186 M, Ramond's Ifistritcttomfof' the AppUcdtion of [SEt»t. 



it f)0^sesses sufficient exactness with ihe precautions which I 

 ridopt. I artive at length at the point of congelation, which, of 

 the two fixed points^ is that which has the greatest influence on 

 that part of the scale which coticertis meteorological observa- 

 tions. It will be readily supposed that two thermometers may 

 be considered as being perfect in their cahbres, if they go 

 through this trial v/ithout disagreeing. 



It can Scarcely be necessary to mention that mercurial ther- 

 mometers alone ought to be employed ; and that in those used 

 for taking the temperature of the air, the bulb ought to be 

 entirely separate from the mounting. Small ones are preferable 

 to large, as being more sensible and easier to use. It will suffice 

 that the scale be lar^e enough to subdivide the centigrade 

 degrees by estimation mto tenths. A greater precision would 

 be superfluous, for the temperature of the air is rarely sufficiently 

 Constant for the uncertainty of the observation not to exceed a 

 tenth of a degree. 



Situation of the Instruments, and Method of observing. 



The barometer ought to be iri a perfectly vertical position ; if 

 it be not constructed in such a way as to take this position of 

 itself, means must be used to place, and keep it so. 



It is proper to keep it in a close place, the temperature of 

 which varies but little, or changes very slowly. We shall thus 

 be the more sure that the correctional thermometer expresses 

 faithfully the temperature of the instrument. In order to profit 

 by this accordance, the correctional thermometer should be 

 observed before the barometer. Since the approach of the 

 observer may modify the superficial temperature, and act on the 

 thermometer before the variation has had time to be propagated 

 to the tube of the barometer, which resists the communication 

 as well by its mounting as by its volume. 



In general, a northern or easterly aspect is preferable to that 

 of the west or south. The impetuous winds which blow from 

 these latter quarters occasion, by ilnpinging on the walls, com- 

 pressions of the air, which tlie mercury mdicates by oscillations 

 often very considerable, and always very inconvenient; the same 

 thing will also happen if the walls or roofs opposite to the place 

 of the barometer reflect, or disturb in different directions the cur- 

 lents of air which strike against them ; and I have seen in such 

 positions the column of mercury not only oscillating so as to 

 render observation impossible, Ijut sustained for whole hours 

 above or below the point at which it stood in moments of calm. 

 We must be guided by the local circumstances ; we have only to 

 choose for the barometer the situation Where these causes will 

 affect it least. 



As to the thermometer, it oUght to be frfeely iexposed to the 

 air, but should nevet* be in the sun. In this last respect a north-t 

 em exposure is the oiily one which is suited to it. But we must 



