1898] SOME MORE ROWING EXPERIMENTS 91 
produced by simply turning the oar, is a horizontal circle, which 
becomes a straight line when the diagram strip is unwrapped. 
At the conclusion of a stroke, when the pencil has reached its 
extreme position to the right, the oar leaves the face and rests 
against the opposite thowl, only touching the instrument at the 
rounded end of the face to carry it into position for the next stroke. 
During this movement the pencil travels over the base line, and a 
lever V engages a spoked wheel 0, which gears into the right hand 
cylinder and winds the strip into a new position, As the winding 
ceases, a cam-wheel P is rotated; this drives a plate attached to 
the bottom of the drum outwards, and thus lifts the pe! off the 
strip and the winding lever out of gear. 
During the stroke the wheels are not engaged, since the levers 
share the backward movement of the face when the spring is com- 
pressed and pass clear of the wheels. While the cam holds the 
plate out no diagrams are drawn and no winding goes on.. Each 
forward swing carries the cam wheel forward one tooth, until at the 
fifth stroke the pencil again reaches the paper, another diagram is 
drawn and the strip is once more wound. Two cords place the 
starting and stopping of the record under the control of the 
coxswain. The upper part of the instrument is protected from 
splashing or rain by a cover. 
Knowing the diameter of the drum, and the strength of the 
spring, the diagram gives a result in pounds pressure per 
degree of turn of the oar. In order to express the result in 
foot-pounds of work, the ratio of the pull on the handle of the oar 
to the corresponding pressure on the rowlock must be known, and 
the distance through which the handle moves per degree of turn. 
These two constants are found from the in and outboard dimensions 
of the oar, and the position of the centre of pressure of the water 
on the blade.t The spring was calibrated in the Millard Laboratory, 
Oxford, by kind permission of Rev. F. J. Jervis-Smith. With the oay 
and spring used in the experiments, a height on the diagram of 1” 
represents a pull of 220 pounds, and 1 sq. inch represents 377 
foot-pounds.? 
When the indicator is fully open it is usually convenient to 
have the spring under some initial compression. This is estimated 
by noting the turns given to the nut F after it has just touched the 
spring, and allowed for by measuring heights from a new base-line 
0”06 per turn below the actual one. Experiments in calibrating 
the spring showed that one end did not completely ‘bed’ itself 
until some pressure was exerted. The error due to this is practi- 
cally eliminated by raising the base line 0”-03. 
1 Neglecting the moment of inertia of the oar about its button. See Appendix I. 
2 See Appendix I. 
