]\gt ; PHILOSOPHICAL TRANSACTIONS .
I. the nination of VERDET'S Constant in bsolnte Units .
By J. .
H. GORDON , 1 ? .
A. , Gonvilte and Caivs College , .\mdash ; lst and 2nd Memoirs* .
nicated by J. CLERK MAXWELL .
Rcceivod Junc June [ PLATE 1 .
] \mdash ; The whole of this work has been done under Profcssor CLERK superintendence ; ho suggested the method and nearly all the details , any merit which the ration may havo belongs to him .
Hc is , however , in no way responsible for any errors there may be in the results .
] INIIIODUCTORY . . . . . . . . . . . . . . . . . . . . . .
Tan . . . . . . . . . .
) Determination of the . . . . . .
1 Verification of the )Ieridiau . . . . .
The Helix . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3 The . . . . . . .
17 The Dynamometers . . . . . . . . .
3 The Optical Expcriments . . . . . . . . . .
19 Polarizing Apparatus . . . . . . . .
3 Value of . . . . . . . .
20 Analyzing Apparatus . . . . . . . . . . . .
3 Vibration Lxpcriments . . . . . . .
The Jellett Prism . . . . . . . . . . . . . . . . .
Vibration Lxperiments at Pixholmc . . . .
The Divided Circle . . . . . . . . . . . . . . . . . . .
4 Galvanometer . . . . . . . . . .
4 Tube . . . . . . . . . .
orsion CctionTcmpcrature Corrcction . .
The Experimenting Table Measurement of Horizontal Distances . . . .
4 Rate . . . . . . . . . . . . . . . .
} Corrected Vibration Timcs . . . . . . .
30 THB of at Pixholmc . . . . .
32 Determination of Number of Windin . . .
4 Determination of Areas . . . . .
) form and valuo of . . . . . .
\amp ; Calculation of Strength of Currcnt in Helix . .
14 ; . .
Formula for . . . . . . . . . . . . . . . . . . .
14 lJ1x.\mdash ; lnalsis of bisulphidc . . . . . . . .
* In a former memoir which Prof. did mc the honour of communicating to the Royal Society , and which was read June , ] , the object of this research explained and a result given .
The method employed , however , was faulty in several respects , the two most important of which were : 1st .
The rotations produced were exceedingly small , so that a slight error in the determination of the plnne of polarization introduced a large error into the results .
2nd .
A tangent-galvanometer used to determine the strength of tho current .
Tn the oxpcriments which arc the subject of the present lIcmoir I haYC , at Prof. , used MR. J. E. H. GORDON ON THE uINATION OF CTORY .
IN the year 1845 ARADAY discovered that if plane polarized light passes through certain media , and these media be acted on by a sufficiently powerful magnetic force , the plane of polarization is rotated .
About the year 1853 .
VERDET commenced a long and exhaustive examination.of the subject , and his first result was published in Ann. de Chimie et de Phys. 3 ser. tom .
In MAXWELL'S lectricity , ' Art .
808 , vol. ii .
p. 400 , M. VERDET'S results are summarized as follows:\mdash ; The which the plane of polarization is turned is proportional ( 1 ) To the distance which the ray travels within the medium ; ( 2 ) To the intensity of the resolved part of the magnetic force in the direction of the ray .
( 3 ) The amount of rotation depends on the nature of the medium .
Prof. CLERK MAXWELL then goes on to say " " hese three statements are included in the more general one that the angular rotation is numerically equal to the amount by which the magnetic potential increases from the point at which the ray enters the medium to that at which it leaves it multia coeffi , cient which for diamagnetic media is generally positive.\ldquo ; The object of the present research to determlne the for a partlcular medium and for light of a particular wave-length .
" " VERDET'S constant " " is this ent for a cular medium and ray , or , otIler words , it is the angular rotation produced on that ray and in that medium by a difference of magnetic potential equal to unity .
In order that the measurements may be expressed absolute units , necessary to modify .
VERDET'S mode of proceeding in several respects .
In particular an electromagnet with an iron core is unsuitable for this investigation , for both the amount and the distribution of the magnetic force between the poles depend on the properties of the iron core , and cannot be deduced from the strength of the current in the helix .
FARADAY'S heavy glass and other media having the highest power of ' the plane of polarization were also unsuitable to be used.as standard media on account of the difficulty of .
specimens exactly alike .
The following lnetllod was therefore adopted:\mdash ; The magnetic force was produced by means of an current in a helix without an iron core , and bisulphide of carbon enclosed in a tube with glass ends placed within the helix was chosen as the medium .
The constants of the helix were measured by the methods described in pages 4-14 .
The strength of the current in it lvas deduced from the deflection of a small magnet -\mdash ; bisulphide of carbon instead of water , which was nsed in the earlier experiments , thus increasing the rotation , and have made the helix act as its own galvanometer by suspending a magnetized needle in its neighbourhood .
In consequence of improvemonts in the mothods , I have , through Prof. MAXWELL , requested the Council of the Royal Society to permit me to withdraw my first Memoir , and have at the same time embodied in the prcscnt paper thosc portions of the first which are necessary for a due understanding of the second .
IN ABSOLUTE UNITS .
suspended near to it and outside it , and the rotation was measured by means of a divided circle .
The methods by which the light was polarized and analyzed are described later .
The investigation then resolved itself into three parts:\mdash ; ( 1 ) The determination of the constants of the helix .
( 2 ) The determination of the ratio which the rotation per unit of length bore to the tangent of the deflection of the suspended needle .
( 3 ) The determination of the horizontal component of the earth 's magnetism at the time and place of observation .
The greater number of the experiments were made in my laboratory at Pixholme , near Dorking , but the detelmination of the number of windings and part of the determination of the are of the helix were in the Cavcndish laboratory in Cambridge .
DESCRIPTION OF IENTS EUPLOYED .
The was about centims. about 5 centims. internal meter , and centims. external , and contained about .
of No. 20 cotton-covered copper wire .
The insulation-resistance was exceedingly , sufficiently so to ble me to neglect leakage .
The resistance of the helix was .
unit .
The dynamometer consisted of an ebonite , about 12 centims. diameter , on which were wound six turns of wire .
It was made for me by Messrs. LLLIOTT , and the mean diameter of the coils accurately gauged .
The large namometer was the instrument belonging to the British Association .
It is described and figured in MAXWELL'S 'Electricity , ' vol. ii .
fig. 54 , p. 330 .
( The suspended coil there shown was not used .
) The pf consisted of a Nicol 's prism , rather more an 1 centim. diameter , fixed into a collimating tube taken from a oscope .
The slit was removed , and the Nicol substituted in its place .
The analyzing apparatus.\mdash ; This , togethel .
with its circle and the Nicol , is the property of the Cavendish laboratory , and was kindly lent to me by Professor .
The analyzer was a Jellett plism , described in the next paraglaph .
It was unted in a tube , together with a lens and some diaphragms , which gave very good definition ( see Plate 1 ) .
The Jellett consisted of a piece of Iceland spar , of which the ends were cut normal to the sidcs .
It was then divided by a plane , making an angle of about with the plane containing the two long diagonaJs .
One half was reversed , and the two cemented together again .
The prism was mounted in a brass tube , a diaphro , with a hole of some 3 or 4 millims. diameter in it , was placed across one end .
On looking in from the other end of the prism , the hole was seen by the ol.dinary rays in the form of a circle divided by a * This prism is described by Prof. JILLBTI in the ' Transactions of the Royal Irish Academy , ' vol. .
Prof. JELLETT also had the kindness to advise the author as to the form of his instrument best suited to this investigation .
line across it , the light of the two semicircles polarized in planes making with each other an of about The image of the holc by the extraordinary rays consisted of two semicircles , one to the right and the other to the lcft of this circle .
A second hid the latter .
Now when light that has pa. sed through a Nicol is examined with the plism , and the latter turned so that the light of one half of the circle is extinguished , the other half is slightly illuminated .
If the Jellett be turned through rather less than degrees , the second half of the circle will become dark , and the first will be slightly illuminated .
It is obvious that there is a position between these two where the illumination over the wholc circle is uniform , and this position can be observed with considerable accuracy .
With the arrangements that were used in this work the probable error was about 1 ' .
divided circle was made for this investigaiion by Mr. BnOlVNING , and is about 14 centims. diameter .
The circle turns against two fixed verniers .
It is divided on brass to , and the verniers read to .
to to By estimation 1 ' can be read with perfect accuracy .
It is moved by a screw of long pitch , working in a thread cut round its .
This gives a sufficiently quick motion to enable changes in the illumination of any part of the field to be noted , and is yet capable of very delicate adjustment .
The ometer consisted of a minute suspended magnet and mirror , weighing about 1 grain , acted on by the helix ( outside of which it was placed ) , and observed by a telescope and scale .
for .
the bisulphide was of brass , with glass ends , and projected some 20 centims. from each cnd of the helix .
The experimenting table was built of massive brickwork , with a calefully levelled top of fine cement .
It was about metre squale .
A continuation of it supported the suspended magnet , and a separate brick pier the telescope by which the magnet was observed ( see Plate 1 ) .
The apparatus for ?
horizontal distances consisted of a number of conical plumb-bobs , which .
by silk threads from laths laid across a horizontal deal framc the same size as the table , and fixed about 1 metre above it .
Distances projected on the table by these were taken by compasses , and measured on a millim. scale engraved on a massive brass plate by ELLIOTT ( for another purpose ) .
The base of the circle , the telescopes , \amp ; c. were all placed on pillars of brickwork built on the tables , and bedded either in cement or plaster of .
The result was a steadiness such as I have never before obtained .
mination of the Number of Windings .
We must first mine the difference of magnetic potential at the two ends of the helix when a unit current passes ough it .
It is a quantity which call , and is a function of the number of windings and their arrangement ; for if know the nctic VERDET'S CONSTANT IN ABSOLUTE UNITS .
intensity at each point of the axis of the helix due to a unit current , viz. what force isexerted by a unit current in the helix on a unit netic pole at that point , then know how much work would have to be done to move this unit pole from one end of the helix to the othel when a unit current was .
in it .
But if be the ends and ?
the force at any point , then the above amount would be equal to , ( 1 ) where is the potential at any point .
But the dimensions of magnetic potential are , in the netic system , ; and these are also the dimensions of the strength of an electlic current .
, which is the ratio of the cf these things to the lattel , ?
a value of for the helix was determined by comparison with the great mometer of the British Association , which is deposited in the Cavendish Lnboratory .
The intensities of the magnetic action were compared at 7 equidistant points in the axis of the helix , and the total was obtained by integrating by formula Finite Differences , , viz. where is the distance between any of the points , and is the nctic i at any point in terms of that of the dynamometer .
The mechanical ements were as follows : dynamometer , which consists of two coils about 60 centimetres in diameter , at a distance of about 25 centimetres one from the other , was placed so that the plane of the coils was accurately vertical and in the magnetic meridian .
The axial line of the coils was then carefully found and marked by means of plumblincs and cross-threads fixed to the table .
A -shaped board supported on three .-scrcws placed so that part corresponding to the stem of the passed through the coils in a planc palallel to their axis ; on this the elix was laid and its axis into exact coincidence with that of the coils .
A boxwood cylinder about ) ccntims .
was turned to fit nicely into helix ; long thick brass wire in a handle was fixed into one end , and a brass about 5 centims. .
was fixed near one of the other .
* Objections have been taken to the use of this formula , which ves , it is observed , much more to and than to and ?
, and docs not furnish an approximation of a itimate analytical ) ) cter .
Whilc fully tho force of these objcctions , I ve not it whilc to make an tion n would involve repeating most of the arithmetic in the paper , for this reason : the expcrimonts , being made with resistance-coils , arc susccptible of such close accuracy that the errors particular determine evcn ] by 5 , cannot perccptibly affect the value of deduced .
To the end of the latter a magnet and mirror weighing only about a grain was hung by a silk fibre , so that when the pin was at the highest point the magnet the axis of the cylinder .
The helix being placed coaxial with the coils was pushed emdways so that the centre point of the coils was just outside one end of it .
The cylinder was then inserted and adjusted , by means of cross-wires , so that the lmirror hung exactly at the centre point , and a mark was then made on the handle corresponding to a mark on the stand of the instrument , by reference to which the could always be brought to the same posi Thus while the helix was slid along the axis , the magnet inside it could always be placed at the centre point of the dynamometer .
The distance between the inside ends of the helix was diyided into six equal parts by means of seven pins stuck as sights into a slip of wood fixed along the top edge .
Thus by sliding the helix along till any one of the was between the vertical threads placed midway between the coils , the force at that point due to the helix could be compared with that due to the dynamometer by means of the needle at the centre .
The comparison was made by sending currents of different intensities through the helix and coils in opposite directions , and varying them till there was no action on the needle .
This was effected by dividing a current and one portion through the coils and the other through the helix , and interposing different resistances in each branch .
The annexed diagram shows the connections .
is the battery , and the resistance-coils .
, the dynampmeter-coils .
And let these lctters also represent , the helix .
their resistances .
the magnet and mirror , whose angular position is observed by , the scalc and lamp .
are contact-keys .
Now when the actions on the needle are equal , the powers of the coils are inversely as the currents in them , that is , directly as the resistances and VERDET'S CONSTANT IN ABSOLUTE UNITS .
The observations being repeated at each of the seven points , we have the intensity of the netic action of the helix at each of those points in terms of that of the dynamometer .
Thus if be the power of the dynamometer with unit current , and the power of the helix with the same current at a point , then the total force of the helix ( that is , the difference of ) ' etic potentials at its ends ) is where is the of the hclix .
In the helix used the results were obtained:\mdash ; centims. , ' take ' Hence we have for the values of , by the of the above , Hence centims. Now Now in order that this may be a number it is necessary that [ P ] should equal ; and if consider the equation of moments we shall see that this is so .
For consider a small magnet ( length ) at the centre of the dynamometer , we have , when a passes , [ PC .
] , or .
( 4 ) Now the value of for any helix with unit current taken with respect to the whole length of its axis produced to an infinite length in both directions is , by Art .
676 of Professor CLERK MAXWELL'S ' Electricity , ' , where is the of windings .
When the length is finite compared lrith the radius , the value of for that part of the axis which is included between the ends is ( see CLERK MAXWELL'S ' Electricity , ' Art .
C76 ) .
Now if we , taking the mean radius , this will give the number of windings in the helix .
as centims. , and A , we have last there al.e 1028 on the helix .
*Professor CLERK method of representing the units by quantities in has been used throughout the paper .
Thus a length is represented by , where is number and [ L ] the unit of length .
IN ABSOLUTE UNITS .
Ver.ification .
Now in the length of the helix there are 91 windings on the outside layer , and the ratio of the length to the difference of the internal and extelnal radii is about .
3.47 assuming that the number of layers per centim. of radius is the same as the number of windings per centim. of length ( it would really be a little greater , as they fit into each other , oooooooo ) , we have which is sufficiently neal the calculated result , viz. , to show that no large mistake has been made , as , for instance , writing down a .
with a wrong index .
Closer agreement could hardly be expected , as the helix was made a different purpose , and no particular pains were taken to wind it .
It is also probable that the instrument-maker took more pains to wind the wires of the outside layer , which could be seen , close together than those of the inside ones , which were hidden .
nination of Areas .
1st \mdash ; To calculate the strength of a current in a helix from the deflection of a needle placed outside it , it is necessary to know 2(A ) where is the area of any winding .
To find 2(A ) the helix so placed that a vertical plane through its centre , and normal to its axis , contained the magnet and mirror when only acted on by the earth 's magnetism .
In this plane , between the helix and magnet , and so arranged that it could be slid along in it , was placed the small dynamometer .
The of magnet , dynamometer , and helix were in the same horizontal line .
The same culrent was sent in opposite directi through both helix and small dynamometer* , and the latter moved till there action in the magnet .
The tances of the centres from the magnet were then found to be:\mdash ; Centre of helix to magnet centims. dynamometer to 2nd method.\mdash ; From these data a value of the area was obtained .
On looking through these experiments , , Professor MAXWELL was of opinion that the dynamometer used was not large enough to ensure accuracy , so he had the kindness to compare the helix with the great B.A. dynamometer at Cambridge .
In his experiments the helix and dynamometer were placed concentric , different currents were sent opposite ways * The circular base of the dynamometer was cemented to a square piece of glass , which slid on a little wooden stand , nished with a guide-bar lying in tho magnetic meridian .
\amp ; e paragraphs on measurement of horizontal distances and determination of meridian , pp. \amp ; through both , till there was no action on a small magnet suspended at a point in a vertical plane perpendicular to and bisecting he common axis of the dynamometer and helix .
Professor MAXWELL found that when ffom to centre was so great that the diameters of the coils could be neglected in comparison , the currents producing equilibrium had the ratio 1 : From these data another value of the area was calculated .
In both cases the following reasoning was used : Calculation of , the sum of the areas of the windings of the Helix .
We proceed as follows:\mdash ; ( 1 ) We first obtain an expression for the force exerted at a point by one winding , at a certain distance and of a certain area , carrying a unit current .
( 2 ) We then , by calculation , find what the action of a certain standard coil of known area would be with a unit current .
( 3 ) We then , by experiment , find what that action is with a certain arbitrary current whose ratio to a second current is known .
( 4 ) We then , by experiment , find what the action of the helix is with this second current , adjusting the ratios till the actions are equal .
( 5 ) This eliminates the variations of the current , and enables us , by means of expression ( 1 ) , to compare the areas in ( 3 ) and ( 4 ) ; one of these being known , we can obtain the other .
To find an expression for the force exercised by a circular current in a given direction at a point , we must first find the potential at that point , and them differentiate along the given direction .
For a winding carrying a current of strength we may substitute a magnetic shell bounded by the winding and of strength .
But by MAXWELL'S ' Electricity ' ( Art .
409 ) , the potential at any point due to a magnetic shell is , . . . . .
( 6 ) where is the .
strength of the shell and the solid angle subtended by it at the point .
We have therefore to find an expression for Now in Arts .
670 and 694 we find .
( P ) ; ( 7 ) for when unity , , where radius of winding ( which in this experimental case becomes a great circle ) is the distance from the centre of the circle to the point where the potential is to be and the potential of a stratum of matter of surface VEBDET'S CONSTANT IN ABSOLUTE UNITS .
density unity spread over a hemisphere bounded by the circle .
Expanding this , we obtain in Art .
695 the expression for the potential of a point outside the circle , , ( 8 ) where and are the angles in Art .
694 .
Let be the plane of the coil ; then the force to it is obtained by differentiating the potential with regard to , and Then for the force we get , changing the independent variables from , to ; .
( 9 ) and lemembering that , and . .
( 10 ) But when , as in case , the point where the potential is required is in the plane of the coil , we have and each , which , and reduces the expression ( 10 ) to Force perpendicular to plane of coil .
( 11 ) Now by Art .
694 , p. 301 , .
( 12 ) The values of this are given in Art .
698 , and in them we must put Hence the force for one winding is .
( 1 ) . .
( 13 ) This is the force exercised by one winding of radius on a unit netic polc in its plane distant from its centre , when the carries a unit current .
The area of the winding is The force exercised when the strength of the current is is of course .
Then we may write .
( 14 ) where is the force on a unit pole due to a unit current .
But the small dynamometer consists of windings all of the same diameter , and approximately all in one plane , and therefore this expression is so nearly trUe that it may be used for it .
When thus used it will be distinguished by dashes on the letters .
The second and third terms are required because in the experiments the small dynamometer is placed very close to the suspended magnet .
The effect of the helix may be got by a much simpler method ( see MAXWELL , Art .
676 ) .
Let be the length of the helix , its mean radius , then the outside effect at the distance of ( say ) a metre or more is that of a disc of magnetism at one end and of \mdash ; magnetism at the other .
For unit current the moment of the magnet is ; hence its strength is , which is the quantity of magnetism at each end .
Then the value of the potential will be , and the value of the force at will be per unit current .
Thus the force exerted by.the dynamometer is , ( 15 ) .
while , that exerted by the helix , is ( 16 ) But in my experiments ; and when we have no action on the at we have Adding the two expressions we have ( 17 ) Now let be the values of for the helix and small dynamometer respectively , we have , when the force on the suspended magnet is zero , , where centims. , and VERDET'S CONSTANT IN ABSOLUTE UNITS .
Also noting that Now , where , is the mean radius of the coils .
Now the diameters of the coil are Internal inches * , External . . .
\ldquo ; which give a mean radius centims. ; whence From ( 17 ) we have , then , ( 18 ) Calculating this from the data jiven , we have as the result of the small dynamometer experiments and From these same experimental data Prof. MAXWELL made a calculation , from which he found Prof. MAXWELL'S Experiments with the large Dynamometer .
In these experiments the helix and dynamometer were placed exactly concentric .
magnet and mirror was suspended rather more than a metre distant from them , first in front and thenbehind , so as to correct any error in centering , and varying currents were sent opposite ways till there was no deflection .
At this distance the helix and dynamometer could each be considered to be replaced by their equivalent magnetic discs , and the difference between and may be neglected in comparison with those quantities themselves .
The formula for then becomes Where there was no action on the suspended magnet , Prof. MAXWELL found Now the area of the great dynamometer is 870200 sq .
centims. * These measurements were made by Messrs. ELLIOIT .
This gives sq .
centims. Thus we obtain the following values for the area of the helix:\mdash ; By Prof. MAXWELL'S experiments with the great B.A. dynamometer\mdash ; sq .
centims. By GORDON'S experiments with small dynamometer ( GOBDON'S sq .
centims. By MAXWELL'S calculation of GORDON'S experiments\mdash ; sq .
centims. Mean of the two calculations of GORDON'S experiments\mdash ; sq .
centims. Finally , Mean sq .
centims. Calculation of the strength of a current in the in terms oj the deflection of the suspended magnet .
Let be the length of the suspended magnet , the horizontal component of the earth 's magnetism , we have from ( 16 ) ml or which gives ( 19 ) Formula for a ) is the rotation expressed in circular measure corresponding to a difference of magnetic potential equal to unity .
Thus , ( 20 ) where is expressed in circular measure , and V the magnetic potential at any part of the tube .
The following formula gives the difference of the magnetic potentials at the ends , due to a curlent C. VEBDET'S CONSTANT IN ABSOLUTE UNITS .
If LM were infinite , it would be , where is the number of windings .
When the length of the tube is finite , and equal to we must subtract the following correctio .
( 21 ) When LA is great compared with Aa , .
; ( 22 ) or , retaining only the first term , we get .
( 23 ) The corrected value is then ( 24 ) Thus the complete mula for is .
( 25 ) The following quantities were taken by measurement:\mdash ; A centims. Distance from to centre of helix centims. ; whence or , and * This formula is due to Professor AXWEIiL .
Tan The scale , which consisted of a wooden board on which was pasted a printed paper scale of millims. , was fixed on a massive wooden tripod-stand fixed to the wooden floor .
telescope was , as before mentioned , bedded with plaster of Paris in a block of stone cemented on to a brick pier .
[ N.B. All the brick piers and tables were built on the solid ground and came up through holes in the wooden floor .
] The scale was so adjusted that , when no current was passing , the line from to the figure seen in the telescope was perpendicular to the scale .
This adjustment was made by hanging a plumb-line near the scale , between it and the mirror , so that on looking into the telescope it was seen in the centre of the field .
A second was then hung in the line between the first and the mirror .
The edge of a square being placed so as just to touch both lines , its base gave the direction of the scale .
As it was not certain whether or not the paper scale had stretched in the pasting , the deflections were determined by taking , with a pair of compasses , the distance between the zero reading and that corresponding to any current , and measuring it on the standard brass scale .
The numbers thus obtained are hereafter spoken of as " " corrected deflections The readings were taken in each direction .
Their mean was called The distance from to the mean zero-point was called , and by measurement we have centims. The values of , where is the deflection of the suspended magnet , were obtained by the formula , due to Professor MAXWELL .
We have \amp ; c. * ( 26 ) Determination of the Meridian .
The best way to have determined the meridian of the helix would have been to have deduced it from that of the suspended mirror .
Owing , however , to the arrangement of the supports , this meridian could not be transferred to other parts of the table .
The meridian was therefore determined by means of a needle furnished with a straw polnter 20 centims. long ( by ELLIOTT ) , belonging to a large tangent galvanometer .
Part of a *This is a very simple formula to use , as we have .
VERDEI'S CONSTANT IN ABSOLUTE UNITS .
sewing-needle was set upright in a hole drilled in the table , and filled with soft clay , at such a height that the magnetic needle when balanced on it just swung clear .
latter was allowed to come to rest under a bell-glass , and then clamped by a simple arrangement .
As the ends of the straw were 5 or 6 millims. from the table , candle was placed close to the centre of the needle , and the ends of the shadow of the straw marked with a pencil .
This meridian was then transferred to any other part of the table by means of parallel rulers* .
The following method was used for placing the helix at a known distance from the suspended magnet and perpendicular to the magnetic meridian , and for making that meridian which passed through the suspended magnet bisect the axis of the helix:\mdash ; A pencil line longer than the helix was drawn through the point on the meridian selected for the centre of the helix and to that meridian .
rrhe plumbbobs wele hung at points , one near each end of this line and in it , and two more in the meridian passing through the suspended magnet .
A mark was made at the centre on the outside of the helix , and cross-wires placed at its ends inside .
When the threads of the .
plumb-bobs hted the cross-wires and the meridian-threads the point on its outside , the helix was properly placed .
Verification of the Meridian .
In order to see that there had been no error , the following verification was made , based on the principle that if the helix was placed symmetrically with regard to the meridian , the deflections produced in opposite directions by reversing the same current should be equal .
Readings .
Current direct .
No current .
Current reversed . . . .
Corrected Deflections . . . .
A difference too small to be wolth correcting The Light .
It was absolutely necessary that the light should be monochromatic , as the rotation of different colours for the same difference of magnetic potential is very different , * See verification of meridian below .
Professor MAXWELL has since pointed out to me that a slight displacement of the helix would not make so large a difference in the deflections as I had expected .
I , however , leave in this raph for what it is worth .
nearly as .
A sodium flame was first tried but found not to be sufficiently steady ; so it was decided to use a white light and a bisulphide-of-carbon prism .
A very powerful paraffin-lamp with a duplex burner was finally chosen as the source of light .
It was used with a lens and a slit , and threw a brilliant spectrum about 8 centims. long on a card .
In the card was a vertical slit which admitted the light into the Nicol .
By moving the prism , or lens , the spectrum could be moved along the card , and any desired colour admitted through the slit into the Nicol .
In order to localize the light thelamp was removed from the lantern , and a spirit-lamp put in its place .
The flame was coloured with thallium , and the green line projected on the card so as to fall upon the slit in it at minimum deviation .
The lantern , lens , and prism being then clamped , the lamp was replaced , and the green light which fell upon the slit was of the same wavelength as that giving the thallium lin. It was intended to make also observations at the sodium and lithium lines , but it was found impossible to isolate the sodium light , while to isolate the lithium the slits had to be made so narrow that the intensity was not sufficient to work by .
In employing the Jellett analyzer , if the light used was not perfectly monochromatic , the effect of the magnetic force was to colour the two halves of the field different colours , rendering all comparison of intensities impossible .
After a great number of preliminary observations ( continued for about seven months ) , all of which were rejected as I found means to improve the methods used , the following sets were made .
In these sets no readings were rejected , except on one occasion , when , finding the first few readings did not agree well with one another , I rejected the whole of them , and after a short rest began them again from the beginning .
In all the observations the Nicolwas so placed that the division line of the " " Jellett\ldquo ; was about horizontal , so that if there were any difference in the light admitted at the right and left of the field it should affect both halves equally .
Thus the light on leaving the lamp passed through the apparatus in the following order :slit , lens , prism , slit , Nicol , collimator , tube , Jellett and its mountings .
" " Current direct\ldquo ; is that direction of current which causes the north-pointing end of the suspended magnet to move toward the east .
After each reading , the circle was slightly displaced at random .
VERDET'S CONSTANT IN ABSOLUTE UNITS .
Set 1 .
February 25.\mdash ; 11 to 11.30 .
Green light centims. 7 cells .
centims. mean of deflections centims. Set 2 .
February 25.\mdash ; 11.30 to 12 .
Same light .
5 cells .
centims. centims. * With regard to the differences between different determinations of the plane of polarization , it must bc remembered that it is impossible to keep the battery-current absolutely constant a series of experiments extending over 10 or 12 minutes .
The seconds only appear in the means .
Set 3 .
February 26.\mdash ; 0 to .
SO Tabulating the results , we have for Set .
2 .
minutes .
133930 $8206 , , 6 1 .
1526 , , 7 See p. 32 .
To reduce the result to the line , we may use one of the formulae given by M.VERDET* ; but as they are only approximate , I have thought it best to state the final result for the thallium line , wave-length on which the experiments were made .
The wa length of the line is Value of H. The ratio of at Dorking to at Kew was determined by vibrating the same magnet at the two places , and determining the times of vibration .
We then know that the values are inversely as the squares of the times of vibration .
Mr. WHIPPLE , Director of the Magnetic Observatory at Kew , had the gleat kindness , Euvres , tom . .
p. 211 .
WATTS ' Index of Spectra .
VERDET'S CONSTANT IN ABSOLUTE UNITS .
not only to lend me a magnet , but also to have it vibrated at Kew , and to determine the temperature coefficient and the correction for torsion of the thread .
The magnet was , at Pixholme , suspended to the same stone support and observed with the same telescope and scale as the smaller magnet used to determine the strength of the currents .
It was therefore in exactly the same place as that magnet had beeIl .
The method used for taking the time of vibration was explained to the author by Prof. MAXWELL , in a letter from which the raphs are extract " " The best way to take the times is to begin with getting the approximate time of an odd number of half-vibrations 1 , 3 , or 5 , according to convenience .
Take such an odd number as will have a period of nearly half a minute or 20 or something easily leckoned , then note down the time of passage through a point near the point of rest , marking it when it goes to right , and \mdash ; when it goes to left , and observe , say , 11 passages , , at convenient intervals .
Then after a time equal to that of 100 vibrations or near it , begin again and observe !
1 passages .
" " From the two pairs of 11 passages find the time of vibration roughly , and thence the number of passages between the first and second set .
Finally , find the time by dividing the mean interval between the first and second set by the numbel of passages .
By taking the passages alternately - , you eliminate the error arising from taking the passage at a wrong point of the swing A large clock beating seconds was so placed that the observer at the telescope could see it by merely looking round .
The times of passage were taken as follows:\mdash ; The hour and minute having been written down by an assistant , the observer noted the second , and went on counting the beats by ear .
Then looking into the telescope , he observed the time of every 7th passage , and called it out to be written down by the assistant .
The passages were thus always alternately - .
The time occupied by 7 passages was sufficient to allow of the clock being looked at to get a fresh beat between each .
The approximate times at which each passage might be expected , within , say , 3 were previously written out and placed in a convenient position .
This prevented the possibility of a 5th or 9th passage being observed instead of a 7th .
A thermometer hung in the case ( wood with glass front ) in which the magnet swung gave the tempel.ature .
The following Tables give the results of the experiments\mdash ; that of ' ' Vibrations at Kew\ldquo ; being sent by Mr. WHIPPLE , and those of ' ' Vibrations at Pixholme\ldquo ; being determined by the author .
* See also MAXWELL'S ' Electricity , ' Art .
456 , vol. u. p. 102 .
Observatory .
Observations of Vibration of Magnet lent to Mr. J. E. H. GORDON .
correction for the decrease of the magnetic moment of the magnet produced by an increase of temperature , and is expressed by a formula of the form\mdash ; correction to ( 27 ) being the observed temperature , and an adopted standard temperature .
Vibration Experiments at Pixholme .
From some preliminary experiments which need not be detailed it was found that the time of a half-vibration at Pixholme was approximately seconds .
The following Table gives the time of every 7th passage taking Set 1 ( A ) .
Apri123 , 1876 , .
Mean temperature C. Passages direction .
* Mean 6 Time of middle negative passage:\mdash ; From observations From \mdash ; , Mean In \mdash ; direction .
*I have here altered the Kew notation to suit the letters commonly used in the C.G.S. system .
VERDBD'S CONSTANT IN ABSOLUTE UNITS .
Set 1 ( B ) .
Passages direction .
In \mdash ; direction .
Mean 6 Time of middle negative passage:\mdash ; From observations From \mdash ; , , 6 Mean 6 Interval between mean negative passages A and sec. If we divide by the approximate period of a half-vibration , the nearest even whole number to the quotient will be the number of passages in the time .
Dividing by we obtain the quotient , which gives 276 half-vibrations .
Dividing sec. by we obtain for the time of one half-vibration seconds .
Set 2 ( A ) .
Apri125 , .
Mean temperature C. Passages direction .
In \mdash ; direction .
Mean 9 5512:0416 9 Time of middle negative passage:\mdash ; From observations From \mdash ; Mean Set 2 ( B ) .
Passages direction .
Mean .
In \mdash ; direction .
Time of middle negative passage:\mdash ; From observations ) From \mdash ; , , Mean Interval between mean negative passages A and sec. Number of passages , that is 260 .
Time sec. Set 3 ( A ) .
Apri128 , .
Mean temperature C. Time of middle negative passage:\mdash ; Flom observations \ldquo ; Mean Passages in direction .
Mean . .
10 S6 In \mdash ; direction .
9222S Interpolated .
CONSTANT ABSOLUTE UNITS .
Set 3 ( B ) .
Time of middle negative passage:\mdash ; From observations From \mdash ; , Mean Passages in direction .
Mean .
10 In \mdash ; direction .
Interval between mean negative passages A and sec. Number of passages Time sec. Set 4 ( A ) .
Apri128 , .
Mean temperature C. Passages direction .
5S Mean 2 Time of middle negative passage:\mdash ; From obselvations From \mdash ; Mean Time of middle passage:\mdash ; From obselvations From \mdash ; Mean Time of middle passage:\mdash ; From obselvations From \mdash ; Mean Time of middle passage:\mdash ; From obselvations From \mdash ; Mean Time of middle passage:\mdash ; From obselvations From \mdash ; Mean Time of middle passage:\mdash ; From obselvations From \mdash ; Mean Time of middle passage:\mdash ; From obselvations From \mdash ; Mean Time of middle passage:\mdash ; From obselvations From \mdash ; Mean Time of middle passage:\mdash ; From obselvations From \mdash ; Mean Time of middle passage:\mdash ; From obselvations From \mdash ; Mean Time of middle passage:\mdash ; From obselvations From \mdash ; Mean Time of middle passage:\mdash ; From obselvations From \mdash ; Mean Time of middle passage:\mdash ; From obselvations From \mdash ; Mean Time of middle passage:\mdash ; From obselvations From \mdash ; Mean Time of middle passage:\mdash ; From obselvations From \mdash ; Mean Time of middle passage:\mdash ; From obselvations From \mdash ; Mean Time of middle passage:\mdash ; From obselvations From \mdash ; Mean Time of middle passage:\mdash ; From obselvations From \mdash ; Mean Time of middle passage:\mdash ; From obselvations From \mdash ; Mean Time of middle passage:\mdash ; From obselvations From \mdash ; Mean Time of middle passage:\mdash ; From obselvations From \mdash ; Mean Time of middle passage:\mdash ; From obselvations From \mdash ; Mean Time of middle passage:\mdash ; From obselvations From \mdash ; Mean Time of middle passage:\mdash ; From obselvations From \mdash ; Mean Time of middle passage:\mdash ; From obselvations From \mdash ; Mean Time of middle passage:\mdash ; From obselvations From \mdash ; Mean Time of middle passage:\mdash ; From obselvations From \mdash ; Mean Time of middle passage:\mdash ; From obselvations From \mdash ; Mean Time of middle passage:\mdash ; From obselvations From \mdash ; Mean Time of middle passage:\mdash ; From obselvations From \mdash ; Mean Time of middle passage:\mdash ; From obselvations From \mdash ; Mean Time of middle passage:\mdash ; From obselvations From \mdash ; Mean Time of middle passage:\mdash ; From obselvations From \mdash ; Mean Time of middle passage:\mdash ; From obselvations From \mdash ; Mean Time of middle passage:\mdash ; From obselvations From \mdash ; Mean Time of middle passage:\mdash ; From obselvations From \mdash ; Mean Time of middle passage:\mdash ; From obselvations From \mdash ; Mean Time of middle passage:\mdash ; From obselvations From \mdash ; Mean Time of middle passage:\mdash ; From obselvations From \mdash ; Mean Time of middle passage:\mdash ; From obselvations From \mdash ; Mean Time of middle passage:\mdash ; From obselvations From \mdash ; Mean Time of middle passage:\mdash ; From obselvations From \mdash ; Mean Time of middle passage:\mdash ; From obselvations From \mdash ; Mean Time of middle passage:\mdash ; From obselvations From \mdash ; Mean In \mdash ; direction .
IDCCCLXXVII .
Set 4 ( B ) .
Time of middle negative passage:\mdash ; From observations From \mdash ; Mean Passages direction .
Mean In directioIL Interval between mean negative A and sec. Number of passages Time .
sec. Set 5 ( A ) .
April 28 , .
Mean temI Passages direction .
Mean 5 Time of middle negatiye passage:\mdash ; From observations From Mean Time of middle negatiye passage:\mdash ; From observations From Mean Time of middle negatiye passage:\mdash ; From observations From Mean Time of middle negatiye passage:\mdash ; From observations From Mean Time of middle negatiye passage:\mdash ; From observations From Mean Time of middle negatiye passage:\mdash ; From observations From Mean Time of middle negatiye passage:\mdash ; From observations From Mean Time of middle negatiye passage:\mdash ; From observations From Mean Time of middle negatiye passage:\mdash ; From observations From Mean Time of middle negatiye passage:\mdash ; From observations From Mean Time of middle negatiye passage:\mdash ; From observations From Mean Time of middle negatiye passage:\mdash ; From observations From Mean Time of middle negatiye passage:\mdash ; From observations From Mean Time of middle negatiye passage:\mdash ; From observations From Mean Time of middle negatiye passage:\mdash ; From observations From Mean Time of middle negatiye passage:\mdash ; From observations From Mean Time of middle negatiye passage:\mdash ; From observations From Mean Time of middle negatiye passage:\mdash ; From observations From Mean Time of middle negatiye passage:\mdash ; From observations From Mean Time of middle negatiye passage:\mdash ; From observations From Mean Time of middle negatiye passage:\mdash ; From observations From Mean Time of middle negatiye passage:\mdash ; From observations From Mean Time of middle negatiye passage:\mdash ; From observations From Mean Time of middle negatiye passage:\mdash ; From observations From Mean Time of middle negatiye passage:\mdash ; From observations From Mean Time of middle negatiye passage:\mdash ; From observations From Mean Time of middle negatiye passage:\mdash ; From observations From Mean Time of middle negatiye passage:\mdash ; From observations From Mean Time of middle negatiye passage:\mdash ; From observations From Mean Time of middle negatiye passage:\mdash ; From observations From Mean Time of middle negatiye passage:\mdash ; From observations From Mean Time of middle negatiye passage:\mdash ; From observations From Mean Time of middle negatiye passage:\mdash ; From observations From Mean Time of middle negatiye passage:\mdash ; From observations From Mean Time of middle negatiye passage:\mdash ; From observations From Mean Time of middle negatiye passage:\mdash ; From observations From Mean Time of middle negatiye passage:\mdash ; From observations From Mean Time of middle negatiye passage:\mdash ; From observations From Mean Time of middle negatiye passage:\mdash ; From observations From Mean Time of middle negatiye passage:\mdash ; From observations From Mean Time of middle negatiye passage:\mdash ; From observations From Mean Time of middle negatiye passage:\mdash ; From observations From Mean Time of middle negatiye passage:\mdash ; From observations From Mean Time of middle negatiye passage:\mdash ; From observations From Mean VERDEI'S CONSTANT IN UNITS .
Set 5 ( B ) .
Passages direction .
5@ 34 Mean 558 Time of middle negative passage:\mdash ; From observations From \mdash ; \ldquo ; Mean In \mdash ; direction .
Interval between mean negative passages A and sec. Number of passages 178 Time sec. CORRECIIONS TO yIBRATION TIME .
Temperature Correction .
The Kew formula was used , and as the coefficients of the magnet had been determined at Kew for the Fahr. scale , that scale was used in the work .
The square of the time of vibration varies inversely as the netic moment .
Hence if the temperature of the magnet be greater than the standald temperature , the observed value of the vibration time will be too great , and its square must be divided by , where is the correction given by the Kew formula .
The temperatures at which the ation observations were taken WCle : ObsK l Temp. F 2 . . . .
S. . . .
4 . . .
5 . . .
G. . .
At Kew .
At Pixholme .
No. Obs. Temp. .
3 . . . .
4 . . . .
5 . . . .
Let the adopted standard temperature be that of , which , being the lowest in either set , will make all the terms positive .
The following are then the values of Similarly:\mdash ; [ standard ] Torsion of the uspending Thread .
We require the ratio of the force of torsion to the netic directive force .
The Kew formula is ( changing the notation to suit the letters commonly used in the C.G.S. system ) where is the force of torsion and is the change of declination produced by a twist of in the torsion-thread .
As the magnet was not suspended from a torsion-circle , the only twist that could be applied was given by twisting the magnet completely round in a direction .
If , however , which for so small a correction as this is we may do , we assume that the change of declination produced by a twist of is 4 times that produced by , the formula still holds , only we must substitute 36 for in the denominator ; we have then , ( 28 ) where is now the torsion produced by of twist .
The following observations were made , the first by reading the scale when the magnet was at rest , the others by taking the limits of swing after stopping the magnet as nearly as possible by means of a magnetized penknife:\mdash ; VERDEI'S CONSTANT IN ABSOLUTE UNITS .
Mean zero from ( 1 ) and ( 3 ) .
From ( 2 ) this gives that 36 of twist causes a deflection corresponding to centim. on the scale .
Mean zero from ( 3 ) and ( 5 ) , which from ( 4 ) gives a corresponding deflection of centim. Mean deflection therefore centim. To obtain the number of minutes corresponding to this , we have the formula of ( 26 ) , p. 16 , where centims. , and centim. This gives which from the tables ives the angle in minutes by which the declination is changed by a twist of , viz. ; taking of this for the torsion produced by of twist we have , from the Kew tables , corresponds to Clock Rate .
The Kew observations are corrected for clock rate .
Owing to an accident to the comparing chronometer , the rate of the Pixholme clock was not determined till two or three days after the conclusion of the vibration experiments . .
J. E. H. GOBDON ON Tm DETERMINATION OF The Kew formula of correction is Real time of vibration obser .
( 29 ) where is the daily rate in seconds , being when clock gains , \mdash ; when it loses .
On Apri129 , at .
4 .
the clock was 7 minutes 5 seconds slo ; on May 3 , at 4 the clock was 9 minutes 30 seconds slow .
Giving a rate of ; viz. it lost seconds per day . !
By interpolation from the Kew Tables we have To obtain the times of vibration at we have vibration time at ( observed time ) .
( 30 ) Also the torsion causes an apparent increase of magnetic force , and therefore decreases the time of vibration , whence ( vibration time corrected for torsion)2 ( observed time ) .
( 31 ) Also ( vibration time corrected for clock error ) ( observed time ) .
( 32 ) Combining these three formulae , we have for the true time of vibration corrected for temperature , torsion , and clock\mdash ; ( true vibration time)2 ( observed time ) . . .
( 33 ) We have the following values for the corrected vibration times:\mdash ; At Kew .
* These observations were taken by the Dorking clock-maker .
VBBDET'8 CONSTANT IN ABSOLUIE At Pixholme .
From these data we get the ratio of the magnetic force at Pixholme to that at Kew , for at Pixholme ( at Kew at time of Kew vibrations ) Vibration time at at Kew ( at Kew at time of Pixholme vibrations ) .
( Vibration time at Pixho and the values of at Pixholme at the times of the optical experiments equal the values at at those times multiplied by the above ratio .
The last columns of the two following Tables respectively the numerator and denominator of the above fraction .
* Rejected .
This gives at Pixholme at Kew .
But at the dates of the optical experiments had , at Kew , the values given in the second column of the following Table , and therefore at Pixholme those in the third .
Dividing the three values of given on page 20 by these numbers respectively , we have , for the three quantities which ought to be identical , ( 2 ) ( 3 ) Mean .
Extreme difference from the mean per cent. We have , finally , then , if be the rotation in bisulphide of carbon of the plane of polarization of the ray whose wave-length is *Rejected .
VERDET'S CONSTANI IN ABSOLUPE UNITS .
between two points whose magnetic potentials differ by unity , , ( 35 ) where is the number of units of circular measure in a minute of arc ; or .
As the constant is the ratio of a number to a current , its dimensions are This is VERDET'S constant in absolute measure .
For light of a given wave-length passing through a given substance , it is a fixed and definite physical quantity , depending only on the units of length , mass , and time* .
It is the number in optical measure which is equivalent to unity in electro-magnetic measure .
In future investigations quantities expressed in electro-magnetic measure can be expressed in optical measure by multiplying them by this number .
, then , is defined to be VERDET'S constant for the thallium ray in bisulphide of calbon , expresS.o in C.G.S. measure .
I here insert , as I have been requested to do , the result of my former paper on the same subject .
It is that for distilled water with white light , I do not , however , attach much value to the result , as the different determinations are in the ratios of showing variations from the mean of per cent. , or only about of the accuracy of the present paper .
The methods used for determining the constants were also susceptible of less accuracy .
Before concluding this paper , I must express my thanks , first and chiefly , to Professor MAXWELL , who has superintended every detail of the work for the year and eight months * The magnetic rotative power of bisulphide of carbon here comes in the same way as the specific of water comes into equivalent .
By an error in arithmetic this was printed in the abstract of my paper published Proc. .
Soc. June 1875 .
IIDCCCLXXVII .
during which it has been in progress ; and secondly , to Mr. WHIPPLE , Director of the Kew Magnetic Observatory , who sent me all the data for calculating the value of , and had several series of experiments and calculations made for me at the Observatory .
APPENDIX .
Analysis of the Bisulphide of Carbon employed.\mdash ; Mr. J. M. THOMSON , Demonstrator of Chemistry at King 's College , London , had the kindness to test the purity of the bisulphide for me .
He writes:\mdash ; " " As far as I can see , the you sent me is perfectly good .
Its .
( the mean of three determinations ) is at , the .
of pure being at .
Its boiling point is rather above , so that the sample may be considered as very nearly pure .
It leaves a trace of sulphur on evaporation , but not a quantity that can be weighed During the year which elapsed between the experiments and the testing of the bisulphide the latter was kept in a stoppered bottle of thick blue glass .
The temperature of the bisulphide , at the time of the optical experiments , was unfortunately not taken ; but I believe it was about F. , as it was a frosty night , while a good fire kept the laboratory pleasantly warm .
The logarithms used in the calculations are ffom Mathematical Tables , Ed. 1873 .
OOrb .
Phil. .
1877 .
Pla\amp ;

