SECTIONAL TRANSACTIONS.— A. 539 



However, Kamerlingh Onnes made a very interesting experiment with a supra- 

 conductor. A wire of copper covered with a very thin sheet of tin and placed in a 

 cryostate at the temperature of liquid helium shows supraconductivity at the ordinary 

 threshold value of the temperature. Hence the electrons do not pass from the tin 

 to the copper. There is no internal friction, as in a gas. 



There are still new phenomena di£Scult to understand under the assumption of 

 a gas of electrons. 



It is a known fact that when a supraconductor is cooled down to the temperature 

 of liquid helium a magnetic field may restore the ordinary qualities of conduction in 

 the metal. Later experiments show that this phenomenon is a real hysteresis 

 phenomenon. The resistance of the metal does not return at the same value of the 

 field at which it vanishes. Now a phenomenon of hysteresis is something of a memory 

 of matter. And all things occurring in time must statistically occur also in the 

 volume of the metal. It is necessary that quite a lot of electrons be involved in the 

 process of hysteresis. It seems possible that the hysteresis, is made by a row of 

 electrons passing rows of atoms. 



Why do the electrons pass in rows in a supraconductor ? I think the phenomenon 

 of supraconductivity has something to do with the zero-point energy. Nature gives 

 in a supraconductor an example of a hidden and profound synchronism only possible 

 with very regularly built atoms. 



Dr. L. F. Richardson, F.R.S. — The Deferred Approach to the Limit. 



In the Infinitesimal Calculus the passage to the limit comes early, namely in the 

 definition of a derivative or integral, and the solution of special problems follows 

 later. As a typical special problem let us consider a differential equation holding 

 throughout a range of the independent variable together with boundary conditions 

 at one or both ends of the range. 



When considerable difficulties have been encountered in special problems treated 

 by this ' previous passage to the limit,' it has been customary instead to pass towards 

 the limit concurrently with the solving of the special problem, by employing finite 

 differences of suitably high order, in the manner described in books on Interpolation. 

 In this ' concurrent approach to the limit ' the difference-equations are always of 

 higher order than the differential-equation which they replace. 



The object of the present note is to call attention to a third process in which the 

 passage towards the limit is deferred until after the special problem has been replaced 

 by a difference-problem of the same order (except for the first step of some solutions) 

 and solved for two or more sizes of the difference of the independent variable. The 

 advantage of the ' deferred ' in comparison with the ' concurrent ' method lies in 

 the lower order of the difference-equation. The advantage becomes important when 

 the given problem becomes complicated. The ' deferred approach ' is an easy routine 

 which should attract those who want numerical solutions of problems that are 

 numerically definite. It may be performed by plotting the numerical results, obtained 

 separately for several different sizes of differences of the independent variable, against 

 the squares of those sizes. Ordinarily the plotted points are nearly on a straight 

 line which cuts one axis close to the desired limit. 



At the Meeting illustrations were given, one of which, concerning a fourfold 

 integral was about to appear in the Philosophical Magazine, under a title beginning 

 ' The Amount of uniformly diffused Light . . . .' 



An extended discussion is to be found in Phil. Trans. A., vol.ccxxvi, pp. 299-361, 

 by L. F. Richardson and J. Arthur Gaunt. It is sometimes possible to retain a 

 variable parameter throughout, as H. Jeffreys has shown {Phil. Mag., October 1926). 



Dr. G. Gkeen. — The Condenser Telephone. 

 Prof. J. G. Gray. — Four new Gyroscopic Tops. 



Dr. A. Ferguson and Mr. J. A. Hakes. — The Simultaneous Determination 



of Surface Tension a)id Density. 



If a capillary tube of r<idius r is immersed vertically to a depth A in a liquid of 

 density p and siarface tension y, the value of y may be deduced from observation of 



