40 JOURNAL OF THE 
command. The drying of large portions and the subse- 
quent ignition would be exceedingly tedious and time con- 
suming, besides requiring such apparatus as could not be 
well afforded. Five or six grams have been about the 
largest amounts that could be well handled at one time. 
Even such an amount as that required from sixteen to 
twenty days for the completion of the experiment. It 
could not safely be hurried through in shorter time. 
The weighings were carried but upon an excellently ~ 
constructed Sartorius balance, intended for a load of 200 
grams. The heaviest apparatus used weighed less than 
sixty grams. ‘The weights were corrected by one which 
had been compared with the standard at Washington. 
All objects were weighed against a tare of as nearly the 
same size, form. and weight as possible, all of the flasks, 
crucibles, etc., being made in pairs. This partly avoided 
the necessity for a reduction of the weighings toa vacuum 
and corrections for moisture, pressure, etc. Such ¢or 
rections would have kad little meaning in comparison with 
the other inaccuracies of the process and manipulation, 
and could only serve to give a false appearance of excess- 
ive accuracy. The objects were left one-half hour in the 
balance-case before weighing, experiments having shown 
that this time was sufficient. Of course the adjustment 
of the balance was carefully watched, and the balance, 
which has been used very little, was put to no other use 
during the progress of these experiments. 
METHOD OF WORK. 
The purified oxychloride was introduced into a small 
olass flask having a capacity of 100cc. This was provided 
with a glass stopper ground to fit, and aiso a second one 
with two tubes arranged for the passage of the hydro- 
chloric acid gas. The arrangement of the tubes was 
similar to that in an ordinary ether wash-bottle, though 
both tubes outside were bent downwards and had little 
