2 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. IIO 



the rays would be immediately absorbed on that cone and would give 

 up their heat there into the flowing water. The remaining 5 percent or 

 less would be scattered over an entire hemisphere, of which nearly 

 the whole solid angle was included in the blackened walls of the 

 chamber. Over 95 percent of the trifling amount of radiation scattered 

 from the cone, impinging upon these walls, would be absorbed on 

 them and this heat also would be communicated to the flowing water. 

 Only' the measured aperture, through which solar rays entered, 

 was open to free escape of the scattered rays As this aperture 

 subtended but 0.012 hemisphere as viewed from the hollow cone le s 

 than 0.012 of 5 percent of the introduced solar radiation could reely 

 escape So, theoretically, the chamber was fully 99-94 percent black 



Lest some unforeseen error should lurk in the device, two coils 

 of insulated wire were wound upon the cone. One coil w^s wound 

 in shellac directly upon the rear wall of the cone, being behind the 

 water stream within the cone, but in front of the water stream in 

 the extreme back wall of the chamber. This coil was more f avorab y 

 situated than solar heating to convey electrically produced heat to 

 the flowing water. The other insulated coil was of several millimeters 

 thickness, was doughnut-shaped, and was stuck on with shellac to tl^ 

 front rim of the hollow cone, outside the area covered by the beam 

 of sunUght. This coil was very unfavorably situated to give up 

 electrically produced heat to the flowing water, since it must hrst 

 give its heat to the air, and then to the walls of the chamber^ 



I have been describing Standard PyrheHometer No. 3. On pages 

 61 and 6^ of Annals of the Smithsonian Astrophysical Observatory, 

 volume 3, 1913, there are given 24 tests, half with each of the two 

 heating coils, where electrically introduced heat was measui-ed by 

 absorption in the flowing water. The results of 12 tests at Wash- 

 ington, April 18, 22, and 23, 1910, showed no certain difference as 

 between the two coils, and gave a mean result of 99-85 percent heat 

 found The results of 12 tests at Mount Wilson, October 10 and n, 

 191 1 also equally divided between the two coils, gave 100^6 percent 

 heat found. These results come to well within their probable error 

 at exactly 100 percent heat found. They therefore indicate that heat 

 introduced in the chamber, no matter whether more or less favorably 

 for measurement than solar heat, is completely absorbed and accu- 

 rately measured by the instrument. This, as we shall see later, is 



a critically important result. , 1 1 ^f ^,.r 



Not content with this method of fixing the standard scale of pyr- 

 heliometry we constructed another instrument of the hollow-chamber 

 type It was called the water-stir pyrheliometer, because, mstead 



