The subject's head was rigidly held by a rest: this rest was adjusted, and the camera focused, before the rotation. The adjustment of the head was greatly facilitated by fastening a fine black thread to pegs that projected forward from the head-rest, on either side; the thread was stretched horizontally, and at such a height that its image in the camera coincided with the middle of the long (horizontal) axis of the open slit. If then the subject, on seating himself in the chair, had his head so adjusted that each eye was directly behind the thread, each eye would certainly be imaged on the sensitive film. Neither the shadow of this thread on the subject's face, nor its image on the film, interfered in the least with the exposure that was made after rotation. This thread was further found very useful by the subject himself, who, after the rotation and just before the exposure was made, could make sure by sighting on the thread that his eyes had not slightly changed position during the rather protracted rotation. The subject was ordinarily turned twenty-five times at about the rate of one turn in two seconds. The kymograph was set in motion and the exposure commenced as soon as the whirling chair was brought to a dead stop. This stopping always took two or three seconds, at the very time when the nystagmus was most pronounced, so that the photographs do not show the maximum eye-movements. The exposure lasted through one rotation of the drum, nine seconds. In the strongest negatives the movements of the eyes can be fairly well made out from the undulatory curve generated on the film by the dark image of the iris as it oscillated from side to side. But this is true only of the best negatives, and almost never of these if the eyes photographed had the iris blue. In order to obtain better definition in the photographs of the eye-movements, small flecks of Chinese white were tried, as invented and described by Judd.[14] A small square of white was laid with a brush on each cornea, on the side toward the lamp, so that its image on the film should be as bright as possible. The flecks were found to adhere to the eyeball even more perfectly than Judd himself has claimed; and they produced so little discomfort that the subject ordinarily forgot their presence on the eyes. Nevertheless their image as produced on the negatives, although much better than that of the iris, was generally not clearly readable, owing to the brief exposure and the illumination by electric light. This light seems not to be well reflected by the Chinese white: but in all cases where daylight can be employed the use of these flecks must be eminently satisfactory.