Mangano-Columbite and Microlite in Maine. 459 



gradually cooled, finally with ice. This operation removed 

 nearly all the caesium, which was filtered on a Gooch crucible, 

 washed with hydrochloric acid saturated with chlorine, dried 

 at 100° and weighed as Cs 2 PbCl 6 . The filtrate was evapo- 

 rated to dryness, the residue taken up in water, the lead re- 

 moved with hydrogen sulphide and the potassium and re- 

 maining csesium precipitated and weighed as platinum 

 chlorides. The platinum was then determined, from which 

 the potassium and caesium were calculated. Lithium and 

 sodium were freed from platinum and separated with amyl 

 alcohol. Only a trace of rubidium could be detected by 

 the spectroscope. 



Water was determined by strong ignition of the substance 

 over a powerful blast lamp. 



Following are the results of the analyses, together with the 

 analysis of the mineral from Hebron, by Wells : 



I. II. Average. Ratio. Hebron. 

 Si0 2 43-75 43-54 43'64 '727 9"00 43*51 

 A1 2 3 16-77 16-90 16-84 -165 2'03 16'30 

 Cs,0 36-25 36-03 36'14 -128") 36*10 

 K 2 0-33 0-42 -037 '004 I . 2'07 '48 

 Na 2 2-06 2-11 2-09 -033 f b 1*68 

 Li 2 0-03 0-13 0-08 '003 J '05 

 H 2 1-57 1'59 1-58 -082 1*01 1*50 

 CaO=0'22 



100-76 100-72 100-74 



99-84 



The results of the analyses are remarkably close to those ob- 

 tained by Professor Wells and the ratio which is nearly 

 9:2:2:1 gives the formula H 2 Cs 4 Al 4 (Si0 3 ) 9 as deduced by 

 him. 



In a recent article, " On the Constitution of the Silicates,"* 

 Clarke has discussed the constitution of pollucite and con- 

 sidered two formulae, one, H 2 Cs 2 Al 2 Si 5 15 , deduced from the 

 analysis of the mineral from Elba by Pammelsberg, the other 

 that of Wells, The latter had shown, however, very con- 

 clusively that his formula was correct and that it agreed better 

 with the results of the earlier analysis than did that of 

 Rammelsberg, and this new analysis fully substantiates his 

 view. 



Many thanks are due Professor Wells for his many helpful 

 suggestions as to methods of analysis. 



*Bull U. S. G-eolog. Survey, No. 125, p. 31, 1895. 



H 



