At< X-ray crystal analysis of certain alloys by the powder method 1 has shown that errors due to the finite dimensions of the samptes employed become appreciable when maximum accuracy is attempted. Since the effect of varying the proportion of the metals forming an alloy, or of varying its properties by any other means, is usually to change slightly the dimensions of the crystal lattice, without changing its kind, the estimation of the magnitude of such errors, and efforts to reduce them, seemed worth while.
The apparatus available was designed and constructed by the staff of the Research Laboratory of the General Electric Company, and permits simultaneous analysis of as many as fifteen samples. The spatial relations of the parts essential for the analysis of one sample are indicated in Fig. I, which is a projection on the (vertical) plane through the centre, M, of the molybdenum target from which the X-rays diverge, the centre, N, of the (horizontal) cylindrical sample, and the centre-line, ST, of the photographic film which is bent over a cylindrical frame coaxical with the sample. The diameter of the sample and the thickness of the cylindrical tube necessary for non-coherent powders or materials which are modified by exposure to the atmosphere, have, for convenience in reference, been enormously exaggerated. Points in Fig. I which are projections of points lying outside the plane of projection are primed. When attention is to be called to the position of these points in space the primes will be suppressed.
The fundamental equation of the method is nX 2d sin 0 * Communicated by Col. John J. Carry, D. Eng.. Associate Editor.