PURKINJE'S EXPERIMENT. 



511 



FIG. 139. 



shadows always fall on the same parts of the retina, and 

 these parts are thus kept sufficiently more excitable than the 

 rest to make up for the less light reaching them through 

 the vessels. To see the latter we must throw the light into 

 the eye in an unusual direction, not through the pupil but 

 laterally through the sclerotic. If v, Fig. 139, be the 

 section of a retinal vessel, ordi- 

 narily its shadow will fall at 

 some point on the line prolonged 

 through it from the centre of 

 the pupil. If a candle be held 

 opposite Z it illuminates that 

 part of the sclerotic and from 

 there light radiates and illumines 

 the eye. The sensation we refer 

 to light entering the eye in the 

 usual manner through the pupil, 

 and accordingly see the surface 

 we look at as if it were illuminated. The shadow of v is 

 now cast on an unusual spot c, and we see it as if at the point 

 d on the wall, on the prolongation of the line joining the 

 nodal point, Tc, of the eye with c. If the candle be moved 

 so as to illuminate the point V of the sclerotic, the shadow 

 of v will be cast on c' and will accordingly seem on the 

 wall to move from d to d'. It is clear that if we know how 

 far I is from #', how far the wall is from the eye, and how 

 far the nodal point is from the retina (15 mm. or 0.6 inch), 

 and measure the distance on the wall from d to d', we can 

 calculate how far c is from c': and then how far the vessel 

 throwing the shadow must be in front of the retinal parts 

 perceiving it. In this way it is found that the part seeing 

 the shadow, that is the layer on which light acts, is just 

 about as far behind the retinal vessels as the main vascular 

 trunks of the retina are in front of the rod and cone layer. 

 It is, therefore, in that layer that the light initiates those 

 changes which give rise to nervous impulses; which is 

 further made obvious by the fact that the seat of most 

 acute vision is thefovea centralis, where this layer and the 

 cone-fibres diverging from it alone are found (p. 490). 



