Be-heating of a Jet of Air, 127 



the truth of the re-heating effect, and give numerical deter- 

 minations of its time, &c.* 



A, a tube, with a stop-cock, leads from a large reservoir 

 of compressed air kept at a uniform pressure and tempera- 

 ture. B, B, two small parallel branch-pipes from A, project 

 within the closed end of a large and long tube c, c, c, c, of very 

 thin zinc sheet, well c/oM^g? outside, with light non-conductors, 

 very small holes are drilled, in the two branch pipes B, B, 

 exactly facing one another, so that the jets from one branch 

 will meet and oppose those from the other branch, and thus 

 the heat will not be developed against the sides of the tube 

 c, c, c, c, but rather in the air itself. The tube f , c, c, should 

 be so large as to let the air move through it as slowly as is 

 compatible with preventing local currents and diffusion in 

 the opposite direction : it may be placed vertical, with the 

 open end upwards. Now, after the jets have been in action 

 until the temperature at each part of the apparatus has be- 

 come constant, the probable effect observed, in accordance 

 with the preceding theory, will be a gradual rise in tempera- 

 ture towards the open end of the tube, — from the maximum 

 cold at the end where the jets enter. Thermometer </, d, d, </, 

 placed at various parts of the tubes' length, will indicate any 

 such differences, and the rate at which the air travels through 

 the tube will give the time occupied in the development of 

 such temperature. The length of tube required for the ex- 

 periment can only be determined by experiment. 



To detect errors of experiment chiefly arising from im- 

 perfect clothing of the tube, a large orifice at the closed end 

 should be connected by a tube, with a source of very quiescent 

 air in a state of natural coldness, equal to that produced at 

 the cold end of the tube by the jets of compressed air ; the 



* It was probably the omission of this element in the calculation from Joule's 

 experiments on the dynamic equivalent of heat, that caused the assigned value 

 of heat to be (as Rankine suspects) too great, to an extent that cannot be ac- 

 counted for by errors of experiment. 



