APPENDIX. 375 



the rate of change in the higher regions ; but it will be obvious 

 that the amount of effect produced by this cause is subject to 

 continual variation from changes in local conditions. The 

 nature of the soil, the extent and character of the vegetation, 

 the form of the surface, are all elements virhich modify the 

 amount of disturbance in the equilibrium of the surrounding 

 atmosphere. As above remarked, in discussing the climate of 

 Western Peru, prevailing winds which impinge upon a range 

 of mountains may indirectly affect the temperature of the 

 higher region by mechanically forcing masses of air to rise 

 along the slopes, and ultimately, by expansion, to be cooled 

 much below the temperature which they possessed when they 

 originally flowed against the slopes. 



One of the most important agencies affecting the distribution 

 of temperature in the atmosphere arises from the presence of 

 aqueous vapour. In its invisible condition it affects the absorp- 

 tive power of the air on the solar rays, and, when condensed in 

 the form of cloud, it acts as a screen, intercepting most of the 

 calorific rays which would otherwise reach the earth. But it is 

 especially through the large amount of heat consumed in con- 

 verting water into, vapour, and set free when vapour returns to 

 the fluid state, that the temperature of the air is largely modified. 

 When we consider that in converting a given volume — say, one 

 cubic metre — of water into vapour, enough heat is consumed to 

 lower about 1,650,000 cubic metres of air by 1° C. in temperature, 

 and that the same amount of heat is liberated when the vapour 

 so produced returns to the liquid state, we perceive how power- 

 fully the ordinary processes of evaporation and condensation 

 must affect the temperature of the air. 



It is needless to analyze further the several agencies which, 

 sometimes co-operating, and sometimes in mutual opposition, 

 determine the vertical distribution of temperature in the atmo- 

 sphere. It is but too obvious that no approach to uniformity 

 can be expected, and it might even be anticipated that any 

 approximation to a regular law of distribution that should be 

 found under one set of conditions — ^as, for instance, in serene 

 weather by day — would be altogether inapplicable under dif- 

 ferent conditions, such as exist in stormy weather, or by night. 



The need for practical application of some empirical rule, or 

 law, of vertical distribution has made it necessary to appeal to 



