88 THE ROYAL SOCIETY OF CANADA 



tight with an asbestos-copper gasket. The inlet tube was dispensed 

 with, and a pressure gauge and rehef valve connected to the outlet 

 tube. 



The coal briquettes employed in each test were weighed out into 

 a 10 gram capacity quartz crucible which fitted inside the iron retort. 

 The heating was done by immersing the retort in a bath, which for 

 tests up to 300°C. was of oil and for those above that temperature of 

 lead. The lead was contained in a 4" length of 4" iron pipe with a 

 cap end, and was heated in a gas fired furnace which gave a very uni- 

 form temperature throughout the bath, and which permitted of rapid 

 heating and easy control. The temperature was followed by two 

 pyrometers immersed in the lead. 



The pyrometers employed were calibrated from time to time by 

 means of the freezing points of lead and zinc. The recorded temper- 

 atures are probably correct to within 5°C. 



Procedure: — The general procedure, modified in the particular 

 cases noted, was as follows. From 3 to 10 grams of the briquetted 

 sample was weighed out into the quartz crucible, this was placed in 

 the iron retort and the top clamped down. A gentle stream of dried 

 coal gas was passed in through the inlet tube to displace the air, the gas 

 was then cut off and the retort immersed in thé bath of oil or lead, 

 previously heated to almost the desired temperature. The progress 

 of carbonization was watched by causing the evolved gas to bubble 

 through a little water; it was thus found that 1| hours was sufficient 

 to complete the operation. In every case after the lead bath had been 

 finally adjusted to the desired temperature, this temperature was 

 maintained for at least half an hour. At the end of the experiment 

 the retort was taken out of the bath and cooled as rapidly as possible, 

 cold water being run over it to complete the cooling. A little dried 

 coal gas was also passed through the apparatus during this period to 

 prevent the oxidation of the hot product by air drawn in as the gases 

 contracted. When quite cold the retort was opened and the quartz 

 crucible weighed to determine the loss in weight of the briquettes 

 during the experiment. The calorific value of the carbonized briq- 

 ettes was then determined fcy means of a Riche adiabatic calorimeter 

 using a platinum resistance thermometer. 



The water content of the original briquettes was determined 

 by heating for one and a half hours in a toluol oven in a stream of 

 carbon dioxide. By means of this the results obtained above were 

 recalculated to a moisture-free basis. 



For reasons given below it was found necessary to standardize the 

 conditions, especially the time factor, in all operations with the dried 

 or carbonized residue. The quartz crucible and contents were always 



