356 PROCEEDINGS OF THE AMERICAN ACADEMY 



the losses produced by conduction and convection, which In a very 

 perfect vacuum are practically negligible ; for the conduction through 

 the metallic connections of the stri|)s tliemselves is wholly insignificant, 

 owing to the excessively small cross-section of tlie latter in proportion 

 to their length. With the value of the loss by radiation thus obtained, 

 the above results, which are of course presented only as a first approxi- 

 mation, can be corrected. 



The preceding remarks comprise what is most essential in a descrip- 

 tion of the working apparatus founded on the method of changed elec- 

 tric resistance. 



I write far from large libraries, and do not pretend to give the bibli- 

 ography of the method, if it have one. The only notices of a use of 

 the principle involved in this instrument that are known to me are 

 the following: — 



1. On the Measurement of Resistance to the Conduction of Electric 

 Currents, and on a Galvanic Differential Thermometer, by A. F. Svan- 

 berg. Poggendorff's Annalen, 1851, vol. Ixxxiv. p. 411. 



2. On Measuring Temperatures by Electricity, by C. W. Siemens. 

 Proceedings of the Royal Institution of Great Britain, March 1, 1872, 

 vol. vi. p. 438. 



Mr. Siemens has made subsequent special applications of this prin- 

 cij^le. 



This paper, as a description of the apparatus for measuring radiant 

 energy, ends here; but I may mention the following results which have 

 been obtained vrith the instrument in other researches, to give a more 

 just idea of its efficiency as a working-tool for the physicist. 



In illustration of the work which the balance may be used for, take 

 the following repetition of one of Melloui's experiments. 



Source of heat, petroleum lamps (argand burner with glass chimney) 

 at 15 cm. distance. Radiation limited by screen with 1 cm. circular 

 aperture at 10 cm. from the steel Balance. 



The unimpeded radiation was compared with that transmitted by 

 distilled water and by an aqueous solution containing ten per cent (by 

 weight) of common alum. The liquids were contained in a glass cell 

 (sides 2.5 mm. thick, distance between sides 19.0 mm.). Temperature 

 21° C. Galvanometer not sensitive. 

 Deflection (mean of five trials) — 



by radiation through distilled water and glass = 1 6.5 div 



unimpeded 

 through 10 

 unimpeded 



= 1 6.5 div. ) 

 = 107.6 „ > 



through 10% alum solution and glass = 13.7 „ ) 



= 103.8 „ ) 



