of the preceding table was then plotted 

 as a function of s, the drum reading, and the ordinates of 

 the energy curve thus formed divided into the corresponding 

 ordinates of the two luminosity curves. The result gave the 

 two visibility curves shown in the next diagram (fig. 3). To 

 determine the shape of these curves was the purpose of the 

 investigation. 



The broken curve is for the weaker intensity, and is plotted 

 only on half the scale of the stronger intensity curve. The 

 maxima for both curves was further towards the violet than 

 had been anticipated, and the energy curve had to be extra- 

 polated in the violet. However, extrapolation gives as great 



