﻿and Fluctuations . 333 



sandwich, is perfectly certain. The balloon observations of Mr. 

 Glaisher have, at any rate, established this beyond the possibility 

 of further doubt*; but they have also established the fact that 

 above all, beyond the reach of these alternations, is a region 

 where the air is intensely cold and intensely dry. I have else- 

 where t discussed at great length the peculiar unwillingness with 

 which two masses of air, or any fluid, mix when their thermome- 

 tric, hygrometric, or other conditions are different. The distinct 

 outline very commonly offered by a stench or a perfume affords 

 a clear and familiar illustration of my meaning : we do not, as a 

 rule, find the smell gradually grow on us ; it comes with such a 

 sharpness that its rate of progress could easily be timed. It is 

 the same with masses of air lifted into the upper regions of the 

 atmosphere. A mass of hot air or a mass of moist air retains as 

 much of its heat or moisture as is not by the forcible expansion 

 withdrawn from it ; and refuses, by virtue of this property, to 

 mix with the cold or dry air which lies above it, or which it drags 

 up after it. 



It must therefore be borne in mind that when a certain mass 

 of air is heated, and a tendency to expand — not to rise — imparted 

 to it, its elastic force, which is the measure of the tendency to 

 expand, is increased, and exerts itself in that direction in which 

 at the moment it finds least opposition. Any one who has 

 watched the course of a stream will know how the most trifling 

 and apparently the most insignificant obstacles deflect its flow in 

 a very marked manner ; a blade of grass lying over from the 

 bank and just tickling the surface will make a ripple and an eddy 

 and a backwater, producing an amount of disturbance which re- 

 peated observation can alone lead us to credit. A tiny bit of 

 gravel thrown into a pool will send waves swelling and circling 

 almost as far as we can see ; and knowing that these effects are 

 produced in a comparatively dense sluggish fluid such as water, 

 we may, without extravagance, maintain that causes infinitesi- 

 mally small, and altogether beyond our power of discernment, 

 will suffice to direct the motion of a fluid so mobile and so elastic 

 as air. It therefore by no means follows that the expansion of 

 heated air takes place in an upward direction. We have, as I 

 have shown, strong reasons for believing that it sometimes takes 

 place laterally; but these cases are exceptional, and may fairly 

 be supposed due to a very extraordinary and sudden increase 

 of temperature; but when an increase takes place slowly and 

 steadily, the elastic force and the expansion would appear to 

 keep pace with each other ; the mass of air immediately over- 



* See also Strzelecki, ' Physical Description of New South Wales/ p. 162. 

 t Physical Geography in its relation to the prevailing Winds and C ur . 

 rents, pp. 149 et seq., 266. 



