1879.] Mr Hart, On two models of parallel motions. 815 



vescence of gas from the Hydrogen pole, and a throw of 18" on 

 raising and — 1" on lowering. 



The E.M.F. was augmented to that of five Daniells in order to 

 obtain if possible a development of Oxygen on the movable 

 mercury surface, the only result however was a film, presumably 

 of sulphate, on the surface. 



The apparatus was left with one Meidinger cell in circuit till 

 the following day, when the effect of raising and lowering of the 

 globe was no longer appr-eciable. The E. M. f. was accordingly 

 diminished gradually in order to ascertain if the effects would 

 increase in the same way as they had diminished. The effect 

 however was not appreciable until. the E.M.F. was reduced to '25 

 Daniell, it then became sensible and increased slowly until at an 

 E.M.F. "05 it suddenly increased, throwing the needle against the 

 stops on either side according as the motion of the globe was up 

 or down. The experiments were repeated with the same appa- 

 ratus, refilled with fluid, giving strictly similar results. 



It appears then that there is no definite E. M. F., which by 

 acting between the surfaces of mercury will cause the movable 

 surface to be ' neutral ' to the fluid. If the surfaces be left with a 

 current passing from the fixed to the movable one, the throw of the 

 galvanometer on the motion of the electrode becomes smaller the 

 longer the current flows ; and generally one may conclude that 

 the diminution of the throw of the galvanometer is, in part at any 

 rate, dependent on the solution of mercury of the movable 

 electrode (which appears always to be dissolved when a current 

 passes), and not entirely upon the E.M.F. of polarization. 



(3) Mr Harry Hart, M.A., On tiuo models of parallel 

 motions. 



Since the discovery by Peaucellier in 1864 of a method of 

 producing an exact parallel motion by means of a combination of 

 jointed rods, many solutions of the problem have been given, of 

 which the two following are theoretically, if not practically, the 

 most simple, inasmuch as they require the use of five links only, 

 whereas in all other cases seven at least are employed. 



The first of these (which I described in the Messenger of 

 Mathematics, vol. iv. page 82) is based on the fact that if a circle 

 be inverted with respect to a point the inverse curve is in 

 general a circle also, but becomes a straight line when the first 

 circle passes through 0. 



This circle is described by means of a link capable of free 

 motion about a fixed point 0', any point P' on it moving in conse- 

 quence in a circle whose centre is 0'. It remains therefore to 

 describe the Reciprocator or system of links which enables us to 

 produce motion in a curve inverse to any given one. 



Vol. III. Pt. vii. 24 



