{ Bai | 
THERE is a method of inveftigating the proportion of ingre- 
dients in this falt, which I once purfued, and which afforded 
a refult not very different from that of Mr. Bergman. 
I constpERED 100 parts of cryftallized mineral alkali, to ap- 
pearance perfe@tly dry, and yet not decompofed, to confift of 16 
parts of fixed air, 20 of pure alkali, and 64 of water, as Mr. 
Bergman found them. However this proportion of fixed air in 
cryftals to appearance perfeQly dry and not efflorefced I often 
failed of finding. Sometimes this proportion was 12 per cent. 
fometimes 13 or 14, and feldom 16 per cent. However, I 
imagined that the falt being perfetly cryftallized neceffarily re- 
quired the fame proportion of fixed air to the alkali, and that 
the deficiency arofe from an excefs ‘of unperceived moifture, 
which made up the weight of roo parts; and therefore when I 
found a fmaller quantity of fixed air in 100 parts of this falt, I 
inferred that it contained a fmaller quantity of alkali, in the pro- 
portion of 16 to 20. So that if 16 parts of air indicated 20 of 
alkali, 12 of fixed air fhould indicate 15 of alkali. I was even 
confirmed in this opinion by the following experiments. I found 
that 100 grs. of cryftallized mineral alkali, which afforded 12 of 
fixed air, were faturated by 52 of vitriolic acid; but that another 
portion of the fame alkali that emitted 13 grs. of fixed air 
required 56 of the fame acid. . Now 12.52::13.56 nearly. 
Again, I found that 100 parts of this alkali which gave out 15 
of fixed air were faturated by 121 of marine acid, whereas 
another portion of the fame alkali which contained but 13,5 of 
fixed air, took up but 108 of the fame acid; and 15.13,6::1r21.109 
nearly. It muft be owned that this conclufion appeared fuffici- 
ently 
