450 Prof. D. E. Hughes .
[ May 27 , Bond in 1858 considered that these areas varied as the time of exposure .
Further investigation is required .
III .
In the last section the more important question is answered , as to whether , on these modern photographic plates , where the times of exposure are reckoned by hours and minutes rather than by minutes or seconds , measures may he made of relative stellar coordinates as exact and trustworthy as those derived from the best astronomical instruments applied directly to the heavens .
For the purposes of this enquiry , resort is made to the same plates of the Pleiades which furnished the results in Section I of this research .
The distances of twenty-five stars from Alcyone were measured with the Oxford macro-micrometer : each measure on each of the four plates was repeated the same number of times as Bessel measured the same distances with his Konigsberg heliometer .
The result of the comparison is slightly in favour of the photographic measures ; the average deviations of the repeated measures from the mean being in the case of photographs 0-24 " and with Bessel 's heliometer measures 029 .
" A remarkable circumstance fortunately occurred in the course of the measures , in the case of one of the four plates ; unmistakable evidence appeared that the photographic film had very slightly but measurably shifted in the neighbourhood of three of the stars , but on no other portions of that plate nor on any portion of any of the other plates .
This indicates the unadvisability of relying on any single plate uncorroborated by others .
The applicability therefore of this form of photography to the most precise astronomical measurements seems to be established , even for plates exposed for considerable times .
I now propose to test the method still further , by applying it to the determination of stellar parallax .
VIII .
" Researches upon the Self-induction of an Electric Current .
" By Professor D. E. Hughes , F.R.S. Received May 24 , 1886 .
Numerous researches have been made upon the self-induction of coils of wire , and but few in relation to the influence exerted by the nature and geometrical sectional form of the electrical conductor when employed in straight wires as in those of a telegraph line with the earth as a return , or those of a single wide loop where the distance of the return wire is sufficient to prevent any appreciable effect from the mutual induction of separate portions of the wire upon 1886 .
] On the Self-induction of an Electric Current .
451 each other ; oar present theories class all non-magnetic metals together , taking only into account their specific resistance and the diameter of the wires ; they admit , however , a certain modification in magnetic metals due to their magnetic permeability , hut we have had but little experimental evidence of the effect produced .
The whole subject seemed to me worthy of experimental investigation , and I have lately* given the results of a first series of experiments made last year .
I have remarked since writing that paper many new and important effects , and I have made a new series of researches , the results of which I wish particularly to point out in this paper .
In my late researches I made use of a modified " Wheatstone 's bridge " together with a portion of my " induction balance , " by means of which the induced or extra currents from the wire under observation were balanced by a secondary current induced from an independent circuit ; this gave excellent results , but the method has been criticised as having the fault of not being clear in its indications .
In order to meet this objection and also verify the results which I had previously obtained , I constructed an entirely new bridge founded upon a most simple and well-known principle , and as the bridge admits of no change in the relative resistance of its sides , its action can be easily understood ; it is with this instrument that I have made my new series of experiments .
The instrument consists of an ordinary Wheatstone 's bridge , with the exception that a telephone replaces the galvanometer ; there are in addition two coils of insulated copper wire , one in each portion of two sides of the bridge , by means of which the mutual self-induction of its convolutions can be increased or decreased as desired .
The electrical contacts are made either by a continuous periodic or tuning fork contact maker , or by a peculiar clockwork rheotome which I have made , in which a contact spring rests lightly on a wheel whose roughened surface is divided into eight equal parts of contact and insulation , by means of which we receive on the telephone eight equal periods of sound and silence each revolution of the wheel .
We are by this means better enabled to appreciate feeble sounds than if they were continuous , and as the wheel can be made to revolve at any between two and ten revolutions per second , we have from sixteen to eighty periods of silence between each rubbing contact per second .
The following figure shows the theoretical plan of the electrical communications of the bridge:\#151 ; * " Self-induction of an Electric Current in relation to the Nature and Form of its Conductor , " " Journal of the Society of Telegraph-Engineers and Electricians , " vol. xt ( 1886 ) , p. 6 .
Prof. D. E. Hughes .
[ May 27 , Fig.1 .
A , B , C , D are the four sides of the bridge , the telephone J replacing the ordinary galvanometer ; the sides CB and CD are formed of German silver wire , each 50 cm .
in length , 0'5 mm. in diameter , and 0-85 ohm resistance ; the sides AB and AD have also 0-85 ohm resistance , so that the four sides have equal resistance , and this remains a constant during the whole series of experiments .
AE and AE is a continuous spiral of copper wire , formed of hard copper silk-covered wire of 1 mm. diameter and 4'80 metres in length , wound loosely on a boxwood cylinder of 3'50 cm .
diameter and 30 cm .
in length , on which it moves freely ; the entire helix has 40 turns of 4 cm .
diameter , each spiral being separated 5 cm .
from each other ; the spiral , however , is separated into two equal portions by giving a greater separation at the centre in order to allow it to be attached to a sliding collar of wood , by means of which we can press the spiral closer together on either side as desired ; at E and F there are also adjustable collars of wood , and as the boxwood cylinder is graduated , we may ( if the central collar is fixed ) approach more or less either side of the helix , and read the degree of approximation of the coil , thus adjusting the mutual induction of each side to a perfect balance or zero .
In practice I prefer moving the central collar , using the end collars only for the perfect adjustment of its zero , as we then have a double effect , viz. , closing the coil say from A to F , increasing its mutual self-induction , and at the same time decreasing the previously balanced induction of the coil AE ; this not only gives a wider range of effect , but renders the scale readings more uniform ; the end F of the helix is joined to about 10 cm .
of German silver wire , completing the circuit from G to D ; this supplementary German silver wire is simply for the purpose of making the resistance of AD equal to DO , and its length should be adjusted to 1886 .
] On the Self-induction of an Electric Current .
this purpose ; the end of the helix E joins directly with the terminal H ; the wire to he tested X is joined to N and H ; from H to I there is a second supplementary German silver wire allowing us by means of the contact side M , which is in direct communication with B , to introduce more or less of the German silver wire in the side AB .
The resistance of the wire to be tested should always be less than that of the opposite side of the bridge , and we then make up the total resistance of AB by sliding the contact slide M until the resistance of AB equals AD .
It will he evident that in the whole series of experiments there can he no change in the resistance of either side whenever zero is found , the induction coils or balance having the same proportional current in all the changes of the wires under observation , the battery and telephone circuit present also a continuous absolute relation ; we may , however , when desirable , keep the resistance from H to M a constant , and vary the resistance or the length of the German silver wire GD , we then balance the wire X by an equal resistance on the opposite side of the bridge , but then the battery and telephone circuits no longer possess the invariable relations which are so extremely necessary in experiments of the nature of which I have been making .
The battery circuit is joined in the usual manner to AC , the current can be interrupted by the rheotome K , or by means of a commutator ( not shown in the diagram ) close the battery circuit , and transfer the contact maker to the telephone circuit BD ; this allows us to observe the effect of an intermittent current compared with that of a constant or steady flow .
M. Gaugain has termed the pei'iod of a steady flow of current the stable period , and that in which a rise or fall of the current takes place the variable period ; these terms have since been generally adopted by telegraph electricians , and in order to avoid introducing new terms , they will be used in this paper .
It would require too much space to enter into the details of the construction of the bridge , special care is required in the construction of the balancing induction coils , and in securing perfect electrical contacts in all parts of the bridge ; the induction balance requires calibration , and for this purpose I introduce as the wire X successive lengths of 10 cm .
1 mm. diameter copper wire , thus forming a table of values throughout the range of the induction balance , or by equal increments of 10 cm .
of copper wire up to 20 metres .
Having no unit of self-induction to which my results could be quickly and practically referred , I have adopted as unit the self-induction of a straight copper wire 1 mm. diameter and 1 metre in length ; this gives on my calibrated induction scale 100 degrees , and it is to this standard that all the comparative force of extra currents mentioned in this paper is compared .
454 Prof. D. E. Hughes .
[ May 27 , The self-induction of a wire is proportional to its length , consequently a source of error might exist in the different lengths of the supplementary resistance wire HI introduced to balance the resistance of GD , but as we are enabled by the high specific resistance of German silver wire to obtain a very great change in resistance by a comparatively small movement of the sliding scale , this error in most cases of comparative experiments is but a fraction of 1 per cent. , and when taken into account as it should be the error no longer exists .
The telephone used should be of the most perfect kind , and adjusted expressly for rapid and feeble sounds .
I have found it best to employ an extremely soft Swedish iron diaphragm , without varnish or anything that can diminish or deaden the sound ; its fundamental note should be higher than those generally in use , or at least 500 double vibrations per second , for we have to deal with extremely rapid effects which on short wires cannot be rendered evident upon a galvanometer , but which with a telephone in perfect adjustment are heard most distinctly with an electromotive force in the battery circuit ranging from O'OOl to 0'250 ampere .
In the sketch of the communications the wire to be tested , X , is shown in the form of a wide loop , but in practice the instrument is constructed on two separate frames of wood articulated together at D , by means of which we can separate the terminals NH , and introduce straight wires , sheets , or tubes of lengths varying from 5 cm .
to 1 metre .
The object of my researches being to observe the self-induction which takes place in straight wires , or in those of a single wide loop where the reaction from any return wire is at such a distance that its influence is not appreciable , I shall , therefore , use the term self-induction to indicate the effects due to the electric current in its own portion of the wire , and mutual induction to those where the reactions of different portions of the current and circuit react on each other , as in the case of coils ; and although some theoreticians consider the two cases as the same , they are , as my experiments prove , entirely distinct , for we have , as will be shown , for copper wires a low coefficient of self-induction with a high mutual induction , whilst in iron wires the reverse is the case , for we there have a high coefficient of self-induction with an extremely feeble coefficient of mutual induction .
Influence of the Nature of the Conductor upon its Self-induction .
I found , as shown in my late paper , that there was a marked difference in the specific inductive capacity of iron and copper , and this entirely agrees with well-known theories ; but I suspected that there might be some difference in the non-magnetic metals independent of their specific resistance .
For this purpose I made a series 1886 .
] On the Self-induction of an Electric Current .
455 of experiments with wires of the same length and diameter , hut of different resistance .
These showed a marked difference provided the current was increased in proportion to its diameter or conductivity , but no difference could be found when the current was kept a constant , and the interior differences in resistance of the wires were compensated by an external added resistance ; but if we take wires of different metals\#151 ; all of the same length and resistance , but of different diameters\#151 ; there is a marked difference due to the mutual reactions of the current in its own wire , being less in wire of large diameter than in small wires .
These effects have been fully explained in my late paper , and have since been completely verified by the method used at present .
I showed by the use of my late method a critical point in the rise and fall of the induced currents ; this is not shown by the present method of compensating by external resistance , but the rapid decrease in the electromotive force of the induced currents , as indicated by the induction balance , is well shown by the present method .
The following table shows the observed force of the extra currents for wires of the same length ( 1 metre ) , but of different diameters:\#151 ; mm. mm. mm. mm. mm. mm. mm. mm. mm. mm. mm. mm. 0-25 0-50 1 -oo 2-00 3-00 4-00 5-00 6*00 7-00 8*00 9-00 10-00 Iron\#187 ; .
760 621 530 360 269 220 190 171 152 138 i 128 124 Copper 129 113 100 89 82 78 75 73 72 71 -5 i 71-2 71 The fall in force is now even more rapid than shown by my late method , and it will be seen that iron is peculiarly sensitive to an increase in diameter , having nearly six times the force of copper in wires of 0'25 mm. , and not twice the force of copper in wires of 10 mm. section .
Reactions of an Electric Current in its own Portion of Conductor .
The phenomenon of a constant decrease in the electromotive force of self-induction , as measured by the induction balance , with each increase of the sectional area of the conductor , is well shown by the present method , and an experimental investigation of its cause has shown that we should not consider a current in a wire as a single element reacting solely upon exterior wires , but that the current acts precisely as would an infinite number of independent streamlets of current reacting upon each other in the interior of its own wire similarly to their known effects upon exterior wires .
My experiments demonstrate this to a degree that leaves no doubt on my mind 456 Prof. D. E. Hughes .
[ May 27 , as to its truth , for according to this view we should be able to reduce the self-induction to a very great extent by employing thin flat sheets where the outlying portions are at a comparatively great distance from the central portions .
This I have experimentally proved to be the fact , but this experiment alone does not show if the reduction of self-induction is due to a different arrangement of the current in sheets as compared with solids .
If the reduction is due to the greater separation of the streamlets , then we should be able to reduce this induction in a still greater degree by employing a conductor composed of numerous small copper wires through which the current is equally divided , and which could he separated or brought close together as desired .
This proves to he also an experimental fact , for the conductor formed of numerous strands has far less self-induction than a thin sheet when its wires are separated , so that they can no longer react on each other , and surpasses the thin sheet and approaches the value of a circular wire when these wires are brought near together , so that their mutual reactions can approximate those of the numerous streamlets in a solid conductor .
Iron shows a still greater reduction in its self-induction when in the form of thin flat sheets or numerous small iron wires separated from each other , with the exception that we cannot restore or approach the value of a solid wire by bringing them in close proximity .
The reduction of the induced currents by employing thin sheets instead of a wire is so great in iron that we could not account for it on the mere separation of contiguous portions of the same current , but if we assume that the comparatively high force in iron wires is due to the induced circular magnetism , and that this almost disappears in flat sheets , we account for the fact that a thin flat sheet of iron has less inductive capacity than a similar sheet of copper of the same resistance , but of different widths ; and if thin strips of copper , brass , lead , and German silver are compared with similar iron strips , and their resistance rendered equal by the added resistance in the bridge , no difference is found between iron and the non-magnetic metals , for under these conditions their inductive capacities appear equal .
Reactions of an Electric Current between Separate Portions of the same Conductor .
In order to distinguish two distinct effects I have defined self-induction as the effect produced by an electric current on its own portion of the wire , and mutual induction as the effect of the reactions between separate portions of the wire on each other .
The mutual induction in iron and copper wires is very different in degree , as shown by the following table :\#151 ; On the Self-induction of an Electric Current .
Table I. Wires , silk-covered , 1 m. in length , 1 mm. in diameter .
Comparative force of the extra currents .
Strips or ribbons , silk-covered , 1 m. in length , 10 cm .
wide , 0*1 mm. thick .
Comparative force of the extra currents .
Copper wire , in a single close loop , where each side is in close proxi-mity ... ... ... .
18 Copper strip , in a single close loop , where each side is in close proximity 14 Copper wire , in a single wide circular loop . . .
100 Copper strip , in a single wide circular loop ... 60 Copper wire , in a coil of 3 cm .
diam. , having ten layers in close proxi-mity 607 Copper strip , in a coil of 3 cm .
diam. , having ten layers in close proximity 580 Iron wire ( soft ) , in a single close loop , as above 440 Iron strip ( soft ) , in a single close loop , as above 16 Iron wire , in a single circular wide loop ... 502 Iron strip , in a single circular wide loop . . .
60 Iron wire , in a coil as in the case of copper .
570 Iron strip , in a coil as - stated for copper ... .
578 There is a remarkable fall in the induction of a straight wire or single wide loop of a copper wire when this wire is doubled upon itself as a return wire in close proximity .
The effect of mutual induction is also shown by the equally remarkable rise when the mutual induction of ten layers of a coil react on each other .
This is well known , but what I believe has not yet been experimentally observed is the remarkably feeble mutual induction of iron wires , either when in parallel return in close proximity or when in coils of numerous layers .
The percentage of increase of induction in a copper wire from a wide single loop to that of the ten layers being 507 per cent. , whilst in iron under precisely the same conditions there was an increase of but 136 per cent. That this remarkable difference as shown between the mutual induction of iron and copper wires is due entirely to the circular magnetism in the iron wire , and that the mutual reactions of the streamlets of the current in a thin flat iron conductor prevent this formation I have proved in various ways , for if the circular magnetism is the cause , flat iron sheets which have as low self-induction as copper should be equally sensitive to mutual induction as the nonmagnetic metals , for they no longer in any appreciable degree possess the protecting magnetic sheath which enfeebles the mutual reactions of iron wires upon each other ; this proves to be the case , a thin " 158 Prof. D. E. Hughes .
[ May 27 , ribbon of iron having the same coefficient of mutual induction as copper , and it no longer behaves as a magnetic body .
Table I shows the difference in the mutual induction of iron wires and strips ; it will be seen that when the iron is in the form of a wire the circular magnetism produces a marked difference between iron and copper , whilst in the form of flat strips the iron resembles copper ( as regards mutual induction ) in every respect .
Influence of Circular Magnetism .
I have shown in my late paper that an iron conductor composed of numerous fine stranded wires ( as in a rope ) behaves like copper , and this I regard as entirely due to the breaking up or prevention of circular magnetism ; I also showed a phenomenon which I could not then explain , viz. , that when an iron wire was heated to a yellow-red heat it lost its previous high specific inductive capacity and behaved like copper .
The effect of heat upon magnetism is well known , consequently we can readily explain the fall in its inductive capacity by the disappearance of its circular magnetism .
I found also that in thin flat strips of iron there was no change whatever in its inductive capacity at the yellow-red heat ; this can now be readily explained : it behaved like copper when cold , and , having but little circular magnetism to destroy , there was no appreciable change produced , except that due to the extra resistance caused by the increased temperature of the strip , and to which copper and iron are almost equally sensitive , consequently we may consider iron ( when in the form of thin flat sheets ) to behave like non-magnetic metals throughout all temperatures .
Influence of Self-induction on the Resistance of the Conductor .
A phenomenon of great importance is the effect which I have observed of the resistance of a wire being greater during the rise of the current , as in the variable period , than that measured or known during the constant flow , as in the stable period ; by resistance I mean a pure ohmic resistance , a resistance which can only be measured , expressed , or balanced in ohms , and , whatever the cause , the effect is one of pure ohmic resistance .
We can imagine that at the first moment of contact there is no current flowing through the wire , its resistance is then infinite , but the current gradually increases in force until it arrives at its maximum , as in the stable period .
We have thus between the moment of contact and period of steady flow a variable period wherein the resistance falls in the form of a curve from infinity to its well-known stable resistance ; the telephone is unable to give the 1886 .
] On the Self-induction of an Electric Current .
\#166 ; exact form of this curve directly , but it gives by the nul method comparative results as to the different duration in time of the curve in different metals .
If we take a straight copper wire 1 metre in length and balance its resistance in the stable period , we shall find that there is only traces of a difference in its resistance in the variable period , and we can balance its self-induction by the induction balance ; but if we replace the copper by an iron wire of the same length , or even a much shorter wire , viz. , 20 cm .
, we find that in the variable period we can no longer balance the wire by the induction balance alone , but we must compensate for its increased resistance by the sliding scale , the amount of subtracted German silver wire expressing in fractions of .
ohms ( the value of the wire being already known ) the additional resistance of the iron wire in the variable period ; we can thus reduce or balance the iron wire to a perfect zero , provided the current from the battery does not exceed O'lO ampere ; but with greater current there will still remain a slight muffled sound , which cannot be reduced to zero either by the induction balance or the resistance slide .
It therefore became important to determine if this muffled sound was due to a difference in resistance which could not be balanced , or to the lengthened duration of the extra currents being slower than the balancing current from the induction balance ; the latter proved to be the case , for on prolonging the duration of the balancing currents from the induction bridge by introducing a core of iron in its coil , the balance or zero became absolutely perfect .
In this case we must choose between cores of different diameter to find one whose reaction on the time effect of the induction coil equals the retardation of the extra currents in the wire tested .
If we observe the method employed in the bridge , we shall see that the induction balance can balance the extra current of the wire X , irrespective of the position of the resistance slide M , or any relation between the sides AB and AD , but we cannot equalise the resistance of these sides except by the necessary adjustment of the resistance slide , consequently when we are forced to adjust the induction balance we are compensating the extra current , and when we are forced to move the resistance slide we are balancing resistance .
The disturbance in the bridge caused by the change of resistance of the wire tested in the variable period , causes a momentary primary current to pass through the telephone in the same direction as the extra current , and if these are not separated by balancing the extra current by an induction balance , the mixed effect would be read as a single effect of the extra current .
To show the importance of separating these two effects , it is only necessary to say that in most of the cases cited in this paper the momentary primary current due to the extra resistance greatly exceeds that of the extra current ; conse460 Prof. D. E. Hughes .
[ May 27 , quently all measurements taken wherein this separation is not complete , gives the result of a mixed effect .
In the method which I have described , the separation of the extra current from the momentary primary current is so complete that the measures given by the induction balance and resistance slide are invariably the same for a given wire ; there is no personal equation , for all observers find precisely the same values both for the resistance and for the extra current .
The method is , however , defective in measuring any small difference of resistance in the variable period in copper wires , as the induction balance itself introduces an additional but opposing resistance by the approaching of its coils , consequently any resistance which we observe is a fraction less than the real amount .
To prove this we have only to use a balancing induction current produced from the battery circuit , as shown in my first paper , when we observe this small difference , and better , observe any small difference of resistance in a straight copper wire , and greater when the wire is formed into a coil .
Soft iron wire gives a far higher resistance in the variable period than hard iron , but each wire , according to its molecular structure , has its own value ; the effect increases with the diameter , ranging from 25 per cent , increase of resistance for wires of 2 mm. in diameter , to 500 per cent , for those of 6 cm .
Table II.\#151 ; Resistance of Iron and Copper in the Stable and Variable Periods .
Comparative Resistance Resistance Percentage lnp'ppnQprl Wires , 1 m. in length , force of the in ohms , in ohms , vPQiQfnnpp in 5 mm. in diameter .
extra stable variable IC/ OlotCllivU X1J .
variable period .
current .
period .
period .
Copper 78 0-001284 0 -001372 7 Soft Swedish iron ... .
American compound 234 1 0-008346 0 -022200 166 wire , copper exterior , steel interior. .
V 83 0 -002247 0 -002696 20 Ditto , steel exterior , copper interior " \lt ; , . .
| 213 0 -007750 0-0248000 220 Table II shows a few illustrative examples : the resistance is given in the fractions of the ohm , indicated by the resistance slide M , upon the supplementary resistance wire HI .
The table shows that copper and the American compound wire coated with copper have an extremely rapid action or curve from an 1886 .
] On the Self-induction of an Electric Current .
infinite to its stable resistance , due to tlieir freedom from circular magnetism , whilst iron shows a comparatively slow curve .
A remarkable result will be seen where a copper wire has been coated with iron , its variable resistance being 220 per cent , above that of its stable period , and 54 per cent , greater than that of a solid iron wire .
I found in my previous researches that the resistance in the variable period could not be changed by a change in speed of a periodic or tuning-fork contact maker , and that the telephone was more sensitive when the rubbing contacts were used without any regard to their frequency of interval .
I also found that the telephone was far more sensitive and alone suitable for these experiments when its diaphragm was entirely free from varnish or coating , and sufficiently thick to give a clear musical dominant tone of about C or 512 double vibrations per second , the electromagnet being also as close as possible without being in actual contact .
I noticed that no matter what number of vibrations were sent , that the tones given by the telephone on each side of the zero of the bridge were invariably those of its dominant note .
Suspecting that the frequency of vibration of the diaphragm had a direct relation with the resistance found , I altered its note through a range of one octave by employing different thickness of diaphragms , and found that the results were in absolute proportion to the frequency of the vibrations of the diaphragm .
The highest tone giving a perfect zero with a high resistance in the variable period and the low tone giving its zero at a marked less resistance .
Thus it was evident that the telephone selected from the mixed vibrations sent by the transmitter those which alone corresponded to its period of vibration , and that we could thus observe the effect of more or less rapid periodic contacts .
The power which vibrating bodies have of selecting from a confused number of vibrations those in accordance with their own , is shown in Helmholtz 's resonators , and still more as regards electrical waves in the remarkable harmonic telegraph of Elisha Gray , who proved that if the armatures of several electromagnets are made to vibrate , each with a different musical tone or rate of vibration , then we may by the use of a vibrating transmitter of a fixed rate of speed select the electromagnet alone whose period of vibrations is similar , and that we may send several different periods of vibrations at the same time without any confusion in the receiving magnets , as each electromagnet selects from the compound vibrations those in accordance with its own .
This is exactly what occurs with a telephone when using a periodic contact maker , but I have found in addition , that a telephone arranged as I have described , will select and respond loudly in its own clear musical dominant tone to a confused mass of rapid contacts such as those produced by a scraping contact , and which in an ordinary telephone gives only the effect of noise , of not 462 Prof. D. E. Hughes .
[ May 27 , one-third the power or loudness of that produced by the tone of the musical telephone .
We may thus have ten or more telephones of different musical notes , each responding separately in their own tone or all at the same time , as each diaphragm selects its own period from the confused periods of contacts sent hy the scraping contact maker .
In order to verify the results , I have used in these experiments both periodic and mixed sound transmitters , but find that owing to the difficulty of keeping the periodic transmitter and telephone in perfect tune with each other , that the mixed sound transmitter is preferable , as it gives no trouble whatever , and we have more accurate results , as we then have only to keep our telephones in accord with the number of vibrations desired .
The experiments mentioned in this paper were all made with a fixed note of vibration of the diaphragm of 512 double vibrations per second .
I will , however , mention a few made with varying rates of vibration , and as I propose in a future paper to give a more extended series of observations through a wider range : I will confine myself to these examples .
Fig. 2 .
m \#166 ; SSS mm \#166 ; ms 18111111 mm. 1W mm mm mm mm i pWMM m R m \#163 ; JS9\#163 ; 1 \#166 ; 55s % 8SI1111 Si iiiiiii The diagram shows the results obtained on iron wires of different diameters ; their stable and variable resistance were all reduced to comparative values , the stable resistance being taken as 100 .
Three telephones were employed , the first having the musical note of G or 1886 .
] On the Self-induction of an Electric Current .
768 double vibrations , the second C sharp or 576 double vibrations , and the third , one octave below the first or G 384 double vibrations per second .
It will be seen that a great change in the variable resistance takes place in the range of a simple octave , and that the fall from the wires of 6 mm. to that of 1 mm. is so regular as to be almost a continuous curve , and the total curve would be probably that taken by the wire of 6 mm. in its approach to its stable resistance .
If this is the case , as future researches may show , then the resistance of this wire with the comparatively slow contact speed of six per second would still have 14 per cent , higher resistance in the variable than in the stable period , or an ordinary Morse telegraph instrument working at a slow speed of only eleven words per minute would experience this addir tional resistance ; consequently this effect should be easy to prove by the ordinary methods with galvanometers , and this can be accomplished as soon as we have found a method of completely separating the extra current from that due to the increased resistance , as I have already succeeded in doing by the method which I have described .
The iron wire of 6 mm. diameter shows for a speed of 384 contacts per second a comparative resistance in the variable period of 638 , or more than six times its stable resistance , but with 192 contacts per second its resistance is but 371 ; the fall of resistance is so rapid here that for a single octave difference in the telephones , the fall is far greater than the whole stable resistance .
The extra current , as well known , is proportional to the length of contact for fine wires , but in large wires the curve indicates that the extra currents have a local reaction on the cessation of the primary current .
Influence of a Magnetic Sheath .
The remarkable phenomenon wherein a non-magnetic metal such as copper shows a higher " variable " resistance when under the influence of an external coating of iron than a wire of solid iron ( and this notwithstanding that the resistance of the side of the bridge is kept a constant or by balancing its resistance on the opposite side GD as mentioned ) seemed worthy of experimental research .
Experiments were made in order to find what form of conductor would give the maximum increased resistance in the variable period , and I found that an external tube of iron insulated from its central core produced this effect , for if the tube was not insulated but joined in electrical contact the results were reduced , due I believe to a transversal neutralisation taking place in its own portion of the wire , as I have mentioned in my reply to the discussion on my first paper .
Table III shows the effect produced upon an electric conductor by surrounding it with an insulated sheath of iron .
It will be seen that 464 Prof. D. E. Hughes .
[ May 27 , Table III .
Influence of an Iron Tube upon an Interior Insulated Wire .
Iron gas tube , 90 cm .
in length , 10 cm .
diameter , 2 mm. thick .
Wires 1 metre in length , 3 mm. diameter .
Comparative force of the extra currents .
Resistance in ohms , stable period .
Resistance in ohms , variable period .
Per centage of increased resistance in variable period .
Absolute increase of resistance in ohms .
Copper wire alone Copper wire insu- 82 0 '00460 0 -00482 7 0-00022 lated in the in- i- 410 0 -00460 0-03220 terior of the iron 600 0-02760 gas tube J Brass wire , ditto. .
410 0-01380 0-03974 188 0 -02594 Iron wire , ditto . .
615 0-02944 0-05888 100 0-02944 Lead wire , ditto. .
410 0 -05750 0-08682 51 0 -02932 Grerman silver | 410 0 07636 wire , ditto 0-10384 36 0-02748 the magnetic reaction of the tube upon the primary current passing in the wire produces a marked effect upon the different metals , the force of the extra currents as measured by the induction balance rise in value from 82 to 410 , this being a constant for all metals except iron , which alone shows the higher force of 615 ; the force of the extra currents , however , has no direct relation to the extra resistance shown in the variable period , but the latter must be in direct relation with the duration or length of time required to pass from an infinite to the stable resistance .
The table shows on comparing the resistance of a metal in its stable period with that found in the variable , that copper has the highest percentage of increased resistance , being 600 per cent , increase , or seven times that found for its stable resistance , each metal in the order of their specific resistance having a less percentage of increased resistance until we arrive at the German silver wire , which shows only 36 per cent , against the 600 per cent , of the copper ; we have also here the remarkable fact that iron even with a circular section stands simply in its order of specific resistance , showing only one-sixth of the increased resistance experienced by copper ; the percentage of increased resistance of a metal when under the influence of an insulated magnetic sheath is directly as its conductivity or inversely as its specific resistance .
If we regard the phenomenon from a different point of view , confining ourselves simply to the additional resistance above that of its specific stable resistance , all metals whether magnetic or non- 1886.J On the Self-induction of an Electric Current .
465 magnetic have an almost invariable quantity dependent entirely upon the coefficient of the electromagnetic action of the iron tube , and independent of the resistance or nature of the wire .
The experiments with an insulated iron sheath may be considered as a forced condition which does not occur or need not be taken into account , but the reaction of the iron tube is evidently electromagnetic , and as our atmosphere is also magnetic , we may assume that its reaction would be similar though in a far less degree to that which I have shown .
Submarine Cables .
The influence of a magnetic sheath upon an external insulated wire is one of vast practical importance in relation to our electrical submarine cables , as they are all constructed writh an insulated wire surrounded by a spiral sheath of exterior iron wires .
In order to study this question I made several cables of short lengths ( 1 metre ) and found ( most fortunately for our practical applications ) that when the exterior iron sheath is divided as in a cable , or when the sheath consists of several large wires in close contact , that its effects are reduced to a mere fraction of what it would be if the external sheath was a continuous tube ; this is entirely due ( as I have shown in the case of stranded wires as compared with solid ) to the imperfect formation of circular magnetism .
I will cite a few examples illustrating this : a cable was formed of a similar copper wire and insulation to that shown in Table III , the iron tube being replaced by eight iron wires ( of 2 mm. diameter each ) wound with a slight spiral of five turns per metre , this showed but 100 per cent , increase in the force of the extra currents against 400 per cent , increase when the same wire was in the iron tube ; the increased resistance in the variable period was but 50 per cent , against 600 per cent , for the iron tube , or only one-twelfth of the additional resistance caused by the tube ; these wires were replaced by galvanised iron wire which prevented magnetic contact , and the variable resistance became much less , and when these were replaced by numerous fine iron wires the effect was reduced to the minimum found , or only 20 per cent , increased resistance in the variable period , or 30 times less than that of a solid iron tube of much less iron .
It is therefore evident that the circular magnetism plays an important role , and that most fortunately our cable manufacturers ( without knowing the immense reaction which would be produced by a continuous sheath of iron ) have constructed our cables with a protection of iron divided into several iron wires instead of a continuous iron sheath ; there are , however , many telegraph lines which in passing through tunnels use a continuous iron tube as the VOL. XL .
2 i 460 Prof. D. E. Hughes .
[ May 27 , protection , and therefore snch lines must feel all the deleterious effects that I have shown to be caused by the reaction of circular magnetism .
Influence of Copper and Iron Cores upon the Induction and Resistance of Coils .
It is well known that a coil of wire has a higher self-induction than the same wire in a single loop , and that the coil has a still higher induction when we introduce an iron core .
I have made , however , a series of experiments in order to measure the influence of a core upon the resistance in the variable period , and also note the influence exerted by the induced or eddy currents circulating in the core : the following table shows some comparative results .
Table IY .
Helix formed of insulated copper wire , 1*50 metre in length , 2 cm .
diameter , and 24 turns .
Comparative force of the extra currents .
Resistance in ohms , stable period .
Resistance in ohms , variable period .
Percentage of increased resistance in variable period .
Absolute increase of resistance in ohms .
Helix alone 460 0 *02632 0 -02896 10 0 -00264 Same with a core of solid copper. .
352 0*02632 0 -04013 52 0-01381 Same with a core of insulated copper wire 460 0 -02632 0-02896 10 0-00264 Same with a core of solid soft iron 2338 0 02632 0-09870 275 0 -07238 Same with a core of 445 separat e fine iron wires , each 0*25 mm. diameter 5360 0*02632 0-04448 69 0-01816 Same with a core of silk-covered fine insulated iron wires 5820 0 *02632 0-04075 55 0 01443 A helix was formed of an insulated silk-covered copper wire of 24 single layers , having an interior diameter of 2 cm .
, the object being to form a coil having as little mutual induction as possible , but which would be readily acted upon by any core of metal introduced ; and in order to measure the high forces obtained an induction balance of great power and range was used , the method , however , being the same as that already described .
The helix alone showed an induction value for its extra currents of 1886 .
] On the Self-induction of an Electric Current .
460 , and a feeble increased resistance in the variable period of but 10 per cent. ; if we now introduce a core of solid copper of 1'75 cm .
diameter and 5 cm .
long , a great change takes place both in the force of the extra currents and the resistance in the variable period , the extra currents fall in value from 460 to 352 , whilst the extra resistance is increased to 52 per cent , above that of the stable period ; we have here a double and contrary effect , the reduction of the force of the extra currents and the increase in the resistance cannot be due to the magnetic nature of copper , but must be due to the induced or so-called " Foucault currents " circulating in the core ; to prove this the core was replaced by another similar in every respect but cut longitudinally to its centre , the currents now ceased to circulate , and the copper core had not the slightest effect either on the extra currents or the variable resistance ; this is shown in the table , where for greater precautions a core of insulated copper wires replaced the solid core .
Evidently the induced current in the core was the cause of the extra resistance ; work was done by the primary current , and a loss of energy at the expense of the electromotive force of the extra currents , but in doing this work a resistance was produced which was no doubt caused by the currents circulating in the core .
These currents required time , passing through the variable stage and thus producing from their electromagnetic inertia an equivalent reaction and electromagnetic inertia in the primary coil itself .
That this inertia'is due to the electromagnetic character of the current and not to an electric current considered apart , is proved by the fact ( which I have verified ) that when we coil a wire into a coil of several superposed layers , its electromagnetic reactions introduce a measurable resistance in the variable period precisely similar though feebler than that which would be produced by the reaction on the conductor of a magnetic body such as iron .
The effect of a solid iron core and a bundle of iron wires on the increase of the force of the extra currents is well known , but the table shows an interesting result as to their effects on the variable resistance ; the solid iron core shows a very high force of extra currents produced by its magnetic reaction on the wires of the helix , the resistance in the variable period has increased 275 per cent. , and its extra currents are extremely high , as we should expect ; now if the extra resistance is due to magnetic reaction alone it should increase when we are enabled to increase this reaction , but if it is due in greater part to the electromagnetic inertia of electric currents circulating in the bar , then by preventing these currents from being formed ( as we did in the case of copper ) we should greatly reduce the extia resistance .
This proved to be the case , for on replacing the solid iron core by a bundle of fine iron wires the force of the extra currents rose from 2338 to 5360 , or more than double the force of that produced by 2 i 2 468 On the Self-induction of an Electric Current .
[ May 27 , a solid bar , whilst the resistance in the variable period fell from 275 to bnt 69 per cent. It is well known that a bundle of fine wires magnetises quicker than a solid bar , and this may have had its effect , although not more important than the suppression of the eddy currents , as the latter may be the cause of the former .
The table shows that where we have suppressed these currents as far as possible by introducing a core of insulated wires containing less iron than in the previous experiment but fully insured against eddy currents , the induction was the highest and the resistance the lowest yet found for a helix containing an iron core ; the experiment proves to my mind that the extra resistance found in the three last experiments is due both to the electromagnetic inertia of the eddy currents , and the inertia of the magnetic molecules of iron .
The experiments related in this paper have been most carefully made and verified , and from the ease and certain action as well as the invariable results obtained by the method which I have described , they should be easy to repeat by others .
If we regard the whole of these researches we cannot fail to notice certain important laws which act in the greater portion of them .
I have shown : 1st .
That the contiguous portions of the same current react upon each other in the interior of its own portion of the con-ductor similar to their known exterior reactions on separate portions of the same conductor .
2nd .
That the coefficient of mutual induction is less in iron than copper wires , but that their coefficient is the same when the conductor is in the form of a ribbon .
3rd .
That the inductive capacity of different metals depends on their specific resistance , on their electromagnetic capacity for circular magnetism , and on the geometrical form of their conductors .
4th .
That the inductive capacity of a conductor of magnetic metal is dependent upon the formation of circular magnetism and not upon its internal magnetic permeability .
5th .
That a magnetic metal can be rendered equally free from circular magnetism as the non-magnetic metals .
6th .
That we have experimental evidence of electromagnetic inertia and the deleterious effects of eddy currents in the cores of electromagnets .
In addition to the above effects , we have the discovery of a large increase in the ohmic resistance during the variable period , allowing us to demonstrate and measure the gradual rise of an electric current in its conductor .
In conclusion , I desire to express my warmest thanks to Lord Rayleigh , Mr. F. L. Pope , Professor Forbes , Dr. Hopkinson , Mr. W. B. Preece , Dr. Fleming , Mr. Fitzgerald , Professor Silvanus Thompson , and Professor Ayrton , for the important theoretical contributions they gave on the discussion of my first paper .
Mr. W. H. Preece , Electrician of the Post Office , not only gave information of great practical On Intestinal Rest and Movement .
1886 .
] value , but has kindly supplied me with the wires used in these experiments .
The discussion proved the necessity of the researches which I have undertaken , and the importance of an experimental determination of the self-induction of an electric current in relation to the nature and form of its conductor .
IX .
" Contribution to the Study of Intestinal Rest and Movement .
" By J. Theodore Cash , M.D. Communicated by T. Lauder Brunton , M.D. , F.R.S. Received May 25 , 1886 .
( Abstract .
) Experiments were made upon a dog in which a Weller 's fistula had been established .
The length of the isolated intestine ( upper part of the jejunum ) was 18'5 cm .
By merely observing the mouths of the fistula , it was ascertained that during a condition of hunger , periods of complete quiescence , varying from two to twelve minutes , occur .
At any time , however , contractions having a very regular rhythm , might make themselves manifest .
On food being presented to the animal , immediately after swallowing and for some time after ingestion the movements became much more active and persistent .
Some hours ( four to five ) after a full meal they fell to their minimum .
For the closer study of the speed of transit of a solid or semi-solid body through the fistula , travelling sounds connected with a registering .
apparatus were employed .
Compressible but fixed sounds for the study of local contraction were also used .
The most important results obtained were the following:\#151 ; 1 .
That the act of swallowing , whether empty , i.e. , produced by external friction of the larynx , or of liquids , is frequently succeeded by contraction , sometimes by distinct peristalsis of the intestine .
Inhalation of sulphuric ether for a few seconds\#151 ; which causes abundant salivation and deglutition\#151 ; is almost invariably followed by active intestinal contraction .
2 .
Mental impressions , tickling the walls of the abdomen or the application of cold to them , are amongst the causes of contraction of the small intestine .
3 .
An unfailing means of producing well-marked peristaltic contraction is the administration of food to the fasting animal .
Not only does this cause the individual contractions to ensue with greater regularity and force than before , but they have less of the pendulum character ( contraction succeeded by complete relaxation ) , and during the pauses the travelling sound is so effectually " gripped " that it

