for obvious reasons called them canal rays.. Later, however, Sir J. J. Thomson, who has made an exhaustive study of some of their properties, changed their name to that of positive rays.

Prof. Thomson's investigations have been chiefly along lines which have led to the use of these rays as a means of analysing the contents of the vessel in which they are produced. The possibility of making such an analysis with any degree of accuracy was greatly helped by the property these rays have of affecting a photographic plate on which they fall. If an ordinary sensitive plate is placed inside the apparatus and the rays are allowed to fall on it for a very short time, on development a sharp image of the spot where the rays hit is obtained. If the rays can enter the observation chamber only

M

1

Figure 3.

2

through a narrow tunnel, such as shown in fig. 3, the spot obtained on a plate placed at cd (fig. 3) is small and circular. Suppose now that our stream of particles pass between two electrically charged plates P, and P, (figure 3) before they hit the photographic plate. On account of the charges borne by the particles they will be deflected from their original path. If each particle carries only one positive charge, or, in other words, if each has lost only one electron (and for the present we shall assume that to be the case), the deflecting force acting on each will be the same. It follows, therefore, that the lighter particles will be deflected more than the heavier, the slower more than the faster. In this way a sifting process takes place, and if, under these conditions, the stream is allowed to fall on the photographic plate, we find a narrow band (see figure 4) instead of a circular spot. The part nearer the origi

nal spot is due to the action of the heavier and the faster particles, the more remote portion of the band to the lighter as well as the slower rays. Suppose, now, that in addition to the electric field, placed so as to deflect the rays horizontally, a magnet MM (figure 3) is placed so as to deflect the particles vertically. Then if both fields act simultaneously, the sifting or sorting process now results in the presence on the photographic plate of curved lines, each line corresponding to a particular kind of particle. To be more accurate, if, as we have assumed above, each particle bears one unit charge, all particles of the same mass, no matter with what speed they move, give rise to a parabolic curve. Further, it can be shown that the atomic weight of each kind of particle may be found directly and simply by measuring the vertical deflection of any portion of the corresponding curve. Reproductions of actual photographs obtained by Prof. Thomson are given in figures 5, 6 and 7. In figure 6 the lines marked a and b correspond respectively to the atom and the molecule of hydrogen.

We have, therefore, a new, simple and direct method of chemical analysis, the advantages of which are apparent, (1) Not only is the presence of an element indicated, as is the case in spectrum analysis, but at the same time the value of its atomic weight is given. Two good examples of the advantage of this may be given. From an analysis of the gases obtained from the residues of liquid air, Prof. Thomson found on his photographs a curve corresponding to an atomic weight 22. Now no known element has this atomic weight. Was it then a new element? In seeking an answer to this question, Prof. Thomson examined several specimens of neon, a rare gas of generally accepted atomic weight 20. In all cases a line 22 occurred on his photographs, in addition to a line 20. He has concluded, therefore, that "there can be little doubt that what has been called neon is not a simple gas but a mixture of two gases, one of which has an atomic weight about 20 and the other about 22."

Again, during the last year or two Prof. Thomson has been investigating the properties of a substance which gave rise to a photographic line corresponding to an atomic weight 3. Now, again, no known element has this atomic weight. Was this too a new element? For over a year this point has been under investigation in the Cavendish laboratory. Ac

cording to the latest available information, the curve is considered to arise from a tri-atomic form of hydrogen (H1).

(2) The method is extremely sensitive, even more so than is the case with the spectroscope. With one-hundredth of a milligram (less than one-six thousandth part of a grain) of a substance in the tube, and with an exposure of less than onemillionth of a second, not only is the corresponding curve found on the plate, but at the same time its atomic weight may be obtained with an accuracy of less than one per cent.

(3) Because of the sensitiveness of the method, substances existing in the vessel only for a very short time leave their record behind them. The importance of this is at once evident.

(4) Impurities, so bothersome in ordinary analysis, have few, if any, harmful effects. They simply give rise to additional curves on the plates.

MULTIPLE CHARGES.

We have assumed that each positively charged ray carries only one unit charge, or, in other words, that it has lost only one electron. But there is nothing in the picture of the atom given above to indicate that this should be the case. Cannot an atom sometimes lose more than one electron and thus bear two, three, four and even more charges of positive electricity? That this is actually the case has also been shown by Prof. Thomson. A careful analysis of all the lines which appear on his photographs has shown that all curves which begin at the line AA' (figure 5) correspond to particles which have lost only one electron, whereas the prolongation of a line extending beyond this distance (see lines marked C, C', figure 5) is due to particles which have lost more than one electron.

To make clear why this is so, a brief consideration of one or two other points is necessary. The particles giving rise to the initial (that is, least deflected) portions of all curves are those which possess the maximum amount of energy, or in other words, those which have acquired their speed by travelling the whole length of the electric field between the electrodes of the discharge chamber. If now we assume that atoms may lose two electrons and so become doubly positively charged, the fastest of these will enter the observation chamber with just twice the amount of energy of the corresponding singly

charged particles. What now will be the deflection of such doubly charged particles? When it is remembered that, although they have a double amount of energy, on account of their double charge the electric deflecting force due to the plates P, P, (figure 3) is also doubled, it will be seen there will be no change in this deflection. Whence, then, comes the prolongation of some of the curves? This is the result of what is known as re-combination, a phenomenon always occurring in ionized gases. Positively charged particles sometimes "pick up" a free negative electron, thus forming either a neutral body or one whose positive charge has been lessened by one unit. That this is going on in positive ray tubes is evident from the fact that there is always an undeflected central spot. (See figures 5, 6, 7.) This spot indicates the presence of rays originally positively charged, which have been neutralized as a result of recombination. Since then we have neutral rays, it is reasonable to suppose that we may have in the observation chamber rays which, while primarily doubly charged, have picked up an electron and possess only one unit charge. Such rays pass through the narrow tunnel with a double supply of energy, but, on account of their ultimate single charge, are subject to a deflecting force by plates P, and P,, of only the normal amount. Their deflection therefore is only half the normal amount and we have a prolongation of the curve corresponding to such particles. In a similar manner the cases of particles which have lost various numbers of electrons and regained some may be worked out.

A good illustration of the evidence of multiple charges is given in figure 7. In this photograph the prolongation of line I corresponds to atoms of the element argon which originally lost three electrons, and later regained two; line III, to those which lost three and remained in this condition. The prolongation in line II is caused by atoms which ,originally minus three electrons, have picked up one electron before entering the deflecting field. It will be noticed that in this case the prolongation extends to a point two-thirds of the normal distance from the vertical.

This method of analysis, therefore, gives us information concerning the physical nature of the substances investigated. A concrete example will perhaps make clear what is meant. A vessel containing a small quantity of oxygen has been shown

to contain (a) neutral atoms; (b) neutral molecules (a union of two atoms); (c) atoms with one positive charge; (d) atoms with one negative charge; (e) atoms with two positive charges; (f) molecules with one positive charge; (g) molecules of ozone (a union of three atoms) with one positive charge; and (h) a group of six atoms (0.) with one positive charge.

As a final illustration of the use which can be made of these positive ray photographs, the following table, giving the complete analysis of all the substances corresponding to the curves shown in figure 5, is appended:

The presence of ions has been demonstrated by recent experiments performed by Mr. C. T. R. Wilson, of the Cavendish Laboratory. Ions, it must be remembered, are much too small to be seen even with the aid of the best ultramicroscope. By making each ion become the centre of a small drop of water, however, Mr. Wilson has succeeded in visualizing their existence, and photographing the appearance of an ionized gas under various conditions. The method employed is based on the principle that when air (or other gas) suddenly expands to fill a larger volume, it becomes cooled. If before the expansion the gas has been saturated with water vapor, after the expansion, on account of the cooling there is more than enough vapor to saturate the gas and some of it condenses.