NO. l8 SMITHSONIAN PYRHELIOMETRY — ABBOT AND ALDRICH 7 



Summary. — A new form of standard pyrheliometer has been 

 devised and tested. In this new instrument, as in the water-flow 

 pyrheliometers, the solar rays are absorbed in a deep chamber 

 approximating to the perfect absorber or " black body." Means are 

 provided for introducing electrically test quantities of heat. 



It is shown that with Standard Water-flow Pyrheliometers Nos. 2 

 and 3, and the new Water-stir Pyrheliometer No. 4, test quantities of 

 heat may be measured to within 1 per cent. 



A summary is given of all definitive comparisons of the three 

 standards just named with Secondary Silver-disk Pyrheliometers, 

 and also the net of inter-comparisons connecting all Smithsonian 

 secondary pyrheliometers now in use. From these data are derived 

 the best values of the constants of all these secondary pyrheliometers. 

 This system of pyrheliometry we call " Smithsonian Revised 

 Pyrheliometry of 1913." 



It rests on 72 comparisons on 20 different days of 3 different years 

 with 3 standard pyrheliometers of different dimensions and 2 widely 

 different principles of measurement, all capable of recovering and 

 measuring within 1 per cent test quantities of heat, and all closely 

 approximating to the " absolutely black body." The 72 comparisons, 

 40 at Washington, 32 at Mount Wilson, were made in 6 groups. The 

 maximum divergence of the mean results of these groups is 1 per cent. 

 Hence it is believed that the mean result of all the comparisons made 

 under such diverse circumstances must be within 0.5 per cent of the 

 truth. The probable error is 0.1 per cent. It is believed that this 

 standard scale is reproducible by the secondary pyrheliometers with 

 the adopted constants given to within 0.5 per cent. The divergence 

 of this scale from that of Angstrom appears to be 3.9 per cent. 



