ELECTRIC-SPARK PHOTOGRAPHS. 175 



follow. Consider a shell of compressed air through which rays of light 

 from a point are made to traverse. These rays travel in straight lines, 

 except where they meet a medium of different density, and the denser 

 this is and the more nearly they meet this at a grazing incidence the 

 more tliey will be bent toward the perpendicular. In comparison with 

 water or glass a layer of compressed air has very little refractive power, 

 and so rays which strike the shell anywhere except at the extreme edge 

 are practically nninfluenced in their course and strike the plate prac- 

 tically in the same place that they would do if the shell of compressed air 

 had not been traversed. But those rays ab, ah, Fl. v, fig. 2, which strike 

 the shell of air almost tangentially, are bent inwards slightly at h and 

 again at c, having traversed what is equivalent to a wide angle prism, 

 and strike the plate at e, leaving the place d, where they would have 

 gone had they not been refracted, dark ; moreover, at e they meet other 

 rays which have been hardly at all refracted since they have passed 

 actually into the shell and out again, and therefore e is doubly illumi- 

 nated. The consequence is that a wave or shell of comi)ressed air gives 

 rise to an image on the plate, iu which there is a dark line and a light 

 line within it. Similarly a wave of rarefaction must produce a light 

 line with a dark one within it.* 



An examination of the photograph (PI. iv, fig. 2) will make it evident 

 that not only is the head wave a wave of compression, but the wave 

 which starts from the end of a kind of vena contracta behind the bullet 

 is also a wave of compression. It is a curious fact which requires ex- 

 planation that the head and tail waves are not parallel to one another, 

 and they do not show any sign that they would become parallel if they 

 were continued indefinitely. This can only be due to either the tail of 

 the bullet travelling considerably faster than the head, or to the actual 

 velocity of propagation of the tail wave being less than that of the head 

 wave The effect observed is true and is not optical, being neither due 

 to the refractive effect of the outer shell disturbing rays which are 

 tangential to the inner shell, nor to an effect of perspective, for though 

 the projection of a cone from a point upon a plane is only seen of the 

 l^roper angle when the perpendicular, dropped from the point upon 

 the plane, passes through the vertex of the cone, yet when, as iu the 

 case of PI. VII, where it passes within both cones, and more within the 

 outer one than the inner one, the effect is to make the projections of 

 both of a greater obtuseness, and of the outer one to a greater extent 

 than tlie inner one; nevertheless an examination of the amount of this 

 effect of perspective made by Mr. Barton showed that the distortion 



* It may be worth while to point out that the dark and light lines are — aud onsht 

 to be — parallel to one another as soon as they are so far away from the shadow of 

 the bullet as to be practically straight lines. For if the thickness of the shell is 

 divided up into a series of elements, the ray passing through any one of these will 

 meet with a refractive medium which is less effective, as the diameter of the part of 

 the shell considered is greater, while the refractive angles of the elementary prisms 

 become inclined more, so as to compensate for the diminished density 



