406 Dr. Thomson'' s Analysis ofNeedlestone. [Dec. 



any of the minerals hitherto reckoned mesotypes have a specific 

 gravity higher than 2-38. Even the specific gravity of stilhite, 

 which Haiiy fixes at 2-5, has not been verified by subsequent 

 eixperiments. The reader may see a hst of the specific gravities 

 of all the zeolites hitherto taken in HofFmaijn's Handbuch der 

 Mineralogie, vol. ii. part i. p. 242. 



The hardness, the structure, the electric properties, and the 

 action of acids on mesolite, agree precisely with the same pro- 

 perties in skolezite. When treated with oxalic acid, it leaves 

 rather a smaller sediment than skolezite. Before the blow-pipe 

 it becomes opaque, and curls up, but not so much as skolezite ; 

 neither does it emit quite so much hght. It does not form a 

 frothy slag, but melts with the extrication of many air bubbles 

 into a porous and almost opaque bead. 



Its constituents, according to the analyses of Fuchs and 

 Gehlen, are as follows : 



Silica 47-0 



Alumina 25*9 



Lime 9-8 



Soda 5-1 



Water 12-2 



100-0 



We see from this analysis that mesolite lies intermediate 

 between natrolite and skolezite. The proportion of silica and 

 alumina in both is nearly the same. The proportion of lime is 

 about two-thirds of that which occurs in skolezite, and the pro- 

 portion of soda about one-third of what occurs in natrolite. 

 Hence it seems to be a mixture or compound of two parts of 

 skolezite and one part of natrolite. 



The constituents of a mineral composed of two parts skolezite 

 and one part natrolite woiUd be as follows : 



SiUca 47-0 



Alumina 25*96 



Lime 9'46 



Soda 5"4 



Water 12-l6" 



100-00 



Now this is so very nearly the constituents of mesolite derived 

 from the analyses of Fuchs and Gehlen that it seems at first 

 sight perfectly satisfactory. I confess, however, that I enter- 

 tam some doubts whether mesolite be any thing more than a 

 mere mixture of the two species. All the analysis of Fuchs and 

 Gehlen indeed approach so near each other that the differences 

 may be fairly ascribed to errors in the exoeriments 5 but they do 



