new Varieties of Ventriculites. 137 



trating material. The plates do not show the striation observed 

 both in Culsageeite and Jefferisite. No evidence of macling 

 could be found. The plates were optically biaxial, with a small 

 optical angle and strong negative double refraction like the other 

 vermiculites and magnesian micas. The hardness of the mineral 

 is about 1*5 ; two determinations of its specific gravity (taken 

 in alcohol at 22° C.) gave 2'] 60 and 2'161 . Heated in a closed 

 tube it gives off water acid in reaction, changes colour, and 

 doubles its volume, and in the forceps before the blowpipe fuses 

 to a dirty enamel. 



The mineral, dried at 100, lost upon ignition 



(1) (2) (3) (4) Mean. 



11-21 11-50 11-13 11*43 11-32 



per cent, of its weight. The details of these determinations will 

 be referred to hereafter. 



In (1), (2), (3), and (4) of the following analyses the ignited 

 mineral was decomposed by fusion with sodic carbonate; in (5) 

 and (6) the mineral, dried at 100°, was decomposed by hydro- 

 chloric acid. In all, the ferric and aluminic oxides were weighed 

 together, and the former subsequently determined by reduction 

 and titration with a solution of potassic bichromate. In (4) the 

 iron and alumina were together separated from magnesia by pre- 

 cipitation by ammonia with the usual precautions. In (5) the 

 bases were converted into nitrates, and alumina and iron sepa- 

 rated from magnesia by Devilled process. In both (4) and (5) 

 the magnesia was weighed as the pyrophosphate. A trace only 

 of ferrous iron was found. The spectroscope indicated small 

 amounts of lithium and potassium*. 



(1) (2) (3) (4) (5) (6) 



Si ........ 41-14 41-28 41-04 4078 4V27 41*23 



¥*e 4-35 4-26 4*49 4*14 



^ i .. 1 15-05 15-19 



Mg ... ( 43 ' 2i) 43 ' 52 128-82 28-25 



j,i, K .... trace 



H 11-32 11-32 11-32 1132 11 -32 



100-20 100-14 100-46 100-17 



Analysis (5), which probably represents the amount of mag- 

 nesia contained in the mineral more correctly than (4), affords 



* In Hallite and Jefferisite, as well as in the Lerni and Pelhnm varieties, 

 small amemnts of lithium were detected. In none of the vermiculites have 

 we been able to detect fluorine. 



