]\gt ; Decay of Surface produced by Snperposed of Viscous Flnid .
By W. J. HARRISON , B.A. , Fellow of Clare College , ( Communicated by Prof. J. Larmor , Sec. R.S. Received May 13 , \mdash ; Read June 17 , 1909 .
) S1 .
In a paper published recently by Prof. R. W. Wood , on " " araboloid as a Telescope , notes the complete damping ipples in the mercury surface produced by a thin superposed layer of 'lycerine .
It has been ested to me by Prof. Larmor that an bpproximate estimate of the effect to be expected might be obtained from the esults in my paper " " The Influence of Viscosity on the Oscillations of uperposed Fluids .
In the case of a film of oil on the surface of water , in which the physical circumstances are diffel.ent on account of the extreme )enuity of the film , Prof. Lamb has made an approximate calculation of 'he effect in his Treatise on Hydrodynamics , by supposing the surface to be nextensible for small In my paper referred to , the approximations were obtained on supposithat the viscosity in both liquids was small\mdash ; the amplitude of the waves eing as usual very small .
In the present case , although the viscosity in nercury is very small , that in glycerine is comparatively large .
Hence a approximation is now required .
S2 .
The characteristics of the wave-motion at the free surface of a single iquid of great depth have been obtained for the two cases of small and of yreat viscosity respectively .
In the second case there are two modes of vibration , both being dead-beat , which the moduli of decay are and , respectively .
S For application to glycerine the viscosity is not reat to Dlake the approximation valid , except for wave-lengths less than 1 cm .
I ) thought it instructive to give case a special consideration , it is a side issue , as it affords an example of a deadbeat type of vibration to a propagated type as the is increased .
' The Astrophys .
Journ vol. 29 , No. 2 .
London Math. Soc. Proc Ser. 2 , vol. 6 , p. 396 .
Edition , 571 .
S Lamb , ' Hydrodynamics , ' 3rd Edition , p. 56 Mr. W. J. Harrison .
[ May 13 , For the type of wave-motion with which we are concerned , the velocity system is given by where ; is the kinematical coefficient of viscosity and is equal to , the period equation can be put into the form ( 1 ) When is great the term / can be ected in the determination of one mode , and the equation reduces to .
( 2 ) We obtain an admissible root , whence For glycerine , , hence the term ected above must be retained when the wave-length is greater than 1 cm .
equation ( 1 ) becomes .
( 3 ) It be solved by Descartes ' method , which leads to the cubic ( 4 ) The two values of , which are admissible , are obtained from the equation , ( 5 ) where is the positive root of ( 4 ) , and itself is taken positively .
It will be seen from the table given below that the rate of decay is always great , but that waves of reater than a certain limit , which lies between 2 and 3 , are When the is very reat , one root of ( 4 ) is iven by , where ; whence We obtain or Thus has the same form when is small as we obtain in the case of a slightly viscous liquid .
Cf .
Basset , ' Hydrodynamics , ' vol. 2 , p. 312 .
1909 .
] The Decay of Surface es , etc. Decay of Wave-motion on Glycerine .
is the wave-length in the modulus of decay in seconds .
the velocity of propagation in .
per second .
S3 .
In the general problem we shall consider the motion of two fluids , both of depth large compared with the wave-length of the motion , in the two cases : ( 1 ) both very viscous ; ( 2 ) the one slightly viscous , .
the upper very viscous .
The period equation for waves at the interface between a liquid of density and viscosity , and a superposed liquid of density and viscosity where , .
Viscosities both Larye .
S4 .
( a ) For the mode in which is large , that is of the order or , the equation determining is The ysis is much nplified by .
With this assumption , after writing , we derive the equation ( 6 ) With this ) be compared the equation ( 2 ) for a single liquid .
'London Math. Soc. Proc Ser. 2 , vol. 6 , p. 398 .
I quoted it again at length as there are two misprints as it has been previously printed .
Mr. W. J. Harrison .
[ May 13 , Some numerical solutions of ( 6 ) are given in the table below:\mdash ; When and are different , we can safely say that the modulus of decay is not greater than , where is the smaller of the two coefficients .
( b ) For the mode in which is small , we assume that is of the order .
We easily find that .
( 7 ) This corresponds to for a liquid .
Upper Liquid Viscous , Lower Slightly Viscous .
S5 .
( a ) For the mode in which is , writing in the terms of hest importance , we obtain the equation ( cf. 6 ) .
( 8 ) The nature of the solution of this equation is shown by the following lable : For the mode in which is small we have .
( 9 ) This can be obtained from ( 7 ) by S6 .
The Case of Thin Layer of superposed on Deep Mercury.\mdash ; The period equation for this case is given in S4 of my paper ( in the first term of the fourth column should be replaced by The effect of capillarity can be included by instead of in the second column , and instead of in the fifth and sixth terms of the fourth column , where is the surface tension of the glycerine-mercury interface , and is the surface tension of ycerine free surface .
For glycerine at , C. , and for mercury , , conse1909 .
] The Decay ojWaves , etc. quently we shall approximate by } ) at once .
This reatly simplifies the period equation , and does affect its validity except when the modulus of decay is very .
It will be further assumed that and are small , so that squares of these quantities will be ected .
With these assumptions the period equation becolues For less than a certain limit this equation has two roots and complex ones .
The two complex roots correspond to the mode in which the two liquids vibrate as a liquid without slip at their common surface .
The two ative roots show that the ways of the other mode are dead-beat .
It appears from the numerical calculation , in the case of mm. , that very nately the two negative roots are obtained from the last and the first two terms of the equation respectively .
Thus they are These approximations are not , vever , sufficiently good to be employed in the determination of the two roots , as they necessarily make the real part of these roots to be positive , equal to \mdash ; For reater than the limit mentioned above all the roots complex , and all the waves are propagated .
This limit is iven , in eneral , by the condition for equality of the loots of the last three terlns of the equation equated to zero .
For a -air surface , ; for a .
surface , It is assumed for the purpose of an illustrative calculation that at the common surface , while .
In the table the effect is exhibited of a layer of glycerine of 1 mm. depth on the modulus of deuay and velocity of ation of wave-motion at the surface of mercury .
The effect of a thin layer of glycerine is very iceable from the table .
The rate of decay of any arbitrary disturbance will depend , in general , on the rates of decay of those of the modes , into which it can be which are .
The other modes disappear instantaneously .
As a consequence of this the curvature of the upper glycerine surface is always the same as that of the mercury surface , a fact which may be of nportance in The Decay of , etc. the application to an optical instrument such as the one which gave rise to this investigation .
Prof. Wood made use of a depth of 4 mm. of lycerine ; a layer of this depth would give a much greater rate of decay .
is the wave-length in is the modulus of decay when capillarity neglected ; that when capillarity is included ; that of waves at the surface of clean mercury .
velocities of- the ponding cases .
*In the case of these quantities the approximation , on the supposition that is small , down ; but the rate of decay is in any case large .
The rate of decay cannot be determined without great labour .
If the wave-length is small compared with the depth , the rates of decay can be derived from ( 8 ) and ( 9 ) .
These approximations are applicable to glycerine , as the wave-length is small .
The moduli of decay for the two modes are and ; both of these are extremely small .
In these investigations the damping effect of the air has been neglected .
That this is very considerable in the case of water waves has been shown in my paper , to which reference has been made ; but to the great density of mercury its effect will be small , especially for small wave-lengths , such as those tabulated .
Its effect in the case of dead-beat modes is absolutely negligible .
It may be mentioned that the used by Prof. Wood was of comparatively small depth ; this would cause considerable damping in the case of the waves .
[ Note added 19.\mdash ; It will be noticed from the foregoing table that the modulus of decay , when the vibrations are influenced by the glycerine layer , increases much more ] with the wave-length than in the case of a clean mercury surface .
This fact may go a little way towards an explanation of the of waves on water by oil .
But a complete explanation is to be looked for in a totally different direction .
Dr. Houstoun notes an mental result which is apparently contradictory to usual experience .
* 'Phil .
Mag January , 1909 .
The Elastic of and Steel , etc. The modulus of decay waves of length 5 .
when the surface of the mercury is covered by 4 mm. is secs .
All modes become propagated for wave-lengths , reater thnn some value between 5 and 10 cms .
For a depth of 1 mm. of glycerine this limit lies between 10 and 20 , and thus we see that as the depth is increased this critical value of the waveapproaches that for wayeH at the surface of deep lycerine , lies between 2 and 3 ] The Elastic Limits of Iro Steel under Cyclical Variations of Stress .
By LEONARD STOW , A.B.C.Sc .
, Wh .
Sch .
Communicated by Dr. Glazebrook , F.R.S. Receiyed Apri120 , \mdash ; Read May 13 , 1909 .
) ( Abstract .
) An explanation of is developed in the paper which is in accordance with all the researches on the fracture of materials by the cyclical repetition of stress .
The theory was put forward by Bauschinger in 1886 , when he ested that the necessary condition of safety was that the repeated stresses applied should be within the limits of ] asticity of the specimen , and that the least variation from this condition introduces and ately fracture occurs .
In order for this to be true for Wohler 's well-known experiments , the elastic limits must be variable within very wide limits , and the present paper describes observations made the adjustment of the linliCs of elasticity to any particular condition of expel.iment .
It is now found that iron or steel is capable of itself to variations of stress , cyclically applied , after a sufficienlt of repetitions .
When the adjustment is complete , the specimen under test is found to become fectly elastic throughout the whole cyclo , and does not occur .
During the adjustment of the elastic limits to a iven cycle of stress , a change of ( occurs in the ) ecimen , which corresponds to th extension observed in an ordinary tensile test when the yield stl.ess is exceeded .
For stress cyclically applied , this extension occurs eve when the maximum stress in the cycle is less than the stntic yield stress .
VOL. LXXXII.\mdash ; A. 2

