than the prism; but the rapid motion of the sun made it difficult to keep the needle in the focus. The effect was produced with equal facility by throwing the spectrum on the floor of the room; but success could not always be depended upon even when the weather seemed most favourable.
I now made the following experiments with blue glass. Three needles free from magnetism, having one half covered with paper, were laid horizontally on the stone outside of a south window, under a dark blue glass coloured by cobalt, in a very hot sun; after remaining in this position between three and four hours, they were found to have become feebly magnetic, the uncovered part being a north pole. On examining these needles the following day, they had lost their magnetism, a circumstance which had not before occurred, though it was observed sometimes to take place afterwards, as the force of the sun diminished from the advance of the season. There was no iron near, and the magnetic needle when placed on various parts of the stone, showed no magnetic influence in it. Next day the experiment was repeated with this difference, that the needles were left exposed to the sun, under the blue glass, six hours; and then the needles had not only acquired very sensible magnetism, but still retain it, at the distance of nearly six months. Pieces of clock spring, which had been heated as formerly mentioned, also became magnetic under the blue glass.
I was desirous of ascertaining whether this kind of glass suffered the chemical rays to pass, and thereby occasion these changes in the steel, therefore I employed a liquid holding muriate of silver in suspension, as a test, in the following manner: a piece of writing paper dipped in the liquid was
