EXPERIMENT STATION BULLETINS. 665 



sary to pass the current through the network or system for more than a 

 second at a time and thus the chance for error froiii polarization is very 

 slight if any. In addition, if one platinizes tlie electrodes well and then 

 selects the proper form of conductivity cell for solutions of high or low 

 resistance, as the case may be, there is little danger of error from 

 polarization. 



INDICATOR OR DRTECTOR. 



There are n number of instruments tliat luive been used to determine 

 wlien the solution wlu)se resistance is desired, is balanced against the 

 known resistance in a Wheatstone Bridge network. The one that has 

 been used more than any other is the telephone. Recent advances in 

 wireless telegrapliy liave had much to do with the improvement of the 

 telephone. It is claimed by some to be the best detector, and because 

 of the improvements made we have the probable reason for its continued 

 use. For precise work the telephone with all its improvements must be 

 used in an absolutely quiet room. An improvised telephone booth should 

 be provided. The operator must also give his undivided attention to the 

 work of listening. The strain on the nerves is severe. It is exceedingly 

 hard even under the.se conditions to detect a "perfect minimum" or a 

 point at which no sound is heard. To determine the exact point where 

 a constant but gradually decreasing sound stops, is difficult. One often 

 imagines he hears a sound when in reality he does not. A continual hum 

 in the head often follows the operator to his recreation for some time after 

 he has left the laboratory. Recently the senior author was told that this 

 trouble in determining the "perfect minimum" was obviated by a certain 

 attachment on the telephone which caused a series of interrupted noises. 

 When these staccato sounds were not heard the balance had been 

 reached. Up to the present time he has been unable to find any article 

 describing the modified telephone but it appears very possible that such 

 a change could be made. An improvement of this sort w^ould be welcomed. 



The telephone at its best is one that is "tuned" to the frequency of the 

 exciting current and this is usually a thousand cycles per second. Even 

 the best telephone, however, is not the best detector. There is a residual 

 tone that cannot be eliminated and which is always present to interfere 

 with the detection of the tone of the machine. The tones most evident in 

 an alternating current are the upper partials, especially the upper third. 

 One can modify the diaphragm of the telephone to eliminate this, but as 

 one corrects one error, one falls into another. This correction would 

 produce a sharp tuning curve and require too close a control of the 

 frequency which in turn is difficult to maintain Avhen there is any varia- 

 tion in the operating current. An extra equipment noted above has to 

 be supplied. With the telephone as the detector a fairly close adiierence 

 to the frequency is necessary. A deviation of 100 cycles per second halves 

 the sensitivity. The difficulties attending the use of the telephone out- 

 lined above led us to select an instrument which would not necessitate the 

 use of the ear. Our best results were obtained with the Alternating Cur- 

 rent Galvanometer with a mirror and scale. This is more satisfactory 

 than the telephone in many respects. 



Probably tlie two greatest factors which influence the sensitivity of 

 any resistance measuring device are the constant of the detector and the 

 magnitude of the current. In the latter case there are certain limitations 



