of the Permanence of Impressions on the "Retina. 44'7 



directed vertically from below upwards, starting from the 

 centre. Let us also consider a slit at the moment when it is 

 equally directed from below upwards. If at this instant the 

 slit is situated before the radius of one of the distorted figures 

 corresponding to the above radius of the regular figure, it is 

 evident that the regular image produced will be seen in the 

 straight position; but if the slit coincides with another radius 

 of the distorted figure, this last will then occupy, in the regular 

 image, the vertical position, and consequently this image will 

 be seen inclined in one direction or the other, or even reversed. 

 Now we shall recall here*, that, unless by a quite peculiar 

 chance, the system of pulleys, however carefully constructed, 

 will never realize in an exact manner the relation which has 

 been assigned to the velocities of the two discs; whence it fol- 

 lows, that, when the slits shall one after the other attain the 

 vertical position from below upwards, they will coincide with 

 radii which occupy, in the partial distorted figures, somewhat 

 different positions, so that the regular images successively pro- 

 duced will not be exactly superposed; but if the apparatus is 

 well made, this displacement of the images will be excessively 

 small, and there will result from it only the continued sensa- 

 tion of a single regular image turning very slowly around the 

 centre, an image which, consequently, will necessarily pass 

 by the straight position. For the rest, this movement of the 

 regular image may be avoided, by substituting for the system 

 of pulleys a system of cog-wheels ; and then if, by attaching 

 the transparent disc upon its axis, it is placed so as to fulfill 

 the condition indicated above, the image will occupy the 

 straight position, and will maintain it invariably. 



In the second place, to observe the image, the spectator 

 must stand at a certain distance from the apparatus, so that 

 the prolongation of the axes of the discs shall pass in the 

 middle of the interval between the two eyes. 



These remarks apply equally to the following case:— 



Ind Case : V^ > V„. -'• 



Here the formula (3.) gives for M a negative value. In 

 order to interpret this change of sign, we resume the reasoning 

 which led us to the formula in question. We start always 

 from a vertical position of the slit and of the radius ;•, and 

 suppose that this slit and this radius proceed afterwards toward 

 the right. Then, when the slit, after describing the angle a, 

 has come before the radius r', the image of the latter will be 

 seen on the right of that of the radius r. But since the trans- 

 parent disc turns more quickly than the black disc, the radius 

 • See the previous note at p. 434 of this number. 



