130 WRITINGS OF JOSEPH HENRY. [1838 



beginning and ending cannot be distinguished from each 

 other, and we can only observe a single impulse, which 

 however may be considered as the difference of two impulses 

 in opposite directions. 



91. The first experiment happened to be made with a cur- 

 rent of the fourth order. The magnetizing spiral (11) was 

 attached to the ends of coil No. 4, Fig. 9, and by the polarity 

 of the needle it was found that this current was in the same 

 direction with the secondary and primary currents. * By a 

 too hasty generalization, I was led to conclude, from this 

 experiment, that the currents of all orders are in the same 

 direction as that of the battery current, and I was the more 

 confirmed in this from the results of my first experiments on 

 the currents of ordinary electricity. The conclusion however 

 caused me much useless labor and perplexity, and was 

 afterwards proved to be erroneous. 



92. By a careful repetition of the last experiment, in refer- 

 ence to each current, the important fact was discovered, that 

 there exists an alternation in the direction of the currents of the 

 several orders, commencing with the secondary. This result was 

 so extraordinary, that it was thought necessary to establish 

 it by a variety of experiments. For this purpose the direc- 

 tion was determined by decomposition, and also by the gal- 

 vanometer, but the result was still the same; and at this 

 stage of the inquiry I was compelled to the conclusion that 

 the directions of the several currents were as follows : 



Pi-imary current + 



Secondary current + 



Current of the third order — 



Current of the fourth order + 



Current of the fifth order — 



93. In the first glance at the above table, we are struck 

 with the fact that the law of alternation is complete, except 

 between the primary and secondary currents, and it appeared 



* It should be recollected that all the inductions which have been men- 

 tioned were produced at the moment of breaking the circuit of the battery 

 current. The induction at the formation of the current is too feeble to pro- 

 duce the effects described. 



