182 1/nlett, Mack and Smyth — 



the gas in the apparatus, the volume readings were very 

 constant and precise. 



The gas volumes were calculated to standard condi- 

 tions, after allowing for the (volume) position of the 

 mercury in the manometer. From these values we 

 readily ohtained the volume of gas liberated per gram of 

 coal for each time interval, and the plot gave information 

 as to whether measurable decomposition of coal sub- 

 stances has taken place during the heating, while the 

 total moisture liberated during the heating was deter- 

 mined at the end of the experiment. It was found that 

 90 minutes heating was sufficient for our purpose and on 

 inspection of these isothermal curves, figs. 5, 6, 7 and 8 

 (the temperature of each curve and total moisture liber- 

 ated are indicated on each curve), it was seen that the 

 gas generally ceased to be liberated after 40 to 60 minutes 

 and the curves from this time tended to run parallel to 

 the time axis, but at the higher temperatures there are 

 distinct indications of a continual liberation of gas as 

 long as the heating of the coal was continued. This is 

 more marked with some coals than with others and 

 clearly indicates decomposition of coal substances. 



The moisture values are seen to be higher, the higher 

 the temperature of the experiment, which is in accord 

 with the conclusions we drew from our previous work 

 on the condition of water in coal. The temperatures 

 used were those of constant boiling liquids and did not 

 give equal temperature intervals, so we plotted water 

 losses against the corresponding temperatures and 

 secured the curves of fig. 9, which represent the water 

 losses when these coals were heated at the various tem- 

 peratures for 90 minutes in a vacuum which was exceed- 

 ingly free from water vapor. Longer heating did not 

 make a noticeable difference in the results. Between 

 184° and 218° the curves show a noticeable increase in 

 the percentage of water lost in all cases, which we regard 

 as the typical behavior of the loss of water from the 

 large surfaces of the colloidal coal. Above 218° this loss 

 is still noticeable but becomes so small in some cases that 

 the curves are nearly parallel with the temperature axis. 

 Above some 275° we soon strike a more or less pro- 

 nounced break in the curves which we attribute to the 

 decomposition of coal substances with the formation of 

 water. These same reactions undoubtedly take place at 



