8 Lord Rayleigh on Reflexion from Liquid Surfaces 



The necessary angle is thus independent of the dispersive 

 power, and does not vary rapidly with the refractive power, 

 of the substance. For water, i=9/2d in circular measure, or 

 about 22°. For glass (p=l:5) we should have 2 = 18°. 



An attempt was made to achromatize the band from bisul- 

 phide of carbon with a 15° prism of that material. So far as 

 could be judged the colour was compensated, but the obser- 

 vation was imperfect on account of the insufficient angular 

 magnitude of the solar disk. 



These experiments on the achromatization of the band had 

 been made in the hope of thereby reducing its apparent width, 

 seeing that according to (8) the difference of position for the 

 lines B and G amounts, in the case of water, to 13', much 

 more than the apparent width of the band. But the width of 

 the achromatized band could not be set at much less than I of 

 the sun's diameter"*. It seems that in estimating the dimen- 

 sions of the uncorrected band the eye instinctively allows for 

 the influence of colour. 



In experimenting upon water various kinds were tried. 

 Usually the tap- water (from an open cistern) behaved after 

 expansion as well as did distilled water. The brass hoop, 

 judiciously applied, appears to be capable of removing ordi- 

 nary surface-contamination ; but the appearance of the band 

 is liable to be deteriorated by suspended matter, which detracts 

 from the central darkness. So far as could be judged by this 

 method of observation, the best bands were sensibly perfect. 

 There was no evidence of any departure from the law of 

 Fresnel. 



Similar results were obtained from other liquids, e. g. strong 

 alcohol, sulphuric acid, benzole. Special interest attached to 

 an observation upon a saturated solution of camphor, of which 

 the superficial tension is much lower ("72) than that of pure 

 water. The band was sensibly perfect. 



Oleate of soda (^-) was troublesome on account of the 

 difficulty of avoiding scum. A pretty good band could be 

 obtained, certainly inferior to the best, possibly owing to 

 residual scum, but much better than from water greased with 

 olive-oil to the point at which the camphor motions are just 

 stopped. 



Tlie results last recorded prove that the optical effect is not 

 determined by surface-tension, for the tension of the oleate 

 solution is much less than that of any merely greased surface. 

 A similar conclusion was suggested by the observed difference 



* A coloured gla?s is still less effective than the compensating prism. 

 A reduction in the intensity of the light necessarily broadens the band. 

 A similar effect occurs if the sun is not quite clear. 



