1884.] Some Experiments on Metallic Reflection. 197 



Table XI. — Amount of Liglit reflected by Speculum Metal Mirror. 





Observed. 



Calculated. 



J 2 . 



I 2 . 



J 2 . 



I 2 . 



o 



30 



OKJ 



64 -55 



59 -16 



69-78 



62-82 



40 



67-74 



54-50 



72-53 



59 -74 



50 



71-45 



50-05 



76-18 



55 -37 



60 



77-70 



43 -18 



80-77 



49-59 



65 



80 -01 



40-10 



83 -42 



46 -38 



70 



83 -29 



37-45 



86 -32 



43 -53 



75 



85-52 



35 -84 



89-44 



42-29 



80 



88-74 



40 -39 



92-77 



45-88 



As far as the general character of the phenomena the agreement is 

 complete and in accordance with the observations of M. Jamin, but 

 the actual values of the observed intensities always fall short of 

 the calculated intensities, the difference being least with the steel 

 mirror. 



The probable errors of the values of the principal incidences and 

 azimuths having been ascertained, the theoretical intensities of the 

 light reflected at an angle of 30° were calculated for the two values 

 obtained by adding and subtracting these sums from the means. 

 The probable errors of the photometric measurements for the same 

 angle were also determined, and Table XII gives the values thus 

 obtained. 



Table XII. 



Steel Mirror. 



Observed. Calculated. 

 J 2 . I 2 . J 2 . I 2 . 



30° .... 61*57 51-09 63 -38 .. 55'17 



59-83 .. 49-29 62 -97 .. 54-72 



Speculum Metal Mirror. 



30° .... 64-95 .. 59-60 70 '24 .. 63 '35 



64-15 .. 58-72 69*37 .. 62*37 



These numbers seem to show that the differences between the 

 calculated and observed results are not merely due to errors of obser- 

 vation, a conclusion that is rendered the more probable by the fact 

 that the difference is always in the same direction. 



The polish of the mirrors was examined at the end of the experi- 

 ments by the method suggested by Professor Stokes, and described in 



