264 CLARKE 



which compares as follows. The 7'cdiiccd analysis was com- 

 puted from Mr. Steiger's analysis by rejecting water, transform- 

 ing Na.,0 into the equivalent amount of Ag.,0, and recalculating 

 to 100 per cent. 



Calculated. Fouud, reduced. 



SiO., 37.39 3^-94 



AUO3 12.67 12.57 



AgjO. . . . : 43.23 43.47 



N2O3 6^ 7-02 



100.00 100.00 



The thallium chabazite is also a nitrate derivative, but the 

 transformation was less complete than in the case of the silver 

 salt. It corresponds to 



3Al,Tl3(Si30,),N03; 



iAl3Tl3(SiO,)3. 



Reducing the actual analysis by exclusion of water, computing 

 CaO and Na.O into Tl^O and recalculating to 100 per cent., we 

 have 



Calculated. Found, reduced. 



SiOj 2S.47 2S.46 



AljOj 10.3S 10.5S 



T1,0 57.49 57-48 



N,05 _3^ _34S 



100.00 100.00 



These agreements are strikingly close, and establish, with a 

 high degree of probability, the existence of the chlorine or nitro- 

 derivatives represented by the formula?. These substances, 

 sometimes mingled with the normal derivatives, seem to exist 

 in the residues obtained in the experiments. Even the barium 

 chabazite " C" agrees roughly with the composition. 



Al,Ba3(Si30,),Cl + Al,Ba3(SiOJ,Cl ; 



although much weight cannot be "given to this coincidence. It 

 may be noted, in passing, that Weyberg ' has recently described 

 compounds obtained by fusing kaolin with calcium chloride or 

 bromide, to which he assigns the formulas 



' Centralblatt Min. Geol. Pal., 1904, p. 729, and 1905, p. 13S. The calcium 

 chloride derivative had previous! v been noted by Gorsreu, Bull. Soc. Min., 10, 

 276. 



