203 
previously been observed floating upon water. In this case 
it would not have been noticed except for the fact that we 
happened to be at the proper angle of observation, on a 
suitable morning, to a bright sun in a cloudless sky, and 
thus caught the dispersed light. 
By bringing a dewdrop to an angle of from 40'’ to 42° with 
the sun we get a similar effect as in the rainbow. Upon 
examining a glass globe filled with water, by our ordinary 
gaslight, two points of light may be observed within the 
globe upon holding it away from you with your back to 
the light. These eclipse one another when brought in 
parallel lines. This can only occur when the body is 
between the light and the globe, but may be easily demon- 
.strated when two or three lights are reflected. When the 
globe is moved until the farthest point of light is brought 
into contact with another light which appears fringing the 
circumference on one side, then a perceptible increase or 
flash of dispersed light, prismatically coloured, is visible. 
Take a meridian line from any fixed light across a table ; 
place the observer under the meridian line with his back to 
the light ; take a glass bulb filled with water, and advance 
it to the right or to the left; mark the point when, at the 
conjunction of the two light spots, there is a perceptible 
generation of colour. This will be found to be at an angle 
of 40° to 42° from the observer to the light source. Direct 
an assistant to move the bulb, with a pencil marking its 
route, according to the observer’s instructions, who, with 
his head immovable, keeps the flash before his eyes, and 
the bulb will be found to describe an arc, as in the ac- 
companying diagram. By placing this diagram on the map 
of Lake Windermere we get a miniature reproduction of 
the phenomenon corresponding to what was seen on the 
lake itself With regard to the secondary bow, I assume 
it to be of similar production to the secondary one in the 
rainbow. (See plan, D.) 
