ON A CONSTANT INDICATOR FOB STEAM-ENGINES. 103 



2ndly. To the continual variation of the level of the water in the well, 

 which variation amounts, according to the statement of the engineer, to 16 

 inches, and often continues for days together. The variation in the registra- 

 tion which results from this cause amounts to 1-0274. Xl| = 1 -284.2 unit of 

 work per square inch of the piston. 



The possible variation in the registration due to these two causes amounts, 



therefore, 



Units per square inch. 



By variation in length of stroke, to 2'543 



By variation in water-load, to .... 1*284. 



Total 3-827 



That further variation in the work of the engine which is due to a different 

 state of the packing of the piston and a different supply of oil, it is of course 

 impossible to estimate ; but it is not too much to assign to it a possible varia- 

 tion of T %ths of a unit of work per square inch of the piston per stroke. This 

 small variation, added to that shown to be due to the length of the stroke and 

 the water-load, accounts for the entire amount of the greatest variation in the 

 daily registration of the indicator, as shown by the table. 



If the friction of the indicator be neglected altogether, the work shown by 

 it per stroke will be 120'2944, whilst the work by experiment is 120-574. 

 The difference is on this supposition, therefore, only -27, or about one quarter 

 of a unit of work out of 120| units. 



In respect to these results it is sufficient to say, that no others ever obtained 

 by any of the instruments applied for a similar purpose have approached to 

 them in the accuracy of their registration of the work, although that regis- 

 tration, which was by those instruments limited to a single stroke, has been 

 by this extended continuously over 179,000 strokes; and that Mr. Wicksteed 

 has expressed an opinion that the results given by the indicator are probably 

 nearer to the true work of the engine than his own experimental results. 



These experiments having sufficiently established the general accuracy of 

 the instrument and its working qualities, as applied to single acting stationary 

 engines, the Committee became desirous of trying it on the marine engine, and 

 the directors of the Great Western steam-ship having kindly given their con- 

 sent, it was fixed upon the engines of that vessel when about to proceed on 

 her first vovage in the present year. 



Unfortunately no trial could be obtained of the particular expedients adopt- 

 ed for this new application of the instrument until the vessel was actually 

 upon her voyage, and the Committee regret to state that this costly experi- 

 ment was rendered useless for their purpose by circumstances which would 

 have been obviated had such a trial been possible, and which would have 

 been wholly unimportant and soon remedied on shore. Of these the princi- 

 pal was the accidental unscrewing of one of the pistons from the extremity 

 of the piston rod. 



The knowledge which would be supplied by a series of registrations of such 

 accuracy as those given by the indicator, as well in regard to the duty of 

 marine engines as in respect to the general conditions of the resistance op- 

 posed under different velocities to the motion of a vessel of the dimensions of 

 a steam-ship, is of vast importance practically and scientifically ; it is there- 

 fore desirable to repeat this experiment under more favourable circumstances. 



The attention of the Committee was directed by the Association, in the 

 second place, to an application of the admirable chronometrical instrument 

 suggested by M. Poncelet, and contrived and applied by M.Morin, to an ad- 

 measurement of the velocity of the piston of the steam-engine at different pe- 



