430 



Mr. J. E. Petavel. An Experimental 



standard must be constant, reproducible, and must emit light of a suit- 

 able colour. We have just seen that the source of light under con- 

 sideration may be regarded as fulfilling the first of these conditions : 

 with regard to the second and third, the conclusions are less favour- 

 able. 



By definition, the ratio between the total radiation and that trans- 

 mitted through the water trough must be as ten to one, 1 per cent, 

 error in the electrical measurements causing 3 per cent, change in the 

 light. To obtain the required degree of precision the galvanometer 

 deflections must be reduced to equality. Three methods are avail- 

 able : — 



1. The bolometer current may be reduced in the ratio of one to ten. 



2. A 1/1 Oth shunt can be introduced into the galvanometer circuit. 



3. The distance between the bolometer and the radiator may be 



varied in the ratio of 1 to ^10. 



None of these methods is entirely satisfactory. The reduction of the 

 current through the bolometer involves a considerable complication of 

 the apparatus, and interferes with the steady working of the instru- 

 ment. It is also by no means impossible that the resulting change in 

 the temperature of the bolometer films may change their coefficient of 

 absorption. With a shunted galvanometer, temperature changes and 

 thermo-cur rents are a frequent source of error. The most reliable 

 method is to change the distance of the radiator ; but here again two 

 objections arise. The enclosure containing the platinum foil is rarely 

 at a temperature so closely approaching that of the bolometer films 

 that shifting this large surface leaves the instrument unaffected. 

 The law of inverse squares cannot strictly be applied; as the rays of 

 light are refracted when passing through the water trough. In prac- 

 tice, marked variations were obtained when the method of determining 

 the ratio of the two radiations was altered. 



The use of another form of bolometer produced a considerable change 

 in the light emitted. The bolometer films are coated with platinum 

 black by electrolysis. The composition of the bath, the electromotive 

 force, and the current have been specified \ but it was found that the 

 temperature of the bath and the resistance of the film affected the 

 nature of the deposit obtained. 



Finally, the spectral composition of the light is unsatisfactory, the 

 colour being much too red. 



We are thus driven to the conclusion that this light does not 

 possess all the qualities required of a primary standard. Under 

 certain circumstances it may be of the greatest value as a standard of 

 reference. It is used in the Eeichsanstalt to check the values of the 

 Hefner lamps; for purposes of this kind it is well suited, and might 

 with advantage be more frequently used. 



