THIRD SERIES.— EEFKANGIBILITY OF HEAT. 201 



69. After the observations made as now described have been projected in the 

 form shewn, Plate XIII, Fig. 3, and Plate XII, the diagonal coiTesponding to the 

 maximum rate of decrease of the intensity was determined, for the pm-pose of de- 

 ducing the index of refraction. The following enumerations of the kinds of heat 

 employed, and the results derived from the several projections, will give a just idea 

 of the confidence due to the results. They are distinguished into those made 

 since, and those previous to the 21st March, because some additional precautions 

 have been taken since that time, which do not, however, appear to have produced 

 a sensible change. Of the experiments made in the way above described, only 

 one series is rejected, on account of its discrepancy from others of the same kind, 

 (the discrepancy was so large as to indicate a displacement of the prism, or some 

 fundamental derangement not perceived at the time) ; and another on account of 

 the irregularity of the points marked out for the ciu-ve, although the general form 

 of the curve did not differ from others smiilarly obtained. 



70. Sources of Heat. — (1.) The direct rays of the LocateUi Lamp. A slightly 

 concave reflector was employed. (2.) The same lamp, with a reflector having the 

 form of a portion of a sphere concentric with the wick ; the heat transmitted 

 through alum. (3.) Heat from the same source transmitted by window-glass .06 

 inch thick. (4.) Heat from the same transmitted by opaque black glass (through 

 which the disk of the unclouded sun is just visible). (5.) Heat from the same 

 transmitted through dark coloured mica, by which direct sunlight is absolutely 

 stopped. This singular substance I long sought for in vain, it is unknown to many 

 practical mineralogists ; it transmits green hght at small thicknesses, Avhen thicker 

 its colom- is hair-brown. By reflected light its colour is between green and black. 

 (6.) Heat from incandescent platinum. (7.) The same sifted by window-glass as 

 above. (8.) The same sifted by opaque mica. (9.) Heat from dark brass about 

 700°. This is obtained from a nearly cylindrical cover of smoked brass placed 

 over the flame of a spirit-lamp, so as entirely to conceal it, and which gives re- 

 markably good results, without increasing considerably the angular breadth of 

 the source (which is gi-eatly to be avoided when a lens is used). It is in fact 

 not much greater in size than the helical coil of platinum wire used in (6). 

 (10.) The same, sifted by clear mica .0044 inch thick. (11.) Heat from a crucible 

 of mercury about 450°. The crucible was about 2 inches in the side, smoked 

 externally, and heated by a spirit-lamp. The temperature of the mercury which 

 it contained (covered with sand) was noted at each observation by means of an 

 inserted thermometer. 



71. The results were the following: 



VOL. Xn^ PART I. c c 



