Miscellanies. 



207 



exempt, since however apparently arid the external air at 120° Fahr. ! 

 may be, the moisture in it may yet be in excess and tending to deposition, 

 when the same air is cooled down to many degrees beneath the freezing 

 point. 



" The data wanting in the case of your Orenburg cave, are the mean 

 temperature of every month in the year of the air, and of thermometers 

 buried say a foot deep, on two or three points of the surface of the hill, 

 which if I understand you right is of gypsum and of small elevation. I 

 do not remember the winter temperature of Orenburg, but for Catherinen- 

 burg (only 5° north of Orenburg) the temperatures are given in Kupp- 

 fer's reports of the returns from the Russian magnetic observatories. If 

 any thing similar obtains at Orenburg I see no difficulty in explaining 

 your phenomenon. Rejecting diurnal fluctuations and confining our- 

 selves to a single summer wave of heat propagated downwards alternately 

 with a single winter wave of cold, every point at the interior of an insu- 

 lated hill rising above the level plain will be invaded by these waves in 

 succession, (converging towards the centre in the form of shells similar 

 to the external surface,) at times which will deviate further from mid- 

 winter and mid-summer the deeper the point is in the interior, so that at 

 certain depths in the interior, the cold-wave will arrive at mid-summer 

 and the heat-wave in mid-winter. A cave (if not very wide-mouthed 

 and very airy) penetrating to such a point will have its temperature de- 

 termined by that of the solid rock which forms its walls, and will of 

 course be so alternately heated and cooled. As the south side of the hill 

 is sunned and the north not, the summer wave will be more intense on 

 that side and the winter less so ; and thus, though the form of the wave 

 will still generally correspond with that of the hill, their intensity will 

 vary at different points of each wave-surface. The analogy of leaves is 

 not strictly that of the progress of heat in solids, but nearly enough so 

 for my present purpose. 



"The mean temperature for the three winter months, December, Jan- 

 uary, February, and the three summer months, June, July, August, for 

 the years 1836, '7, '8, and the mean of the year, are for Catherinenburg 

 as follows : — 



" The means of the intermediate months are almost exactly that of the 

 whole year, and the temperature during the three winter as well as the 

 three summer months most remarkably uniform. 



