in Labrador Felspar. 325 



nada balsam interposed, and I found that the tint diminished as 

 the angle of incidence increased, and that the same effect took 

 place in the same degree in all azimuths. This experiment 

 proved incontestibly that the colours were not those of polarised 

 light. That the cavities do not contain a gas or a fluid of any 

 kind, is obvious from the fact, that the felspar does not decre- 

 pitate or burst with heat. Hence, it follows, that the parallelo- 

 grams must be either empty, or must contain indurated matter. 



In order to ascertain, upon the supposition of the parallelo- 

 grams being solid, whether the colour arose from the thinness of 

 the solid matter, or from the thin open space which separated 

 the surface of the parallelograms from the adjacent felspar *, I 

 observed the particular tints which a number of individual pa- 

 rallelograms produced at a given incidence ; and upon reversing 

 the specimen, I found, that, in every case, the very same tints 

 were developed at the same angle of incidence. This result 

 clearly proves, that the tints were due to the thickness of the 

 cavity, whether they were empty or filled with indurated mat- 

 ter. 



The examination of individual parallelograms presents some 

 instructive peculiarities. While the greater number give an 

 uniform uninterrupted tint, several have the appearance shewn 

 in Fig. 2. In No. 1, the parallelogram is imperfect. In No. 2, 

 it is more so, though the individual patches of colour fill up 

 its outline. In No. 3, they are smaller still, and more un- 

 equal. In No. 4, we can still discover the outline of each in- 

 dividual patch ; but in No. 5, the patches are so minute, that the 

 surface of the parallelogram produces all the variety of mottled 

 colours. These phenomena indicate a general resemblance to 



* It is from this cause that the splendid colours arise which accompany the den- 

 dritic crystallizations of titanium in mica, which I have examined with much atten- 

 tion. 



s s 2 



