After an experimental study of the various factors governing the performance of the Bucky diaphragm grid, there appeared to be but one feature of design where appreciable improvement could be made on existing models, and that was in regard to the distance between the patient and the film. It is a well-known principle in radiographic technique that the part radiographed should be as close as possible to the film, in order to obtain the sharpest definition and detail and to minimize distortion. For instance, in the radiography of the frontal sinuses, the forehead rather than the back of the head is placed in contact with the film holder or cassette, and in the radiography of the lumbar spine, the knees are raised in order to lower the arch of the spine and bring it closer to the film. The size of the focal spot and the “undercutting” of scattered radiation both act to diffuse the outlines of portions of the subject which are distant from the film. Undoubtedly one of the reasons why Dr. Potter succeeded where others had failed in devising a practical form of Bucky diaphragm was that he brought the distance between the patient and the film to the point where it gave improvement in definition as well as in contrast. This factor seems often to be neglected or ignored in the construction of Bucky diaphragms as well as in various types of plate changers. In the models of Bucky diaphragm that the writer has had opportunity to examine, the distance between the cover of the diaphragm and the film ranges from 7/8 inch to 1¼ inches. As a rule, the thickness of the grid does not account for more than half the distance between the patient and the film, so that there is opportunity for reducing this distance by reducing both the thickness of the grid and the free space in which it travels. This requirement for the best definition conflicts somewhat with the conditions necessary for high contrast. The highest contrast is secured by making the slits between the lead strips as narrow as possible in proportion to their depth. Since there is a limit as to how close together it is practical to mount the lead strips, the easiest way to secure a small ratio of slit width to slit depth is to increase the depth of the slit; but the experiments showed that to do this, keeping the slit width constant, diminished the definition. The problem in grid design, therefore, is to make the grid as thin and as fine meshed as possible in order to obtain both good definition and high contrast. To secure the same excellence of definition over the whole radiograph, the cassette should be curved to fit the curvature of the grid. The use of the curved cassette with the Potter-Bucky diaphragm has been suggested by Dr. H. W. Van Allen (Amer. Jour. Roentgenol., 1921, viii, 340). In attempting to put into practice the conclusions of this study, two experimental Bucky grids have been constructed.