334< M. MELLONI ON THE POLARIZATION OF HEAT. 



their axes perpendicular. The observations on this pair, as well as 

 those on each of the remaining pairs in each of the two positions, were 

 several times repeated. This method affords a compensation for any 

 possible errors of observation, as well as for those that may arise from 

 the slight variations of intensity to which the radiation of the calorific 

 source is liable, and which, owing to the perfection of the apparatus, 

 do not, even in their extreme limits, exceed a fiftieth part of the mean 

 value. As to the last column, we mean by the index of polarization 

 that portion of heat which disappears when the axes are perpendicular, 

 as compared with the quantity transmitted by the system when the axes 

 are parallel. Thus, the first pair of tourmalines transmits 27"50 when 

 the axes are parallel, and 26"4'8 when they are perpendicular : the dif- 

 ference between these two numbers, which is 1"02, repi-esents the quan- 

 tity of heat that has disappeared in consequence of the axes being 

 crossed. In order to obtain the index of calorific polarization in this 

 pair of tourmalines, expressed in hundredth parts of the quantity trans- 

 mitted when the axes are parallel, we must evidently resort to the fol- 

 lowing proportion, 26'48 : 1*02 : : 100 : x, which gives x = 3*71. 



The column of the indices shows that the proportion of heat polarized 

 varies with the qualities of the tourmalines which compose each pair of 

 plates employed. These variations, already numerous enough, if we 

 consider the number of pairs submitted to experiment, led me to 

 think that they might be found yet more numerous with other plates, 

 and that they depended very probably on the diathermancy of each 

 species of tourmaline, that is, that they arose from these different spe- 

 cies of the same mineral substance being each permeable to a differ- 

 ently constituted calorific stream. In order to verify this conjecture, 

 I placed on the apparatus that pair which polarized the greatest pro- 

 portion of heat ; and, after having made such arrangements as to render 

 the quantity of heat transmitted as great as possible, I successively in- 

 terposed, in the passage of the rays concentrated by the first lens, plates 

 of different substances. The heat which fell on the tourmalines was 

 thus more or less diminished by the partial absorption of the interposed 

 screen ; but I took care to make a suitable change in the reciprocal di- 

 stance of the two rock-salt lenses, in order to obtain an almost constant 

 calorific transmission through these different systems when the axes of 

 the two tourmalines were parallel. 



The results of this second series of experiments, performed on rays 

 emanating from the same source, and with the same pair of tourmalines, 

 are registered in the following table: 



