of the Solar Rays. '227 



In all these experiments the focus of the violet ray was made 

 to traverse one-half of the needle 200 times, excepting on the 

 11th July and 12th August, when it was done only 100 

 times, and on the 25th July, when it was done 525 times. 



In the experiments with the needles a, 6, c, the spectrum 

 was fixed by means of a heliostate. By this means it was 

 protected against the agitations which the motion of the sun 

 and the displacement of the needles rendered inevitable. We 

 did not, however, always use this instrument, because, in the 

 experiments which we wished to verify, no mention was made 

 of the action of reflected light. It is proper to observe, that 

 the needle had been exposed 17J hours to the action of the 

 sun without becoming magnetic, though M. Morichini re- 

 quired only from fifteen to thirty minutes for completely mag- 

 netising it. 



In order to ascertain that the magnetism had not undergone 

 any change while the needles remained in the violet light, we 

 suspended, in a small earthen vessel, a needle strongly mag- 

 netised and two inches long, and whose south pole (that is the 

 pole directed to the south of the earth) oscillated before the 

 needle under experiment, fixed vertically and submitted to the 

 action of a violet ray, which the heliostate rendered immove- 

 able. The following were the results : 



Hour of day 



In order to repeat the experiments of M. Baumgartner, {Zeit- 

 schrif'ty tom. i. p. 263,) which the magnetic action of the sun 

 presents under another form, we took steel wires from 3 to 

 3.4 inches long, and 0.04 in diameter, polished in different 

 parts, and we fixed them vertically before and after each ex- 

 periment, before the north pole of the small trial needle, 

 situated in the cylindrical tube, and oscillating. We could 

 thus bring the latter to different heights of the steel wires. 

 The numbers in the third column refer to the length of these 

 wires reckoned vertically. i 



li '^, ^ m 



