Page:EB1911 - Volume 02.djvu/920

 The diagrams show that Dalton formed a very definite conception of the nature of chemical combination; it was the union of a small number of atoms of one kind with a small number of another kind to form a compound atom, or as we now say a “molecule,” this identical process being repeated millions of times to form a perceptible amount of a compound. The conceptions of “element,” “compound” and “mixture” became more precise than they had been hitherto; in an element all the atoms are alike, in a compound all the molecules are alike, in a mixture there are different kinds of molecules. If we accept the hypothesis that each kind of atom has a specific and invariable weight, we can, with the aid of the above theory, make most important inferences concerning the proportions by weight in which substances combine to form compounds. These inferences are often summarized as the laws of constant, multiple and reciprocal proportions.

The law of constant proportions asserts that when two elements unite to form a compound the weights that combine are in an invariable ratio, a ratio that is characteristic of that compound. Thus if Dalton’s diagram for the molecule, or compound atom, of water be correct, it follows that in all samples of water the total number of the hydrogen atoms is equal to that of the oxygen atoms; consequently, the ratio of the weight of oxygen to that of hydrogen in water is the same as the ratio of the weights of an oxygen and a hydrogen atom, and this is invariable. Different samples of water cannot therefore differ ever so little in percentage composition, and the same must be true for every compound as distinguished from a mixture. Apart from the atomic theory there is no obvious reason why this should be so. We give the name bread to a substance containing variable proportions of flour and water. Similarly the substance we call wine is undeniably variable in composition. Why should not the substance we call water also vary more or less? The Aristotelian would find no difficulty in such a variability; it is only the disciple of Dalton to whom it seems impossible. It is evident that we have in this law a definite prediction that can be tested by experiment.

The law of multiple proportions asserts that if two elements form more than one compound, then the weights of the one element which are found combined with unit weight of the other in the different compounds, must be in the ratio of two or more whole numbers. If we compare Dalton’s diagrams of the two oxides of carbon or of the three oxides of nitrogen that are given in the preceding table, we at once see the necessity of this law; for the more complex molecule has to be formed from the simpler one by the addition of one or more whole atoms. In the oxides of carbon the same weight of carbon must be combined with weights of oxygen that are as 1 : 2, and in the oxides of nitrogen a fixed weight of nitrogen must be in union with weights of oxygen that are as 1 : 2 :, which are the same ratios as 2 : 4 : 1. This law has been abundantly verified by experiment; for example, five oxides of nitrogen are known, and independent analyses show that, if we consider the same weight of nitrogen in every case, the weights of oxygen combined with it are to one another as 1 : 2 : 3 : 4 : 5. The discovery of this law is due to Dalton; it is a direct deduction from his atomic theory. Here again, apart from this theory, there is no obvious reason why the composition of different substances should be related in so simple a way. As Dalton said, “The doctrine of definite proportions appears mysterious unless we adopt the atomic hypothesis.” “It appears like the mystical ratios of Kepler which Newton so happily elucidated.” The chemists of Dalton’s time were not unanimous in accepting these laws; indeed C. L. Berthollet (Essai de statique chimique, 1803) expressly controverted them. He maintained that, under varying conditions, two substances could combine in an indefinitely large number of different ratios, that there could in fact be a continuous variation in the combining ratio. This view is clearly inconsistent with the atomic theory, which requires that when the combining ratio of two substances changes it should do so, per saltum, to quite another value.

The law of reciprocal proportions, or, as it might well be named, the law of equivalence, cannot be adequately enunciated in a few words. The following gives a partial statement of it. If we know the weights a and b of two elements that are found in union with unit weight of a third element, then we can predict the composition of the compounds which the first two elements can form with each other; either the weights a and b will combine exactly, or if not, these weights must be multiplied by integers to obtain the composition of a compound. To see how this law follows from Dalton’s theory let us consider his diagrams for the molecules of water, ethylene and the oxides of carbon. In water and in ethylene experiment shows that 8 parts by weight of oxygen and 6 parts of carbon, respectively, are in union with one part of hydrogen; also, if the diagrams are correct, these numbers must be in the ratio of the atomic weights of oxygen and carbon. We can therefore predict that all oxides of carbon will have compositions represented by the ratio of 8m parts of oxygen to 6n parts of carbon, where m and n are whole numbers. This prediction is verified by the result of analysis. Similarly, if we know by experiment the composition of water and of ammonia, we can predict the probable composition of the oxides of nitrogen. Experiment shows that, in water and ammonia, we have, respectively, 8 parts of oxygen and 4·67 parts of nitrogen in union with one part of hydrogen; we can therefore infer that the oxides of nitrogen will all have the composition of 8m parts of oxygen to 4·67n parts of nitrogen. Experiment alone can tell us the values of m and n; all that the theory tells us is that they are whole numbers. In this particular case, n turns out to be 3, and m has in succession the values 1, 2, 3, 4, 5.

It is evident that these laws all follow from the idea that a compound molecule can only alter through the addition or subtraction of one or more complete atoms, together with the idea that all the molecules in a pure substance are alike. Fortunately, the compounds at first examined by the chemists engaged in verifying these laws were comparatively simple, so that the whole numbers referred to above were small. The astonishing variety of ratios in which carbon and hydrogen combine was not at first realized. Otherwise Berthollet’s position would have been a much stronger one, and the atomic theory might have had to wait a long while for acceptance. Even at the present time, it would be too much to say that all the complex organic substances have been proved by analysis to obey these laws; all we can assert is that their composition and properties can be satisfactorily explained on the assumption that they do so.

The above statement does not by any means exhaust the possible predictions that can be made from the atomic theory, but it shows how to test the theory. If chemical compounds can be proved by experiment to obey these laws, then the atomic theory acquires a high degree of probability; if they are contradicted by experiment then the atomic theory must be abandoned, or very much modified. Dalton himself made many analyses with the purpose of establishing his views, but his skill as an analyst was not very great. It is in the work of the great Swedish chemist J. J. Berzelius, and somewhat later, in the experiments of the Belgian chemist J. S. Stas, that we find the most brilliant and vigorous verification of these laws, and therefore of the atomic theory.

We shall now give an outline of the experimental evidence for the truth of these laws.

The law of the conservation of matter, an important element in the atomic theory, has been roughly verified by innumerable analyses, in which, a given weight of a substance having been taken, each ingredient in it is isolated and its weight separately determined; the total weight of the ingredients is always found to be very nearly equal to the weight of the original substance. But on account of experimental errors in weighing and measuring, and through loss of material in the transfer of substances from one vessel to another, such analyses are rarely trustworthy to more than one part in about 500; so that small changes in weight consequent on the chemical change could not with certainty be proved or