266 Dr Brewster's Description of Gmelinite. 



with violence from the fragment. The continued application 

 of the heat drives off the water of crystallization, and reduces 

 the fragment to a white fibrous-looking powder. In perform- 

 ing this experiment, by exposing the fragment on a piece of 

 glass to the fire, I was surprised to observe, upon looking at 

 the powder with a microscope, that many of the particles were 

 in a state of restlessness, some of them leaping from the glass, 

 and others endeavouring to separate themselves from the 

 larger particles to which they were attached. This effect was 

 no doubt owing to the heat of the glass, which continued to ex- 

 pel the water of crystallization which still remained in some 

 of the particles, for I could not discover in the powder any 

 trace of pyro-electricity. The property which has now been 

 described is possessed also by the Gmelinite from Glenarm, 

 but it is not possessed by Analcime or Chabasie, or, so far as I 

 know, by any other mineral, and may be regarded as an infal- 

 lible chemical character of this species. 



The following is the composition of the Gmelinites from the 

 Vicentine, according to Vauquclin. 



« Gmelinite from Gmelinite from 

 RIontecchio Maggiore. Castel. 



Silex SO - - 60 



Alumine 20 - - 20 



Lime - 4.5 - - 4.25 



Soda - '• - " ' *£/ - 4.25 



Water 21 - - 20 



Loss 1.5 



100 - - 100 



I cannot conclude this notice without directing the attention 

 of the philosophical mineralogist to the peculiar value of opti- 

 cal characters. The analysis of the Vicentine minerals by 

 Vauquelin gave results so like those obtained from the Cha- 

 basies, that the chemical mineralogists even never felt them- 

 selves authorized to consider them as new. In hardness and 

 specific gravity these minerals were almost exactly the same as 

 Chabasie, and the obtuse rhomboid from which the six-sid- 

 ed prisms from Castel are derivable, has almost the same 

 angle as that of Chabasie. Hence, Mr Haidinger was led to 

 consider them as Chabasics, and, indeed, in any system which 

 does not take cognizance of chemical and optical characters, 



