1884.] 



Some Experiments on Metallic Reflection. 



193 



In Table VITT the numbers obtained in this way are given in the 

 2nd and 5th columns, the values as determined by the experiments 

 described in the paper already referred to in the 3rd and 6th columns, 

 and in the 4th and 7th columns the results calculated out, as described 

 in a subsequent part of the present paper, from the measurements 

 made with polarised light. 



Table VIII. 



Steel. 



Speculum metal. 



Observations made 

 with polarised light. 

 J 2 + I 2 



Observations made 

 with ordinary light. 

 rJ 2 + I 2 



Observa- 

 tions made 



with 

 polarised 

 light, 

 J- + 1 2 

 2 



Observations made 

 with ordinary light. 

 rJ 2 + l 2 



2 



r + 1 



r + 1 



o 





Observed. 



Calculated. 





Observed. 



Calculated. 



30 



55-44 



54-93 



56-62 



61 -85 



66-87 



62 39 



40 



55-24 



55 '62 



57-26 



61-12 



67-26 



62-61 



50 



54-75 



56-74 



57 -84 



60-75 



67-26 



63-15 



60 



54-60 



57-63 



59 -04 



60-44 



66-32 



64-31 



65 



53 -70 



58-37 



59 -0 



60-05 



66-53 



64-53 



70 



54-40 



58-09 



60-65 



60-37 



67-65 



65 51 



75 



55 -37 



58-69 



62 -33 



60-68 



67 43 



66-22 



80 



57 23 



63-56 



64-10 



64-56 



70-17 



69 -98 



In addition to the actual numerical differences between the values, 

 the two sets of observations appear to differ fundamentally, for whilst 

 the numbers in the second and fifth columns diminish slightly, and 

 then increase again, as the angle of incidence increases, those in the 

 third and sixth increase with the angle, a result that ; as was pointed 

 out in the former paper, is not in accordance with either theory, or 

 previous observations, and which, as was stated in the paper, if 

 erroneous, must have been caused by some defect in the method 

 employed, and therefore common to all the determinations. 



Further consideration showed that such was really the case, and 

 that the defect in the method was the one pointed out by Professor 

 Stokes in the note appended by him to the paper (" Proc. Roy. Soc," 

 vol. 35, p. 39). As has already been stated, the apparent brightness 

 of the paper varied with the polarisation of the incident light, or, in 

 other words, the amount of light irregularly reflected, or diffused, by 

 the paper was different for light polarised in, or perpendicularly to, 

 the plane of incidence. The light being incident upon the paper at 

 an angle of about 30°, whilst the line of sight formed an angle of 



