THE LUMINOUS EFFICIENCY OF ILLUMINANTS 543 



source producing only visible white light and no extraneous 

 non-luminous vibrations. According to the latest determination 

 of P. G. Nutting, who has studied this matter with special care 

 at the Bureau of Standards in Washington, U.S.A., such a 

 source would yield 26 C.P. per watt. Seeing that the most 

 efficient artificial illuminant in use to-day does not produce 

 more than about 4-5 C.P. per watt, even when considerable 

 latitude is allowed as regards colour, it is evident that we are 

 still a long way from the desired goal. 



In what has been said above, it has been assumed that our 

 ideal illuminant should not only emit all the visible vibrations 

 in the spectrum, but should emit these in the proportions in 

 which they constitute daylight. In a sense, a source which 

 achieved this might be said to have a radiant efficiency of 100 

 per cent., seeing that all this output would be within the visible 

 spectrum limits. Yet it is clear that by disregarding colour 

 and aiming only at the production of as bright a light as 

 possible, i.e. the generation of a maximum amount of light for 

 a given consumption of energy, considerably more efficient 

 results might be obtained. It is obviously uneconomical to 

 include the rays at the extreme ends of the spectrum, the deep 

 red and violet, which are just on the borderland of visibility. 

 Clearly if quantity of light were the only consideration, our 

 right course would be to ascertain the precise ray for which 

 the sensitiveness of the eye was a maximum and then produce 

 only this variety of radiation. 



The question arises therefore where in the spectrum this 

 position occurs. There has been much misunderstanding on 

 this point, partly on account of the difficulties involved in 

 heterochromatic photometry and also on account of the com- 

 plexities introduced by the physiological peculiarities of the 

 human eye. In order to locate this point of maximum sensi- 

 tiveness, it is necessary to obtain the curve of luminosity for 

 an illuminant and also the curve representing the distribution 

 of energy; then to divide the ordinates of the former by the 

 latter. In this way a curve of luminous efficiency regarded 

 as a function of the eye, which would be independent of the 

 illuminant studied, would be obtained. 



Now the difficulty that arises in this connexion is that the 

 eye itself is not constant in this respect. At a high illumination 

 its behaviour is profoundly different from that characteristic of 



