1823 .J. the Barometer to the l^edsuremeiit of tteigtits. lH 



perhaps come currents of air which modify the causes of error ; 

 an interval of calm restores to them their energy ; the interven- 

 tioa of clouds instantly suspends their action : in the midst of 

 such a complication of effects nothing seems clear except motives 

 for doubt ; and the observer is neither unsuccessful nor unskilful 

 if he knows, within one or two degrees, the mean temperature 

 of his instrument. 



I mention these inconveniences, because it is necessary to 

 have a just idea of them in order to be in a situation to obviate 

 them according to the exigency of the case, and the means 

 which the situation or chance place at our disposal, when fore- 

 sight has not been able to provide against them. A rock or a " 

 tree frequently afford a convenient shelter. We may supply the 

 want of them, at least in part by a man placed between the sun 

 and the instrument : by a piece of Hnen fastened round the tripod = 

 which supports the barometer ; or for want of any other resource, 

 by making the shadow of one of the legs fall along the tube, and 

 especially over the cistern. The thermometer ought always to 

 be turned away from the sun. When the alternations of wind 

 and calm cause variations too sudden and too great, I cover the 

 bulb in such a way as to defend it to a certain extent from these 

 passing and capricious sources of variation. The action which 

 they exert on the thermometer will induce an error upon the 

 temperature of the instrument, because these very transient 

 variations may have time to make an impression on the surface 

 of the mounting, but not to be communicated to the entire mass , 

 of mercury. 



With respect to the thermometer employed to mark the tern- ■ 

 perature of the air, it is always in the most elevated, the most 

 exposed, the most airy situation, that its place ought to be 

 chosen. This condition is much more easily fulfilled in an open 

 country, and on the summit of a mountain exposed to all winds; - 

 than in buildings where we commonly make meteorological* 

 observations. There our stationary thermometers have commu- 

 nication only with half the surrounding atmosphere : the other 

 half is kept from them by the wall against which they are placed; 

 but this disadvantage is compensated by the facility of placing 

 them at an elevation where they are secure from the effects of 

 reverberation from the earth : this resource is wanting in moun- 

 tain observations. We cannot place the thermometer higher 

 than the point where we can observe it without parallax ; and at 

 this elevation, which does not exceed that of the human body, 

 the instrument is far from being out of the reach of the earth's 

 influence ; this inconvenience is unavoidable : we ought, there- 

 fore, to lose none of the advantages which accompany it. It is 

 not without reason that Saussure condemns the practice of sus- 

 pending the thermometer from a body of any magnitude. He 

 attaches it to a simple staff, the shadow of which, directed upon 

 the bulb, is sufficient to shade it from the sun, and the diameter 



