344 On the Photography of Ripples. 



place of uniform depth. The Lloyd bands due to reflexion are 

 also shown. 



Fig. 7. Rectilinear ripples are refracted at a semicircular 

 boundary, and appear approximately circular afterwards. The 

 analogue of optical opacity is so strong that the motion is 

 quenched before the ripples come to a focus. The squares 

 with dark shaded edges represent two slips of cover-glass put 

 into the water to make it so shallow as to refuse to allow of 

 diffraction through these parts. The diffraction effects, if 

 not prevented, would completely mask the refracted ripples as 

 the latter would be so much the weaker. 



By using a shallow region shaped like the section of a double- 

 convex lens, in combination with a circular shallow region, a 

 real image has been obtained ; but the real foci have never 

 been obtained with sufficient success to warrant the repro- 

 duction of the photographs. 



By using the analogue of a double-concave lens the phe- 

 nomenon of the virtual image has been obtained. While the 

 effects of refraction at both concave and convex boundaries 

 have been obtained on the s:;me plate by using a piece of 

 glass with a sinuous contour, 



Fig. 8. In this case the source was placed by trial at such 

 a distance as to produce refracted ripples which are nearly 

 rectilinear. Diffraction was again warded off by thin glass 

 at the sides of the semicircle. Before producing this photo- 

 graph cases in which the wave-front indicated the presence 

 of a virtual image of the source were obtained. 



Conclusion. 



The method used in the second and third papers is easy to 

 a l J ply> an d if the experiments had to be repeated no funda- 

 mental alteration would 16 adopted. For the production of 

 photographs illustrating refraction the method described in 

 this paper suffers from many practical disadvantages. The 

 use of liquids other than water, such as carbon tetrachloride 

 and carbon bisulphide, might be tried. But the best method 

 would perhaps be to abandon the use of small waves altogether, 

 to be content with using non-parallel light, and to carry out 

 the experiments on a larger scale. 



1 wish to thank Mr. Boys and Prof. J. J. Thomson for much 

 help in the whole of the work. 



Cavendish Laboratory, 

 Cambridge. 



