32 Br Brewster on the Register of the Thermometer , fyc. 



Hours. Difference. Hours. Difference. 



H. II. 



8 — 0°.184 8 27 o°.000 



9 —0 .082 



These parabolic abscissa? were calculated by the following 

 formula?. By the property of the parabola, we have 

 BH : Bm = AH 2 : m?i 2 ; and 

 Bm= BH_x ot»?. 

 AH 2 

 But since A£ is the line of mean temperature, pn the de- 

 pression of the temperature below the mean at the point of 

 time p, and pn = Hm ~ HB — Bm, then, calling m the mi- 

 nimum temperature, and y the ordinate mn, we have the re- 

 quired temperature T at the time p, thus : 



~ T AH 2 

 For the semi-parabola BC, 



T CH 2 

 For the semi-parabola CD, M being the maximum tempe- 

 rature, 



T--M — GD *^ 

 CG 2 



For the semi-parabola DE, 



T = M — GD x ^ 2 

 EG 2 ' 



Upon comparing the differences in the preceding tables, it 

 appears, that the greatest is a quarter of a degree of Fahren- 

 heit, and that they are most perceptible in the afternoon 

 branch of the curve, between 4 P. m. and 8 p. m. 



I have no hesitation, however, in saying, that the mean of 

 a greater number of years will produce a close approximation 

 to the parabola. In 1824, the afternoon branch is irregular. 

 In 1825, which was a year of uniform character, the after- 

 noon branch becomes more convex, and approaches closely to 

 the parabolic branch ; so that the mean of 1824 and 1825, 

 which we have contrasted with the parabolic abscissa?, partakes 

 of the irregularities of 1824, and thus occasions a flatness in 

 the curve, and consequently the differences observed between 

 4 r. m. and 8 v. m. 



