594 
a needle, while yellow, orange, and red, produce no sen- 
sible effect. As many philosophers have failedin repeating 
these experiments, we were induced, in the course of the 
summer, to undertake the investigation of this subject, 
‘* which has so often disturbed science.” Having procured 
several hundred needles, of different lengths and thicknesses, 
and having ascertained that they were perfectly free from 
magnetism, we enveloped them in white paper, leaving one 
of their extreme ends uncovered. Taking advantage of a 
favourable day for trying experiments upon the chemical 
ray, (known by the few seconds required to blacken chloride 
of silver,) we placed the needles at right angles to the mag- 
netic meridian, and exposed them for,three hours, from 
eleven to one, to the differently refrangible rays of the sun, 
under coloured glasses. Those beneath the red, orange, 
and yellow, showed no trace of magnetism, while those be- 
neath the blue, green, and violet, exhibited, the two first 
feeble, but the last strong traces of magnetism. 
To determine how far the oxidating power of the violet ray 
is concerned in the phenomena, we exposed to the dif- 
ferent coloured lights needles whose extremities had been 
previously dipped in nitric acid, and found that they became 
magnetic (the exposed end having been made a north pole) 
in a much shorter time than the others, and that this effect 
was produced in a slight degree, under the red (when ex- 
posed a sufficient length of time) strongly under white glass, 
and so strong under violet glass, that the effect took place 
even when the needles were placed in such a position along 
the magnetic meridian, as would tend to produce, by the 
earth’s influence, a south pole in the exposed extremity. 
Conceiving that the inactive state produced in iron (as 
observed by Schcenbein) when plunged into nitric acid, 
S. G. 1.386, or by being made the positive pole of a battery, 
or by any other means (which Dr. Faraday supposed to be 
due to a slight oxide formed on the surface, and which may 
