Experiments refpccling Heat. 32$ 



allowed to run itfelf empty, othervvife the atmofpheric air 

 getting into the bottom pipe, a portion of it might be forced 

 down with the next fupply of water. The receiver being 

 charged with gas, fhnt the cocks c and b for the prefent. 



Should any experiment require more gas than the receiver 

 will contain, the flream may be kept up direct from the cafk 

 by opening the communication UT ba, and taking care to 

 keep the funnel W well filled with water. Another cafk 

 may be joined to the firft by a mode of communication fo 

 obvious that we need not defcrine it; fo that oxygen gas may 

 eafily be fupplied for a length of time fufficient to melt down 

 completely the whole internal arrangements of any furnace. 



In none of the experiments, except the 3 2d, were the table- . 

 furnace and its appendages employed. In the firft feven, a 

 fimple blow-pipe was joined to the tube of the gafometer, 

 and the fubftance to be operated upon, laid on a piece of 

 ignited charcoal, was expofed to the iffuing ftream of gas. 

 It was obferved in the courfe of thefe experiments that the 

 fubftances feemed fometimes to be partially cooled, by the 

 ftream falling on them while the charcoal was moving about 

 to expofe the different parts round the ftone to the action of 

 the gas. To obviate this inconvenience, the double blow- 

 pipe alreadv mentioned was conftrufted, the nozles of which 

 formed fuch an angle as to make the ftreams of blaft crofs 

 each other. From this it is obvious, that the charcoal placed 

 at any diftance from the nozles fhort of that where the ftreams 

 would crofs, received two ftreams of gas; and that the nearer 

 the charcoal was brought to the blow-pipe, the wider was 

 The fpace between the centres of the fpots on which the blaft 

 Kll. By this means it became eafy to expofe the gems and 

 Other fubftances to the full action of the caloric liberated bv 

 the decompofition of the oxygen gas, without putting them 

 in the way of the undecompoled ttreams. 



This double blow-pipe is represented in fig. 2. The 

 part A, which joins the tube of the gafometer, turns air- 

 light in a collar B, on the end of the tube BC. On BC. are 

 two brafs boxes d,e, into which are fitted the tubes f,g, 

 which alfo turn in fockets air-tight for the purpofe of en- 

 abling the operator to move the blow-pipes b,i, nearer or 



Vot, VIII. U u fvrther 



