Measured by the Black Bulb Thermometer. 25 



reflected back from the surface of the mercury, and lost, 

 while the coating of the blackened bulb being nearly 

 or quite opaque to all rays, the whole heat falling on the 

 bulb becomes rapidly absorbed. 



The polished surface exhibited by the black glass bulb 

 too, must reflect many rays which would be absorbed by the 

 dead surface of the blackened bulb, and further, the coated 

 bulb would radiate heat very rapidly as compared with the 

 polished glass surface, and give rise to variations due to the 

 interposition of visible or invisible aqueous vapour in the 

 atmosphere, which the plain glass bulb would be insensible to. 



In the numerous observations made with these thermo- 

 meters, many cases occurred in which the differences already 

 given would not represent the actual observed differences — 

 in some of which there could be no doubt the more rapid 

 radiation of the blackened bulb during the drifting of thin 

 filmy clouds across the sun was the cause of this, and 

 I am inclined to attribute all such differences to the absorp- 

 tion of heat rays by aqueous vapour in the atmosphere, the 

 more sensitive blackened bulb indicating the loss, while the 

 black glass, with its badly radiating surface, remains in- 

 sensible to such variations. In these fluctuations the lowest 

 temperatures indicated by the coated bulb, were never so 

 low as the highest shown by the plain black glass bulb. 



Last December, in a communication I received from Mr. 

 Todd, the director of the Adelaide Observatory, he makes 

 the following remarks : — 



" I recently had occasion to substitute a fresh solar 

 thermometer for the one that had been in use for several 

 years. They were both Negretti and Zambra's make, having 

 black glass bulbs, and enclosed in vacuum tubes — the only 

 difference being that the new one had a smaller bulb, and 

 was enclosed in a smaller bulb. They were both compared 

 with Greenwich standards, and had no appreciable error. 

 They also both read exactly alike in the shade, or in hot 

 water up to limits of scales. Yet, in the sun, the old one 

 reads often 10° or 12° higher than the other, according to 

 the intensity of solar radiation. How is this to be accounted 

 for ? It is a rather interesting question, because solar 

 radiation observations are no longer comparable if the 

 instruments differ, and I have for a long time noticed, with 

 surprise, that yours and Sydney's observations are much 

 lower than mine, which I have hitherto attributed to our 

 much drier and consequently more heat transparent atmo- 



