1835.] Meteorological Observations. 15 



the basin ; but instead of being divided through out its 

 whole length, a few inches near the top are only markedir 

 except in mountain barometers which require a scale of a 

 xnuch greater range. 



To render the instrument portable, and at the same time 

 delicate, various contrivances have been adopted. One of 

 its most simple forms is that recommended by Delus, coia- 

 sisting of a bent glass tube a 5.,(Fig. 2.) closed at both ends, 

 and with a small hole at d to admit the air. When in- 

 verted, the whole of the mercury occupies the long limb of 

 the tube ; but when in the position represented in the figure, 

 the mercury descends, and part of it flows into the short 

 limb, until the column, between the upper surface e,, and 

 the lower surface exactly equals the pressure of the at- 

 mosphere. The length of this column is very easily mea- 

 sured by means of two scales on the brass case which in-, 

 closes the tube, one of them being placed along the upper, 

 the other along the lov/er limb, and the figures on which, 

 run in opposite directions, and require to be added together. 

 This barometer is very portable, and (with the improve- 

 ments that have been made by M. Bun ten of Paris) is 

 well adapted for the measurement of heights,* but it i^.. 

 scarcely delicate enough for meteorological purposes. 



Another form of the barometer is that in which the lower 

 end of a straight tube, like that in the first figure, dips into 

 the mercury included in a small cistern attached to it. In 

 this case, it is evident that when the mercury descends in 

 the tube it must rise in the cistern, and vice versa ; and ac- 

 cordingly, if the scale be fixed, the mercury will, in the former 

 instance, stand at a higher point, in the latter, at a lower 

 point, than it ought to do. In order to prevent this inac- 

 curacy, most barometers are so constructed, that by means^ 

 of a screw, the surface of the mercury in the cistern may 

 be raised or depressed to the commencement of the scale. 



As the mercury in the barometer expands by heat, a 

 slight inaccuracy would arise in observations made at dif- 

 ferent temperatures, if not corrected. In order to exclude 

 such irregularities, every observation might be reduced to 

 what it would be at the freezing point of water (32"^ Fah- 



* It may be procured, With its latest improvemeiits from Mr. Robiuson, Opti- 

 ciaa, 38, Devonshire Street, London. 



