318 Messrs. H. E. Ives and E. F. Kingsbury on the 



If there are no mechanical irregularities whatever, the speed 

 necessary to eliminate flicker on the alternation of the two 

 surfaces by the rotation of the disk is zero. Suppose, now, 

 the reflecting-power of one of the 90-degree sectors is reduced 

 by ten per cent. From equation (15) we obtain at once 

 that the critical speed at what was the no-flicker point, due 

 to a 90-degree sector of 9/10 illumination, is raised to about 

 12 cycles per second (using the value *001 for 8). Now 

 comes an interesting point. The computation as made 

 applies to a disk similar to those for which all the data were 

 obtained, namely in which the 90-degree sector is actually 

 two 45-degree sectors. In other words, our speed counter 

 being calibrated to read twice the actual number of rotations, 

 it follows that the speed called for by the above computation 

 must be doubled to make flicker disappear, or 24 cycles are 

 •called for. (For smaller values of S this speed is considerably 

 higher.) Similarly, any single spot or other irregularity 

 exerts a two-, four-, or more-told disturbing effect according 

 as the rotating disk is composed of two, four, or more 

 sectors, for its period is reduced in that proportion with 

 respect to the period of alternation of the illuminations. 

 This multiplying effect would, of course, be reduced to a 

 minimum by the use of a single sector in place of the two 

 as here used, an arrangement more difficult mechanically. 



The computation just made applies to the equality point 

 ns determined before the one sector is darkened. The com- 

 putation of intermediate positions, where the field presents 

 three different degrees of brightness in succession, is not 

 capable of treatment by the theory in its present state of 

 development. The theory has, however, served its purpose 

 in showing how the most striking effect, namely, the very 

 great increase of critical speed at the mid position, is brought 

 about by mechanical defects. The facts found by experi- 

 ment are thus in accord with the theory, although here, as 

 has happened before, the actual phenomena are of that- 

 exaggerated magnitude which call for surprisingly low values 

 of the Fechner fraction. 



Escape from the effects of irregularities of this sort lies 

 primarily, of course, in the mechanical symmetry of the 

 apparatus. But no matter how perfect the flicker photo- 

 meter, it is evident that a strict substitution method of 

 measurement should be adhered to. A mechanically im- 

 perfect flicker photometer has its own individual setting 

 for lights of the same colour, which may be considerably 

 different from that of the same or another instrument used 

 on the same photometer track as an equality instrument. 



