Mr. Drach on the Hourly Observations at Leith in 1824-25. 43 



This process is cheaper than that of heating chlorate of 

 potash ; for two parts of bichromate of potash will produce as 

 much oxygen as one of chlorate of potash, while the latter is 

 nearly three times the price of the former; and besides this, 

 the residue of the first is valuable, and may be reconverted 

 into bichromate of potash. It is likewise a more convenient 

 process than any at present known, since it may be conducted 

 at so low a temperature that an ordinary retort and lamp may 

 be used for the production of a considerable quantity of oxygen. 

 Mechanics' Institution, W. H. BALMAIN. 



Liverpool, May 10, 1842. 



[Note. — I have tried this process and find that it answers 

 very well, the gas being given off, I think, with greater readi- 

 ness than when sulphuric acid and binoxide of manganese are 

 employed. Occasions I have no doubt will occur in which 

 this method may be advantageously substituted for others. — 

 R. P.] 



X. On Sir D. Brewster's Deductions from the Hourly Ob- 

 servations at Leith in 1824-25. By S. M. Drach, Esq., 

 F.R.A.S. 



To the Editors of the Philosophical Magazine and Journal. 

 Gentlemen, 

 HPHE deductions alluded to in the title of this article, as 

 -■- detailed in the Edinburgh Philosophical Transactions, 

 vol. x., flow from any expression of the temperature in func- 

 tions of the time. Let v = the temperature, / = the time ; 

 — T = a fixed instant ; then to be real v = function of 



J.f+T\\ ( ' + T) \ log (t + T), *™ i (t + T), constant \, 



which is developable into the series 



r; = A+B(* + T) + C (t + T) 2 + D(* + T) 3 + &c. 



A, B, C, &c. are functions independent of the time, and com- 

 prehending the latitude, declination, radiation, &c. 



When t = - T, v = A. 



First. If A = the daily mean temperature, t = — T = time 

 of morning mean, and = B + C {t + T) + D (t + T) 2 , + &c. 

 gives the other times of mean daily temperature. 



There being only one (evening) mean, this series must be 

 very convergent, and 



B B 



t — — T — j^t or more correctly, t = — T ~ ; 



thus B D is very much less than C 2 . 



c -c- D 



