Thomas Thomson (1773-1852), M.D. F.R.S.

On the Daltonian Theory of Definite Proportions in Chemical Combinations

Annals of Philosophy 2, 32 (1813).[from David M. Knight, ed., Classical Scientific Papers: Chemistry (New York: American Elsevier, 1968)]

I promised in an early number of the Annals of Philosophy some observations on Mr. Dalton's theory of definite proportions, and I now sit down to fulfil that promise. Too much attention cannot be paid to this important theory, the developement of which I consider as the greatest step which chemistry has yet made as a science. It puts us in the way of establishing principles of rigid accuracy as the foundation of our reasoning, and to call in the assistance of mathematics to promote the progress of a science which has hitherto eluded the aid of that unrivalled instrument of improvement. The idea of definite proportions seems to have struck the mind of Richter, though the methods which he took to determine them were far from successful; and Mr. Higgins, in his wok on phlogiston, maintained the opinion that chemical bodies unite atom to atom. But the generalization of the doctrine, and the striking and irresistible proofs deduced from the combinations of the simple substances, and the acids and bases belong entirely to Mr. Dalton; without whose labours the theory would probably have still remained unknown. On the Continent the notions originally established by Dalton have been adopted, and ingeniously extended in certain cases by Gay-Lussac; and Berzelius has published a most elaborate, extensive, and accurate set of experiments on the same subject, which fully confirm the Daltonian doctrine, while he has deduced several subordinate laws from his analyses, which, though in some measure empyrical, are nevertheless of very considerable importance in determining the constitution of bodies.

In this country less attention has hitherto been bestowed upon Dalton's theory than could have been anticipated from the sagacity and enlightened views of many of our chemists. Dr. Bostock has even written an essay against it; but from the well- known candour and liberality of this able philosopher, I have no doubt that he will embrace it with ardour as soon as his doubts are removed. Sir Humphry Davy has embraced the Daltonian theory with some modifications and alterations of terms; but his notions are not quite so perspicuous as those of Mr. Dalton, and they do not appear to me so agreeable to the principles of sound philosophy. These, as far as I recollect, are the only chemists in this country who have written upon the subject (some observations of Dr. Wollaston, and of myself, excepted); though not the only persons who have embraced the opinions of Mr. Dalton. I shall in this essay state, in the first place, the nature of the theory, and the grounds upon which it has been advanced; in the second place, I shall state the laws, or subordinate canons, which have been deduced from analysis, in consequence of the introduction of this theory; and in the third place, I shall give a table of the numbers representing the proportions in which substances combine, deduced from the application of the Daltonian theory to the most accurate analyses hitherto made.

I. Outline of the Daltonian Theory.

It may be necessary to mention in the outset that I propose to give the view which I have accustomed myself to take of the subject, and that I would not be understood to make Mr. Dalton Answerable for the opinions which I shall state. I call it the Daltonian theory because I consider it as belonging to Mr. Dalton; because he first suggested it to me, and set me to think on the subject; and, of course, every thing here stated originated from him, either directly, or at least indirectly.

1. With respect to the nature of the ultimate elements of bodies, we have no means of obtaining accurate information; but it is the general opinion that they consist of atoms, or minute solids, incapable of farther division. That these atoms are mere mathematical points surrounded with spheres of attraction and repulsion, as Boscovich supposed, appears to me incomprehensible. They must, I think, be physical points, as minute as you will, but still possessed of length, breadth, and thickness. This opinion, I say, is generally received by philosophers; and I cannot, for my part, conceive any other. It is taken for granted as he foundation of the Daltonian theory; and, I presume, will be readily admitted by every one without hesitation.

2. In cases of the chemical union of one body with another, the substances combined are dispersed every where through the whole mass. Thus chalk is a compound of lime and carbonic acid. Now how minute a portion soever of chalk we tae, we shall find it to contain both lime and carbonic acid. How minute a portion soever of water we take, we shall find it to contain both oxygen and hydrogen. How minute a portion soever of saltpetre we examine, we shall find it to contain both nitric acid and potash. Now this could not be the case unless the atoms of the combining bodies united with each other. This accordingly is the opinion universally entertained respecting chemical combinations. It has been long generally admitted, and does not therefore require any farther illustration.

3. All chemical compounds contain the same constant proportion of constituents with the most rigid accuracy, no variation whatever ever taking place. Water is universally composed of 1 part of hydrogen and 7.5 parts oxygen; sulphuric acid, of 1 part of sulphur and 1.5 part of oxygen; carbonic acid, of 1 part of carbon and 2.7 parts nearly of oxygen, by weight. This permanency of chemical compounds is generally admitted. Indeed, the whole science of chemistry is founded on it, and depends upon it. Even Berthollet, who contends for indefinite proportions in the abstract, admits the incontrovertible fact that the proportions of chemical combinations in general are permanent.

4. This permanency of chemical compounds cannot be owing to any thing else than to the union of a certain determinate number of the atoms of one constituent with a certain determinate number of the atoms of the other. Let us suppose water the compound. Let the number of atoms of oxygen which unite be x, and of hydrogen y, then an integrant particle of water will in every case be x+y.

5. Oxygen has the property of uniting with different bases in various proportions, sometimes in two, sometimes in three, four, or even six proportions with the same base. Thus with azote it unites in four proportions, with carbon in two, with mercury in two, and so on. Now if we represent the weight of base with which the oxygen unites by a, and suppose all the different proportions of oxygen to unite with this portion of base; and if we denote the first portion of oxygen by b; then, in general, the constituents of the different compounds formed by the union of the different doses of oxygen with the base will be as follows:--

Suppose 10 parts of oxygen enter into the first compound, then 20 parts enter into the second compound, 30 parts into he third compound, and 40 parts into the fourth compound. Hence, whatever number of atoms of oxygen enter into the first compound, twice that number enters into the second, thrice into the third, and four times that number into the fourth.

Hence it is clear that there is a determinate number of atoms of oxygen which always enter into these combinations. If we represent this number by x, then a + x is the first compound, a + 2x the second, a + 3x the third, and a + 4x the fourth. Now it would be singular if 2, 3, 4, &c. atoms of oxygen were to be always inseparably linked together, so as never to be able to enter into combinations separate. It is much more simple to conceive that x represents only one atom. Even though the opinion should not be mathematically true, till it would be proper to adopt it: for as far as our calculations are concerned, a number of atoms of oxygen constantly and invariably united constitute a compound atom, about which we may reason as accurately and justly as we could about the simple atoms themselves. Indeed, I think that x certainly represents one atom only; for oxygen gas being a permanently elastic fluid, must consist of atoms that repel each other. Hence I conceive that a compound atom of oxygen, or a number of atoms of it united together, is impossible. And if x consisted of atoms not united together, I can see no reason why the same number should unite in every case (or a multiple of it) with other bodies.

This reasoning may be applied to hydrogen as well as oxygen. Hydrogen has the property of uniting in different proportions with various bodies, as with carbon, phosphorus, sulphur, &c. In these different proportions we find the hydrogen always denoted by y, 2y, &c. Hence we have every reason to conclude that y, which represents the proportion of hydrogen which unites with the other constituent in these cases is an atom.

The numbers x and y are easily discovered, by making an accurate analysis of the different compounds into which various proportions of oxygen and hydrogen enter, and when reduced to their lowest terms they are very nearly x = 7.5 and y = 1. Hence these numbers represent the ratios of the weight of an atom of oxygen and an atom of hydrogen to each other. Now it deserves attention that these numbers represent the composition of water. For it has been ascertained, I think, with precision, that water is composed of 100 measures of oxygen gas, and 200 measures of hydrogen gas. Now the specific gravity of these gases are as follows:--

Hence water is composed by weight of

From this coincidence we are entitled to conclude that water is formed by the union of an atom of oxygen to an atom of hydrogen. This very important conclusion is supported by other considerations. Oxygen and hydrogen have never been mad to combine in any other proportion than that in which they exist in water. Hence this proportion must be that which unites most readily, and with the greatest force. Now as the atoms of hydrogen repel each other, as is the case also with the atoms of oxygen; and as hydrogen is attracted by oxygen; it is obvious that when they are mixed equably, as is the case when 200 measures of hydrogen gas, and 100 measures of oxygen gas, are put into a tube, and fired by electricity, they will most readily unite atom to atom. This, though not in itself decisive, is a corroborating circumstance. It follows from it that a given bulk of hydrogen gas contains only one-half the number of atoms that exist in the same bulk of oxygen gas.

6. Knowing the weight of an atom of oxygen and of an atom of hydrogen, we have it in our power to determine the weight of an atom of the other substances which unite with oxygen, or with hydrogen, or with both. For example, 100 parts of sulphur unite with two proportions of oxygen, the first consisting of 100, the second of 150 parts, both in weight. Here the proportions of oxygen being to each other as the numbers 1, 1 1/2, or as 2, 3, it is reasonable to suppose that the firs portion represents two atoms of oxygen, and the second three atoms; and that there is another compound, consisting of sulphur united with one atom of oxygen, not yet discovered. If this supposition be reasonable, it follows, that the weight of sulphur which enters into these combinations represents an atom of that substance. Therefore 100 represents an atom of sulphur, and 100 two atoms of oxygen; so that an atom of sulphur, it appears, is just double the weight of an atom of oxygen.

We have it in our power to verify this reasoning, by means of the combinations which sulphur makes with hydrogen. It has been ascertained that 100 measures of hydrogen gas, when they unite with sulphur, do not alter their bulk, but merely their specific gravity. Hence, in order to determine the constituents of sulphureted hydrogen gas with perfect accuracy, we have only to ascertain correctly the specific gravity of hydrogen gas and of sulphureted hydrogen gas. Now
100 cubic inches of hydrogen gas weigh2.230 gr.
100 cubic inches of sulphureted hydrogen35.890
Hence it follows that sulphureted hydrogen gas is composed of
This shows us that if sulphureted hydrogen gas is composed of an atom of hydrogen united to an atom of sulphur (and hardly any other supposition seems admissible), then if an atom of hydrogen weigh 1, an atom of sulphur will weigh 15.09. The combination of sulphur and oxygen gave us 15.12 for the weight of an atom of sulphur. Thus the two processes of reasoning lead to the same conclusion, since the difference between 15.09 and 15.12 is only 3/1000. This is as near a coincidence as it is possible to obtain from chemical experiments, where absolute precision, from the nature of our processes, is impossible.

By a similar mode of reasoning, we may determine with considerable accuracy the weight of an atom of azote, phosphorus, carbon, and the metals. It would be tedious to state the methods here at full length; but some of the most important of them will be given afterwards.

It is hardly necessary to observe, how very powerfully a particular conclusion is confirmed when we arrive at it by different processes. This advantage we have in full perfection when we set about determining the weight of the atoms of the simple substances. In most cases we come to the same conclusion by two, three, or four different methods. These coincidences, I thinks, could not exist, unless the conclusion were well founded.

7. I shall terminate this part of the subject with Mr. Dalton's canons for the combination of the atoms of bodies with each other. They are very ingeniously contrived, and their truth, I conceive, will be readily admitted by every person who pays due attention to the subject:--

If the observations of Gay-Lussac be correct, nitrous gas constitutes an exception to Mr. Dalton's 5th rule. It will come under our examination hereafter.

II. Chemical Canons founded on the above Theory, but deduced from Analysis.

1. When gaseous bodies combine they always unite in determinate proportions; and if we represent the bulk of the gas that enters into the compound in the smallest quantity in bulk by 1, then the bulk of the other constituent is either 1, 2, or 3. Thus muriate of ammonia is composed of 1 muriatic acid + 1 ammonia in bulk; carbonate of ammonia, of 1 carbonic acid + 1 ammonia; nitrous gas, of 1 azote + 1 oxygen; water, of 1 oxygen + 2 hydrogen; gaseous oxide of azote, of 1 oxygen + 2 azote; nitrous acid, of 1 azote + 2 oxygen, or of 1 oxygen + 2 nitrous gas; sulphuric acid, of 1 oxygen + 2 sulphurous acid; carbonic acid, of 1 oxygen + 2 oxide of carbon; ammonia, of 1 azote + 3 hydrogen; nitrous acid gas, of 1 oxygen + 3 nitrous gas. This canon has been established by Gay-Lussac, I think, in a satisfactory manner.[3] The only one of his conclusions which is still doubtful is that nitrous acid is a compound of 2 nitrous gas + 1 oxygen gas. At least I have not been able to make the two gases unite exactly in that proportion. This canon is obviously connected with the Daltonian theory. It is simple and beautiful, and of considerable utility in practical chemistry.

2. The quantity of acid requisite to saturate the different metals is directly as the quantity of oxygen which these metals require to convert them into oxides. Thus 100 parts of mercury require 4.16 parts of oxygen, and 100 parts of silver require 7.9 parts of oxygen, to convert them into oxides. Therefore the quantity of acid necessary to saturate 100 parts of mercury is to the quantity necessary to saturate 100 parts of silver as the number 4.16 to 7.9. This law was first pointed out by Gay-Lussac.[4] It may be expressed in the following manner, which adapts it better to the purposes of the chemist. When different metallic oxides saturate the same weight of acid, each contains exactly the same weight of oxygen. According to Berzelius, in order to saturate 100 parts of muriatic acid, a metal must be combined with 42 parts of oxygen; to saturate 100 of sulphuric acid, it must be combined with 20 parts of oxygen.

I believe that this law applies only to those metals which are precipitated by each other; namely, gold, silver, mercury, copper, lead, cobalt, and perhaps iron, zinc, and one or two others. The other metals, I conceive, follow a different law; and it is because they follow a different law that they are not precipitated. this will appear more obviously hereafter, when we come to examine the constitution of the metallic salts.

3. When sulphur combines with a metal, the proportion remains unchanged, though the sulphur be converted into an acid, and the metal into an oxide. Thus the proportion of metal and sulphur in sulphate of copper is the same as in sulphuret of copper. Hence sulphuret of lead, when treated with nitric acid, is converted into neutral sulphate of lead, sulphuret of antimony into sulphate of antimony, and so on. This law, which is of great importance in practical chemistry, and very much facilitates the analysis of the metalline salts, was first pointed out by Berzelius.

4. The oxygen in a metallic protoxide is equal to half the sulphur in the sulphuret of the same metal, supposing the weight of the metal in both cases 100. This canon was first specified by Berzelius. It depends obviously upon the fact above established, that an atom of sulphur is twice the weight of an atom of oxygen, and holds only in those cases where the protoxide is a compound of one atom of metal and one atom of oxygen; and the sulphuret, of one atom of metal and one atom of sulphur. It may hold also when the oxide contains two atoms of oxygen, and the sulphuret two atoms of sulphur. This is the case with the black oxide of iron and magnetic pyrites. Hence the canon is of some utility, by enabling us the better to determine the constitution of the sulphurets; which, like the oxides, are susceptible of considerable variation.

5. In combinations of two bodies containing each a quantity of oxygen, the weight of oxygen in each body is either equal, or one contains twice, thrice, four times, &c. as far as eight times the quantity of oxygen in the other. This law has been laid down by Berzelius;[5] but I must acknowledge I entertain considerable doubts about its accuracy. It will be better to leave the investigation of the subject till we come to examine the composition of the different salts, in all of which the two constituents contain oxygen. If it hold it will indicate a certain regularity in the relative weights of the atoms of bodies which I have not yet observed.

6. Water is capable of combining both with acids and bases. When it units with an acid it acts the part of a base, and contains the same quantity of oxygen that the base would contain. Therefore the least quantity of water that can combine with sulphuric acid must contain 20 parts of oxygen. Hence the strongest possible sulphuric acid is a compound of
Acid77 1/3
Water22 2/3

When the water combines with a base, it acts the part of an acid, and combines in the same proportion. Hence such compounds are called hydrates. This canon has also been laid down by Berzelius. I must confess that I have not hitherto met with sufficient evidence of its accuracy to induce me to put much confidence in it; but we shall be able to judge better when we come to examine the constitution of the hydrates, than we can at present.

7. In combinations composed of more than two bodies containing oxygen, the oxygen of that constituent which contains the least of it is a common divisor of all the portions of oxygen found in the other bodies. This law, likewise laid down by Berzelius, evidently depends upon the kind of combination which these bases make with oxygen. If they are each combinations of one atom of base with one atom of oxygen, the quantity of oxygen present in all will be the same. If one is a protoxide, and another a deutoxide, then the one will contain double the quantity of oxygen in the other. And since oxygen always units by atoms, it is obvious that all the quantities of oxygen will be divisible by one atom of oxygen. Hence the law.

If we were accurately acquainted with the constitution of the earths, this law would be of great importance to the mineralogist. It would enable him to distinguish between chemical combinations and mechanical mixtures. I have no doubt that it will ultimately throw a new light upon the chemical analysis of minerals; even at present it may be applied with some success, taking the imperfect knowledge that we have as a basis.

8. When two combustible bases unite they always combine in such a proportion that, when oxidized, either the quantity of oxygen uniting with each would be the same, or the oxygen in the one would be twice, thrice, &c. that in the other. This is another law laid down by Berzelius, and obviously depends upon this fact, that the two bodies must either unite atom to atom, or a certain number of atoms of the one must combine with one atom of the other.

This law might be applied successfully to determine which of the metallic alloys are chemical combinations, and which are mechanical mixtures. For example, there can be no doubt that copper and zinc combine chemically. Now from the following table it will appear that the weight of an atom of these metals is as follows:--

Therefore, if they unite atom to atom, brass ought to be a compound of 100 copper and 53.93 zinc. Now if any person will be at the trouble to analyse brass, he will find that this is very nearly the proportion of the ingredients. In like manner bell-metal seems to be a compound of 5 atoms of copper to 1 atom of tin; and the metal for mirrors, of 4 atoms of copper and one atom of tin. If mercury unites atom to atom with tin, it ought to dissolve somewhat less than half its weight of that metal. And if the same law holds with zinc, it ought to dissolve about 1/5 of its weight of that metal.

III. Relative Weight of the Atoms of different Substances determined from Chemical Analysis.

Before we can draw up a table of the relative weights of the atoms of bodies, we must fix upon some one whose atom shall be represented by unity. Mr. Dalton has made choice of hydrogen for that purpose, because it is the lightest of all known bodies. Sir Humphry Davy has followed his example; but he has doubled the weight of an atom of oxygen, and consequently of all other bodies, by the arbitrary supposition that water is composed of two atoms of hydrogen and one of oxygen. Dr. Wollaston and Professor Berzelius have both proposed the atom of oxygen as the most convenient unit: nor can there be any hesitation in embracing their plan. Oxygen is in fact the substance by means of which the weight of the atoms of almost all other bodies is determined. It enters into a much greater number of combinations than any other known body; hence the great advantage attending a convenient number for that body to the practical chemist.

It would remove a great deal of confusion, which is at present very conspicuous in this department of the science, if chemists would agree to represent the weight of the atoms by the same numbers. The following table is submitted to the chemical world as more convenient than the methods hitherto followed; and as the means employed in determining the numbers is every where stated, it were to be wished that they were adopted by chemists in general, as far as they are accurate. If the same numbers were steadily employed by all persons, they would soon be recollected by chemists, who would thus be able to state the composition of every chemical compound without being obliged to refer to a book. The utility of such a recollection to the practical chemist is too obvious to be pointed out.

Weight of an atom.
1. Oxygen1.000
2. Hydrogen0.132
3. Carbon0.751[a]
4. Azote0.878[b]
5. Phosphorus1.320[c]
6. Sulphur2.000
7. Boron

Number of atoms.Weight of a particle.
8. Water, composed of1 o + 1 h1.132[d]
9. Carbonic oxide1 o + 1 c1.751[e]
10. Carbonic acid2 o + 1 c2.751
11. Nitrous gas[f]1 o + 1 a1.878

[1]By binary, Mr. Dalton means a compound of one atom of one body with one atom of another; by ternary, a compound of one atom of one body with two atoms of another; and so on.
[2]Dalton's New System of Chemistry, vol. i, p. 214.
[3]Mem. d'Arcueil, vol. ii. p. 207.
[4]Mem. d'Arcueil, ii. 157.
[5]Ann. de. Chim. lxxxiii, 119.

[a] This is obtained from carbonic acid, which I conceive, with Dalton, to be a compound of 2 atoms of oxygen and 1 atom of carbon. When charcoal is burned in oxygen gas, the bulk of the gas is not altered, it is merely changed into carbonic acid. Hence if from the weight of 100 cubic inches of carbonic acid gas = 46.313 grains, we subtract the weight of 100 cubic inches of oxygen gas = 33.372 grains, the remainder = 12.641 grains gives us the weight of carbon in 100 cubic inches of carbonic acid gas. This shows us that the carbonic acid is composed of 27.29 parts of carbon and 72.71 of oxygen. My analysis of olefiant gas gives 0.748 for the weight of an atom of carbon, which comes very near the preceding.
[b] This number is obtained from nitrous gas, which, I think, Mr. Dalton has successfully shown to be a compound of 1 atom azote and 1 atom oxygen; and Gay-Lussac has proved it to be a compound of equal measures of azotic gas and oxygen gas. The number adopted by Sir. H. Davy to represent an atom of azote when reduced to our proportions would be 0.860, and that of Mr. Dalton 0.714. Ammonia gives us a different ratio if we suppose it, with Dalton, to be a compound of 1 atom of hydrogen and 1 atom of azote. The weight of an atom of azote in that case would be 0.584; but if we consider ammonia as composed of 2 atoms of hydrogen and 1 atom of azote, then the weight of an atom of azote comes out 0.876. Hence I conceive ammonia to be a ternary compound.
[c] This number is deduced from phosphoric acid. Sir H. Davy has ascertained that the quantity of oxygen in phosphoric acid is double of that in phosphorous acid. Hence, I think, it follows, that phosphoric acid is a compound of 1 atom of phosphorus and 2 atoms of oxygen. Farther, it has been ascertained that when phosphorus is burnt in oxygen gas each grain of phosphorus absorbs 4 1/2 cubic inches of oxygen. Therefore phosphoric acid is composed of 100 phosphorus and 151.52 of oxygen, and phosphorous acid of 100 phosphorus and 75.76 of oxygen: and an atom of oxygen is to an atom of phosphorus as 75.76 to 100, or as 1 to 1.320 nearly.
[d] I conceive water to be a binary compound. The opinion advanced by Sir H. Davy, that it is a ternary compound of an atom of oxygen united to 2 atoms of hydrogen, cannot, I think, be supported.
[e] It has been ascertained by accurate experiments that 100 measures of carbonic oxide for complete combustion require 50 measures of oxygen gas, and that the residue is 100 measures of carbonic acid gas. Hence it follows that the quantity of oxygen in carbonic acid is just double what exists in carbonic oxide, while the quantity of carbon in both is the same. Berthollet's objections to this conclusion are inadmissible. He seems, indeed, to have given up the point himself.
[f] The determination of the compounds into which azote and oxygen enter is attended with considerable difficulty. The reason for considering nitrous gas as a binary compound, and its analysis, have been stated before. I consider it as exactly determined. The reasons assigned by Dalton for considering nitrous oxide as a compound of 2 atoms of azote and 1 atom of oxygen, seem to me conclusive: and Gay-Lussac has shown that it contains just twice as much azote as nitrous gas, supposing the oxygen in both the same. Hence its composition must be as stated in the table. The acid analysed by Cavendish and Davy, and called nitric acid, I conceive to have been in fact nitrous acid. What Mr. Dalton calls oxy-nitric acid is true colourless nitric acid, which seems only to exist in combination with water or a base.

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