Low Frequency and their Resonance. 513 



before, I prefer for the present to discuss those cases in which 

 there is an apparently striking disagreement between theory 

 and experiment. 



When the maximum point is reached then the slightest 

 variation of the capacity one way or the other causes a very 

 large variation in the potentials. The curve of potentials looks 

 and behaves just like a sensitive flame. The maximum po- 

 tential is at about 1"66 microfarads. There is however another 

 maximum at about 0*18 microfarads which would correspond 

 to the first upper harmonic. This second maximum has been 

 determined with exceedingly great care, so that there is not 

 the slightest doubt about its existence. From the shape of the 

 curve one is led to infer that the form of the impressed e. m. f. 

 is given by 



E =z a i sin pt -f- a 3 sin 3 pt. 



where the amplitude a 3 is exceedingly small. This was rather 

 surprising, since, owing to the peculiar shape of the armature 

 one would have expected a much stronger development of the 

 upper harmonics. 



Before I obtained the alternator just described my experi- 

 ments on the rise of potential by resonance were all performed 

 by means of alternating currents obtained from the interrupter 

 described in my first paper.* In these experiments, an ac- 

 count of which will be given at some future time, the upper 

 harmonics appear very strongly in curves corresponding to the 

 curve in fig. 4. In fact, the crest corresponding to an upper 

 harmonic can be made considerably higher than that corres- 

 ponding to the fundamental. In the experiment with our 

 large alternator giving about 100 periods per second, which 

 experiment I described in my last paper f the upper har- 

 monics seemed to be present in strong force although the 

 armature of this alternator has no iron projections nor is the 

 machine in any other respect as apt to generate a complex 

 e. m. f. as the small machine described above. The considera- 

 tion of the circumstance that since the small machine generated 

 an e. m. f. of nearly four times the frequency of that generated 

 by the large machine and that therefore in the case of the 

 small machine the self-induction is much more effective in 

 destroying upper harmonics did not seem to explain matters 

 quite satisfactorily. There was evidently something going on 

 in my circuits during those experimentsf that I did not under- 

 stand clearly. 



* This Journal, April, 1893. 



f This Journal. May, 1893, pp. 426, 427 and 429. 



