836 Transactions of the American Institute. 



springs, and tlieir velocity, where the force i& applied. Neer's rotary 

 dj^amometer, on this principle, may be taken as a type of its kind, 

 and has i^iven general satisfaction. 



The accuracy of an instrument of this kind can easily be tested 

 by weighing the springs, measuring the distance from the center at 

 which they act, and correcting the dials accordingly. 



A good dynamometer is the only instrument that can be depended 

 upon to accurately measure the nseful work which an engine is 

 capable of performing ; still, the best instruments of this kind have 

 many disadvantages for every day practical use. 



In the first place, especially when great power is to be measured, 

 the dynamometer must needs be a large, heavy, and expensive 

 measuring machine, rather than an instrument ; consequently, but 

 few can afford to purchase it. The dynamometers, at present in the 

 market, are sold chiefly to establishments that rent rooms with power, 

 where a small machine can be shifted about the building in the night, 

 and so attached as next day to indicate the power use4 by one of the 

 tenants. 



The steam indicator, on the contrary, is neat and compact, and 

 can be easily applied to nearly every kind of steam engine. Its use 

 has, therefore, become so general, that it is acknowledged throughout 

 the world as the standard measure of the power of the steam engine. 

 This instrument is not perfect ; still we cannot point out another, fit 

 in every respect, to take its place. The dynamometer is the more 

 perfect instrument ; but we acknowledge that, in a majority of cases, 

 it is impracticable to apply it. Then, as we have proposed two 

 methods of investigation, one for careful scientific experiment, and 

 the other for practical and tolerably accurate comparison, we conclude 

 that the first would always require tlie use of the dynamometei', and 

 the latter whenever it is practicable to employ it. Generally, how- 

 ever, until a new instrument is perfected, we must use tlie indicator 

 alone in ordinary practical trials. 



When th^ indicated power alone is used, it is important to know 

 the probable friction of the engine, so that the net power, or tliat 

 portion available for useful work, may be estimated. A favorite 

 method is to take an indicator friction diagram from the engine, when 

 disconnected from its load, and running at its working speed. The 

 mean friction pressu^re thus obtained is supposed to be constant at all 

 loads. Hence it is usual to deduct from the indicated working pres- 

 sure the indicated friction pressure previously obtained, when the 



