=a ee 
890 
A uniform temperature was maintained in the combustion furnace 
throughout all the distillations and this was shown by the pyrometer to 
be 625°."8 
The possibility of maintaining an average, uniform temperature in the 
Tropics is not a serious problem. All gas measurements and gas analyses 
were made at 30° and all data are given at this temperature. 
It was not possible to exhaust the air completely from the apparatus, 
but from its known capacity (864 cubic centimeters) and the pressure, 
as read by the manometer (F), the amount of air contaminating the gas 
was determined and is given in column 3 of the experimental results. 
The coal analyses were made according to the directions recommended 
by the committee appointed by the American Chemical Society? In 
the determination of volatile combustible matter it has been found that 
these give very inaccurate results. ‘The committee state that the most 
serious objection brought against their method is that the rapid heating 
causes mechanical loss in the case of certain non-coking coals; that no 
evidence has been given as to the amount of such loss, while in the light 
of certain experimental determinations which are described, they state 
that the loss can only have been insignificant. It has been observed in 
this laboratory that the error from this source on our coals is very large, 
possibly amounting to a few per cent in some cases. It has also been 
found that this could be largely, if not entirely, eliminated by expelling 
the moisture and most of the volatile matter at a low heat before sub- 
jecting to “the full flame of a Bunsen burner for seven minutes.” Four 
to five minutes’ gentle heating is sufficient to do this. With this excep- 
tion the official method has been followed in detail. 
8 A value of 600° was approximated by judging the color of the iron tube in 
the furnace. However, such results are unsatisfactory, as there is a large per- 
sonal equation to be considered as well as the degree of illumination under which 
the observation is made. The table used was that constructed by M. Pouillet, 
given in. Kent’s Mechanical Hngineer’s Pocket-book, New York (1903), page 454, 
as follows: 
Color of iron. Séarees Cc. 
Incipient red heat 625 
Dull red heat 700 
Incipient cherry-red heat 800 
Cherry-red heat 900 
Clear cherry-red heat 1,000 
Deep orange heat 1,100 
Clear orange heat 1,200 
White heat 1,300 
Bright white heat 1,400 
Dazzling white heat pe 
1,600 
7, Am. Chem. Soc. (1899), 21, 1116-1132. J. Soc. Chem. Ind. (1900), 
19, 174. ; 
