Prof. B. Silliman, Jr., on some American Minerals. 381 

 This gives the ratios 



••• ■•• • #••• .. . . . • # 



Si 3 Al« R -f-H=R Si+2Ai* Si+H. 



Atoms. Required. Found. 



3 atoms silica = 173194=perct. 36-31 35 708 



4 « alumina = 2569 32 53-87 53] 31 

 1 atom lime = 356-02 7-46 8926 

 1 " water = 112-48 2 36 2-303 



4769-76 100-00 100068 



species 



shown on further examination that margarite is a hydrous min- 

 eral. At present it is reported as anhydrous, and its proportions 

 of silica and alumina are different from the present species. Its 

 analysis given by Hausmann on the authority of the Gottingen 

 Laboratory is 



Silica, 33-50= 8atoms. Silica, =4618-48 perct. 34-47 



Alumina, 58- 12 " Alumina, 7708-00 " 57-55 



Lime, 7-50 3 " Lime, 1068-06 " 7-98 



Protox. iron, -42 



Manganese, 0-03 13394-54 10000 

 Magnesia, 0-05 



99-50 



R 3 AP« Si 3 =R 3 Si»+6A1» Si. 



Possibly a new analysis may bring these species together. 



The species Corundellite occurs not only in the broad foliated 

 masses above alluded to, but also in small scales disseminated 

 throughout the mass of granular corundum, at Unionville, Pa., 

 and in this form is quite abundant. Not unfrequently these scales 

 have a delicate shade of violet, especially when wet. The rock 

 J s difficult to break and the corundellite appears to adhere very 

 strongly to the associated minerals, and the laminae are not so 

 easily separable as in the foliated masses.* 



Euphyllite. 



This beautiful pearly white mineral is found associated with 

 black tourmaline and corundum, at Unionville, Pa. Form appa- 

 rently hexagonal, cleavage eminent on basal plane, the laminae 



* The species Barsawite (G, Rose), appears in the Urals to hold the same geog- 

 nostic relations to corundum as do the minerals of the present memoir, in this coun- 

 ty- Its composition, however, is quite distinct, (silica 40-01, alumina 33-85, lime 

 5 "W» magnet in 1*55=99-87, Varrentrapp,) while its hardness 6, and abeeoc of mi- 

 caceous structure render it entirely distinct. It approaches scapolite in composi- 

 J**, but with a smaller quantity of protoxyd. I am led to allude to this species 

 Irom the fact, that an intelligent foreign mineralogist to whom I showed some of the 

 corundellite, remarked that* there appeared to be a similarity between the species. 

 1 «ere is however a most marked difference in that corundellite is a mica. 



Second Series, Vol. VIII, No. 24.— Nov., 1849. 49 



