and Resistance of a Galvanic Circuit. a ee 
chord, besides which the circuit ene contained 200 centi- 
meters of thin copper wire. As to th ee of the internal 
resistance I have to make the following statements. 
The vertical columns (a) vt cme i exhibit this resistance 
when calculated after the rule, viz: combining the 
— nora 4 with every lowest inecinie 
able v1, the column (a) gives an increase of internal resist- 
a if only the mean values oof the first, and last, five acne 
vations are taken, from 103-95 to 128: 51, or from 1 to 1:24, 
In table vit, the corresponding column (a) gives, from the first 
and last, three values, an increase from 6°5 to 10°55, or from 1 to 
1:61. For the whole r range of intensities, eee the limits 
1:982 and 07844, the increase of resistance therefore is as 1 to 
1 24X1°61, or as 1 to 1996. The two centimeters of aoeide 
count for. 
But thinking the matter over I found why this way of calcu- 
lation is wrong. The first direct observation belongs to a cir- 
pi with high intensity, and where there is no part of it heated 
nsiderably. With this observation there are to be combined, 
first, the observation of an intensity somewhat lower, but the 
measuring part of the circuit heated very much; second, the 
measuring unit of resistance hented aa cenit and Pe 
two observations of low intensities at which this unit of resist- 
ance is heated but very little. Hstimating then the ratio of 
nerease, da ue to the difference of temperature of the unit of 
give the ratio of increase of the internal resistance, due to the 
difference of intensity of the current. Of course, it would be 
much better to provide means for keeping the unit of resistance 
at constant temperature, but I refer to the method of determin- 
ing the constants of the battery as commonly p 
‘In order to eliminate as much as possible, the errors of obser- 
