This pattern shows one line which fits the cubic (331) spacing much more closely than it does the nearest hexagonal spacing (2130) indicating the presence of the cubic crystal form. The data on this line will be considered more fully in a later paragraph.

The mean value of the side of the hexagonal prism, taken from nine exposures on various samples, was 4.593 × 10−8 cm.

The data do not afford a very precise direct determination of the axial ratio. The only line giving an independent measurement of the height of the elementary hexagonal prism is the second-order reflexion from the (0001) plane; the small angular deviation of this line gives a relatively low precision in the value of the planer spacing. Independent measurements of the side of the hexagon are given by four reflexions (1010), (1120), (2020) and 2210). The mean valne of the axial ratio, as determined from these reflexions in the nine exposures, was 1633+008.

Table III. gives the complete data on the line designated as (331) in Table II., showing how this fits the two nearest hexagonal spacings as compared with the cubic (331) spacing. The cubic (331) spacing lies nearly midway between the hexagonal (2130) and (2131) spacings; the reflexion from the (2131) plane would hardly be expected to appear on account of its low value of the product N x F. In Table III. the edge of the elementary tetrahedron is computed from the position of the observed line, assuming it to be reflected from each of the three planes in turn, and these values are compared with the mean value of the tetrahedral edge for the given film. In every case the observed line fits the (331) spacing much more closely than either of the nearest hexagonal spacings.

The evidence appears quite conclusive that every sample of silver iodide studied contained the cubic form; in a few cases none but the cubic structure was evident; in some, one or two faint lines were present which belonged only to the hexagonal form; while in most of the samples the hexagonal form predominated. No systematic study was made of the conditions governing the production of one crystal form or the other; the first sample showing the cubic form was prepared by precipitation; the other cubic crystals were produced by fusion of the salt and quenching in cold water; as a rule, the predominantly hexagonal form was produced by every method tried, whether precipitation or the quenching or slow cooling of a fused sample.

The lattice constants of silver chloride and silver bromide were also re-determined. Silver chloride gave the lines of a face-centered cubic structure in which the reflexions from planes of all odd indices were of lower intensity than would be expected from such a structure composed of like atoms; this effect would be expected in a structure of the sodium chloride type where the two kinds of atoms differ considerably

in atomic number. The mean value of d (100) from four exposures was 2.770 × 10-8 cm.

Silver bromide gave the diffraction pattern of a simple cubic lattice. This pattern is accounted for by the sodium chloride type of structure where the two kinds of atoms are of sufficiently similar atomic numbers as to act as

similar diffracting centres. The mean value of d (100) from four films was 2.884 × 10-8 cm.

The deduction of the sodium chloride structure for both silver chloride and silver bromide follows simply from the published methods of analysis of powdered crystal diffraction patterns. The possibility that either of these simple compounds has any other structure seems so remote that it is considered superfluous to publish the data in detail.

Several samples of metallic silver of known high purity were also examined by the powder method; this substance gave photographs having unusually well-defined lines. The pattern was that of a face-centred cubic structure, as has been reported by other investigators *; the mean value of the side of the elementary cube from six exposures was 4.078 × 10-8 cm. Very good diffraction photographs were also obtained from the blackened silver of a developed photographic film.

The lattice constants of the structures studied are summarized in Table IV.

The

In conclusion, it may be pointed out that while the X-ray powder method of crystal analysis has its limitations, as have other methods, it has also its advantages, The great majority of crystalline substances do not ordinarily occur in large crystals, and the powder method is the only method applicable to crystals as they exist in a finely divided state. convenience with which it may be used to examine a large number of samples prepared in various ways has made it possible in the present case to obtain more complete information about the crystalline characteristics of silver iodide than has been obtained by other methods depending upon single large crystals.

Rochester, N.Y.,

May 8, 1923.

*L. Vegard, Phil. Mag., xxxi. p. 83 (1916). L. W. McKeehan, Phys. Rev. xx. p. 424 (1922).

LIV. Studies in the Electron Theory of Chemistry. On the changes in chemical properties produced by the substitution of one element or radicle by another, with applications to benzene substitutions. By Sir J. J. THOMSON, O.M., F.R.S., Master of Trinity College, Cambridge *.

WE

THEN a monovalent atom or radicle-such as F, Cl, Br, I, OH, CH3, CN-is substituted for an atom of hydrogen the number of electrons is increased; these additional electrons, together with the corresponding positive charges, will change the electric field in and around the molecule. It is the object of this paper to investigate the nature of this change and the effect it might be expected to have on the chemical properties of the molecule.

I will begin with a case which I have already discussed to some extent in my lectures on "The Electron in Chemistry," given at the Franklin Institute, Philadelphia (Journal of the Franklin Institute, June 1923), that of the replacement of hydrogen by halogens. As the halogen atom has seven electrons in the outer layer, while the hydrogen atom has only one, there will be six more disposal electrons after the halogen atom has been introduced, the negative charges on these will be balanced by the excess of six units of positive charge which the positive core of the halogen atom has over that of the hydrogen one.

The dispositions of the electrons before and after the substitution are represented in figs. 1 a, b. H, fig. 1a, is the Fig. 1b. b

positive core of the hydrogen atom, d the projection on the plane of the paper of the line joining the two electrons which bind it to the rest of the molecule. Cl, fig. 1 b, is the positive core of the chlorine atom, a, b, c, d the projections of the sides Communicated by the Author. Phil. Mag. S. 6. Vol. 46. No. 273. Sept. 1923.

2 K