Thus it was found necessary to fall back on the image of the arc as reflected from the cornea. This corneal image invariably traced a clear, strong curve on the negative, and would have been appropriated at the outset, were it not that its movements are not, as is well known, a true register of the _amplitude_ of the corresponding eye-movements; a fact that was shown clearly from a comparison in these negatives of the curves produced respectively by the flecks of Chinese white and by the corneal image. The former showed a much greater amplitude of movement. But the corneal reflection is a perfect register of the _time_ and _direction_ of the eye-movements; and in the following tables these features alone are studied. This reflection traced on the film a perfectly readable curve, although in some of the films, owing to a shifting of the carbons in the lamp taking place during the rotation, one of the eyes would be badly illuminated and a good record would be obtained from the other eye alone. The arc ran on an alternating circuit of 60 phases per second, and owing to these interruptions of the illumination the curve of the corneal image showed on the negative as a dotted line in which the distance between any two dots represented one sixtieth of a second. Since the constancy of this alternation in the current has been measured in the Jefferson Physical Laboratory (of Harvard), and found to vary within a few tenths of one per cent only, the spacing of the dots on the negatives formed the most convenient possible means for determining the durations of the nystagmiform movements. These dots are shown in Figs. 3, 4, and 5 (Plates I and II). [Illustration: PLATE I. Fig. 3 (By an error Fig. 4 is shown reversed; the lettering is correct.) Fig. 4]