﻿124 



Mr. J. S. Dow on Colour 



the connexion between the apparent relative candle-power 

 and the distance away of the eye with a Joly photometer, 

 when a Nernst lamp was compared with a glow-lamp (running 

 at 3*7 watts per c.p.), and when an incandescent mantle was 

 compared with a Methven burner. 



In the Joly photometer the distance away of the eye is 

 left entirely to the inclination of the observer, and it will be 

 seen from the above that differences of 5 per cent, or more 

 might easily be introduced between the readings of different 

 observers in this way. 



In the Lummer-Brodhun photometer the distance of the 

 eye is limited, to some extent, by the use of the telescope. 

 But the position of the telescope can be varied between 

 wide limits without putting the field out of focus, and this 

 latitude allows of a considerable difference in the readings. 



In the table below the extreme differences are given for 

 these two limiting positions of the telescope : — 



Nature of Lights 

 compared. 



Eatio of C.P. 



Percentage 

 Difference. 



Telescope in. 



Telescope out. 



Kuby-Red to Signal- 

 Green. 



1-65 



2-20 



25-0 



Glow-lamp (3'Twatts per 

 c.p.) to Nernst lamp. 



1-38 



1-43 



3-5 



Metbven Gas Standard 

 to Incandes. mantle. 



0-129 



0-135 



4-4 

 0-77 



Harcourt 10 c.p. Pen- 

 tane Standard to 

 Fleming Standard 

 Glow-lamp. 



0-652 



0-657 



Here again a distinct difference in reading is produced in 

 several practical cases. 



It is also remarkable that a small but distinct effect was 

 produced in the last case, even though the flame of the 

 Harcourt lamp is only very slightly redder to the eye than 

 the light from the Fleming glow-lamp. 



It is difficult, of course, to speak with certainty of such a 

 small change as this — a change which would be produced by 

 moving a photometer, set midway between two lights 2 

 metres apart, a distance of less than 2 millimetres. But the 

 writer has usually found that the mean of a set of readings, 

 taken with the telescope out, worked out to a value slightly 



