of water to it, nor by pouring it several times. through the water, and letting it stand in water two days.

Another mixture, which had stood about fix hours on the quick filver, was diminished a little more upon the admiffion of water, but was never lefs than the original quantity of common air. In another case however, in which the mixture had stood but a very fhort time in quickfilver, the further diminution which took place upon the admiffion of water, was much more confiderable; fo that the diminution, upon the whole, was very nearly as great as if the procels had been intirely in water.

It is evident from these experiments, that the diminution is in part owing to the abforption by the water; but that when the mixture is kept a long time, in a fituation in which there is no water to abiorb any part of it, it acquires a conftitution, by which it is afterwards incapable of being abforbed by water, or rather, there is an addition to the quantity of air by nitrous air produced by the folution of the quickfilver.

It will be feen, in the fecond part of this work, that, in the decompofition of nitrous air by its mixture with common air, there is I nothing

nothing at hand when the process is made in quickfilver, with which the acid that entered into its compofition can readily unite.

In order to determine whether the fixed part of common air was depofited in the diminution of it by nitrous air, I inclofed a veffel full of lime-water, in the jar in which the process was made, but it occafioned no precipitation of the lime; and when the veffel was taken out, after it had been in that fituation a whole day, the lime was easily precipitated by breathing into it as ufual,

But though the precipitation of the lime was not fenfible in this method of making the experiment, it is fufficiently fo when the whole process is made in lime-water, as will be seen in the second part of this work; fo that we have here another evidence of the depofition of fixed air from common air. I have made no alteration, however, in the preceding paragraph, because it may not be unuseful, as a caution to future experimenters.

It is exceedingly remarkable that this effervefcence and diminution, occafioned by the mixture of nitrous air, is peculiar to common air, or air fit for refpiration; and, as far as I can judge, from a great number of obferva

tions, is at least very nearly, if not exactly, in proportion to its fitness for this purpose; fo that by this means the goodness of air may be diftinguifhed much more accurately than it can be done by putting mice, or any other animals, to breathe in it.

This was a moft agreeable difcovery to me, as I hope it may be an useful one to the public; especially as, from this time, I had no occafion for fo large a ftock of mice as I had been used to keep for the purpose of these experiments, ufing them only in those which required to be very decifive; and in thefe cafes I have feldom failed to know beforehand in what manner they would be affected.

It is alfo remarkable that, on whatever account air is unfit for refpiration, this same test is equally applicable. Thus there is not the leaft effervefcence between nitrous and fixed air, or inflammable air, or any species of diminished air. Also the degree of diminution being from nothing at all to more than one third of the whole of any quantity of air, we are, by this means, in poffeffion of a prodigiously large fcale, by which we may diftinguish very small degrees of difference in the goodness of air.

I have not attended much to this circumftance, having ufed this teft chiefly for greater

1

differences; but, if I did not deceive myself, I have perceived a real difference in the air of my study, after a few perfons have been with me in it, and the air on the outfide of the house. Also a phial of air having been fent me, from the neighbourhood of York, it appeared not to be so good as the air near Leeds; that is, it was not diminished so much by an equal mixture of nitrous air, every other circumftance being as nearly the fame as I could contrive. It may perhaps be poffible, but I have not yet attempted it, to distinguish fome of the different winds, or the air of different times of the year, &c. &c. by this test,

By means of this teft I was able to determine what I was before in doubt about, viz. the kind as well as the degree of injury done to air by candles burning in it. I could not tell with certainty, by means of mice, whether it was at all injured with refpect to refpiration; and yet if nitrous air may be depended upon for'furnishing an accurate teft, it must be rather more than one third worse than common air, and have been diminished by the fame general cause of the other diminutions of air. For when, after many trials, I put one measure of thoroughly putrid and highly noxious air, into the fame veffel with two measures of good wholesome air, and into another veffel an equal quan

tity, viz. three measures of air in which a candle had burned out; and then put equal quantities of nitrous air to each of them, the latter was diminished rather more than the former.

It agrees with this obfervation, that burned air is farther diminished both by putrefaction, and a mixture of iron filings and brimftone; and I therefore take it for granted by every other cause of the diminution of air. It is probable, therefore, that burned air is air fo far loaded with phlogifton, as to be able to extin guish a candle, which it may do long before it is fully faturated.

Inflammable air with a mixture of nitrous air burns with a green flame. This makes a very pleafing experiment when it is properly conducted. As, for fome time, I chiefly made use

of

copper for the generation of nitrous air, I first afcribed this circumftance to that property of this metal, by which it burns with a green flame; but I was presently fatisfied that it must arife from the spirit of nitre, for the effect is the very fame from which ever of the metals the nitrous air is extracted, all of which I tried for this purpose, even filver and gold.