410 Experimental Researches on the Depressions [July, 



agree remarkably well, and it does not materially signify, nor is it 

 perhaps possible to certify which multiplier is to be preferred. Pro- 

 fessor Apjohn's has the merit of coinciding precisely at the tempera- 

 ture of 1 90° with my steam experiment ; but for the range of lower 

 and more practical temperatures it is perhaps slightly in excess. 

 The simpler expression of " one- eightieth of the depression = the 

 aqueous tension at t'" would there be nearer the mark ; and would 

 be easier of application. From my own experiments I deduced a 

 mean of D = 84 /' with which I constructed the table at the conclu- 

 sion of this paper, but I must in fairness acknowledge that its prefer- 

 ence to Professor Apjohn's rule is nearly evanescent in practice. 



§ 2. — Value of depressions less than the maximum, in centesimal hygro- 



metric tension. 



We are now arrived at the second subject of inquiry, which is in 

 fact of more practical importance than the first, since it includes every 

 observation that can be made in an atmosphere never reduced to a 

 state of absolute siccity. 



The simplest condition of the case of intermediate depressions would 

 be that assumed by Dr. Hudson, viz., that the maximum depression 

 being divided into 100 parts, each part should indicate one hundredth 

 of the moisture of saturation at the given temperature, or D : d : :f : 



/'-/"■ 



But such a law is not found to prevail in reality : nor is it analo- 

 gous to the course of nature that it should exist in the case of the 

 wet-bulb thermometer, when the hair-hygrometer and the law of evapo- 

 ration require different ratios. It is more consonant with theory*, as 

 it proves to be with practice, that the tendency to evaporation, 

 and the cold consequent upon it, should increase in a geometrical 

 ratio to the dryness of the air. 



* The depressions will, ceteris paribus, be less, the more aqueous vapour is 

 previously contained in the air, because the specific heat of a given volume of 

 vapour being .529 (or .847X-625) while that of air is .267, the specific heat 

 of any mixture of the two must exceed that of air alone. But the curvature 

 imparted to the line of depressions from this cause may easily be shewn to be tri- 

 fling. Thus at the temperature of 80° where/ = 1.00 inch ; the capacity of dry 

 air being e, that of moistened air will be c Xp—f'+c'X.f"; whence, calling e=l, 



p 

 for saturated air we should have the specific heat 1.053 ; and for half- saturated 



air, 1.031 ; and the depressional degrees at those points will be inversely so 



much less than those at the dry extremity of the curve. Were the other agents 



easily evaluated, we might through this means verify the specific heat of 



aqueous vapour. 



