Proceedings of the Polytechnic Association. 845 



to thirty per cent of the steam is unacconnted for by the indicator at 

 full stroke ; and as high as sixty to eiglity per cent when the steam 

 is expanded considerably. Large engines show a small discrepancy 

 at full stroke, which rises to thirty, and often fifty per cent, with 

 shorter admissions. The best examples of the English double cylin- 

 der pumping engines, with steam-jacketed cylinders, use thirty-three 

 per cent more steam than is shown by the indicator on the cylinders. 

 This method of determination is, therefore, absolutely worthless for 

 our purpose, as it furnishes no basis for reliable comparative tests. 

 These discrepancies show us where a great loss takes place in the use 

 of the steam engine. They have been ascertained, in practice, by 

 indicating the engine and measuring the water pumped into the 

 boiler, and evaporated there, to furnish steam. In other cases, the 

 exhaust steam of the engine has, by surface condensation, been 

 reduced to water, and its quantity determined by measuring or 

 weighing it. The weight of feed-water, or what is the same thing, 

 of steam used in any case, to produce a given power, may, by either 

 of these plans, be ascertained with scrupulous accuracy ; and if the 

 coal be weighed at the same time, the evaporative efHciency of the 

 boiler can also be determined, and the excellence of both engine and 

 boiler be detected and credited aright. 



In addition to the standards above given, expressing the economy 

 of tlie engine, others of special application are used, which give the 

 cost in terms of that for which money is paid, namely, the coal, and 

 the result in that which returns the money. For instance, the miller 

 speaks of the number of pounds of coal it requires to grind a barrel 

 of flour; a thing, by the way, which may depend as much upon the 

 condition of the mill as of the steam machinery. Locomotives are 

 rated by the number of pounds of coal or coke burned per ton per 

 mile. So, also, what is known as the " duty " of a pumping engine, 

 is the number of foot-pounds of w^ork derived from the consumption 

 of a certain quantity of coal. 



Having discussed the various measures and means that may be 

 employed for our purpose, we desire next to select such as Avill be 

 useful in particular cases, and show their practical application, which 

 leads us to 



The Methods of coNDUCTiisrG Experiments.— I. Testing Boileks. 



The power of an engine can never exceed that of the boiler which 



furnishes it with steam; hence, it is eminently proper that we 



