THE results of the investigations by Schlesinger, Perrine, and Albrecht are summarized. In this investigation photographs of artificial stars were made with a precision camera reducing in the ratio of I to 20, on plates one inch wide pressed firmly against a rigid metal frame. In agreement with previous results, no evidence of systematic errors in distance was found, indicating that, in general, distortions are of purely local origin and extent. While the average measurements of ten exposures on a triple-coated plate are practically the same as for ten exposures on a thin emulsion, the local distortions are greater on the thick film. Yet the greatest variation from the mean was only o.oo44 ram. Faint and normal exposures gave the same average measurements within o.ooo 3 mm. Examination of the images showed that the effects of local variations in distribution of the silver bromide grains and of variations in sensitiveness and in development are negligible. Intensification modifies the minor details of image structure, but does not remove the larger irregularities in the original images nor change the centre of gravity appreciably, l~Ieasurements of the same plates, when wet and dry, agree accurately and indicate that, on drying, the images move perpendicular to the surface. However, this is not the case for images within a few millimetres of the edge, as these move outward on drying with alcohol or by hardening with formalin. It is concluded that the principal factor in producing the slight image displacements observed is local non-homogeneity of the gelatin. The probable error in accuracy of measurements of the position of stellar images is found to be considerably less in the case of plates with a fine grain thin emulsion (astronomical) than for those with thicker emulsions (Seed 3 Β° and triple-coated). Intensification of weak * Communicated by the Director. ~Communication No.