The Rev. Dr. Robinson on the Constant of Refraction. 179 



assuming the density to decrease uniformly as the height increases; that of 

 Laplace, expressing the density by a product of two factors, representing the 



preceding hypotheses, and that of Ivory, supposing it asf 1 — —y- J .* Now these 



are obviously mere arbitrary assumptions, and the verifications which some of 

 them are supposed to receive by exhibiting the decrease of temperature at a 

 small elevation, and the barometric formula for heights, are worth little ; the 

 first being unknown at any given place,f and the second being a consequence of 

 any law which will make the temperature decrease nearly uniformly within a few 

 thousand feet. The slightest attention to meteorological facts will show that 

 there cannot be any general formula expressing the density in terms oHhe height 

 alone, and that even could it be found for one place by experiment, it would be 

 entirely inapplicable to any other. It is certain, that between the tropics there 

 is an ascending current of heated air, replaced by a stream of cooler from the 

 north, while it flows towards the poles, descending in its turn and giving out its 

 heat ; and it is therefore equally certain that the law of atmospheric temperature 

 must depend on the latitude. It is not impossible, that in the arctic regions we 

 may find a uniform temperature, or even an increase on ascending. Such must 

 indeed be the case, if there be any truth in the conclusions of Fourier, or Poisson, 

 respecting the temperature at the termination of our atmosphere ; for if with the 

 former we suppose it = — 58° of Fahrenheit, or with the latter, much more 

 elevated, approaching 32°, yet cold below either has been observed by northern 

 travellers. At a given place we might, perhaps, by aeronautic investigations, 

 ascertain the law of decreasing density and temperature, for a certain epoch ; but 

 it is highly probable, that this would not obtain when the sun had a different 

 declination, or the weather was different ;| it is unquestionable, that it would be 



* The last appears the best, but it is to be regretted that Mr. Ivory has assumed the use of the 

 internal thermometer, and not given separate reductions for the temperature of the barometer. 

 This last also applies to the very convenient tables of Bessel's Refractions, given by Mr. Airy. 



f Because the decrease in free air cannot be the same as that observed on the side of a mountain, 

 and in contact with a mass of matter influenced both by the air and the earth's internal heat. 



{ In the celebrated ascent of Gay Lussac, the temperature at Paris was 87o Fahrenheit, so that 

 the air cannot have been in a normal condition : the meteorological instruments below should have 

 been noted every few minutes, and the times of observation above given. In the published 



2 a2 



