1836.] of the Wet-bulb Hygrometer. 413 



On inspection of the columns of complementry centesimal depres- 

 sion and centesimal tension in all the foregoing tables, the constant 

 excess of the former is their first predominate feature ; whence the cer- 

 tain conclusion that the ratio is not direct. But to arrive quicker at a 

 conclusion of what it may he, let us view the position of the whole series 

 in diagram 6, PI. XXII. Here the base line designates the hygrometric 

 tensions f"-r-f and the ordinates denote the corresponding centesimal 

 depressions D — d -J-D. If amid such a straggling and scattered nebula 

 it be allowable to trace a normal line, the curve D d will have a pre- 

 ference over any other. Pursuing its dubious course, it passes through 

 the two principal test groupes, upon which more dependence ought to 

 be placed than upon isolated comparisons with the dew-point in still air. 

 Now this line D d nearly coincides with the curve I suggested in 

 1829, from my Benares experiments, making H (or f ,! -±f) follow the 



ratio of D — d ' ; or, calling D = 100, H = D — d '' ; in other 



loo - 

 words, the centesimal tension is as the difference of the actual and the 

 maximum depression raised to the 1.5th power; a form obviously very 

 convenient to be worked by logarithms. This formula has been used 

 for constructing my general table ; and its errors may be judged of by 

 the last two columns of the preceding experiments : but it need by no 

 means supersede the elegant formula d = 87/' — /" when the table 

 is not at hand. The curve corresponding to the latter formula at 90° 

 is also entered in fig. 6. At lower temperature it will have less flexure. 



On the same diagram I have traced the curve of the hair-hygro- 

 meter indications, both according to Gay Lussac's data and those of 

 my original plate in Brandr's Journal, on purpose to shew that the 

 depression curve passes between the two near^the summit : — it was 

 hence I derived the rule for correction of the rough maximum de- 

 pressions, (Table I. II.) by taking it in the direct ratio of the hair- 

 hygrometer indications : and the near accordance of the maxima so 

 deduced, with the observed maxima in dry air, is an additional testi- 

 mony in favor of the assumed parabolic curve. 



It seems an unmerciful increase of the tax upon my reader's patience 

 to extend this train of comparison further : yet it would be hardly 

 fair to omit any thing that can tend to elucidate the subject or assist 

 future investigation : I will not, therefore, forego, through a false and 

 unphilosophical delicacy, the insertion of an abstract I had prepared 

 for my own satisfaction, of three years' comparative deductions from 

 the wet-bulb and hair-hygrometer. It detracts somewhat from its 

 value, that a constant index error of 4 degrees has to be substracted 

 from the readings of the hair-hygrometer during the period in ques- 



