discuss it, but on the understanding that no settlement must yet be attempted, no doctrine laid down. He favours a view which deduces all the phenomena of nature from matter and local motion'-a philosophy 'which, because it explicates things by corpuscles, or minute bodies, may not very unfitly be called corpuscular.' But he will not pronounce on the philosophy of a fullblown atomic theory, or commit himself to the doctrine that it is more than a useful working hypothesis with other possible alternatives. Finally he rejects the philosophic materialism with which the ancient atomists connected their theory. To explain the mechanism of the construction of matter does not lay bare the mysteries of the Universe.

Descartes (1596-1650) held that matter was unlimited in extension, infinitely divisible and continuous, the possibility of motion being secured by the celebrated hypothesis of 'vortices,' whereby all movement was conceived to take place in a closed ring, matter simultaneously moving up to fill the room of matter displaced. The discoveries of Newton (1642-1727) destroyed the Cartesian theory of physics, swept away the planetsustaining vortices,' and pointed to the conclusion that the space in which the planets moved without resistance under the action of gravity must be void of matter. With void in space as a necessity for planetary motion, the discontinuous view of matter returned as a complementary proposition. Newton held that differences in density were due to differences in the closeness with which the atomic particles were packed, and believed that cohesion and chemical affinity were manifestations of the forces whereby the atoms attracted each other across the intervening void. Both Boyle and Newton looked to vibrations of the atoms as an ultimate explanation of the nature of heat, a view which thus became interwoven with the atomic theory. Towards the end of the 18th century Newton's views had prevailed; and Voltaire, as the spokesman of the French Encyclopædists, could write: 'the plenum is to-day considered a chimera . . . void is recognised; bodies the most hard are looked upon as full of holes like sieves, and, in fact, this is what they are. Atoms are accepted, indivisible and unchangeable.'

It will thus be seen that in 1801, when John Dalton began his epoch-making researches on the phenomena of chemical combination, the revived atomic theory was ready to his hand. Its possibilities had been considered and many of its probable effects had been pointed out. There is some uncertainty whether the theory was brought to Dalton's mind by the experimental facts, or whether he used the pre-existing theory to deduce consequences, and verified those consequences by experiment. He was working on two different gaseous compounds of carbon and hydrogen, and found that, for the same contents of carbon, one gas contained just twice as much hydrogen as the other. This fact is explained at once by, and indeed naturally suggests, the supposition that changes in composition can only take place by definite steps, and that one atom of carbon is linked to one atom of hydrogen in the first case and to two atoms of hydrogen in the second. Moreover, it follows that the relative weights, as determined by analysis, of the carbon and the hydrogen in these compounds have a simple connexion with the relative weights of the respective atoms of carbon and hydrogen. Since one part by weight of hydrogen is linked to six parts of carbon in one compound, while two parts of hydrogen go to six of carbon in the other, we see that on these data the simplest hypothesis, giving the hydrogen atom the value of unity or one, is that carbon must possess an atomic weight of six. Similarly, from other series of compounds, Dalton assigned relative atomic weights to oxygen, nitrogen, and, in all, to some twenty different elements.

*

Thus Dalton's great achievement was the introduction of quantitative measurement into the atomic theory. Weight is perhaps the most striking and fundamental property of matter, and such ascertained differences of weight led to the sense of some definite material reality underlying the phenomena. The theory thereupon ceased to be merely a philosophical speculation, and became a clear-cut chemical and physical conception, capable of co-ordinating experimental facts, and subject to quantitative and exact examination. The weak point

* Atomic weight may be defined as the relative weight of an atom, taking the weight of the hydrogen atom as unity.

of the theory, as left by Dalton, is seen when we ask why it is assumed that the simplest known compound of two elements A and B, a compound to which the constitution AB is assigned, is in reality the simplest possible? It may well be that the compound which is the simplest of those known at a given time has two atoms of one or other constituent, and should be expressed as A,B or AB2. This difficulty could not be solved by a consideration of Dalton's combining weights alone; a new series of experiments was needed.

Now the composition of water by volume had been determined in 1781 by Cavendish, who, with characteristic accuracy, got results giving 201.5 volumes of hydrogen to 100 volumes of oxygen. It was not till 1805, when Gay-Lussac and Humboldt undertook a long series of careful and exact experiments, that this result was improved upon. With an accuracy reducing the probable error to about one part in a thousand, they found that the volume ratio of the component elements of water was two to one. Struck by the simplicity of this result, Gay-Lussac examined other cases of gaseous combination, and found similar simple ratios. In all cases, when gases combined, if unity represented the volume of one gas, the volume of the other was found to be equal to, or twice, or at most three times as large as, that of the first. Of course, no such relations are found to exist in liquids and solids. It is only when matter is widely extended in the free molecular state, as it is in gases, that these simple phenomena appear.

Gay-Lussac's experiments, interpreted in the light of Dalton's atomic theory, suggest at once that equal volumes of different gases must contain numbers of atoms which bear simple ratios to each other, and probably are equal. This result was grasped by Berzelius, but Dalton pointed out a difficulty. When combination between two gases A and B occurs, the product AB of the reaction is often found to occupy a volume equal to that of the mixed reagents. The two volumes, on uniting to form the new compound, frequently occupy the space previously taken up by both of them together. According to Gay-Lussac and Berzelius, this result means that the number of particles in AB is equal to the sum of those in A and B separately. If Dalton's idea that all three

substances A, B and AB consist of indivisible atoms be correct, there can be of AB atoms but half the number of A and B atoms together, since one A and one B unite to form but one AB.

The difficulty was met and overcome in 1811 by Avogadro, who pointed out that everything was explained if we accepted a middle term or intermediate condition, and supposed the elementary particles or molecules of A and B to consist each of two atoms, which separated from each other before combining with the opposite kind to form an AB molecule. The reaction was thus represented by the equation A2 + B2 = 2 AB, and the known volume relations were satisfied.

This presentation of the subject introduced for the first time a clear distinction between the physical molecule of a substance, whether simple or compound, and the chemical atom. The molecule became the smallest particle which could exist in the free state as gas, and might be composed of one, two, or more atoms; while the atom became the smallest particle which could enter into chemical combination. Avogadro's hypothesis is best formulated by saying that equal volumes of all gases, elementary or compound, at the same temperature and pressure, contain equal numbers of molecules. By an extension of the means here indicated it has been found possible to assign relative atomic weights to all the elements yet discovered, calculated from the startingpoint of the lightest known element, hydrogen, taken as unity. On this basis has been built all the marvellous superstructure of molecular chemistry. The complicated structure of organic substances is successfully represented by picturing atoms linked to each other in open chains or closed rings, while the wonders of isomerism, where two different substances have the same percentage composition measured by chemical analysis, are well explained by imagining differences in the arrangement of similar atoms to form different complex molecules.

But the particular line of research we have now to follow lies in physical rather than in chemical fields. We have already said that the history of the atomic hypothesis is closely connected with that of the theory which regards heat as the vibratory energy of the ultimate particles of Vol. 219.-No. 436.

I

bodies; and, when the experiments of Joule had made it clear that heat-energy and mechanical work were mutually convertible and equivalent, that theory was developed in a mathematical form for the simple case of a gas. A gas was pictured as consisting of a number of molecules flying through void with great velocity, colliding with each other and with the walls of the containing vessel, and undergoing perpetual changes in speed and direction as the result of these collisions. The outward pressure of a gas is due to the impact of the molecules on the walls; and the temperature of the gas is a measure of the mean energy of movement of the molecules. On this theory it has been shown that the pressure of a gas is directly proportional to the number of molecules in unit volume; that is, it is inversely proportional to the volume in which the unit mass of the gas is confined-the smaller the confining space, the greater the pressure exercised by the gas. It has also been shown that the pressure of all gases will rise at the same rate in proportion to the temperature; and that all gases at equal temperatures and pressures contain the same number of molecules. Thus the experimental properties of a gas, as well as Avogadro's hypothesis, are explained by this conception of gaseous movement, which is called the kinetic molecular theory. It is possible, too, to calculate the speed of movement necessary to give the observed pressures, and to show that the average velocity of a hydrogen molecule, as it whisks about in the gas at the temperature of freezing water, is about 2000 yards a second, while the heavier oxygen molecule moves at one quarter of that rate.

It

Hitherto nothing has been said about the sizes of atoms and molecules or the distance between them. is clear that in a gas the molecules must be much further apart than in a liquid or solid. The fact that one volume of water gives some 1600 volumes of steam shows that, since space is three-dimensional, in steam on the average the molecules must be 1600, that is, the cube root of 1600, or about twelve times more distant from each other in a gas than in water. But how distant are they? What is the fineness or coarseness of molecular structure?

It has proved a simple enough problem to find an upper limit. It is easy to say that the molecular