50 J. W. Mallet on Brewsterite. 
weighed; the other specimens were treated directly with hydro- 
chloric acid, which seems of ae to be capable of effecting 
complete decomposition, The baryta was precipitated by hydro- 
fluosilicic acid,* and the relative amounts of lime and — 
were determined indirectly, by = ighing the mixed earths 
as sulphates and then as carbona 
The following are the results abiineds 
qd) Q) (3) (4) ) Mean. Atoms. 
- - 54°49 53°66 54°31 54°84 .... 65442 1°209 4°08 
. 15°42: 36°29: 16°05 = tac - ei 1 te ee i: 
Peroayd of _ trace 08 trees 255.25 SO eae 
O16 - 684 (etétresd Ahi ORE 
Strat - - 899; 990. “47-4. a ee Oe *304—1°'03 
Lim - - “92 TG See a ce 042 
Water, - - 18°89: vue Se SS eS IS 2. 2969 4°96 
99°67 99°87 
Analysis (4) was ed by an accident; and in (8) the deter- 
mination oer e earths was abandoned on ascertaining the ne- 
cessity for the removal of ammoniacal salts before precipitating 
— vid. note), a precaution which had not been take 
The a silicic acid, alumina, protoxyds and gg are clearly pres- 
ent in the ratio 4:1:1:5, giving the form 
(BaO, SrO, CaO), SiOs+4+Al20s, #6 3+5HO. 
The atomic relation between the lime, baryta and strontia is near 
4, 
. 
. 
i 
the cea of pee or lime, e found no notice taken in any work on pos 
ical the — ernte seantoaiane! male upon silico-fluorid of barium. 
Wrens states hat the latter dissolves in 3400 to 3800 parts of Paton and in 
640 to 733 parts of saath acidified by hydrochloric acid, but does no tion salts 
of ammonia. 
I digested pure silico-fluorid of barium in the cold, with frequent stirring, “for 
forty-eight hours—(a) with a saturated solution of chlorid of ammonium, (5) with 
thie sai ion di ice i The fluid was in each 
case filtered off perfectly clear, 1 ¢ centimetres were measured 
was ined ‘5 a 
(a) gave "1942 grm. ep: grm. of BaF, SiFs. Hence 1 part 
of the latter salt eaves in eee parte of a saturated solution of sal-ammoniac. 
(6) gave "1409 grm. of BaO, S "1697 grm. of BaF, SiF,, or 1 part in 589 
ecessity of ng ammoniacal salts from a fluid in which baryta is to 
a as silico-fluorid is sufficiently obvious. 
