THIRD SERIES.— REFRANGIBILITY OF HEAT. I99. 



(made between the 21st March and the present date) ; but it is important for the 

 credit of the results to observe, that they not only derive a general confirmation, 

 but exhibit an almost exact numerical coincidence with those obtained formerly, 

 and with less careful adjustments. 



65. Avoiding, then (as in these papers I have habitually done), the tedious 

 detail of minute precautions which the experienced operator will soon discover, 

 and which to others would be of little use, it is to be understood that in the fol- 

 lowing experiments on the law of the Transition from Pai*tial to Total Reflection, 

 the arrangement was that shewn in Plate XIII. Fig. 1, and described (with the 

 adjustments) in arts. 51-57 ; — that the centre of the pile p was 13 inches from 

 the prism P, and the distance of the som-ce of heat S from P was 12 inches ; — 

 that a diaphragm T, whose apertiu'e was 1 inch by 1', was placed in the path of 

 the ray usually between P and L near P ; — that the aperture of the pile was con- 

 tracted to a breadth of one inch, whose centre was exactly in the line ad; — 

 and that only that part of the prism was employed which was free from flaws 

 capable of producing total reflection. 



QQ. The diagonal of the lozenge frame was varied from 14.5 inches up to 

 16.5 or 17.0, about eight observations of the intensity of reflected light being 

 made at intervals. The series was then frequently reversed, and the mean re- 

 sults of the going and retm-ning series taken to allow for any change which might 

 have occurred in the intensity of the course. In aU cases an observation of veri- 

 fication was made and such change allowed for. The dynamical eflect on the 

 galvanometer (II. 8) was obsen^ed and noted. 



67. In reducing the observations the following plan was adopted. The in- 

 tensity coiTesponding to the diagonal 14.5 inches being assumed ■=. 100, the other 

 intensities were reduced relatively to it, and projected, as explained in art. 60. By 

 this means different series of observations became at once comparable with each 

 other, and the beauty and regularitj^ of the curves thus formed, and the almost 

 perfect identity of those obtained on difierent days, and with different adjustments, 

 give a degree of confidence in the results which is extremely satisfactory. When 

 from the nature of the heat the effect was very small (as in the case of alum being 

 interposed, or the som-ce being of low temperature), I have endeavoured to supply 

 the deficiency by multiplying observations, and the uniformity of the curves thus 

 obtained has been the test of my success. Where this test has failed (as in the 

 attempt to work with heat of 212°), I have suppressed the results. 



68. I am unwilling to swell this paper by a quotation of individual experi- 

 ments, of which the number is very great,* but I think it fair to give specimens 

 of the actual work in a few cases. 



* It may not be superfluous to state, that during the course of the experiments referred to in this 

 series of papers, I have adopted a uniform and clear system of recording my experiments, which admits 

 of subsequent reference, and, if necessary, of publication. The experiments have been fairly written out, 



