IV .
" On the Elasticity and Viscosity of Metals .
" By Prof. W. THOMSON , LL. D. , F.R.S. , F.R.S.E. Received May 18 , 1865 .
Among the experimental exercises performed by students in the physical laboratory of the University of Glasgow , observations on the elasticity of metals have been continued during many years .
Numerous questions of great interest , requiring more thorough and accurate investigation , have been suggested by these observations ; and recently they have brought to light some very unexpected properties of metallic wires .
The results stated in the present communication are , however , with one or two exceptions , due to the careful experimenting of Mr. Donald Macfarlane , official assistant to the Professor of Natural Philosophy , whose interested and skilful cooperation have been most valuable in almost everything I have been able to attempt in the way of experimental investigation .
The subject has naturally fallen into two divisions , Viscosity , and Moduli of Elasticity .
Viscosity.-By induction from a great variety of observed phenomena , we are compelled to conclude that Ino change of volume or of shape can be produced in any kind of matter without dissipation of energy .
Even in dealing with the ab8soltely perfect elasticity of volume presented by every fluid , and possibly by some solids , as for instance homogeneouis crystals , dissipation of energy is an inevitable result of every change of volume , because of the accompanying change of temperature , and consequent dissipation of heat by conduction or radiation .
The same cause gives rise necessarily to some degree of dissipation in connexion with every chanige of shape of an elastic solid .
But estimates founded on the thermodynamic theory of elastic solids , which I have given elsewhere * , have sufficed to prove that the loss of energy due to this cause is small in comparison with the whole loss of energy which I have observed in many cases of vibration .
I have also found , by vibrating a spring alternately in air of ordinary pressure , and in the exhausted receiver of an air-pump , that there is an internal resistance to its motions immensely greater than the resistance of the air .
The same conclusion is to be drawn from the observation made by Kupffer in his great work on the elasticity of metals , that his vibrating springs subsided much more rapidly in their vibrations than rigid pendulums supported on kniife-edges .
The subsidence of vibrations is probably more rapid in glass than in some of the most elastic metals , as copper , iron , silver , aluminium t ; but it is much more rapid than in glass , marvellously rapid indeed , in some metals ( as for instanice zinc ) t , and in india rubber , and even in homogeneous jellies .
" On the Thermo-elastic Properties of Solids , " Quarterly Journal of Mathematics , April 1857 .
t We have no evidence that the precious metals are more elastic than copper , iron , or brass .
One of the new bronze pen-nies gives quite as clear a ring as a two-shilling silver piece tested in the usual m-lan-ner .
f Torsional vibrations of a weight hug on a zinc wire subside so rapidly , that it Thefrictional resistance against change of slhape must in every solid be infinitely small when the change of shape is made at an infinitely slow rate , since , if it were finite for an infinitely slow change of shape , there would be infinite rigidity , which we may be sure does not exist in nature* .
Hence there is in elastic solids a molecularfriction which may be properly called viscosity o'fsolids , because , as being an internal resistance to change of shape depending on the rapidity of the change , it must be classed with fluid molecular friction , which by general consent is called viscosity of/ luids .
But , at the same time , it ought to be remarked that the word viscosity , as used hitherto by the best writers , when solids or heterogeneous semisolid-semifluid masses are referred to , has not been distinctly applied to molecular friction , especially not to the molecular friction of a highly elastic solid within its limits of high elasticity , but has rather been employed to designate a property of slow continual yielding through very great , or altogether unlimited , extent of change of shape , under the action of continued stress .
It is in this sense that Forbes , for instance , has used the word in stating that " Viscous Theory of Glacial Motion " which he demonstrated by his grand observationis on glaciers .
As , however , he , and many other writers after him , have used the words plasticity and plastic , both with reference to homogeneous solids ( such as wax or pitch even though also brittle , soft metals , &amp;c. ) , and to heterogeneous semisolid-semifluid masses ( as mud , moist earth , mortar , glacial ice , &amp;c. ) , to designate the property t commoln to all those cases of experiencing , under continued stress , either quite corntinued and unlimited change of shape , or gradually very great change at a diminishing ( asymptotic ) rate through infinite time , and as the use of the term plasticity implies no more than does viscosity any physical theory or explanation of the property , the word viscosity is without inconvenience left available for the definition I propose .
To investigate the viscosity of metals , I have in the first place taken them in the for-m of round wires , and have chosen torsional vibrations , after the manner of Coulomb , for observation , as being much the easiest way to arrive at defirnite results .
In every case one end of the wire was attached to a rigid vibrator with sufficient firmness ( thorough and smooth soldering I find to be always the best plan when the wire is thick enough ) ; has been found scarcely possible to count more than twenty of them in one case experimented on .
and the other to a fixed rigid body , from which the wire halngs , bearing the vibrator at its lower end .
I arranged sets of observations to be made for the separate comparisons of the following classes:(a ) The same wire with different vibrators of equal weights ( to give equal stretching-tractions ) , but different moments of inertia ( to test the relation between viscous resistances agailnst motions with different velocities through the same range and under the same stress ) .
( b ) The same wire with different vibrators of equal moments of inertia but unequal weights ( to test the effect of different longitudinal tractions on the viscous resistance to torsioni unider circumstances similar iu all other respects ) .
( c ) The same wire and the same vibrator , but different initial ranges in successive experiments ( to test an effect uniexpectedly discovered , by which the subsidelnce of vibrations from any amplitude takes place at very different rates according to the immediately previous molecular condition , whether of quiescence or of recurring change of shape through a wider range ) .
( d ) Two equal and similar wires , with equal and similar vibrators , one of them kept as continually as possible in a state of vibration , from day to day ; the other kept at rest , except when vibrated in an experimenit once a day ( to test the effect of continued vibration on the viscosity of a metal ) .
Results .
( a ) It was found that the loss of energy in a vibration through one ranige was greater the greater the velocity ( within the limits of the experiments ) ; but the difference between the losses at low and high speeds was much less than it would have been had the resistance been , as Stokes has proved it to be in fluid friction , approximately as the rapidity of the change of shape .
The irreguilarities in the results of the experiments which up to this time I have made , seem to prove that much smaller vibrations ( producing less absolute amounts of distortion in the parts of the wires most stressed ) must be observed before any simple law of relation between molecular friction and velocity can be discovered .
( b ) When the weight was illereased , 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 .
It has not yet been practicable to continue the experiments long enough in any case to find the limit to this variation .
( c ) The vibrationi subsided in aluminium wires much more rapidly from amplitude 20 to amplitude 10 , when the initial amplitude was 40 , than when it was 20 .
Thus , with a certain aluminiium Nvire , and vibrator No. 1 ( time of vibration one way 1P757 second ) , in three trials the numbers of vibrations counted wereVibrationis .
Vibrations .
Vibrations .
Subsidence from 40 initial am64 64 plitude to 20 ... ... ... .
And from 20 ( in course of the 6 98 96 same experiments ) to 10 .
9 The same wire and same vibrator showedSubsidence from 20 initial aniplitude to 10 1 ( average of four trials). .
12 vibrations .
Again the same wire with vibrator No. 2* ( time of vibration one way 1'236 ) , showed in two trialsVibrationis .
Vibrations .
Subsidenice from 40 initial amplitude } 54 52 to 2 ... And continued from 20 to 10 ... 90 90 Again same wire and vibrator , From initial anmplitude 20 to 10. .
103 vibrations ( mean of eight trials ) .
This remarkable result suggested the question ( d ) .
( d ) Only one comparisoni was made .
It showed in a wire which was kept vibrating nearly all day , from day to day , after several days , -very much more molecular friction than in .
another kept quiescent except during each experiment .
Thius two equal and similar pieces of wire were put up about the 26th of April , hanging with equal and similar lead weights , the tops and bottoms of the two wires being similarly fixed by soldering .
No. 2 was more frequiently vibrated than No. 1 for a few days at first , but no comparisoln of viscosities was made till May 15 .
Then No. 1 subsided from 20 initial ranige to 10 in 97 vibrations .
No. 2 , the same suibsidence in 77 vibrations .
During the greater part of A-ly 16 and 17 , No. 2 was kept vibrating , and No. 1 quiescent , and late on May 17 experiiments with the following results were made Timiie per vibration .
No. 1 .
Subsided from 20 to 10 after 99 vibrationis in , 237 seconds ... ... 2-4 98 , , 235 , ... ... 24 98 , , 235 , ... ... 24 No. 2 .
Subsided fromii 20 to 10 after 58 , , 142 , ... ... 2-45 , , , , , , 60 , .
147,.245 57 , , 139 , ... ... 2-45 60 , , 147 , ... ... 245 [ Addition , i 'a.y 27 , since the reading of the paper.]-No .
1 has been kept at rest from May 17 , while No. 2 has been kept oscillating more or less every day , till yesterday , May 26 , when both were oscillated , with the following results Time per vibration .
No. 1 .
Subsided from 20 to 10 after 100 vibrations in 242 seconds .
242 , , 2 .
, , , , 44 or 45 vibrations ... ... ... ... ... ... 2-495 Moduli of Elasticity.-A modulus of elasticity is the nuiimber by which the amount of any specified stress , or component of a stress , must be divided to find the strain , or any stated componenit of the strain , which it produces .
Thus the cubic compressibility of water being w-1l - , { per atmosphere , its " moduluis of compressibility " or its c volume modulus of elasticity , " is 21000 atmospheres , or 76 x 13-596 x 21 O0( ) =2 17x 10 ' grammes weight *Of saime weight.as Nxo .
1 , but lifferent miioment of inertia .
per square centimetre ( as 13'596 is the density or specific gravity* of mercury , and 76 centimetres the height of the barometric columni corresponding to the pressure defined as " one atmosphere " ) .
Or , again , Young 's " modulus , " which has generally been called simply the modulus of elasticity of a solid , is the longitudinal traction of a stretched rod or wire of the substance , divided by the extension produced by it .
Or , lastly , the " modulus of rigidity , " or , as it is conveniently called , simply " the rigidity " of an isotropic solid , is the amount of tangential stress divided T by the deformation it prodJices , the former being measured in units of force per unlit of area applied , as showni T in the diagram , to each of four faces of a cube , and the T latter by the variation of each of the four right angles , L_ reckoned in circular measure .
T Measurements of Young 's mooduluis have been made for many bodies by many experimenters ; but hitherto there have been very few determinations of rigidity , notwithstanding the great ease with which this can be done for wires by Coulomb 's method .
Accordingly , although several accurate determinations of Young 's modulus have been made uponx wires of different substances hung in the College Tower of the University of Glasgow ( which , by giving 80 feet of clear protected vertical space , affords great facilities for the investigation ) , I shall in this paper only refer to some of the results as bearing on the question , how are moduli of elasticity affected in one substance by permanentt changes in its molecular condition ?
which was my startingpoint for all I have attempted to do experimentally regarding the elasticity of solids .
To determine rigidities by torsional vibrations , taking advantage of an obvious but milost valuable suggestion made to me by Dr. Joule , I used as vibrator in each case a thiin cylinder of sheet brass , turned true outside and inside ( of which the radius of gyration must be , to a very close degree of approximation , the arithmetic mean of the radii of the outer and inner cylindrical surfaces ) , supported by a thin flat rectangular bar , of which the square of the radius of gyration is one-third of the square of the distance from the centre to the corners .
The wire to be tested passed perpendicularly through a hole in the middle of the bar , and was there firmly soldered .
The cylinder was tied to the horizontal bar by light silk threads , so as to hang with its axis vertical .
The following particulars show the dimensions of the vibrators of this kind which I have used .
__oment of Cylinders .
~~~~~~~~~~~~~~~~inertia round |Cylinders .
Outer diameter .
Inner dian-cter .
j Mean radius .
XVeight in axis in grammies .
grammecentimetres .
No. 1 15 3 ceintimns .
1]4 8 centimiis .
7 525 527P92 29894 , , 2 15 3 14-8 , , 7525 523 45 29641 , , 3 10 295 979 , , 5 021 360-54 9089 , 4 10 3 9-81 , , 5-027 726-40 18357 , ,9 5 10-25 9 745 , 4-999 718 36 17952 , , 6 10295 9 805 , 5-025 342-45 8647 !
Aoment of inertia round axis througlI Length .
Brcadtli .
Weight .
middle , perpendiclr to length and breadth .
Bar 1 24-03 centims. 965 centim. 38-955 grms. 1877 5 , 2 24 11 , *95 , , 46-68 , , 2255-5 Towards carrying out the chief object of the inivestigation , each wire , after having been suspended and stretched with just force enough to make it as nearly straight as was necessary for accuracy , was vibrated .
Then it was stretched by hand ( applied to the cross bar soldered to its lower end ) and vibrated again , stretched again and vibrated again , and so till it broke .
The results , as shown in the following Table , were most surprising .
Length of T ~~~~~~~~~~~Rigidity , in Vegholum , n oment ofvirtoI wire , in ubn-lieti flI gramimies weight centimeEtiietres sity .
v Yibrator ( hapl pers ae Substa.ces i tres , peri ' od ) , cetmr , T. gT2V2_ 60 3 1-1845 2-764 31771 114 241 X 106 Alumiiituia .
304-9 2 351 1 7105 31896 4 31 3596 x 106 Zinc b. 237 7 4 ... ... ... ... .
I4 76 410 3x 106 | Brass .
248 3 5 ... ... ... .
I5 456 3548 X 106 , Ptemcuiks .
a Only forty vibrations from initial aic of convenient amplitude could be couinted .
Ila(d been stretched considerably before this experiment .
So viscous that only twernty vibrations could be counted .
Broke in stretchiii( .
TABLE ( continued ) .
Time of Length of vbain Rgdt , i wire , in Volume , in Moment one way Rigidity , in centime .
cubic cenDest .
inertia of onea rammsquweit tree , ietres , enty vibrator , ( or halfpe sqae Substances .
tres , ~~~~~~~period ) , centimetre , V. Wk2 .
2.,.2 1 .
~~~ ~~~Wk .
in seconds , 2r/ jk T. gTV 261-9 1P703 8-398 ... ... 5-96 350 1X 106 Brass .
2435 0 15-30 8-91 38186 16-375 448-7 x 106 Copper .
, , 61412 20-77 448.4 X 106 214 4 1-348 8-864 31771 5.015 433-0x106 Coppere .
, , ... , , 61412 6-982 431-8 x 106 143-7 -9096 8-674 3-381 393-4 x Io( Copper d. 286-8 ... ... ... ... 20612 4a245 442-9 x 106 Coppere .
291 ... ... ... ... .
i 4-375 435-6 X o6 ' , , 293 ... ... ... ... 4-417 436-2 X 106 it 296-1 ... ... ... ... 4500 433-8 x 106 , 300-0 ... ... ... ... 4-588 434-0 X 106 i 303-4 ... ... .4 ... . .
4646 437-8 x 106 , , 309-3 ... ... ... ... 4-833 428-6 x 106 , 313-2 ... ... .4 ... . .
4931 427-5 x 106 317-4 1-962 8-835 , , 5 040 425-9 x 106 , , 215-6 ... ... ... ... 31771 8-155 442-3 x 106 Copper f. 235-5. .
9-425 432-2 X06 , , 251-9 -827 8-872 , , 10-463 428-6 X 106 , 253-2 1-580 8'91 ... ... 5-285 472-9 x 106 Copper .
262-8 ... ... ... ... ... ... 5-640 464.3 x lo , , 270-4 ... ... ... ... ... ... 5-910 460-4 x 106 , , 278 -7 - ... ... ... . .
620 458-5 X 10 ?
, , 287-9 ... ... ... ... ... ... 6-5325 455-0 X 105. .
297 5 ... ... ... ... ... ... .
68195 451-0X106 308-8 ... ... ... .
73075 448-9 x106 , , 256-5 1-6145 8-90 4-226 463-5 X 106 Copperh .
267-9 ... ... ... .
45625 453-3 X 105 , , 280-1 ... ... ... ... .
4915 446-2 x 106 , 292-2 ... ... ... ... ... ... 5-240 445-5 x 106 !
301-9 ... ... ... ... ... ... 5-532 438-2X10 ?
, , 316-8 ... ... ... ... ... ... 6-655 791-4 x 103 Soft iron i. 322-1 ... ... ... ... ... ... 6-88 778 3X 106 , , 335-1 ... ... ... ... ... ... 7-301 779-0 x 10 ?
, 347-4 ... ... 7.768 766-6 X 10 , 366-0 1-357 7G57 8-455 756-0 X 1063 39-4 *1745 20-805 20612 2-05 622-25 X 106 Platinum k. 65-9 -1895 19-8 1 0902 ... ... 281 X 106 Gold ' .
75-7 *1185 10-21 10967 270 x 106 Silver .
Read703s .
A piece of the preceding stretched .
d The preceding made red-hot in a crucible filled with powdered charcoal and allowed to cool slowly , became very brittle : a part of it with difficulty saved for the experiment .
e Another piece of the long ( 2435 centims. ) wire ; stretched by successive simple tractions .
fA finer-gauge copper wire ; stretched by successive tractions .
C Old copper wire , softened by being heated to redness and plunged in water .
A lengthof260 centims. cutfi-om this , suspended , and elonigated by successive tractions .
h ' Another length of 260 centims. cut from the same and similarly treated .
One piece , successively elongated by simple tractions till it broke .
k Not stretched yet for a second experiment .
Added , May 27 , after the reading of the paper .
Thus it appears that that specific rigidity which is concerned in torsioni is very markedly diminished in copper , brass , and iron wire when the wire is elongated permanently by a simple longitucdinal traction .
When I first observed indications of this resuilt , I suspected that the diminution in the torsional rigidity on the whole length of the wire might be due to inequalities in its niormal section produced by the stretching .
To test this , I cut the wire into several pieces after each series of experiments , and weighed the pieces separately .
The result prov ethat in nlo case were there any such inequalities in the gauge of the wire in different parts as could possiblyaccount for the diminutitionin the torsional rigidity of the whole , which was thus proved to be due to a real diminLution in the specific rigidity of the substance .
The following , sets of weighings , for the cases of the wires of the two last series of experimclnts on copper , may suffice for example WVire of 308-8 centirns .
long , cut into four pieces .
) Lltilllength , in WI ; tesight , ~ Weight per centhimetres .
gl"me .
cenitimetrec , gramme , i granaiines .
No. 1 109 25 023 -04600 2 66 3 050 *0457/ 3 3 63 22865 *04.533 46 94 3 143 -045 17 308 5 14 408 1 Wire of 301-9 when last vibrated ; further elongated by about 8 centimetres , when it broke ; then cut into five pieces in all .
Length , in WTeight , ill Weight per I cenitmietres .
graiimes .
cenititnetre , in grammes .
No. 1 66i3 3-183 *04801 2 66-4 3 083 -04643 36 63 039 -04570 4 668 3 072 04599 5 43X4 1-986 -04576 By several determinations of observations on the elongations within the limnits of elasticity produced by h-anging weights on long wires ( about 80 feet ) suspended in the College tower , it seemed that Youlng 's modulus was not nearly so much ( if at all sensibly ) altered by the change of molecular condition so largely affecting the rigidity ; but this question requires further investigation .
The amount of the Young 's modulus thus found was , in grammes weight per square cenitimetre , 1159 x l10 for one copper wire , and 11 53 X 106 for another which had beeni very differently treated .
The highest and lowest rigidities which I have found for copper ( extracted from the preceding Table ) are as follows:_ Highest rigidity , 473 x 106 , being that of a wire which had been softened by heating it to redniess and plunging it into water , and which was found to be of density 8 91 .
Lowest rigidity 3934 x 106 , being that of a wire which had beeni rendered so brittle by heating it to redness surrounded by powdered charcoal in a crucible and letting it cool very slowly , that it could scarcely be touched without breaking it , and which had been found to be reduced in density by this process to as low as 8 674 .
The wires used were all commercial specimens-those of copper being all , or nearly all , cut from hanks supplied by the Gutta Percha Company , having been selected as of high electric conductivity , and of good mechanical quality , for submarine cables .
It ought to be remarked that the change of molecular condition produced by permanently stretching a wire or solid cylinder of metal is certainly a change from a condition which , if originally isotropic , becomes aeolotropic* as to some qualitiest , and that the changed conditions may therefore be presumed to be eolotropic as to elasticity .
If so , the rigidities corresponding to the direct and diagonal distortions ( inidicated by No. 1 and No. 2 in the sketch ) must in all probability become different from one another when a wire is permanently stretched , instead of being equal as they must be when its substance is isotropic .
It 7 becomes , therefore , a question of extreme interest to find whether rigidity No. 2 is not increased by this process , which , as is proved by the experiments above described , diminishes , to a very remarkable NO .
I degree , the rigidity No. 1 .
The most obvious experiment , and indeed the only practicable experiment , adapted to answer this question , will require an accurate determination of the difference produced in the volume of a wire by applying and removing longitudinal traction within its limnits of elasticity .
With the requisite apparatus a most important and interesting investigation might thus be made .

