24 REPORT 1871. 



tirely unprotected against the pressure of the water in which it is immersed, 

 and it is important to consider what effect this pressure will have. 



In thermometers of the ordinary construction this pi'essure acts only ex- 

 ternally, and produces much greater diminiition of the internal volume than 

 when, as in Prof. Lubimoff's thermometer, it acts both externally and in- 

 ternally, a mode of action with which we are familiar in the case of CErsted's 

 piezometer. 



From Eegnault's experiments it appears that the apparent compression of 

 mercury in glass, when the pressure is thus applied, is -000001234 per atmo- 

 sphere, whereas the apparent expansion of mercury in glass for heat is -0000857 

 per degree Fahi-enheit. The latter number is 69 times the former ; it there- 

 fore appears that a pressure of 69 atmospheres would be required to falsify 

 the indications of Prof. Lubimoff's thermometer to the extent of 1° F. The 

 actual pressure at the bottom of the well is less than the half of this, and 

 therefore should only produce an error of a few tenths of a degree. This, 

 however, is on the assumption that the glass undergoes no change of figure, a 

 condition which may easily fail of being fulfilled, owing to the want of perfect 

 uniformity in the glass. 



Mr. Donaldson has written from Calcutta to the effect that the thermo- 

 meter which was sent to him has been entrusted to a competent observer, 

 who has taken numerous observations with it, which will be sent shortly. 



M. Erman's letter above referred to is immediately followed in the ' Comptes 

 Rendus' by an account, by M. Walferdin, of some observations, which appear 

 to be very reliable, taken in artesian weUs in the basin in which Paris is 

 situated. They were taken with maximum thermometers of the kind in- 

 vented by Walferdin himself, in which the mercur3' overflows into a reservoir 

 when the temperature exceeds a certain limit, the thermometers being her- 

 metically sealed in glas^ tubes to protect them from pressure. 



The observations which he first describes were taken in a well, newly 

 sunk to the depth of 263 metres, at St. Andre, about 50 miles to the west of 

 Paris, and which failed to yield a supply of water. The temperature was 

 carefully observed at the depth of 253 metres by means of two thermometers, 

 which were allowed to remain at that depth for ten hours. Their indications 

 agreed to -03 of a degree Centigrade, and gave a mean of 17°-95 C. For the 

 sake of comparison, M. Walferdin observed the temperature at the bottom of 

 a well 75 metres deep, situated at a distance of only 13 metres from the other 

 well, and found it 12°-2C., showing a difference of 5°-75C. in 178 metres, which 

 is at the rate of 1° C. in 30-95 metres, or 1° F. in 56-4 feet. He mentions that 

 he also employed two Six's thermometers (deux thermometrographes) enclosed 

 in copper tubes to protect them from pressure, but both of these gave erroneous 

 indications. The copper case of one was imperfect, and allowed a little water 

 to enter. This one read l°-25 too high, owing probably to the effect of 

 pressure ; the other read 2°-15 too low, owing probably to the index being 

 shaken down. 



The temperature at the depth of 400 metres in the puits de Grenelle at 

 Paris was observed on two difierent occasions. The indications were 23°-5 

 on the first and 23°-75 on the second occasion; and these M. Walferdin com- 

 pares with the constant temperature ll°-7 in the caves of the Observatory at 

 the depth of 28 metres. Taking the mean of the two observations, 23°-6, we 

 have a difference of ll°-9 in 372 metres, which is at the rate of 1° C. in 31-2 

 metres, or 1° F. in 56-9 feet. 



Observations in the well of the Military School, at a distance of 600 metres 

 from the puits de Grenelle, showed a temperature of 16°-4 C. at the depth of 



