130 CAUSES which qffeSl the ACCURACY, &c. 



the unavoidable errors of obfervation, (which may amount to 

 five feet), may be confiderable in refpect of the whole ; nor fo 

 great as to introduce error from other caufes. It ought not, 

 therefore, to be lefs than 100, nor much greater than 500 feet. 



45. IN this manner, we (hall have a meafure, not indeed of 

 the abfolute quantity of humidity diilblved in the air at a given 

 time, but of the differences of the humidity diflblved in it at 

 different times. Our hygrometer, therefore, will afford a fcale 

 for the meafuring of moifture, not unlike that which the ther- 

 mometer affords for the meafuring of heat ; and both deduced 

 from the changes produced on the bulk, or the fpecific gravity 

 of certain bodies. The beginning, or zero, of this fcale may al- 

 fo be fixed by a certain and invariable rule, if we aflume m, in 

 the preceding formula, (or the expanfion of air for one degree 

 of heat), of a given magnitude, as, for inflance, .00245, and 

 conceive the fcale to begin when o, or when the formula, 

 thus adjufled, gives the true height. 



THE hygrometer with which we will be thus furnifhed, 

 feems well adapted to the purpofes of aftronomy. For it mea- 

 fures the humidity chemically united with the air, and not 

 merely the difpofition of the air to depofit that humidity, 

 which, though much connected with the changes of the wea- 

 ther, has little to do with the aftronomical refraction. It is 

 true, that the fractions n may not be directly proportional to 

 the differences of the humidity of the air, nor to the changes 

 of refracting power, which thofe differences of humidity may 

 produce ; but they are probably connected with thefe laft, 

 by fome fixed and invariable law, which future experiments 

 may be able to afcertain. Nor can this application of the ba- 

 rometer fail of leading to fome ufeful conchifion ; for if, on 

 trial, it fhall be found, that the operation of humidity in 

 changing the fpecific gravity of the air, is over-ruled or con- 

 cealed by the action of more powerful caufes, the difcovery, 

 even of this fact, will give a value to the obfervations. 



IV. 



