John Dalton (1766-1844)

Experimental Enquiry into the Proportion of the Several Gases or Elastic Fluids, Constituting the Atmosphere

Memoirs of the Literary and Philosophical Society of Manchester 1, 244-58 (1805)

Read Nov. 12, 1802

In a former paper which I submitted to this Society, "On the constitution of mixed gases," I adopted such proportions of the simple elastic fluids to constitute the atmosphere as were then current, not intending to warrant the accuracy of them all, as stated in the said paper; my principal object in that essay was, to point out the manner in which mixed elastic fluids exist together, and to insist upon what I think a very important and fundamental position in the doctrine of such fluids:--namely, that the elastic or repulsive power of each particle is confined to those of its own kind; and consequently the force of such fluid, retained in a given vessel, or gravitating, is the same in a separate as in a mixed state, depending upon its proper density and temperature. This principle accords with all experience, and I have no doubt will soon be perceived and acknowledged by chemists and philosophers in general; and its application will elucidate a variety of facts, which are otherwise involved in obscurity.

The objects of the present essay are,

  1. To determine the weight of each simple atmosphere, abstractedly; or, in other words, what part of the weight of the whole compound atmosphere is due to azote, what to oxygen, &c. &c.
  2. To determine the relative weights of the different gases in a given volume of atmospheric air, such as it is at the earth's surface.
  3. To investigate the proportions of the gases to each other, such as they ought to be found at different elevations above the earth's surface.
To those who consider the atmosphere as a chemical compound, these threeobjects are but one; others, who adopt my hypothesis, will see they are essentially distinct.-- With respect to the first: It is obvious, that, on my hypothesis, the density and elastic force of each gas at the earth's surface, are the effects of the weight of the atmosphere of that gas solely, the different atmospheres not gravitating one upon another. Whence the first object will be obtained by ascertaining what share of elastic force is due to each gas in a given volume of the compound atmosphere; or, which amounts to the same thing, by finding how much the given volume is diminished under a constant pressure, by the abstraction of each of its ingredients singly. Thus, if it should appear that by extracting the oxygenous gas from any mass of atmospheric air, the whole was diminished 1/5 in bulk, still being subject to a pressure of 30 inches of mercury; then it ought to be inferred that the oxygenous atmosphere presses the earth with a force of 6 inches of mercury, &c.

In order to ascertain the second point, it will be further necessary to obtain the specific gravity of each gas; that is, the relative weights of a given volume of each in a pure state, subject to the same pressure and temperature. For, the weight of each gas in any given portion of atmospheric air, must be in the compound ratio of its force and specific gravity.

With respect to the third object, it may be observed, that those gases which are specifically the heaviest must decrease in density the quickest in ascending. If the earth's atmosphere had been a homogeneous elastic fluid of the same weight it is, but ten times the specific gravity, it might easily be demonstrated that no sensible portion of it could have arisen to the summits of the highest mountains. On the other hand, an atmosphere of hydrogenous gas, of the same weight, would support a column of mercury nearly 29 inches on the summit of Mount Blanc.

The several gases constantly found in every portion of atmospheric air, and in such quantities as are capable of being appreciated, are azotic, oxygenous, aqueous vapour, and carbonic acid. It is probable that hydrogenous gas also is constantly present; but in so small proportion as not to be detected by any test we are acquainted with; it must therefore be confounded in the large mass of azotic gas.

1. Of the Weight of the Oxygenous and Azotic Atmospheres.

Various processes have been used to determine the quantity of oxygenous gas.
  1. The mixture of nitrous gas and air over water.[*]
  2. Exposing the air to liquid sulphuret of potash or lime, with or without agitation.
  3. Exploding hydrogen gas and air by electricity.
  4. Exposing the air to a solution of green sulphat or muriat of iron in water, strongly impregnated with nitrous gas.
  5. Burning phosphorus in the air.
In all these cases the oxygen enters into combination and loses its elasticity; and if the several processes be conducted skillfully, the results are precisely the same from all. In all parts of the earth and at every season of the year, the bulk of any given quantity of atmospheric air appears to be reduced nearly 21 per cent. by abstracting its oxygen. This fact, indeed, has not been generally admitted till lately; some chemists having found, as they apprehended, a great difference in the quantity of oxygen in the air at different times and places; on some occasions 20 per cent. and on others 30, and more of oxygen are said to have been found. This I have no doubt was owing to their not understanding the nature of the operation and of the circumstances influencing it. Indeed it is difficult to see, on any hypothesis, how a disproportion of these two elements should ever subsist in the atmosphere.

As the first of the processes above-mentioned has been much discredited by late authors, and as it appears from my experience to be not only the most elegant and expeditious of all the methods hitherto used, but also as correct as any of them, when properly conducted, I shall, on this occasion, animadvert upon it.

  1. Nitrous gas may be obtained pure by nitric acid diluted with an equal bulk of water poured upon copper or mercury; little or no artificial heat should be applied.-- The last product of gas this way obtained, does not contain any sensible portion of azotic gas; at least it may easily be got with less than 2 or 3 per cent. of that gas: It is probably nearly free from nitrous oxide also, when thus obtained.
  2. If 100 measures of common air be put to 36 of pure nitrous gas in a tube 3.10th of an inch wide and 5 inches long, after a few minutes the whole will be reduced to 79 or 80 measures, and exhibit no signs of either oxygenous or nitrous gas.
  3. If 100 measures of common air be admitted to 72 of nitrous gas in a wide vessel over water, such as to form a thin stratum of air, and an immediate momentary agitation be used, there will, as before, be found 79 or 80 measures of pure azotic gas for a residuum.
  4. If, in the last experiment, less than 72 measures of nitrous gas be used, there will be a residuum containing oxygenous gas; if more, then some residuary nitrous gas will be found.
These facts clearly point out the theory of the process: the elements of oxygen may combine with a certain portion of nitrous gas, or with twice that portion, but with no intermediate quantity. In the former case nitric acid is the result; in the latter nitrous acid: but as both these may be formed at the same time, one part of the oxygen going to one of nitrous gas, and another to two, the quantity of nitrous gas absorbed should be variable; from 36 to 72 per cent. for common air. This is the principal cause of that diversity which has so much appeared in the results of chemists on this subject. In fact, all the gradation in quantity of nitrous gas from 36 to 72 may actually be observed with atmospheric air of the same purity; the wider the tube or vessel the mixture is made in, the quicker the combination is effected, and the more exposed to water, the greater is the quantity of nitrous acid and the less of nitric that is formed.

To use nitrous gas for the purpose of eudiometry therefore, we must attempt to form nitric acid or nitrous wholly, and without a mixture of the other. Of these the former appears from my experiments to be most easily and most accurately effected. In order to do this a narrow tube is necessary; one that is just wide enough to let air pass water without requiring the tube to be agitated, is best. Let little more nitrous gas than is sufficient to form nitric acid be admitted to the oxygenous gas; let no agitation be used; and as soon as the diminution appears to be over for a moment let the residuary gas be transferred to another tube, and it will remain without any further diminution of consequence. Then 7/19 of the loss will be due to oxygen.-- The transferring is necessary to prevent the nitric acid formed and combined with the water, from absorbing the remainder of the nitrous gas to form nitrous acid.

Sulphuret of lime is a good test of the proportion of oxygen in a given mixture, provided the liquid be not more than 20 or 30 per cent. for the gas (atmospheric air); if the liquid exceed this, there is a portion of azotic gas imbibed somewhat uncertain in quantity.

Volta's eudiometer is very accurate as well as elegant and expeditious: according to Monge, 100 oxygen require 196 measures of hydrogen; according to Davy 192; but from the most attentive observations of my own, 185 are sufficient. In atmospheric air I always find 50 per cent. diminution when fired with an excess of hydrogen; that is, 100 common air with 60 hydrogen, become 100 after the explosion, and no oxygen is found in the residuum; here 21 oxygen take 39 hydrogen.

2. Of the Weight of the Aqueous Vapour Atmosphere

I have, in a former essay, (Manchester Mem. vol. 5, p. 2, page 559.) given a table of the force of vapour in vacuo for every degree of temperature, determined by experiment; and in the sequel of the essay, have shown that the force of vapour in the atmosphere is the very same as in vacuo, when they are both at their utmost for any given temperature. To find the force of aqueous vapour in the atmosphere, therefore, we have nothing more to do than to find that degree of cold at which it begins to be condensed, and opposite to it in the table above mentioned, will be found the force of vapour. From the various facts mentioned in the essay it is obvious, that vapour contracts no chemical union with any of the gases of the atmosphere; this fact has since been enforced in the Annales de Chimie, vol. xlii. by Clement and Desorme.

M. de Saussure found by an excellent experiment, that dry air of 64° will admit so much vapour as to increase its elasticity, 1/54.-- This I have repeated nearly in his manner, and found a similar result. But the table he has given us of aqueous vapour at other temperatures is very far wrong, especially at temperatures distant from 64°.-- The numbers were not the result of direct experiment, like the one above.-- If we could obtain the temperatures of all parts of the earth's surface, for any given time, a mean of them would probably be 57° or 58°. Now if we may suppose the force of vapour equivalent to that of 55°, at a medium, it will, from the table, be equal to .443 of mercury; or, nearly 1/70 of the whole atmosphere. This it will be perceived is calculated to be the weight of vapour in the whole atmosphere of the earth. If that incumbent over any place at any time be required, it may be found as directed above.

3. Of the Weight of the Carbonic Acid Atmosphere.

From some observations of Humboldt, I was led to expect about 1/100 part of the weight of the atmosphere to be carbonic acid gas: but I soon found that the proportion was immensely overrated. From repeated experiments, all nearly agreeing in their results, and made at different seasons of the year, I have found, that if a glass vessel filled with 102,400 grains of rain water be emptied in the open air, and 125 grains of strong lime water be poured in, and the mouth then closed; by sufficient time and agitation, the whole of the lime water is just saturated by the acid gas it finds in that volume of air. But 125 grains of the lime water used require 70 grain measures of carbonic acid gas to saturate it; therefore, the 102,400 grain measures of common air contain 70 of carbonic acid; or 1/1400 of the whole.-- The weight of the carbonic acid atmosphere then is to that of the whole compound as 1:1460; but the weight of carbonic acid gas in a given portion of air at the earth's surface, is nearly 1/1000 of the whole; because the specific gravity of the gas is 1-1/2 that of common air. I have since found that the air in an assembly, in which two hundred people had breathed for two hours, with the windows and doors shut, contained little more than 1 per cent. of carbonic acid gas.

Having now determined the force with which each atmosphere presses on the earth's surface, or in other words, its weight; it remains next to enquire into their specific gravities.

These may be seen in the following Table.

Atmospheric air,1.000
Azotic gas,.966
Oxygenous gas,1.127
Carbonic acid gas,1.500
Aqueous vapour,.700
Hydrogenous gas,0.77[1]

Kirwan and Lavoisier are my authorities for these numbers; except oxygenous gas and aqueous vapour. For the former I am indebted to Mr. Davy's Chemical Researches; his number is something greater than theirs: I prefer it, because, being determined with a least equal attention to accuracy with the others, it has this further claim for credit, that 21 parts of gas of this specific gravity, mixed with 79 parts of azotic gas, make a compound of exactly the same specific gravity as the atmosphere, as they evidently ought to do, setting aside the unfounded notion of their forming a chemical compound. The specific gravity of aqueous vapour I have determined myself both by analytic and synthetic methods, after the manner of De Saussure; that is, by abstracting aqueous vapour of a known force from a given quantity of air, and weighing the water obtained--and admitting a given weight of water to dry air and comparing the loss with the increased elasticity. De Saussure makes the specific gravity to be ,71 or ,75; but he used caustic alkali as the absorbent, which would extract the carbonic acid as well as the aqueous vapour from the air. From the experiments of Pictet and Watt, I deduce the specific gravity of aqueous vapour to be ,61 and ,67 respectively. Upon the whole, therefore, it is probable that ,7 is very nearly accurate.

We have now sufficient data to form tables answering to the two first objects of our enquiry.

I. Table of the Weights of the different Gases constituting the Atmosphere.

Inch of Mercury.
Azotic gas23.36
Oxygenous gas6.18
Aqueous vapour.44
Carbonic acid gas.02
[total]30.00

II. Table of the proportional Weights of the different Gases in a given volume of Atmospheric Air, taken at the Surface of the Earth

Per Cent.
Azotic gas75.55
Oxygenous gas23.32
Aqueous vapour1.03[2]
Carbonic acid gas.10
[total]100.00

III. On the Proportion of Gases at different Elevations.

M. Berthollet seems to think that the lower strata of the atmosphere ought to contain more oxygen than the upper, because of the greater specific gravity of oxygenous gas, and the slight affinity of the two gases for each other. (See Annal. de Chimie, Tom. 34. page 85.) As I am unable to conceive even the possibility of two gases being held together by affinity, unless their particles unite so as to form one centre of repulsion out of two or more (in which case they become one gas) I cannot see why rarefaction should either decrease or increase this supposed affinity. I have little doubt, however, as to the fact of oxygenous gas observing a diminishing ratio in ascending; for, the atmospheres being independent on each other, their densities at different heights must be regulated by their specific gravities.-- Hence, if we take the azotic atmosphere as a standard, the oxygenous and the carbonic acid will observe a decreasing ratio to it in ascending, and the aqueous vapour an increasing one. The specific gravity of oxygenous and azotic gases being as 7 to 6 nearly, their diminution in density will be the same at heights reciprocally as their specific gravities. Hence it would be found, that at the height of Mount Blanc (nearly three English miles) the ratio of oxygenous gas to azotic in a given volume of air, would be nearly as 20 to 80; --consequently it follows that at any ordinary heights the difference in the proportions will be scarcely if at all perceptible.[3]


[*]This "nitrous air test" plays an important role in Priestley's characterization of oxygen.--CJG

[1] The specific gravity of hydrogen must be rated too low: if 100 oxygen require 185 hydrogen by measure, according to this 89 oxygen would require only 11 hydrogen to form water; whereas 85 require 15. Hydrogen ought to be found about 1/10 part of the weight of common air.

[2] The proportion of aqueous vapour must be understood to be variable for any one place: the others are permanent or nearly so.

[3] Air brought from the summit of Helvelyn, in Cumberland (1100 yards above the sea--Barometer being 26.60) in July 1804, gave no perceptible differences from the air taken in Manchester. --M. Gay-Lussac determines the constitution of air brought from an elevation of four miles to be the same as that at the earth's surface.


Back to the list of selected historical papers.
Back to the top of Classic Chemistry.