EFFICIENCY OF ELECTRIC GENERATORS. 99 



The fall of potential at the terminals of the machine, which 

 was 133 volts, came down to 27 volts. 



The maximum of mechanical efficiency was * 56 ; it corre- 

 sponded to a capacity of 35 75 kilogrammetres and to B = 2 1 r. 

 It is evident that under this regimen, the machine was very 

 badly utilised ; but the partisans of the equality of resistances 

 could object that should wires of a suitable diameter be wound 

 on the skeleton a much larger capacity could be obtained without 

 increasing the value of E. Let us admit that the electromotive 

 force obtained in the second series of experiments is sufficient 

 for the use for which the machine is intended, and let us try to 

 obtain it with a practical speed as great as possible ; the normal 

 type of electroplating machine is precisely constructed with 

 such objects in view, and it is in experimenting upon it that we 

 have established table No. 3. 



The machine could give 154 amperes without any dangerous 

 heating. The maximum working capacity was 2908 watts, with 

 an electrical efficiency of 84 and a mechanical efficiency of 

 78. The speed was, as in the first case, 1750 revolutions per 

 minute. The external resistance was five times greater than the 

 internal resistance. The work absorbed by friction was 23 kilo- 

 grammetres. 



This last series of experiments proves that the size of the 

 wound wire has no influence on the equality of the resistances. 

 With any given well-studied skeleton, a winding suitable for 

 the work to be effected can be combined by using either short 

 thick wires or long thin ones. In any case the maximum 

 capacity will be obtained with an external resistance much 

 superior to the internal resistance. 



It is true that if on the last machine experimented upon we 

 had stripped the electro-magnets a little, we might have made 

 the maximum capacity correspond to the equality of resistances ; 

 but the economy would only have resulted in a few kilogrammes 

 of wire, and the work would have been three or four times less 

 than that obtained from a well equilibrated machine. 



Nothing is more difficult than to uproot an error, and especi- 

 ally when it is based on some apparently convincing comparisons 

 or on an incomplete although not false theory. We therefore 



H 2 



