216 
and there is a further error introduced by the combustion 
of the copper of the tube — the tube, becoming heated to 
bright redness, becomes covered with a scale of the oxide ; 
and, last of all, the potassium chloride remaining from the 
combustion absorbs a considerable quantity of heat in dis- 
solving. Mr. Lewis Thompson recognised that his process 
did not register the total quantity of heat produced, and he 
says it is necessary to add 10 per cent, to the result 
obtained. According to Mr. Thomas’s experiments, if the 
coal be of a graphitic nature, the chlorate and nitrate 
mixture will not burn it completely; yet the process is 
employed by the Italian Government as a test, and by 
railway and other companies in England. I have further 
found it difficult to obtain concordant results from this 
process with the same coal tested at different tim es. 
The process which I have devised consists in burning the 
coal in oxygen. In a short communication which I brought 
before the Physical and Mathematical Section of this Society 
on the unsatisfactory nature of Lewis Thompson’s process, 
our excellent Vice-President, Dr. Joule, suggested the idea 
to me of burning the coal in pure oxygen. I endeavoured 
to accomplish this in many different ways, but failed, a 
certain amount of unconsumed carbon, under some con- 
ditions, being liberated which adhered to the sides of the 
vessel containing the oxygen, whilst under other conditions 
the carbon consumed very slowly and usually imperfectly, 
leaving some unconsumed carbon behind. I ultimately 
succeeded in completely consuming the coal within a few 
minutes, and measuring the heat produced therefrom by 
the apparatus and method which I give, as follows. 
I took the ordinary stand, with brass springs attached, 
which was used for the Lewis Thompson apparatus. In it 
I fitted the bowl of an ordinary clay tobacco pipe, rather 
less than Jin. internal diameter by IJin. long. This I used 
as a stand for a small platinum crucible Jin. diameter by 
