294 Prof. A. Gray and Mr. A. Wood .
Effect of a [ May 1 12 .
By touching different points of the wire with different reagents , the excitability of these portions are rendered unequal .
Hence a resultant electromotive variation may be obtained by vibrating the wire as a whole .
The current in the wire is from the less to the more excitable .
13 .
By this method , invisible traces of physico-chemical change in a wire may be detected .
14 .
Chemical reagents not only change the excitability but the quickness of response .
Two points having two different rates of excitation will thus , under proper conditions , give rise to diphasic effects .
I take this opportunity to thank the Managers of the Royal Institution for the facilities offered me to carry on the investigation at the Davy-Faraday Laboratory .
" On the Effect of a Longitudinal Magnetic Field on the Internal Viscosity of Wires of Nickel and Iron , as shown by Change of the Rate of Subsidence of Torsional Oscillations .
" By Professor Andrew Gray , F.R.S. , and Alexander Wood , B.Sc. , Houldsworth Research Student in the University of Glasgow .
Received May 1 , \#151 ; Read-May 15 , 1902 .
We can obtain information as to the nature of the magnetisation of magnetisable bodies only by testing the various hypotheses with reference to effects which it seems likely should , under these hypotheses , be produced on the physical properties of the substance .
Thus , for example , the internal friction of the different parts of a solid must depend upon the size and mode of arrangement of these parts , and any alteration in their dimensions or relative arrangement ought in general to produce some change in the amount of the internal friction .
Magnetisation of iron and other substances has with great probability been supposed to consist in a rearrangement and general alignment of the particles of the substance , already themselves elementary magnets , but so arranged in the unmagnetised metal as to be unproductive of any external magnetic field .
It is not unusual to suppose that this unmagnetised state is one of what we may call complete absence of arrangement , and it is sometimes so represented in text-books on the subject of magnetism , where pictures are given of a perfectly confused distribution of elementary magnets , so completely mixed up as to have no preponderating magnetic moment in any one direction .
Any such distribution , it is clear at once from the most elementary considerations , is impossible , as a large majority of the elementary magnets would otherwise have to remain in stable equilibrium in 1902 .
] Longitudinal Magnetic Field on Internal Viscosity .
295 unstable conditions of magnetic forces , a state of things which could only be brought about with the assistance of forces independent ot magnetic action\#151 ; forces which there is otherwise good reason to believe do not exist to anything like the degree which would be necessary .
There are other reasons for believing that in the unmagnetised metal the elementary magnets have a perfectly orderly arrangement in closed chains , each of which may be composed of a large group of molecules of the substance .
These molecules , when the substanee is unmagnetised , form a closely coherent arrangement which is broken up into an entirely different collocation under the influence of an externally applied magnetic field .
Thus if the wire of the unmagnetised material perform torsional oscillations , the relative motions of the parts of the material will naturally be affected by the internal forces existing between the parts , and by the arrangement of the parts in more or less distinct chains or groups .
It is reasonable to suppose that the effect of magnetisation being to entirely alter these internal magnetic forces , and to break up and rearrange the groups o elementary magnets , will be also to alter considerably the internal friction shown by the material .
With this idea in view , we have subjected wires of iron , nickel , and steel to longitudinal magnetising forces , while they were compelled to perform torsional oscillations round their axes , and have observed the rate of subsidence of these oscillations with magnetising forces of different values , the amounts o which will be found stated in curves illustrative of the results obtained .
The effects of the fields were found to be considerable in amount and very curious in character .
For example , in the case of iron , as the material was more intensely magnetised , the rate of subsidence of the oscillations continually diminished ; on the other hand , with wires of nickel in the state in which they were received from the maker the effect was to increase the rate of subsidence until a certain value of the magnetising force was reached , after which the rate of subsidence l diminished with further increase of the magnetising force .
Ihese s : results were modified considerably by alteration in the state of the material , produced by drawing the wires through a draw-plate .
Since the completion of the experiments described m this paper , we have become aware of a paper by Mr. Herbert Tomlinson , on the subject of the effect of magnetisation on the internal viscosity of iron* The results obtained by Mr. Tomlinson , show an increase of viscosity produced by a field of 35 units , which is contrary to the effects described below .
But the extreme amplitude in Mr. Tomlinson 's experiments was only 10 for a wire 1 metre ong , as against 90 ' in our experiments ; and as we have pointed out below , the effect seems to depend in an important way upon the amplitude , thoug * * Phil. Trans. , ' A , vol. 179 , 1888 .
296 Prof. A. Gray and Mr. A. Wood .
Effect of a [ May jf whether it changes sign in our specimens at very low amplitudes we are as yet unable to say .
It will be convenient to give here a short description of the apparatus , and then to state in detail the results obtained from the experiments made on the three substances referred to .
The specif mens were wires about 1 metre long , and about 1*3 mm. in diameter The exact dimensions are given below where the particular specimens are referred to in the arrangement adopted .
The wire is suspended vertically along the axis of a magnetising coil A ( Diagram I ) .
Across the upper end of the coil is placed a cross-piece of brass , to which the upper end of the wire is attached , while the coil itself is supported by the upper cross-bar of a Willis frame which stands on the laboratory ' Diagram I. floor .
About 3 or 4 cm .
below the lower end of the coil , the wire is attached to a brass fitting cemented on to the upper end of a glass 1902 .
] Longitudinal Magnetic Field on Internal Viscosity.297 rod B about a metre long and 5 mm. in diameter .
To the lovrer end of this glass rod is rigidly fastened a cylindrical vibrator C. The axes of the wire , the glass rod , and the vibrator are as nearly as possible in ,0ne vertical straight line , and the three bodies are rigidly connected , Lo that if the vibrator is turned round its axis , no slipping takes .place at the attachments .
The wire and rod are thus subjected to ( twist , the amount of which is infinitesimal in the glass rod , which turns round its axis practically as a rigid body ; so that the whole twist may be regarded as contained in the wire .
This arrangement was adopted to keep the vibrator , which in all cases was of iron ( being selected from a series of vibrators in the collection of apparatus ) , sufficiently remote from the coil to obviate the possibility of any action between the field of the coil and the moving vibrator , while the whole of the twisting and untwisting material was contained within the field .
Bound the lower edge of the vibrator is gummed a millimetre scale , which is read by a telescope , and the vibrations were always sufficiently slow to allow of the scale being read at the beginning and end of a semi vibration , the difference between the two readings giving the amplitude .
The coil consisted of 3080 turns of copper wire wound on a double core of brass tubing .
This double core was made of two coaxial tubes , the outer 1^ inches and the inner \#163 ; inch in diameter .
The space between the tubes forms a water-jacket , through which a stream of water can be forced so as to shield the wire from the heating effect of the current in the coil .
The current produced by the battery E , and adjusted by the resistance D , was measured by a Kelvin graded galvanometer F , and the field was calculated from the value AxrCn for the intensity of the field produced by a cunent of C absolute units flowing in a coil of 'n turns per cm .
of length .
Results far Nickel.\#151 ; The wire used was obtained from Messrs. Johnson and Matthey , London , and was stated by them to contain only a very small percentage of impurity .
Its diameter was 1 4 mm. The results of the experiments are shown in Diagram II .
In the curves of that diagram the abscissae , drawn from left to right , show amplitude of vibration in degrees ) the ordinates , drawn downwards , represent the number of periods which have elapsed from the beginning of the set of observations .
Thus we obtain the amplitude left after different numbers of oscillations have been performed , and are able to estimate from the curve the rate of subsidence .
The ordinates may be taken as giving the time from the commencement of the set of observations , as it was found that the alteration of period produced by the imposition of the field was negligible , the actual periods being 7'21 seconds : with a field of 132 C.Gr .
S. and 7*17 seconds with no field .
It will be seen that the effect of small fields is to greatly increase the rate of subsidence , but that at a field of about 160 C.G.S. the maximum effect of the field in increasing the rate of subsidence is produced : thus 298 Prof. A. Gray and Mr. A. Wood .
Effect of a [ May 1 , Curves 1 , 2 , 3 , 4 for fields up to 156 C.G.S. , lie successively farther to the left in the diagram , while the Curves 5 , 6 , 7 , 8 for fields from 187 to 360 C.G.S. lie in succession to the right of one another , showing in the first case a continual increase , and in the second a continual diminution in the rate of subsidence .
AmpUbude .
Diagram II .
Curve , Field .
Curve .
Field .
1 0 C.G.S. 5 187 C.G.S. 2 22 " 6 257 " 3 50 " 7 330 " 4 156 " 8 360 " A curve for zero field was taken after these eight curves had been obtained , and it was found to agree closely with Curve 1 of the series shown in the diagram .
The logarithmic decrements were calculated for these curves and were found to diminish in each curve as the amplitude diminished .
This diminution was about half the initial value of the logarithmic decrement for each of the first four curves , but then fell off until in curve 8 it was only i of the value at the initial amplitude .
A series of experiments was made for the purpose of finding more precisely the value of that magnetising field which had maximum effect upon the rate of subsidence , and curves of subsidence were obtained for values of the field ranging from about 100 to about 200 C.G.S. The curves form a close series , and intersect one another in a way which it is rather difficult to disentangle , and which renders the critical value of the magnetic field in a sense indefinite .
The crossing of the curves in the immediate neighbourhood of the critical 1902 .
] Longitudinal Magnetic Field on Internal Viscosity .
299 field was traced to a slight extent to small uncompensated effects of heating but it is extremely unlikely that the crossing of the curves | due to this cause alone , as no doubt the effect of the magnetic field on the rate of subsidence depends upon the amplitude .
Further experiments with more nearly perfect prevention of heating effects , are required to enable the dependence of the rate of subsidence on Cuplitude to clearly disclose itself .
It may be mentioned , however , .that the actual variations in temperature , as determined by observations of the temperature of the water issuing from the jacket , were very slight For example , the temperatures at the end of each of nine successive experiments\#151 ; the first and last of which were made with no current in the coil , and the others with currents from 2'56 to 6 amperes , varied from 6''5 G. to 8 ' C. It will be noticed that in Diagram II the shape of the curves undergoes change as the magnetic field is increased .
This is most obvious near the critical value of the field , as shown by Curves 4 and 5 .
For Curve 4 the rate of subsidence at first is distinctly greater than m | Curve 5 , but as the amplitude diminishes the two curves approach one | another , so that while , after ten , twenty , and thirty vibrations there is distinctly more amplitude left in Curve 5 than in Curve 4 , there is practically the same amplitude left after fifty vibrations , the curves being then on the point of crossing .
This effect seems , as we have said , to be , in part at least , due to the dependence of the magnetic effect on amplitude .
FieLot .
Diagbam III .
Prof. A. Gray and Mr. A. Wood .
a [ May j ?
f Diagram III brings out the point just alluded to more clearly . !
These curves show the effect of the different fields upon the amount of subsidence effected in the times corresponding to different numbers of vibrations and to different amplitudes .
For example , Curve 1 gives a comparison of the subsidences ( from an initial amplitude of 90 ' ) 1 effected in twenty vibrations under different fields varying from about 100 to 200 C.G.S. Curve 2 gives a comparison of the subsidences from the same initial amplitudes effected in thirty vibrations for the same rahge of field .
Curves 3 and 4 show the subsidences for forty-five and seventy-five vibrations respectively in the same circumstances .
,1 It will be seen that the points of maximum of the curves trend to I the right from Curve 1 to Curve 4\#151 ; that is to say , the value of the field which produces maximum effect is greater the smaller the j amplitude of the vibration to which it is applied .
fl Results for Iron.\#151 ; Diagram IV shows curves of subsidence obtained for iron .
Six curves are given showing the results for fields varying AmpLibude .
Z/ 3/ 4/ /6 : Diagkram IY .
Curve .
Field .
Curve .
Field .
1 0 C.G-.S .
4 81 O.G-.S .
2 22 " 5 170 " 3 44 " 6 229 " from 0 to 230 C.G.S. It will be seen that the curves fall successively to the right of one another with the exception of 5 and 6 , which are practically coincident .
Experiments were also made for fields up to " 1902 .
] Longitudinal Magnetic Field on Internal Viscosity .
301 i 400 C G S. , but the curves obtained showed only a very slight tendency to move to the right , so slight that if they had been drawn Tev could not have been distinguished in the diagram from Curves 5 and 6 The same vibrator was used for the iron wire as had been used in the case of nickel , and the period of vibration was 5-27 seconds .
For iron , then , we have the remarkable result that the effect of a magnetic field on the rate of subsidence of torsional oscillations is to diminish that rate , and by an amount diminishing with increasing e until a field of about 160 or 170 C.G.S. is reached , after which the effect of the field , however great , is practically constant .
This is , of course , what one would expect in consequence of saturation of the iron , but i is a result very remarkably different from that obtained for mcke .
The totally different behaviours in the two cases seem to point to an entirely different collocation of elementary magnets m the nickel and the iron , i.e. , perfectly distinct molecular constitution , consistent , however , in each case with magnetisability .
Results for Steel\#151 ; Diagram V shows three curves obtained for pianoforte steel wire .
The same vibrator was employed as m the former Amplitude .
Diagkam Y. Curve .
Yield .
1 0 C.G.S. 2 HO " 3 230 " cases , and it gave with this wire a period of .16-81 seconds .
The fields were respectively 0,110 , and 230 C.G.S. , and it will be noticed that the results are very similar to those obtained for iron , except t .e 302 Effect of a Longitudinal Magnetic Field on Internal Viscosity ] effects are much smaller .
The curves for higher fields moved coni tinually to the right , and showed no tendency to become coincident i In both iron and steel the logarithmic decrements , as the curves ' show , were much more nearly constant throughout each curve than was the case for nickel .
In steel the decrement fell off rapidly at first so that throughout the greater part of the curve it was nearly constant at a distinctly smaller value than at the beginning of the curve Experiments were also made with a view to ascertaining whether the change in the physical properties of the wires produced by drawing them through a draw-plate altered in any way the effects of magnetisation on the rate of subsidence .
In the case of drawn nickel wire the most striking alteration was the disappearance of the reversal of the effect of magnetisation .
The magnetisation increased slightly the rate of subsidence , and this increase continued quite regular to the highest values of the field obtainable ( about 400 C.Gr .
S. ) , no critical value of the field similar to that observed for the undrawn wire being obtained .
In the case of drawn iron , on the other hand , the effect of magnetisation was to regularly diminish the rate of subsidence .
The effect was thus similar to that obtained for the undrawn wire , the chief point of difference being that , in the case of the drawn wire , further increase of the field gave farther perceptible diminution in the rate of subsidence , so that there was not that approximation to constancy of effect of field , apparently depending on magnetic saturation , which was observed in the case of the undrawn wire .
The drawing of the wires was found to result in a considerable increase of their rigidity .
As a check upon the experiments detailed above , a copper wire was substituted for the iron or nickel wire along the axis of the coil , and a field of 230 C.Gr .
S. units was applied .
No perceptible effect upon the rate of subsidence was found to be produced .
\Note , added May 30 , 1902.\#151 ; The effect of annealing the nickel wire which had been hardened by drawing has now been observed .
The wire was heated to a bright red heat and then suddenly cooled by being plunged into water .
It was found that the wire had returned to its former state , as all the features of the curves of Diagram II were again observed , with the difference that the curve of maximum rate of subsidence occurred with a field of only about half the former value .
Experiments on the effect of annealing the iron wires are in progress , and show that the effect of annealing , by heating and slow cooling , is to annul the change produced by drawing .
]

