30 RICHARDS—SOLVENT IN CRYSTALS. [April 4, 
ranted assumption is made that every hidden cell containing*the 
mother-liquor has been split open by the pestle, and in the next 
place the equally unwarranted assumption is made that all the 
mother-liquor thus exposed evaporates into the uncertain mixture 
constituting the atmosphere of the laboratory. 
The former of these assumptions will be considered first. Is it 
possible to open all the cells enclosing mother-liquor by means of 
any finite amount of powdering ? 
This question cannot be answered a priori ; accurate experiments 
are needed to decide it, and no published work known to me 
seems to furnish the needed data. Accordingly a series of experi- 
ments was planned which involved the progressive powdering of a 
typical substance. In order to separate the powders of different 
degrees of fineness four pieces of brass netting were used, having 
openings about 0.5, 0.3, 0.23, and o.16 mm. in diameter respec- 
tively. This netting was cleaned with acid and ammonia, and 
showed a brilliantly clean surface in the microscope.’ 
The test substance chosen was baric chloride, because careful ex- 
periments on the atomic weight of barium had shown that it may be 
analyzed with ease and accuracy. A pure finely-crystallized speci- 
men of the salt was slowly and carefully powdered and sifted 
through the successive sieves. That which went through the finest 
sieve was still more finely powdered, until the average diameter of 
the particles as estimated in the microscope was perhaps one- 
twentieth of a millimeter, some being coarser and some finer. 
Each specimen was thoroughly air-dried. 
Thus were obtained four examples of baric chloride containing 
crystal-water, of four different degrees of fineness, the coarsest 
particles averaging about a thousand times the bulk of the finest. 
Upon analysis by heating to constant weight at 400° these samples 
yielded respectively 14.780, 14.771, 14.763, and 14.760 per cent. 
of water. 
The data, reduced to the vacuum standard, were as follows. 
The experiments were made in 1893: 
Average diameter Weight of Loss on Per cent. of 
No. of particles. salt, heating. water found, 
Bs Set acme ee 0.45 mm. 4.15929 0.61475 14.780 
2, wis siiteeene tats 0.27 mm. 3.65127 0.53933 14.771 
Bie ale ovat eee 0.20 mm. 4.54136 0.67047 14.763 
AGS hs; < atone 0.05 mm. 10.13720 1.49620 14.760 
' The material sifted through this netting was tested for copper, with satisfac- 
torily negative results. 
