164 KEPOET— 1892. 



transmitted light, that of the incident being 1- 



2/x 



On multiplying the calculated final excess got from the observations 

 without any plate by the above factor, it came, within the limits of errors 

 of observation, the same as the calculated final excess obtained from the 

 observations with the quartz plate on. It follows that there is no 

 seusible loss due to absorption in the quartz plate. It is to be re- 

 membered that the rays that fell upon the quartz had already passed 

 thi-ough the glass lens, and also that in the radiation from the sun it is a 

 CO inpai-atively small proportion of heat rays that are absorbed by glass 

 and similar substances. 



It remains to be explained in what way we were led to the conclusion 

 tliat the employment of green instead of coloui-less glass for the bulb of 

 tlie insolation thermometer must have made but little difference in the 

 results obtained. 



Imagine a thermometer to be suddenly exposed to solar radiation, as 

 in Stewart's second actiuometer, and consider what its behaviour ought 

 to be on the two extreme suppositions: (I) that the mercury in contact 

 with the glass reflects perfectly all the rays that fall upon it, but that the 

 shell is partially opaque ; (2) that the mercury reflects only partially, but 

 that the shell is perfectly diathei-manous. 



On the first supposition the mercury would not be warmed at all by 

 the rays which fell upon it, but only by conduction from the shell, which 

 itself would be heated by absorption of a portion of the rays that fell 

 upon it, either as they came from the sun or as they were on their way 

 back after reflection at the surface of the mei-cury. The rise of tempera- 

 ture of the shell would ultimately vary as the time elapsed. But if the 

 shell were at a given temperature the total heat received by the mercury 

 from the shell would vary ultimately as the time during which it has been 

 passing in. But as the temperature of the shell is not constant, but its 

 rise varies ultimately as the time since exposure, the total heat received 

 by the mercury will vary ultimately as the integral of a quantity which 

 varies as the time, and will therefore vary ultimately as the square of 

 the time. 



On the second supposition the mercury receives its heat directly from 

 the sun, and the total heat received varies ultimately as the time daring 

 which it has been receiving it. 



Now in the actual observation the gain of heat was found to be 

 ultimately sensibly proportional to the time elapsed, not to the square 

 of the time, as may be inferred from the fact that the rate of increase 

 was decreasing from the first. We may conclude therefore that the gain 

 of heat was due almost entirely to the imperfection of the reflection from 

 the mercury, which entails direct absorption by the mercury of the 

 portion which failed to be reflected, and only in a comparatively in- 

 significant degree due to absorption of heat by the shell in the passage of 

 the heat through it. "We may therefore infer that the substitution of 

 green for colourless glass in the shell of the bulb would make but little 

 difference in the results obtained. This agrees with the experience of 

 Captain Abney, who was led by his experiments on the diathermancy 

 of various kinds of glass to suppose that a thermometer with a bulb of 

 green glass would rise decidedly higher in sunshine than one with a shell 



