A10 Prof. A. Ewing .
[ June 10 , VII .
" On certain Effects of Stress on Soft Iron Wires .
" By J. A. Ewing , B.Sc. , F.R.S.E. , Professor of Mechanical Engineering in the University of Tokio , Japan .
Communicated by Fleeming Jenkin , F.R.S. , Professor of Civil Engineering in the University of Edinburgh .
Received May 24 , 1880 .
The following is a preliminary notice of some of the results of experiments which are still in progress .
As I hope to publish later a ^detailed account of the whole investigation , it is unnecessary to describe here , in more than brief outline , the apparatus which has been employed .
The wires experimented on were of soft Japanese iron , drawn in Japan , and annealed shortly before being tested .
Each wire was -stressed separately by being hung vertically from a strong frame , a tank capable of holding 100 kilos , of water being attached to the lower end of the wire .
The weight of the empty tank was balanced , so that the stress on the wire was due only to the weight of water in the tank .
The tank was circular and of uniform diameter from top to bottom , and the stress on the wire was therefore proportional to the height of water in the tank .
This height " vyas recorded -on a sheet of paper drawn along by a float , which rose as the water was run in .
A pencil travelled transversely across the paper through distances proportional to the elongation of the wire .
In this way a continuous diagram was automatically drawn , showing the relation of the strain or elongation to the stress from zero stress up to the .breaking point .
Fio .
1 . .
Stress increased uniformly up to breaking point , at the rate of 1 kilogramme in 15 seconds .
Stress on wire in kilogrammes .
1880 .
] On certain Effects of Stress on Soft Iron Wires .
Water was run into the hanging tank from a cistern at a considerable height above it .
This gave a convenient means of applying stress slowly or quickly at will , and at a sensibly uniform rate .
When not otherwise specified , the rate of flow of water into the tank is to be understood to have been about 1 kilo , in 15 seconds .
The iron wires were IT millim. in diameter , and the length undei test was 250 millims. .
, Fig. 1 shows one of the diagrams of a continuous test , m which the flow of water was continued uniformly at the above rate until the wiie broke .
The horizontal ordinates give the stress in kilogrammes , the vertical ordinates give the elongation in percentages of the original leno-th .
The line is sensibly straight until the stress is about 21 kilos .
; after that it curves rapidly but rather irregularly .
The breaking stress in this example was 37'5 kilos , and the final elongation 16*3 per cent. Fig. 2 shows the effect of maintaining a constant stress on the wire for a considerable time .
When the stress reached 35 kilos , the flow of Fin .
2 .
Showing effect of maintaining a stress of 35 kilogrammes constant for 45\#163 ; hours .
water was stopped , and this load was allowed to remain on the wire for an interval of 45| hours , after which the flow of water was resumed at the same rate as before , and continued until the wire broke .
T ie effect was most remarkable .
Instead of continuing to lengthen at once when the flow of water was resumed , the wire refused to stretch further until the stress rose to about 40 kilos .
There was , in fact , a new " limit of elasticity " at this point .
When the stress exceeded 40 kilos , the wire ran down rapidly , and finally broke with a load of 41 kilos , and a total elongation of 13\#163 ; per cent. The breaking strength was therefore greater , and the elongation less , than w en t e wire was tested continuously .
This effect , which I shall call the " hardening effect , " depends on the length of time during which the test is interrupted .
Fig. 3 shows 512 Prof. A. Ewing .
[ June 10\gt ; the effect when the same stress ( 35 kilos .
) was maintained constant for 5 minutes only .
I have also obtained diagrams showing the effect of keeping a load of 35 kilos , on wires for 2 minutes , 10 minutes , Fig. 3 .
Showing effect of maintaining a stress of 35 kilogrammes constant for 5 minutes .
30 minutes , 1 hour , 3 hours , and 141 hours .
In all these cases the hardening effect is greater for greater lengths of time , but the increase of the effect is much more rapid at first than it is later .
A curve showing the relation of the hardening effect ( measured by the length of the straight line in the diagram after the stress has been kept constant , and before the wire begins to stretch again ) to the time rises rapidly at first , but afterwards becomes nearly parallel to the line along which time is measured .
Fig. 4 .
Showing effect of removing the load when it readied 35 kilogrammes and replacing it after 19| hours , during which the wire was free from stress .
An interval of constant stress of even five seconds produces a per1880 .
] On certain Effects of Stress on Soft Iron Wires .
ceptible " knee " in the curve similar to , hut of course much smaller than , that shown in fig. 3 .
Experiments have also been made to determine the effect produced by completely relieving the wire of stress for a time , and then continuing the test .
Fig. 4 is an example of the effect observed .
In this example , when the stress reached 35 kilos .
, the flow of water into the tank was stopped , and the tank was emptied by opening a stop-cock at the bottom .
This brought the line in the diagram back to the line of no stress .
The wire was then left without stress for 19| hours , after which water was run into the tank again , at the usual rate , until the wire broke .
The result , as may be seen from the diagram , was that no considerable lengthening took place until a stress far exceeding 35 kilos , was reached .
It was not until the stress reached nearly 40 kilos , that the wire began to stretch again , and then it stretched rapidly , and broke with the comparatively small final elongation of less than 13 per cent. Here the interval of no stress had an hardening effect exceedingly similar to the effect produced by an interval of constant stress .
In the test shown in fig. 4 there was a gradually diminishing stress on the wire while the tank was being emptied , and a gradually increasing stress on it while the tank was being refilled .
To show that the hardening effect observed was not due to these stresses , I made the experiment whose diagram is given in fig. 5 .
Here the wire was Fia .
5 .
Showing effect of removing the load when it reached 35 kilogrammes , and replacing it without leaving any sensible interval during which the wire was free from stress .
stressed up to the same point as before ( 35 kilos .
) , the water was then run out of the tank at the same rate as in fig. 4 , and immediately run in again .
The circumstances of this experiment were , therefore , identical with those of the preceding one , except that now there was no sensible interval during which the stress was There was in this case a small hardening effect , such as might have been anticipated 514 On certain Effects of Stress on Soft Eon Wires .
[ June 10 from the results mentioned above , but nothing like so much as in fig. 4 .
The difference can only be due to the long interval of no stress .
An investigation is now being made of the hardening effects of equal intervals of constant stress and no stress at different points along the curve .
So far as the experiments have yet gone , it appears that outside the so-called elastic limit , that is , for stresses greater than about 22 kilos .
, the effect of 19 hours of constant stress is not greatly different , whether the stress is great or small .
Inside the elastic limit the effect of constant stress is to prolong that limit , that is , the wire then requires a greater stress to make it begin to stretch considerably .
This effect is greater the nearer to the elastic limit the interval of constant stress occurs .
By intermitting the flow of water for a succession of short intervals during one test we get a stepped curve , having a number of " knees " upon it , like those in figs. 2 and 3 .
A wire to which a continuously but very slowly increased stress is applied , gives a curve in which the elongation corresponding to each stress ( beyond the elastic limit ) is less than when the increase of stress is rapid .
In other words , when the flow of water is very slow , the curve is bent upwards to the right .
The final elongation is diminished , and the breaking strength is increased .
It will be observed that the last-mentioned results are in perfect agreement with the remarkable experiments described by Mr. J. T. Bottomley , in the Report of the British Association Committee for commencing secular experiments upon the elasticity of wires ( 1879 ) .
When a stress is applied and maintained constant it produces two effects .
It produces a gradual viscous elongation , at first rapid , after-wai*ds slow .
This I shall call for brevity the viscous effect .
It also produces the hardening effect described in this paper .
The hardening effect , like the viscous effect , is greatest at first .
The viscous effect is quite visible in the diagrams .
Thus , in figs. 2 and 3 , especially in fig. 2 , the wires have stretched considerably during the time that the stress has been constant ; and in figs. 4 and 5 the wires have gone on stretching during part of the time that the water took to run out of the tank .
If the viscous effect existed alone , and there were no hardening effect , a wire would show a higher breaking strength and less final elongation the more quickly the increase of stress took place .
For , with a quickly increased stress , the wire would not have time to take the elongation , and consequently the diminished cross-section proper to each value of the stress .
If , on the other hand , the hardening effect existed alone , a wire would show a higher breaking strength and less final elongation the more slowly the increase of stress took place .
1880 .
] Mr. R. C. Row .
Memoir on Abel 's Theorem .
We may , therefore , expect that whenever the viscous effect is great relatively to the hardening effect , a wire will show to greater advantage as regards its breaking strength , and to less advantage as regards its elongation , when tested quickly than when tested slowly .
This is probably the case with tin , copper , and other metals in which the viscous effect is great .
[ I have not tested tin wires , but copper and also brass wires show a perceptible hardening effect under constant stress , much less , however , than that shown by soft iron .
] When , on the contrary , the hardening effect is great relatively to the viscous effect , a wire will show to greater advantage as regards breaking strength , and to less advantage as regards elongation the more slowly it is tested .
This is the case with soft iron , at least in circumstances like those of the above experiments .
The experiments show that even such variations in the rate of applying stress , or such intermissions of stress as are liable to occur in practical testing , are sufficient to affect most materially the results of the test .
A batch of specimens cut from the same coil , and similarly treated in every way , give very closely accordant results when tested with the same rate of increment of stress .
The examples quoted above are all taken from the same batch of specimens .
One peculiarity is common to them and to all the other specimens in the same batch , namely , the double curvature , which appears in the diagrams soon after the limit of elasticity has been passed .
On the other hand , other specimens cat from the same bundle of wire , but annealed at another time , do not show this peculiarity .
The examples given may fairly be taken as representative of a large number of tests .
I am now extending the experiments ; but it appeared to me that the results already arrived at were of sufficient interest to excuse their presentation in this somewhat crude form .
VIII .
" Memoir on Abel 's Theorem .
" By R. C. Row , Fellow of Trinity College , Cambridge .
Communicated by A. Cayley , LL. D. , F.R.S. , Sadlerian Professor of Pure Mathematics in the University of Cambridge .
Received May 27 , 1880 .
( Abstract .
) The object of this paper is to present in a shortened and simplified form the processes and the results of Abel 's famous memoir , " Sir un propriety generale d'une class tres-etendue de functions trans-cendaotes , " composed and offered to the French Institute in 1826 , but first published in the " Memoires des Savans Etrangers " for

