556 Proceedings of the Royal Society of Edinburgh, [Sess. 
limit of the visible spectrum. The lower limit is immaterial ; there is so 
little energy in the violet and ultraviolet spectrum of glow-lamps. The 
luminous efficiency does not measure the commercial value of a lamp, because 
the eye is very much more sensitive to green than to red light, and it is 
possible for one lamp to have much more energy in its spectrum than a 
second one but yet to have that energy so disadvantageously distributed 
that the second produces a stronger effect on the eye. The luminous 
efficiency is of more theoretical interest. It tells us how much of our 
energy is wasted in dark heat and how far we can with advantage hope to 
improve our lamps. 
The object of this paper is to describe some measurements made on 
the radiant efficiency of carbon, osmium, tantalum, and tungsten glow- 
lamps. 
An old method of measuring the efficiency was the calorimetric one. 
The lamp was placed in a glass calorimeter and then in a similar copper 
one, and from the difference in the rise of temperature, with various correc- 
tions, the percentage of light radiated could be calculated. The method is 
in principle bad ; the whole energy and the dark heat are measured, and 
the light is obtained as the difference of two much larger quantities, both 
subject to considerable error of observation. Another method was to 
measure the total radiation by a thermopile. A water filter was then placed 
in front of the thermopile and the radiation that penetrated through the 
filter measured. The latter was supposed to be light, and if we made a 
correction for the light absorbed in, or reflected by, the filter — which could 
easily be determined photometrically — the radiant efficiency could be 
obtained. The objection to this method is that the water filter does not 
absorb all the heat. A solution of alum is no better than water, although 
popularly supposed to be.* Hence corrections were made for the dark heat 
transmitted by the water by means of solutions of iodine in carbon disul- 
phide, which were supposed to absorb all the light, and let the heat pass. 
The utter untrustworthiness of all such methods has, however, been demon- 
strated by E. L. Nichols and W. W. Coblentz.j- They have shown that they 
give values much too large. 
The older methods being thus discredited, there remain two modern 
methods. One consists in determining the energy radiated for different 
* It has been shown recently by K. A. Houstoun and J. Logie, Phys. Zs., xi. p. 672, that 
an aqueous solution of ferrous ammonium sulphate absorbs the dark heat much more 
effectively than water alone while allowing the light to pass. Nevertheless, it is not 
satisfactory enough as a means of separating the dark heat from the visible radiation. 
t “On Methods of Measuring Eadiant Efficiencies,” Physical Review , xvii., 1903, 
p. 267. 
