272 



Our heliotropes were circular glass mirrors, 8 inches in diameter ; 

 and for the smaller apertures, diaphragms were used between the 

 heliotropes and the observer. At the distances stated the light was 

 just visible to the naked eye in clear weather, and when seen over a 

 valley : if the ray grazed near the surface, the light was much re- 

 duced. On one occasion I employed a heliotrope at 6 J miles, and 

 used an aperture of \ of an inch, and found it rather brighter than 

 usual, so that probably 6i or 7 miles would be the normal distance 

 for that size. 



This agrees well enough with the rest of the scale, but there is no 

 need to employ a conjectural quantity ; and if the rate of absorption 

 corresponding to the above be computed, so close an agreement will 

 be found, as may entitle the numbers to be looked on as something 

 better than mere estimates, — as the results, indeed, of a species of 

 observation. 



The mean of the whole shews a loss of -0610 in passing through 

 one mile of atmosphere ; with the barometer at 27'0 inches (that 

 being about the average height of my stations), but reduced to 30' 

 inches, the quantity will be •0671. 



Hence the loss of light in passing from the zenith through a homo- 

 geneous atmosphere of 5'2 miles will be '303, or only about one per 

 cent, less than Professor Forbes's result. And as my air was con- 

 siderably drier than his (the mean humidity being not much above 

 ■30 instead of "56), this will probably account for the difference ; 

 and, at any rate, the agreement is much closer than could have been 

 expected. 



I once mentioned this matter to Captain Waugh, the pi'esent 

 Surveyor-General of India, then my fellow-assistant ; but he not 

 only had not noticed the thing, but did not even apprehend my mean- 

 ing. He assented to my remark on the loss of light in passing 

 through the atmosphere, but asserted that the aperture should vary 

 as the distance, thus allowing for no loss ! O'l inch per mile answered, 



