]\gt ; PHILOSOPHICAL TRANSACTIONS .
\mdash ; I. On the Effect of on the Speciftc Inductive Capacity of a Dielectric .
W. CASSIE , M.A. , College , CaJnbridge , in tlw ersities of Aberdeen } } .
by J. J. THOMSON , M.A. , F.R.S. , endish Professor of Physics in the Unirersity of Received May 24 , \mdash ; Read June 20 , 1889 .
THE object of the experiments describe in this paper was to learn how the specific inductive capacity of a ic is affected by change of temperature .
CAVENDISH observed an increase in the capacity of a glass condenser hen it was heated , but gave no measure of the effect .
Dr. HOPKINSON observed in light flint glass an increase of per cent. in the capacity between and C. And Messrs. GIBSON and BARCLAY showed there is no eciable change in the ) of paraffin between temperatures \mdash ; 1 and C. Except these , no measurements of the efiect appear to have been hitherto published .
present investigation shows an increase of specific inductive capacity with rise of temperature in all the solids* examine , and a decrease in all the liquids except one .
As paraffin , which is a substance comparatively near its melting point at ordinary temperatures , shows no change , these results seem to indicate , as far as they go , that the specific inductive capacity of a substance has a maximum value about the meltinoi point .
But it may be questioned whether the are sufficienlt as yet to 11 arrant so general an induction .
The relation which CLERK MAXWELL'S electromagnetic tlJeory of light indicates between specific inductive capacity refractive index makes it interesting to compare the effects of temperature on these two quantities .
Four of the * The results given here for mica , ebonite , and the first specimen of arc all less fhall those mentioned by Professor J. J. THOIISON , in his treatise on " " Applications of Dynamics to sics and Chemistry\ldquo ; ; because the results quoted there wore obtained from an earlier set of obscrvations , which the precautions for ins ulation were inferior to thosc describccl ON OF RATURE ON .
which DALE and GLADSTONE have observed the ctiv indices ( 1 ) ternperatures .
For one of these four the temperature efTects no simil rity 1tever , for another the relation is fairly close to that by the 'd other not agreeing exactly with the , the elati oloes not differ it very greatly , more , ) , than explained ) di ces in the specimens used .
Tho ents done in lihe Cavendisb idge , and of nking ) Professor J. J. , F.R.S. , , at whose estio they cing the resources of the ttory at my for advice the course of the work .
I. SOLIDS .
in of solids to with the to be nented 1 observe its ) different temperatures .
The condenser consisted of pile of alternate thin sheets of the dielectric and discs of lead foil , with of iron above and below , and a weight on the top to keep them .
The condenser could llot be fixed together in any more permanent way , because the unequal nsion with heat of the different materials lvou]d have altered the compression of the pile , and so produced a change of capacity greater the effect under investigation .
Special were required to secure that the tion between the sides of the condenser be as far as independent of the temperature .
To secure this the conden ser was suspended in a stirrup from a cket about four feet above the air bath in which it heated .
The suspending wires , which formed the connecti o. the side of condenser in cont , act with the stirl.up , passed without holes in the top of the air bath , and over a short glass tube varnished with sbellac , which was fixed across the bracket above .
The connection for the other side of the condenser consisted of a stiff wire , which also passed without touching through a hole in the top of the air bath , and was cari'ied through the glass on the ) racket .
Thus the condenser had no contact with anything about the bath , supported solely by the glass tube on the bracket , which was always cold and perfectly insulating .
The perfect freedom of the condenser and its easily tested before each observation by tapping the air bath and seeing that no motion was produced in the suspending wires .
The condenser was heated by ising the bath quickly to a temperature considerably that imed at , then ving under it a small me adjusted by experience to .
the right .
The temperature was read by a thermometer passing through a cork in the top of the bath .
About three hours was the shortest time required to ensure that the condenser was heated throughout ; it was usually fFIC INDUCTIVF , CITY O ) The condensers needed before settlino b ' to it city in the cold state .
The tures in these preliminary ] bove the highest used in the actual observations .
The capacities were measured by the method used by Professor J. J. }IsoN in his determination of ratio of the electrostatic and gnetic units ( ' Phil. Trans 1883 ) .
It is thus described in aper : " " In a Wheatstone 's bridge , A , with the galvanometel at , and the battery between A and , the circuit is not closed , but the points and are connected with the two poles , and , of a commutator , between which a trayelling piece , moves backwards and forwards ; is connected witI ] ) ) of condenser , the other plate of which is connected with D. Thus when is in ) tact with the condenser will be , and until it is charged , electricity will flow into it from the battery ; this will produce a momentary the various of the bridge .
the moving piece is contact with two plates of the condenser are connected , and the condenser will discharge itself through , and as the resistance of is infinitesimal in comparison with the resistance of Fig. 1 .
any other circuit , the discharge of the condenser will not send an ] moun t ectricity through the galvanometer .
Thus , if we make the moving piece oscillate quickly from to , there will , owing to the flow of electricity to the condenser , be succession of momentary currents through the galvanometer .
The resistances are so adjusted that the deflection of the galvanometer produced by these momentary currents is balanced by bhe deflection due to the steady cul.rent through the galvanometer , and the resultant deflection is zero .
When this is the case there is a relation between the capacity of the condenser , the number of times the condenser is charged and discharged per second , and the nces in the various of the bridge This relation , is worked out in the , is expressed by taking to represent the istances of A , A , A respectively .
The resistance MR. W. CASSIE ON THE ]FFECT OF TEMPERATURE ON of and so small as to be negligible .
Then if the condenser has a capacity , and is and discharged n times per cond The resistances used were about between 3000 itnd 6000 , , ohms .
With these values is correct to within2 per cen and is the formula used here .
The commutator P ( fig. 2 ) was that used by Professor THOMSON and is thus described in his paper:\mdash ; " " The current from two cells passes .
through a tuning-fork interruptor , and then through the coils of an electromagnet .
is a strip of brass with a piece of iron attached to it .
When there is no current passing the electromagnet , the elasticity of the rod makes it press against a screw , which is electrically connected with a binding screw : when the current passes through the electromagnet , the magnet attracts the iron attached to the rod and brings it into contact with stop , which is electrically connected with the binding screw S. The letters indicate the same points in this figure as in fig. 1 .
All the ] aces where contact is made by the ating piece are covered with platinum , and the TH SPECIFIC CAPACTTY A DI whole aArang.ement is fastened down to an ebon ite board .
As the current passes through the coils of , the piece strikes alternately yainst the parts and ; when it strikes the opposite plates of the condenser are connected with the two poles of the battery ; when it strikes against the condenser is discharged To be able to allow for conduction or absorption in the diclectric by means of observations with forks of different rates , we must know the ative times during which the vibrating piece is in contact with , in contact with , and travelling from one to the othel .
I therefore fixed stiff arm to the end of and took with it on stnoked paper , lnoved across while interruptor was .
These tracings showed that when the interl.uptor is going steadily ( this is easily known by the sound ) is in contact with and for times .
and the time occupied in Qsing irom one to the other is to the time of contact .
The battery used six LECLANCH cells , resistances taken from a box by ELLIOTT .
TTuning forks were used , , C4 , and 49 complete vibrations per second .
The forks were worked by a current the storage cells in the The coefficients of expansion with temperature of lead , ebonite , and glass are 000029 , 000077 , and ctively .
That of mica does not seem to have been determined ; but it is probably less than that of lead .
So that the combined effects of expansion with rise of temperature do not affect the capacity before the fifth decimal place ; and , as the results do not profess to be correct to more than four decimal , the expansion of the materials may be neglected .
Some observations were also taken with ebonite by an electrostatic method based upon that described in MAXWELL'S " " Electricity netis m\ldquo ; S229 ( second edition ) ; the only difference being that there was no guard .
The condenser was with a variable condenser by dividing a charge between them , separating them , and connecting opposite poles together to the electrometer ; then , if they were unequal , the difference deflected the electrometer .
For this method a special key was required to make the connections quickly and to keep the electrometer to earth until everything except the charge to be measured been discharged .
If the electrometer were not kept to earth it would be deflected in spite of ordinary screens by induction inside the key on account of the potentials required .
When neither of the condensers being any absorption , there is no difficulty with this method , and any ordinary quadrant electrometer may be used .
But when this is not the case the residual discharge will to come out as soon as the condensers are connected , and this makes the needle lways go ultimately to the same side .
The only way to get a balance is bo make the variable condenser too large , and gradually diminish it until the initial motion of the needle due bo it disappears .
This requires a needle with a short of , because with a slow R. W. CASS IOF ON needle the ) overpowers the first effect before the needle to it .
With a small ebonite it as just possible to make an electrometer sufficient by making the connection as short as possible ; but with a large condenser it was useless .
Consequently , this experiment would require an electrometer with a needle of very short time of swing .
One with small cylindrical needle quadrants seems most suitable .
The condenser was made of twenty-two sheets of brown Muscoyite with markings .
The sheets were about inches diameter , and the thicknesses varied from about a quarter to about a tenth of a millimetre .
Some difficulty found in getting rid of Stll.face conduction on the mica .
No cleaning would make it insulate , but in the end perfect insulation was secured by putting a border of shellac val.nish round the edges of the sheets .
The shellac did not come between the plates of the condenser , and so did not affect the capacity .
The highest temperatures in the preliminary were less than 14 C. The insulation was constantly tested by an electrometer , and was perfect throughout for temperatures below 1 C. Aboye this it began to give way .
The results are given in the following table and diagram ( fig. 3 ) .
THE SPECIFIC CAPACITY OF A The fact that both periods give the same value shows that , to be expected from its crystalline structure , there is no absorption in for those short times of charging .
Ebol ?
ite .
The ebonite condenser was made of venty discs of nite , about third of a [ nillimetre thick , and three and a quarter inches in diameter .
The ebonite sheets were carefuIly cleaned with paraffin dissolved in benzol , and this gave perfect insulation without any shellac border .
At temperatures above C. the capacity began to increase rapidly owing to softening .
The highest temperature reached in the preliminary heating was a little under C. The insulation was constantly tested as in the case of mica , and found perfect up to about C. The results are given in the following table and diagram ( hg .
4 ) .
As the insulation was perfect , the greater capacity shown by the longel time of charging must be due to absorption\mdash ; an effect which the structure of would lead us to expect .
The colunln of values for ntaneons charge have been calculated on the assumption that the absorption goes on at the same from the beginning of the charge , as during the interval observed . .
W. CASSIE ON ] ON of of change corrected for instantaneous charge .
4 lo C. 00037 .\mdash ; In the obseryations on ebonite by the electrostatic method the condenser consisted of two sheets of ebonite between the iron discs , and with copper between them .
The condenser rested upon a glass tripod inside the the leading entered the air bath through ] tubes .
The city of this adjusted so as to be within the of a fine sliding condenser in the atory against which it balanced ; and the readings were taken on an Elliott electronleter .
It required care and patience to make so slow an electrometer suflice , for the reasons already stated .
The insulation of all parts of the also required much ttention .
The cold temperature , obtained by .
iced round the bath contain ing the condenser .
Subsequent observations showed that this ffected the tion and so increased the apparent diIninution of capacity .
The greater rate of shown by the electrostatic results is probably due to the time of ging by the key worked by hand greater with the tuning fork .
Any leak would produce an apparent decrease of capacity by this method ; so that , although the electrostatic results are of less weight than the others , the reement of the results of two methods in which main of error ten ds in opposite directions is a ) confirmation of their accuracy .
The ostatic results are given in the liate of change of Temperature .
capacity per C. C. C. C. C. C. C. C. C. C. C. C. C. C. C. C. C. 0008 0010 0016 \mdash ; \mdash ; THE SPECIFIC ) OF A glass condenser was made with seventeen she of thin -slide -glass , about three inches diametel .
They consisted of a soda glass of high conductivity ; so that , although the discs were carefully cleaned and bordered with shellac as in the case of mica , the insulation of the condenser , as tested by the electrometer , was never perfect .
Since it is only the of the condenser that affects the galvanometer , the discharge passing sensibly ] through ( fig. 1 ) , the defective insulation introduces simply a steady current during each time of arging .
Each time the condenser is charged , the quantity of electricity passing and consists of ( 1 ) the charge of the condenser , and ( 2 ) the current through the condenser , which sts during the time the vibrator is in contact with S. For our present , the current m.ay be considered as immediately established at full strength when the circuit is closed .
So that , neglecting absorption , the apparent capacity exceeds the true capacity by the conductivity of the condenser multiplied by the duration of the contact between and S. And from observations on the apparent capacity with two forks of known speed , the true capacity can easily be found .
The highest temperature the preliminary heating was less than 11 C. The results are given in the following table and ( fig. ) C. 4795 4320 C. 4795 4320 C. 4795 4320 C. 4795 4320 C. 4795 4320 C. 4795 4320 C. 4795 4320 C. 4795 4320 C. 4795 4320 C. 4795 4320 C. 4795 4320 C. 4795 4320 C. 4795 4320 C. 4795 4320 C. 4795 4320 C. 4795 4320 034 0016 053 0017 077 Fork making 64 complete vibrations per secolld .
CCCXC . .
W. ON EFFFCT OF fURE ON of change corrected for instantaneous charge . .
Fig. 5 .
next three tables and diagram give the esults of exper.iments upon another condenser of microscope-slide cover-glass .
This specimen of the glass had less conductivity than the previous , but the precautions for general insulation were not equal to those in the other experiments described .
The condenser rested on a glass tripod inside the air bath , and the leading wires passed into the bath through glass tubes .
The condenser consisted of twenty-two discs of glass cleaned and bordered with shellac in the usual way , with discs oflead foil between them .
The curves almost exactly parallel up to between and , and diverge at higher temperatures in consequence of conduction ; so that we may take the temperature change of specific inductive capacity for this specimen of glass to be about per cent. up to An attempt was made to observe this temperature effect for shellac .
The condenser was made by dipping the lead discs in shellac varnish , and carefully and thoroughly evaporating the alcohol .
But after the condeIlser was made it was found that at a temperature below the shellac softened , so that the plates were pressed together by the weights on the top of the condenser .
To diminish the weight would have been useless ; because , in any case , it would have been impossible to say how far a change of capacity was dne to softening , and without a weight all the results are quite unreliable .
CAPA C OF A C. emperature .
apacity aRate ocper d With 99 C. C. C. C. C. C. C. C. C. C. C. C. C. C. C. C. 4010 C. 4010 C. 4010 C. 4010 C. 4010 C. 4010 C. 4010 C. 4010 C. 4010 C. 4010 C. 4010 C. 4010 C. 4010 C. 4010 C. 4010 C. 4010 C. 4010 C. 4010 C. 4010 C. 4010 00230 00369 00251 With ' ' 64 fork " " ( between 64 and 6.5 ) .
With 49 fork .
Fig. 6 .
II .
paratus for liquids consisted essentia]]y of a quadrant electrometer immersed .
in the liquid .
The heating done by a water bath ; and on account of the highly inflammable character of some of the liquids experimented upon , the water was heated in a separate vessel at a ) distance , from the top and bottom of which pipes went to and in fig. 7 .
The hot water was pumped through in the direction Fig. 7 .
The provision for the insulation was on a similar principle to that adopted in the case of solids .
Each quadrant was supported by a long Qtiff rod , the upper part of which was fixed in a varnished glass tube .
These four glass tubes were separately SPECIFIC INDUCTIVI CAPACTTY OF A ) I.3 clamped to a support about 18 inches above surface of the liquid ; so that the quadrants touched except the liquid and this perfectly insulating support .
The needle was suspended by a fine wire from the same height so as to oscillate under torsion .
The hollow top of the water batb had seven holes through it , as shown in the figure .
Through the outside four passed , without touching , the rods supporting the quadrants , through the centre one hung the needle , and through the two next the centre thermometers could be inserted .
This top was permanently fixed to the support of the needle and quadrants ; and the lower part of the bath containing the liquid was oved up on a smooth sliding arrangement into contact with the top so as to immerse the needle and quadrants .
Above the fixed top of the bath was a box with a window on one side , through which the movements of the needle and mirror were read by a scale and telescope .
The deflections of the needle were observed when first to one pair of quadrants and then to the pair .
As the needle and quadrants were parts of cylinders between three and four inches diameter and about a quarter of an inch apart , a large electromotive force was required to produce a deflection .
The electromotive force had also to be rapidly reversed in direction to avoid as far as possible polarisation , convection , \amp ; c. , in the liquid , The electromotive force was obtained from a Ruhmkorff coil without a condenser , and with a high resistance between the terminals to prevent sparking .
This high resistance consisted of a wide glass tube , about 6 inches lon , filled with distilled water , and having a thick copper wire through a cork at each end .
By altering the distance of the ends of the copper wires in the water the could be adjusted and the deflection controlled as desired .
The coil was worked by a current from the storage cells in the Laboratory .
As the electromotive force given by this arrangement was variable , and also the loss of electromotive force by conduction was different for each liquid and for each temperature , it was necessary to be independent of such changes .
dingly a second electrometer was placed between the terminals just outside the liquid one , which being always in the same state gave the comparative values of the electromotive force .
This second electrometer was an Elliott 's lecture-room pattel.n with two quadrants removed and the needle connected to one of the remaining quadrants .
The advantage of this form for the present purpose was that by moving the needle up from the quadrants the deflection could be diminished to any desired extent .
Then the quotient of the deflection of the liquid electrometer by the deflection of the second electrometer was proportional to the specific inductive capacity of the liquid .
For every deflection of the liquid electrometer a pair of deflections of the second electrometer were taken with the needle connected first to one pair of then to the other .
The genel.al arrangement of the apparatus is shown in the figure .
The liquids experimented upon were turpentine , carboll , glycerine , benzene , benzylene , olive oil , and paraffin oil .
Methylated spirit , also tried ; but cond so great that deflection could be obtain ed. All except oil showed decrease of specific inductive capacity with rise of temperature .
The oil was that used in the in the Laboratory , and its exceptional behaviour may have been due to some secondary action arising from impurity .
The results are given in the following tables and diagram .
And to show the way in which the observations were made , the readings are given for turpentine at two eratures .
to that at the , SPECTFIC CAPACITY OF A ] Temperatul inductive capacityRatio opecificto t Rate of lowest temperaParaffin oil : ture EADINGS for Turpentine .
ElectrometelLiquid .
Ratio ometer r 16.5 6.2 6.2 16.5 6.2 6.2 Needle cnnected t\ldquo ; , othe .
' 16.5 6.2 6.2 Needle cnnected t\ldquo ; , othe .
' 16.5 6.2 6.2 Needle cnnected t\ldquo ; , othe .
' 16.5 6.2 6.2 Needle cnnected t\ldquo ; , othe .
' Needle cnnected t\ldquo ; , othe .
098 Mean 100 098 098 Mettn .
W. ON UltL ON If the relation indicated by MAXWELL'S ] omagrnetic theory of light held good , the specific inductive capacity were equal to the square of the ctive index , then the rate of change of specific inductive capacity with tenll ) ) ought to be twice of the efractive index .
Amongst the liquids , for which DALE and have observed the refracbive indices at erent t four of those dealt with here .
The ates of change , with temperature of the efractive index for the A in the spect1um , deduced their tions , THE SPECIFIC INDUCTIVE CAPACITY OF A DIEL .
ECTRIC .
Turpentine .
00035 for rature range from to Benzene 00040 , , , , , , 10 , , { 39 Benzylene 00037 , , , , , , 25 , , Glycerine 00018 , , , , , , 20 , , 48 Thus , it appears that although the two rates of change for Glycerine present no similarity whatever , those of the rest of the four are in a ratio not very frolu 1 to 2 , approach being nearest in the case of Benzylene .
APPENDIX .
( Received October 18 , 1889 ) .
A suggestion having been made that , as the opposite effects of rise of temperature upon solids and liquids wele observed by different methods , it would be well that both should be tested by the same method , a experiment was made on a solid by the method used for see whether the result would agree with that already obtained .
A cylinder of glass was placed between the quadrants and needle of the liquid electrometer , leaving the needle free to oscillate ; and observations were taken at different temperatures , exactly as already described for the case of a liquid dielectric .
The result was always an increase of specific inductive capacity with rise of temperature .
All the precautions for insulation , \amp ; c. , were observed as in the experiments already described .
But the heating was not maintained long enough to secure that the glass had acquired the full temperature of the air in the bath ; so that the results obtained are only qualitative , the change less than that corresponding to the temperature indicated by the air inside the bath .
In one case , where the heating had lasted several hours , the rate of change rose as high as 0024 per degree for a range of C. , a result sufficiently close to that obtained for ] by the other method .
IJDCCCXC .
) PHILOSOPHICAL TRANSACTIONS .
\mdash ; I. On the of Temperature on the Speciftc Inductive Capacity of a Dielectric .
W. CASSIE , M.A. , College , CaJnbridge , in the ersities of Aberdeen } by J. J. THOMSON , M.A. , F.R.S. , endish Professor of Physics in the Unirersity of Received May 24 , \mdash ; Read June 20 , 1889 .
THE object of the experiments describe in this paper was to learn how the specific inductive capacity of a dielectric is affected by change of temperature .
CAVENDISH observed an increase in the capacity of a glass con denser when it was heated , but gave no measure of the effect .
Dr. HOPKINSON observed in light flint glass an increase of per cent. in the capacity between and C. And Messrs. GIBSON and BARCLAY showed there is no appreciable .
in the ) of paraffin between temperatures \mdash ; 1 and C. Except these , no measurements of the efiect appear to have been hitherto published .
present investigation shows an increase of specific inductive capacity with rise of temperature in all the solids* examine , and a decrease in all the liquids except one .
As paraffin , which is a substance comparatively near its melting point at ordinary temperatures , shows no change , these results seem to indicate , as far as they go , that the specific inductive capacity of a substance has a maximum value about the meltinoi point .
But it may be questioned whether the are sufficienlt as yet to warrant so general an induction .
The relation which CLERK MAXWELL'S electromagnetic tlJeory of light indicates between specific inductive capacity refractive index makes it interesting to compare the effects of temperature on these two quantities .
Four of the liquids *The results given here for mica , ebonite , and the first specimen of arc all less fhall those mentioned by Professor J. J. THOIISON , in his treatise on " " Applications of Dynamics to Physics and Chemistry\ldquo ; ; because the results quoted there wore obtained from an earlier set of observations , which the precautions for instllation were inferior to thosc described ON OF RATURE ON .
which DALE and GLADSTONE have observed the indices ternperatures .
For one of these four the temperature efTects no simil rity 1tever , for another the relation is fairly close to that by the 'llld other not agreeing exactly with the , the elati oloes not differ it very greatly , ot , more , ) , than explained ) di ces in the specimens used .
Tho ents done in lihe Cavendisb idge , and of nking ) Professor J. J. , F.R.S. , , at whose estio they cing the resources of the ttory at my for advice the course of the work .
I. SOLIDS .
in of solids to with the to be nented 1 observe its ) different temperatures .
The condenser consisted of pile of alternate thin sheets of the dielectric and discs of lead foil , with of iron above and below , and a weight on the top to keep them .
The condenser could llot be fixed together in any more permanent way , because the unequal nsion with heat of the different materials lV0U]d have altered the compression of the pile , and so produced a change of capacity greater the effect under investigation .
Special were required to secure that the tion between the sides of the condenser be as far as independent of the temperature .
To secure this the conden ser was suspended in a stirrup from a cket about four feet above the air bath in which it heated .
The suspending wires , which formed the connecti o. the side of condenser in cont , act with the stirl.up , passed without holes in the top of the air bath , and over a short glass tube varnished with sbellac , which was fixed across the bracket above .
The connection for the other side of the condenser consisted of a stiff wire , which also passed without touching through a hole in the top of the air bath , and was cari'ied through the glass on the ) racket .
Thus the condenser had no contact with anything about the bath , supported solely by the glass tube on the bracket , which was always cold and perfectly insulating .
The perfect freedom of the condenser and its easily tested before each observation by tapping the air bath and seeing that no motion was produced in the suspending wires .
The condenser was heated by ising the bath quickly to a temperature considerably that imed at , then ving under it a small me adjusted by experience to .
the right .
The temperature was read by a thermometer passing through a cork in the top of the bath .
About three hours was the shortest time required to ensure that the condenser was heated throughout ; it was usually fFIC ICITY O The condensers needed before settlino b ' to it city in the cold state .
The tures in these preliminary ] ] bove the highest used in the actual observations .
The capacities were measured by the method used by Professor J. J. }ISON in his determination of ratio of the electrostatic and gnetic units ( ' Phil. Trans 1883 ) .
It is thus described in aper : " " In a Wheatstone 's bridge , A , with the galvanometel at , and the battery between A and , the circuit is not closed , but the points and are connected with the two poles , and , of a commutator , between which a travelling piece , moves backwards and forwards ; is connected witI ] ) ) of condenser , the other plate of which is connected with D. Thus when is in ) tact with the condenser will be ] , and until it is charged , electricity will flow into it from the battery ; this will produce a momentary the various of the bridge .
the moving piece is contact with two plates of the condenser are connected , and the condenser will discharge itself through , and as the of is infinitesimal in comparison with the resistance of Fig. 1 .
any other circuit , the discharge of the condenser will not send an ] moun t ectricity through the galvanometer .
Thus , if we make the moving piece oscillate quickly from to , there will , owing to the flow of electricity to the condenser , be succession of momentary currents through the galvanometer .
The resistances are so adjusted that the deflection of the galvanometer produced by these momentary currents is balanced by bhe deflection due to the steady cul.rent through the galvanometer , and the resultant deflection is zero .
When this is the case there is a relation between the capacity of the condenser , the number of times the condenser is charged and discharged per second , and the nces in the various of the bridge This relation , is worked out in the , is expressed by taking to represent the istances of A , A , A respectively .
The resistance MR. W. CASSIE ON THE ]FFECT OF TEMPERATURE ON of and so small as to be negligible .
Then if the condenser has a capacity , and is and discharged n times per cond The resistances used were about between 3000 itnd 6000 , , ohms .
With these values is correct to within2 per cen and is the formula used here .
The commutator P ( fig. 2 ) was that used by Professor THOMSON and is thus described in his paper:\mdash ; " " The current from two cells passes .
through a tuning-fork interruptor , and then through the coils of an electromagnet .
is a strip of brass with a piece of iron attached to it .
When there is no current passing the electromagnet , the elasticity of the rod makes it press against a screw , which is electrically connected with a binding screw : when the current passes through the electromagnet , the magnet attracts the iron attached to the rod and brings it into contact with stop , which is electrically connected with the binding screw S. The letters indicate the same points in this figure as in fig. 1 .
All the ] aces where contact is made by the ating piece are covered with platinum , and the TH SPECIFIC CAPACTTY A DI whole aArang.ement is fastened down to an ebon ite board .
As the current passes through the coils of , the piece strikes alternately yainst the parts and ; when it strikes the opposite plates of the condenser are connected with the two poles of the battery ; when it strikes against the condenser is discharged To be able to allow for conduction or absorption in the diclectric by means of observations with forks of different rates , we must know the ative times during which the vibrating piece is in contact with , in contact with , and travelling from one to the othel .
I therefore fixed stiff arm to the end of and took with it on stnoked paper , lnoved across while interruptor was .
These tracings showed that when the interl.uptor is going steadily ( this is easily known by the sound ) is in contact with and for times .
and the time occupied in Qsing irom one to the other is to the time of contact .
The battery used six LECLANCH cells , resistances taken from a box by ELLIOTT .
TTuning forks were used , , C4 , and 49 complete vibrations per second .
The forks were worked by a current the storage cells in the The coefficients of expansion with temperature of lead , ebonite , and glass are 000029 , 000077 , and ctively .
That of mica does not seem to have been determined ; but it is probably less than that of lead .
So that the combined effects of expansion with rise of temperature do not affect the capacity before the fifth decimal place ; and , as the results do not profess to be correct to more than four decimal , the expansion of the materials may be neglected .
Some observations were also taken with ebonite by an electrostatic method based upon that described in MAXWELL'S " " Electricity netis m S229 ( second edition ) ; the only difference being that there was no guard .
The condenser was with a variable condenser by dividing a charge between them , separating them , and connecting opposite poles together to the electrometer ; then , if they were unequal , the difference deflected the electrometer .
For this method a special key was required to make the connections quickly and to keep the electrometer to earth until everything except the charge to be measured been discharged .
If the electrometer were not kept to earth it would be deflected in spite of ordinary screens by induction inside the key on account of the potentials required .
When neither of the condensers being any absorption , there is no difficulty with this method , and any ordinary quadrant electrometer may be used .
But when this is not the case the residual discharge will to come out as soon as the condensers are connected , and this makes the needle lways go ultimately to the same side .
The only way to get a balance is bo make the variable condenser too large , and gradually diminish it until the initial motion of the needle due bo it disappears .
This requires a needle with a short of , because with a slow R. W. CASS IOF ON needle the ) overpowers the first effect before the needle to it .
With a small ebonite it as just possible to make an electrometer sufficient by making the connection as short as possible ; but with a large condenser it was useless .
Consequently , this experiment would require an electrometer with a needle of very short time of swing .
One with small cylindrical needle quadrants seems most suitable .
The condenser was made of twenty-two sheets of brown Muscoyite with markings .
The sheets were about inches diameter , and the thicknesses varied from about a quarter to about a tenth of a millimetre .
Some difficulty found in getting rid of slllface conduction on the mica .
No cleaning would make it insulate , but in the end perfect insulation was secured by putting a border of shellac val.nish round the edges of the sheets .
The shellac did not come between the plates of the condenser , and so did not affect the capacity .
The highest temperatures in the preliminary were less than 14 C. The insulation was constantly tested by an electrometer , and was perfect throughout for temperatures below 1 C. Aboye this it began to give way .
The results are given in the following table and diagram ( fig. 3 ) .
THE SPECIFIC CAPACITY OF A The fact that both periods give the same value shows that , to be expected from its crystalline structure , there is no absorption in for those short times of charging .
Ebol ?
ite .
The ebonite condenser was made of venty discs of nite , about third of a [ nillimetre thick , and three and a quarter inches in diameter .
The ebonite sheets were carefuIly cleaned with paraffin dissolved in benzol , and this gave perfect insulation without any shellac border .
At temperatures above C. the capacity began to increase rapidly owing to softening .
The highest temperature reached in the preliminary heating was a little under C. The insulation was constantly tested as in the case of mica , and found perfect up to about C. The results are given in the following table and diagram ( hg .
4 ) .
As the insulation was perfect , the greater capacity shown by the longel time of charging must be due to absorption\mdash ; an effect which the structure of would lead us to expect .
The colunln of values for ntaneons charge have been calculated on the assumption that the absorption goes on at the same from the beginning of the charge , as during the interval observed . .
W. CASSIE ON ] ON of of change corrected for instantaneous charge .
4 lo C. 00037 .\mdash ; In the obseryations on ebonite by the electrostatic method the condenser consisted of two sheets of ebonite between the iron discs , and with copper between them .
The condenser rested upon a glass tripod inside the the leading entered the air bath through ] tubes .
The city of this adjusted so as to be within the of a fine sliding condenser in the atory against which it balanced ; and the readings were taken on an Elliott electronleter .
It required care and patience to make so slow an electrometer suflice , for the reasons already stated .
The insulation of all parts of the also required much ttention .
The cold temperature , obtained by .
iced round the bath contain ing the condenser .
Subsequent observations showed that this ffected the tion and so increased the apparent diIninution of capacity .
The greater rate of shown by the electrostatic results is probably due to the time of ging by the key worked by hand greater with the tuning fork .
Any leak would produce an apparent decrease of capacity by this method ; so that , although the electrostatic results are of less weight than the others , the reement of the results of two methods in which main of error ten ds in opposite directions is a ) confirmation of their accuracy .
The ostatic results are given in the liate of change of Temperature .
capacity per C. C. C. C. C. C. C. C. C. C. C. C. C. C. C. C. C. 0008 0010 0016 \mdash ; \mdash ; THE SPECIFIC ) OF A glass condenser was made with seventeen she of thin -slide -glass , about three inches diametel .
They consisted of a soda glass of high conductivity ; so that , although the discs were carefully cleaned and bordered with shellac as in the case of mica , the insulation of the condenser , as tested by the electrometer , was never perfect .
Since it is only the of the condenser that affects the galvanometer , the discharge passing sensibly ] through ( fig. 1 ) , the defective insulation introduces simply a steady current during each time of arging .
Each time the condenser is charged , the quantity of electricity passing and consists of ( 1 ) the charge of the condenser , and ( 2 ) the current through the condenser , which sts during the time the vibrator is in contact with S. For our present , the current m.ay be considered as immediately established at full strength when the circuit is closed .
So that , neglecting absorption , the apparent capacity exceeds the true capacity by the conductivity of the condenser multiplied by the duration of the contact between and S. And from observations on the apparent capacity with two forks of known speed , the true capacity can easily be found .
The highest temperature the preliminary heating was less than 11 C. The results are given in the following table and ( fig. ) C. 4795 4320 C. 4795 4320 C. 4795 4320 C. 4795 4320 C. 4795 4320 C. 4795 4320 C. 4795 4320 C. 4795 4320 C. 4795 4320 C. 4795 4320 C. 4795 4320 C. 4795 4320 C. 4795 4320 C. 4795 4320 C. 4795 4320 C. 4795 4320 034 0016 053 0017 077 Fork making 64 complete vibrations per secolld .
CCCXC . .
W. ON EFFFCT OF fURE ON of change corrected for instantaneous charge . .
Fig. 5 .
next three tables and diagram give the esults of exper.iments upon another condenser of microscope-slide cover-glass .
This specimen of the glass had less conductivity than the previous , but the precautions for general insulation were not equal to those in the other experiments described .
The condenser rested on a glass tripod inside the air bath , and the leading wires passed into the bath through glass tubes .
The condenser consisted of twenty-two discs of glass cleaned and bordered with shellac in the usual way , with discs oflead foil between them .
The curves almost exactly parallel up to between and , and diverge at higher temperatures in consequence of conduction ; so that we may take the temperature change of specific inductive capacity for this specimen of glass to be about per cent. up to An attempt was made to observe this temperature effect for shellac .
The condenser was made by dipping the lead discs in shellac varnish , and carefully and thoroughly evaporating the alcohol .
But after the condeIlser was made it was found that at a temperature below the shellac softened , so that the plates were pressed together by the weights on the top of the condenser .
To diminish the weight would have been useless ; because , in any case , it would have been impossible to say how far a change of capacity was dne to softening , and without a weight all the results are quite unreliable .
CAPA C OF A C. emperature .
apacity aRate ocper d With 99 C. C. C. C. C. C. C. C. C. C. C. C. C. C. C. C. 4010 C. 4010 C. 4010 C. 4010 C. 4010 C. 4010 C. 4010 C. 4010 C. 4010 C. 4010 C. 4010 C. 4010 C. 4010 C. 4010 C. 4010 C. 4010 C. 4010 C. 4010 C. 4010 C. 4010 00230 00369 00251 With ' ' 64 fork " " ( between 64 and 6.5 ) .
With 49 fork .
Fig. 6 .
II .
paratus for liquids consisted essentia]]y of a quadrant electrometer immersed .
in the liquid .
The heating done by a water bath ; and on account of the highly inflammable character of some of the liquids experimented upon , the water was heated in a separate vessel at a ) distance , from the top and bottom of which pipes went to and in fig. 7 .
The hot water was pumped through in the direction Fig. 7 .
The provision for the insulation was on a similar principle to that adopted in the case of solids .
Each quadrant was supported by a long Qtiff rod , the upper part of which was fixed in a varnished glass tube .
These four glass tubes were separately SPECIFIC INDUCTIVI CAPACTTY OF A ) I.3 clamped to a support about 18 inches above surface of the liquid ; so that the quadrants touched except the liquid and this perfectly insulating support .
The needle was suspended by a fine wire from the same height so as to oscillate under torsion .
The hollow top of the water batb had seven holes through it , as shown in the figure .
Through the outside four passed , without touching , the rods supporting the quadrants , through the centre one hung the needle , and through the two next the centre thermometers could be inserted .
This top was permanently fixed to the support of the needle and quadrants ; and the lower part of the bath containing the liquid was oved up on a smooth sliding arrangement into contact with the top so as to immerse the needle and quadrants .
Above the fixed top of the bath was a box with a window on one side , through which the movements of the needle and mirror were read by a scale and telescope .
The deflections of the needle were observed when first to one pair of quadrants and then to the pair .
As the needle and quadrants were parts of cylinders between three and four inches diameter and about a quarter of an inch apart , a large electromotive force was required to produce a deflection .
The electromotive force had also to be rapidly reversed in direction to avoid as far as possible polarisation , convection , \amp ; c. , in the liquid , The electromotive force was obtained from a Ruhmkorff coil without a condenser , and with a high resistance between the terminals to prevent sparking .
This high resistance consisted of a wide glass tube , about 6 inches lon , filled with distilled water , and having a thick copper wire through a cork at each end .
By altering the distance of the ends of the copper wires in the water the could be adjusted and the deflection controlled as desired .
The coil was worked by a current from the storage cells in the Laboratory .
As the electromotive force given by this arrangement was variable , and also the loss of electromotive force by conduction was different for each liquid and for each temperature , it was necessary to be independent of such changes .
dingly a second electrometer was placed between the terminals just outside the liquid one , which being always in the same state gave the comparative values of the electromotive force .
This second electrometer was an Elliott 's lecture-room pattel.n with two quadrants removed and the needle connected to one of the remaining quadrants .
The advantage of this form for the present purpose was that by moving the needle up from the quadrants the deflection could be diminished to any desired extent .
Then the quotient of the deflection of the liquid electrometer by the deflection of the second electrometer was proportional to the specific inductive capacity of the liquid .
For every deflection of the liquid electrometer a pair of deflections of the second electrometer were taken with the needle connected first to one pair of then to the other .
The genel.al arrangement of the apparatus is shown in the figure .
The liquids experimented upon were turpentine , carboll , glycerine , benzene , benzylene , olive oil , and paraffin oil .
Methylated spirit , also tried ; but cond so great that deflection could be obtain ed. All except oil showed decrease of specific inductive capacity with rise of temperature .
The oil was that used in the in the Laboratory , and its exceptional behaviour may have been due to some secondary action arising from impurity .
The results are given in the following tables and diagram .
And to show the way in which the observations were made , the readings are given for turpentine at two eratures .
to that at the , SPECTFIC CAPACITY OF A ] Temperatul inductive capacityRatio opecificto t Rate of lowest temperaParaffin oil : ture EADINGS for Turpentine .
ElectrometelLiquid .
Ratio ometer r 16.5 6.2 6.2 16.5 6.2 6.2 Needle cnnected t\ldquo ; , othe .
' 16.5 6.2 6.2 Needle cnnected t\ldquo ; , othe .
' 16.5 6.2 6.2 Needle cnnected t\ldquo ; , othe .
' 16.5 6.2 6.2 Needle cnnected t\ldquo ; , othe .
' Needle cnnected t\ldquo ; , othe .
098 Mean 100 098 098 Mettn .
W. ON UltL ON If the relation indicated by MAXWELL'S ] omagrnetic theory of light held good , the specific inductive capacity were equal to the square of the ctive index , then the rate of change of specific inductive capacity with tenll ) ) ought to be twice of the efractive index .
Amongst the liquids , for which DALE and have observed the refracbive indices at erent t four of those dealt with here .
The ates of change , with temperature of the efractive index for the A in the spect1um , deduced their tions , THE SPECIFIC INDUCTIVE CAPACITY OF A DIEL .
ECTRIC .
Turpentine .
00035 for rature range from to Benzene 00040 , , , , , , 10 , , { 39 Benzylene 00037 , , , , , , 25 , , Glycerine 00018 , , , , , , 20 , , 48 Thus , it appears that although the two rates of change for Glycerine present no similarity whatever , those of the rest of the four are in a ratio not very frolu 1 to 2 , approach being nearest in the case of Benzylene .
APPENDIX .
( Received October 18 , 1889 ) .
A suggestion having been made that , as the opposite effects of rise of temperature upon solids and liquids wele observed by different methods , it would be well that both should be tested by the same method , a experiment was made on a solid by the method used for see whether the result would agree with that already obtained .
A cylinder of glass was placed between the quadrants and needle of the liquid electrometer , leaving the needle free to oscillate ; and observations were taken at different temperatures , exactly as already described for the case of a liquid dielectric .
The result was always an increase of specific inductive capacity with rise of temperature .
All the precautions for insulation , \amp ; c. , were observed as in the experiments already described .
But the heating was not maintained long enough to secure that the glass had acquired the full temperature of the air in the bath ; so that the results obtained are only qualitative , the change less than that corresponding to the temperature indicated by the air inside the bath .
In one case , where the heating had lasted several hours , the rate of change rose as high as 0024 per degree for a range of C. , a result sufficiently close to that obtained for ] by the other method .
IJDCCCXC .
)

