360 Sir John Herschel on the Phenomena of the 



i. e. the difference of temperature) is greatest, the reverse process 

 is going on. Caves in moderately free communication with the air 

 are dry and (to the feelings) warm in winter, wet or damp and cold 

 in summer. And from the general course of this law I do not con- 

 sider even your Orenburg caves 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 ther- 

 mometers buried say a foot deep, on two or three points of the sur- 

 face 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 Catherinenburg (only 5° north of Orenburg), 

 the temperatures are given in Kuppfer's reports of the returns from 

 the Russian magnetic observatories. If anything similar obtains at 

 Orenburg I see no difficulty in explaining your phenomenon. Re- 

 jecting diurnal fluctuations and confining ourselves to a single sum- 

 mer wave of heat propagated downwards alternately with a single 

 winter wave of cold, every point at the interior of an insulated hill 

 rising above the level plain will be invaded by these waves in suc- 

 cession (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 determined 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 sum- 

 mer 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 naves is not strictly that of the pro- 

 gress of heat in solids, but nearly enough so for my present purpose. 



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

 January, February, and the three summer months, June, July, Au- 

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

 Catherinenburg as follows : — 



M The means of the intermediate months are almost exactly that 



