Almnsphere and the Earth. 239 



in Konigsbei-g (See these Annals, July 1827, part x. p. 302.), 

 will be useful, which are contained in the following table : 



The four last numbers are calculated. 



We see, from these observations, that the minimum of tem- 

 perature falls in March, the maximum in September ; the mean 

 of both gives 46^.76 : therefore, very near the mean tempera- 

 ture of the earth, calculated from all the observations which 

 Erman finds, 46^.70. The difference of temperature for Ja- 

 nuary and October, is, for the Konigsbcrg spring, 4°.39 ; for the 

 Kasan spring, 0°.9, that is nearly five times less ; it is therefore 

 probable that the fifth part of the changes of the Konigsberg 

 spring very nearly expresses the changes for the Kasan spring. 

 We thus obtain for Kasan : 



September, . 43.7 + 1 (49°.r7 — 49'.4l) = 43!'77. 



March, . . 42.8 — \ (45.03 — 43.74) = 42.55. 



Mean, . = 43.16. 



We may therefore regard 43°.2 as the true mean tempera- 

 ture of the earth at Kasan. 



The temperature of springs depends on many, often variable, 

 circumstances, and especially on the quantity of water which is 

 renewed in a certain time. A well not far from the second 

 spring, the level of whose water was only about twelve feet un- 

 der the surface of the ground, and was scarcely two feet deep, 

 whose water was therefore continually renewed (for this well 

 supplies the whole village, at the Archiepiscopal Palace), 

 shewed, on the 19th July 1828, 43°.81. Another well, in the 

 village of Butiska, likewise not far from the second spring, but 

 which was little used, and the level of its water twenty feet un- 

 der that of the earth, shewed, on the 19th July 1828, 41 °.56; 

 on the 29th October of the same year, 42°.12. This tempera- 

 ture of 42^.12 was also that of a small spring 011 the 4th No- 



