i82 NATURAL SCIENCE. March. 



area enclosed by the true base-line, the ordinate at the end of the 

 curve, and the curve produced down to the base-line. The two 

 corrections together seldom exceed i per cent, of the whole. 



The reduction factor is obtained as follows. To know the relation 

 between pull on the handle of the oar and pressure on the rowlock, the 

 " centre of pressure " of the blade on the water must be found. This 

 was obtained by calculating the statical centre of pressure and allow- 

 ing for the disturbing rotation (the position is thus known within 

 2 per cent, or 3 per cent.). The required result then follows from the 

 principle of the lever. The ratio is found to be '697. Finally, 5 degs. 

 turn of the oar represents a motion of the handle of -293 feet. Conse- 

 quently, multiplying the results in Table I. by '293 x '697 = "203 gives 

 the result in foot-pounds. 



The present form of indicator has various sources of error, all of 

 vvrhich have been considered and found to have only a small effect on 

 the result. The effect of friction was found at a pressure of 200 lb. 

 to be about i per cent., and even this largely cancels, since the effect 

 at one end of the stroke is opposite to that at the other. The com- 

 pression of the spring is not strictly proportioned to the pressure, 

 owing to the peculiar form of the instrument, but the variations were 

 too small to be observed in testing the springs. Finally, at the 

 commencement of the stroke, the outward pressure of the button of 

 the oar very slightly exaggerates the " beginning." This has no 

 appreciable effect on the whole work done, and in comparing different 

 curves the effect on all is similar. 



To pass now to some of the results obtained with the indicator. 

 The first point to be noticed is that each oarsman has a marked 

 individuality, both with regard to style and amount of work done. 

 This fact makes generalisations difficult, and reliable figures of the 

 average value of a man-power can only be obtained by a very 

 extended series of observations. Table II. exhibits some of the 

 results already obtained. The division into "rowing" and "paddling" 

 is necessarily somewhat arbitrary. In the latter class are included 

 all cases in which the rower was not working his hardest. 



It has been found possible to indicate three or four strokes on the 

 same card, and thus to obtain various stroke-diagrams in one 

 piece of rowing. The stroke-rate was counted in the case of 

 "rowing," but only estimated in the second class, and the 18 rate is 

 possibly in cases 3 and 7 in excess or defect. 



The results obtained show that in a " tub pair," in which it is 

 only possible to row about 24 strokes to the minute, the horse-power 

 at high pressure and short duration varies between '23 and -40. It 

 should be said that these results have been obtained from oarsmen, 

 all of moderate experience. All but one have rowed in their College 

 Eight, and two— the reader will easily see which — have rowed in the 

 'Varsity Boat. 



The diagrams reproduced will show the use of the indicator in 



