30 368. L. Penfield—Occurrence of Alkalies in Beryl. 
Disregarding the water we should have, according to the 
usually accepted formula for beryl, Al,Be,Si,O,,, SiO, : Al,O, : 
O=6:1:3. As silica is probably the most accurately deter- 
mined of any of the constituents, one-sixth of its quotient can 
be taken as unity and the ratio of the sesquioxides and pro- 
toxides is as follows: 
¥. ue: Til. Vi as VI, VIL. 
SiO. 6:00 6:00 6°00 6:00 6°00 6:00 6:00 
Al,O; 1:06 1°03 1°08 lll 1°08 1°08 1°08 
-Protoxides 2°90 24s 2°59 2°54 2°62 2°62 2°58 
It will be seen that the alumina is in all cases too high and 
the protoxides too low. The latter might indicate that part or 
all of the water is basic, replacing the beryllium, and in all but 
the last analysis there is more than enough water, if calculated 
as replacing the protoxides, to bring up the ratio to 6:1: 
If the error in the separation of aluminum and beryllium is, as 
it is said to be, that the alumina is too low, the excess of that 
constituent in all the analyses cannot be explained if we accept 
the present formula. 
As regards the water, it certainly cannot be called accidental, 
for it is always present in amounts varying from o 2°5 
percent. It is also given off completely only by very eicong 
igniti fl chow. indicating that it is 
very firmly united in the ne PO The table is arranged as 
follows: Ist. Loss by drying for one hour at 100°C. 2d. By 
igniting for fifteen minutes at a faint red heat, so that the 
bottom of the crucible only was red. 3d. By igniting for 
fifteen minutes at a full red heat. 4th. By igniting for fifteen 
minutes at the full heat of a ring burner. 5th. By heating for 
ve minutes over a blast lamp. 
Loss at 100° low red full red white heat blast 
Hebron, Me... 60.2. 0°12 0: 0°84 1°65 2°33 
- 0°42 1°37 13 2°44 
Norway, Me... c.2ccs 0°12 { 0°42 1-49 178 9-44 
Branchville, Conn. --.-- 0°10 0°55 154 2°62 2°69 
Amelia, Virginia ...... 0-10 SON 1°58 2°10 2°19 
Pit MASS, coc 0:00 isle 1°54 2°00 2°04 
Stoneham, Me. ........ 0°08 oo 1°68 2°02 2°08 
‘Aduntwchion, Siberia _. 0°06 Pyle 0°82 1:10 114 
With the exception of analyses I and IJ it will be seen that 
the ratio of SiO, : Al,O,: protoxides is nearly 6 : 1: 2°5, and there 
is also in all cases water enough to make the ratio of SiO,: 
1,0, : protoxides : HO=6:1: 2°: 5 or12:2:5:1. As wecan- 
not disre ard the water entirely, this would seem to be a plau- 
’ sible ratio giving the ries Al,Be,H,Si,,0,,. A compound 
of the above composition ould r require the westehie% per- 
centages: SiO,67°36, Al,O, 19: 7, BeO 11°69, H,O 1: ie 100-00. 
