222 PROCEEDINGS OF THE AMERICAN ACADEMY. 



given spark distance at A, was 0.6. When the zigzag was replaced 

 by the spiral, the critical spark length at B rose to 6.4. A bundle of 

 finely divided iron was now inserted in the spiral, and the critical 

 length continued still to be 6.4, 



He remarks on this result as follows : '* Here is magnetic time- 

 lag raised to an extreme. ... It may be said that the iron fails to 

 get magnetized because of the opposing action of the inverse ' Fou- 

 cault ' current induced in it, just half a period behind the inducing 

 currents. I thought this would be so, of course, with thick iron rods, 

 but with a bundle of thin wires I felt doubtful. . . . Whatever the 

 explanation, the fact of time-lag is patent. Yet there is something 

 strange about it, for that a steel knitting-needle can be magnetized by 

 discharging a Leyden jar round it is mentioned in every text-book, 

 and it is certainly true. There are points here requiring further 

 examination." * 



So far the investigations that had in view the effect of iron upon 

 extremely rapid rates of oscillation have given but negative results, 

 though both investigators quoted expected to find that the magnetic 

 properties of the iron would be shown under such conditions. Some 

 positive results showing that the magnetic properties of iron still have 

 some effect upon rapid electric discharges have been obtained by the 

 following observers. 



Professor John Trowbridge has proved f that the magnetic char- 

 acter of a conductor is by no means unimportant with 1,000,000 

 double oscillations per second. In brief the experiment and the 

 results were as follows. The oscillating circuit consisted of a Leyden 

 jar and two parallel wires 30 cm. apart and 510 cm. long. These 

 parallel wires could be replaced by others of different diameter and 

 material. A spark micrometer with tin terminals was included in 

 the circuit, and when the discharge occurred the spark w^as photo- 

 graphed by means of a rapidly revolving mirror.f 



The following results bear upon the subject under investigation. 

 When the parallel wires were of copper (diam. 0.087 cm.), the number 

 of double oscillations on the negative averaged quite uniformly 9 or 

 9.5 ; but when an annealed iron wire (0.087 cm. diam.) was substi- 

 tuted, only the first return oscillation was distinctly visible, with 

 sometimes a trace of the first duplicate. 



* Lightning Conductors and Lightning Guards, p. 48. 

 t Proceedinfrs of the American Academy, XXVL 115. 

 t Ibid., XXV. 10':), 



