EXPERIMENT STATION BULLETINS. 563 



coil does not give a current whose integral value is zero. In other words, 

 the current is not alternating; merely pulsating and having a uni- 

 directional component. This condition of affairs gives rise to polar- 

 ization because of a counter e. m. f. and consequently a false reading. 

 Altho we had assumed that such errors existed, and our assumption ap- 

 peared justified, they had not been experimentally verified. Dr. Wash- 

 burn, however, has recently shown by means of curves obtained from 

 oscillograms, that these errors were present. Fig. 3 is a curve taken from 

 an oscillogram of an induction coil current. It shows clearly that the 

 current is not alternating. Fig. 4 represents the e. m. f. wave when the 

 induction coil is operating well. By plotting the areas of the positive 

 and negative loops with a suitable measuring device it will be seen that 

 they are not equal, therefore the e. m. f. in one direction is greater than 

 in the other and we have a unidirectional current. This same conclusion 

 was reached by Dr. Washburn from another experiment. When the cur- 

 rent from the secondary of an induction coil is passed through a silver 

 coulometer a deposition of silver takes place, hence there is a polarization 

 at the electrode and the current is unidirectional. The introduction of 

 another electrolytic cell in the adjacent bridge arm has been suggested as 

 a means of balancing out the polarization effect. This unfortunately does 

 not correct the trouble because the cell itself is a source of error especi- 

 ally on account of the temperature uncertainties. 



When used with the telephone the induction coil is not free from over 

 tones and the frequency is not always constant for a sufficient length of 

 time. If it should happen to maintain a constancy for a sufficient time 

 it is apt to change in tone unexpectedly. Thus an improvement in the 

 telephone, which recent work in wireless telegraphy has made ])ossible 

 would be of no advantage. Therefore, for precise work, the induction 

 coil is not satisfactory. 



The Siemens, and Halske high frequency generator, in connection with 

 the telephone, has been used, but this only produces an api>roximate sine 

 wave even when operated on a current of constant potential. To protect 

 it from varying loads and to detect such variations, quite an amount of 

 auxiliary apparatus is necessary. A 110 volt generator to drive the ma- 

 chine; two mica condensers to vary the capacity of the circuit and dis- 

 tribute it symmetrically in resi)ect to the generator; a device for detecting 

 the frequency and to indicate when corrections should be made; two 

 rheostats, one to control the speed of tlie motor and the other to vary the 

 excitino- current emoloved durina: a conductance measurement : and a hot 

 wire ammeter to indicate the condition of the current, constitute the 

 necessarv apDaratus. Such an array of apparatus will need constant 

 care and watchiuir to keen in aood workiuir order. It cannot be assembled 

 in one place. The operation of certain parts interferes with the proper 

 working of others. The generator because of the noise produced by the 

 rapidly revolving toothed wheel must be put away in a room some dis- 

 tance from that in which the measurements are being made. The operator 

 must be in a quiet place in order to determine the minimum tone through 

 the telephone. This auxiliary apparatus costs in the neighborhood of 

 |300 and if it constituted all that was necessary for determining a con- 

 ductance measurement, an objection might not he forthcoming. This is, 

 however, only a small part, and simply arranges for a current of constant 

 potential to the bridge. Probably the best source of constant potential 



