1884.] The Thomson Effect and Properties of Metals. 27 



The first column contains the names of all the metals given in 

 Everett's table, with the exception of alloys. 



The second column gives the numerical coefficients of t in Everett's 

 table ; these numbers are proportional to the coefficients of the 

 Thomson effect. 



The specific heats in the third column are Begnault's determi- 

 nations, as given in Pickering's " Physical Manipulation," vol. ii, 

 p. 287. 



The specific resistances in column IV (except that of tin) [are 

 Matthiessen's, as given in Jenkin's " Electricity " and Pickering's 

 " Physical Manipulation." The resistance of tin is reduced from 

 E. Becquerel's determination, on the assumption that the specific 

 resistance of silver at 15 C. is G'016. Matthiessen's is considerably 

 higher, viz., 0'134. 



The fifth column contains the products of the third and fourth 

 multiplied by 10 6 to get rid of decimals. 



The coefficients of expansion in the sixth column are taken from 

 the "Encyclopaedia Britannica," 9th edition, article " Heat," by Sir 

 William Thomson. The authorities are given in the seventh 

 column.* 



The eighth column gives the squares of the coefficients of expansion 

 divided by 34 2 .f 



The numbers in the ninth column are obtained by subtracting 

 those in the eighth from those in the fifth column. 



In the last column the numbers in column IX are divided by 

 2,400. 



It will be seen that, with one exception, the order of magnitude of 

 the numbers in column IX (proportional to spec, heat X spec. res. X 10 6 

 (Exp.-f-34) 2 ) is exactly the same as the order of those in column II, 

 which are proportional to the coefficient of the Thomson effect. The 

 rate of decrease in column X is, however, not the same as that in 

 column II, the numbers diminishing too rapidly in the upper half of 

 the column and too slowly in the lower half. Although, therefore, it 

 appears very probable that the direction and magnitude of the 



* Where more than one value was given the first was always used except in the 

 cases of silver and zinc. The expansions for silver are Muschenbrock's, 2120 ; 

 Kupffer's, 1910; Matthiessen's, 1973; those for zinc are Calvert's, 2200; 

 Matthiessen's, 2976 ; Fizeau's, 2918. In both cases there is a fair agreement 

 between the second and third values, while the first differs from them considerably. 

 The second values give silver and zinc the same places in column X as in 

 column II ; the first would change their order. 



t The divisor was so chosen that, while the ratio of the first number to the last 

 in column IX should be as nearly as possible equal to the ratio of the first number 

 to the last in column II, the number corresponding to lead in column IX should at 

 the same time be as near zero as possible. Both conditions could not be exactly 

 fulfilled at once. 



