350 Mr. J. D. Chorlton. Examination of some of the [Mar. 19, 



ments, the constant of the old balance as determined by means of the 

 "Kelvin balance " is 0*9766, while that of the new balance is 06316. 

 So that in the case of the old balance I find a constant which is 

 greater than that given by Joule by 2*5 parts in a thousand, while 

 the constant of the new balance seems to come out about 4*6 parts in 

 a thousand less than the value given by Joule. 



My determination of the constant of the new balance is, however, 

 not to be relied upon, because I found it exceedingly difficult to 

 replace the upper fixed coil, which had become unwound and had to 

 be removed in order to be rewound in exactly the same position as it 

 formerly occupied, so that the distance between the two fixed coils 

 may have been altered, and in making an experiment to determine 

 the amount of error which a slight increase in the distance between 

 the two fixed coils would be likely to produce in the constant, I 

 found that on increasing the distance, which is about 1 in., by 1/16 of 

 an inch, an increase of about 1 per cent, appeared in the constant. 



Fortunately, although the upper coil of the old balance had also 

 become loose, there was no difficulty in replacing it exactly in its 

 former position. 



The " Kelvin " balance had been tested by silver electrolysis some 

 months previously, and found to be correct, and in order to assure 

 myself that it had not changed in the meantime, I afterwards again 

 tested it, with the same result. As no trace of leakage could be 

 detected in the " Joule " meter, it would therefore seem that some 

 error must have been made in the measurement of the radius of the 

 galvanometer or in the determination of H. Unfortunately, on 

 examining the tangent galvanometer, I found that the coil used in 

 these experiments, which consisted of a single circle of 1/10 in. 

 copper wire, had got displaced, and could not be replaced exactly 

 in its former position; in fact the measurement of the diameter 

 would have been liable to an error of 1 per cent. 



I have not, then , been able definitely to discover an error in any of 

 Joule's measurements, and can only indicate the probability that on 

 account of error either in the measurement of the radius of the 

 galvanometer or in the determination of the earth's horizontal mag- 

 netic force, the currents in these experiments were estimated about 

 2 5 parts in a thousand less than their real value, and that as the 

 heat developed depends on the square of the current, the value of J 

 given by these experiments may be too small by five parts in a 

 thousand. 



The experiments are divided into three series. The value obtained 

 from the first series was 786*3, from the second 787, from the third 

 782*4; but Joule says that the extra precaution taken in the last 

 series entitles the last figure to be taken as the result of the inquiry. 



The value then obtained for J from these experiments was 782'4 



