INFLUENCE OF AURORA ON THE TELEGRAPH. 431 



color up to about 11 o'clock, when the whole northwestern part of the 

 heavens was a brilliant red, reaching to the zenith, the color fading 

 gradually to the north to a strong white light, and then rising again in 

 the northeast, but not so brilliantly as in the northwest. At this time 

 (11 p. m.) I came to the of&ce ami had one of our lines disconnected from 

 the batteries and the ends grounded at Harrisburg and Philadelphia. 

 This arrangement gave a current a trifle stronger than the regular bat- 

 teries and in the same direction, galvanometer deflecting to the right. 

 At 11.45, after one or two breaks, the current changed, the galvanometer 

 needle deflecting to the left ; at 11.55 back to the right ; at 12.10 to the 

 left, where it remained until 1 o'clock, when I went home. 



After the galvanometer needle deflected to the left, at 11.45, the current 

 became very unsteady and weak; was only sufficiently strong to move 

 the relay for a few minutes at a time. The galvanometer showed a 

 current of varying strength all the time. 



Sunday, September 25th, the aurora again appeared about 8.30 p. m., 

 but not near so brilliant as on the former evening. 



I had a wire connected as before, viz, to the ground at Harrisburg and 

 Philadelphia. This produced no effect whatever on the relay. The gal- 

 vanometer at 8.40 deflected to the left; 8.45, right; 9, left; 9.20, right; 

 9.21, left ; here the current was very changeable, the needle flattering 

 from side to side ; at 9.22 aga,in to the right, where it remained until 

 10 p. m., when I went home. 



The galvanometer I have is Dr. Werner Siemen's universal galvan- 

 ometer. 



[The change in direction of the current may possibly have been 

 due to the greater action of a beam of the aurora on the easterly part 

 of the wire between Harrisburg and Philadelphia, and afterward of 

 another beam on the westerly part of the wire. It is therefore import- 

 ant to note whether a change of a similar kind takes place in a wire 

 extending north and south. — J. H.l 



