AND THE OTHER PLANETS OF THE SOLAR SYSTEM. 643 



14. It appears then that in whichever of the above methods we solve this problem, we 

 arrive at the same result respecting the temperature of the atmosphere at its upper limit. The 

 only reasonable cause of doubt, I think, as to that temperature being so low as here determined, 

 is to be found in the assumption that the equation 



p = a*p (1 + at), 



where a and a are constant, holds true for extremely low values of p and t. But whether this 

 be accurately so or not, there can be no sufficient reason to doubt the accuracy of the 

 conclusion that the temperature t 2 must at least approximate to some such low value as that 

 determined above. 



15. The best data we possess for the determination of the constant ( — ] are furnished by 



Mr Welsh's observations made in four balloon ascents from Kew, of which two were made in 

 the month of August 1852, one in October, and another in December. He reached the height 

 of very nearly 23000 feet above the Earth's surface. In each ascent the temperature decreased 

 in ascending a few thousand feet, and then became stationary for some three or four thou- 

 sand feet, after which a regular decrease again commenced. This anomaly was manifestly 

 due to peculiar atmospheric causes, and ought to be abstracted in determining the proper 

 decrease of temperature due to general causes. This being done, Mr Welsh shews that the 

 law of observed temperature is very nearly as well expressed by taking, in the series for 

 1 + a£, only the term involving the first power of z, as by including also the term in *' ; 

 and he obtains the following decrements of temperature for 1000 feet of ascent, in the four 

 ascents respectively : 



Aug. 17 ... 3°,434 (F.) 

 Aug. 26 ... 3°,440 

 Oct. 21 ... 3°,431 

 Nov. 10 ... 3°,205 



Mean ... 3°,377 (F.) 

 or ... 1°,85 (C.) 



Retaining the centigrade degrees, and taking 100 feet for our unit of length, we shall have 



- = ,185. 

 a 



Mr Welsh also gives the results of his observations, freed as far as possible from the effects 

 of anomalous causes, in the following form, shewing the mean ascent corresponding to one 

 degree of Fahrenheit : 



Aug. 17 ... 292,0 feet. 



Aug. 26 ... 290,7 ... 



Oct. 21 ... 291,5 ... 



Nov. 10 ... 312,0 ... 



