Two methods are described for measuring the differences in the time of appearance of the spectrum lines in sparks. The first is a refinement of the Kerr cell method previously used and a detailed description of its use is included. The criticisms of Gaviola of the method are shown not to apply. The second method makes use of a rapidly rotating mirror which either reflects the dispersed light of the spark to a photographic plate or the undispersed light to the slit of a spectrograph. In either case the position of the beginning of a line on the photographic plate gives its time of appearance. The method of Henriot and Hunguenard was used to rotate the mirror. Photographs of sparks are shown with the mirror rotating 1830 r. p. s. However, higher rotational speeds have been used. It is concluded that the Kerr cell method is superior for examining the air lines in the initial stages of the spark; while the rotating mirror method is better in studying the appearance and duration of the metallic lines which are not present in the very first stages of the spark discharge.