438 PHILOSOPHICAL TRANSACTIONS. [aNNO 1754. 



spindle aiul shoulder to be of a less diameter. To break the jerks of the waves, 

 next to the end of the spindle he fixed a spiral spring of wire, to which the cord 

 was fastened; which, by this means, was capable of playing backwards and for- 

 wards, and giving way to the irregularities of the sea: and lest the plate should 

 lay fast hold of any thing, or any extraordinary jerk should damage the spindle 

 or spring, a knob or button was fastened on the cord, at a small distance from 

 the spring, which stopped on a hole in a piece of wood, and prevented the spring 

 from being pulled out to above a certain length ; so that all addition of force, 

 beyond this, could only tend to break the cord, and carry away the plate. The 

 spindle, being thus guarded from accidents, will allow of a still further diminution 

 of its size; so that, at last, he ventured to make the spindle-pivot no more than 

 -jV of an inch diameter, and that of the shoulder J-, being of tempered steel, 

 and sufficiently smooth. The hole, in which the pivot, and against which the 

 shoulder worked, was of agate likewise, well polished. 



Being thus provided, in May 1751} he procured a boat, on the serpentine 

 river in Hyde-park, to try how far the turns of the machine would be consistent 

 with themselves, when the same space was measured over with different velocities. 

 The course was determined at each end, by observing the coincidence of two 

 trees, in a line nearly at right angles to the river. We however rowed beyond 

 the mark, that the machine might be in full play when the course was begun: 

 the spindle was stopped at the beginning and end; the numbers read off were as 

 follow : 



The space between the marks was, by estimation, about half a mile. 

 1st rowing up the river, in 11 min. the plate made 6l5 revol. 



2d down 14 645 



3d up 18i 6l2 



4th down Q-J- 603 



5th up 18 620 



6th down 10 600 



It is observable, that the greatest difference among the above observations, is 

 between the 2d and 6th, being 645 and 60O; the difference being about a 4th 

 part of the whole; the times being 14 minutes and 10, both in going down the 

 river; whereas those observations, which differ most in point of time, viz. the 

 3d and 4th, being performed in I8-1- minutes, and 9-r minutes, have their revo- 

 lutions more nearly alike, being 6l2 and 603; which differ only by a 68th of 

 the whole. From these observations he was led to think, that the different ve- 

 locities, with which a vessel moves forwards, would make no material difference 

 in the number of rotations of the plate; or at least that those differences would 

 be less than the irregularities arising from other causes, even in trials nearly 

 similar. 



