145 



In the second section, the observations formerly published on the 

 polarizing action of tourmaline are confirmed, including the case 

 where heat entirely dark was employed. Of all the observations 

 this alone has been attended with any difficulty in the verification. 



The third section treats of the laws of the polarization of heat by 

 refraction or transmission. The author denies that the results 

 given in the first series of these researches were held out as nume- 

 rically accurate ; and he took occasion to state verbally, that whilst 

 an attempt has been made to depreciate the credit due to his obser- 

 vations, on the ground of the numerical inaccuracy of his first re- 

 sults, the objectors themselves seem to have fallen into a graver 

 mistake, that of considering all kinds of heat equally polarizable at 

 the same angle of incidence. A great number of experiments have 

 confirmed the opinion which the author formerly entertained, that 

 heat from non-luminous sources is less polarizable by a given plate 

 of mica at a given angle of incidence than wheu accompanied by 

 light. He has constructed polarizing mica bundles of great tenuity, 

 by using sudden heat to split that substance into a vast number 

 of minute laminae. With one pair of such plates he obtained the 

 following percentages of heat stopped when the planes of refraction 

 in the two bundles were rectangular : 



Sources of Heat. Ray, out of lOO polarized. 



Argand lamp, 

 Incandescent platiniun, . 

 Heat from brass about 700°, 

 Do. transiiiitled through glass, 

 From mercury in a crucible at 410 

 From boiling water. 



72 to 74 



72 



63 



72 



48 



44 



These observations were repeatedly made, and were confirmed by 

 others made with a different pair of mica plates. The results are 

 shewn to agree with those obtained in the case of light ; the heat 

 at the bottom of the list being the least refrangible. 



In the fourth section the law of polarization by reflection is dis- 

 cussed. A number of reflecting surfaces were tried, and split mica 

 was preferred. The amount of polarization, by reflection at a given 

 angle, is shewn to vary with the source of heat : and it is probable 

 that the kinds of heat do not rank in the same order when the an- 

 gle is changed. This is the case in light. The change of the plane 

 of polarization by subsequent reflection is similar to that which oc- 

 curs when light is used. (See page 130.) 



