564 THE SENSES Ai\D SENSORY 



Schmidt), the crystalline cones are drawn out into long fine 

 threads, often curved more or less in their long axes. Schmidt 

 1 220], and myself independently [219], made experiments on 

 glass rods, models on a large scale of the cones of the Phroni- 

 niid.'u. Schmidt found that when the convex conical end of 

 the rod is illuminated the light is emitted in a divergent pencil 

 from the other end of the glass thread. In my experiments 

 I used rods 5^5- of an inch in diameter and an inch in length : 

 these were enclosed in a capillary glass tube filled with 

 water, and whether curved or straight they transmitted light, 

 although so little passed through the lumen of the tube that 

 it appeared black when seen with a half-inch objective. The 

 capillary tube was placed in the axis of the microscope so 

 that the end nearest the object-glass \vas level with the stage; 

 the lower end was illuminated directly by the lamp. When 

 the focus of the microscope was altered, so that the ends of 

 the rods were beyond it, the discs of light enlarged, showing 

 that the emitted light was divergent. 



The light which enters a cylinder with divergent rays after 

 any number of total reflections, emerges from it with divergent 

 rays and the angle of divergence remains the same. When 

 the cylinder is also a refractive cylinder it will have a focus 

 like a lens, and this focus will be a recurring function of its 

 length (see p. 572) ; so that whether the thread conducting 

 the light-rays is a simple cylinder, with a lens situated near or 

 at its extremity, or whether it is a refractive cylinder, in 

 Exner's sense, a small image is capable of being formed 

 beyond its emergent surface. 



The conditions necessary for the formation of such an image 

 are that the conducting-rod shall be a cylinder, and not a cone, 

 as total reflection from the inner surface of a cone produces 

 divergent pencils at its apex, and the internal surface of the 

 rod must be capable of reflecting the incident light. 



I have found that small glass threads do not possess this 

 power when the outer surface is blackened with soot, or even 

 with black varnish. The blackened surface absorbs and does not 

 reflect incident pencils; even a narrow ring of pigment prevents 



