jyQ PROFESSOR FORBES'S RESEARCHES ON HEAT. 



use of it. But the quality of the heat from incandescent platinum varies between 

 the widest hmits. Nor is this wonderful ; it is composed of heat from two xerj 

 different soui-ces combined in uncertain proportions, that from the incandescent 

 coil of wire, and that from the alcohol flame which heats it. The intensity of in- 

 candescence, too, varies exceedingly. On one occasion when the incandescence was 

 unusually bright, and the alcohol flame very low, I obtained a higher degi-ee of 

 polarization than I have ever done before or since. The ordinary proportion be- 

 tween the indications with I and K parallel and crossed, is with incandescent pla- 

 tinum 100 to -20 or 27. In this case it was 100 : 20 ; and when the heat was lifted 

 by an interposed plate of thin glass, it rose as high as 100 : 13. 



15. The general results obtained in the way above described are stated in the 

 following Table, in which I have included the numbers for mercury heated to 

 410°, and for boding water taken from the Second Series, art. 22; those experi- 

 ments not having been repeated because the use of a lens is in those cases of little 

 avail. 



Polarizing Plates I and K. 



Source of Heat. Rays out of 100 polarized. 



Argand-Lamp, . ... . . .78 



Locatclli-Lamp, . . • . • . 76 to 77 



Incandescent Platinum (usually) . . . . 74 to 76 



Incandescent Platinunij with Glass .06 inch thick, interposed, 6 to 7 per 



cent, more, or . . . . . . 80 to 82 



Alcohol Flame, . . . . • .78 



Brass heated to about 700°, ..... 66.6 



Ditto, with a plate of Mica .016 inch thick interposed, (between K and B) 80 



Mercury in a Crucible at 410°, . . . . .48 



Boiling Water, . . . . . .44 



16. I next tried whether the closest possible approximation of the mica plate 

 1 to the pile would produce any effect. The pasteboard A was removed, and the 

 mica plate I was brought up until it touched the funnel-shaped reflector of the 

 pile. In this extreme case the apparent polarization was found to be diminished 

 about 2 per cent., whether in the case of incandescent platinum or of dark heat. 

 1 shall not inquke whether any or how much of this effect was due to the heating 

 of the mica surface, and how much to the reflection of heat from the interior of 

 the cylindrical tubes containing the mica plates, since it is evident that this could 

 not have produced the variation of effect shewn in the above experiments. 



17. I presume that it wiU be conceded, that the experiments now cited, in- 

 controvertibly establish the unequal polarizability of heat from different sources. 

 Yet, I confess, I should have felt uneasy, could I have thrown no light upon the 

 cause of the discrepancy between M. Melloni's results and my own. This I be- 

 lieve, that I am able completely and satisfactorily to do, allowing hhn every cre- 

 dit for the perfect exactitude of his experiments. For the sake of clearness, I 

 will state the course by which I myself arrived at this result. 



