]\gt ; Electrical on Thin Anchor-Ring .
It will be seen above that the heavy barium crown glasses are those which show the greatest variation in homogeneity in the same melting .
My thanks are due to Sir David Gill for his great kindness and help in arranging the paper .
Etectrical brations on Thin Anchor-Ring .
By Lord RAYLEIGH , O.M. , F.B.S. Received Read June Although much attention has been bestowed upon the , subjec of electric oscillations , there are comparatively few examples in which definite mathematical solutions have been gained .
These problems are much simplified when conductors are supposed to be perfect , but even then the difficulties usually remain formidable .
Apart from cases where the propagation may be regarded as being in one dimension , * we have Sir J. Thomson 's solutions for electrical vibrations upon a conducting sphere or cylinder .
But these vibrations have so little persistence as hardly to deserve their name .
A more instructive example is afforded by a conductor in the form a circular ring , whose circular section is supposed small .
There is then in the neighbourhood of the conductor a considerable store of energy which is more or less entrapped , and so allows of vibrations of reasonable persistence .
This problem was very ably treated in 1897 , but with deficient explanations .
S Moreover , Pocklington limits his detailed conclusions to one particular mode of free vibration .
I think I shall be doing a service in calling attention to this investigation , and in exhibiting the result for the radiation of vibrations in the higher But I do not attempt a complete re-statement of the argument .
Pocklington starts from Hertz 's formulae for an elementary vibrator at the origin of co-ordinates , ( 1 ) where , ( 2 ) in denote the components of electromotive intensity , is ' Phil. Mag 1897 , vol. 43 , p. 125 ; 1897 , vol , 44 , ; ' Scientific Papers , ' vol. 4 , pp. 276 , 327 .
'Recent Researches , ' 1893 , SS301 , 312 .
'Camb .
Proceedings , ' 1897 , vol. 9 , p. 324 .
S Compare W. .
Orr , ' Phil. Mag 1903 , vol. 6 , p. 667 .
Lord Rayleigh .
[ May 24 , the period of the disturbance , and the wave-length corresponding in free aether to this period .
At a great distance from the source , we have from ( 1 ) .
( 3 ) The resultant is perpendicular to , and in the plane contaimng and Its magnitude is , ( 4 ) where is the angle between and The required solution is obtained by a distribution of elementary vibrators of this kind along the circular axis of the , the axis of the vibrator being everywhere tangential to the axis of the ring and the coef of intensity proportional to where is an integer and defines a point upon the .
The calculation proceeds in terms of semi-polar co-ordinates , the axis of symmetry being that of , and the origin being at the centre of the circular axis .
The radius of the circular axis is and the radius of the circular section is .
very small relatively to The condition to be satisfied is that at every.point of the surface of the ring , where , the tangential component of , shall vanish .
It is not satisfied absolutely by the above specification ; but Pocklington shows that to the order of approximation required the specification suffices , provided be suitably chosen .
The equation determining expresses the evanescence of that tangential component which is parallel to the circular axis , and it takes the form , ( 5 ) where .
( 6 ) In ( 5 ) we are to retain the large term , arising in the integral when is small , and the finite term , but we may reject small quantities .
Thus Pocklington finds , ( 7 ) the condition being to this order of approximation the same at all points of a cross-section .
1912 .
] Electri , Vibrcations on a Anchor-Ring .
The first integral in ( 7 ) may be evaluated for any ( integral ) value of Writing , we have .
( 8 ) The large part of the integral arises from small values of .
We divide the range of integration into two parts , the first from to where though small , is compared with , and the second from to For the first part we may eplace by unity , and by .
We thus obtain .
( 9 ) Thus to a first approximation .
In the second part of the range of integration we may neglect in comparison with , thus obtaining .
( 10 ) The numerator may be expressed as a sum of terms such as , and for each of these the integral may be evaluated by taking , in virtue of Accordingly , ( 11 ) when small quantities are neglected .
For example , The sum of the coefficients in the series of terms ( analogous to ) which represents the numerator of ( 10 ) is necessarily , since this is the value of the numerator itself when .
The particular value of chosen for the division of the range of integration thus disappears from the sum of ( 9 ) and ( 10 ) , as of course it ought to do .
When , corresponding to the gravest mode of vibration specially considered by Pocklington , the numerator in 10 ) is Lord Rayleigh .
[ May 24 , and the value of the integral is accordingly To this is to be added from ( 9 ) making altogether for the value of ( 8 ) .
( 12 ) The second integral in ( 7 ) contributes only finite terms , but it is important as determining the ginary part of and thus the rate of dissipation .
We may write it , ( 13 ) where approximately .
Pocklington shows that the imaginary part of ( 13 ) can be expressed by means of Bessel 's functions .
We may take , ( 14 ) whence } .
( 15 ) ( 13 ) may be replaced by .
( 16 ) Now so that .
( 17 ) The imaginary part of ( 13 ) is thus simply .
( 18 ) A corresponding theory for the functions does not appear to have been developed When , our equation becomes 1\mdash ; 2 cos2 , ( 19 ) *Compare ' Theory of Sound , ' S302 .
Gray and Mathews , ' Bessel 's Functions , ' p. 13 .
1912 .
] Etectrical Vibrations on Thin and on the right we may replace by its first approximate value .
Referring to ( 2 ) we see that the negative must be chosen for and , so that .
The imaginary term on the is thus For the real term Pocklington calculates , so that , being written for , .
( 20 ) " " Hence the period of the oscillation is equal to the time required for a free wave to traverse a distance equal to the circumference of the circle multiplied by , and the ratio of the amplitudes of consecutive vibrations is 1 : or \ldquo ; For the general value of is eplaced by , ( 21 ) where is a real finite number , and finally .
( 22 ) The ratio of the mplitudes of successive vibrations is thus in which the values of can be taken from the tables ( see and Mathews ) .
We have as far as equal to 12:\mdash ; . .
It appears that the damping during a single vibraf , ion diminishes as ?
increases , vi the greater the number of subdivisions of the circumference .
An proximate expression for the tabulated quantity when is large may be at once derived from a formula due to Nicholson , * who shows that when and are large and nearly equal , is related to Airy 's integral .
In fact , , ( 24 ) 'Phil .
Mag 1908 , vol. 16 , pp. 276 , 277 .
Lord Rayleigh .
[ May 24 , so that .
( 25 ) If we apply this formula to .
we get as compared with the tabular It follows from ( 25 ) that the damping in each vibration diminishes without limit as increases .
On the other hand , the damping in a given time varies as increases indefinitely , if slowly , with We proceed to examine mole in detail the character at a great distance of the vibration radiated from the ring .
For this purpose we choose axes of and in the plane of the ring , and the coordinates of any point may also be expressed as .
The contribution of an element at is given by ( 4 ) .
The direction cosines of this element are ; and those of the due to it are taken to be .
The direction of this disturbance is perpendicular to and in the plane containing and the element of arc .
The first condition gives , and the second gives ; so that The sum of the squares of the denominators in ( 26 ) is Also in ( 4 ) ; and thus .nsin To these quantities the components due to the element are proportional .
Before we can proceed to an integration there are two other factors to be rarded .
The first relates to the intensity of the source situated at To represent this we must introduce .
Again , there is the question of phase .
In we have ; and in the denominator of ( 4 ) we may the difference between and S 0.13166 .
1912 .
] on a Thin Anchor-Ring .
Thus , as the components due to , we have , ( 29 ) with similar expressions for and corresponding to the right-hand members of ( 28 ) .
The integrals to be considered may be temporarily denoted by , where , ( 30 ) being written for .
Here .
and in this , if we write for We thus find , ( 31 ) where .
( 32 ) In like manner , .
( 33 ) Now When is even , the imaginary part vanishes , and ( 34 ) On the other hand , when is odd , the real part vanishes , and .
( 35 ) Thus , when even , and are both odd and and are both pure imaginaries .
But wheu is odd , and are both real .
As functions of direction we may take to be proportional to Whether be odd or even , the three components are in the same phase .
On the same scale the intensity of disturbance , represented by is in terms of , ( 36 ) V0L .
LXXXVII .
Lord Rayleigh .
[ May 24 , an expression whose sign should be changed when is even .
Introducing the values of and in terms of from ( 31 ) , ( 33 ) , we find that ' is proportional to From this it appears that for directions lying in the plane of the ring the radiation vanishes with .
The expression ( 37 ) may also be written or , in terms of , by ( 34 ) , ( 35 ) , ( 39 ) and this whether be odd or even .
The argument of the 's is Along the axis of symmetry the expression ( 39 ) should be independent of .
That this is so is verified when we remember that vanishes except .
The expression ( 39 ) thus vanishes altogether with unless , when it reduces to simply .
* In the neighbourhood of the axis the intensity is of the order In the plane of the ring the general expression reduces to , or It is of interest to consider also the mean value of ( 39 ) reckoned over angular space .
The mean with respect to is evidently By a known formula in Bessel 's functions .
For the present purpose ; and ( 41 ) becomes .
( 43 ) To obtain the mean over angular space we have to multiply this by , and integrate from to .
For this purpose we require an integral which does not seem to have been evaluated .
[ June 20.\mdash ; Reciprocally , plane waves , travelling parallel to the axis of symmetry and incident upon the ring , excite none of the higher modes of vibration .
] 1912 .
] Electrical on a Thin Anchor-Ring .
By a known expansion*we have , .
, whence .
( 45 ) for the integral on the left and thus , ( 46 ) as in ( 15 ) .
Thus the mean value of is , ( 47 ) as before .
In order to express fully the mean value of at distance we have to introduce additional factors from ( 29 ) .
If , and these factors may be taken to be .
The occurrence of the factor , where is , has a strange appearance ; but , as Lamb has shown , is to be expected in such cases as the present , where the vibrations to be found at any time at a greater distance correspond to an earlier vibration at the nucleus .
The calculations just effected afford an independent estimate of the dissipation .
The rate at which energy is propagated outwards away from the sphere of great radius , is , ( 48 ) Gray and Mathews , p. 28 .
' Enc .
Brit " " Wave Theory of Light Equation ( 43 ) , 1888 ; 'Scientific Papers , ' vol. 3 , p. 98 .
' Proc. Math. Soc 1900 , vol. 32 , p. 208 .
Mr. J. J. Manley .
Apparent Change in [ May 29 , , since ( the period ) , the loss of energy in one comoAlete vibration is given by .
( 49 ) With this we have to compare the total energy to be found within the sphere .
The occurrence of the factor is a complication from which we may emancipate ourselves by choosing great in comparison with , but still small enough to justify the omission of , conditions which are Teconcilable when is sufficiently small .
The mean value of at a small distance from the circular axis is .
This is to be multiplied by , and integrated from to a value of comparable with which need not be further specified .
Thus ( 50 ) and , ( o1 ) in agreement with ( 23 ) .
On the Weight during By J. J. , Hon. M.A. Oxon .
, Daubeny Curator , Magdalen College , Oxford .
( Communicated by Prof. J. H. Pointing , F.R.S. Received hIay 29 , \mdash ; Read June 27 , 1912 .
) ( AbstracG .
) In this communication are considered some of the causes which to apparent chances in the total mass of chemically reacting substances .
A short review of the work of and the conclusions drawn by the late Prof. Landolt is first given , and then we deal with the conditions under which one of Landolt 's final experiments was repeated by us .
Next our balance and reaction vessels are briefly described , together with the preliminary treatment to which the latter were subjected prior to their being weighed .
Following this , our first plan for weighing is outlined and then illustrated with the aid of data given in the form of tables and otherwise ; then , by means of an actual : example , the method adopted for calculating final results from our two of preliminary experiments is explained .

