184 Mr. F. Y. Edgeworth on a New Method of 



Table II. gives the apparent electrochemical equivalent of 

 copper for the different current-densities and temperatures 

 stated in the table; it is simply the numbers in Table I. mul- 

 tiplied by the proper constant. The results in this table are, 

 as in the former paper, based on the results obtained by Lord 

 Rayleigh and Mrs. Sidgwick for silver, or '001118 grm. of 

 silver deposited per coulomb of electricity passed through the 

 cell. Tbe ratio of the electrochemical equivalents of copper 

 and silver obtained in the experiments described in my former 

 paper has been assumed as correct (Phil. Mag. Nov. 188(5). 

 It is '2940 when the copper plate presents a surface of 50 square 

 centimetres per ampere of current. 







Table 11. 







Area of cathode, 

 in square centi- 

 me! ns per am- 

 pere of current. 



Tempera- 

 ture, 

 2° C. 



Tem pera- 



ture, 



12° C. 



Tempera- 

 ture, 

 23° C. 



Tempera- 

 ture. 

 28° C. 



Tempera- 

 ture, 



35° C. 



50 



100 



150 



200 



250 



300 



•0003288 

 •0003288 

 •0003287 

 ■0003285 

 ■0003283 

 ■0003282 



•0003287 

 ■0003284 

 •0003281 

 ■0003279 



■0003278 

 •0003278 



•0003286 

 •0003283 

 •0003280 

 ■001 13277 

 •000.3275 

 ■0003272 



■0003286 

 ■0003281 

 •0003278 

 •0003274 

 ■0003268 

 •00U32G2 



•0003282 

 ■0003274 

 ■0003267 

 ■0003269 

 ■0003252 

 •0003245 



XXII. On a New Method of Reducing Observations relating to 

 Severed Quantities. By F. Y. Edgeworth, M.A., Lecturer 



at King's College, London*. 



A SUBSTITUTE for the Method of Least Squares has 

 been proposed by mef, based upon the following prin- 

 ciple. The data being of the form 



OxW + biy . ..— «j|=0 ; a»:r + h. : y . . . — r. 2 =0 ; &.c, 

 (where v v v. 2 , &c. are observations of equal worth), a solu- 

 tion is obtainable by taking an, y . . . such that the sum of the 

 residuals (the left-hand members of the above-written equa- 

 tions), each residual taken positively, should be a minimum \. 



* Communicated by the Author. 



t Philosophical Magazine, August 1887; and Hevmathena (Dublin ), 



1887- 



X This rule is derivable from the hypothesis that the law of error, the 



facility-curve under which the observations range, is of the form y= 9 e~ hl \ 



x taken positively in both directions. But the use of the rule does nut 

 commit us to the assumption of the hypothesis. The Method of Least 

 Sum is in this respect exactly on a par with the Method of Least Squares. 

 The rule of the latter Method is — Determine .?• and y so that the sum of 



