— 79 — 
effect, if there was any, of drying the sample at ioo degrees 
C. and then exposing the dried hay to the air until it had 
saturated itself with moisture under the usual atmospheric 
conditions. We found this much more convenient and 
without effect upon the analytical results. The samples 
however they might have been prepared, had to be sealed’ 
and every precaution taken to prevent their absorbing 
more moisture, which in the closed bottle, they did not so 
readily give up. 
PREPARATION OF THE ASH. 
I he method pursued in preparing the ash was, to heat 
a large platinum dish over the flame of a small Arcand 
burner so strongly that the bottom, the room being partially 
darkened, began to show a dull redness over an area from 
one to one and a half inches in diameter; the weighed 
sample was introduced in separate portions until the char¬ 
red mass filled the dish to rather more than two-thirds full 
when it was allowed to continue heating until the volatile 
matter was nearly or quite expelled; the bulky mass was 
then transferred to a porcelain casserole and allowed to 
burn of its own accord so long as it would. The mass was 
stiried frequently and new portions of the sample were 
created in like manner until the whole of it was brought 
into the casserole When it had burned out and cooled 
sufficiently the still highly carbonaceous ash was extracted 
with water and washed so long as the wash water showed 
h? P resenceof chlorine. I he insoluble portion free from 
chlorine was then burned to whiteness at as low a heat as 
was feasible. This often proved to be a tedious operation. 
1 he solution containing the -alkalies was evaporated to dry- 
ness in a platinum dish after the addition of the insoluble 
portion and enough amnionic carbonate to convert the cal¬ 
cic oxide formed into calcic carbonate. The ash was 
eventually dried at 200 degrees C., at last with addition of 
solid amnionic carbonate and bottled while still hot and 
carefully corked. If the ash thus prepared has to be kept 
tor any length of time, it is necessary that it should be 
sealed. 1 his degree of care was taken to avoid loss of 
chlorine and also possibly of sulphur. During the course of 
the preparation, however, we became convinced that the 
precautions taken were inadequate, because the loss of 
chlorine was not due to the volatilization of the sodic or 
potassic chlorides, but due to the formation of ammonic 
chloride. I he odor of ammonia was very marked at cer¬ 
tain stages of the process and was present in sufficient 
quantities to react upon red litmus paper, and a glass cylin- 
