300 M. F. Kohlrausch on the Absolute Value of the 



Less ground for distrust would no doubt be afforded if the 

 expression that has been quoted were interpreted as meaning 

 that a change had taken place, through gradual elastic yielding, 

 in the position of torsional equilibrium of the cocoon-fibre, as 

 might be the case, for instance, if the fibre had not been sus- 

 pended for a sufficiently long time. But even this I do not 

 like to suppose ; for in such a case the observations ought to have 

 been postponed, or at least the yielding should have been elimi- 

 nated by means of special observations. 



In short, we can hardly avoid coming to the conclusion that 

 the feeble magnetism of the needle gave rise to considerable 

 discrepancies ; and it does not appear to me allowable to apply 

 the rules for the calculation of probabilities to such observations 

 unless we have explicit proof of the absence of constant sources 

 of error. 



In any case, however, the probable error of 0*1 per cent., 

 which is calculated for the final result, can only apply to the 

 scale-readings ; to extend it to the entire measurement is pre- 

 supposing that there were no other sources of error. Even the 

 very noteworthy agreement between the numbers found in the 

 two years 1863 and 1864, which differ by only 0*16 per cent., 

 cannot be regarded as unconditionally decisive. If we now in- 

 quire for other sources of error, we. encounter first of all the ob- 

 jection raised by Dr. W. Siemens against calculating the mean 

 radius of the coil from the length of the wire and the number 

 of turns. I think I may conclude, from careful experiments 

 of my own, that this process may be employed without hesita- 

 tion in the case of thick wires; but the section of the wire 

 which we have to deal with here (to judge from the weight and 

 total resistance of the wire, as well as from the dimensions 

 of the coil) cannot have been more than about 1 millim.; so 

 that the above objection may perhaps be regarded as not out of 

 place. But, in any case, the error thus arising cannot have 

 been great. 



But possibly another circumstance (which is not alluded to 

 in the Reports of the Committee, though in general the correc- 

 tions are treated with a comprehensiveness and completeness 

 that is worthy of imitation) may have been of greater import- 

 ance. The support in which the coil revolved consisted of a 

 " strong brass frame," which, as appears from the figure, 

 formed within itself a closed circuit. It is nowhere stated how 

 the observers convinced themselves of the unimportance of the 

 currents which the closed inductor must have caused in these 

 stationary metallic parts when rotating. No doubt the experi- 

 mental detection of these currents would have been difficult ; but 



