1823.] the Barometei" to the Measurement of Heights, l&f[ 



also take care that it be beyond the reach of heat reflected by 

 the ground, by walls, and roofs opposite. In our houses, we 

 cannot place it too high : it does not do well except in the upper 

 stories ; and we may very well fix it on the sash frame, the 

 shutter, or side of a window. The air ought to circulate freely 

 round it. I usually suspend it by a hook, the arm of which is 

 about six inches long. A ring fixed at the end of another arm of 

 the same length eral3races the instrument at its lower end, and 

 secures it from being moved by the wind. I fix this small appa- 

 ratus on the outside of a window upon the sash frame, in sucn a 

 manner that the thermometer may be easily observed without 

 opening the window. 



But in thus exposing the thermometer to a free circulation of 

 air, we must at the same time take care to defend it from the 

 immediate contact of snow, hail, or rain ; as often as it is touched 

 by any of these, it is no longer the temperature of the air, but 

 that of the meteor in question which it indicates. This object 

 is attained if the roof have a sufficient projection. I prefer, 

 however, a small moveable pent-house, placed at a convenient 

 height, and which we may let down only when requisite ; except 

 in these cases, a shelter is rather pernicious than useful. In 

 those winter nights, for example, when the calm of the atmo- 

 sphere, the serenity of the sky, and the twinkling of the stars, 

 announce a sharp frost, the thermometer will not indicate the 

 whole intensity of the cold if a shelter be interposed between it 

 and the particles of air, which, after being condensed in the 

 middle regions, fall vertically upon the earth in an invisible 

 shower. It ought in this case to be uncovered for the same 

 reason that we cover an espalier which we wish to preserve from 

 the frost. 



When once the thermometer is well placed, the observation of 

 it does not present any difficulty. The only attention which it 

 requires is that of holding the eye exactly on the level of the 

 point observed ; for if we raise or depress it, if the visual ray 

 deviate from the perpendicular to the axis of the tube, the sur- 

 face of the mercury will successively correspond to different 

 divisions of the scale. It will appear lower if we look at it from 

 a higher situation ; and higher if from a lower ; and the error 

 will be proportional to the angle which the visual ray forms with 

 the perpendicular : this angle is what is termed the parallax. 

 In observing the barometer, this source of error is annulled by 

 means of the index or ring which fixes the fine of sight. This 

 method cannot be applied to the thermometer, which must be 

 observed at a distance, and never handled. Attention will 

 supply the want of it, and gradually become a habit. 



It is useless to take greater precautions than these. The 

 temperature of the air is often so inconstant, and undergoes so 

 many alterations in the places in which we are often obliged to 

 make our observations, that it would be in vain to seek in the 



