Mean Results of Observations. 97 



are there made equal to zero ; which gives as many equations 

 as there are unknown quantities to be determined. This me- 

 thod, if it possessed only the advantage of uniformity, and of 

 freeing the steps of the calculation from all that is indetermi- 

 nate, would be an important service rendered by our brother 

 academician to the sciences of observation ; but it is also that 

 which leaves the minimum of error to be apprehended in the 

 value of each element, as Laplace has proved by the calculus 

 of probabilities. Let us add, in conclusion, that after having 

 calculated the corrections of the elements by the method of 

 least squares, and having substituted their values in the linear 

 expressions of the errors of the observations, if we take the 

 sum of the uneven powers of all these errors, and divide it by 

 the square root of the sum of their double powers, the magni- 

 tude of the quotient will furnish a criterion, according to which 

 observations should be rejected, or their results adopted, if they 

 have, in other respects, a sufficient probability. 



On a Method of rendering Platina malleable. The Bakerian 

 Lecture. By the late William Hyde Wollaston, M.D. 

 F.R.S., &c. 



[From the Philosophical Transactions for 1829. Part I.] 



As, from long experience, I probably am better acquainted 

 with the treatment of Platina, so as to render it perfectly mal- 

 leable, than any other member of this Society, I will endea- 

 vour to describe, as briefly as is consistent with perspicuity, 

 the processes which I put in practice for this purpose, during 

 a series of years, without seeing any occasion to wish for fur- 

 ther improvement. 



The usual means of giving chemical purity to this metal, by 

 solution in aqua regia and precipitation with sal ammoniac, are 

 known to every chemist ; but I doubt whether sufficient care 

 is usually taken to avoid dissolving the iridium contained in 

 the ore, by due dilution of the solvent. In an account which 

 I gave in the Philosophical Transactions for 1804, of a new 

 metal, Rhodium, contained in crude platina, I have mentioned 

 this precaution, but omitted to state to what degree the acids 



JULY— sept., 1829. H 



