102 BULLETIN OF THE BUSSEY INSTITUTION. 
boiling temperature. In practice, 100 grams of the powdered 
material to be tested are packed in a percolator a foot long and 
nearly two inches in diameter, made of tin ware. The lower part 
of this cylinder is constricted to an opening three quarters of an 
inch in diameter into which a glass tube is fitted by means of a 
perforated cork. In order to keep the percolator hot it is placed 
in a tin can six inches in diameter, filled with boiling water. The 
glass tube, that serves as the shank of the percolator, passes 
through a hole in the bottom of the can and is fastened there with 
a short piece of rubber tubing acting as a perforated cork. Into 
the water of the can, which acts as a jacket to the percolator, 
steam is blown constantly to maintain it at the temperature of 
boiling. Repeated small quantities of boiling water are poured 
in upon the matter in the percolator until some 200 to 400, or 
even 500,°c.c. of percolate have been collected. Care is taken 
to add the water little by little, so that each new portion of it 
may ‘‘ displace” the solution formed by the preceding portion of 
water. . 
To test the delicacy of this process of percolation when used in 
conjunction with the foregoing method of precipitating with am- 
moniacal copper sulphate, three grams of pure mannite were dis- 
solved in water, and a sufficient. number of small pieces of filter 
paper to absorb the whole of the liquid were soaked in this solu- 
tion. The paper thus charged with mannite was placed in the 
percolator and 200 ¢.c. of hot water were made to pass through it 
little by little. The percolate was evaporated to one half its 
original volume, and mixed, while still warm, with a quantity of 
the ammoniated copper solution. The light blue precipitate was 
collected on a filter, washed with cold water, and then suspended 
in warm water through which sulphuretted hydrogen was passed 
to saturation, as has been explained already. In this case the 
black colloidal precipitate of copper sulphide was made to settle 
by adding four or five drops of a dilute solution of barium chlo- 
ride and boiling. ‘The filtrate, evaporated to a small volume over 
a water bath, was finally left to dry out spontaneously in a warm 
place. Needle-like crystals of mannite separated, and were found 
to weigh 0.293 grm.; i. e., there was recovered somewhat less 
‘than ten per cent of the mannite taken. On being dissolved in 
water, treated with ferrous sulphate and hydrogen dioxide, and 
