132 Prof. How on the Mmeralogy of Nona Scotia* 



two recorded analyses of the mineral (in my paper allusion is 

 made to the probable replacements), viz. : — 



Gyrolite. 



! ; s 



Anderson. How. 



Lime . . . 33-24 29-95 



Alumina . . 1-48 127 



Magnesia . 0-18 O08 



Potassa . . none 1*60 



Water . . 14-18 15-05 



Silica . . . 50-70 51-90 



99-78 99-85 



I wish to draw attention to this concordance particularly, 

 because I find in Dana (Mineralogy, fifth edition, p. 398) the 

 formula 



(§Ca + JH)Si + H=2CaO,3Si0 2 + 4HO 



assigned to gyrolyte : the percentages calculated from this ex- 

 pression, 



Si0 2 =30 "... 49-45 

 ^CaO =18-66 . . . 30-76 

 1J HO = 12-00 . . . 19-78 

 99-99 

 by no means correspond with our experimental results. Again, 

 I find that Greg and Lettsom (Manual of Mineralogy of Great 

 Britain and Ireland, p. 217) give for the same mineral the 

 formula 



3CaSi + 4H=3CaOSi0 3 + 4H 2 0, 



with the calculated percentages 



Si0 3 .... 53-29 

 CaO .... 32-86 

 H 2 O . . . . 13-85 



100-00 



which agree no better with our experimental numbers than the 

 last. Neither of these formulae, therefore, has the validity of 

 Anderson's, which it affords me a melancholy pleasure to re- 

 affirm on behalf of my lamented friend, under whom I did far 

 too much work not to know what delight he took in accuracy, 

 both of work and its representation. The water is much tco 

 high, the silica and especially the lime are too low, allowing 

 for replacements in the latter, in Dana's formula ; while the 

 other shows too little water and too much silica. • 



Considering that Anderson states (loc, eit.) that he foui;d 



