Prof. H. W. Dove on a New Photometer. 21 



1 



sin x u 



Assuming the strict validity of a diffusion taking place 



uniformly in all directions, the determination of the brightness 

 of a body which causes this is reduced to the simple measure- 

 ment of an angle. The strict validity, or its limitation, may, 

 however, be ascertained empirically on comparing the values 

 obtained by diminishing the aperture by means of the slider, 

 the inclination remaining the same, with the values obtained by 

 altering the inclination while the aperture remains the same. 

 When an investigation is not made in this manner, the use of 

 the slider is the most reliable. 



If the brightness resulting from the combination of the light 

 emitted by two differently coloured surfaces is to be determined, 

 it is best obtained by means of Fechner's discs : in these, in 

 concentric rings, the magnitude of the sector belonging to one 

 colour gradually increases from 0° to 360°, while that of the 

 other decreases simultaneously from 360° to 0°. The other 

 rings- being deadened off, the horizontal microscope is succes- 

 sively directed towards the individual concentric rings, and the 

 brightness compensated. 



In rotating a Newton's coloured disc, of course the brightness 

 of a white one is not obtained, but only that corresponding to 

 the absorption of all individual colours. It is only necessary to 

 direct the photometer upon the rotating disk, and then ascertain 

 the angle which gives compensation, and subsequently to repeat 

 the same experiment with the white back of the disk, to see that 

 in the first case much light was absorbed. To determine the 

 quantity, a dark sector is gradually enlarged upon a white disc 

 until the brightness of both discs is the same. 



This gives rise to a physiological question, the answer to which 

 is not without interest. If a disc is divided into five white and 

 five dark sectors, the quantity of light which it sends to the 

 eye, while rotating with a definite velocity, is the same as if the 

 disc were divided into twenty-five dark sectors alternating with 

 twenty-five white ones ; but the intervals in the first case last 

 five times as long as in the second. Has this an influence on 

 the determination of the brightness ? It has always appeared to 

 me as if the brightness increased until the velocity of rotation 

 exceeded a certain limit. Possibly the limit of the interval at 

 which no increase takes place is different with different individuals. 



Streaked surfaces reflect more light in the direction of the 

 stripes than in one at right angles to it, because in the latter 

 case there is a partial overshadowing. This is distinctly observed 

 by producing a grating by passing chalk several times over a tile. 

 In these respects the most remarkable differences are obtained 

 on certain plates of mother-of-pearl. 



