238 SCIENTIFIC AND TECHNICAL LABOURS TO 1860. 



mechanician. This attempt, however, fell through, 

 because the resistance of the wire changed in course 



o 



of time and the copies supplied showed values varying 

 as much as ten per cent from one another. The 

 resistance of a German mile of copper wire of one 

 millimetre diameter at first employed as unit by Halske 

 and myself, and pretty generally adopted in Germany 

 and other countries for practical telegraphy, proved 

 also to be only a makeshift. I soon became convinced 

 that it is quite impracticable to set up an empirical 

 standard in the manner of Jacobi, as the electrical 

 resistance is not such a fixed and controllable property 

 of bodies as (say) the dimension and mass of solid 

 bodies. There was also no prospect of inducing the 

 whole world to accept a standard of resistance de- 

 posited in any particular place. 



On these grounds the choice remained between 

 the absolute unit of resistance of Weber and an 

 empirical unit everywhere reproducible with the greatest 

 exactitude. Unfortunately the adoption of the absolute 

 unit was not then to be thought of, its reproduction 

 being too difficult, so that William Weber himself de- 

 clared to me that errors amounting to a considerable 

 percentage were unavoidable. I decided therefore to 

 take, as the basis of a reproducible standard of re- 

 sistance, the only metal fluid at ordinary temperatures, 

 mercury, whose resistance cannot be affected by mole- 

 cular variations and is influenced less by changes of 

 temperature than that of the solid metals available 

 for the gauging of resistances. In the year 1860 my 



