■i'2 Penfield and Stanley — 



The hornblendes represented by analyses VI to X constitute 

 a group by themselves, which may next be considered. They 

 di titer from the tremolite-actinolite group in that the ratio of 

 SiO„ : RO, instead of being as 1 : 1, indicates an excess of RO 

 o^er Si0 2 and with the increase of RO there is a correspond- 

 ing increase in the amounts of A1 2 0„ and Fe 2 3 . The relations 

 are best shown in the accompanying tables, where the R„0 3 

 quotients, (the sum of the results obtained by dividing the per- 

 centages of A1 2 3 and Fe 2 3 by the corresponding molecular 

 weights) are given together with the Si0 2 : RO ratios, RO 

 including Na 2 0, K 2 0, H,0 and F 2 . 







R 2 3 



quotients 



Si0 2 



RO 



VI. 



Pierrepont 





•045 



1 : 



1-05 



vi r. 



Renfrew 





•124 



1 : 



1-08 



VIII. 



Edenville 





•129 



1 : 



1-10 



IX. 



Cornwall 





•165 



1 : 



1-11 



X. 



Mte. S omnia 





•172 



1 : 



1-14 



The first analysis of this group forms a connecting link 

 between the actinolite and hornblende groups. If it is com- 

 pared with analysis V no essential differences will be found ; 

 it is only when the ratios are examined that a slight excess of 

 RO is noted in the mineral from Pierrepont, and this occurs 

 with a smaller R 2 3 quotient than that found in the actinolite 

 from Kragero ; as previously stated, the two groups grade 

 imperceptibly into one another. The other analyses are, how- 

 ever, considerably different from any thus far considered ; 

 they are all of dark green to black varieties, the color due to 

 the presence of iron ; the percentages of sesquioxides are high, 

 about 20 per cent in IX and X, and the excess of RO as shown 

 by the ratios is very noticeable. Another marked difference 

 is seen on comparing the silica percentages ; starting with 

 about 57'5 per cent in tremolite, it falls below 37 per cent in 

 analysis IX. 



It is evident that if the minerals of the amphibole group 

 are all to be regarded as salts of one acid, we must in some 

 way be able to account for the introduction into the amphibole 

 molecule of both sesquioxides and increasing amounts of pro- 

 toxides. This mav be done bv assuming in addition to the 

 fluor-hydroxyl radical [R /7/ 2 (F,OHl]^thus far considered, 

 some other radicals containing both R 2 3 and RO. Thus there 

 might be considered the following alumina and the correspond- 

 ing ferric iron radicals. 



"Al-CV -Al<^- R/ 



>0 >R" or 1>R" 



-Al— (K — Al< 



*0— R' 



