202 



TRANSACTIONS OF THE 



[mar. 1,. 



to give the least violent flicker with the colored disk was chosen, 

 and this flicker used as a third standard for comparison in the 

 second selection of small sjray disks giving a less violent flicker 

 with the colored disk. This method of successive selection of 

 gray disks giving less violent flickers with the colored disk than 

 preceding gra^^ disks have given was continued till all but a 

 single large and small gray disk had been excluded. The lumi- 

 nosities of these two disks were then measured and the average 

 of the two luminosities taken as the luminosity of the colored 

 disk. 



The comparison of the flicker given by a small gray disk com- 

 bined pair-wise with the colored disk, with that given by a large 

 gray disk similarly combined was made in the following way. 

 The small gray disk was placed concentrically on the large gray 

 disk as in Fig. II. A. and the two combined pair-wise with the 



Fig. II. 



colored disk as in Fig. II. B. This combination could be placed' 

 on the axis of the rotation apparatus, rotated, and the flickers 

 inside and outside compared. The direct comparison of the 

 relative violence of the two flickers was found very difficult to 

 make. It was observed, however, that the flicker which required 

 the highest rate of rotation to extinguish it was always, as near 

 as could be judged, the most violent. In comparing the relative 

 violence of the inside and outside flickers produced by such a 

 combination as is shown in Fig. II. B. the rate of rotation of the 

 disks was increased till all flickering disappeared. The speed 

 was then gradually diminished and the disk was observed care- 

 fully so as to detect whether the flickering began first on the in- 

 side or outside. The flicker which appeared first was the most 

 violent. The flickering was observed through a rectangular 

 opening cut in black cardboard, the cardboard being held in. 



