Having a trough containing a large quantity of quicksilver, he filled a glass jar (sufficiently capacious to contain his hand and wrist) with that metal. He inverted it into the trough in the usual way of proceeding in collecting gases. He fixed the glass jar in a sloping position, that he might introduce his hand the more readily beneath the quicksilver. In this way, whatever was given off from the skin of the hand, rising through the quicksilver to the top of the glass, and of course displacing a proportionate quantity of quicksilver, could be made the subject of analysis. He describes his first experiment as follows: "I held my hand ten minutes in the jar beneath the surface of the quicksilver, and frequently moved it in that situation, in order to detach any atmospheric air that might accidentally adhere to it, and afterwards introduced it into the inverted jar. The quicksilver _soon acquired_ a degree of warmth which rendered it not unpleasant. Minute air-bubbles ascended to the top of the quicksilver, more speedily in the beginning of the experiment, more tardily towards the conclusion. After an hour had elapsed, I withdrew my hand; the bubbles of air, which now appeared on the top of the quicksilver, were, I suppose, in bulk equal to one scruple of water. "In _sixteen hours_, I collected a half-once measure of air, which makes fifteen grains the average product of an hour. No kind of moisture appeared on the surface of the quicksilver. Some sucking-paper was put up, which was withdrawn unmoistened. My hand was always damp when taken out of the quicksilver. Whatever aqueous perspiration was produced, adhered to its surface, whilst the aeriform ascended to the top of the jar. To the air I had thus collected, I threw up some lime-water[20], when about two thirds of it were rapidly absorbed; to the remainder, I added a bubble of nitrous gas[21]; but could not discover any red fumes, nor any diminution of the quantity. I repeated this experiment six times, with similar though not uniform results. I believe it will be found that the air perspired consists of carbonic acid gas, or fixed air, a little more than two thirds; of nitrogenous gas, a little less than one-third. In one experiment, the nitrogen made only one-fourth part of the air collected; in another, I thought it exceeded one-third."