Experiments on the Eudiometrical Means, and on the Proportion of the Constituent Principles of the Atmosphere.

By Messrs. Humboldt and Gay-Lussac.

From the Journal de Physique.

IF philosophers and chemists are at present agreed respecting the nature of the constituent principles of the atmosphere, they are not yet so with regard to their absolute quantity.

Since Scheele and Lavoisier, who had found twenty-seven hundredths of oxygene in the air, numerous experiments, for which we are indebted to Messrs. Cavendish, Marti, Berthollet, Fourcroy, and Dery, have greatly modified this proportion, by fixing it at between twenty and twenty-three hundredths.

Nevertheless it must be admitted that these proportions are still far remote from that degree of accuracy which the present state of Science allows of; or, if these limits are well established, it must be concluded that the atmosphere experiences considerable oscillations in its composition.

Although with respect to most chemical phenomena, it is not necessary that we should have an accurate knowledge of the absolute quantity of its principles, this knowledge is not less interesting in itself, than important with regard to the history of our globe. If all the facts of geology tend to prove that the earth is no more what it has formerly been; that the highest mountains have been covered with water, and that the Arctic regions have nourished animals which at present are to be found only between the tropics, these very changes prove of what great utility it would be to future ages actually to ascertain the physical state of the globe at the present day; and even though the grand catastrophes which it has already experienced should never more be reiterated, it is possible that it may undergo gradual modifications which man would not be able of himself to appreciate, unless he found incontestible documents in the annals of science.

It would therefore be of the highest importance to fix in an authentic manner the grand phenomena of nature, which are to be supposed subject to variation; such as the intensity of the magnetic forces, the elevation of the barometer at the level of the sea, that of the sea itself, the medium temperature of every climate, and the proportion of the constituent principles of the atmosphere.

We have directed our attention to the latter question, and although we have not yet solved it in a manner entirely satisfactory to ourselves, we venture to give an account of the commencement of the investigation which we have undertaken upon this subject, and the researches to which it has led us.

But the eudiometrical means which must serve to determine the proportions of the constituent principles of the air are not all susceptible of equal precision, and some distinguished chemists give the preference to a means which is rejected by others.

It therefore was indispensably necessary that we should put to the trial the known eudiometrical methods, in order that we might be enabled properly to appreciate their value; for we are well convinced that accuracy in experiments depends less upon the faithful observation of the divisions of an instrument, than upon the accuracy of the method itself.

Although the nitrous gas appears indeed at first to be the most uncertain eudiometric means that can be chosen, we have convinced ourselves that by combining its action with that of the sulphate of iron, or of the oxygenated muriatic acid and of potash, it is capable of indicating with great precision the quantity of oxygene contained in the air.

All the eudiometrical means would give the same results, if we were equally acquainted with them all, and it is only because it is very difficult to make all the corrections they allow of, that we naturally give the preference to such as present less occasion for them, although they are not always the most simple in their use.

We shall therefore first give an account of the eudiometrical researches with which we have occupied ourselves, and then apply them to the analysis of the atmospheric air and of that obtained from water under different circumstances, or placed in contact with it.

We find it necessary again to repeat that we shall not treat the question which we have proposed to ourselves in so extensive a manner as it merits.

Compelled to break off our researches before we were able to bring them to their termination, our purpose is only to give an account of their principal results.

During nearly two months since the time that we commenced them in one of the laboratories of the Polytechnic-school at Paris, we pursued them, in spite of the coldness of the weather, which is very disagreeable in researches of this kind, with so much the greater assiduity, as M.

Humboldt found a peculiar interest in them.

In the year 6, he had presented to the Institute two memoirs on the analysis of the air, which contain a great number of experiments, which he now considers (it is he himself that avows it) not only as very inaccurate, but also as very justly controverted by Mr. Davy, and by a chemist with whose particular friendship we are both honoured, namely, M. Berthollet.

Zealous for the advancement of science, M. Humboldt, when he commenced his researches, wished to substitute in the place of those investigations of his early youth, others founded upon more solid foundations, and as he desired to have me for his associate in them; I ought to consider myself the more highly honoured by his proposition from our having been, since his return from the tropics, linked together in the closest bonds of friendship.

Observations on several of the Eudiometric Means.

It is not our intention in this memoir to record all the researches which we have undertaken upon different eudiometric means: most of them are still in too imperfect a state; but having occupied ourselves more particularly with the alkaline sulphurets, and especially with hydrogene gas, we shall here exhibit the result of our observations on these two eudiometrical means.

Although the alkaline sulphurets have in general a sufficiently-constant action for the analysis of the air, which has caused them to be justly preferred to the other eudiometrical means, they nevertheless present some sources of error, which it is indispensibly necessary we should be well acquainted with, in order to justify us in placing an entire confidence in their results.

It has long been believed that they have no action upon the azote; and although M. Marti announced, as early as in the year 1790, that they absorb this gas, no farther attention has since been paid to this property.

It is true M. Marti at the same time announced, that by saturating them with azote we might employ them to advantage for the analysis of the air, and constantly obtain for the oxygene a proportion comprehended between 0.21 and 0.23.

On the other hand, this chemist not having indicated the details of his experiment with sufficient precision, M. Berthollet, who repeated it after him under different circumstances, announced in his Chemical Statics, that he had not found the alkaline sulphurets to possess the property of absorbing azote, whereby he gave new sanction to their employment for the analysis of the air.

When we began to employ this means we placed great confidence in it, and had nothing to object to it but the great length of time which it requires, and which had long rendered it a desirable object that, notwithstanding its accuracy, some other might be substituted in its place that were not attended with the same inconveniences; but we soon discovered that it did not always act in an uniform manner, and herein we were favoured by accident.

Having placed 100 parts of atmospheric air in contact with a solution of sulphuret of potash, made with hot water, in three vessels of unequal capacities, we observed at the end of eight days that the air had lost twenty-three parts of its volume in one of the vessels, and 23,6 26,0 in the two others.

This great inequality gave us at first much surprise; but having remarked that the absorption had been most considerable in the largest flask, we suspected that azote had been absorbed, and, in order the better to confirm our suspicion, we repeated the same experiment, employing two vessels more unequal in capacity, but otherwise placed in the same circumstances.

At the end of ten days we found that in the small flask the absorption had been only 22,5 parts, whilst in the large one it was 30,6.

But the most decisive experiment which we made on this subject, was the placing of a solution of sulphuret of potash, which had been heated to ebullition, in contact with azote in unequal vessels, when it was discovered that the absorption was proportionate to their capacities.

It would therefore be possible to cause a determinate quantity of atmospheric air to be absorbed by a solution of alkaline sulphuret, and to make it be considered as pure oxygene, if it were supposed that the whole diminution of volume were owing to the oxygene gas.

But if, instead of employing a solution of sulphuret made with hot water, we employ one made with cold, as M. Berthollet has always done, the solution of the azote takes place no longer, at least not in a perceptible manner, and the results of the analysis of the air made by this means become then much more susceptible of comparison.

This variable action of the alkaline sulphurets, dissolved at different temperatures, requires to be farther elucidated, which we are about to do by citing phenomena of an analogous but more comprehensible nature.

As water always holds in solution a certain quantity of air in which the proportion of oxygene is more considerable than it is in the atmospheric air, it happens that when we heat it, or dissolve a salt in it, a part of its air is disengaged from it whilst another is retained, which may be separated from it by a more intense heat.

If, therefore, we place this water, which has been deprived of its air by the last-mentioned means in contact with atmospheric air, it will absorb, in returning to its original temperature, a quantity equal to that which it has lost; and if we were not aware of this absorption, but judged upon appearances, we should suppose that the water alone, or charged with salt, had made the analysis of the air.

Thus M. Hetter has very recently announced that a solution of muriate of soda absorbed all the oxygene of the air, though on repeating his experiment with a highly-charged solution of the same salt, but made with cold water, we did not find the slightest difference between the ordinary atmospheric air and that which had been in contact with the solation of muriate of soda for a month and a half.

The very same thing happens with a sulphuret as with a salt.

At the moment of its solution in water, a portion of air is expelled, and an equilibrium of saturation is established between the water, the sulphuret, and the air which it holds in solution, so that if the circumstances are not changed, there is now no reason why it should absorb a fresh quantity of air; but if we heat the solution, there is disengaged from it a part of the gas which it contained, and, in returning to its original temperature, it is necessary that it absorb what it had lost, in order that its equilibrium may be re-established . We think, therefore, that the difference between the results of Messrs. Marti and Berthollet may be explained by the difference of the circumstances under which they operated; but it appears to us, that M. Marti believed that it was the nature of the sulphuret to absorb azote, whereas it does not absorb it all, but rather prevents the water with which it has been boiled from absorbing as much as it would do without it.

The absorption of which we mean to speak in this place is independent of that of oxygene by the sulphuret, which is thereby converted into sulphate. But as the sulphuret absorbs the oxygene which the water holds in solution, it will very probably happen that the water will be able to absorb a larger quantity of azote; so that if we employed a solution made with cold water, and very recently prepared, there would be a still greater diminution of volume than that which is owing to the absorption of the oxygene.

We say very probably; for we have not yet made the experiment.

Thus if we take care to dissolve the sulphurets in cold water, and leave them for some time in contact with azote or with air, we may employ them with advantage for the analysis of the atmosphere.

We shall however observe, that as they are attended with the inconvenience of requiring a great length of time before their action is completed, we are obliged to have recourse to the corrections of the thermometer and barometer, which often are very uncertain.

The best way of remedying this inconvenience is undoubtedly to follow the method of Messrs. Berthollet and Marti, which consists in placing, for comparison, upon water a determinate quantity of air, in order to judge, from its variations of volume, of that of the air which we analyze; but this method has not appeared to us to be attended in practice with all the advantage which it seems to promise.

We must still remark with regard to all the eudiometrical means, where the absorbing substance is solid or liquid, that if we commit an error, either in observing the divisions of the instrument, or in the appreciation of the uncertainties of the method, this error necessarily falls altogether upon the quantity of oxygene; and since with all possible exactness we cannot answer for its amounting to much less than a hundredth part, it would follow that the proportion of oxygene contained in the air cannot be determined within this quantity.

In fact, we find that chemists who have employed such means have found very considerable variations in the quantity of oxygene of the air; and M. Marti himself, who appears to have made a great number of experiments with the alcaline sulphurets, and who was acquainted with the precautions which they require, fixes it between 0,21 and 0,23.

We shall see hereafter that the eudiometrical means, in which the substance which combines with the oxygene is gaseous, admit of a much greater precision.

As we had proposed to ourselves from the commencement of our investigation, to ascertain whether the eudiometer of Volta was fit to be employed for the analysis of the air, we principally directed our attention to it.

This instrument has been accused of being fallacious, of indicating too small quantities of oxygene in the air; but it appeared to us that supposing it required corrections, we might by appreciating them, as well as the law of their variations, render it very exact and convenient.

With this view we proposed to ourselves the following questions:

1. When a mixture of hydrogene gas and oxygene gas is inflamed in Volta’s eudiometer, can the absorption of one of the gases be complete?

2. Is the product of their combination of a constant nature?

3. What is the exact proportion of the two gases for forming water?

4. What are the limits of error that Volta’s eudiometer admits?

We must examine these four questions in succession, but first of all we think it incumbent upon us to give an account of the manner in which we prepared the gases which we employed in our experiments.

to be continued in our next

Experiments on the Eudiometrical Means, and on the Proportion of the Constituent Principles of the Atmosphere.

By Messrs. Humboldt and Gay-Lussac.

(Continued from Page 239.)

THE oxygene gas we obtained from the super-oxygenated muriate of mercury.

To obtain it we employed a glass retort, to which had been soldered, by the flame of a lamp, the curved tube through which the gas was to pass; and, in order that we might have it as exempt from azote as possible, we filled the retort about one-fourth with water.

This water being entirely reduced to vapour before the decomposition of the salt, soon expelled all the air out of the retort; but, in order to prevent the absorption which would have taken place before the disengagement of the oxygene gas, we plunged the extremity of the tube in a saucer filled with mercury, which we removed as soon as the gas began to be disengaged.

In order that the oxygene may not, passing through the water, expel azote from it, we convey it directly into the upper part of the recipient which is to receive it, by means of a tube bent at a right angle, which at the one end ascends to the upper part of the recipient, and at the other is adapted to the first tube by means of a cork stopper, common to both.

This process, which is very simple in its application, is particularly advantageous for the gases which are soluble in water, such as the carbonic acid gas, the azote oxyd gas, &c.

We obtained our hydrogene gas by decomposing water by means of zinc and muriatic or sulphuric acid diluted with about six parts of water: we took care to fill exactly with acid the vessel from which the gas was to be disengaged, and not to make it pass through the water; but notwithstanding all these precautions, our oxygene left with the sulphuret four thousandth parts of azote, and the hydrogene, analysed by other means, shewed six thousandths.

After these elucidations let us proceed to the questions which we have proposed to ourselves to resolve, beginning with this:

When we inflame a mixture of oxygene and hydrogene gas in the eudiometer of Volta, can the absorption of one of the gases be complete?

In order to ascertain whether all the oxygene or all the hydrogene could be entirely destroyed, we thought that if two gases were perfectly pure, or if we knew their degree of purity, and that their absorption must be complete, we should be able to find the same proportion for the principles of water, whether the hydrogene or the oxygene were predominant.

In fact, by detonating mixtures of 300 parts hydrogene and 100 oxygene, and of 200 of the first and 200 of the second, in which the hydrogene and oxygene alternately predominate, and making the corrections due to the impurity of the gases, we obtained very nearly the same proportion.

Although the absorption of the two gases might be complete, it were however possible that the proportions obtained, in making them alternately predominate, might not be identical, and this would be the case if, according to the predominance of either of the gases, there were formed an oxygenated or hydrogenated water; but since the proportions have become identical, it must necessarily be concluded, that the hydrogene and the oxygene were entirely absorbed.

But though the absorption of the two gases may be complete under certain circumstances, it must not be supposed that it is so with any quantities; there are not only such proportions of hydrogene and oxygene, or of their mixture with azote, or even with any other gas, that it is impossible to inflame them by means of the electric spark; but there are also others with which the inflammation having been commenced, stops before the combustion is completed.

We proceed to cite experiments to this effect, which appear to us to be conclusive.

We mixed 100 parts of hydrogene with 200, 300—900 of oxygene, and inflamed them by the electric spark: with these different proportions the absorption constantly amounted to 146 parts; but with 1000 of oxygene it was at once reduced to 55; with 1,200 and 1,400 it was reduced to 24 and 14, and with 1,600 it was reduced to 0; that is to say, no inflammation took place.

These different results are exhibited in the following table.

Hydrogene.

Oxygene.

Absorption.

100 ‒ ‒ ‒ 200 ‒ ‒ ‒ 146 100 ‒ ‒ ‒ 300 ‒ ‒ ‒ 146 100 ‒ ‒ ‒ 600 ‒ ‒ ‒ 146 100 ‒ ‒ ‒ 900 ‒ ‒ ‒ 146 100 ‒ ‒ ‒ 950 ‒ ‒ ‒ 68 100 ‒ ‒ ‒ 1000 ‒ ‒ ‒ 55 100 ‒ ‒ ‒ 1200 ‒ ‒ ‒ 24 100 ‒ ‒ ‒ 1400 ‒ ‒ ‒ 14 100 ‒ ‒ ‒ 1600 ‒ ‒ ‒ 0

The absorptions 68, 55, 24, and 14, are possibly not exact within two or three hundredths, for our instruments being too small for the corresponding proportions, we were obliged to measure them several times; but this is of no moment with respect to the phenomenon in general.

What is striking in these different experiments is to see, 1, a constant absorption, with very different proportions, change suddenly into a decreasing absorption; 2, the combustion of hydrogene gas which had commenced, stop before it was completed; 3, that there are such proportions of hydrogene and oxygene as cannot be inflamed.

These different phenomena will be in some measure explained in the sequel; but in the mean time we shall remark, that there are even very extended proportions with which the combustion of the hydrogene gas may be complete.

The above-mentioned phenomena are not peculiar to the hydrogene and oxygene gases mixed together, under the circumstances of which we have been speaking: they also take place when we inflame 100 parts of oxygene with 200, 300—1000, &c. of hydrogene; only it then happens, that the term when the absorption ceases to be constant is more remote; and to comprehend the reason of this it is sufficient to observe, that in this case there disappear about 300 parts by the inflammation, whereas there disappeared only half the quantity in the preceding experiments.

Azote gas and carbonic acid gas present also analogous results.

If, for example, we inflame a mixture of 900 parts of azote, 100 of hydrogene, and 100 of oxygene, the absorption, which ought to be 146 parts, if the combustion were complete, was in one experiment, which is that which we take as example, only 50 parts, though in others we have seen it a little above or below this quantity.

With inferior proportions of azote, we have constantly had the same absorption of 146. Although the azote appears here to comport itself like the oxygene, since with 100 of hydrogene and 1000 of oxygene, we had nearly the same result as with 100 of hydrogene, 100 of oxygene, and 900 of azote, we shall not draw from hence any inference, because we have not sufficiently multiplied and varied our experiments.

Nevertheless, those which we have made tend to prove, that when determinate proportions of oxygene gas and hydrogene gas are mixed with different gases, the absorption may be constant as far as to a certain point, beyond which it diminishes very rapidly.

The absorption of the oxygene and of the hydrogene being complete in determinate proportions, and not so in others, it will always be possible, when a gaseous mixture is given, which alone would not be able to inflame, to reduce it to another with which the absorption of one of the gases would be complete, by adding to it oxygene or hydrogene, or even both together.

The combustion of the 100 parts of hydrogene in the preceding experiment not having been complete, we analyzed the residue.

100 parts, placed in contact with phosphorus, diminished by 7 in the space of four hours, an evident proof that the residue contained oxygene.

In order to ascertain whether it had retained hydrogene, we inflamed, in Volta’s eudiometer, a mixture of 200 parts of the preceding residue, 200 of oxygene gas, and 200 of hydrogene gas; in all 600 parts.

After the inflammation 312 parts had disappeared; and as, according to experiments of which we shall give an account hereafter, 100 of pure oxygene require for their saturation 200 of hydrogene gas, the absorption which, with the hydrogene gas which we have employed, ought only to have been 292 parts, amounted to 312, the residuum must necessarily have furnished a sufficient quantity of it to carry the absorption from 292 to 312; that is to say, it must have contained 13,3 parts.

Now calculation shews that it ought to have contained 12; it is therefore clearly proved, that though inflammation took place, the combustion was not complete; and that all the hydrogene did not enter into combination, since we have found that which had not been absorbed in the residuum.

We must observe, that in all cases in which the absorption was not complete the inflammation was languid.

From comparing, in the inflammation of hydrogene and oxygene gas, the effects of electricity with those of a high temperature, we have been led to believe that the inflammation produced by the electric shock might very likely be owing to the heat produced by the instantaneous compression which the electric spark occasions in its passage.

In fact we know, from our own experience, that the inflammation of a mixture of hydrogene and oxygene gas depends solely upon the temperature when this inflammation is produced by heat.

For if we cause this mixture to pass very slowly through a tube heated very gradually, from its extremity to its central part, without opposing the free dilation of the gases, the inflammation will take place as soon as the temperature shall be raised to a sufficient degree.

This being admitted as fact, that the inflammation of the oxygene and the hydrogene gas takes place only at a certain temperature, let us see what passes in their inflammation by the electric spark.

When this passes through a mixture of oxygene and hydrogene, it displaces it by its rapid passage, which does not permit the gaseous particles to communicate to each other the motion as quickly as they have received it; hence results a very strong instantaneous compression, which produces an elevation of temperature superior to that which is requisite for the combination of the gases, and the inflammation being thus commenced, must be propagated very rapidly.

According to this mode of accounting for the effects of electricity, we thought that when a weak spark produces only an imperfect combustion in a mixture of hydrogene and oxygene gas, a stronger one would produce a more complete combustion; but whether it was that we did not employ a sufficient brisk electricity, or that we did not multiply our experiments sufficiently, we obtained no sensible differences in employing the spark of an electrophorus, three decimeters in diameter, or the shock of a highly-charged Leyden flask; but the construction of our eudiometer did not permit us to draw very brisk sparks, on which account we shall reserve our opinion respecting the influence of the force of the electricity in the inflammation of the hydrogene and oxygene, till we shall have made farther researches upon the subject.

In the above-described experiment on the inflammation of a mixture of 900 parts of azote, 100 of oxygene, and 100 of hydrogene, the absorption was not so considerable as it ought to have been, and we have proved that the residuum ought to contain what had escaped combustion; that is to say, it ought to be composed of six parts of hydrogene, eight of oxygene, and eighty-six of azote in the hundred.

Therefore, since the combustion was interrupted when these proportions took place, it may be concluded that another electric spark would no more be able to inflame this mixture.

Consequently, in the atmosphere, which contains much less than six hundredths of hydrogene, the electric spark will not be able to inflame it, or if it does it at the place of its passage, by reason of its great force, the inflammation will not be able to propagate itself, but will be in a manner confined to the places which it traverses.

Hence, finally, the inflammation of hydrogene gas, by lightning, and à fortiori by weaker charges of electricity, will not serve to explain the igneous meteoric phenomena; or if these phenomena are actually results of the inflammation of hydrogene gas, we must conclude that there are more than six hundredths of it in the air at the moment when they are produced.; which is contrary to all probability, especially when we recollect that air, collected at a very great elevation, presented no appreciable quantity of hydrogene above that contained in atmospheric air collected at the surface of the earth.

But if every time that we cause an electric spark to pass into a mixture of hydrogene and oxygene, or of azote, hydrogene, and oxygene, which is not capable of inflaming, there is actually produced a local and instantaneous heat by the compression which the spark occasions in its passage, it is possible that by directing a succession of sparks into one of the mixtures of which we have been speaking, a slight local inflammation might be produced each time upon the passage of the spark, and that thus it might be practicable to destroy a determinable quantity of hydrogene inveloped in a large proportion of azote and oxygene, or of oxygene only.

What seems to confirm this supposition is, that it is well known that ether and ammoniac, which are decomposed by heat when they are made to pass in vapours through a red-hot tube, are likewise decomposed by repeated electric shocks.

It would also be interesting to know, whether it be possible to inflame by the electric spark a proper mixture of oxygene and hydrogene, after having dilated it by means of the pneumatic machine.

If its inflammation by the electric spark really depends upon the heat which this produces by compression, it would be natural to suppose, that when these gases are dilated, the compression by the spark being less considerable, the heat which is produced by it must also be much slighter, and that there may be a degree of dilatation of the gases at which the inflammation cannot take place.

We have not yet had time to try these different experiments; but we do not abandon our design to attempt them, which we hope we shall be able to do shortly.

To recapitulate: there exist certain proportions of hydrogene and of oxygene, or of these two gases with azote, in which combustion can be complete.

There exist also others, in which it stops spontaneously before being completed; and, finally there are proportions in which it cannot take place at all.

The hydrogene gas which escapes the combustion is found again entire in the residue.

When we cannot produce by the electric spark a complete inflammation of the hydrogene gas, or even when we cannot commence it, nothing more is necessary than to augment the proportions of the oxygene or of the hydrogene.

The igneous meteoric phenomena cannot be the result of the inflammation of the hydrogene gas, because in the regions where the principal of them are supposed to take place, such as the sudden and abundant torrents of rain which sometimes succeed a clap of thunder, it would be necessary that there should then be more than six hundredths of hydrogene in the atmosphere, without which the inflammation could not take place; besides which, only the quantity exceeding this proportion could pass into inflammation.

We may account for the cases in which the combustion was not complete according to the laws of the affinities, by saying, that when one of the gases becomes very predominant it may defend the other by its affinity, and guard it in part from combustion.

Although this affinity may be very weak, we conceive, with M. Berthollet, how the quantity of gas may compensate for it; and if there be in the different gases peculiar properties of stopping the combustion sooner or later, this may be explained by their different nature.

But when we consider the case in which the hydrogene is mixed with oxygene only, and suppose the phenomena of its combustion with different proportions of oxygene to depend upon affinity, how explain the sudden transition from a constant absorption to a decreasing absorption, when it is agreed that if the hydrogene can be prevented from the combination by the oxygene, the effect of the latter must follow a regular law?

How conceive that these two gases, after having been placed in circumstances favourable to their combination, can by their affinity maintain themselves in their elastic state, when they might form a much more dense combination, namely, water?

How conceive, finally, that an affinity which produces a very great condensation and saturation, can be inferior to an affinity which produces no change in the dimensions of the two gases, no saturation?

Hydrogene and oxygene, in whatever state they may be, have the same degree of affinity, as this affinity is measured by their capacity of saturation; only the state in which they are may be more or less favourable to their combination.

Now, to say that hydrogene and oxygene have a greater affinity in the state of gas than in the liquid state, is to say that their molecules attract each other more when they are very remote than when they are very near to one another.

These objections against an explanation founded solely upon the affinities, having appeared to us to be of some weight, we have endeavoured to present one which, in our opinion, did not involve the same difficulties.

All combustible bodies require in general a certain elevation of temperature, in order to combine with oxygene.

Carbon, for example, is not converted into carbonic acid until it is red-hot; and this same substance, which at a high temperature can continue to burn when it is exposed to a current of aqueous vapour, is extinguished as soon as it is immersed in water.

This principle, that bodies in general require a certain elevation of temperature in order to burn, being once admitted, let us suppose that we have a body which burns in a given volume of atmospheric air, and that the temperature necessary for the combustion is maintained solely by the heat due to the absorption of the oxygene: let us also suppose, that at the commencement of the combustion the heat due to the fixation of the oxygene contained in a cubic centimeter of air is =1, and that the heat lost during the fixation, whether in radiating heat, or by the absorption which is made of it by the azotic gas or other bodies, is = ½, not taking here into consideration the law according to which it decreases.

According to these premises, it is very evident that in the first moments of the combustion the temperature of the body must rise; but, in proportion as the quantity of oxygene shall diminish, and that of the azote proportionally increase, the heat communicated will also diminish.

A period will therefore arrive, at which the heat lost will be equal to the heat communicated, and below which, the temperature being too low, the combustion must cease.

What evidently proves that the combustion stops only because the temperature is too low, is that if we artificially maintain the temperature sufficiently high, the body will continue to burn.

Now this explanation will still hold good, when, instead of azote, sulphurous gas, hydrogene gas, carbonic acid, or any other gas, is mixed with the oxygene; only the combustion may cease sooner or later than with azote gas.

For it is very evident, that if the sulphurous gas, or carbonic acid gas, had a capacity for caloric, much greater than that of azote, supposing them to be mixed with oxygene in the same proportions as the latter, the loss of heat would be much greater, and consequently the cessation of the combustion must take place sooner.

But if the gases had equal capacities for caloric, they must all of them stop the combustion at the same period, as we have seen has nearly been done by oxygene and azote with hydrogene, and this would perhaps afford a solution of the important question, whether the gases have equal or different capacities.

Thus a combustible body, sulphur for example, would cease to burn in a determinate volume of air, not because the affinity which the azote or the gases produced have for oxygene were more powerful than that of the combustible body; but because the heat absorbed by these gases, which tend to place themselves in an equilibrium of temperature with the burning body, would be greater than the heat proceeding from the fixation of the oxygene; whence it would result, that the temperature would soon be reduced below that necessary for the combustion.

In fact, we know that sulphur can continue to burn in air in which it had been extinguished if, we raise the temperature to a sufficient degree.

What takes place in the instantaneous combustion of the hydrogene in Volta’s eudiometer, is perfectly analogous to what passes in its successive combustion in a given volume of air, or in that of any other body.

If we place a lamp, the flame of which is supplied by hydrogene gas, under a bell-glass filled with oxygene gas, the flame will become small, bright and slightly coloured.

If we replace the oxygene by atmospheric air, the flame will be more voluminous, less bright, and more coloured.

In proportion, especially as the relative quantity of oxygene diminishes, the flame will increase in size, because the hydrogene will be obliged to go farther to come at the oxygene, and the flame will soon be extinguished, although the air still contain some hundredth parts of oxygene.

The phenomena which take place in Volta’s eudiometer are of the same nature.

When the proportions of oxygene and hydrogene do not deviate much from those which constitute water, the flame is still very bright, notwithstanding its dilatation; but if we mix, for example, 1000 oxygene with 100 of hydrogene, the flame is then weak, of a blueish-green colour, and the combustion of the hydrogene is far from being complete, for we still find nearly two-thirds of it in the residue.

What farther proves that the combustions’ not having been complete was owing to the temperature not having been sufficiently elevated, is that if the residue is made to pass, as was done by us, through a red-hot tube of porcellain, the whole of the hydrogene will be absorbed.

We must observe, that in the combination of the hydrogene and oxygene gases, a very singular phenomenon takes place, which has long since engaged the attention of M. Monge.

“How does it come to pass,” says this distinguished philosopher, “that by elevating the temperature of the two gases, or, in other words, by increasing the proportion of the solvent, we diminish the adhesion which it had with its bases?”

Far from conceiving that the present state of our knowledge is competent to give a satisfactory solution of this question, we wish to recommend it anew to the attention of philosophers.

In fact, according to the idea which we are able to form to ourselves of the force which produces the combinations, and of those which are opposed to it, the elastic state indicates that the force of cohesion is destroyed, and that two bodies in this state are in the condition most favourable to combination; so that now that the attractive force of their molecules has been changed into a repulsive force, every cause which shall favour the latter will be in opposition to the former.

It happens, however, that by raising the temperature of the two gases, that is to say, by augmenting their repulsive force, their attractive force is increased.

It cannot be believed, that the heat does nothing more than separate their molecules to greater distances from each other; for, in this case, why should a mixture of hydrogene and oxygene gas not inflame under the receiver of an air-pump, where it may be indefinitely dilated?

Neither is it to be supposed that the heat acting instantaneously can produce a compression which favours the combination of the two gases by bringing their molecules into closer contact; for it is easy to convince ourselves that a mixture of oxygene gas and hydrogene gas, heated very gradually without opposing its dilatation, will nevertheless inflame when the temperature shall be sufficiently elevated.

Having now proved, that under determined circumstances the combustion of hydrogene and oxygene may be complete, we shall proceed to examine whether its products are constant.

According to all the experiments which have been made on the composition of water, the result has been considered to be uniform.

There has, however, sometimes been obtained a small quantity of nitric acid; but it has been ascertained that this acid is not a constant product of the combustion of the hydrogene, but, on the contrary, merely accidental.

Cavendish was the first who discovered this formation of nitric acid, and Messrs. Foureroy, Seguin, and Vauquelin, have taught us how we may avoid it, and obtain water without any acidity.

It has not, indeed, been demonstrated, that oxygenated or hydrogenated waters have not been formed, because, in all the exact experiments that have been made, the combustion of the hydrogene gas has always been effected in the same manner, and at most it would be proved that those which have been obtained are constant under the same circumstances.

If we compare the combustion of hydrogene gas with that of nitrous gas, the products of which are so variable, we shall be still more justified in concluding, that since oxygene has always predominated in the experiments which have been made, an oxygenated water may have been formed; whereas, if hydrogene had been predominant, a hydrogenated water would have been the result.

Let it then be admitted, that an oxygenated water may be formed; if, for example, we obtain it in all circumstances, and it be constant, this will be of no consequence with regard to the proportion of its principles, which must serve for the analysis of the air; but if it be so only because the oxygene predominated, it is manifest that we shall no longer obtain the same proportions when we alternately cause both gases to predominate.

to be continued in our next

Experiments on the Eudiometrical Means, and on the Proportion of the Constituent Principles of the Atmosphere.

By Messrs. Humboldt and Gay-Lussac.

(Continued from Page 318.)

NOW as we have made a great number of experiments which prove, that when we employ them alternately in excess we constantly, obtain the same proportions, it follows, that the combustion of hydrogene gas is of an uniform nature.

The Galvanic phenomena of the decomposition of water seem however to prove that water is capable of oxygenating or hydrogenating itself; and it is upon this supposition that Messrs. Laplace and Berthollet have accounted for the singular experiment of the decomposition of water by two metallic wires immersed at one end in this liquid, and communicating by the other with the two poles of a Galvanic pile.

But without wishing to controvert this explanation, which appears to us the most satisfactory that has hitherto been offered, we shall observe that the complete absorption of the hydrogene at one of the wires, and of the oxygene at the other, proves that the water does not become oxygenated or hydrogenated, because, in order to become so, it would be necessary that it should absorb one of the gases in a proportion greater than that required for the composition of water.

If therefore it absorbs oxygene and hydrogene in exact proportions for forming water, it must be concluded that the properties of one of the gases are neutralized by those of the other.

Accordingly, in the circumstances of which we are treating, the water might become instantaneously oxygenated at one of the wires, and hydrogenated at the other; but the two gases, being deprived of elasticity, and in exact proportions, must soon enter again into combination.

If it is well demonstrated that, under given circumstances, hydrogene or oxygene may be completely absorbed, and, if it is equally so, that the product of their combination is constant, nothing more remains to be done, in order to solve the third question, which we have proposed to ourselves, than to determine the proportions of oxygene and hydrogene which constitute water.

To 100 parts of oxygene gas we added 300 parts of hydrogene gas; and after having inflamed them by the electric spark, we obtained in twelve experiments the following residues:

100,8 101,0 102,0 101,4 101,7 102,0 100,5 102,0 101,0 101,0 101,5 101,5 The mean term of which is ‒ ‒ 101,3

Thus 100 parts of oxygene, supposed to be very pure, required 98,7 of hydrogene: but on placing our oxygene gas in contact with sulphuret, we found that it had all been absorbed within about 0,004; hence it follows, that 99,6 of oxygene absorbed 199,1 of hydrogene, or that 100 absorbed 199,89, or, to use round numbers, that 100 of oxygene requires 200 of hydrogene for their saturation.

In the preceding experiments it was the oxygene that disappeared.

Let us now reverse the experiment, by inflaming a mixture of 200 parts of each gas; the residue of the different inflammations will be the following:

101,5 102,0 101,5 101,3 102,0 102,3 102,2 101,0 102,0 102,0 101,0 102,0 Mean term ‒ ‒ 101,7 Mean absorption 298,8

200 parts of hydrogene, supposed to be pure, required therefore 98,3 of oxygene; whereas, according to the proportion which we have just established, they ought to have had 100. But if we admit this same proportion to be exact in the 298,3 of absorption, there would be only 198,8 of hydrogene, which would indicate 0,006 of azote in this gas.

Supposing even the hydrogene to be perfectly pure, the two proportions obtained by causing the oxygene or the hydrogene to predominate, agree sufficiently with each other to confirm all that has been said in the course of this memoir: to render them identical, nothing more is necessary than to suppose 0,006 of azote in the hydrogene, in which, in fact, we can demonstrate its presence.

We have just seen, from the preceding experiments, that 200 parts of hydrogene, without making any correction, absorbed 98,3 of oxygene.

Let us then take the residues 101,0 and 101,5, proceeding from the combustion of 100 of oxygene and of 300 of hydrogene, and detonate them with 200 of oxygene gas.

In these two residues there must be 0,008 of azote belonging to the 200 parts of oxygene gas; and if the remainder 201,7 was pure hydrogene it ought to absorb 99,1 of oxygene, and consequently there ought to have disappeared by the inflammation 300,8 parts, but there disappeared only 295,0: it follows, therefore, that the residuum 201,7 was not pure hydrogene; and that, according to the proportion of 100 of oxygene to 200 of hydrogene, it contained 5,0 of azote, proceeding from 600 of hydrogene; that is to say, that this last-mentioned gas contained 0,008 of azote.

We therefore think it proved, that 100 parts, in volume, of oxygene gas require very nearly 200 parts of hydrogene gas for their saturation.

According to the experiment of Messrs. Fourcroy, Vauquelin, and Seguin, 100 parts of the former would require 205 of the latter; but we shall observe, that, whichever of the two proportions we adopt, the error cannot amount to more than 0,0035 upon the absolute quantity of the oxygene of the air, and when the question respects relative quantities, the error will be still much smaller.

We have ascertained, that the proportion does not vary in consequence of changes of the temperature.

It is evident that it must be so, since, as the heat equally dilates both gases, and causes them to dissolve equal quantities of water, the real weights of hydrogene and oxygene contained in equal volumes remain always in the same proportion.

It would therefore be more accurate to say, supposing our proportion by the volumes to be well established, that 100 parts of oxygene require 200 of hydrogene, than to indicate the proportions of water by the weight.

If the oxygene and hydrogene employed for the composition of the water had been perfectly dry, or if a correction had been made according to the quantity of moisture which they might contain, it would be a matter of indifference whether the proportions of its principles were enunciated according to their volumes or their weights; but since hydrogene combines with oxygene in double the volume of the latter, and as they both dissolve the same proportion of water, it is evident that they do not carry into the combination quantities of water which are in the same reciprocal proportions as the quantities, by weight, of oxygene and of hydrogene, and that consequently the proportion of the principles of the water must thereby be altered.

Thus the proportion, according to the volumes, has the property of remaining constant, notwithstanding the changes with regard to temperature and moisture, whilst that, according to the weights, varies under the same circumstances.

And let it not be imagined that this consideration is of little importance: for it would be very easy to prove that it has very considerable influence upon the proportion of the principles of water.

According to the experiment of Messrs. Fourcroy, Vauquelin, and Seguin, the most accurate that has hitherto been made upon this subject, water contains by weight 85,662 of oxygene and 14,338 of hydrogene.

But the experiment having been made at a temperature of about 14°, and the correction due to the water held in solution by the gases not having been made, it follows, that if we adopt their specific weight of the oxygene and hydrogene gases, as well as the proportion of their volumes in their combination, and moreover admit with Saussure, that a cubic foot of air, at the temperature of 14°, contains very nearly 10 grains of water in solution, the proportion by weight between the oxygene and hydrogene, instead of being 85,662 to 14,338, would be 87,41 to 12,59; a very remarkable difference, and which especially must be of great influence in analyses where the real weight of the hydrogene is to be determined.

The same consideration applies also to the specific weight of the gases, and principally to that of the hydrogene, of which about a sixth part is due to the water which it holds in solution when the temperature, as here supposed, is 14° Reaumur.

We do not therefore doubt, that if we had hydrogene gas perfectly dry, and deprived of the azote gas, which appears very often to accompany it, we should find its specific lightness to be at least 15 times greater than that of the atmospheric air.

It still remains for us to answer the last question which we have proposed to ourselves, and to shew what are the limits of the error in Volta’s eudiometer; and next, what are the smallest quantities of oxygene or of hydrogene that can be estimated by his method.

The effects obtained with this instrument, being instantaneous, are independent of the thermometer and barometer.

In this point of view, it has the very decided advantage over phosphorus and the alkaline sulphurets, of giving results very capable of comparison; but this is not the only one; it has also that of the eudiometrical means which give multiples of the quantity to be estimated.

As in this instrument each hundredth of oxygene is represented by a three times greater absorption, the error that can be committed amounts only to a third upon this gas; and now, especially as we have instruments which divide the measure into three hundred parts, it is evident, that if we err even by one division, the accuracy, as to the quantity of oxygene, may be carried to near a thousandth part of the quantity of air analyzed.

If therefore the results of the combustion of hydrogene gas admit so well of comparison, and the errors to which they are liable are restricted within such narrow bounds, it is evident that we may not only find the slight differences which exist between two portions of atmospheric air, but also determine less than three thousandth parts of oxygene which shall be lost in azote or hydrogene gas: but in this case, in order that the inflammation take place, it would be necessary to add a given quantity of oxygene, the absorption of which with hydrogene gas had been determined by previous experiments, and then the excess of the first absorption above the second would be attributed for a third to the oxygene gas contained in the air which is analyzed.

On the other hand, in order to determine whether one portion of hydrogene be more pure than another portion, or whether there exist very small quantities of it in a gas or in the atmospheric air, it would be necessary, in the first case, to mix 100 parts of hydrogene gas with 100 of oxygene: the quantities of real hydrogene would be in the direct ratio of the absorptions.

But if the proportion of hydrogene were very small, for example one two hundredth, it would be necessary, in order to effect its combustion, to add 100 parts of this gas to 200 of the air to be analyzed, and to detonate the mixture with a sufficient proportion of oxygene.

By this means, and with the practice which we have now acquired, we have been able to find again three thousandth parts of hydrogene gas which we had mixed with atmospheric air.

It might be objected against Volta’s eudiometer, that the hydrogene not being always the same, we might incur errors difficult to be appreciated.

We shall first observe that it is indifferent whether or not it contain azote; but if it contained oxygene, its quantity confounding itself with that which we wish to appreciate, would alter the results.

In order to avoid this inconvenience, we may first of all detonate separately 500 parts of hydrogene with 100 of oxygene; by this means its oxygene will be destroyed, and we may then employ it for the analysis of the air.

With this precaution we may employ a gas prepared as inaccurately as possible.

It is sufficient if it have been extracted from the water by means of zinc and sulphuric acid or muriatic acid; for it is well known that if we employ another metal, such as iron, the gas is no longer of the same nature.

After all the experiments which we have recorded, we might have been justified in concluding that the eudiometer of Volta must indicate the whole of the oxygene contained in the atmospheric air; but we wished to determine this point by direct experiment.

We analyzed an air composed of 20 parts of very pure oxygene and 80 of azote, obtained from the decomposition of ammoniac by oxygenated muriatic acid, all possible precautions being taken to prevent its becoming mixed with atmospheric air. 200 parts of this air, inflamed with 200 of hydrogene gas, gave five absorptions, the greatest of which differed from the smallest only by 5,1000 and the mean term of which was 124,9.

These 124,9 parts indicate 41,6 of oxygene, of which the half 20,8 corresponds to 100 of our factitious air.

We find therefore 0,008 of oxygene more than we had employed, which might seem to indicate that the proportion of 100 of oxygene to 200 of hydrogene is rather too great; but we must observe that our hydrogene, though carefully prepared, still emitted light with phosphorus, and that in order to explain our result, it is sufficient to suppose that the azote contained a hundredth part of oxygene, which is sufficiently probable if we consider that the oxygenated muriatic acid is very speedily decomposed by the action of light.

It appears from what has been said, that the results afforded by Volta’s eudiometer admit very well of comparison, and that the limits of their differences may be reduced very nearly to a thousandth part of the air analyzed.

It appears also that by its means we may estimate very minute differences between two airs or very small quantities of hydrogene mixed with the atmospheric air.

Independently of the property which this instrument possesses of indicating the whole quantity of oxygene contained in an air, it is the only one with which we can ascertain the proportion of hydrogene in a gaseous mixture, and in this point of view it still merited to have its mode of acting attended to and studied.

Thus the illustrious Volta, to whom Natural Philosophy is indebted for so many beautiful discoveries, has also the merit of having furnished Chemistry with the most accurate and valuable instrument for analyses.

Analysis of the Atmospheric Air by Volta’s Eudiometer.

As we have now proved that Volta’s eudiometer gives results very capable of comparison, that it can indicate the whole quantity of oxygene contained in the air, and that it has over the other eudiometrical means in which the absorbing substance is solid or liquid, the advantage of giving multiples of the quantity of oxygene to be estimated; we shall proceed to apply it to the analysis of the air.

If the proportion of 100 of oxygene and 200 of hydrogene which we have established be strictly correct, we shall obtain the proportion of the oxygene to the azote likewise correct; but supposing even that the quantity of hydrogene were too large or too small by 5 parts, the error would not amount to more than three thousandth parts of the air analyzed, and we should have the advantage of obtaining a greater precision than by any of the other known eudiometrical means.

The atmospheric air which we have analyzed was collected over the middle of the Seine in cold, temperate and rainy weather, and during the prevalence of different winds.

In order to obtain a greater parity between the circumstances, and better to appreciate the differences in the nature of the air, if there were any, we analyzed on the same day the different portions of air which we had collected in different weathers, and which we had preserved in glass vessels well closed and inverted over the water.

For brevity’s sake, we have comprized in the table annexed to this memoir, the absorptions produced by the inflammation of 200 parts of air and 200 of hydrogene gas, and have at the same time indicated the correspondent quantities of oxygene.

It will there appear that our experiments prove, in the first place, that there are no variations exceeding a thousandth part in the quantity of the oxygene of the air, though that which we analyzed, having been collected during the prevalence of different winds, came from regions very remote from each otber; and in the second place, that the proportion of the volume of the oxygene to the other gases which exist in the air, is as 21 to 79. The first result, that the air does not vary in its composition, is rigorously exact, because it is independent of the proportion of the hydrogene and oxygene gas which constitutes water; but the second result, that the air contains 21 hundredths of oxygene, can also deviate but very little from the real truth; for if we suppose that the quantity of hydrogene required for saturating 100 parts of oxygene were larger or smaller by 5 parts than that which we have assigned (and which we have reason to believe is correct within a much less difference), the error with respect to the proportion of oxygene which we have found in the air, would not amount, as we have already observed, to more than three thousandth parts of the air analyzed.

But as many meteoric phenomena may be attributed to inflammation of hydrogene gas, it has been endeavoured to explain them by admitting the existence of this gas in the atmosphere.

We therefore thought it a very interesting enquiry, to investigate whether the air actually contains hydrogene gas; and in order the more easily to discover it, we made a gaseous mixture, in which we were sure that there was none of it, and we made a comparative analysis of the two airs.

We made a mixture of 20 parts of oxygene and 80 parts of azote, obtained from ammoniac by means of the oxygenated muriatic acid, and we detonated 300 parts of each of the two airs with 100 of hydrogne; but the result of six experiments made with the atmospheric air was exactly the same as that of six others made with the factitious air. And as we have shewn that we could appreciate less than three thousandth parts of hydrogene, it must be concluded that the atmosphere does not contain this gas, or if it contains it, its quantity does not amount to three thousandth parts.

It cannot however be doubted that there exists a little hydrogene in the air; some is daily evolved from marshes, but its quantity may be so small (a thousandth for example) as to elude all our means.

The proportion of carbonic acid existing in it ought to be much more considerable, if we reflect upon the abundance of the sources which furnish it, and yet, if it did not form insoluble combinations with lime and barytes, it would perhaps still remain to be ascertained by the determination of its volume whether any of it existed in the air.

The carbonic acid, it is true, cannot accumulate in the air, because vegetation decomposes it; but is it proved that there do not exist causes which return the hydrogene to the earth, and thereby prevent it from accumulating in the air?

To draw a conclusion from the preceding experiments, we shall say: 1, that the atmosphere does not vary generally in its composition; 2, that the quantity of oxygene which it contains is 21 hundredths; 3, finally, that it does not contain any hydrogene which we are able to appreciate.

This identity of composition in which the principles of the atmosphere are constantly maintained, and this absence of hydrogene which our experiments prove, must give confidence to the Geometrician with respect to the theory of refractions.

The refractive power of the different gases being different, and that of hydrogene being stronger than that of oxygene and of azote, the theory of refractions, which is founded only upon the variations of the barometer and thermometer, would be very imperfect if the atmosphere changed in its constituent principles; but fortunately it is easy to prove that these changes do not take place in a sensible degree, and that the hydrogene gas, the refractive power of which is very great, does not exceed 0,003, at least as far as the greatest heights to which men have ever ascended.

The Geometrician will therefore have nothing else to consider in the Theory of Refractions than the barometer, the thermometer, and the hygrometer.

In fact, a little reflexion will be sufficient to convince us that the atmosphere cannot vary very considerably in the space of some years, and still loss of some days, at least if we do not speak of some very particular local variations.

For if it varied thus in so short a space of time, by what miracle should it do so, and return suddenly to its original state?

How conceive a cause sufficiently powerful to change from one day to another the proportion of oxygene by a thousandth part only, unless we should admit the existence of a magnetic or electric power, or any other equally imaginary, which could change, by unknown modifications, the oxygene into azote, and vice versâ?

It is possible that the atmosphere varies very slowly both in the proportion of its principles and in its weight; but these variations, for being so insensible, ought not the less to engage the attention of philosophers.

If it is now well proved that in general the atmosphere does not vary in its composition, we must seek for the cause of the differences which some have imagined they had discovered in it, in the local circumstances under which it was analyzed.

Volcanoes upon high mountains, particular fermentations, water issuing from a marsh or a lake, might perhaps in some degree impair the purity of the air in contact with them, either by depriving it of oxygene, or by exhaling into it non-respirable elastic fluids; but how trifling must not this diminution of the proportion of the oxygene be in so large a mass of continually agitated air, when we consider that in places where a great number of individuals is collected, or in those where there seems to exist a focus of infection, the air nevertheless experiences but very slight variations.

We have analyzed two portions of air, one of which was collected in the pit of the Theatre Français, immediately before the curtain was raised for performing the after-piece, three hours and a half after a great number of spectators had assembled, and the other of which was collected three minutes after the entertainment had ended in the most elevated part of the house.

These two portions scarcely rendered lime water turbid; the atmospheric air indicating 0,210 of oxygene, the air of the pit indicated only 0,202 and that of the highest part of the house 0,204.

Analysis of the atmospheric air. Of the air of the pit.

Of the air of the highest part of the building.

200 atmosph. air. 200 atmosph. air. 200 atmosph. air. 200 hydrogene.

200 hydr.

200 hydr.

126 air absorbed.

121,5 abs. 122,5 abs. 21 oxygene.

20,2 oxyg.

20,4 oxyg.

M. Seguin has also analyzed the air of hospital wards which he had kept closely shut up for the space of twelve hours, and found it to be almost as pure as the atmospheric air, although it had an insupportably infectious smell.

If therefore, even under circumstances the most favourable to the absorption of the oxygene, the air does not lose one hundredth part of it, we cannot thereby account for the sense of anxiety which we feel in close and crowded apartments, or the maladies which are peculiar to the vicinity of lakes and marshes, or to certain countries.

Under some circumstances they will be produced by emanations which elude all our eudiometrical means, and which act in a peculiar manner upon the human body.

Thus a bubble of sulphurated hydrogene gas, of oxygenated muriatic acid, a putrid exhalation, even a flower, may fill an immense space with this odour, and astonish our imagination by their extreme subtility, even when we are ready to sink under their action.

The pestilential miasmata may be equally subtile without being the less deleterious, and equally elude all our means of analysis.

Fortunately, if we cannot seize these atomic substances and determine their nature, we may at least, after the labours of M. Guyton, which have been productive of such great benefit to mankind, destroy their action.

But under other circumstances the maladies may arise from the humidity of the air, from its temperature, from its electric state, or in general from the state of the atmosphere with respect to the peculiar state of the individual affected; and in these cases, which may be very frequent, the malady may make great ravages without its being possible to arrest its progress: it would therefore be illusory to attribute all to a single cause, when the state of human health depends upon the concurrence of all the circumstances under which men are placed.

But let us now recapitulate the principal facts contained in this first part of our Memoir, and call to mind some of the explanations which we have offered, hoping we may be permitted to consider them as expressing the real state of the facts.

The solution of an alkaline sulphuret, when made cold, does not absorb the azote, and it may be employed with advantage for the analysis of the air: when made hot, it absorbs it, and indicates a greater diminution of volume in the air than that which proceeds from the absorption of the oxygene.

It is to the water only, and not to the sulphuret, that we are to attribute this property.

There are certain proportions of oxygene and hydrogene in which the combustion produced by the electric spark may be complete; there are others in which the combustion ceases before being completed; and finally there are others in which it cannot take place at all.

These last phenomena seem to depend upon the circumstance that the temperature necessary for the combustion is not sufficiently elevated, and not upon the mutual affinity of the gases: for in all cases in which the combustion is not complete, we need only raise the temperature artificially in order that it may become so.

When the hydrogene and oxygene are not entirely absorbed, we find them again in the residues, and prove that they have not formed new combinations.

When we cannot inflame a gaseous mixture which contains oxygene and hydrogene, it will be sufficient to augment the proportion of these two gases.

The meteoric phenomena cannot be results of the inflammation of hydrogene gas, since even in an air consisting merely of pure oxygene, it would require more than 6 hundredths of hydrogene for combustion to take place, and even then it would be only partial.

Electricity seems to act in the inflammation of oxygene and hydrogene gas by the heat arising from the compression which it produces in its passage through their mixture.

These two gases, by their combination, form water, which is constant in its nature.

If the galvanic phenomena seem to prove that water is susceptible of oxygenation or hydrogenation, they may be equally well accounted for without the aid of such an hypothesis.

One hundred parts by volume of oxygene require for their saturation 200 of hydrogene.

This proportion is independent of the changes of temperature and moisture, whereas that determined by weight varies under the same circumstances, because the two gases do not carry into the combination quantities of water which are in the same proportion as their quantities by weight; whence it results that the proportions of the water which have been established must be modified.

The Eudiometer of Volta is capable of indicating the whole quantity of oxygene contained in a given volume of air within about a thousandth part of that volume, and its results admit very well of comparison.

In the present state of our knowledge, it is the most exact of our eudiometrical means; it is not only capable of indicating very small quantities of oxygene or of hydrogene, and determining the purity of the lastmentioned gas, but it has likewise the advantage of giving multiples of the quantity to be estimated.

It has therefore in all these respects a very decided superiority over the other eudiometrical means.

The atmosphere contains only 0,21 in volume of oxygene, and it does not vary in its composition.

It contains no hydrogene, or if it contains any, its quantity cannot amount to 3,003.

to be concluded in our next

Experiments on the Eudiometrical Means, and on the Proportion of the Constituent Principles of the Atmosphere.

By Messrs. Humboldt and Gay-Lussac.

(Concluded from Page 381.)

Of the Nature of the Air extracted from Water, and of the Action of Water upon the Gases, both pure and mixed.

WE have hitherto examined the eudiometrical means which lead to the exact analysis of the atmospheric air.

We should undoubtedly have confined ourselves to the enunciation of the principal facts to which the first part of our investigation has conducted us, had we not remarked in the course of these experiments, and particularly of those upon the sulphurets, that water and other liquids exert an action upon the air which may frequently become a cause of error, the more important as it has hitherto been little attended to.

We therefore apprehended, that we should leave our labours in a still more imperfect state than we have already done, if we had not directed our inquiries to this action of water upon the gases, both pure and mixed, which are subjected to it.

The experiments which we have made with a view to this object, shall form the conclusion of our memoir.

It is generally known that water can hold air in solution.

Boyle, Huggens, and Mairan, have discussed this fact; but they did not possess the means of discovering that this dissolved air differs chemically from the atmospheric air.

The celebrated Priestley was the first who observed that the air extracted from water contains more oxygene than common air. M. Hassenfratz afterwards asserted that rain-water evolved an air which contained nearly forty-hundredths of oxygene; and Messrs. Ingenhoutz and Breda, in their experiments upon nitrous gas, were led to similar results.

But if it is already known that the air contained in water is more pure than atmospheric air, it has also been maintained that water absorbs the oxygene gas more abundantly and more readily than the azote. M.

Fourcroy even mentions the curious fact, which however he himself does not conceive to be sufficiently authenticated, that water charged with oxygene gas absorbs hydrogene gas, upon which common water has hardly any action.

We shall see hereafter, that that which it exercises upon any gas is modified by the nature of the air which it already holds in solution.

Mr. Henry, in a memoir lately published in England, has examined the absorption of different gases by water deprived of air. He effected these absorptions under a pressure equal to that of two or three atmospheres; but he has not treated of the mixture of different gases, nor of the affinity which water has for this mixture: he confines himself to the examination of the quantity absorbed, according to the difference of temperature and of barometric pressure, without directing his inquiries to the action of water already saturated with other gases.

We conceived that we ought not to neglect a subject so intimately connected with eudiometrical inquiries, and to which chemists have hitherto paid so little attention.

We have examined the degree of affinity by which the oxygene dissolved in water is retained in it, according to the temperature and the salts which it may contain.

We placed in contact with water equal quantities of gases, both pure and mixed, and we observed the changes which these mixtures undergo in their chemical composition.

Finally, we have begun to examine a problem of great importance to meteorology, namely, whether rain water holds hydrogene in solution.

All these inquiries, which we intend to prosecute during the course of this year, and particularly upon the mountains which we purpose to visit, are not as yet in a very advanced state; we shall therefore content ourselves for the present with offering some leading facts, which, we trust, will not be deemed altogether destitute of interest by philosophers.

After mixing the whole mass of air which water yields by boiling, without separating the portions first disengaged from those evolved at the end of the operation, we found by Volta’s eudiometer that distilled water, which has re-absorbed atmospheric air, gives an air which contains in the hundred parts 32,8 of oxygene; the water of the Seine, one which contains 31,9 oxygene; and rain-water an air containing 31,0 of the same principle.

From these experiments it follows, that we may extract from these three waters air nearly equally rich in oxygene, and purer by ten-hundredths than the atmospheric air.

This quantity of oxygene is more variable in the waters of wells, which in the bowels of the earth are in contact with substances that exercise an affinity upon oxygene.

Water of the Seine, collected at another period, yielded an air containing only 29,1 of oxygene, an air somewhat less pure than that of the rain water.

If distilled water which has again absorbed air, rain water, and river water, yield airs all of which are much purer than the atmospheric air, it will be still more interesting to examine the nature of the gaseous mixtures which water yields when heated gradually.

It is in these experiments that the great affinity of oxygene for this liquid displays itself in its strongest light.

We gradually heated water of the Seine to ebullition, and collected the air which is disengaged by successive and unequal portions.

We took 200 parts of these portions, and having detonated them with 200 parts of hydrogene gas, they gave us the following results.

Portions of air, according to the order of their evolution.

Absorption.

Oxygene gas contained in 100 parts of the air evolved.

First ‒ ‒ ‒ 142,0 ‒ ‒ ‒ 23,7 Second ‒ ‒ ‒ 164,5 ‒ ‒ ‒ 27,4 Third ‒ ‒ ‒ 185,0 ‒ ‒ ‒ 30,2 Fourth ‒ ‒ ‒ 195,0 ‒ ‒ ‒ 32,5

These experiments, several times repeated, prove that water abandons at first an air, the purity of which is only a little superior to that of atmospheric air; after which the purity of this air, or the disengagement of oxygene, progressively increases, and the last gaseous portions which the heat separates contain the most oxygene.

On repeating this experiment upon snow-water, the first portions of air had 24,0; the last 34,8 of oxygene.

Possibly if the mass of water were heated still more slowly, and the small portion of air which is first evolved carefully separated, we should have at the beginning of the operation an air still less pure than that which we obtained.

Water therefore does not exercise an uniform action upon oxygene and upon azote, and elevation of temperature diminishes its action upon the first less than upon the last.

It is even probable that the portion of air which is disengaged towards the end of the operation would be of greater purity than that of 32 or 34 per cent. of oxygene, if the water contained in the vessel which receives the gaseous mixture did not begin to be heated, and to evolve its air, which then is only at 23 per cent. of oxygene.

This disengagement takes place especially when the aqueous vapour begins to pass, and it is this diminution of the purity of the air expelled last and the inequality of volume of the four separate portions, which explain how the whole mass of air extracted at once contains to the amount of 31 hundredths of oxygene.

This unequal action of water upon the oxygene and the azote manifests itself also in the solution of salts.

We have observed that the pure water of the Seine gave by ebullition nearly one half more of air than the same water charged with muriate of soda.

The cause of this diminution consists in the very considerable quantity of air which already disengaged from the water, still cold, whilst the solution of the salt is taking place.

This air, accurately analyzed, shewed only 0,225 of oxygene, whilst the air obtained by boiling the water charged with muriate of soda, contained 0,305.

Hence it appears that the water, in dissolving the salt, abandons a part of the air which it holds in solution, but that this part contains oxygene in a less proportion than that which it retains.

The condensation which water experiences in passing from the liquid to the solid state presents to us a third class of phenomena analogous to those which we have just described.

Melted ice yields only about half the quantity of air which we obtain from common water, and it is to be observed that it does not begin to evolve its air until its temperature has been raised as high as above the sixtieth degree of the centigrade thermometer.

The air obtained, divided into two unequal portions, shewed in Volta’s eudiometer 27,5 and 33,5 of oxygene.

Then the purest air was again evolved the last.

The small quantity and the great purity of the air evolved from melted ice, proves that water, as it passes into the solid state, abandons a large portion of its air, and that this part, separated during the congelation, is air of much less purity than that which it retains.

Thus three phenomena, which at the first view appear different, water at a temperature from 35° to 40° centigram, water dissolving salts cold, and pure water congealing into ice, present results entirely similar to each other in their action upon oxygene and azote.

A moderate temperature acts like the solution of a salt, and this like the transition from the liquid into the solid state.

The water in these three cases evolves an air of less purity than that which it holds in solution.

It is a very remarkable phenomenon that the congelation of water into the state of snow expels less air from it than the formation of ice.

We melted some newly fallen snow, and heating it gradually, we obtained a volume of air nearly twice as great as that which melted ice yields.

The air extracted from the snow-water was almost in equal abundance with that evolved from the water of the Seine; for the latter gave by ebullition 1940 measures of air, while the same volume of snow-water yielded 1892.

These 1892 parts collected into 5 portions, according to the periods at which they were expelled by the heat, shewed successively in Volta’s eudiometer.

The first ‒ ‒ ‒ ‒ ‒ 24,0 of oxygene.

Second ‒ ‒ ‒ ‒ 26,8.

Third ‒ ‒ ‒ ‒ 29,6.

Fourth ‒ ‒ ‒ ‒ 32,0.

Fifth ‒ ‒ ‒ ‒ 34,8.

This last portion of air is the purest we have yet extracted from any water.

The volumes of each portion being known, calculation gives us for the purity of the air considered collectively 28,7 of oxygene.

The water of the Seine yielded on the same day an air which was less pure by [Formel] . Otherwise the two waters, that of melted snow and that of the river, give a volume of air equal to nearly [Formel] of their own.

These experiments upon snow-water and upon melted ice, which we design hereafter to vary in many ways, present considerations of great importance to the study of meteorology.

Snow is nothing else than an aggregate of small crystals of ice which are formed in the higher regions of the atmosphere, and yet these small crystals melted give a volume of air nearly twice as great as that which the ice formed upon our rivers yields.

Hence it is to be concluded that when the water dissolved in the air congeals into snow, it does not expel that large portion of air which it disengages in its congelation upon the surface of the earth, unless we might be allowed to suppose that snow retains between its small crystals a certain quantity of air which it absorbs as it melts; for it appears that it is principally at the moment of its congelation that water abandons the greater part of its air.

The beautiful vegetation which surrounds the Glaciers, the rapid developement of plants when the snow melts in the spring, and several phenomena which are supposed to have been observed in agriculture and in bleaching, have led to the notion that the waters of ice, snow and rain, produced peculiar effects by a large quantity of dissolved oxygene which was evolved from them. The experiments which we have hitherto made do not seem favourable to these conjectures.

There undoubtedly exist wells, the waters of which contain air inferior in purity to the atmospheric air, and we do not doubt that these waters, charged besides with salts and carbonic acid, must have an influence upon vegetation and bleaching very different from that of snow-water.

But the differences produced by distilled water exposed to the air, rain-water, snow-water and the water of the Seine, are not easily to be accounted for by the oxygene dissolved in them, if we recollect that all these airs contain water of nearly equal purity, and that they contain it in almost equal quantity.

The phenomena of vegetation, like those of meteorology, are so complicated; they depend upon the concurrence of so many causes at once, that we must be very careful not to attribute to one what is the effect of many.

The experiments which we have recorded on the force with which the last particles of oxygene dissolved are retained by water, throw an additional light upon the state in which air exists in liquids.

The specific gravity of distilled water and of that charged with air, being apparently the same, Mairan has concluded with reason that this air cannot be lodged in the fluids in an elastic state.

The chemical phenomena confirm this opinion.

If water deprived of its air by distillation, or by the air-pump, could be considered as a sponge the pores of which are empty, how should these pores not fill themselves on the first contact with atmospheric air?

But this solution of air in water can be considered only as the effect of a chemical affinity; for how but by this affinity could the absorption of the gases by the water deprived of air be so slow, and above all, how else should the water dissolve one gas in preference to another?

How should water charged with one species of air abandon a part of it, as we shall see hereafter that it does, to receive another of a different nature?

After having examined the air which may be extracted from water under different circumstances, we shall conclude our Memoir with the experiments which we have made in placing gases, both pure and mixed, in contact with water.

The gases which we employed were exactly of the same volume, and the quantity of filtrated water of the Seine was nearly equal.

After a space of from 6 to 8 days, we not only measured the quantity of the volumes absorbed, but we also analyzed the residues.

This analysis was the more necessary as one might often be tempted to conclude from a very small change in the volume of the gas placed in contact with the water, that the latter has no sensible action upon it, whilst the nature of the residue indicates that this action has been very strong, but disguised by the quantity of air evolved from the water in lieu of that which has been absorbed.

Of all the gases, the oxygene is that which the water of the Seine absorbs in the most considerable quantity.

If we place in contact with this water, already charged with air, 100 parts of oxygene gas, 100 of azote, and 100 of hydrogene, the oxygene gas will have diminished by 40 parts when the two others have lost only 5 and 3 parts.

But the real absorption of the oxygene gas is still much more considerable than its apparent diminution indicates.

The 60 parts of residue, instead of being pure oxygene, contained 37 parts of azote and only 24 of oxygene; so that the 100 parts of oxygene employed, had lost upon the Seine water 77 parts, which had expelled 37 parts of azote.

Thus a river water which had been exposed for a long time to the atmosphere, and which might be considered as saturated with air, absorbs a large quantity of oxygene when it is presented to it.

It takes it, without abandoning a portion of azote equal in volume to the oxygene absorbed.

The action of water upon the volume of hydrogene gas appears to amount almost to nothing.

The inequality of the results which we obtained prevent our saying any thing concerning the slight changes which it may undergo during their contact.

The volume of pure azote gas diminishes upon the water by between 2 and 3 hundredth parts; but the residue is no longer pure azote: we discovered in it 11 parts of oxygene which had been displaced from the water by 14 parts of azote.

The azote therefore dislodges ozygene from the water, as oxygene dislodges azote.

The action is analogous, but the quantities absorbed and dislodged are different.

The contact of the river-water with a mixture of hydrogene and oxygene gases was examined under various circumstances.

Sometimes we mixed the two gases in equal quantities, sometimes we caused one or the other to predominate.

The diminution of the volume of the gases is greater when the oxygene predominates; that is to say, when we expose to the water a mixture of 200 parts of oxygene with 100 of hydrogene.

In all these experiments azote is again dislodged from the water.

In analyzing the residuum of a mixture of equal parts of oxygene and of hydrogene, we found in 100 parts 20 parts azote, 50 hydrogene, and 30 oxygene.

The greater the absorption of oxygene was, the more azote we found to be dislodged.

Having mixed 400 parts of oxygene with 200 of hydrogene, this volume was reduced upon Seine water, in the space of ten days, from 600 parts to 562. If this residue had undergone no chemical change in its proportions, if no other gas had been dislodged, it ought to have contained 375 parts of oxygene and 187 of hydrogene; but our analysis shewed it to contain 246 azote, 142 hydrogene, and 174 oxygene.

These experiments prove that hydrogene, which, when placed alone in contact with water, is not sensibly absorbed by it, is dissolved in it, and that in a very considerable proportion, when it is mixed with oxygene.

On this subject a question of great importance to Natural Philosophy presents itself; namely, whether this hydrogene absorbed by the water exists in it in the state of hydrogene, or whether it combines in it with oxygene and so forms water.

We have endeavoured to solve this question by leaving a mixture of hydrogene and oxygene in contact with water recently deprived of all its air by ebullition.

After the space of 12 days, we distilled this water, and on analyzing the air disengaged from it, we found the hydrogene to exist in it in such abundance that we could inflame it in Volta’s eudiometer without adding to it any other gas.

This experiment proves beyond a doubt that the hydrogene absorbed is found again in the water.

But did this water give up again the same quantity which it had absorbed?

Would not this hydrogene dissolved in the water unite with the oxygene if it had been lodged in it for several months?

We design to make a long series of experiments upon this subject.

If the hydrogene and oxygene contained in the water could combine in it, we should be able more readily to conceive how the hydrogene gas which rises from the surface of the earth is not to be discovered either in the air which surrounds us, or in the high regions of the atmosphere into which we have ascended.

We must recollect upon this subject that having carefully examined rain-water, in order to discover hydrogene in it, we have ascertained that the air disengaged from this water did not contain a quantity of it amounting to [Formel] . We shall repeat these experiments upon rain that has fallen in different seasons, and particularly during storms.

River water placed in contact with mixtures of gases has in general acted less upon the mixtures of oxygene and azote than those of oxygene and of hydrogene.

This result will appear the less surprizing if we take a general view of what takes place in these phenomena.

We find that water has a continual tendency to place itself in a state of equilibrium with the gases presented to it.

If we present oxygene to it, it abandons azote.

If we place it in contact with azote, it abandons oxygene.

If we present to it a mixture of oxygene and hydrogene, it absorbs a part of these two gases, and replaces them with azote.

It every where tends to modify the proportions of the air which it holds in solution according to the nature of the gases which are presented to it.

Now the water of the Seine being charged with a mixture of azote and of oxygene, it seems natural that it should excite more action upon a mixture of hydrogene and oxygene than upon that of azote and oxygene, which is analogous to the air which it holds in solution.

In order to obtain an accurate idea of these phenomena, we shall make experiments upon water recently deprived of air, by charging it with different gases both pure and mixed, and examining the action of this water after a long space of time; for frequently it is not till after a long repose that nature is enabled to overcome the obstacles which oppose themselves to the action of the affinities.

Here we shall conclude the account of the inquiries into which we have been engaged for several months past.

The more extensive the field is which we purpose to explore, the more strongly we feel the imperfection of the work which we now present to the public; but this sentiment, so far from discouraging us, will only render us the more assiduous in interrogating nature, in order that we may carry these researches to a higher degree of perfection.

TABLE Representing the Analysis of the Air.

Days when the Air was collected.

Temperature expressed in Degrees of the centigrade Thermometer.

State of the Atmosphere.

Absorption proceeding from the Inflammation of 200 of Air and 200 of Hydrogene.

Quantity of oxygene contained in 100 Parts of Air. Brumaire 26 ....... 7°.3 Sky overcast; wind E. .......... .... 126,0 .. 21,0 126,0 .. 21,0 27 ........ 4,5 Sky overcast; wind E. S. E. ...... .... 126,0 .. 21,0 126,0 .. 21,0 28 ........ 4,7 Light rains; very high wind S. W. .... 126,0 .. 21,0 126,0 .. 21,0 29 ....... 10,0 Light rain; wind S. ........... .... 126,0 .. 21,0 126,5 .. 21,1 30 ....... 12,5 Sky overcast; wind S. W. ........ .... 126,8 .. 21,2 126,0 .. 21,0 Frimaire 1 ........ 6,7 Sky cloudy, slight rain; wind S. W. .... 126,0 .. 21,0 126,0 .. 21,0 2 ........ 1,5 Cloudy sky; wind W. .......... .... 126,0 .. 21,0 126,0 .. 21,0 3 ........ 8,5 Rain; wind S. ................ .... 126,3 .. 21,0 126,5 .. 21,1 4 ....... 10,6 Sky overcast; wind S. W. ........ .... 126,2 .. 21,0 126,5 .. 21,1 5 ........ 3,3 Cloudy sky; wind E. .......... .... 126,5 .. 21,1 126,0 .. 21,0 6 ...... —1,6 White frost; wind N. .......... ...... 126,0 .. 21,1 7 ...... —1,3 Snow; wind N. ................ ...... 12,65 .. 21,0 10 ...... —4,1 Fog; wind N. N. E. ............ ...... 126,0 .. 21,0 12 ...... —2,8 Cloudy sky, mists; wind E. .... ...... 136,5 .. 20,9 14 ........ 4,2 Rain; wind S. ................ ...... 126,0 .. 21,0 16 ........ 3,1 Thick fog .................... ...... 126,0 .. 21,0 22 ........ 9,6 Rain; wind S. S. W. ............ ...... 126,0 .. 21,0 28 ...... —2,2 Sky overcast; wind N. E. ........ ...... 126,0 .. 21,0 Nivose 2 ........ 1,0 Hoar frost, thick fog; wind S. E. .. ...... 126,0 .. 21,0