]\gt ; [ 801 ] Influence or Stress Strain on the sical ) HERBERT TOMLINSON , Communicated by Professor W. GRYLLS ADAMS , M.A. , Received February 18 , \mdash ; Read March 11 , 1886 .
( Revised by the Author October 19 , 1886 .
) [ PLATE 43 .
] PART I.\mdash ; ELASTICITY ( continued ) .
THE FRICTION METALS .
Origin Pmpose of the I HAVE already had the honour of presenting to the Royal Society*an account of researches carried on with a view of determining the effects of stress and strain on the elasticity and electrical resistance of metals .
Side by side with these researches were conducted others on magnetic induction and thermo-electricity , and a very considerable amount of experimental detail was collected with reference to the two last physical properties of matter .
The results of the last-mentioned investigations have , however , now remained unpublished for several years , for it seemed desirable , previously to publication , certain phenomena should , if this were possible , have light thrown upon them by further pxperiments on elasticity .
More especially was I anxious to examine into the causes of the loss of energy experienced by ) wire when vibrating torsionally , for the interesting memoirs of G. WIEDEMANN and D. E. HUGHES led me to expect that my doing so would cause some insight to be gained regarding the above-mentioned phenomena .
The results of these labours , which have now occupied almost the whole of my spare time for the last three years , I offer to the Society in the hope that they may prove as interesting to others as they are to myself of and MANN .
Under the title of ' ' The Elasticity and Viscosity of Metals Sir W. THOMSON published a memoir , first portion of which deals with the loss of energy of a ) wire * Phil. Trans Part 1 , 1883 .
'Wiedemann 's Annalen , ' vol. , 1879 . .
Soc. Proc vol. 14 , 1865 , p. 289 .
NIJCCCLXXXVI .
5 cause is small in comparison with the whole loss of energy which has been observed in many cases of vibration .
( b ) That , as a result of experiments in which a spring was vibrated alternately air of ordinary pressure and in the exhausted receiver of an , there is an internal resistance to its motions immensely greater than the resistance of the air .
Hence it is concluded that with solids with liquids there exists internal resistance to change of shape depending upon the rapidity of the change .
The results of THOMSON'S experiments are briefly as follows:\mdash ; The loss of energy in a vibration through one range was greater , the greater the velocity ; but the difference between the losses at low and high speeds was much less than it would have been , had the resistance been approximately as the rapidity of the change of shape .
When the weight of the vibrator was increased , whilst its moment of inertia was maintained constant , the viscosity was always at first much increased ; but then , day after day , it gradually diminished and became as small in amount as it had been with the lighter weight .
A wire which was kept vibrating nearly all day , from day to day , after several days showed very much more molecular friction than another kept quiescent except during each experiment .
The investigation was continued with much smaller degrees of maximum angular distortion , to discover , if possible , the law of the molecular friction , and , so far as the irregularities depending upon previous conditions of the elastic substance allowed any simple law to be indicated , it was proved that , as with fluids , the diminution of range per equal numbers of oscillations bore a constant ratio to the dimimshing range .
But , on the contrary , as with the larger angular distortions , the relation between the law of subsidence in two sets of oscillations having different periods , with the same elastio body in the same circumstances , was not that which would occur if the molecular resistance were simply proportional to the velocity of the change of shape in the different cases .
If the molecular friction followed this simple law , the proportionate diminution of range per period would be inversely as the periods .
This proportion was not found to hold good , but the loss of energy was , in fact , as it would be if the result were wholly or partially due to the " " elastische Nachwirkung\ldquo ; \mdash ; elastic afterworking\mdash ; as the Germans call it .
* I shall be able to show that this is not the case if we eliminate the effect of the resistance of the air .
This so-called " " fatigue of elasticity\ldquo ; does not occur if the oscillations are kept well within the limits of elasticity.\mdash ; H. T. There is a slip in S36 of 'Encycl .
Brit 9th edit .
, Art .
" " Elasticity in which article Sir WILLIAM THOMSON refers to his previously-mentioned memoir ; the words ' directly as the square roots of the periods " " should be " " inversely as the periods STRAIN ON THE PHYSICAL PROPERTIES OF GUSTAV WIEDEMANN , at the end of a very interesting paper on\ldquo ; Torsion discusses the question of the loss of energy experienced by a wire vibrating torsionally .
He had by previous statical experiments proved\mdash ; That permanent torsion occurs even after the slightest temporary torsion .
That when a wire is repeatedly twisted by a given torsional first one way and then the other , the distance between the permanent positions of equilibrium on each side gradually diminishes to a minimum ; the period during which this diminution is taking place is called " " the accommodation period That when the wire has been " " accommodated\ldquo ; by stresses repeatedly acting in opposite directions , till the temporary and permanent torsions and respectively are constant , if the wire be again twisted by increasing stress in the direction of the last torsion , increases nearly in proportion to , where , and represent any temporary torsion , permanent torsion , and torsional stress respectively between and , and WlEDEMANN , from these and other results , ] himself justified in concluding that " " the hypothesis according to which the diminution of the amplitude is to be referred to an internal friction which for the entire course of the oscillations is a function of the velocity can no longer be maintained Neither does he consider that the elastic after-action will alone account for what is observed .
The chief cause of diminution of the amplitude lies rather , according to him , in the shifting of the position of permanent torsion of the molecules the end of each oscillation ; what takes place at the alternate oscillations may be well described in his own words as follows:\mdash ; " " In the absolute first position of rest of the wire , before any torsion in its molecules , we will draw axes with the axis of the wire .
Let the wire be next ' accommodated ' by frequent vibrations to and fro , and at last be temporarily twisted the positive direction , describing an angle of , while the molecules may be rotated so that the lower ends of beir axes , looked at from the axis of the wire , describe an angle , say , to the left .
If the wire be slowly brought back into the permanent torsionposition , the axes of the molecules will retain a portion of their rotation to the left .
If the wire now receive an impulse in the positive direction , which again sends it to , it will , according to the of perfect elasticity , swing back to If it now swing beyond this position further to the right , and if the molecules in their rotation had no friction to overcome , it would arrive at the position the axes of the molecules would be rotated to .
Again , with perfect elasticity the wire would go back to the position , in which the molecules would be rotated and so forth .
The to-and-fro motions of the wire between and are perfectly elastic ; therefore the performances of work in the swingings backward and forward ' Wiedemann 's Annalen , ' vol. 6 , 1879 , pp. ; 'Phil .
Mag Jan. and Feb. , 1880 .
WIEDEMANN employs the term " " temporary torsion\ldquo ; so as to include not only the strain which disappears on the removal of the stress , but also the permanent torsion .
I , on the contrary , will always denote the torsion which disappears on the removal of the torsional stress as the temporary torsion .
positions of equilibrium , or the rotation of the molecules from to which determine it .
The diminution of amplitude may therefore be taken as a measure of this work .
Since the shifting of the rmanent torsion-position is , within certain narrow limits , proportional to the temporary torsion , and therefore the rotation of the molecules therewith is likewise approximately so , the amplitudes must , within those limits , diminish according to law of a geometrical series Again , Professor WIEDEMANN agrees with Sir W. THOMSON that , whilst an increase of load at first increases the loss of energy , this loss , under repeated oscillations , becomes less and less , and finally approaches in magnitude that obtained with the lighter load .
explanation of the matter is that , just as in his own statical experiments the torsional displacements were found to be initia ] greater with a heavier load than with a lighter one , so also must the decrements attendant upon oscillations be greater in the former case , but , like those , gradually approach the magnitude which is observed with the lighter loading .
Briefly expressed , then , the views of Sir W. THOMSON and Professor WIEDEMANN are as \mdash ; The former , whilst acknowledging that his experiments do not show that the proportionate loss of energy of a wire when vibrating torsionally is inversely as the period of vibration , which should be the case if the internal friction of a solid resemble bhat of a fluid , still employs the term viscosity with reference to metals , and appears to think that the internal friction of a metal partakes of the nature of the viscosity of a fluid ; the latter is of opinion that the loss of energy is mainly due to a shiftin from side to side of the positions of stable equilibrium of the molecules caused by the torsional vibrations of the wire .
I now proceed to give an account of my own researches relating to the internal friction of metals .
Descriptio of and Mode of imenting .
The wire under examination was firmly clamped at its upper extremity to a stout brass block , which was itself clamped to the end of a strong iron bracket projecting from the outside of the wall at the back of my house , the height of the bracket above the ground being centims. A wooden box A , nearly 600 centims. in length and 12 centims. square inside , protected the vertically suspended wire from currents of air , and passed through the roof of a small wooden house .
The box was secured at its upper extremity to the iron bracket , and rested with its base upon another box ( fig. 1 ) , which measured about 40 centims. in length and breadth and about 86 centims. in height .
The box had in front a glass window and a door , and at each of the two sides a door .
* All the doors were lined with caoutchouc , so *These doors are not shown in the figure .
STRAIN ON rPHE PHYSICAL PROPERIIES OF MATTEiR .
as to fit air-tight , and were only opened for the purpose of adjustment .
The top of the long box A A was also closed by a metal cap , and every precaution taken to ensure that the vibrations of the wire should be undisturbed by currents of air .
The wire near its lower extremity passed through the centre of an aperture made in the top of the box S. The bar V V of the vibrator was clamped to the lower extremity of the wire , and to it were suspended two cylinders of equal mass and dimensions ; by sliding the cylinders backwards or forwards on the bar , the moment of inertia of the whole vibrator could be altered without changing its mass .
The bar V V was graduated in milIimetres , and the cylinders could be clamped at any required equal distances from the axis of the wire ; the position of the cylinders could readily be estimated to the tenth of a millimetre .
The mode of attaching the cylinders is sufficiently shown in the figure , and it will be noticed that each cylinder is capable of being divided into two equal parts , each part having a mass of grms. , a length of centims. , and a diameter of centim. These cylinders , either complete or in halves , were used for nearly all the wires , except those of tin and lead , with which last lighter cylinders and bars were employed .
The bar V V consisted in some experiments of a piece of thin , drawn , tubing , of which the length was centims. and the diameter centims. Two br caps provided with screws , about 8 centims. in length and 2 millims. in diameter , fit one into each end of this hollow bar , and can be easily removed from or placed in it .
This bar will be designated as the bar .
In other experiments V V consisted of a thin ] brass tube , closed at both ends , of length centims. , and , except at the centre , of mean diameter centim. The central portion of the bar had soldered on to it a solid concentric brass ring centims. in length and of the same diameter as the bar so that either bar might be fitted into the brass block , which was itself clatnped to the wire .
The bar last described will be designated as the ; both and can just pass through a circular aperture in the block and be clamped at their centre to K. When the cylinders suspended to V are used complete they ] be designated as the cylinders , and when they are used in halves as the cylinders The torsional vibrations of the wire were observed by the usual mirror-and-scale arrangement , which is sufficiently shown in fig. 1 , where is the light lnirror reflecting an illuminated circle of light crossed by a vertical , fine , dark line on to scale placed at the distance of 1 metre from the mirror .
The scale was bent into an arc of a circle 1 metre in radius , and was supported on brickwork , which last , together with box , rested on the flagstones which formed the floor of the small wooden house previously mentioned .
So perfectly firm and free from disturbing causes of whatever sort was the whole arrangement that the spot of light , when at rest , remain perfectly so for any required length of time .
* * This , of course , shows that the arrangements made for securing freedom from currents of answered their purpose satisfactorily .
and had been tested recently at Kew ; the thermometers and were carefully compared by myself with degree by degree , and were graduated in degrees Centigrade .
A delicate aneroid barometer , placed on a level with the middle of the box served to determine the pressure of the air in the box ; the aneroid had been recently compared with a standard mercury barometer , and was graduated to one-hundredth of an inch .
Readings of the three therrnometers , and , and of the aneroid , were taken at regular intervals dul.ing the experiments .
When all the adjustments had been completed , the wire , whether in the nnealed or hard-drawn condition , was allowed to rest for a period of time varying , with the nature of the metal , from one day to several weeks , until there was no shifting of the position of torsional equilibrium and no appreciable " " running down\ldquo ; of the wire either before or after vibration .
It was necessary to adopt this precaution inasmuch as , though great care had been used in adjusting the wires to avoid imparting permanent or rather sub-permanent torsion , such was frequently found to exist , and manifested itself by a gradual shifting with time of the position of torsional equili- brium .
Now , when a wire is either possessed of sub-permanent torsion or is " " running down or even on the point of doing so , the loss of energy which occurs during torsional vibration is appreciably greater than when there are no such things .
Even when there may be no apparent permanent torsion , or " " running down , it is necessary to allow the wire to rest for some time , and to cause it to undergo frequent vibrations before the internal friction assumes constancy .
When it was necessary to excite vibrations a side-door of the box was opened , and slight impulses were given by means of a worsted thread to the bar V V until the amplitude of the oscillations was of the required magnitude ; the door was then shut , and another at the bottom of the long box A A was opened , so that any accidental pendulous oscillations might be checked : finally this door was closed , and when the amplitude of the oscillations had become reduced to the requisite extent the observer commenced to write down the extreme limits of several consecutive vibrations , as shown by the numbers on the scale .
At intervals the time of ansit of the light across the middle point of the scale was noted .
When the amplitudes decreased rapidly the observations were continued throughout the experiment ; but when the decrement was small the observer frequently left the room till the amplitude was sufficiently reduced , when he again noted down the extreme limits of the same number of consecutive vibrations as before , and so on .
In ] ] cases the experishown in the figure .
STRAIN ON THE PHYSICAL PROPERTIES OF MATTER .
ments were repeated nd again until the logarithmic decrement becanle sensibly constant in the differenut experiments .
* In reducing the observations , the following plan was adopted in those cases where the decrement of the amplitude was comparatively slight .
Suppose that ; , and are seven consecutive readings , the six corresponding arcs from rest to rest will be .
The means of , were written down , and if these agreed well with each other , which was almost invariably the case , the logarithm of the total mean was taken .
Now , say that single vibrations have taken place between the end of this and the beginning of the next readings , the difference of the logarithms of the first and second total means will , when divided by , give the mean logarithnlic decrement for a single vibration .
When the decrements of amplitude were comparatively large , the following mode of reduction was used : lppose a , and to be , as before , seven consecutive readings , so that the arcs from rest to rest , and : the difference of the logarithms of and ; of and ; of and , were put down , and , if they agreed well , the mean of these differences , divided by three , was taken as the mean logarithmic decrement of a single vibration .
The following experiments will furnish illustrations of the methods pursued in the two cases , and will also give a fair idea of the accuracy to be at , tained .
*This is very necessary , as sometimes , from slight fluctuations of temperature , the internal friction is very variable .
I have not unfrequently had to go on experimenting for a couple of hours or more efore the logarithmic decrement became constant .
In the earlier experiments seven readings were taken ; in the later ones eleven readings .
Experiment A W1RE of pure copper , hard-drawn , 6025 centims. in length and square centim. in section , allowed to rest for 24 hours after ffijustment .
Time of a single vibration , seconds .
isionsarc.akcnP i STRAIN ON PHYSICAL PROPERTIES MATTER .
Experiment II .
.A TIN wire , as pure as could be obtained by the ordinary process of distillation , centims. in length and 04972 centim. in diameter , tested about two months after suspension , having been in the meantime vibrated almost every day .
Time of a single vibration , seconds .
* The great variations which occur in the numbers due to the fact that in this case the period of torsio period of pendulous vibration .
( See note on pp. 809 , 8 Remarks Experiments I. It is worth while to pause here and to consider what we can glean by a careful examination of these two experiments , because they are typical of most of the others , whose number now can be reckoned by hundreds .
I had in my mind rather a strong conviction that there was an intimate relationship between the loss of energy from internal molecular friction and the electrical resistance of a metal , and have already observed to that effect .
Accordingly , regardin this apparent relationship as a matter of considerable importance , I spared no pains in attempting to get accurate values for the coefficients of viscosity of metals , such as have been obtained by MAXWELL and others in the case of fluids .
All the wires mentioned in this paper have been examined over and over again ; nearly all have been set up a second time , and most of them a third time , and , whilst in no case has a wire been under examination for less than a week , in some instances the experiments have extended over several weeks .
As the work progressed , the conviction became more and more forced upon me that I was dealing something very distinctly different from what is met with in fluid friction .
In Experiment I. we see apparently that one of the laws of fluid friction is exactly carried out , for according to the seventh column , in which is given the logarithmic decrement for one vibration , the diminution of range per equal numbers of oscillations bears a constant ratio to the diminishing range , the departures from the mean value of the logarithmic decrements being in no case greater than per cent. A closer scrutiny , however , shows us that this is only the case because , in calculating each of the logarithmic decrements , the mean of a great number of oscillations is taken .
If we look into each set of six arcs , we may notice that in some instances two consecutive arcs will show no diminution whatever , whilst in others the proportionate diminution is much more than the average .
These are not the results of errors of observation , because the differences are appreciab]y greater than any which could be accounted for in this way , and when we turn to Experiment II .
the phenomenon is more marked ; we have most indubitable evidence of a continued rise and fall of the logarithmic decrement .
This rise and fall would probably escape notice , were we able to take the mean of as many oscillations as were taken with copper ; moreover , if we commence a fresh experiment altogether , we find a total mean logarithmic decrement agreeing , within the limits oferrors ofobservation , with the previous one .
[ Note added Sept. 18 , 1886 .
have since discovered the cause of this rise and faJl : it is due to the fact that it is impossible to ensure that the axis of rotation always passes exactly through the centre of mass .
Of course , if the wire were perfectly flexible , and we could entirely avoid pendnlous vibrations when we started * Phil. Trans Part 1 , 1883 , p. 168 .
It will be observed that in Experiment 1 .
the mean logarithmic decrement for 70 oscillations is taken in all the series except the last .
where we have the mean for 210 oscillations given .
STRAIN ON TH PHYSICAL PROPERTIES OF MATTER .
the torsional vibrations , this would not be the case ; but , as a matter of fact , even if : we could start the wire without any pendulous , the axis of the wire never exactly passes through the centre of mass , and , as a consequence , when the wire vibrates torsionally , so-called " " centrifugal force\ldquo ; causes torsional motion to be partially converted into pendulous motion , or vice versed .
When , as in Experiment I. , the period of pendulous vibration does not approximately synchronise with the period of torsional vibration , the phenomenon is not very pronounced ; but when , as in Experiment II .
, the two periods approach to synchronism , it is very apparent .
Indeed , when the two periods approach sufficiently to synchronism , and the damping is not great , the phenomenon may become so marked that , after a time , almost the whole of the torsional motion becomes converted into pendulous motion , and presently reconverted again , so that the torsional vibrations , having first dimimshed to nearly zero , begin to increase again until nearly the initial amplitude has been reached , when once more the torsional motion begins to be converted into pendulous motion , and so on .
* An interesting fact in connexion with this phenomenon is that , if we allow the wire to rest , with oscillations at intervals , the rise and fall becomes much less pronounced , in consequence , probably , of the wire gradually allowing its axis to pass more nearly through the centre of mass , but we can easily reproduce it in all its intensity by jarring the wire , or even sometimes by change of temperature .
Unless there is nearly synchronism between the pendulous and torsional vibrations , there is little or no effect on the value of the logarithmic decrement , and , if we reckon from maximum to maximum , I have found that the logarithmic decrement is sensibly constant , the amplitude , provided the latter does not exceed a certain limit .
Similarly there will be little error introduced in our estimation of the internal friction of the metal from the mean logarithmic decrement deduced from a number of trials , which number was in my experiments very considerable .
When , however , the synchronism between the two periods is sufficient , there may be a very serious lterference with the logarithmic decrement , and even a sensible interference with the torsional vibration-period ; and I hold it as a matter of some considerable importance that , in all experiments connected with the damping , or even , when great accuracy is required , with the vibration-period of torsionally vibrating bodies , this point should be well looked to .
] * I had the pleasure of showing this phenomenon to the Physical Society , June 26 , 1886 .
A very slight shock sometimes will elfect this . .
In the case of a suspended with its axis horizontal , the axis of rotation will of not pass through the of mass anywhere except at the magnetic equator .
Formula for determining the Loss of Energy experienced by a Metal \ldquo ; under the Influence of Elasticity encountering Resistance proto the Velocity .
The diminution of amplitude of a wire vibrating as in the present experiments is due to three causes:\mdash ; 1 .
Energy impart , ed to the support to which the upper extremity of the wire is attached , or to the parts of the vibrator which may be capable of independent motion .
* .
Energy imparted to the air surrounding the wire and vibrator .
3 .
Internal molecular friction .
Of these causes , 1 may be dismissed at once as not affecting the results within the limits of errors of observation , and a means of eliminating the effect produced by the resistance of the air will be shown presently .
1 propose , therefore , in dealing with the problem before us , to confine myself to the consideration of Cause 3 , on the assumption that the internal friction varies as the first power of the velocity .
Let be the moment of the couple necessary to twist the wire through one radian ; the resistance due to internal friction , when the wire is twisting or untwisting with unit angular velocity ; the period of vibration from rest to rest ; and the moment of inertia : then it can readily be shown that very approximately , ( 1 ) and that the proportionate diminution of amplitude is ( 2 ) From ( 2 ) it follows that the logarithmic decrement should be independent of the amplitude of vibration , and should vary inversely as the period of vibration .
Discussion of THOMSON'S concerning thoe internal Friction of Me tals .
According to the mathematical formula just given , the logarithmic decrement should , with the same vibration-period , be a constant for all amplitudes if the friction between the molecules of a metal resemble fluid friction .
This we have seen is the case .
When I turned , however , to examine the effect of changing the moment of inertia of *If , for example , the cylinders attacl , ed to the bar of the vibrator be capable of a motion independent of the bar , as they would be if suspended by fine wires or threads .
Except , of course , in such cases as that of Experiment II .
, where the synchronism between torsional and pendulous vibration.periods causes the logarithmic decrement to vary ; or , rather , prevents there being any such thing as a logarithmic decrement .
@ STRAIN ON THE PHYSICAL OF MATTER .
813 the vibrator without at the same time altering the mass in order to prove the second law deduced above , I failed entirely , and , though I adopted numerous changes of the mass of the vibratcr , of the amplitude of vibration , of the vibration-period , and of the nature of the metal , in no case was I able to get a step further towards proving the above-mentioned law .
the contrary , when I had eliminated the effect of the resistance of the air in the manner now to be described , I found the logarithmic decrement nearly independent of the vibration-period in some cases increasing with it in others .
Mode of eliminating the of the Resistance of the In a paper quite recently presented to the Royal Society , I have shown that the following formulae for finding the logarithmic decrement due to the resistance of the air may be deduced from Prof. G. G. STOKES'S valuable memoir " " On the Effect of the Internal Friction of Fluids on the Motion of Pendulums For a cylinder vibrating horizontally dec. , .
( 3 ) where is the density of the air , the radius of the cylinder , the length of the cylinder , I the moment of inertia of the vibrator , and is a constant , provided the vibration-period , the diameter of the cylinder , and the nature of the fluid remain unchanged .
For a pair of cylinders of equal mass and dimensions , suspended vertically at equal distances from the axis of the wire , dec. , ( 4 ) where is the diameter and the length of each cylinder , and is the distance between the two axes of the cylinders .
We further have from Prof. STOKES'S paper the formula , ( 5 ) where is the vibration-period and is the coefficient of viscosity of air .
When has been determined , can also be determined by the aid of the Table given on p. 46 of Prof. STOKES'S paper .
By several very carefully executed series of experiments with vibrating cylinders ' Camb .
Phil. Soc. Trams vol. 9 No. X. where is the temperature of the air in degrees Centigrade and the units C.G.S. units .
From these four formulae we can readily determine the logarithmic decrement due to the resistance of the air on the horizontal bar V and on the two suspended cylinders .
The mode of proceeding is as follows :First , from equation ( 6 ) , is found the value of , which enables to be obtained from equation ( 5 ) .
Then we obtain by interpolation , from the Table given on p. 46 of Prof. STOKES'S paper , the value of , and , by substituting this last in equations ( 3 ) or ( 4 ) , according as we are determining the effect of the resistance of the air on the bar V V or on the two vertical cylinders , the logarithmic decrement required .
Th vertical cylinders are connected to the bar V V by the two suspenders , the cylindrical portions of which are affected by the air to an extent which must be calculated independently of the long cylinders , which they connect with V V. The effect of the resistance of the air on the other portions of the suspenders was determined experimentally by using the bar for V V , and observing the logarithmic decrement , first when the bar was vibrating with the suspenders attached , and afterwards with the suspenders removed , but with brass cylinders introduced into the hollow bar V V , of equal mass with the suspenders , and so adjusted that the moment of inertia ained unaltered .
By this means it was ascertained that the portions of the suspenders which were not cylindncal were equivalent to a certain length of the long vertical cylinders .
This equivalent length would vary so slightly with the vibration-period that we may without sensible error consider it a constant ; accordingly , in order to allow for this portion of the suspenders , the length of each of the long vertical cylinders was estimated at so much more .
* For the coefficients of and in this formula I have taken my authority HOLNAN ( ' Phil. Mag vol. 21 , 1886 , p. 199 ) , as not only do Prof. HOLMAN'S investigations seem to have been very carefully conducted , and to agree as regards the effect of change of temperature on the coefficient of viscosity with those of MEYEN , PULUJ , and others , but also they are more in accordance with my own in this respect than are those of MAXWELL , who found the coefficient to vary as the absolute temperature .
MEYER and PULUJ have obtained values for the absolute coefficient of viscosity of air which differ from my own , but I prefer to use my own results , as these were obtained by mentmg with cylinders and bars of somewhat similar dimensions to those used here , and also under similar conditions to those holding here .
I believe the equations given above to be capable of expressing the logarithmic decrement due to the resistance of the air with certainly a less error than 1 per cent. Because varies for cylinders of different diameters .
For this purpose the caps were removed from the ends of .
and the brass ] inders , by means of companion.screws cut along their axes , were adjusted at the required distance on the screws attached to the caps ; the caps were then replaced .
STRAIN ON THE PHYSICAL PROPERTIES OF MATTER .
Equation ( 4 ) does not take into account the effect of the air due to the rotation of the cylinders abont their axes ; the correction required for this is , however , very slight .
* When necessary , this correction was made by means of the following formula : .
( 7 ) where is the natural logarithmic decrement due to the rotation of both cylinders about their axes , is the mass of fluid displaced by cylinder , and , ( 8 ) being put , by way of abbreviation , for , where is the radius of the cylinder .
The Extent to which the Friction of Metals depends upon the Vibration-Period .
We are now in a position to examine how far the loss of , resulting from internal friction , depends upon the of the wire .
I propose , tlJerefore , to give , in illustration , the results of four sets of experiments on -steel , on copper , on zin , and on tin .
III .
UNANNEALED Steel 0 centim. in diameter and 602 centims. in length .
Bar and Cylinders * Would not in any case exceed per cent. of the whole .
Kindly furnished by Prof. G. G. STOKES .
will be observed that the numbers in this column at first decrease with the time of swing and then increase .
This arises from the fact that the time of swing , as well as the dimensions of the vibrator , come into the equations used for determining the logarithmic decrement due to the resistal ) of the air .
Experiment I ANNEALED Copper Wire centim. in diameter and 602 centims. in length .
Bar and Cylinders Experiment UNANNEALED Zinc Wire 07913 centim. in diameter and 602 centims. in length .
Bar and Cylinders used .
Temperature i Experiment VI .
A wire of tin , unannealed , centims. in length and09944 centim. in diameter .
The vibrator consisted of a solid cylindrical bar 233 centims. in length , centim. in diameter , and grms. in mass ; suspended from the bar were two brass inders , each centim. in diameter , centim. in height , and grnla in mass .
The total mass of the vibrator was 49 grms. * The temperatures varied from C. to C. , data were at hand for ecting the of the logarithmic decrement when the temperature differed from the average , C. A similar remark applies to the temperatures C. and C. of the next iment .
STRAIN ON THE FHYSICAL PROPERTIES OF MATTER .
The brass cylinders replaced by lead cylinders , so that the total mass of the vibrator was 168 grms. 018784 023664 7 .
024885 020583 026035 027316 on Experiments lclusive .
Let us now see what we can gather from the data given in these experiments .
With tin the logarithmic decrement dec , idedly increases with the vibration-period , but not in the proportion as the latter , and may be expressed by the following formula : where is the vibration-period and the logarithmic decrement , whilst , and are constants , the last being a negative quantity .
Thus , when the mass of the vibrator in periment V. was 49 , and telnperature the formula becanle , ( 9 ) whilst for same temperature , and with the mass of the vibrator equal to 168 grms. , it . . . .
( 10 ) The logarithmic decrement would thus to be capable of being divided into two parts , the one being independent of the load and of the vibration-period , and the other dependent upon the load and the vibration-period .
When the mass of the vibrator was 168 grms. , and the temperature C. , the formula became .
( 11 ) The effect of change of temperature is very considerable , for a rise of C. caused an increase of per cent. in the internal friction , and proved to be nearly the same for the three vibration-periods : 1913 second , 4366 seconds , and seconds .
With zinc the decrement also increases with the vibration-period , but not to the same extent as with tin , the formula in this case becoming . . . . .
With the metals possessing less internal friction , steel and annealed copper , the variation of the logarithmic decrement with the period of vibration is so slight as to be masked by errors due to small fluctuations of temperature*and other causes ; and , had not a great many other observations than those ere recorded proved that the internal friction does slightly increase with the , it would have almost appeared not to do so .
If , however , we compare in Experiments III .
and IV .
the value of the first half of the logarithmic decrements with that of the second half , we may trace indications of the truth of the above statement .
To sum up , then , these and many other experiments of a like kind have nearly satisfied me that , with all metals , the logarithmic decrement does increase with the period of vibration , but that , the ] internal friction the metal possesses , the more nearly is the former independent of the latter .
Effect of of Temperature on the Internal Friction .
Perhaps one of the most remarkable features presented by a torsionally vibrating wire is the ease with which the molecules are disturbed by changes of temperature from those positions of stable equilibrium into which they will , after a period of time , whose length depends upon the nature of the metal , naturally creep .
It is truly astonishing to observe the profound change in the interIJal molecular friction of such an elastic metal as unannealed piano-steel which is wrought by a sudden rise or fall of one or two degrees of temperature .
I have explained that , though the wires were shielded by a long box , this box was itself exposed , for the greater part of its length , to the external air .
Now , the first wire examined was one of unannea]ed piano-steel , and some hundreds of trials were made with it during a period of three months , and * See what follows } I do not know that unannealed piano-steel is more remarkable than other metals in the respect mentioned.\mdash ; H. T. STRAIN ON TIIE PHYSICAL PROPERTIES OF R. 819 at various hours of the day .
Sometimes the proportionate diminution of amplitude would seem quite as constant for different ranges of amplitude as could be expected , but at others there would be variations far exceeding anything which could be attributed to errors of observation .
For some weeks I was quite at a loss to account for these variations , but at length was successful .
Luckily , I had determined to make frequent observations of all three thermometers , and , whilst examining the contents of my note-book , it occurred to me that the greatest variations took place on those days and hours at which the variations of temperature were greatest .
I also noticed that the decrement , whaether the of temperature was of the nature of a rise or a fall .
At first this explanation seemed almost incredible , for the variations of temperature would sometimes be only 1o C. , or even less , and yet at the same time the value of the logarithmic decrement would be increased 30 , 40 , or even 100 per cent. My suspicions having been thus roused , I found it comparatively easy to verify them , and among several other experiments , made with the same object , tried the The same piano-steel wire as in Experiment III .
The sun was shining rather faintly , and at intervals on that side of the house on which the wire was suspended , but before the end of the experiment passed over to the other side .
Civil time .
Temperature o ] .
10.33 I .
10.33 I .
10.33 I .
10.33 I .
10.33 I .
10.33 I 1.98\mdash ; From 7 .
to 9 , when the experiment terminated , the logarithmic decrement and the temperature became nearly constant , the value of the former being The changes of temperature , as shown in the second column , are very ] , but we see what a large effect they have on the internal friction , which in one case became twice as great as when under a fairly steady temperature .
The effects of fluctuations of temperature on the internal friction resemble those of mectJanical jarrings , such as would be produced when the mass of the vibrator is altered , or even when the cylinders are shifted along the bar without any alteration of mass ; *they also , as I shall show in a future paper , can be imitated by means of magnetic stress .
Attention should also be directed to the fourth column of the experiment , where it will be seen that not only is the loss of energy increased by change of temperature , but so also is the ation .period ; the latter is , however , so much less altered than the former that it requires careful observation to tect any such similarity between them .
In consequence of the mentioned effect of fluctuations of temperature on the internal friction , it was necessary to choose a day , or time of day , when the temperature was fairly steady , and all the results which will be recorded in the following portion of this memoir obtained when this was the case , with all the metak which I have examined the phenomenon is more or less pronounced .
Since it is impossible , with the arrangements which I have described , to obtain a perfectly steady temperature , it seems probable that most of the values of the logarithmic decrements previously recorded are slightly in excess of what they would have been , could we have secured greater uniformity in this respect .
The question next arises : will any change of ) , however , introduce increase ) internal friction ?
I think not ; for not on\yacute ; in some instances have I obtained values for the logarithmic decrements , which were as constant as the limits of errors of obser would allow , and yet where the experiments the temperature did change several times , and in an irregular manner , though by small amounts , but I shall in another paper bring forward instances where the molecules were slightly twisted backwards and forwards by small magnetic stress without appreciably affecting the value of the logarithmic decrement : The Effect of Loading on the Friction .
We have seen that according to both Sir WILLIAM THOMSON and Professor G. , when the load on the wire is increased , there is at first an increase in the internal friction , but this increase diminishes with time , so that the logarithmic decrement tends to become the same as with the lighter load .
Neither of the above observers has , however , atbempted to ascertain whether or not the internal friction is entirely independent of the load , provided sufficient time be allowed , and sufficient account be taken of the effect of the resistance of the air in diminishing the amplitude of vibration .
The angements were therefore fflade for the purpose of scttling this point by experiment:\mdash ; The brass block , which was clamped the bar V V of the vibrator ( fig. 2 ) , *This is one of the reasons why the logarithmic decrement at inoreased when the mass of the vibrator is decreased or increased , even the change of mass produce no perceptible permanent alteration whatever in the length of the wire , and why it is prudent to allow the wire to oscillate for some time after shifting the cylinders , as in Experiments III .
, STRAIN ON THE PHYSICAL PROPERTfES OF MATTER .
terminated at its lower extremity in a ring , through which a pin could just pass .
The pin served to connect the brass box A with the bar of the vibrator by passing through two rings , attached to the top of , and placed one on each side of the ring at S. In this way the box A was well secured to V , and without the least chance of any shifting taking place , as the wire vibrated torsionally .
Two smaller boxes were suspended from V , and could be clamped , as shown in the figure , to any of it .
All three boxes were provided with hinged doors*of the same height as the boxes , so as to admit of the introduction of two sets of cylinders and , the one set of brass and the other of lead ; these doors , after the cylinders had been arranged in the boxes , were kept shut by means of springs and clasps : The mass of each of the brass ] indersC was 100 grms. , and that of each of the lead cylindets 2000 grms. The masses of each of the cylinders in the two sets were adjusted by myself , so as to agree with some standard masses within part in 10,000 .
Each of the cylinders had been casb in the same mould , and the dimensions had been carefully measured by means of to inch .
The moment of inertia of each cylinder was detel.mined from the dimensions and mass , and afterwards by certain indirect methods , which last , when suitable precautions were taken , results in good accordance with those which had been obtained by the direct method .
Moreover , the moments of inertia of all the lead cylinders were found to be very nearly equal to each other .
The mode of proceeding was as follows :Eight of the lead cylinders were placed in the box , and on the top of these was put a wooden cylinder , fitting rather tightly in A. On the top of , and fitting into a circular cavity let into the centre of it , were placed sixteen of the brass cylinders , arranged in sets of four each , each set having its four constituents bound together by an elastic band .
The wooden cylinder served not only to prevent any rocking of the lead cylinders as the wire vibrated , but also to enable the experimenter to readily adjust the brass cylinders in their proper positions in the box A. When all the necessary adjustments had been carefully completed , the whole was allowed to remain for five days at rest , except that the wire was caused to vibrate several hundred times on each day ; finally , logarithmic decrement and the vibration-period were determined from a very large number of oscillations .
As soon as these had been determined in a sufficiently satisfactory manner , two of the lead cylinders were removed from , and four of the little brass cylinders were transferred from A to the boxes , two being placed in each box .
The boxes , had been at the outset so placed on V V that the above transference would exactly make up for the loss of the moment of inertia which would follow from the removal of bhe two lead cylinders .
After this second adjustment and a rest of twenty-four hours , followed by a large number of preliminary vibrations , the logarithmic decrement and the vibration-period were again determined .
two more cylinders were * Not shown in removed from A and four more transferred A to all the lead cylinders had been removed from A and all the brass cylinders transferred from A to the boxes , B. The next experiments show the results obtained by treating steel and copper wires in this manner .
Experiment VIJI .
nannealed piano steel , centim. in diameter and 602 centims. in length .
Mean vibration-period , seconds .
The temperature ranged from C. to C. Experiment IX .
Annealed copper wire , centim. in diameter and 602 centims. in length .
Mean vibration-period , seconds .
The temperature ranged from C. to C. Logarithmic Load on wire ( Iecrement due to in kilos .
internal friction .
The variations , occur in the values of the logarithmic decrements dffierent loads , are to be attributed to differences in the rate at which the temperature varied on different days .
Had a greater number of observations been made with each of the loads on different days , the mean values of the logarithmic decrements would have varied still less .
We may say , then , that for comparatively large loads the internal friction is independent of the load , whether the wire be annealed or hard-drawn .
We will now turn to consider the effect of comparatively light loading on the internal friction .
The { igure shows th arrangements of the cylinders after this last operation .
STRAIN ON THE PHYSICAL PROPERTIES OF MATTER .
L'xperiment The piano-steel wire used in Experiment VII .
was , after the removal of the boxes , tested with suspended cylinders of different masses .
The temperature ranged from C. to C. , and the vibration-period from seconds to ecor ) The annealed copper wire of Experiment IX .
, after having been tested in the manner described in this experiment , was tried with lighter loads , The temperature ranged from C. to C. , and the vibration-period from seconds to seconds .
Experiments X. and XI .
teach us that both for hard-drawn and annealed ires the internal friction is as independent of the load for light loads as we have seen it to be for heavy ones .
We must , however , in such experiments as these be careful to begin with the heaviest load first , as loading tends more or less to produce a diminution of the internal friction , which is partly permanent and partly sub-permanent , provided no sensible permanent elongation results from the loading .
Care should also be taken , in changing from the heavier load to the lighter one , to avoid shocks .
If these precautions be not taken , the logarithmic decrement will probably be found less for the heavier load than the lighter one .
The Effect of on the Torsional It is , I believe , generally assumed that the torsional elasticity of a wire is independent of the load , but apparently no experimental evidence has been brought forward lpport of this assumption , at any rate as fnr as large concern We might , perhaps , be inclined to infer that , because the internal friction is independent of the load , so also be the torsional elastici , but my own experience had taught me that it was by no means safe to do so .
It is true that Experiments VIII .
and IX .
had shown that the torsional elasticity must be nearly independent of the load when the latter is at all arge , inasmuch as the period of vibration was nearly the same for all the loads .
There were , however , slight differences in the vibration-period for the different loads , which I felt inclined to attribute to the fact that the boxes might not always hang quite vertically .
The following experiments seem to show conclusively that for loads rangin from moderate amounts to great pnes the period of vibration is independent of the load .
Experiment XII .
A. pan made of brass was suspended by two hook .
( fig. 3 ) ; the bottom of the pan was circular , and tho two suspenders .
, made of drawn brass , were soldered into A. The lead cylinders before-mentioned were in the first instance placed on the pan , so that their centres were in the same line as the axis of the wire .
brass cylinders , , were placed near the ends of the bar , at equal distances from the wire , and had in consequence of their position and mass a very considerable moment of inertia as compared with the rest of the vibrator .
Afte the arrangements had been completed the apparatus was allowed to remain for several days , and was during this period frequently vibrated .
The vibration-period was then determined , and , two of the lead cylinders been rerooved , a further rest of 24 hours was given , when the vibration-period was determined ; and so on , until all the lead cylinders had been removed .
In this experiment the same piano- steel wire as before was The ternperature ranged from C. to C. in .wire in * G. has shown in his paper already quoted that for moderate loads the torsional elasticity is independent of the load .
Rest should always be given either after on or taking off load , as , though the elasticity is not so much affected by shocks as the internal friction , it is affected .
STRAIN ON THE PHYSICAL PROPERTIES OF MATTER .
Experiment XIII .
copper wire used in Experiment IX .
was treated in the same manner as the steel in the last experiment , except that now the cylinders were dispensed with .
The temperature throughout remained nearly constant at C.Remarks on Experiments XII .
and XIIl .
Experiment XII .
shows that for the uuannealed piano-steel the couple required to twist the wire through one radian is quite independent of the load for loads ranging from 4 to 20 kilos .
For smaller loads than 4 kilos .
the value of the couple is slightly less .
With the annealed copper , the couple required to twist the wire one radian is independent of the load for all the loads .
Having thus endeavoured to prove that the torsional elasticity does not depend upon the load when the latter is comparatively ] arge , both for hard-drawn annealed wires , I wiJl consider the effect of light loading .
As regards annealed wire , Experiment XIII .
furnishes evidence that the elasticity is very arly independent of the load , even for light loads , for the lightest load in this experiment does not much exceed kilo .
* Care must , however , be taken , with annealed wire , to begin with the heaviest load first , as loading , if not carried to the extent of causing perceptible permanent elongation , invariably produces an increase of elasticity , which is partly permanent and partly sub-permanent , and this last must not be disturbed by shocks in unloading if we are to strictly compare the values of the torsional elasticity with different loads .
As regards the unannealed piano-steel , the case is otherwise , for if the loads be small the torsional elasticity may vary very perceptibly with them .
would produce only a very small stress per square centimetre as compared with the stress used with the unannealed piano-steel .
See similar remarks respecting the internal friction .
With the heavier loads still remaining on the wire , the sub-permanent is not so readily shaken out as with the lighter ones .
I ] sinco thoroughly satisfied self that torbional elasticity is independent of the load , even loads , the case of annealed wires , provided sufficient time be allowed after changing each load .
Experiment same piano-steel wire as before .
The temperature ranged from C. to C. It will be noticed that there is a comparatively rapid rise in the torsional elasticity as the load increases from 164 grammes to 1208 grammes , and a much less proportionate increase as the load rises still further .
Now we find this phenomenon to be closely associated with the permanent torsion , which is , to a greater or less extent , always imparted during the process of wire-drawing .
Permanent torsion produces decrease of torsional elastic.ity , but , as I have already shown , *unannealed piano-steel , if permanently twisted by hand , begins to temporarily untwist when loaded , the amount of untwist being , for small loads , very rapid , and consequently we might perhaps expect the elasticity to be increased by loading , rapidly at first , and afterwards much more slowly as the loads were increased in amount .
Similarly pianosteel behaves when permanently twisted by wire-drawing ; this is apparent from the next experiment .
Experiment The unannealed piano-steel used in the previous experiments , and which had been , during a period of several weeks , frequently loaded and unloaded with weights reaching to upwards of 20 kilos .
, was allowed to lest for some time with only the bar of the vibrator S attached ; a pan whose mass was grammes was then suspended * ' Phil. Trans Part I. , 1883 , p. 21 .
Tbis expression needs a little explanation .
By the words " " permanently twisted\ldquo ; is meant that the torsion will remain in the wire so long as we do not ] the load under the influence of which the wire was when the torsion was originally imparted , or , if we do alter the , but afterwards restore it to the original amount , the twist originally imparted will be restored .
I have since examined the matter more closely , and find that , with all very hard-drawn wires , there is a slight difference between the torsional elasticity , as inferred from torsional vibrations , when ereIlt light loads are used .
This is also the case with annealed metals which , after annealing , have suffered considerable permanent torsion .
Whether the torsional sticity increases or diminishes with the load depends apparently upon whether loading causes the wire to temporarily untwist or twist more .
S The mass of this bar was 164 grammes .
STRAIN ON THE PHYSICAL PROPERTIES OF MATTHiR .
) by means of hooks from the bar , and afterwards loaded carefully with weights , which eventuall .
reached 16 kilos .
Load dded.ebar i The of Permanent on the Frlction .
Experiment An iron wire , 602 centimetres in length and 0.960 centimetre in diameter , was provided with a vibrator whose mass was 3606 grammes ; a rest of 24 hours was now allowed , and after this the wire was tested , the results of the experiments showing a logarithmic decrement of 001131 .
Next the wire was permanently elongated to the extent of per cent. , alud , having beep to its original length , was again tested about ten minutes afterwards ; the logarithmic decrement was now found to be 001422 .
After a rest of 24 hours , the internal ction had so far diminished that the logarithmic decrement had become , i.e. , nearly 8 per cent. greater than it was before the permanent extension had taken place .
The vibration-period before the permanent strain was seconds , and aftel the extension seconds ; therefore , allowing for the diminution of section , resulting from the strain , the effect of the latter was to cause a decrease of the torsional rigidity of per cent. We must observe , with reference to such an experiment as above , that , in order that permanent extension may produce increase of internal friction , the wire must be loaded , when tested before the extension , to a degree which , though not sufficient to cause sensible permanent elongation , is yet enough to secure a certain amounlt of permanent strain among the molecules ; if the wire be only tested with a light vibrator , permanent extension may possibly produce a diminution of internJ friction , as the following experiment shows .
The load produced by the pan itSf , .
H. TOMLINSON ON THE NFLUENCH ] OF STRh'SS AND Experiment An annealed copper wire 602 centims. in length and centim. in diameter was provided with a vibrator whose mass was 430 grammes , and after resting for several days was tested ; the logarithmic decrement proving to be 000297 .
The wire was now permanently lengthened per cent. , and ten minutes after it had been shortened to its length and -suspenderJ was again tested , with the followin results : \mdash ; Thus we see that , though the permanent extension increased considerably the internal friction , when the wire was vibrated shortly after the strain had been produced , the effect was rapidly lessened by rest , so that after two days the logarithmic decrement became slightly less than it was before the extension .
The time of ration was , before the extension , seconds , and after the extension seconds .
If we allow for the diminution of section , caused by the strain , there results from extension a of the torsional ?
.igidity of , about per cent. VIII .
The same wire was twisted through complete revolutions , of which came out , leaving a permanent torsion of fourteen reyolutions ; ten minutes afterwards the wire , which had been kept vibrating , was again examined , with the following results:\mdash ; Time of lion in seconds .
Logarithmic decrement due to internal friction .
Remarks .
Mean value for the first 100 vibrations lean value for the next 200 vibrations After a rest of 24 hours .
Mean value for 400 vibrations *observe t diminution of onal rigidity is not always attended by incre , as of internal friction .
STRAIN ON PHYSICAL PROPERTIES OF MATTER .
It would appear that permanent torsion has much more effect than extension increasing the internal friction .
The torsional rigidity of metal is also much more diminished by permanent torsion than by permanent extension , but the effect on the rigidity is nothing like so great as that on the internal friction , which last amounts to nearly 76 per cent. Having shown that when sensible permanent deformation is produced , either by traction or torsion , thele an increase of internal friction , I will now consider the effect of stress , which is not sufficient to produce sensible molar permanent deformation , but which causes molecular .
deformation .
Experiment XIX .
An annealed iron wire 601 centims. in length and centim. in diameter .
Attached to the bar of the vibrator was a pan , which , together with the bar , weighed rather more than 2 kilos .
The wire was , after adjustment , allowed to remain for five days , and was in the meantime frequently vibrated .
After period weights were put into the pan , and after each addition of load , followed by a rest of about ten minutes , the logarithmic decrement was determined ; before each determination the wire was suffered to make about 100 vibrations .
After having vibrate .
Started athrough a.or about ten minutes-An additional load of 1 kilo .
was now put on and caused a slight permanent extension , which , after a rest of two days , was ascertained to be centims. The wire was then reduced to its original length , and , 14 kilos .
being the load , was allowed to vibrate for ten minutes .
After this the logarithmic decrement was again determined .
* Notice the " " fatigue of elasticity\ldquo ; ( see p. 832 ) .
a verylargenumber The above experiment suffices to show the great extent to which the internal friction can be dimimished by the permanent molecular strain resulting from loads which are not sufficient to cause any appreciable permanent extension .
For we see that even with aload of 15 kilos .
the permanent extension was only about per cent. , whereas a load of half this was sufficient to cause a permanent molecular strain which reduced the friction to less than one-half of the original amount .
The Effect of Strain on the Rigidity .
I have already pointed out in Experiments XVI .
, XVII .
, and XVJII .
that when very sensible permanent molar deformation is produced , either by traction or torsion , the torsional rigidity is decreased , that we need only now consider the effect of stress which , without producing sensible molar deformation , suffices to cause molecular permanent strain .
Experiment An annealed copper wire 602 centims. in length and centim. in diameter .
The mass of the vibrator attached to the wire was 1554 grammes , and the determinations of the logarithmic decrements were all made with only the vibrator .
When , however , it wa .
necessary to strain the wire , a pan was suspended from the bar of the vibrator , and on this pan weights were placed which were suffered to remain for some time .
The pan and weights were then removed , and 24 hours after the removal the testing , preceded by a large number of preliminary vibrations , began .
The temperature ranged from C. to C. *Sir W. THOVSON had previously shown this to be the case Encycl .
Brit 9th edit .
, Art .
' ' Elasticity S81 ) .
STRAIN ON PHYSICAL PROPERTIES OF MATTER .
831 There was no perceptible permanent elongation produced by the loading , so that the increase of torsional elasticity , which is very apparent , must be due to molecular and not molar strain .
Part of the strain is only sub-permanent and can be shaken out ; when this done there is an immediate loss of elasticity , which is regained if the wire be loaded to the same extent as before , and subjected to the same previous treatment .
Effect of on the Internal We have seen that both permanent extension and permanent torsion are attended with permanent increase of internal friction .
Now , in the process of wire-drawing there occurs both permanent extension and torsion , combined with lateral pressure .
As a consequence , wire-drawing increases the internal friction in an astonishing manner .
I shall have occasion to dwell more fully on the effects of annealing in a memoir which I hope will shortly follow this one , but the following Table will bear witness to the truth of the above statement:\mdash ; TABLE .
We thus see that annealing the hard-drawn metals reduced the internal friction to an amount which varied from to one-thirtieth of the original .
The Logarithmic Decrement due to is independent of the of the nent The annealed copper wire of Experiment .
was drawn to a diameter which was nearly one-half of the original and afterwards re-annealed .
With the same length of wire in both cases , and with nearly the same preliminary treatment , the logarithmic decrement due to internal friction only was for the thicker wire and for the thinner one The temperature of the wire was nearly the same in each case , * Great care must be used in an experiment of this kind to ensure uniformity of temperature during the period of experimenting , otherwise the thinner will show a gl'eater internal friction than ) thicker one see Experiment ifference idecrements ireater.than c wires , whose diameters were also nearly in the same ratio , namely 2 : 1 , gave values for the logarithnlic decrements due to internal friction which agreed with each other within the errols of observation .
xperiment XXII .
Two pieces of the same specimen of annealed zinc wire , one 602 centims. in length and the other only 97 centims. in length , were both tried under nearly the same conditions as regards preliminary treatment and time of vibration , \amp ; c. The logarithmic decrement due to internal friction was , in the case of the longer wire , 003207 , and , in the case of the shorter one , 003327 .
The difference between these two numbers can quite be accounted for by the fact that the temperature of the shorter wire was about C. higher than that of the longer .
Two pieces of iron wire , cut from the same hank , but one having a length of 96 centims. and the other of 602 centims. , gay , when treated in a similar manner , exactly the same logarithmic decrements .
We may , therefore , regard the logarithmic decrement as independent both of the length and of the diameter of the wire , provided care be taken that the two pieces of different lengths which are compared shall have suffered the same preliminary treatment .
" " Fatigue of Elasticity Sir WILLIAM THOMSON writes , Experimental exercises performed by students in the physical ]aboratory of the University of Glasgow , during the session 1864-65 , brought to light some remarkable and interesting results , proving a loss of energy in elastic vibrators incomparably greater than anything which could be due to imperfections in their elasticity , and showing also a very remarkable fatigue of elasticity , according to which a wire , which had been kept vibrating for several hours or days through a certain range , came to rest much more quickly when left to itself than when set in vibration after it had been at rest for several days then immediately left to itself This so-called " " fatigue of elasticity\ldquo ; seemed to me such a very remarkable phenomenon that , throughout the whole of these investigations , I was constantly on the look-out for traces of it ; but , owing no doubt to the very slight maximum deformations employed in these experiments , I could detect nothing of the kind with any metalS except nickel .
The whole of the other metals examined , namely , * It is necessary to take great care that the temperature should fluctuate very little , as any fluctuation will afTect a thin wire more than a thicker one .
The effect of change of temperature on the internal friction of zinc is rather considerable .
'Encycl .
Brit 9th edit .
, .
" " Hilasticity S30 .
S Except in Experiment XIX .
With nickel , however , the case was different , met with most marked signs of " " fatigue of elasticity even for such small molecular displacements as attended arcs ranging from 300 to 700 scale-divisions .
Experimeni XXIII .
A nickel wire , unannealed , 602 centims. , and 018622 square centim. in section , after having been tested for several days with three rent vibrators , was finally tried with the mosb massive of the three .
* Immediately after adjustment .
005689 further r After ten minutes ' rest .
After 2 hours ' further rest .
After hours ' further rest .
After hours ' further rest .
005644 Stnrted again 006023 . .
An arc of scale-divisions maintained for 2minutes , and then to subside .
006427 . .
An arc of 700 scale-divisions maintained for 30 minutes , and then allowed to subside .
In all the above trials the initial arc from which the logm.ithmic decrements were computed was about 600 scale-divisions , and the final arc about 150 scale-divisio1ls .
I next give the whole of the observations which were taken in the last trial from the initial to the final arc .
* The total mass of this vibrator was 366 grammes .
DCCCLXXXVI .
5 calculations necessary for obtaining the above results bad been finished , the vibrator was again started , but now with only an initial arc of about 300 scale-divisions .
By this time the vibration-period had diminished to seconds .
After a rest of five hours the vibrator was started with an initial arc of just over 100 scale-divisions .
of arc .
ternat Remarks on Experiment XXIII .
In the first place , we may direct our attention to the diminution of the vibrationperiod produced by rest after repeated vibrations ; this seems to be almost , if not quite , as pronounced as it is in the case of , the vibration-period gradually diminishing from seconds to seconds , and this , too , in spite of the fact that the same vibrator had been used several times previously with the wire , and had been carefully adjusted at the commencement of the experiment .
* In the second place may be noticed the diminution of the logarithmic decrement which attends the above-mentioned diminution of the vibration-period , and that the former of the two is very much greater than the latter .
This last is easily accoumted for if we ssume with WIEDEMANN that the internal friction is due to the to-and-fro rotation of the molecules about their axes , and not , as is the torsional elasticity of the wire , mainly upon the force necessary to alter the distances separating the centres of the molecules .
Thirdly , the effect of elastic fatigue\ldquo ; becomes less and less apparent as the ampli*It should be understood that the bar of the vibrator was not changed in any of the alterations of the mass of the vibrator , but only the cylinders .
tudes through which the logarithmic decrement is calculated inish , and , moreover , the range through which the logarithmic decrement is a constant becomes wider as the amount of " " elastic fatigue\ldquo ; is diminished .
We see here analogies between the increase of proportionate tion of amplitude caused by " " elastic fatigue " " and the increase of the ratio of the temporary molar deformation , whether or longitudinal , to the stress producing it , as the latter increases whenever the vire acted upon has recently suffered permanent longitudinal extension or torsion .
Lastly , it is desira , ble to draw attention to the marvellously small deformations which can produce " " elastic fatigue\ldquo ; in nickel as compared with those which are attended with corresponding effects in the case of the other metals .
The Effect of Rest of equent Oon the Internal Friction .
It has been already observed that rest , er strain of any kind invariably diminishes the internal friction , but rest may be considerably aided in this effect by agitating the molecules either by thermal or by mechanical agency .
With the part played by thermal agents shall deal in a later memoir , and now turn to agitations by oscillating the wire torsionally .
Experiment XXIV .
The wire used in Experiment , and which had suffered the extension mentioned in this experiment , was allowed to rest for 24 hours with a load of kilos .
on , and was then tested with the same load still on .
ationsb Started again with an initial arc of ( i00 scale-divisions , and left vibrating until the arc was reduced to scale-divisions .
* ' Phil. Trans 1883 , Part I. , ) .
, 8 .
See also paper .
cit. In this experiment the effect of the resistance of the air was not eliminated , it was impossible to have done so , and the elimination would have been of little importance .
tion .
When , however , rest for one or more days was now given , the internal friction would rise again until once more repeated oscillations would reduce it to 001972 or thereabout .
No doubt the increase of internal friction attending on rest in this case is owing to changes of temperature which occur in the interval , such changes , as we have seen , tending to disturb the molecular arrangements which existed at the end of the fi.equent oscillations .
1 .
The proportionate diminution of amplitude due to internal molecular friction in a torsionally vibrating wire is independent of the amplitude , provided the deformations produced do not exceed a certain limit .
This limit varies with the nature of the metal , and is , for nickel wire , very low .
2 .
The logarithmic decrement of amplitude increases with the length of the vibration-period , but in a less proportion than the latter , and in diminishing ratio .
The amount of increase of the logarithmic decrement attending on a given increase through a given range of the vibration-period varies with the nature of the metal , and with those metals in which the internal friction is small becomes nearly insensible .
3 .
It follows from 2 that the internal friction of a solid does not resemble the viscosity of fluids .
4 .
Permanent molecular strain resulting from loading , not carried to a sufficient extent to produce sensible permanent extension , diminishes the internal friction and increases the torsional elasticity .
5 .
Considelable permanent longitudinal extension and permanent torsion produce increase of internal friction and diminution of torsional elasticity .
The effect of torsion is much greater than that of extension , and the increase of internal friction is much greater than the decrease of torsional elasticity .
6 .
The internal friction of a metal is independent of the temporary load .
7 .
Annealing a hard-drawn wire may enormously diminish the internal friction , especially if , after annealing , the wire be kept suspended for some days and frequently vibrated .
8 .
Rest after suspension , aided by not too large oscillations at intervals , diminishes the internal friction of a wire which has been recently suspended , or which , after suspension , has been subjected to considerable molecular agitations by mechanical or thermal agency .
9 .
" " Fatigue of elasticity\ldquo ; is not felt , provided the deformations produced do not exceed a certain limit .
This hmit varies considerably with the nature of the metal , STRAIN ON THE PHYSICAL PROPERT ] OF MATTER .
being very low for nickel , so that with this metal is difficult to avoid ' ' elastic fatigue 10 .
Mechanical shocks and rapid fluctuations of tempelature beyond certain limits may increase very considerably the internal friction , and , though to a much less extent , diminish the torsional elasticity .
11 .
The logarithrnic decrement is independent both of the length and of the diameter of the wire .

