therefore, the chamber containing the grate should be completely surrounded 

 by water, and should be below the level of the water in the boiler. The 

 magnitude of the surface exposed to radiation should be as great as is consistent 

 with the whole magnitude of the machine. In the next place, it is necessary 

 that the heat, which is absorbed by the air passing through the fuel, and keep- 

 ing it in a state of combustion, should be transferred to the water before the 

 air escapes into the chimney. Air being a bad conductor of heat, to accom- 

 plish this it is necessary that the air in the flues should be exposed to as great 

 an extent of surface in contact with the water as possible. No contrivance 

 can be less adapted for the attainment of this end than one or two large tubes 

 traversing the boiler, as in the earliest locomotive engines : the body of air 

 which passed through the centre of these tubes had no contact with their 

 surface, and, consequently, passed into the chimney at nearly the same tem- 

 perature as that which it had when it quitted the fire. The only portion of air 

 which imparted its heat to the water was that portion which passed next to the 

 surface of the tube. 



Several methods suggest themselves to increase the surface of water in con- 

 tact with a given quantity of air passing through it. This would be accom- 

 plished by causing the air to pass between plates placed near each other, so 

 as to divide the current into thin strata, having between them strata of water, 

 or it might be made to pass between tubes differing slightly in diameter, the 

 water passing through an inner tube, and being also in contact with the exter- 

 nal surface of the outer tube. Such a method would be similar in principle to 

 the steam-jacket used in Watt's steam-engines, or to the condenser of Cart- 

 wright's engine. But, considering the facility of constructing small tubes, and 

 of placing them in the boiler, that method, perhaps, is, on the whole, the best 

 in practice ; although the shape of a tube, geometrically considered, is most 

 unfavorable for the exposure of a fluid contained in it to its surface. The air 

 which passes from the fire-chamber, being subdivided as it passes through the 

 boiler by a great number of very small tubes, may be made to impart all its 

 excess of heat to the water before it issues into the chimney. This is all 

 which the most refined contrivance can effect. The Rocket engine was 

 traversed by twenty-five tubes, each three inches in diameter ; and the principle 

 has since been carried to a much greater extent. 



The abstraction of a great quantity of heat from the air before it reaches the 

 chimney is attended with one consequence, which, at first view, would present 

 a difficulty apparently insurmountable ; the chimney would, in fact, lose its 

 power of draught. This difficulty, however, was removed by using the waste 

 steam, which had passed from the cylinder after working the engine, for the 

 purpose of producing a draught. This steam was urged through a jet presented 

 upward in the chimney, and driven out with such force in that direction as to 

 create a sufficient draught to work the furnace. 



The importance of this subject will be understood, when it is considered 

 that the only limit to the attainment of speed by locomotive engines is the 

 power to produce, in a given time, a certain quantity of steam. Each stroke 

 of the piston causes one revolution of the wheels, and consumes four cylinders 

 full of steam : consequently, a cylinder of steam corresponds to a certain 

 number of feet of road travelled over : hence it is that the production of a rapid 

 and abundant supply of heat, and the imparting of that heat quickly and effectu- 

 ally to the water, is the key to the solution of the problem to construct an engine 

 capable of rapid motion. 



The method of subdividing the flue into tubes was carried much further by 

 Mr. Stephenson after the construction of the Rocket ; and, indeed, the princi- 

 ple was so obvious, it is only surprising that, in the first instance, tubes of 



