]\gt ; INDEX SLIP .
, and rrrx .
Researches on the Electric Properties of Pure ubetances .
No. I.\mdash ; The Electrical Properties of Pure Sulphur .
PhiI .
frans .
A 1896 , pp. BBnABLBY , } HnrRY DRAPILR , and Researcbes on Electric Properties of Pure Substances .
No. I. \mdash ; The Electrical Properties of Sulphur .
Phil. Trans. A 1896 , pp. Annealing , effect of , on electrical conductivity of sulphur filnls .
Threlfall and Brearley .
Phil. Trans. A , p. 136 .
Galvanometer , construction and management of a seositive resistance .
Threlfall and Br.earley .
Phil. rans .
A 1896 , p. 80 .
's law not applicable to conduction in solid sulphur .
Threlfall and earley .
Phil. Trans. A 1896 , pp. 124 , 129 .
Residual charge of ]lighly pnrified sulphur .
IhrelfaU and Brearley .
Phil. A 1896 , 1 ) .
112 .
ecific inductive capacity of highly purified sulphur .
Threlfall and Brearley .
Phil. Trans. A 1896 , p. , 76 , 97 , resistances , methods of determining Threlfall and Brearley .
Phil. Trams .
A 1896 , p. 74 .
Specific resistance of highly purified sulphur ; behaviour of of modifications .
Threlfall and Brearley .
Phil. A 1896 , p. 97 .
Sulphur , a new monoclinic variety of ( ' ' aged\ldquo ; monoclinic ) .
Threlfall and Brearley .
Phil. A 1896 , p. 71 .
Sulphur , electrical properties specific capacity ) of solid highly , and of mixtures of modifications of .
Threlfall and Brearley .
Phil. Trans. A 1890 , pp. 72 , , 149 .
Sulphur , general physical properties of highly purifieA .
and rley .
Phil. Tr A 1896 , p. 69 .
, purification and tssting of .
Threlfnll and Brearley .
Phil. A 1896 , p. 59 .
Sulphur , .
Phil. Traus .
A ) , 129 .
57 on the Electric Properties of Substances.\mdash ; No. I. The of Pure RICHARD THRELFALL , M.A. , Professor of Phy.sics in the Univer.of Sydney , and JOSEPH HENRY DRAPIER BREARLEY , Deas-Thornson .
in the University of Sydney .
by Professor J. J. , F.It .
S. Received April 19 , \mdash ; Read May 24 , 1894 .
PRELIMINARY bTATEMEhT BY PROFLSSOR R. THRELFALL .
IN 1886 I began to make some experiments on the specific resistance of certain Australian gums , with a view to finding new insulating matel.ial .
In connection this I constructed anci tested varlous galvanometers .
In 1887 I was by Mr. J. A. POLLOCK , and in 1889 we published conjointly in the 'Philosophical Magazine ' a paper on the " " Specific Resistance of Imperfectly Purified Sulphur and on the " " CLARK as a source of Small Constant , Currents I also published in the same volume a paper " " On the Measurements of Higlr Specific Resistances All these papers may be considered as nary to the present in on of the behaviour of pure sulphur under electric stress , which has gone on steadily since June , 1889 .
The remainder of that year , the whole of 1890 , and most of 1891 , were occupied in preliminary experiments on galvanometers and in making and testing various arrangements of electrodes between which the phur to be examined was placed .
This work was carrled out rely by Mr. POLLOCK .
Early I was j.oined in the work by Mr. J. H. D. BREARLEY , who worked continuously at the subject with me up to the present time .
I began the study of sulphur originally with the object of discovering the exact electrical properties of a pure , non-metallic substance , and chose sulphur because it appeared to offer ges in melting at a moderate temperature , in be.ing capable of existing in several allotropic forms , and , above all , in being capable of being brought to a high degree of purity with tively little trouble .
With respect to the galvanometer , which has enabled us to perform the experiments described , so far as its general features are concerned , it was designed by me , in 1890 , on the laid down in the ' Philosophica ] Magazine , ' , vol. 28 , p. 473 ; but the detailed design , as well as the actual construction , I owe to , who made the insbrument in the Laboratory workshop .
Most of the work in connection with mounting the moving parts and in installing the instrument was done by Mr. BREARLEY .
A great , if not the cnief , difficulty , in attempting to form a theory of electrochemistry lies in the uncertainty surrounding the data .
The memorable researches of KOHLRAUSCH on the conductivity of water , as well as those of DIXON on the effect of a small impurity on the of combination of gases , together with the quite allied .and equally ) variations produced in thermoelectric power and contact force minute differences of chemical composition , have had a destructive effect the theories of twenty years .
It is not too much to say that the electrical action of mosb bodies in a pure state is entirely unknown at present , although the fact of the continuity of the electrical properties of certain substances\mdash ; e.g. mercury , gold , and copper\mdash ; as they are more and more purified , raises a hope that the properties of substances as at present investigated may only differ quantitatively from their properties when pure ; but whether this be so or not can only be decided by experiment .
The question as to the effect of purification is one which must be discussed in each separate case , and probably over and over again , for the word " " pure\ldquo ; has no significance except with respect to a definite state of the art of chemistry .
It is to be noted that the final stages in a process of purification must always depend more or less upon hypothesis\mdash ; for a point is soon reached where the npurities become too small in amount to menable to ordinary chemical examination .
Theoretically this can bs got over by operating on largel quantities of substances , but in practice such is not found to be the case , for the greater the bulk of substance to be prepared , the more ti me is required for its and the greater is the chance of small and unnoticed accidents occurring .
In the case of sulphur this is particularly well marked , for the chief source of impurity is in dust from the air\mdash ; which leads to a contamination most proportional to the time required in making the preparations .
We look forward to the time when the exclusion of dust during operations of precision will be regarded as of as much importance as its exclusion in bacterioiogical research .
The sulphur we have used is probably as pure as it can be got without taking ) scial precautions to avoid all contact with dusty air .
The absence of such precautions vitiates to a great extent .
observations of discharge in air , and led QUINCKE to explain away the diminution of the resistance of benzene and other liquids of small conductivity with increased voltage per unit length : In the case we have studied , the physical action of dust particles is not to be feared , and appropriate means were adopted for minimising the possible chemical action bebween dust and sulphur .
One of the results we have attained with sulphur in a certain condition\mdash ; is a , " " Recberches QUINCKE , " " Wied .
Ann 1886 .
vol. 28 , p. ON THE ELECTRICAL PROPLRTIES OF PURE SULPHUR .
similar diminution of resistance , and in ihis case explanation is impossible .
The suggestion is obvious that the phenomena observed by QUINCKE may possibly be real phenomena\mdash ; not depending on dust particles\mdash ; and if this be the case , then it is clear that a valuable discovery has been rendered unavailable by the source of uncertaint .
we are dealing with .
In the following statement we shall explain with some minuteness the methods adopted for purifying our material , and give a summary of the conclusions to which previous observers have been led as to its physical properties .
We shall then explain in a general way the methods we have used for determining the electrical properties .
This be followed by a short treatise on the construction of sensitive galvanometers of .
high resistance , ogether with an account of the experiments which have led us to the uction of our instrument , of which full details and drawings will be shed .
We shall then proceed to deal with our various experiments .
At present time experimental work is still in full progress , and is likely to continue , for in investiga ing properties of substances the end is only reached when the observer 's interest falls off .
Our chief reason for publishing is that we consider the main features are now pretty well established .
SECTION I. On the and Properties of Pure Sulphur .
S1 .
General Discussion of Causes , Procedure , .\mdash ; Our information on the transformations of sulphur is chiefly derived from the following authors:\mdash ; 1856 .
C. -CLAIRE DEVILLE , 'Ann .
de Chim .
et de ' Ser. vol. I ) .
94 .
1854 .
MAGNUS , 'Po 1854 , vol. 91 , p. 308 .
1854 .
B. C. BRODIE , ' Proc. Roy .
Soc. , ' vol. 7 , p. 24 ( March 30th , 1854 ) .
1856 .
MITSCHERLICH , ' Jour .
p.rakt .
Chem vol. 67 , p. 369 .
1857 .
BERTHELOT , Ann. de Chim .
et de Phys. ' ( 3 ) , vol. 49 ( two papers ) .
O. , ' Zeitschr .
Phys. Chem vol. 8 , p. 608 .
1884 .
, 'Chem .
entr .
Blatt , ' 1884 , p. 450 .
1889 .
J. MONCKMAN , 'Proc .
Roy .
Soc. ' for 1889 , vol. 46 , p. 136 . .
D. GERNEZ , ' C.R. , ' voJs .
, p. 219 , 82 , , p. 217 .
There are a large number of other observers whose work does not bear so precisely on the subject of the present investigation , and whose names are recorded in the " " Sachregister zu den Annalen der Physik und Chemie published in 1888 , p. 563 .
Amongst the authors mentiOned , we have found the papers of BRODIE and BERTIIELOT of most service. .
BRODIE gives an account of the transformations of sulphur , but unfortunately omits many details , particular]y as to the mode of SSOR R. THRELFALL AND .
J. H. D. BREARLEY purification adopted .
BERTHELOT attempts to classify the varieties of sulphur on electro-chemical principles .
According to him , sulphur exists in two principal modifications , one , electro-positive sulphur , liberated at the positive-electrode during the electrolysis of chloride of .
the other electro-negative sulphur liberated at the negative electrode during the electrolysis of metallic sulphides .
The itive sulphur comprises all hose varieties which are amorphous ; the electro-negative all those which are crystalline .
The distinction between crystalline and amorphous sulphur appears to us to be much more fundamental than between those varieties which are soluble and those which are insoluble in carbon bisulphide .
Indeed , as is well , mere crystallization from carbon bisulphide often involves the formation of insoluble sulphur .
The distinction between the main varieties is considerable from the dynamical point of view\mdash ; for the thermal changes involved in passing from one to the other are very great\mdash ; as was shown by BRODIE in a simple experiment and by PETERSEN in an elaborate thermo-chemical research .
We have chiefly examined those varieties of sulphur which are sufficiently definite and permanent to us to be sure that we are dealing with pure substances .
Thus we have examined the variety of sulphur pretty thoroughly , together with what we believe we are the first to recognize as a new variety of monoclinic sulphur , as well as mixtures of this variety with amorphous sulphur obtained by cooling idly from a high temperature .
The permanence of monoclinic sulphur we find to be much greater than is generally supposed , at least , if the melting point may be taken as a criterion , *while as for octahedral sulphur , the only possible means at our disposal for obtaining it pure\mdash ; viz. , by the method of hitherto failed in our hands .
This method is based on the obtaining of sulphur in a state of " " surfusion The sulphur , which is perfecbly crystalline to begin with , is melted below 13 C. , cooled to , say , C. , and then caused to crystaIlize in the octahedral form by sprink-ling the sulphur with a trace of the dust of octahedral crystals .
This appears be a matter of some delicacy , and in the course of a good many trials we have not succeeded to our satisfaction , even when following M. GERNEZ'S instructions most minutely .
The fact is , that melted sulphur has a very strong tendency to crystallize in the monoclinic form , and this teudency can only be overcome under the most favourable circumstances ( if at all ) and by the exercise of great care .
For most electrical experiments , however , it is necessary to obtain a film of the material between two conducting plates which serve as electrodes , and the conditions attending the production of such ffims are experimentally unfavourable to forced octahedral crystallization .
Again , both FOUSSEREAU ( ' C.R. ' 97 , p. 996 ) and BOLTZMANN , RONICH and FAJDIGA , and NOWAK ( ' sber .
der Wiener Akad vols .
68 and 70 ) found no appreciable conductivity in crystals of octahedral sulphur , though their methods were not very searching .
* The phenomena attending the transformation of monoclinic into octahedral sulphur win be dealt with during the discussion of cases .
ON THE ELECTRICAL PROPERTIES OF PURE SULPHUR .
SECTION II .
The Preparation Testing of of Sulphur : The earliest note we have found on the influence of minute traces of impurity on the properties of sulphur is by MITSCHERLICH in the ' Journal Praktische Chemie ' for 1856 , vol. 67 , p. 369 .
This chemist states that the discoloration of ] phur which been heated to the boiling-point is to be traced to the presence of organic compounds .
One three-thousandth part of " " grease\ldquo ; will produce a very deep red coloration in sulphur which has been heated to the boiling-point , and even the greasiness produced by handIing is said to notably affect the colour of sulphur .
The strongly coloured organic compound does not come over when the sulphur is distilled , and its presence may be detected on the bottom of the ret , ort , where it forms black spots .
MITSCHERLICH considers the supposed coloured varieties of sulphur discovered by MAGNUS ( ' Ann. de Chim .
et de Phy@ [ .3 ] , vo ] .
47 , p. 194 ) to be merely sulphur contaminated by organic matter ; he gives , however , no of his experiments .
MONCXIIAN ( ' Proc. Roy .
Soc 1889 , vol. 46 , p. 136 ) the black spots which are left in a retort from which sulphur has been distilled , considers they are produced oniy when the retort is heated by the bare flame\mdash ; a process leading , of course , to the heating of the glass much above the boiling-point of the sulphur .
BERTHELOT .
cit. is inclined to with MITSCHERLICH as to the influence of small organic impurities on the colour of sulphur .
STAS ( ' Recherches.\mdash ; Bull .
] ' Acad Royale de Belgique , ' 2eme serie , vol. 10 , p. 253 ) made use of the apour of ' ' pure\ldquo ; sulphur , but gives no details as to how it was obtained .
We know of no discussion as to the minute purification of sulphur except the one contained in MONCKMAN'S paper\mdash ; in which he refers to a method proposed by one of us , and carries out that method and several others in a careful manner .
We must therefore consider the purification of sulphur from the commencement , since there is no previous work to help us\mdash ; we have , in faot , to solve the problem of purif.ying sulphur and of nizing it in the state of purity .
Our experiments lead to the following result .
sulphur can be recognized chiefly by its negative istics .
It evaporates from a platinum basin without residue , it has little or no smell , it has a clear yellow lour even when suddenly cooled from above 18 C. by water ; it gives off no gas when heated vacuo , it leaves no black spots when distilled from a retort with or without the bare.flame ; it has a specific resistance of above C.G.S. units when crystallized in inic p ; it exhibits , to all intents and purposes , perfect freedom from " " electric absorption\ldquo ; when in this state , and it is perfectly soluble in carbon hide , forming a solution which is absolutely clear .
We will now give a short account of how these characteristics were determined .
The question of the colour of sulphur occupied one of us previous to 1888 , and in that , and as a result of a good many trials , it was found that when sulphur is distilJed over and over again in glass retorts , black spots always make their appearance in the residue unless the operation is carefully shielded from the action of dusty air .
If sulphur is distilled and then exposed to air so that dust can settle on it , then , on repeating the distillation , black spots make their reappearance .
We have only once succeeded in absolutely preventing the presence of black spots in a retort from which sulphur was distilled , but we have noted over and over again that just in proportion to the precautions that are taken to exclude dust , so do the black spots become smaller and fewer , and the colour of sulphur suddenly cooled from a high temperature clearer and clearer .
That pure sulphur has no smell we know , because we have prepared a sample by precipitation of hyposulphite , and five distillations smell could be detected from it .
This observation is important , for we have large quantities of sulphur twice distilled from Sicilian roll sulphur which le.aves no residue when burned away in a platinum dish , which remains bright yellow on cooling quickly , which will yield no ' gas when heated in up to near its boiIing-point , and which generally might be considered to be had the well-known sulphury smell , which , by the way , .
we consider to be possibly due to traces of chloride of sulphur .
The smell is to be removed by exposure in vacuo at any temperature and during many hours , and samples of sulphur which possess it had better be left on one side .
sulphur , of course , leaves no residue on burning in a platinum basin , but we have never succeeded in reaching this stage with sulphur which is burned in the ordinary way .
We have , however , got as far as this\mdash ; that the combustion of . .
200 grams of ] phur in a ] atinum dish will leave a stain so small that it cannot be observed unless the platinum is ightly polished .
stain disappears at a red heat , and may be safely attributed to the dust which gets into the dish during the time the sulphur is being introduced , and while the process of burning is being observed .
In order to obtain the maximum residue from any sample of sulphur , the last cubic centinletre or so must be boiled off very carefully and without overheating .
The residue which , with our purest sulphur , such as we have used in electrical experiments\mdash ; is after all .
mere stain , is far beyond the powers of the balance to detect .
It is dark in colour and burns off at a red heat , indicating that it is composed of the same organic compound which was discovered by MITSCHERLICH .
We have observed that after a sample of sulphur has been exposed to the air during electrical experiments , it will give a comparatively large residue , although when first , prepared the residue could only be observed by close attention .
Those of sulphur which when suddenly cooled appeared of a buff colour , always gave comparatively large residues , and these were the samples which had been most exposed to the action of air .
Of course ordinary sulphur leaves alarge quantity of earthy residue besides , which caIl be got rid of by about two \mdash ; but we are not discussing such preliminary ncations .
Impurities , which can bedetected by means above described may be considered to be the more obvious rities ; not give rise ON THE ELECTRICAL PROPERTIES OF PURE SULPHUR .
to any great concern .
Above this point , unhappi]y , we are to proceed by way of hypothesis , and so conduct the process of purification as to tend to eliminate impurities , which our general knowledge of the chemistry of sulphur leads us to suppose may be in quantities too small for detection by chemical means .
Only those impurities are to be feared which will distil with sulphur , or be mechanicaly c:.arried over during distillation , and dissolve in carbon bisulphide .
practically narrows the investigation down to the consideration of selenium as an impurity .
In order.to cause .
any residual impurity to vary in different samples , we have tried a large number .
of different sources for our sulphur .
Of these we retain .
The first is by decomposing commelcial sodium-thiosulphate by pure ydrochloric acid , washing the deposited sulphur with platinum distilled water , and distilling till the on combustion is inappreciable .
This method has the advantage of involving no contact with carbon compounds , and of giving rise to no impurities except those that are soluble in water , or left on distillation .
As to the impurities which may be euspected to be present in the thiosulphate originally , these will be dealt with when we consider soda waste as a source of sulphur or thiosulphate .
The second kind of sulphur which we have employed is that derived from the polysulphides o.f calcium .
Some rolI sulphur , in powder , was digested with marble-lime in the usual manner .
The yellow liquor obtained was decomposed by pure chloric acid , and the washed and distilled .
The third kind of sulphur is some recovered sulphur that ca1ne into our hands by the idness of Mr. F. WRIGHT , of ELLTOTT BROS .
, importers of drugs in this city .
We could get no information at all as to where it came from , or what the process of recovery was ; but we were told that the sample ( amountiug to about ) lbs. ) was recovered and very pure .
A very few trials showed us that , as far as the negative characteristics , the claim of purity was well founded , for the sulphur was at least at good as our precipitated sulphur after about three distillations .
To find out where it came from , with a to investigating its history , we wrote to Mr. J. F. CHANCE who was known to us as an authority on recovered sulphur ) askin him whether he knew of any ordin.ary process by which sulphur of such a high degree of purity was produced commercially , and whether he could obtain a fresh supply for us .
In reply ffi .
CHANCE gave us a general account of the method of sulphur recovery invented by himself , and most kindly presented us with a large quantity of sulphur recovered by his process .
On comparing this sulphur with that which we had received from Mr. WRIGHT , it became obvious at once that the latter had been produced by the Chance process : the appearance alone is almost unmistakable .
Aocording to Mr. CHANCE , the moisture and non-volatile impurities amount together to not more than per cent.\mdash ; and the non-volatile impurities to perhaps only per with .
regard to the latter number we can say that our experience roughly confirms this estimate , for we never obtained a weighable residue even from the evaporation of several ounces of the fused and filtered sulphur .
Traces of selenium and arsenic , however , might occur as volatile impurity , it .
was therefore necessary for us to ascertain their presence ( or absence ) by means of a direct'experiment .
After a considerable of preliminary work we fell back on the combustion of quantities of about twelve grammes of the sulphur to be tested , in a platinum boat in a combustion tube .
The products of combustion being mixed with excess of oxygen and passed over spongy platinum , the resulting trioxide was received in a large globe cooled by ice and salt .
The trioxide was then converted into sulphuric acid , and tested for arsenic and selenium .
We made two sucoessful combustions , using in one case the residue from a ] tion of Chance sulphur , and in the other the sulphur in the state ( twice distilled ) in which it was used in our experiments .
The combustion requires some precaution .
In our arrangements , the tube employed was of hard glass , and was heated for a length of 80 centims. , it was centim. in internal diameter .
In this tube 15 centims. were occupied by the platinum boat containing the sulphur , this was followed by a " " mixing\ldquo ; space of red-hot asbestos for 25 centims. , and 35 centims. were occupied by spongy platinum and platinized asbestos .
Oxygen was admitted by two tubes , one only just entered the combusiion and provided the oxygen for the burning of the sulphur in the boat ; the other ran over the boat and delivered a large excess of oxygen into the " " ulixing\ldquo ; space .
Oxygen was observed to be in excess during the whole process .
By regulating the oxygen supply and the temperature of the sulphur the boat , the combustion can be got to proceed arly .
The combustion tube was bent at the delivery end , and dipped into the receiver .
At the close of the operation , stream of oxygen was passed for some time , the tube being now red-hot throughout\mdash ; up to the bend\mdash ; by this means the whole of the contents of the boat were burned and all residues chased out of the combustion tube .
Finally the part of the tube which projected into the globe was washed out with platinum-distiIIed water and the solution added to the solution in the same water of the sulphur trioxide contained in the globe .
In the first combustion the oxygen supply was not in sufficient excess during a few moments and the trioxide was coloured blue by combination with a trace of free sulphur .
In the second combustion , in which the oxygen was in great excess throughout , the resulting sulphuric acid contained sulphurous acid , and the excess of oxygen also carried off sulphur dioxide during the process .
The nclusion seems to be that Sulphur Trioxide is dissociable by spongy platinum at a red heat , even in presence of oxygen\mdash ; this we did not know , it is not mentioned by LENOINB ( ' Etudes sir les Equilibres Chimiques The solution of trioxide from the first combustion required to be filtered ; the second did not , and was absolutely clear and colourless .
We remark on this , for there is a chance of selenium separating during the solution of the trioxide , in the presence of sulphurous acid .
In both cases the solution was next boiled with a large excess of sulphurous acid and hydrochlorio acid .
In neither case was there the ghtest precipitate .
We are satisfied that , under the conditions of the experiments , this means that there could not have been more than ON THE ELECTRICAL PROPERTIES OF PURE SULPHUR .
milIigram of selenic acid present , or say milligram of selenium , or about one part in a million .
As at first we under estimated the delicacy of the selenium test , it may be worth while to indicate how the test should be performed in order to give the best results .
The points are : ( 1 ) a large quantity of hydrochloric acid should be used ; ( 2 ) the liquid after some minutes ' boiling should be cooled before the sulphurous acid is added ; ) the mixture should then be well shaken ; ( 4 ) on heating to boiling , a precipitate will appear in two or three minutes if selenium is present\mdash ; .at the limit a yellowish rose-colour will be produced , but no actual precipitate .
The essential point is to have the volume of liquid as small as possible .
For instance , with one cubic centimetre of liquid 01 milligram of selenic acid may be detected with certainty , and still smaller quantities iu proportion as the volume of liquid is smaller .
After testing for selenium , and taking the most refined precautions as to re-agents , the MARSH'S test for arsenic was applied .
In no case was arsenic discovered .
We found by direct experiment that the minimum amount of arsenious oxide we could detect by our appliances was 02 milligram , and that by the smell of the gas only .
The test , however , as made by the nose is just as certain as that made by the eye when a mirror is produced ; the latter , however , is less delicate .
With respect to precautions , we will only say that the instructions of FRESENIUS were carefully fOlJowed , and that this was not the first occasion on which we had made the test .
* We may conclude that if arsenic is present in the sulphur , it is only so to an extent of about one part in a million .
The enium test will be observed to be more delicate than the arsenic test\mdash ; a fact not generally recognized\mdash ; and this when the test is made by noting the smell of the evolved hydrogen .
If a recognizable mirror be required .
the arsenic test is a good deal less delicate .
A test for tellurium and selenium was made for one of us in 1888 by Dr. HELMS , on a sample of sulphur recovered from thiosulphate .
These results were ative .
Aaregards arsenic and selenium , we wish to further , and to show that it is very obable that they could have been present even in quantities much less than we aould have detected .
The results of the enquiry will apply alike to sulphur recovered by th6 Chance process and to sulphm recovered in the laboratory from commercial sodium thiosulphate ( assuming the latter to have come from soda waste ) .
For this purpose it is necessary to consider the process in some deta-il .
The sutphuric acid employed in the salt-cake process is the source from which arsenic and possibly seleuium may be derived .
The greater part of the arsenic is got rid of in the salt-cake furnace in the form of chlorides .
With respect to selenium , we believe that it will be almost all precipitated in the sulphuric acid chamber , but as the acid is not filtered , as a rule , this cannot be regarded as a purification .
* We may note here that the decomposition of arsine by strong sulphuric acid is a test , though not quite so delioate as the mirror test , is not by any means a bad one .
It is not usually given as a test at all .
( The snlphuric acid must be rated .
) ORGE ELIOTT informs us that when burning some samples of sulphur from , containing about 1 per cent. of selenium , for the purpose of making acid , the selenium separated in the denitrating tower and in the chamber , in the nJanner we suggest .
In order to find selenium is got rid the -cake process , we burned a quantity of selenium in a properly bent tube , and passed products of combustion into water , burning the selenium part of the time in a brisk current of oxygen , and part of the time with only just enough oxygen to keep the selenium alight .
Part of the ulting selenious acid was converted into selenic acid , and a mixture of these aoids and selenium , together with any other of the products of combustion of selenium which might have been formed , were added to some sulphuric acid in an experiment imitating the salt-cake process .
We found that all the selenium remaining in the salt-cake did not amount to 1 per cent. of that mixed with the sulphuric acid at the commencement of the process , so that the salt-cake process gets rid of most of the selenium as well as the arsenic .
It is possible for some small trace of selenium to get into the black-ash furnace .
If it does , it is presumably in the condition of free selenium , sodium selenate , or selenite .
Now the trace of free selenium will probably be voIatilized to a great extent , and off with the other gaseous products .
As for the selenite and selenat an experiment to be described indicates that they will be decomposed , yielding selenium , with which some of the lime will form calcium selenide .
We find , however , calcium selenide is decomposed when suspended in water by carbonic acid gas yielding free selenium , but not a trace of selenium hydride .
These facts were established as follows : About 4 grams ofprecipitated chalk were converted into calcium selenate , and this was reduced by hydrogen ( pure ) to calcium selenide .
During the operation much of the selenide decomposed into free selenium and lime , and there was a considerable evolution of selenium dride .
The process was stopped while the evolution in full progress .
On washing the contents out of the tube we probably had a mixture of free selenium , calcium selenide , calcium selenate , and perhaps other products .
stream of carbon dioxide gas ( pure ) was paesed through this mixture , including water , contained in a set of potash bulbs , and as a result a large quantity of selenium was deposited , but not the slightest trace of seleniu hydride could be discovered either by the smell of the emergent gas or by heat.ing the delivery tube red-hot .
It will be noticed that the great instability of calcium selenide for we presume that the hydrogen selenide came from the reduction of this salt ) , when heated to redness , gives us another reason for suspecting that very little calcium selenide is formed in the black-ash rather that very little can survive , unless reduction *A subsequent note from Dr. ELLIOTT throws some doubt on this , for a red itate being noticod at the foot of the denitrating when sulphur noarly , if not quite , free from selenium was being burned as a source of sulphur \mdash ; an analysiq of the precipitate lybS m. It then turned out be ferric oxide .
ON THE ELECTRICAL PROPERTIES OF PURE SULPHUR .
by carbon and carbon gases differs very unexpectedly from reduction by hydrogen .
Any alkaline selenide that might be formed is unimportant , because , firstly , it would be soluble in water , and , secondly , it is not stable in solution but deposits selenium , fOrming , perhaps , some oxyacid of selenium .
If this be the case , carbonic acid gas might liberate an oxide of selenium , but this would not at ordinary temperatures .
It is possible , finally , that free selenium in the waste might gradually eombine with the excess of lime , if there be an excess , and re-form Qelenide of calcium .
This was tried by boiling fineIy powdered seleniura with excess of lime\mdash ; but no combination was produced during an hour , and operating on about half a gram of lelenium .
The case for the probable absence of selenium from Chance sulphur seems therefore to be pretty well made out , as well as for its absence from commercial thiosulphate of sodium .
Other Methods of We have purposely omitted to use any method based on the crystallization of sulphur from organic liquids or sulphur dichloride .
Our reasons are very simple\mdash ; anyone who has crystallized sulphur carbon disulphide must have noticed that on reducing such crystals to powder , even long after they have appeared perfectly dry , a strong and horrible smell , characterized by all the worst features of the odour of impure carbon bisulphide , becomes immediately perceptible .
Since powdering the orystals reduces them only to smaller crystalline fragments , there is no reason why volatile compounds should not continue to linger in them .
If this be the case , what are we to expect on distillin Surely it is gratuitous to assume that organic compounds could not be formed and distil with the sulphur .
Supposing these compounds to be without smeIl , how could they be detected ? .
Again , carbon bisulphide is a liquid almost impossible to purify , it begins to decompose immediately under the action of light , and forms unknown compounds .
Judging by the smell still clinging to carbo1l disulphide very carefully prepared by one of us , we may hazard a guess that some selenium compound still lingered .
As to toluene , turpentine , \amp ; c. , conidered as solvents of sulphur , we have never seriously entertained the idea of using any of them , for fear of introducing volatile carbon compounds\mdash ; and aJso because we have an idea , owing to the comparatively small solubility of suIphur in these substances and the necessarily large amount of the solvent , traces of impurity might possibly accumulate in the sulphur .
It stlrprises us that anyone conversant with , say , the curiously complex ontents of rock crystals , should rely on crystallization to temove impurity .
Worst of all possible solvents , in our opinion , is chloride of sulphur .
Chlorine is most difficult to purify , and the process of preparation of chloride of sulphur involves uncertainties to which exception may be .
Even supposing this solvent to be obtained in a pure state , there is no more difficult substance to remove from the sulphur* ( compare MONCKMAN , .
cit It is ether a most unsatisfactory body to deal with .
The Contents of Sulphur.\mdash ; On heating sulphur in vanuo , there is no doubt that gases are given off , even from the purest sampIes .
The following is an account of some experiments made on this subject in connection with what will later on be described as the " " Film of November 18th , 1892 It may be premised that this wats a film which had been rapidly cooled from a high temperature , after having been art of some sulphur which had caught fire in an atmosphere containing much carbonic acid gas , and which looked rather dirty .
The sulphur composing this film was found to give off a good deal of gas on heating ; so , on November , 1892 , a quantit of recovered sulphur , twice distilled , was selected for an experiment having for its object the determination of the conditions under which gas is given off and absorbed by sulphur .
The sulphur was melted under an air-pressure of about 5 .
centims. of mercury ; the results were\mdash ; Hardly any gas was given off the thermometer rose to 1 C. , and then the evolution of gas increased up to 19 C. At this point the increased visco of the sulphur prevented the bubbles coming off , but they were still being formed .
At about 23 C. the bubbles began to separate from the sulphur , and continued to do so up to the boiling-point .
There was nothing like a sudden evolution of at temperature .
The next day the sulphur was re-heated to C. under similar conditions , but not a single bubble of gas could be detected .
A question now arose as to the nature of the gas .
Is it air If so , shaking up the sulphur with air while it cools from the boiling-point might be considered to be likely to induce more air to get absorbed .
This was tried , but no bubbles were given off on re-heating er an air-pressure of 5 centims. of mercury only .
Is the as carbon dioxide ?
The last experiment was repeated in an atmosphere of carbon dioxide , but again no gas was driven off on re-heating .
Has the gas anything to do with the sulphur being on fire ?
, is the gas an oxide of dissolved from an atmosphere mainly composed of this oxide To test this , the sulphur was heated to boiling-point and set on fire in a platinum dish , a bell jar was placed over it , and the sulphur was agitated freely .
It was then re-trawferred to the vacuum-bath and tested as before , but little or no gas came off .
These experiments give no clue as to what the gas given off really is , but a }able result is that , when once the gas has been caused to come off , remelting win not cause more gas to be absorbed .
In view of the more important awaiting solution , we judged it unadvisable to spend any time in further determining the -nature of the gas , as we possibly might have done in a vacuum-tube by the help of spectroscopy .
We conclude , however , the purest distilled sulphur ought to be \mdash ; It can be done however by continued ction with carbon bisulphido\mdash ; at allevents fairly well ( see end of ON THE ELECTRICAL PROPERTIES OF PURE SULPHUR .
remelted in before being used\mdash ; but this we only began to do in 1893 , after the above facts were discovered .
We should like to hazard the mere guess that the gas is water vapour , and that it is t , aken up during the formation of the sulphur snow , which is always present when sulphur vapour condenses in a retort and receiver .
Physical Properties of Purified ur .
We have not any serious corrections to make to the observations cf previous workers\mdash ; in other words , very pure sulphur does not differ appreciably in ordinary characteristics from sulphur which is only moderately pure .
The most obvious physical datum of a fusible solid is the melting-point , and this is , in the case of sulphur , still rather a matter of uncertainty .
he most important contribution to our knowledge of this constant was supplied by B. C. BRODIB Proc. in the year 1854 .
The main results , as as melting-point is concerned , are as follows :\mdash ; 1 .
Pure ) octahedral sulphur tallized from carbon bisulphide melts at C. 2 .
But octahedral sulphur begins to .
into monoclinic sulphur below this temperature ; and 3 .
The melting-point of pure monoclinic sulphur is 12 C. 4 .
Mixtures of octahedral and monoclinic sulphur melb at temperatures between and 12 C. No are given as to how the Phur was purified , nor as to the method adopted for finding the melting-point .
We will add to this statement the owing account , which may be regarded as summarising a large number of experiments .
Method .
have employed the method of heating small particIes in thin .ss tubes alongside a thermometer .
We are aware ( though we cannot find reference ) that a recent observer who carefully examined the relative accuracy of different methods , decided that the best method to employ in determining meltingpoints is that of using a large quantity of the substance and plotting a cooling curve .
In the case of sulphur , however , this method is rather difficult for many reasons\mdash ; some of whid ] will be evident on referring to BRODIE'S paper\mdash ; and with any method the very small conductivity of sulphur constitutes a grave difficulty .
Our thermometers , which we had hoped to have standardized at Kew , were , by an oversight , only standardized up to C. , and to send them to Kew for the oversight to be remedied would have required too much time .
They were .
however , very good .
One was six years old at least , and its B.P. was too high by .
We have confidence that the observations are correct , as far as the thermometers go , within about.1o C. , within which limit the thermometers were in accordance at the temperatures dealt with , the boiling-point corrections being made .
Octahedral Sulphur.\mdash ; The observations were as follow for octahedral sulphur crystallized ( and well dried ) from carbon bisulphide:\mdash ; PROFESSOR R. THRELFALL AND MR. J. H. D. BREARLEiY Sulphur heated in a glycerine bath at rate of aboub per minute .
I.\mdash ; Melting-point of Octahedral Sulphur .
As BRODIE has shown , octahedral sulphur heated beyond about 10 C..begins to turn into monoclinic sulphur .
If the sulphur is in the state of fine powder , the change cannot be avoided even if the sulphur be heated " " for the shortest between 10 and As our sulphur was in small fragments , it was possible that the bove discrepancies could be attributed to a more or less partial change havin occurred .
In order to test this , the glycerine bath brought up to different temperatures near the melting-point before the sulphur was introduced .
The results are as follow:\mdash ; TABLE lI.\mdash ; Melting-point of pure Octahedral Sulphur .
Number oWent anceRemarks .
These observations were made by Miss FLORENCB MARfflN , a most careful observer , to whom our thanks are due .
She recommends that Experiment No. 8 should be discarded , on the ground of the point of commencement not having been s6 sharply marked .
It will be noticed that the above tables indicate two things : firstIy , that the whole the melting-point is lower when the sulphur.is raised to it suddenly ; and secondly , that all the melting-points are considerably above those given by BRODIE .
The reason probably is to be found either in tho..faot that BRODIE used imperfectly ON THE ELECTRICAL PROPERTIES OF PURE SULPHUR .
purified sulphur , or that there is a constant error in our thermometers , which we have no reason to believe .
The fact probably is that BRODIE did not sufficiently purify his crystals from the evil-smelling compounds which they car1y with them from the bisulphide .
Monodinic Sulphur.\mdash ; The melting-poinl of perfectly soluble monoclinic sulphur was taken by the same observer in connection with a film which had been experi- ented upon .
TABLE III.\mdash ; Melting-point of Aged Monoclinic Sulphur .
* This agrees as weII as can be expected with the observations of other observers on monoclinic sulphur .
The fact that the meiting-point of monoclinic sulphur , even when it has lost its crystallographic properties , is about , may be taken as established by these observations .
The want of sharpness in melting-point , which is to be noticed even in the case of monoclinic sulphur , is marked in all the samples we have and is on the whole more pronounced when we deal with mixtures , but not markedly so .
We think it is established that sulphur , like selenium , has not a perfectly , sharp melting-point\mdash ; another point of resemblance between these substances which we have not seen notice\amp ; Of course , the effect is much more marked in the case of selenium than in the case of sulphur ; but even with the latter the corners of fragments become less sharp and assume a greasy appearance at temperatures a good deal below those at which the fragment is definitely liquid .
The freezing points of sulphur have not been investigated by us , as they have been dealt with most elaborately by GERNEZ ( ' C.R. , ' .
cit One disadvantage of the method of publication adopted in those famous " " Comptes\ldquo ; is that it effectually prevents any estimate being formed of the reliability of experimental work , the scarcity of space not allowing sufficient details to be given .
The melting points of mixtures of mon.oclinic and electro-positive sulphur are *This sulphur was afterwards found by Professor DAVID to have lost its crystallo-optic properties , and could not be distinguished by the polariscope from octahedral sulphur .
We have ventnred to call this variety " " aged monoclim.c\ldquo ; .
It is probably generally formed as a preliminary step in Formation of octahedral sulphur monoclinic at ordinary temperatures , rather below , and will be dealt with when the films forming the subject of experiment are considered .
The detail of some other physical will be dealt with as occasion arises .
SECTION III .
Methods of Determining the Constants of Pure Sulphur .
The chief constants determined by us are the Specific Resistance and Specific Inductive Capacity .
Observations have also been made on the residual charge of suIphur regarued as a dielectric .
No observations on the conduction of electricity through liquid sulphur have as yet been made , this having been provisionally accomplished by previous workers .
On the other.hana ' , the phenomena attendin conduction through solid sulphur have been made out by us , at least , in their more important features .
It will be convenient here to give a sketch of the information obtained by previous observers .
Resistance of Sulphur .
: WIEDEMANN , ' Lehrbucl vol. 1 , p. 498 .
Sulphur is mentioned as a non-conductor .
BOLTZMANN , 'Wien .
Sitzb vol. 70 , p. 342 .
In this paper , which deals more particularly with dielectric constants , it is stated that sulphur is a perfect insulator , and is free from " " dielectrische Nachwirkung G. FOUSSmREAU , ' C.R. , ' vol. 95 , 1882 .
This paper contains an account of a method of determining high specific resistances , together with results as to the resistance of glass , which do not concern us .
The method employed is that of placing the substance to be examined between two cylindrical electrodes .
These electrodes are in series with a battery and condenser , the latter having its armatures connected to the poles of a LIPPMANN electrometer .
The observation consist , ed in noting the time required for the er to register a known difference of potential .
From this , uand the voltage acting on the resistance cell , and the capacity of the condenser , .
the specific resistance can be calculated .
In ' C.R. , ' vol. 97 , p. 996 , 1883 , this method is applied to liquid sulphur , the electrodes being concentric zinc cylinders .
We did not believe in the freedom of zinc from action on sulphur , but after a good many experi on bright , in no case observing any effect even up to the boilingpoint , we must admit thab zinc is probably a safe material to use .
No details are given by FOUSSEREAU as to the methods ( if any ) used to purify the sulphur .
The method adopted for examining the conductivity is free from error so far as the action of the containing vessel goes , but not with respect to the surface conductivity .
The sulphur experimented upon is as having existed in various states .
which are named , but the paper not suffiCiently detailed to admit of any dgment bein formed as to whether these states were actually attained .
The results are in ON THE ELECTRICAL PROPERTIES OF PURE SULPHUR .
Specific resistance of prismatic sulphur at 11 1 C. ohms .
Specific resistance of prismatic sulphur C. ohms .
Below C. , the conductivity ceases to be measurable .
With regard to the at C. , we show that this is not the value for pure sulphur perfectly atic .
With respect to the value al C. , all depends on whether the sulphur is heating or cooling ; but , in any case , we consider it is probably too low .
It is further stated that the same prismatic sulphur left to itself at ordinary temperatures , devitrifies and assumes an appreciable conductivity .
Thus : After one day , specific resistance C. ohms .
After two days , specific resistance at ohms .
This is contrary to our results .
Our impression is that probably the sulphur was a mixture of monoclinic and insoluble sulphur , that surface effects were not sufficiently guarded against , and that the sulphur was probably impure .
But here again the brevity of the account disarms criticism ; we can only say that the results do not agree with ours in any way .
There are other remarks on the conductivity of octahedral sulphur , which is said to be zero , and on the conductivity of sulphur in the liquid state , which we have not yet investigated. .
E. DUTER , ' vol. 106 , p. 836 .
Some experiments with an induction coilon melted sulphur .
The author found traces of action both on gold and platinum electrodes , J. MONCKMAN , ' Proc. Roy .
Soc 1889 , vol. 46 , .
The most complete record of the electrical properties of sulphur hitherto published .
He used a voltage of about 60 volts , and a hi gh resistance galvanometer .
The electrodes were of gold or carbon , and the paper is morae concerned in establishing qualitative properties than absolute quantitative results .
The results are more valuable than any of those hitherto mentioned , because the purification of the sulphur is more fully dealt with .
All the numerical results refer to melted sulphur , and therefore will not be dealt with here .
The most novel observation is that the resistance of\ldquo ; soluble \ldquo ; sulphur depends on the light upon it , a result which is inconsistent with that of a number of experiments , designed to test the point by one of the authors , with the assistance of .
J. B. ALLEN .
Specific Inductive of Sulphur .
An exhaustive list of results up to 1883 is given by GORDON E]ectricity and Magnetism vol. 1 , p. 134 ( Table ) ) .
The lowest results are those of FARADAY and GORDON , viz. , and ; and the highest recorded values are those of BOLTZMANN , from to , along a particular axis in octahedral sulphur .
Values given by oLTZlfANN ( by condenser method ) lie between the limits .
It is clear that the value of the specific inductive capacity of sulphur requires further XDCCCXCVJ .
\mdash ; A. investigation , if only from the great discrepancies which appear .
above .
Other values which have been obtained for the dielectric constant of sulphur will be found in WIEDEMANN'S ' Lehrbuch , ' vol. 2 , p. 25 , and onwards .
The only determination of the constant for sulphur in a special state is that of BOLTZMANN , but the purity of the natural crystal he used was not established .
Methods of Determining ffigh Specifc Resistances .
( 1 .
) The most direct method is that by means of a galvanometer .
This method was used by one of us in 1888 ( ' Phil. Mag 1889 ; vol. 28 ) .
The advantages of the method ( which will be fully discussed presently ) are that it involves a direct determination of voltage and current , and can be made practically independent of any leakage other than the one under investigation .
It is also fairly rapid and very flexible .
Its drawback is that large voltages are required , and , as generally practised , the method is wanting in sensibility on the one hand , and is more less dependent on some assumed galvanometric formula on the other .
These difficulties we have , however , overcome .
The galvanometric method was used with success by the Brothers GRAY , 'Proc .
Roy .
Soc 1884 , vol. 36 , p. 287 ; MONCKMAN , .
: BROOKS , ' Journ. Soc. Tel. Eng vol. 9 , p. 5 ( 1881 ) ; AYRTON and , ' Proc. Roy .
Soc vol. 27 , p. 219 , 1878 , and many others .
Leakage Methods.\mdash ; These are of two classes , and are so well dealt with in Professor GRAY'S book on ' Electric Measurements ' that very little remains to be said here , and that little is only by way of criticism .
In the first , place , the advantage of the method is that it is very sensitive and convenient , and can be carried out without special training by anyone possessing a good electrometer .
, it works well over a large range of voltages .
The disadvant .
for very high resistances are 1 ) the insulation of the instrument is only made sufficient with great difficulty , ( 2 ) the time required for an observation is very long , and this is apt to mask very important effects .
Again , if the calculation of the results be made from rate of fall of potential with time ( as was done by J. J. THOMSON and NEWALL ; ' Proc. Rcly .
Soc vol. 42 , p. 410 , 1887 ) , a certain want of sensitiveness is to be erved .
on the other hand , two potentials and two times sufficiently far apart for accuracy are observed , an assu.mption of a of resistance ( generally OHM'S ) requires to be made in order to integrate the equation GRAY , Absolute ents , , since , in the case of a very high resistance , the fall of potential is a small fraction of the total voltage only .
This assumption , although philosophically repugnant , probably does not produce any very disastrous effects on the subsequent statement of results .
An objection of greater gravity lies in the fact that very often the capacity of the quadrants of the electrometer cannot be neglected , and as these are affected by electric absorption ( with electrometers constructed with glass insulation ) , a source of uncertainty is introduced .
If the potential-time-curve-slope ) used to givo a value of the equivaTHE ELECTRTCAL PROPERT I OF PURE SULPHUR .
lent resistance for a definite potential , when neither capacity of the quadrants nor the leakage of the electrometer can bo neglected , then the formula of calculation becomes arithmetically very objectionable .
This is by no means a slight disadvantage if many results have to be reduced .
The capacity of the resistance cell also has to be known with fair accuracy , which , at all events , involves more work .
The gravest disadvantage , in our experience , lies in the variability of rate of leakage of the quadrants .
If the experiment last for an hour , as is often necessary , then undisooverable changes ma .
have been taking place in the insulation resistarlces of the electrometer .
If a large condenser be employed in order to reduce the magnitude of the uncertainties arising from want of accurate knowledge of the other capacities involved , we have to face the difficulty of making a proper allowance for its probably very variable insulation resistance .
Witl ] our high resistances this difficulty proved insuperable .
Taking everything into account , we do not consider this method favourable for the determination of absolute resistances , particularly when the word resistance has no meaning except with respect to an instantaneous voltage , In the second class of experiments the leakage method , the charge which ] eaks is taken up by an auxiliary condenser .
This method has been used by BOUTY with fair success , on standard condensers , for the purpose of investigating the phenomena of residual charge .
For this purpose , and when the capacities of the different parts of the apparatus are known , the method is suitable ; but , as a method for measuring.absolut.e resistances , it merely to the uncertainties mentioned above\mdash ; those due to the uncertainty of capacity and insulation of the receiving condenser .
In our case these methods were out of the question , as we had to deal with resistances greater than the insulation resistance of our mica condeI ) sers .
In the application of this method made by FOUSSEREAU matters were complicated by the use of a LIPPMANN electrometer , the poles of which can surely only insulate fairly well for small voltages , and which presumably has a not necessarily vanishing equivalent capacity .
However , we have not had much experience with the LIPPMANN instrument , so our criticism must 1lot be taken for more than it is worth .
A quite different method has been employed by H. KOLLER ( ' Sitzungsber .
Wien .
Akad vol. 98 , p. 201 ) .
This method consists in placing the substance undergoing investigation , in series with a high resistance constructed either of a tube of hate of zinc solution , with zinc electrodes , or of a quantity of saturated solution of iodine in carbon bisuiphide with platinum electrodes .
The latter resistances may be co1lstructed ( with proper precautions ) as high as SIEMEI$S ' units .
A divided circuit is used to give any desired voltage to the circuit containing the artificial istance in series with the resistance to be measured .
The fall of potential between one terminal of the cell or derivation circuit and the resistance cell , and between the same terminal and the artificial resistanc .
, allow ( proper connections being made ) of a comparison of the resistance in question with that of the artihcial resistance .
KOLLER was looking for qualitative rather than absolute quantitative results , and for this purpose his method has advantages in its quickness and simpIicity .
If absolute values are required , the metbod has many drawbacks .
The artificial resistances are subject to considerable variation , and are themselves calibrated one from another in such a manner as to give every opportunity for the accumulation of errors .
There is reason to suppose that the resistance to feeble currents , such as would be produced by a few cells , will depend on the previous history , and will not be by any means the same when the current is reversed .
In KOLLER'S experiments the current was , in general , not reversed ; in fact , the peculiar behaviour of the liquids investigated did not permit of the method of reversal being employed in all cases .
Since the resistance comparison depends on the comparison of two potential differences , it is necessary to eliminate the electrode effects without taking advantage of the method of reversals .
KOLLER allowed for the ectrode effects by observing them when no current was flowing , which was , perhaps , as much as the circumstances permitted , but hardly be considered sufficient when absolute measurements are in question .
KOLLER does not appear to have reversed his ectrometer ( loc. cit. , p. 204 ) in taking his readings ; an oversight which , we are persuaded , must damage the authority of his work .
We are , however , more concerned to explain why KOLLER'S apparently simple and advantageous method was not suitable for the ends we had in view , than to criticise his work .
We may add that the power of the method is finally limited by the want of insulation of the electrometer , but it remains probably the most sensitive method known for the examination of high resistances .
Method Adopted for Measuring Bigh Specific Specific Inductive The method employed by is substantially the same as that explained by one of u.s and J. A. POLLOCK , in a paper in the ' Philosophical Magazine , ' vol. 28 , 1889 , p. 469 .
The sulphur forms a between two aluminium plates .
A up to about 300 volts is supplied by a set of small test-tube accumulators .
The sulphur , storage cells , and galvanometer , are ed in series , keys being provided for cutting the galvanometer out of the circuit , and for reversin the current through the galvanometer without reversing it in the rest of the circuit .
Readings are always taken by the method of reversing the current through the galvanometer , and observing thefirst elongation of the needle .
Two galvanometers are always in commission\mdash ; one at a very high gl.ade of sensitiveness known as the " " new galvanometer\ldquo ; and the other at a lower grade , read by a spot of light on a scale , known as the " " old galvanometer In order to evaluate the of the galvanometers , a second set of apparatus is prepared , consisting of a standard Clark cell , a megohm , and various resistance boxes , whereby any voltage , from that of the cell down to one hundredON THE ELECTRICAL PROPERTIES OF PURIB SULPHUR .
thousandth of the same , may be included in the galvanometer circuit in series with a megohm .
A reversing key for the galvanometer and for the cell is provided .
The resistances are adjusted till the first deflection of the galvanometer needle on reversal is , as nearly as may be , the same as the deflection when the sulphur is in circuit .
The keys are arranged in such a way that the change from the sulphur circuit t , o the megohm circuit can be made almost instantaneously .
In this way a direct value can be assigned to the current causing any deflection of the galvanometer needle , without relying on any particular galvanometer law ; both circuits are practically induction]ess .
remains to measure the of the storage cells .
In our earlier experiments cells were permanently so as to be tested in sections against a set of 50 Clark cells , or a smaller number by the method of differences employed by Lord RAYLEIGH in bis experiments on the absolute electromotive force of the Clark cell ( ' Phil. Trans 1885 ) .
The resistance ofsulphur , however , depends so much ) previous history , voltage , \amp ; c. , that a refinement of this kind was soon seen to be unnecessary , and in :all the later experiments the voltage was obtained simply by the cells in series with a megohm and low stance galvanometer , and observing the double steady deflection .
This was compared with the double steady deflection produced by the 100 volts of the electric light circuit , derived from 31 plate .
accumulators , and measured by a FLEMING and GIMINGHAM voltmeter , which had been calibrated against Clark cells , but which was found to require no correction at or near 100 volts .
The.proportionality of the readings of this galvanomeber to the currents passing , within the limits employed , was established byaspecial seriesof experiments .
The galvanometer was resistance reflecting galvanometer , very strongly controlled , and was read with a lamp and scale .
It very soon became obvious\mdash ; especially when the sulphur was heated\mdash ; that some check must be devised , capable of detecting any breaking away of the film from the electrodes , owing to unequal expansion , \amp ; c. An arrangement of resistances for comparing capa(jities by DE SAUTY'S method ( see GLAZEBROOK , 'B.A .
Reports , ' Leeds , 1890 , p. 102 , or ' Electrician , ' October 3 , 1890 ) was accordingly set up in a permanent manner , and provided with keys , \amp ; c. , so that the capacity of the sulphur cell could be obtairled within a few minutes of the determination of i.ts resistance .
A certain 2 microfarad division of a condenser by ELLIOTT BROS .
was used as the standard , and this was afterwaxds compared with a condenser by MUIRHEAD , whose corrections were obtained and furnished by Dr. A. MUIRHEAD .
The method used by Dr. MUIRHEAD is described by GLAZEBROOK ( loc. cit uncertainties still remaining in the value of the capacity are probably much too small to require further discussion for the following reasons : When we used the capacity test merely to check the adherence of the film , the absolute value of the standard did not matter , and when we used it to determine the capacity with a view to obtaining the specific inductive of the sulphur , the small thickness of the film ( about 25 millim. ) prevented its actual thickness being measured with sufficient accuracy to make it worth while to endeavour to go behind Dr. MUIRHEAD'S corrections .
It is of course , easy to criticise the DE SAUTY method of comparing capacities as applied to two condensers , one showing absorption and the other not ( for sulphur is almost free from this phenomenon ) , but the largeness of our available voltage and the extreme perfection of our galvanometer permitted us to experiment under very advantageous conditions .
For instance , with 100,000 ohms out in one branch , and about 7000 in the other , an appreciable kick was produced when the balance was upset by 1 ohm .
Of course , it is necessary to use a thin fiim of sulphur in order to examine the resistance , and this is not the most satisfactory arrangement for arriving at the best value of the specific inductive capacity .
There are two incidental advantages , however , which are of importance in practice .
The first is that the correction for the effect of the edges of the plates is small .
Thus , the actual area of the opposing plates in one case found to be square centims. ; and with a film thickness of 0241886 centim. , the correction for the edges ( the plates being nearly square ) amounted to 26 square centim. only , or about 14 per cent. As it was impossible to insure the plates being exactly over one another , and as any deviation from exactitude would diminish the capacity , and as , moreover , it was difficult to avoid the presence of small bubbles ( whose area was approximately allowed for ) , it will be understood that the correction for the edges may be in general left on one side .
It was , however , usually applied .
the capacity being large\mdash ; between and .
electromagnetic unit\mdash ; the capacity of the connec and keys could be neglected without prejudicing the results ; they were , however , included .
The disturba ce produced in this case will be understood by the statement that the capacity of the keys and leads was measured by the ballistic method , as well as by the DE SAUTY method .
The throw was divisions for a voltage which gave with the sulphur condenser in as well , 388 divisions .
This corresponded to a capacity for the sulphur microfarad .
Qequently the capacity of the leads and keys , \amp ; c. , was about 2 per cent. of the capacity of the sulphur condenser .
Now we shall show that the ecific inductive capacity of prismatic sulphur is about , and of the same sulphur when mixed with about 35 per cent. of sulphur is as much as in one set of observations .
* Now , whether our results are right or wron , this shows that the correction of 2 per cent. is not of great importance .
It can , however , be supplied if desired , as the key , leads , \amp ; c. , remained the same throughout , and is supplied with the other corrections in the tables .
The reader may possibly and naturally think that we ought to have observed the specific inductive capacity of sulphur in thick pIates in order to get over the uncertainties mentioned , but it was one of our principal objects to discuss the specific inductive capacity related to the conduction phenomena , and this could only be done with security by making both sets of observations the same ffim .
That this was a necessity , will be understood after an examination of the tables , which show *Oclober , 1893.\mdash ; We believe this high value is due to some error whioh , however , we cannot trace .
The values of are elaborately dealt with by the method of eighing later on .
ON THE ELECTRICAL PROPERTIES OF PURE SULPHUR .
that an admixture of less than 1 per cent. of insoluble sulphur with monoclinic sulphur reduces the specific resistance of the latter ( in one case ) to at least the millionth part .
There is still room , however , for a careful study of the specific inductive capacity of sulphur in its several allotropic states , and with varying duration of charge , in order to discover whether the great differences between the obtained can be accounted for by what KOLLER calls the " " Schliessungs Strom for , from our experiments , residual effect exists only to a very small extent , even in tolerably conducting mixtures of soluble and insoluble sulphur .
This is one of the matters which requires further investigation and is dealt with later .
The drawings which accompany this paper will show the exact arrangements of the experiments , as well as the dimensions of the various parts of the apparatus , for the drawings are to scale with the exception of the resistance boxes in the ground plan .
We will now mention one or two points in which we have found it requisite to exercise particular caution ; the first of these is the insulation of the apparatus .
We have found that if the } ) aratus is to be left in position for any great length of time there appears to be a distinct gain in making use of a of different insulating substances .
Thus the " " old which is very well insulated itself , stands on three bits of ebonite placed on a clean sheet of glass .
This sheet of glass is supported by three combined glass and paraffin insulators , made by placing a small cylinder of paraffin on the top of a flint-glass bottle , and inverting over it a ical glass dish .
The outer dish serves to protect the paraffin against dust ; the whole stands on a slab of ebonite 1 inch thick .
Al ] the is insulated with extraordinary care , including the galvanometer and set of test-tube storage cells .
The\amp ; nacute ; , as a precaution against electrostatic effects , one end of the galvanometer is put definitely to earth by means of a wire soldered to the water supply .
The galvanometer is enclosed in iron case which is kept in connection with the earth , and affords a screen against external electrostatic eflects .
Referring to the yram ( see Plates ) , it will be seen that the effect is to put the lower plate of the sulphur condenser and one pole of the battery to earth as wel ] , which insures that except for the small P.D. corresponding to the small current flowing , the greater part of the circuit and all the coils of the galvanometer are at zero potential .
This has the effect of throwing the whole stress of insulation on the sulphur film , and on one insulating pole which the wire coming from the upper plate , and on two terminals of the reversing key .
The effect of accidental leakage , if any , is always observed by making all the connections except that leading to the top plate of the sulphur condenser , the wire simply hanging from the pole .
Though this pole is of splendid flint-giass most carefully varnished and having a resisting length of about one metre , and is provided with an ebonite cross-arm about 20 centims. long , to the ends of which the wire isfastened , the insulation is seldom sufficient till the glass has been dried and the long , thin cross-arm of ebonite scraped .
The connecting and reversing keys ( see Plate 3 , figs. 3 and 2 ) are formed of pillars of ebonite of from 8 to 20 centims. long , and the two points of the reversing key requiring special insulation are made by placing ebonite cups on the ends pillars of fused quartz , each about 10 centims. long .
The advantage o.f fused quartz as an insulator\mdash ; where it is exposed to dusty air\mdash ; lies not so much in its great freedom from surface effect , as observed by .
( ' Phil. Mag. ' [ 5 ] , vol. 28 , p. 14 ) ; as in the facb that a small oxy-hydrogen flame can be applied to it without .
of breakage , and thus the insulation be made practically perfect for the day , without any rubbing or deranging of contacts .
The advantages of the general arrangements are\mdash ; ( .1 .
) The points at which the insulation requires to be perfect are reduced to the fewest possible .
( 2 .
) The system being definitely earthed at the galvanometer , gives a complete protection against electrostatic effects , which are very noticeable when this precaution is not taken .
( 3 .
) The general insulation of the system is necessary when the ion of the galvanometer sensitiveness , or the capacity of the sulphur condenser , has to be determined , and it is a great convenience to secure this by disconnecting a single wire .
The insulation of the " " new galvanometer by means of fused quartz rods , will be explained in its proper place ( see p. 88 ) .
The necessity for insulating the lower sulphur plate well s'nown on one occasion , when it waes omitted , to give to some trouble .
Though this non-insulation produces no effect in so far as the applied voltage goes , .yet it has practically the.effec of short-circuiting the galvanometer through earth in one position of the rever-sing key and insulating one terminal of the galvanometer in the other .
There are always minute voltages in the circuit , and these will cause a current the galvanometer when it is short circuited .
Hence , on reversing the key a throw will be observed which might be mistaken for a conduction throw to conduction by the s.ulphur .
As these voltages are very small , however , a sufficient insulation is easily obtained by a system of ebonite , paraffin and glass .
SEUrION IV .
On the Construction of Sensitive High Galvanometers .
In a paper , " " On the Measurement of High Specific Resistances\ldquo ; ( ' Phil. Mag. ' [ 5 ] , vol. , p. 466 , one of us laid down what he considered to be the most important direction in which to look for improvement in galvanometers .
Since then work the subject has been going on , though with many interruptions , both in the workshop and .
During the time that has elapsed , the matter has received a good deal of discussion , notably from Professor AYRTON , and Messrs. and SUMPNER ( ' Phil. Mag. ' [ 5 ] , vol. 30 , p. 58 ) .
As to the corrugations on paraffined ebonite pillars , our experience that.they are ON THE ECTRICAL PROPERTIES OF PURE SULPH R. not to be recommended with bare ebonite , unless the pillar be arranged ( as should always be the case ) so that they can easily be dismounted and cleaned up with a outting tool on a Iathe ; mere washing or sand-papering does little good when very perfect insulation is required .
Theory , which is well understood , indicates the following conditions for maximum sensitiveness :\mdash ; ( 1 .
) The magnetic force per unit current should be a maximum .
( 2 .
) The magnetic moment of each member of the suspended system should be a maximum .
( 3 .
) The mirror should , if possible , be so good that its defining power is only limited by its size ; it should be optically perfect .
( 4 .
) The optical magnification should be a maximum .
( 5 .
) The astaticism of the suspended system should be a maximum .
( 6 .
) The field in which the suspended system moves should be zero when the astaticism is perfect , and the directional force entirely due to the unavoidab]e torsion of the suspending fibre .
Since he astaticism is never perfect , and the field can never be zero , the torsion of the fibre can only operate so as to reduce the sensitiveness , and must therefore be made as nearly zero as possible .
It is obvious that the above conditions are partially incompatible , and , as a matter of fact , it is the last condition which provides a starting-point .
The line of argument is this .
In order to prevent changes of zero , we are compelled to use quartz threads .
These threads are , however , very stiff .
In order to prevent the stiffness being such as to seriously affect the sensitiveness , the threads must be very fine , and , consequently , the suspended apparatus very light .
In order to beyond this , it becomes necessary to determine to what point the optical sensitiveness can profitably be carried\mdash ; or , what comes to the same thing , the limit of the angular value of the fluctuations of the zero .
Under fixed conditions , the weight of the mirror will simply depend on the defining power required .
The maker must take into account the optical limit as to smallness , depending on the wave-length of light ; and ( 2 ) the fact that the mirror must be thick enough neither to deform under its own weight , nor to deform owing to the freeing of internal stresses during its manufacture .
Before exact theory can be applied to the construction of a galvanometer it is , therefore , necessary to fix on the weight of the mirror , and though , with the experience we have now acquired with magnetically and thermally screened instruments , we can an estimate of the minimum weight of the mirror , we could not do so before the instrument was constructed .
It was necessary , therefore , to proceed along the lines indicated by the general principles involved , and to make such alterations from time to time in the design as increased knowledge suggested .
An alteration which is now in progress is the application of stronger material than glass to the preparation of .
Even hard glass , considered as a mineral is very weak , and is far surpassed in strength by many substances whose specific gravity is about the same .
Fused quartz , for instance , is much stronger , and several discs have been cut from this substance , and are awaiting manufacture .
We are trying with slices of bloodstone , \amp ; c. , and expect to obtain a considerable reduction in weight .
It must never be sight of that the weight of the mirror is an unmixed evil .
The mirror in use at present is of glass silvered on the back , with a diameterof centim. and a weight of 0485 gram\mdash ; it is slightly concave , having a radius of curvature of one metre , which was measured pproximately .
When we discuss the lines along which further improvements can be made it will be shown that .
owing to a peculiar mode of scale illumination which we have discovered , these dimensions can probably be a good deal reduced , but as our coils are wound at present , no advantage would be gained by effectingthis reduction .
The mirror is supported above the magnet system , but we are not sure that this is a good arrangement with such a large mirror\mdash ; with a slllaller one it would not matter .
The definition of the mirror* is so good that the image of a millimetre scale in it is read at a distance of metres , by an improvised telescope and micrometer eyepiece , whose scale is graduated to 2 millim. Each division of this scale can be divided by eye into about 5 parts ; and by means of the diffraction fringes surrounding the scale images , the latter can be located this of accuracy in the micrometer , With regard to the designing of the instrument we will take the conditions for maximum sensitiveness as laid down above , in order .
We have experienced so much inconvenience ourselves when reading papers , in having to refer from lettering in the text to letters on the drawings , that we have prepared notes explanatory of the drawings to be placed beside them , and will here to give what explanations are required without making it necessary for the reader to refer to the drawings at every moment .
( 1 .
) The magnetic bforce for unit current sho.uld be maximum\mdash ; resistance of the coil , or trouble and expense of winding , not to be considered .
We have endeayoured to make an improvement in this direction .
by using four pairs of coils , one above the other , instead of two , as has hitherto been the case .
Preliminary trials showed .
us that we could hope to make the suspended sysbem about 20 centims. long , and conseq.uently the maximum diameter of the coils was approximately fixed .
It will be noticed that the coils are much smaller than those ordinarily employed .
This may be taken as a realization of a well-known principle , that when it is a question of ance or quantity of.wire , it iS better to have many small coils than large ones .
We by making a drawing of an instrument with eight tiers of coils , but abandoned it on account of the difficulty of a suitable magnetic system of the necessary length .
It is necessary to wind the wire regularly at the commencement of the winding , in order to get most turns in the most important part .
We decided to wind the coils with cylindrical holes .
The dimensions of the coils are shown in the This mirror was for us by our assistnnt , Mr. J. Coox .
ON THE ELECTRICAL PROPERTIES OF PURE SULPHUR .
drawing ( Plate 1 ) .
It might possibly be orth while to use fine silver wire , and would certainly be advantageous to have the wire specially lightly insulated , but we are too far from manufacturing countries to be able to take advantage of such expedients .
The wire employed by us has a diameter millim. It was frequently necessary to make joints , owing to irregularities in the insulation .
The material for reinsulating the joints is a film of collodion , as it takes up less room than anything else we know of .
( 2 .
) T'he magnetic moment of each member of the suspended system should be a .
This again is not purely a matter of design .
Our magnets were hardened circular discs stamped oub of thin sheet steel .
Each magnet consisted of two discs , one on each side of the support .
Their dimensio1ls were : Diameter .
876 centim. Weight 0809 gram each .
Thickness 008 centim. So that the weight of the suspended part is made up thus : gram .
8 discs weigh 2470 1 mirror weighs 0485 Wire and cement 0660 Total weight . . . .
3615 So that the weight of the mirror is only about one-eighth of the whole weight , and the magnets form about five-sevenths of the whole weight .
At first we tried mica as a supporting material , but found it very much inferior to flattened alun ) inium wire , which can be twisted .
A strip of mica , moreover , makes the combination more dead-beat ; which , with our method of observation , is equivalent to a reduction of sensitiveness .
We also rolled some sheet steel , 004 centim. thick , into small cylinders as described by Professor LANGLEY , and compared the intensity of magnetization .
obtained with that of the small discs .
The result is given in the accompanying table ; it will be noted the advantage is on the side of the disc .
The experiments were made by Mr. POLLOCK .
We dare say that better steel can be got ; but , unless Mr. ELLERY , F.R.S. ( Melbourne ) , had kindly come to our assistance , we should have none at TABLE IV.\mdash ; Comparison of netic Moments of Discs and Cylinders .
Condition ( 3 ) has been ready treated .
( 4 .
) The optical should be maximum .
No part of our work has given us more trouble than the production of good scale images , and , in the end , in no part have we had more success .
We began by using the telescope and scale set up in the usual manner , and , as usual , were troubled by want of illumination some parts of the scale , and excess of illumination over others .
The sources of light were also sources of heat , and caused the galvanometer to be unsteady , though it was well boxed in .
We will not describe the many ineffectual experiments we made .
Finally , the following plan was hit upon .
A minute point of light ( a kerosene lamp turned down very low ) was placed edge-on about metres away from the galvanometer .
An image of this source was focussed on the centre of the galvanometer mirror by a large lens , centims. in diameter , placed at a distance of 55 centims. from the mirror .
Directly behind the lens , between it and the mirror , was placed a scale , which , in our experiments , and after many trials , was made coating a slip of glass with a film of Canada balsam stained " " \ldquo ; and cutting divisions on this film on the dividing engine .
The divisions were transparent , the general field being dark , and carefully kept so by placing the engrave scale in a proper recess cut in a black wooden screen .
The advantage of the Canada balsam film is not obvious in this case , but the scale had been used in other experiments where double images gave trouble , and then the Canada was of the greateqt advantage in helping to prevent their formation , much in the same manner M. CORNU'S varnish cures photographic plates of giving.a halo round the image of a bright star ( ' Journ. de Phys vol. 9 , p. 275 ) .
The film was , of course , turned towar.ds the mirror , .
and the light fell upon normally .
It will be understood that , by we mean that the'scale was a gent to a circle described about the centre of the mirror .
The object , of courSe , to let the light pass through each division of the scale normally ; and , since the scale was only about 10 centirns .
long , this condition was sufficiently fulfilled , and the scale did not require to be bent .
Rea ] images of the scale divisions were formed at about ON THE ELECTRICAL PROPERTIES OF PURIC SULPHUR .
260 centims. from the mirror .
The reflected beam was , however , intercepted by the telescope , whose object-glass , consisting of a spectacle lens of 10 centims. focal length , was 267 centims. from the mirror .
The image was observed in the micrometer eye- piece .
When the light was not too bright , and everything well adjusted , the images vere exceedingly bright on a dark field , and each was the centre of a system of the most perfectly and fine interference .
By moving the lamp slightly , these fringes can be got to themselves somewhat unsymmetrically about the luminous image , and thus enable an observer to distinguish a particuJar fringe with .
eertainty ; one crossing the brightest part of the luminous image was always selected ; and if the fringes were unsymmetrical one was always much the most sharply marked .
The similarity of the tracings of the dividing engine was such that each division had a set of fringes , as nearly as we could see identical with the fringes of the images on each side of it , so that a particular fringe could be selected as the fiducial mark of the image , or the distance between two similar fringes could be taken as the distance between the millimetre divisions on the scale .
At first the micrometer eye-piece was provided with a toothed or notched scale , but this not being sufficiently capable of , it was replaced by a scale divided on thin glass by a diamond , on the dividing engine , to 2 millim. A candle was then arranged so as to throw a suitable diffused light into the telescope , the micrometer scale presenting a edge , the coincidence of its divisions or otherwise with the interference fringes could be made out clearly and with great accuracy .
The spider-line of the micromet , er was not nearly so fine or so well defined as the interference fringes .
By properly adjusting the fringes and illumination , everything became so distinct that it was much easier to read to micrometer scale divisions than it usually is to read with a lamp and scale to one division .
We are quite sure that this method of illumination and scale reading will be found valuable by any one requiring to read small deflexions .
The only points about which we would gest caution are the followinoo:\mdash ; The light falling on the :nirror must come from a source as small as possible , and a person setting up the mrangement for the first time would probably make the light much too bright and furge for the interference fringes ; the source must be a mere spark .
The lens must be adjusted so as to be normal to the joining the mirror and Iamp , and this line must pass through its centre .
The lens we used was an ordinary one , but it had good adjustments .
The divisions on the nada balsam or other varnish must be fine ( we tried photographic transparencies , \amp ; c. , instead , but they always looked woolly ) .
The eye-piece of the telescope must be capable of very good adjustment .
Ours was..mounted on a fine tuning-fork stand by KoNIG ; it is no use unless the stand be very steady , and the adjustments , both vertically , horizontally , and in the altitude must be smooth and good .
We made ours out of good brass tube without any diffieulty .
The source of light should be within reach of the observer at the telescope , in order that the final adjustment may be made by slightly moving the source , and by the lamp up and down till the best effect is .
this.involves ving the reflected ray separated from the incident ray by an is not too large .
The method is suitable only for the observation of small angular deflexions , fcr we have not been able to rave our scale with numbers finely enough permit us to see the figures and divisions in the field at the same time .
Consequently , a deflexion can only be observed by counting the images of the millimetre di.visions as they flash past .
With our present arrangements wo can only count the divisions when they pass a certain rate ; when the whole deflexion amounts to more than about 20 millims. , the divisions pass too quickly to be counted .
However finely we might succeed in numbers on the scale , we fear that the interference images would be too blurred to be read .
The best method wouId be to omit every tenth division , say .
We did not require to trouble about this , however , for the " " old\ldquo ; galvanometer was always set up at hand , and , as soon as the deflexions became too large for the method of observation mentioned , recourse was had to it .
The ( Plate 4 ) will make the arrangement clear .
A deflexion of one millimetre division is in angular measure , and since each millimetre division of the scale covered about 8 micrometer scale divisions , each divisible into 5 parts , we may consider that the smallest deflexion which could be read with certainty was say one-fortieth of this , or .
Since the magnification depended on the adjustments , it .
varied slightly from day to day , but the general magnification was , as we say , such that the distance between two consecutive scale images covered from 8 to 9 micrometer scale divisions .
We ought to add that the definition of the mi1ror was a good deal better than we have ever seen before in a mirror of about the same weight and dimensions .
It was worked for us by Mr. COOK , together with a large number of similar mirrors , all about equally good .
The best unworked glass we have tried was incomparably worse .
The difficulty in making such thin mirrors out of glass arises from the that they always change their shape when they come off the polishing tool ; but in our experience the change of shape is very regular , thus a flat mirror almost always becomes concave or convex ; but it does not do so irregularly , and consequently the definition nains .satisfactory .
The glass , of course , is ground and polished on both sides .
The window through which the light passes to and from the mirror , is a strip of potent plate-glass , selected by the method of observing the reflected images from the front and ba , ck , and cutting the glass so that it forms a prism of small angle along t , he top or bottom edge .
Patent plate.-glass can be got with fairly flat surfaces , whirh are , however , generally more or less incline:d , and so in cases of this kind the ilst be cut in such .
a way that the inclination of the surfaces does not disturb the of the .
ob@eivations .
For this method of selecting glass , we are indebted to Professor WRIGHT , of Yale .
( 5 .
) staticism of the suspended system should be a This is a most tiresome condition to satisfy , .
and have nothing.to add to what was said about it by one of us Phil. .
In order to magnotize the system , four agnets had to be used ; an attempt- that was ON THE ELECTRICAL PROPERTIES OF PURE SULPHUR .
to screen half the magnetic } , while the other half was being magnetized , iled , even when the screening was attained by pushing half the system into a cylindrical hole bored in a thick soft iron .
During the process of mounting the discs on aluminium wire , they genelally got magnetized , and hence required to demagnetized before the final magnetizing process was carried out .
The demagnetizing was carried out by means of the arrangement of four siulilar electrowhich were suppJied with alternating current .
If we had foreseen the necessity , we should have provided these magnets with laminated cores ; as it was , care had to be exercised to prevent them becoming so hot as to melt the paraffin used as a cement for the disc magnets .
The final adjustments for astaticism were made by twisting the aluminium wire .
This wire was only millim. in diameter , and so the difficulty of obtaining a perfectly straight system of magnets of about 20 centims. long , and so astatic that the natural period of vibration was ffom to 2 will be understood by anyone engaged in similar work .
The natural period of vibration of two magnetized discs was a fraction of a second .
We had hoped that it would turn out more easy to secure good astaticism with four-magnet systems than with two , but this hope was not justified by the event .
The adjusting and mounting of the astatic system is the most difficult part of the manipulation ; the final adjust .
meant mnst be made by hand twisting , as descri.bed in the ' Phi ] .
Mag. ' ( loc. cit ( 6 .
) The Jibre must have a minimum torsional coefficient .
In plain English , the fibre must be as thin and long as possible ; and it is , of course , much more important to have the fibre thin than long .
The proper thickness was calculated from the results obtained by one of us in a paper on " " Quartz Threads\ldquo ; ( ' Phil. [ 5 ] , vo ] .
30 , p. 99 , 1890 ) , and the thread was then picked out by measurement with a microscope ; a factor of safety of about 2 being generally allowed .
As a matter of fact , the thread required is very fine , but not unmanageably so : pieces about 25 centims. long were generally used .
We find that the thread can be fastened more securely by a little melted shellac than by hard paraffin ; we have had several accidents from the thread slipping out of the paraffin .
Of course shellac that , has not been overheated or dissolved in alcohol , must be used .
Care was always taken to use the thread in an untwisted state .
On the Design of the for High Deyrees of Sensitiveness .
The general principles had been arrived at in 1889 from experience with the old galvanometer , from which fair results had been obtained .
The following is the result our general experience .
( 1 .
) It is of the first importance that the coils be adjustable to the suspended part , as well as the suspended parts to the coils .
Unless provision be made for this it is practically beyond the powers of even the most skilful manipulator to make use of magnets properly filling up the coil space , and yet free to turn .
This condition PROFESSOR RTHREbFALL AYduse oofter oramework onstrument which h coils , can be lifted up and carried away from the other the plane of ' being the plane containing the magnetic system .
In order that the adjustment of the magnet system to the coils may be perfectly made , it is necessary to be able both to look through the axial holes of the coils and also into the narrow space ( millim. ) separating the two systems of .
The method of mounting the coils will be clear from the drawings and description ( Plate , but it is necessary here to indicate the general procedure .
The coils , when wound , were embedded concentrically in rings of ebonite , a little paraffin being poured in and fixstened both to the coil and ebonite by means of a hot wire .
Four bars of fused quartz had been previously fastened as spokes into the ebonite\mdash ; regarded as a hub\mdash ; and by means of a centering apparatus devised by Mr. COOK , each coil with its ebonite and quartz was laid in a proper position in its coil carriage .
This was a matter requiring much ingenuity , and was accomphshed by means of the adjustable catches into which the ends of the quartz rods drop , and which are screwed to the carriage .
The coil being in the centering apparatus , the catches of the carriage are adjusted to the quartz spokes , the catches are then made fast to the carria.ge , and then the quartz is made permanently fast to the catches by means of a little plaster of Paris .
The quartz spokes are like nails , i.e. , they have heads , and are fastened to the ebonite rings by bein pushed through them from inside before the coil is centred and cemented by paraffin , which also , of course , cements the head of the quartz spoke .
Each coil carriage is screwed to the frame by separate screws , which allow a little latitude for final adjustment to the magnetic system when this is suspended .
The detail of the whole of the devices for adjusting coils insulated by quartz rods , we owe to Mr. COOK , and we are not sure that the system would succeed with less perfect workmanship than he put into the construction .
The arrangement proved in practice to be everything that could be desired .
( 2 .
) The apparatus carrying the coils and magnet system must be absolutely separate and independent of the parts carrying the oontrolling magnets .
This is a condition of the first importance \mdash ; however good all the other arrangements may be , they will fail in practice unless this condition is fulfilled .
We have tried over and over again with the controlling system more or less connected to the suspended system , but have never had any success with such an arrangement , however massive the brass work This condition may be regarded as the result of all our experience .
At one time we tried supporting the coil system on rubber bungs mounted with more or less elaboration , and on discs of rubber mounted in different ways ; but observations with a mercury trough convinced us that with our concrete and ashphaIt block floors , the rubber always ON THE ELECTRICAL PROPERTIES OF PURE SULPHUR .
did more harm than good .
At present the instrument is mounted on a hard smooth slab of concrete , which reposes on a sandstone pillar , on the concrete on which the floors are laid .
The coil system is mounted on a sheet of ebonite supported by insulators on the concrete , and the controlling magnet system is supported by leaden pillars passing through the ebonite without touching it , and rest , ing on the concrete slab .
The natural stiffness of the slab , or rather its is much increased by the fact that it carries the magnetic shield of cast iron about three hundred pounds .
This materially lessens any tendency there might be to produce disturbance in the uspended system by adjusting the controlling magnets .
The Magnetie Control.\mdash ; The whole secret of success lies in this .
stead of regarding the control as a subsidiary part of th apparatus , it must be regarded as the most important part , must be capable of the finest adjustment .
We have fOund it advisable to use controlling nets at such it distance that the residual field is due to them , and not to the earth .
In this case , small changes in the direc ion of the earth 's field produce less effect than when the residual field is that due to the magnetic action of the earth .
We have a simple but important improvement to report in the matter of control .
It turns out that when the sensitiveness is very high , the uniformity of the magnetic field becomes very important .
When the field is not sufficiently uniform , the following fatal effect is observed .
We will { { suppose that the earth 's in excess , and that the magnets are in the magnetic meridian .
The control magnet , we will suppose , is lowered gradually ; , a point is reached at which the period is fairly long , and the suspended part , if nearly at rest , behaves apparently ( well ; now let it be caused to vibrate ( in practice , the control can never be lowered without causing the suspended magnets to more or less ) .
It is always noticed that if the excursion rises beyond a certain very small value , theh the suspended magnet swings round perhaps or , and takes up a new position of greater stability than the old one .
In order to bring it back the control magnets have to be rotated , with the invariable effect of making the suspended maguet swing past the north and south position , and take up .
a lpore or less ical position on the other side .
Attempts to bring the needles back by moving external or working damping arrangements , Iseldom succeed , and in the end the controlling magnet has usually to be raised , and Tthe tiresome business begun afresh .
We have wasted many days over this untoward phenomenon .
Suppose , however , that the difficulty is apparently surmounted and the image got on the scale and fairly steady , and observations begun .
It ways found that after a time the needle begins to drift , and finally goes over to one of tho side positions .
This was first noted by Mr. POLLOCK with the old galvanometer in 1889 .
The cause of these effects is not far to seek .
The field produced by a simple control magnet is never uniform .
In such a long system as ours , and with the controlling magnet at the top , the lower suspended magnet is directed by the earth 's field , and the upper one by the control , the centre ones being probably in a nearl .
NDCOCXCVL neutral field .
The diagram MAXWELL , Elec .
, shows how very far from uniform the field is , at or near the neutral point .
If the magnetic system be rotated through a very small distance , it comes to a place where the resultant field is differently directed , and then sails away till a new position 01 equilibrium is attained .
The usual practice of considering the field as only influenced by the earth and controlling magnet , and of regarding the suspended magnet as without influence , is fallacious .
The direction of the lines of force depends on three things , the earth ; the controlling magnet , and the suspending magnet .
MAXWELL'S , represents the case of the discovery well , but the matter is complicated by the fact that the resultant field of earth and control magnet is far from uniform even before it is distorted by the suspended magnet .
It seemed likely that much of the difficulty would be removed by making residual field more uniform , and with this object two additional and similar control magnets were introduced below the suspended system , and adjusted by trial till they were as nearly as possible symmetrically placed with respect to it and the uppei acIjustable system .
The magnets were of the ions and at the distances in the drawing ( figs. 1 and 2 ) , and it was expected that they would produce a field lrface uniformity sufficient for the purpose , after the manner that bhe HELMHoLTlr coil arrangement produces a solid uniformity ( MAXWELL , vol. 2 , fig. 19 ) .
The arrangement figured was set up as a trial , but as its success immediate and complete , we never went on to the selection of magnets of proper length\mdash ; a step we expected to have to take .
Since using the arrangement all the wearisome and trouble has been diminished so much that we now adjust our sensitiveness and observations with no more concern than if we were operating with a wooden galvano meter for rough resistance balance .
* We must not forget to mention the prime importance of having the fine movements of the adjustable part of the control\mdash ; the upper magnets in our .
case\mdash ; as good workmanship can make them .
The rough adjustment for distance is made by sliding the nets , screw system and all , up and down .
The adjustment is by of an excellent nut and fine screw thread\mdash ; a device described in 1889 .
The adjustment in azimuth is by means of the lever and screw device , the rough adjustment by releasing a clamp and twisting the magnet carriage .
As the ma.gnets and nut weigh several pounds , we should have done better had we arranged a counterpoise , sc that the freedom of rotation of the screw should not be affected by the great pressure on it , as is the case at present ; for , though the bearing surface of the nut was well ground , and was kept well oiled , still , on screwing the magnets up , the friction was to slightly twist the wide and thick brass tube , and so to displace the magnets in azimuth .
* Oct. , 1893.\mdash ; After more tha , a year 's experience we have never had the slightest difficulty obbaining the eusitiveuees desired Sinc9 we made the field unif-orm , ON THE ELECTRICAL PROPERTIES OF PURE SULPHUR .
We believe that it is not feasible but advisable to further increase the uniformity of the controlled field .
of the Coils from the .\mdash ; If the quartz is protected from dust it ateS very well , better than anything else we know of in places which can not be got at constantly for treatment .
After in use for some months , the total insulation of the coils and terIninals from the frame ohms , the measurement being made by means of an electrometer .
The method adopted for insulating the terminals and connecting wires does not differ much from ordinary practice .
The terminals are of corrugated ebonite , of slender section ; the wire passing into the terminal from the coils passes up the ebonite column without touching it , till it reaches the binding screw .
The ebonite terminals themselves are screwed into the glass forming the front and back of the galvanometer , and the insulation is improved by drying and varnishing this glass .
Where the wires pass from coil to coil inside , they are supported as little as possible ; but where supports can not be avoided , these are made of needles of fused quartz .
All the connections are permanently soldered , and it is to be noted that we have carried out in insulating the coils what we COIlsider to be an essential condition of success , viz. , the use of a variety of insulating substances .
There are 12,280 turns in each coil .
Total turns , 98,240 .
resistance , 43,985 ohms , at C. Screening from Effects.\mdash ; We expected have to interpose a film of mica silvered on one side , between the coils and magnetic system , but this has turned out to be unnecessary .
The insulation would certainly suffer if the magnets have to be screened .
of Coils.\mdash ; The total capacity of the coils and terminals , the inner casing being put to earth , is lnicrofarad , by measurement .
of Thermo-electric Effects.\mdash ; When the instrument was first set up , some small pieces of stiff German-silver wire were used to pass down inside ths ebonite terminals .
On short circuiting enormous thermo-electric effects were always noted even when all precautions to secure a constant temperature had been taken .
These effects could only be removed by replacing the German-silver by copper , and by electroplating the brass binding screws with copper , and by soldering ] joints ( except the contact at the binding screws ) right up to the galvanometer key .
Screening Air Currents and Convection Currents due to diffe , rence of Temperature.\mdash ; It was found necessary to seal the coil cases with soft wax , so that interior of the galvanometer becomes almost -tight .
The sealing is easily accom , because , as has been mentioned , there are four glass sides , and the wax was run round the edges of the glass .
Thermal effects were guarded against by enclosing the whole apparatus in a mill-board box , coated with tin-foil inside an out , and very conveniently built up around the heavy gullmetal frame carrying the controlling system .
The iron slabs of the nuagnetic screen were also tiu-foil coated and a roof was erected on them of mill-board covered with tin-foil .
This roof had holes cut in it for the connecting wires to through , and for the stem of the controlling system : All these precautions were the result of attempts to remedy faults as they made their appearance .
Magnetic Screen.\mdash ; Four iron slabs were prepared from cast iron and arranged so as to a symmetrical box , without top or bottom , round the instrument .
The screening was not very perfect , and if we could have got soft iron we would have used it .
However , the improvement made by this screening was very great , and it was no use trying to irllprove it by cast iron , for we used as much as we could trust the concrete slab to carry , and put it as near to the instrument as possible .
It was supported on a wooden frame .
When further improvenIents are made , the most tant of them will be in increasing the netic screening , even if it takes half a ton of soft iron .
The approximate reduction of the earth 's horizontal field due to our screen is from 26 C.G.S. to about one-twentieth of this .
The screen weighed Fluctnations of \mdash ; These always make their appearance when the single steady deflexion is above 1 micrometer for ampe .
We have endeavoured to the cause of this fluctuation .
We made a number of experiments by heating the instrument uuequaly by means of gas furnaces placed in different positions with respect to it , with the result that we discovered that the tempetatut.e differences had a considerable , but not enough to account for the .
We habitually use the instrument in a room of fairly constant ature , and the method of illumina , tion cannot be charged with giving rise to variations ; for the lamp may be turned up so as to radiate , say , a hundredfold as much energ.y without producing any increased effect .
The only radiation reitching the interior from the is that which can pass through the fine scale divisions , and this all npinges on the surface of the mirror .
The fact that the instrument is practically air-tight , precludes the theory that external curren ts of air cause the Fact .
Besides , there is the mill-board box coated with tin-foil , and outside that the iron and mill-board box .
It remains to consider variations in external magnetic force , and it is to these variations that we attl.ibute the fluctuations observed .
Our reasons are : ( 1 ) The magnetic screen made the needle much steadier .
( 2 ) The residual unst.eadiness varies much from day to day ; on some days it is impossible to use the instrument with a period of more than 10 sec. , and on others the zero is absolutely still and constant during , say half-an-hour , when the penod reaches sec. or 30 sec. , and , of course , it vays happens that these periods of magnetic peace only occur when we want to measure comparatively low resistances .
The fluctuations have , moreover , all the appearance of being due to general causes , for they appear equally at night-time when ON THE ELECTRICAL PROPERTIES OF PURE SULPHUR .
nothing in the Laboratory is being moved , and when there is no traffic in the road .
They are always bad on a windy day ; but this may possibly be attributed to small changes of barometric pressure occurring from wind pressure .
We are placed at about 150 yards from a road along which a steam tram runs , and every time a tram eomes past , the needle moves about violently ; in fact , we often obtain the indicat , ion of the approach of a tram from the motion of the needle .
All these drawbacks cau only be eliminated by increasing the efficiency of the magnetic screen , and this is therefore the next step to take .
Method of Observation.\mdash ; In order to obtain the maximum reading for a given current , we always obtain a double deflexion , and read the first elongation of the needle .
Whether this gation be produced by the current through the sulphur or by the derived circuit from the Clark cell , there is practically no induction in the circuit , except that in the galvanometer itself If the derived circuit current be adjusted to give the same throw of the galvanometer as the leakage current , we consider that the currents must be the same .
As an example of the behaviour of the instrument , we will give one complete set of experiments , at a normal sensitiveness , taken on October 17 , 1892 , a very unsteady day .
TABLE V.\mdash ; Throws of Galvanometer in Micrometer Scale Divisions .
Mean deflection , micrometer divisions .
Period of a double vibration , sec. Values of \mdash ; The ratio of the 2nd elongation to the 1st .
( 1 ) ( 2 ) PROFESSOR R. THRELFALL AND .
J. H. D. BREARLEY ( 3 ) ( 4 ) Mean , If be the steady deflection corresponding to the throw , then ( MAXWELL , S745 ) Now , and therefore .
But is the deflection on reversal , hence deflectiou from the zero for the current employed is ions .
current ampe ampere .
Hence a current of ampere produces on reversal a throw of micro .
meter divisionls .
Leaving the damping out of consideration , this is the same as a current of ampere , producing a throw of 1 micrometer division on reversal .
Taking damping into consideration , this would be slightly increased , since the damping is less the smaller the deflection .
Now on several occasions ( though not on this particular one ) in times of exceptional magnetic quiet , readings could be taken to one-fifth of a ometer division .
If , therefore , we consider that a throw of onefifth of a micrometer division is the lesst observable , it follows that we could detect a current of one-fifth of the above , viz. , ampere , and this with a time of swing of only sec. In practice we do not that a current so small as ampere could be detected with ainty , even when the galvanometer is steady , because of the small thermo-electric effects that are always exhibited by the contact keys , and which would probably give rise to much greater currents than this .
* One of the most important matters awaiting solution is the construction of a reversing key , which shall be free from contact effects .
If we estimate the current required to produce a steady single deflection of 1 micro meter division , we find it is and the current which would produce the smallest perceptible deflection is one-fifth of this , or ampere .
On the same day a similar series of experiments were made with currents corresponding to and .
part of the voltage of a Clark cell .
* Oct. , 1893.\mdash ; Since then we habitually measured currents of this order .
ON THE ELECTRICAL PROPERTIES OF PURE SULPHUR .
The following table is a summary of the results of these and other ions , and shows how necessary it is to calibrate the galvanometer for each particular throw , when these are not large:\mdash ; TABLE VI .
, giving Summary of Experiments made on October 17 , 1892 ; October 25 , 1892 ; and December 1 , 1892 .
Galvanometer period , sec. on October 17 , 21 sec. on October 25 , and 26 sec. on December 1 .
8 Micrometer divisions cover 1 millim. in the scale image .
This table requires little comment .
If the sensitiveness were reckoned in terms of the least visible deflexion , it would appear to be five times greater .
The deflexions are obviously rather greater for smaller currents than for larger .
If we consider what has been said about the state of the resultant field , even under such a fairly strong ntrol as this , it appears that any simple nometric law is not to be expected .
It is not worth while exhibiting any other tables of sensitiveness , for , though all the sulphur observations might be used for the purpose , in general the sensitiveness for steady deflexion was not required , and was not calculated .
* In the tables giving the results by the measurements of the sulphur resistance , the sensitiveness per double deflexion is given , corresponding to each set of observations .
For the sake of rapidity of observation , the sensitiveness was adjusted to suit the work in hand .
We do not wish to attempt any comparison between this instrument and those employed by other experimenters : our object was to attempt to combine ) a high degree of sellNote added September 23 , 1893:\mdash ; We have lately had occasion to use the galvanometer at the highest degree of sensitiveness conveniently attainable , in order to study the conduction in films which could not be made very thin .
For this purpose the period was raised to 25 sec. The terrestrial magnetic conditions were only fairly steady .
Hence , mean deflexion for a resistance of 2 ohm between the ts of derivation is is less thau amperes , since the megohm was in series with th The sensitiveness , therefore in amperes per scale division , is On a really steady day , 2 micrometer division can be read with absotute oertainty ; dition had obtained on this occasion , the sensitiveness for least observable xion wou ave 3 ampere .
ON THE ELEC ICAL IES OF PURE SULPHUR .
of adjustment .
The improvements which can still be made are ( 1 ) greater strength in the support of controlling system ; ( 2 ) increased fineness of adjustment of the controlling system ; ( 3 ) better arrangements for twisting the suspended fibre when the controlling system is mounted ; ( 4 ) increased uniformity of residual field ; ( 5 ) more perfect magnetic screening ; ( 6 ) numbering the divisions of the millimetre scale ; ( 7 ) observing in some place not affected by traffic along streets or roads .
Of these , and ( 7 ) are much the most.important .
Perhaps we ought to add that we are quite aware that at least as high a degree of sensitiveness as we have attained , may be got out of a much less elaborate instrument \mdash ; on occasion .
We have obtained such results ourselves with the\ldquo ; old\ldquo ; galvanometer .
The point we wish to make is not that the sensitiveness is extraordinary , but that , such as it is , it is daily and hourly availabJe with all the ease and certainty tion generally associated with instruments of , say , a thousand times less sensitiveness .
SECTION V. Account of Experiments on Sulphur in Chronological Order .
The first experiments were made with the arrangement of plates figured in the ' Philosophical azineo ' [ 5 ] , vol. 28 , Plate 14 , and the diagram of connections was the same as is given in the same volume , p. 470 .
The plates were of brass truly faced by scraping to a surface plate : they were first platinated by ROSELEUR'S method , and afterwards gilt and burnished many times .
The object of the continual gilding and burnishing was to make the gold film as solid as possible ; for it was found that if the film was at all spongy , the sulphur penetrated it at a high temperature and acted on the brass below .
There were also some few spots that seemed disinclined to gild .
These spots were finally drilled out and plugs of gold were inserted by a dentist 's apparatus .
The result was never perfectly satisfactory .
The flatness of the plates , however , was preserved to such a degree that the mere contact of the upper cleanplatewas strong enough to lift the lower one , which weighed about kilograms .
hese plates were prepared by Mr. POLLOCK .
{ ilm of sulphur was got between the plates by placing the lower one in the gold dish , and pouring in sufficient melted sulphur to cover it to a depth of about half a centimetre .
The sulphur used was in in very subsequent experiment unless the contrary be specifically stated\mdash ; the sulphur recovered by the Cbance process , distilled three , times in hard-glass retorts .
A subsequent examination of the film showed that it was almost entirely monoclinic oontaining less than 1 per cent. of insoJuble sulphur .
The screws used for separating plates had gone a slight , so that the effective area as deduced f.rom a series of measurements by a firlQ pair of vernier callipers was .
centims. From this it was estimated that the actual area was. .
sq .
centims. \mdash ; 3 sq .
centims. being taken up by bubbles , and by a corner of the film which was CXCVI .
broken away in cutting down into the sulphur round the plates in order to increa6s the surface resistance , We may call the area 190 sq .
centims. micrometer screws were adjusted so as to leave a space of 05 centim. This was therefore thickness cf the filtn .
So much difficulty was experienced in screwing the screws back that it was deterinined for the future to use some other method of separating the gilded plates .
the attempts that were made to get a value of the resistance of the film , several facts came to light on which our sequent practice was founded .
The was with respect to the necessity for insulating every part of the appa ratus .
Our connecting keys proving unfit for such hi gh resistance work , new ones were 1nade out of rods of ebonite about 20 centims. long , and were so arranged that it was easy to get at the ebonite in every part for the purpose of scraping it with of broken glass\mdash ; a very effective way of it .
It was also found that there was a considerable amount of surface conductivity over the sulphur , and we finally were obliged to use phosphorus pentoxide as a drying agent , anything less active fail .
ing to give satisfactory results .
The battery of test-tube storage cells was reinforcffi by a water consisting of zinc and copper plates , as described by Professo ] ROWLANI ) Phil. .
In the final experiment with the film under a voltage of 551 was used .
It was measured\mdash ; not very accurately but sufficiently -by the simple means of charging a fraction of a microfarad condenser with it , and observing throws with a galvanometer .
Another division of the microfirad was afterwards charged by 40 Clark cells , and the throws similarly observed .
The galvauometer employed was the one described by one of us in the ' Philosophica ] Magazine , ' vol. 28 , which we now call the " " ol galvanometer .
It had a sensitive .
ness of 1 scale division for ampere , with a deflexion of about 23 scale battery is very convenient when it is once got into good order , but it not suitable vork of this kind , because many days often elapse between consecutive measurements , and then , as a rule a good deal has to be done to the battery to get it into good order again .
For this reason we no longel use it .
The little storage cells are also a great source of trouble and annoyance .
A battery of storag cells , to be reliable , requires much attention and ought not to be too small .
However , as this batterJ was first set up in 1887 , and is still ( October , 1893 ) in use , perhaps we ought not to complain .
The resuil of our experience is that if we had to make such a battery again , we would attend to the following ( 1 ) The test-tubes might be advantageously replaced by strong glass cells\mdash ; of sqnare section , and nol Jess than.3 centims. on each side\mdash ; inside .
The lead plates should be at least 3 millims. thick ( 3,1 They should be formed plates\mdash ; not pasted in any way .
( 4 ) The tops of the cells and the place : where the plates bend over should be dipped in marine glue .
( 5 ) The cells should not be crowded together , so that they can only be examined by being taken out .
It is a mistake to savo spare at expense of the satisfaction of this condition .
( 6 ) The plates should be separated in each oell by a sub stantial partition of celluloid or other suitable substance , and it is better to pnt np with an iucreased resistance than to cut a Nay most of the partition for the sake of reducing it .
( 7 ) The terminal wiref should be thick and well coated with marine glue or gutta-percha .
The cells should be enclosed in a space which caIl be sbut up , so as to prevent excessive evaporation .
9 ) lead plates approach the bottom of the cells nearel than about 6 .
ON THE ELECTRICAL PROPERTIES OF SULPHUR .
visions .
Since it was described in the ' Philosophical Magazine , ' and before it made appearance in connection with this work , one of the coils developed a .
It was therefore replaced by a coil of much fewer turns and only a few hundred ohms nsistance , which we happened to have by us .
diminished sensitiveness is due to this cause .
It was rather unstable and consequently difficult to observe ; .
to this , for the sake of security we have considered 10 scale divisions , instead of one , as the limit of effective sensitiveness .
We could detect no current whatever the sulphur with this instrument and the voltage mentioned .
The resistance is therefore certainly greater than ohms at ( say ) C. And the specific resistance is greater than ohms ( say ) .
S. This will be in connection with other results .
The capacity of the condenser formed by the sulphur plates was ballistically : or rather it was compared with the capacity of the same condenser , when the film had been removed and the plates separated to the same distance by bits of quartz needles .
AB this was an operation requiring time , one of the divisions of the microfarad condenser was used as a step .
The result was that the specific inductive capacity arrived at was , which is about the mean of the values obtained by BOLTZMANN and , who probably used sulphur of this kind , i.e. , melted at a temperature not much above the melting-point , withou any special precautions to ensure solubility .
At this time we were not examining sulphur with respect to the resistance of its modifications .
and the specific inductive capacity was taken as a check in order to be sure that the plates had not come apart .
We were also anxious to see whether we could detect any sort of excessive current on chal.ging , bat the value obtained showed us that the closing current was normal .
Of course the sulphur had been frequently stressed by the repeated application of a large voltage to insure its having reached a steady state .
Tbese experiments were partly preliminary ; and we took no notice of the change of resistance on reversal , or of the action of time , for the very good reason that we got no measurable deflexion even with the large voltage applied .
On taking the plates apart the film was almost perfect , and showed no sign of staining or of any action at all .
The gilding of the , tes , however , showed a few black specks where the copper had been acted on through the ( gold .
is the only objection we have to this series of experiments .
The were afterwards confirtned so far as resist.qnce goes by the new galvanometer .
The next two series of experiments failed , owing to the spots on the plates becoming more and more and though some of the res..ults are probably unvitiated by this source of error , we have decided to leave them on one side .
So far as they went they were very variable , the variability probably depending on the action of the current itself and on the time .
To remove the source of due to the perforation of the gold film , we reluctantly decided to prepare two new plates , and after some difficulty ) tained two brass castings , which appeared to be perfect after they were shaped .
These were then gilt very heavily\mdash ; several ounces of gold must have been deposited on them\mdash ; and the gilding , burnishing , and regilding were carried out as before .
A ] iminary experiment was then made to find the sulphur had any action , when it turned out that though the spots were fewer and smaller than before , they were still present .
This induced us to put the plates on one slde , and some thick aluminium having by this time come to hand , two new plates were made of aluminium .
These plates have remained in use ever since .
They were got flat on the surface plat , , and we trusted to keeping them apart by bits of quartz thread instead of screws , a process that turned out quite satisfactory .
A large number of experiments were made on bright and scraped pieces of aluminium , with the object of finding out whether sulphur had any action on the plates during the time when and at the temperature at which they vere exposed to its action .
No trace of action having been , even the sulphur was burned on the plates , it was decided that they ] be used with confidence .
The upper )late w } provided with an aluminium handle attached to it by aluminium screws ; the handle was always removed during the electrical measurements .
The dimensions of the plates were obtained by a series of measurements with the vernier callipers , which will be set down here , as the other absolute measurements depend upon them .
These callipers had been found on a previous occasion to agree with the standard metre within very narrow limits .
Top MEASUREMENT in terms of the scale of brass callipers by ELLIOTT BROS .
Corners named , in order .
Temperature , C. Side centims. At one end At middle other end Side centims. At one end .
At middle At other end Mean Mean ON THE ELECTRICAL PROPERTIES OF PURE SULPHUR .
Side centims. At one end .
At middle . . .
At other end Mean Side centims. At one end .
At middle . . .
- !
-2367 At other end 12367 Mean breadth of top plate , centims. Mean tength of top plate , centims. of top sq .
centims. Dimensions of Bottom Plate .
Mean Mean Mean Mean Mean Mean Mean Mean Mean Mean Mean Mean Mean Mean Mean Mean Mean Mean Mean Mean Mean Mean Mean Mean Mean Mean Mean Mean Mean Mean Mean Mean Mean Mean Side centims. At one end . . .
At middle . . .
At other end Side centims. At one end . . .
At middle . . .
At other end Side centims. At one end . . .
At middle . . .
At other end Side centims. At one end . . .
At middle . . .
At other end Side centims. At one end . . .
At middle . . .
At other end Side centims. At one end . . .
At middle . . .
At other end Side centims. At one end . . .
At middle . . .
At other end Side centims. At one end . . .
At middle . . .
At other end Side centims. At one end . . .
At middle . . .
At other end Side centims. At one end . . .
At middle . . .
At other end Side centims. At one end . . .
At middle . . .
At other end Side centims. At one end . . .
At middle . . .
At other end Side centims. At one end . . .
At middle . . .
At other end Side centims. At one end . . .
At middle . . .
At other end Side centims. At one end . . .
At middle . . .
At other end Side centims. At one end . . .
At middle . . .
At other end Side centims. At one end . . .
At middle . . .
At other end Side centims. At one end . . .
At middle . . .
At other end Side centims. At one end . . .
At middle . . .
At other end Side centims. At one end . . .
At middle . . .
At other end Side centims. At one end . . .
At middle . . .
At other end Side centims. At one end . . .
At middle . . .
At other end Side centims. At one end . . .
At middle . . .
At other end Side centims. At one end . . .
At middle . . .
At other end Side centims. At one end . . .
At middle . . .
At other end Side centims. At one end . . .
At middle . . .
At other end Side centims. At one end . . .
At middle . . .
At other end Side centims. At one end . . .
At middle . . .
At other end Side centims. At one end . . .
At middle . . .
At other end Side centims. At one end . . .
At middle . . .
At other end Side centims. At one end . . .
At middle . . .
At other end Side centims. At one end . . .
At middle . . .
At other end Side centims. At one end . . .
At middle . . .
At other end Side centims. At one end . . .
At middle . . .
At other end Side centims. At one end . . .
At middle . . .
At other end Mean Side centims. At one end At middle . . .
At other end Side centims. At one end At middle . . .
At other end Side centims. At one end At middle . . .
At other end Side centims. At one end At middle . . .
At other end Side centims. At one end At middle . . .
At other end Side centims. At one end At middle . . .
At other end Side centims. At one end At middle . . .
At other end Side centims. At one end At middle . . .
At other end Side centims. At one end At middle . . .
At other end Side centims. At one end At middle . . .
At other end Side centims. At one end At middle . . .
At other end Side centims. At one end At middle . . .
At other end Side centims. At one end At middle . . .
At other end Side centims. At one end At middle . . .
At other end Side centims. At one end At middle . . .
At other end Side centims. At one end At middle . . .
At other end Side centims. At one end At middle . . .
At other end Side centims. At one end At middle . . .
At other end Side centims. At one end At middle . . .
At other end Side centims. At one end At middle . . .
At other end Side centims. At one end At middle . . .
At other end Side centims. At one end At middle . . .
At other end Side centims. At one end At middle . . .
At other end Side centims. At one end At middle . . .
At other end Side centims. At one end At middle . . .
At other end Side centims. At one end At middle . . .
At other end Side centims. At one end At middle . . .
At other end Side centims. At one end At middle . . .
At other end Side centims. At one end At middle . . .
At other end Side centims. At one end At middle . . .
At other end Side centims. At one end At middle . . .
At other end Side centims. At one end At middle . . .
At other end Side centims. At one end At middle . . .
At other end Side centims. At one end At middle . . .
At other end Mean Mean Mean Mean Mean Mean Mean Mean Mean Mean Mean Mean Mean Mean Mean Mean Mean Mean Mean Mean Mean Mean Mean Mean Mean Mean Mean Side centims. At one end . . .
At middle . . .
At other end Mean Side centims. At one end . . .
At middle At other end Mean of bottom centims. of bottom centims. of bottom centims. of bottom centims. of bottom centims. of bottom centims. of bottom centims. of bottom centims. of bottom centims. of bottom centims. of bottom centims. of bottom centims. of bottom centims. of bottom centims. of bottom centims. of bottom centims. of bottom centims. of bottom centims. of bottom centims. of bottom centims. of bottom centims. of bottom centims. of bottom centims. of bottom centims. of bottom centims. of bottom centims. of bottom centims. of bottom centims. of bottom centims. of bottom centims. of bottom centims. of bottom centims. of bottom centims. of bottom centims. of bottom centims. of bottom centims. of bottom centims. of bottom centims. of bottom centims. of bottom centims. of bottom centims. of bottom centims. of bottom centims. of bottom centims. of bottom centims. of bottom centims. of bottom centims. of bottom centims. of bottom centims. length of bottom centims. length of bottom centims. length of bottom centims. length of bottom centims. length of bottom centims. length of bottom centims. length of bottom centims. length of bottom centims. length of bottom centims. length of bottom centims. length of bottom centims. length of bottom centims. length of bottom centims. length of bottom centims. length of bottom centims. length of bottom centims. length of bottom centims. length of bottom centims. length of bottom centims. length of bottom centims. length of bottom centims. length of bottom centims. length of bottom centims. length of bottom centims. length of bottom centims. length of bottom centims. length of bottom centims. length of bottom centims. length of bottom centims. length of bottom centims. length of bottom centims. length of bottom centims. length of bottom centims. length of bottom centims. length of bottom centims. length of bottom centims. length of bottom centims. length of bottom centims. length of bottom centims. length of bottom centims. length of bottom centims. length of bottom centims. length of bottom centims. length of bottom centims. length of bottom centims. length of bottom centims. length of bottom centims. Area of bottom plate Mean area , plates Thickness of top plate .
hickness of bottom Area of bottom plate plates Thickness of top hickness of bottom Area of bottom plate plates Thickness of top hickness of bottom Area of bottom plate plates Thickness of top hickness of bottom Area of bottom plate plates Thickness of top hickness of bottom Area of bottom plate plates Thickness of top hickness of bottom Area of bottom plate plates Thickness of top hickness of bottom Area of bottom plate plates Thickness of top hickness of bottom Area of bottom plate plates Thickness of top hickness of bottom Area of bottom plate plates Thickness of top hickness of bottom Area of bottom plate plates Thickness of top hickness of bottom Area of bottom plate plates Thickness of top hickness of bottom Area of bottom plate plates Thickness of top hickness of bottom Area of bottom plate plates Thickness of top hickness of bottom Area of bottom plate plates Thickness of top hickness of bottom Area of bottom plate plates Thickness of top hickness of bottom Area of bottom plate plates Thickness of top hickness of bottom Area of bottom plate plates Thickness of top hickness of bottom Area of bottom plate plates Thickness of top hickness of bottom Area of bottom plate plates Thickness of top hickness of bottom Area of bottom plate plates Thickness of top hickness of bottom Area of bottom plate plates Thickness of top hickness of bottom Area of bottom plate plates Thickness of top hickness of bottom Area of bottom plate plates Thickness of top hickness of bottom Area of bottom plate plates Thickness of top hickness of bottom Area of bottom plate plates Thickness of top hickness of bottom Area of bottom plate plates Thickness of top hickness of bottom sq .
centims. 156.415 58 centim. The preparation and testing of these different plates lasted from till October , ON THE ELECTRICAL PROPERTIES OF PURE SULPHUR .
experiment was then made in order to reduce the surface conductivity ( if it existed ) to zero .
A number of thick rods of fused quartz were prepared , carefully cleaIJed and heated to .
when they were sufficiently cool , i.e. , when they ] just melt the sulphur , they were arranged to lie in the surface of the sulphur between the upper plate and the gold dish , and were caused to lie half embedded in the sulphur .
Now , if any surface conductivity exists in the sulphur , it will be checked where the surface of the sulphur is interrupted by the quartz\mdash ; in other words , the current would have to climb over the.exposed part of the quartz rods .
Now , the surface conductivity of the quartz is certainly very small\mdash ; indefinitely less than anything we could detect by means of the old galvanometer .
The quartz rods were intended to form a complete fence round the upper plate , and practically formed a feIlce which was very nearly perfect , being only interrupted gaps a few millimetres wide at the corners .
The result was a ] ight apparent increase of resistance , but the irregularities were so great that no certain conclusion could be drawn except the follo w conductivity observed was almost entirely , if not entirely , in the sulphur itself , and not on the surface .
This was further proved by removing the drying substance ( phosphorus pentoxide ) and .
in a CUtTent of warm air saturated with water vapour , when it gain noticed that no appreciable change of resistance took place .
We Bhall see later on that the explanation of result lies in the fact that sulphur does not appear to condense moisture so as to reduce its apparent resistance , as glass does .
With the arrangement described the conductivity of the condensed layer was smaJl compared with that of the sulphur itself This result was true whether the quartz rods were in or out .
The facts as to the conductivity of the surface of annealed sulphur were confirmed and completed at a later time .
With respect to the tests of specific inductive capacity and residual charge , we will only state that by a comparison of this film with a bit of speoially-prepared mica , we observed , for the first time , that striking absence of residual effect which 1 will be described when we deal with a film from the objections we have raised to this one .
It now became clear that we must reduce the electric stress on the sulphur if we desired be free from the risks of breaking the film .
Consequently a number of films were prepared during February , 1891 , some soluble , some a mixture of eoluble and insoluble ] phur , and these were tested with lower voltages .
In one cane the film was cooled by carbon dioxide snow .
All these hlnls went wrong .
In a thin one the plates were found to touch at one corner .
One insoluble film was \amp ; om having caught fire .
The film cooled by carbon dioxide snow was got so cold that it became wet round the edges with condensed moisture , and some of the snow had been in contact with the , and aces of grease were suspected ia the snow .
Oue good soluble film gave no conductivity at all , but our galvanotneter was not sensitive enougb with the reduced voltage to make the negative result worth anything .
A mixed film showed no conductivity til ] it was heated nearly to melting , when it began to conduct at C. , and the conductivityr then diminished to about 1 C. , when it increased rapidly as the sulphur melted .
Some films separated from the plates on heating .
The result of all this was to show us the necessity for a good many small precautions ; and induce us to concentrate our efforts on the new galvanometer .
This work , which has already been described , occupied us until September , 1892 , when the new instrument was got into good working order and the investigation of the sulphur proceeded with .
Preliminary experime.nts , therefore , took nearly two years , tbough the work was not continuous , and we blundered a good dea ] .
With increased galvanometer sensitiveness came increased difficulty from want of thermoelectric ects , \amp ; c. These took some time and trouble to overcome ; and , as our experienca and the sensitiveness of the galvanometer increased , we were continually obliged to repeat our observations .
The film was the subject of experiment until October 28 , 1892 , at which time , the plates having come apart , the thickness of the film was measured by a large spherometer .
The threads having been measured before they were put between the plates , and previous work having shown that the plates ested exactly on the threads , it was imagined that the thickness of the film would be the same as the diameter of the threads .
This , however , turned out to be an erroneous assumption , for the spherometer showed that the tbickness varied slightly from point to point , and was more than 1 per cent. greater than the diameter of the threads .
In the progress of the work it was necessary to reduce most of the results with wrong ( i.e. , the supposed thickness of the film ) , but they have been recalculated for this paper .
The exact statement of the preparation of the film follows .
Preparation Properties of Film of September 14 , 1892 .
The plates having been scrupulousl cleaned , as well as the pure gold dish , some twice distilled sulphur ( Chance ) was melted at a temperature of in the large oil bath .
This oil bath was eally a large oil oven , heated by two long perforated Bunsen gas tubes .
It took about four hours to heat to , and was a proportio1lally long time in cooling ; it was not the same oil bath that was used for heating the sulphur when examination .
Eight threads , each broken off to a diameter of 24 millim. , were placed on the lower plate .
These threads were about 1 to 2 centims. long , and their diameter was gauged by a BROWN and SHARP vernier callipers .
The upper plate was adjusted to the } , care being taken avoid the presence of air bubbles by lowering the plate from one end like a$ mioroscope cover-slip , and at the same time the plate was lowered gently enough not to wash out the threads .
The sulphur stood in the gold dish nearly level with the top of the upper plate .
After the plates were adjusted as perfectly as possible ( examination .
afterwards showed that the adjustment was practioally perfect , the whole arra geON THE ELECTRICAL PROPERTIES OF PURE SULPHUR .
meant was put back into the oil bath , again heated to , after which the temperature might be considered uniform .
We desired to avoid distorting the plates by having any want of uniformity in the temperature .
The bath was then allowed to cool slowly to C. , when the sulphur crystallized , but was still soft .
The temperature was then raised to 11 C. , kept at this for about an hour , and the gas then turned off , and the bath allowed to cool slowly .
It takes six or seven hours to reach the temperature of , say , C. In order to have a check on the nature of the sulphur employed , two auxiliary aluminium plates were prepared at the same time , nd furnished with a film of sulphur , lder identical conditions with the actual experimental plates .
The next day these plates were forced apart , and about 1 gram weight of the sulphur tested by carbon bisulphide .
It was found that the Phur was all but absolutely soluble .
We would not like to say it was not perfectly soluble , for the minute particles left could only be seen when the test-tube was in bright , and they might have been merely dust motes .
However , our impression was that a trace of insoluble sulphur was present ; it could not have amounted to more than about01 milligram , so that the film may be considered to contain , say , 99,999 parts of soluble sulphur out of 100,000 .
After the observations were complete , the plates taken apart on October 25 , it was found that 3 sq .
millims. ought to be allowed for the space taken up by the bubbles .
The film appeared to be perfectly soluble , nothing in the way of residue oould be detected , but only about 1 decigram was available for testing .
The melting point was taken five times , and found to be from C. to C. , consequently the sulphur was prismatic .
A microscopic investigation showed that there were larger crystals present in the film cemented together by smaller ones , or , possibly , by uncrystalline matter .
We were so fortunate as to secure the assistance of an expert mineralogist , Professor DAVID , in the microscopic examination of the film .
The investigation of the film at the end of the experiments , so kindly undertaken by Professor DAVID , involved grinding pieces of the film down to very fine slices , in the same manner as rock sections are ground down , and subsequent examination by the microscope .
The results were as follows:\mdash ; 1 .
) The greater portion of the film , probably the whole , is crystalline and anisotropic .
( 2 .
) Most of the crystals , but not all , have their principal axes lying in the plane of the film .
( 8 .
) So far as could be seen , most of the cr stals appear to have their principal axes paralleL ( 4 .
) The question as to whether all the crystals were monoclinic was investigated with great care by observing the extinction angles , and comparing the action of the film with that of fresh and aged monoclinic crystals .
The result was , that the crystals had certainly ceased to be monoclinic , and within the limit of observation were entirely octahedral .
The ' extinction angle\ldquo ; of monoclinic sulphur is at to the long axis of the cl.ystal , in the case of the film sulphur the extinction angle had changed to about with the original monoclinic axis .
As the extinction orientation of the supposed octahedra is ur\amp ; nown , and the crystals are too small to observe singly , no further conclusion can be obtained .
We desire to express our thanks to Professor DAVID for the foregoing uable addition to our knowledge of the structure of this film .
It clearly consists either of a of sulphur intermediate between monoclinic and rhombic sulphur with extinction angles of or , and which still preserves the meItin -point of prismatic sulphur , or , on the other hano1 , it may perhaps turn out that the octahedra formed in this way are merely very unstable and return to the monoclinic state so rapidly that in melting-point experiments the reconversion always occurs before the melting can take place .
Aged monoclinic crystals of sulphur from various sources were found to behave in the same way .
Against the instability theory we have the fact that the melting-point was the same , whatever the initial temperature of the glycerine bath .
During the electrical observations the film was probably partly normal mono- clinic and partly ' aged\ldquo ; monoclinic .
No quartz rods were used to reduce the surface conduction .
The film was dried for some days in position in the oil bath by sulphuric acid and phosphorus pentoxide .
The colour of the sulphur was a pure lemon-yellow .
Most area of film , sq .
sq .
centims. Thickness of Film.\mdash ; The threads were slightly conical and broken , so that the thickest parts were centim. in diameter , as has been stated .
When the plates were taken apart , a series of measurements of the film by the spherometer were made .
Bits of the film were taken from each of the sides near the edge , and a bit from the middle .
The bit to be measured was placed on the spherometer plate in a marked position\mdash ; with careful dusbing\mdash ; and then covered by a small square of plane parallel glass , not larger than the fragment of film employed .
The spherometer was well oiled and got into a steady state of temperature by handling before the zero obtained\mdash ; precautions very necessary , but generally omitted .
Ten settings were taken on each portion of the film .
The five means of the readings were , in inches , '337840 Grand mean 3371898 ON THE RICAL PROPERI'IES OF PURE SULPHUR .
setting on bit of glass .
10 observations , 346773 .
Difference , 0095832 inch centim. thickness of film .
The differences probably due to imperfect sting of the bits of film .
The film was very crumbly and the oken bits were strongly electrified and difficult to remove .
This measurement of film thickness is the most unsafisfactory part of the work .
In calculating specific resistance and specific inductive capacity , the ratio of thickness to area has to be known .
The area has to be corrected for the effect of the edges .
In the case of the electrostatic measurement , i.e. , the determination of the specific inductive capacity , correction to the area can be calculated as already mentioned- , and may be taken at centim. In view of the specific resistance being dependent on so many different factors , it was not worth while to try to calculate a separate correction , so that the corrected area has been used both for the calculation of specific inductive capacity and specific nce .
This area is or .
centims. The ratio of corrected area to thickness is , therefore , The arrangement of apparatus adopted for making the nent of the resistance of this will be understood from he diagram .
Two days experimenting were absorbed in curing leaks , conbact effects , \amp ; c. The .gaJvanometer was simply used with a lamp and scale .
A BCD ( fig. 1 ) is a reversing key , put to earth at E. is a contact key .
are insulating supports of glass and ebonite .
The terminals of the battery are reversed when necessary by interchanging the wires .
A key , not shown , is provided for connecting A and or and , so as to cut out the galvanometer .
Before and after making an observation the wire is disconnected from the top plate at , and hung on the support at T. Nothing touches the upper plate except the sulphur and connecting wire .
Consequently , if no deflexion occurs when the wire is disconnected from and hung on , even .
when the reversing key is put over , it shows that the insulation is sufficient .
divisions .
Finally , we decided that the ] lay , as was to be expected , in the smaller divisions of our microfarad ; we were therefore obliged to make use of the larger divisions .
This necessitated the finding of a way to conveniently bridge the great difference in the capacity of the Jlphur condenser and that of 2 microfarad , the smallest division available .
We finally arrived at what we found to be a very satisfactory method .
prelimioary expel.lmen owe , us ng showed that , usin a certain number of Clark cells , we obtained , by simple charge and discharge of the condenser through the galvanometer , a convenient deflexion .
The comparison was then made by subdividing the voltage of one Clark cell and charging and discharging the selected division of the microfarad with it until a nearly equal deflexion was observed .
This avoids the necessity for determining the logarithmic decrement .
The voltage was divided by using points of derivation in a circuit of 100,000 ohms , including a Clark cell .
The capacity of the key and connections was found to be very nearly two hun re dredths ( .O2 ) of the combined capacity of the key , condenser , and connections .
This correction is amply accurate enough for the results got by the ballistic method , and is rather too large\mdash ; uncertainty due to large absorption in the standard prevents really good comparisons being made .
There being a difference between the divisions of the microfarad , one em one of th was selected as a standard pending calibration when our standard shall arrive .
Let be the first elongation of the galvanometer needle .
Let be the capacity of the condenser .
Let be the potential difference of the coatings in terms of the Clark cell .
Let be the ballistic corJstant of the galvanometer .
diminated , ON ELECTRICAL PROPERTIES OF PURE SULPHUR .
109 Then when the deflexion is small and the logarithmic decrement can be neglected , or CV We considered that the best way of obtaining a mean value in terms of the divisions of the microfarad would be to take several sets of observations , calculate kom each and then use this mean value to enable us to find the capacity of the condenser .
The results are as below .
.\mdash ; Values of ( Galvanometer constant ) .
Mean value of The observations on the sulphur were repeated with 21 Clark cells , as it was vanometer sensitiveness hhanged since twhenthe nbtained.divisionsHence .
C.G.S. Corrected capacity , therefore , is If A be the " " effective area\ldquo ; of the ; , , thickness of the film ; ratio nits ; specific inductive capaoity ; ; Hence PROFESSOR R. THRELFALL AND MR. J. H. D. ; ; ; a number very like that previously obtained for the same kind of film .
Having obtained a provisional value for the specific inductive capacity of the sulphur , we considered that a suitable time had come to repeat some experiments made a year before on the residual effect , and endeavour to obtain comparison between sulphur and mica , the latter having been much recommended by M. BOUTY ( ' vol. 110 , p. 1362 ; 'C .
vol. 112 , p. 931 ; Journ. de Phys vo ] .
9 , p. 288 , June , 1890 ; 'Ann .
de Chim .
et de Phys. ' [ 6 ] , vol. 24 , p. 394 ) .
We propose to describe our experiments first , and then to discuss this very important matter ( the literature of which is overwhelming ) when we have described a mber of nate phenomena .
A preliminary experiment was made to find the capacity of the quadrantls of our electrometer and of the key and connections used in the investigation of the resid al The capacity of the quadrants was rather less than microfarad , and of the key and connectious about microfarad .
The sulphur plates have , as has been shown , a capacity of about microfarad .
The smallness of the capacity of the quadrants and key and leads , in comparison with the capacity of the condenser , enables us to disregard them in estimating the residual effect .
The mica condenser was prepared from a specially selected bit of mica , such as is used for covering photographs ; it was silvered on both sides , and cleaned alon.g the edges as described by BOUTY , the edges being lightly varnished by dipping in a dilute filtered solution of shellac .
This condenser plate was dried at the temperature of the water bath for twel hours , and was then mounted on quartz insulators in a desiccator over sulphuric acid .
It had been in this for a year when our experiments began .
Connection was made to the insulated plate by means of an aluminium wire on a fused quartz support .
The dimensions of the mica condenser were : Length of rectangular silver films centims. Breadth Thickness of mica . .
005 centim. The capacity was of the order of that of the sulphur condenser .
We will not describe any preliminary experiments on residual effect , because there was nothing novel either in our method or in the difficulties we encountered .
The investigation of residual effect is not difficult if the effect be large , but if it be vanishingly small , as it was with the sulphur , if it be required to assign a limit to this small effect , ON THE ELECTRICAL PROPERT[ES OF PURE SULPHUR .
then numerous difficulties arise which are to be got over by electrostatic screening , $voidance of contact forces , insulation , \amp ; c. The following diagram shows our arrangements :\mdash ; Fig. 2 .
, are the six points of a reversing key .
The points and insulated by supports of fused quartz .
The condenser is inserted at S. GH is a pillar make-and-break key with mercury cups ( see Plate 3 , fig. { 3 ) .
Between and ae placed.a wire of aluminium , so arranged that.when the rocking key is thrown over charging position A B to its discharging position B , this wire momentarily touches part of the rocker connected with , the other end being :ermanently conrJected at E. In this way the discharge can be made to occupy a very time ; say 01 sec. , and includes the key .
A short-circuiting key is sometimes kept in and till the discharge has taken place .
During the and are connected .
HA is a long wire enabling the battery .
\amp ; c. , to be placed far :nough away not to complicate matters by direct electrostatic induction .
As soon as the rocker is moved over far enough to be insulated from , this point is earthed and he rocker then moved over to , the short-circuiting key being removed after ischarge has taken place , and , of course , before the rocker makes contact at or F. one time an additional key was used for short of the electometer , but it was given up as merely increasing the trouble of insulation .
The electrometer *as adjusted to a sensitiveness of about 40 scale divisions per Clark cell , and this id not vary by more than 2 divisions over the scale .
Thermo-electric effects , \amp ; c. , prevented the sulphur from ever giving absolutely no deflexion after the removal of charge .
About two hundred } ) eriments were made on } sulphur and micB , hnd we know nothing whatever against the summary , which we consider to be a ufficient statement .
A2 PROFESSOR R. THRELFALL AND MR. J. H. D. BREARLEY .
went tcharge whout 1ccount fxperiment ceptember 2adIt oticed fabove tcbarge iosure teven weated , ffect increased .
been heated .
BOTJTY attributes the residual charge almost entirely to the action of the edges , which are varnished , and we are inclined to agree with this , but we do not how the polarisation can be galvanic in the ordinary sense , for in our case it must B$ have amounted to more than 10 volts .
We think it is due to a creeping of the charge NICOLL , Phil. .
The sulphur is really phenomenal .
We are inclined to attribute most of the small effects observed to the residual sources of uncertainty , contact action , imperfect shielding , imperfect insulation , imperfect prevention of the creeping of the charge , in spite of these effects having been eliminated apparently before the sulphur condenser was inserted .
Because these uncertainties are suggested , it must not be thought that they were overlooked at the time .
We believe that everything was done that could be done , and that we are , in fact , at about the limit of the applicability of these methods .
However , taking the numbers as they stand , they are sufficiently remarkable , and show that the residual charge , it exists , is less than , say , four paarts in ten thousand of the original even when the duration of charge amounts to ten minutes .
In our previous experiments using the film between the ] gilded plates , we made use of enormously greater voltages derived from an electrophorus , but some uncertainty arose from the possible parting of the plates , and from the burning of the film where the spark had jumped across .
However , we never detected any residual eftect at all with the ffim in question , under circumstances where the residual chal from the amounted to more than 50 divisions ; this was with a voltage estimated from the spark length at about 5000 volts .
: The electric strength of sulphur from the broken film is in the neighbourhood of and 'probably greater than 730 volts per millim. , or at the rate of about 3300 volts per millim. , or 33,000 volts per centimetre , but too much weight must not be attached to this , for the influence of a small air bubble in disturbin the field might be very considerable , and would reduce the strength very materially .
We defer the discussion of these results till the similar data for other films have been dealt with .
The residual charge having been found to be so small , rather greater interest attached to the specific inductive capacity of the film , which had to be taken again in an case in order to give us security as to the permanence of the contact between the plates and the sulphur .
As the results of a complete series of very careful experiments by the ballistic method , usin one division of the microfarad only , the one chosen as a provisional standard , the capacity of the sulphur was found to be\mdash ; mococxcvI .
( uncorrected ) .
( corrected for standard ) .
A comparison between condenser and one supplied by Messrs. CLARK , MUIRHEAD , and Co. , may now be given ; it was made on Jan. 24th , 1893 , at a temperature of C. , by means first of the old , and afterwards of the new galvanometer , with 40 Clark cells by DE SAUTY'S method .
The leakage through the two condensers was measured , and found to be so small as only to affect the result in the fifth place ; it was therefore neglected .
We used two 100000 ohm boxes , and a dial box of 10000 ohms , all by ELLIOTT , but the divisions were not compared with our standards : Dr. MUIRHEAD'S value for the microfarad was microfarad at C. As we do not know whether the mica plates are shellacked or only paraffined , we cannot apply any temperature correction\mdash ; in any case the two condensers probably have about the same temperature coefficients .
The result of the companson was that with 100000 out in one branch , the balance lay between 98815 and 98800 in the ot , her .
Taking 98808 as the mean , the capacity ofour .
division for instantaneous charge is 20186 microfarad , or say 1 per cent. too large .
For the future this value will be employed .
The difference between this value and that formerly obtained is to be ascribed chiefly to the difference between the standard selected and the mean of the three divisions taken before as a standard .
It was thought desirable to check this result by another method , both for the lrpose of obtaining independent evidence , and because the istic method takes so long to carry out that it is unsuited for measuring .
the capacity at any given temperature .
The well-known method of DE SAUTY was selected after consideration , the sensitiveness of our galvanometers giving us great advantages in carrying it out .
By observing the initial kick we also hoped to get the instantaneo'us capacity comparison ( see GLAZEBROOK , ' B.A. Report , ' Leeds , 1890 , p. 102 ) .
By using 40 Clark cells , a balance for instantaneous charge was given by 100000 and 1065 ohms out in the two branches , one a dial box , the other from an ordinary box , both by ELLIOTT , but not compared with our standards .
final corrected result is The sensitiveness of the g.alvanometer having been increased , as well as our experience in using it , it was thought worth while to determine the resistance again for the purpose of obtaining a higher limit .
Date , October 1 , 1892\mdash ; Voltage , 304 volts .
Sensitiveness of galvanometer , ampere per\ldquo ; ] ) tooth of the .
micro .
meter .
ON CTRICAL PROPERTIES OF PURE SULPHUR .
Limit of certain discrimination , owing to thermal effects , was about '25 tooth only , but it is pu6 at 5 tooth for certainty .
Specific resistance is greater than C.G.S. ( The capacity was taken again just after the test , and found to be the same as before .
) In order to check this , it was made the subject of experiment by the elsctrometer method .
At first we intended to experiment with different voltages in order to test the sulphur with respect to its law of conduction , and made a number of experiments on the lines of the experiments of THOMSON and NEWALL ( ' Proc. R.S. , ' vol. 42 , p. 410 , 1887 ) .
For reasons formerly given , however , we ended by observing the leak during a given time , and then assuming Ohm 's law , we will not enlarge on our troubles in this work , for they must have been such as would come to anyone trying to get at a very high resistance by this method .
We will , however , ask the reader to believe that all such sources of error as are introduced by incorreot knowledge of capacities and electrometer law\mdash ; imperfect insulation , \amp ; c. , were attended to by us with great care , and the results we offer are , we believe , free from objection .
Mr. LOCK was kind enough to check the algebraical and arithmetical work for us , and as he used no approximations , we give his results rather than ours\mdash ; with which they agreed within the hmit of approximation adopted by us .
From the results , the specific resistance 01 the sulphur is as follows , neglecting fractions:\mdash ; 1 .
Oct. 6 , 1892 Specific C.G.S. 2 .
Oct. 7 , 1892 Specific Resistance , C.G.S. 8 .
Oct. 7 , 1892 Specifc Resistance , C.G.S. These results were obtained with a volt , of about six Clark cells .
All these observations were taken on sulphur after it had been in its silvered brass box with phosphorus pentoxide for at least twelve hours .
The effect of exposing the sulphur to air was tried on one occasion , when the resistance fell at once ( i.e. , in ten minutes ) about one thousand-fold , and seemed to stay there .
After replacing the lid of the box and for about twenty minutes ; the resistance had again increased to six times its value when exposed freely to air .
This explains partly how it was that we got only a small ] or negligible effect by exposing to moist air when we used the galvanometer .
At the time the plates were expoBed to air for ten minutes , as just recorded , the thermometer wet bulb was at F. , and the dry bulb at F.\mdash ; a rather damp day .
Nearly all the reduction in resistance took place during the first few minutes of exposure , but the method does not admit of this matter being treated satisfactorily , for it takes too long to get a good measurement .
influence of moisture seems to be to reduce the resistance of the sulphur condenser , but not to such a point as would enable the conductivity to be detected in any measurements hitherto by means of the galvanometer .
It is possible that the small conductivity observed by the electrometer may be wholly due to residual surface action , though later experiments lead us to think that this is not the case .
The capacity test was applied , giving the same result as before , showing that the plates are still in position .
It was decided to heat the phur condenser , and observe change of resistance and capacity as the temperature rose .
The galvanometer was now more sensitive than before , giving one micrometer division of double elongation for ampSre .
With 291 volts no effeot could be detected .
The result is to place the specific resistance above C.G.S. The galvanometel now approaches the electromete in its power of discrimi- .
nating small conductivities .
The temperature was O. The temperature of the bath was then raised to C. , at a rate of about minute , and was kept wit , hin , degree on either side for about an hour while the resistance and capacity tests were .
made .
The capacity appeared to be exactly the same as before \mdash ; i.e. , it did not change by as much as 3 per cent. , or we should have detected its variation .
It will be shown later that this probably indicates an appreoiable , but .
small , positive temperature coefficient of the specific inductive capacity .
The resistance test also failed to show any signs of conduction , and the result may be oonsidered to be the same as at C. The temperature was then raised to C. , and kept there for several hours .
slight conductivity .
showed- itself , corresponding to a deflex of divisions , with a voltage of about 285 volts .
The specific resistance is now C.G.S. After the battery had been the sulphur for about an hour and a half at the temperature of the deflexion fell from to abont 5 divisions , showing that the resistance had increased , say , four times by the continued application of electric stress .
This being the first clear case of conductivity we had encountered , we were anxious to note any phenomenon which might present itself The most striking effec was the discontinuity of the conduction .
At first we thought that the sudden sharp kicks gi every now and then by the galvanometer were.due to accidental causes of some kind , but it soon became clear that the phenome..non was a real one .
When the voltage is first applied the discontinuity is more marked than when it has been on for some time .
In the case we have before us the galvanometer would take up a definite position of deflexion , and every two or three minutes this would suddenly increase or diminish by a large amoimt .
The only possible of estimating the conductivity is , therefore , to take a long series of observations and obtain a mean value , and that what was done above .
The impression forced on one while observing at the galvanometer is that a succession of irregular changes are ing place ; for instance , that groups of molecules , one after the other , succumb the electrio stress , and signaJise their destruction by allowing a current to pass .
The conductivity in these cases , however , never falls zero .
It is , perhaps , not too much to .
say that all the observations accordanee with a theory that there is a small regular and continuous conduction , superimposed on which there is something like a disruptive conduction .
We have since discovered that the discontinuous ON THE ELECTRICAL RTIES OF SULPHUR .
conduction occurs also when the conduction is superficial , but that it is a real phenomenon of the conduction through sulphur as well .
The temperature was constant between and C. The capacity appears to be absolutely the same as before , i.e. , within3 per cent. The temperature raised to C. There now appeared to be a slight fall in capacity , say per cent. , but it was so small as to be quite uncertain .
The following lesistance results were obtained:\mdash ; Double deflexion of divisions at C. Six experiments .
, , C. Four experiments .
, , , , , .
Six eipperiments .
, , C. Five expe , riments .
\ldquo ; The sensitiveness of the galvanometer had been reduced to one division for ampere for the sake of rapidity of observation .
The specific resistance ( mean of above ) is about C.G.S. , but we lay no stress on it , for on applying the capacity test it was found that the plates had come apart .
Th6 results with this film may now be summarized as follows:\mdash ; Film of September 14 , 1892 .
Thickness , 02431 centim. ( effective ) , square centims. Kind of sulphur , either perfectly soluble , or containing less than 1 part in 100,000 insoluble .
, monoclinic ( aged monoclinic ) and octahedral .
Optic axis in plane of film .
Melting point , C. inductive at C. , corrected in terms of Dr. MuIRHIiiAD'S standards .
By ballistic method , By instantaneous DE SAUTY method , .
The latter value to be preferred ; capacity constant up to C. , indicating small temperature increase of specific inductive capacity .
These values were obtained with a voltage of 40 Clark cells .
Residual charge with 300 volts , after charging one minute and discharging one minute , 03 per cent. , say .
Specific resistance at C. by electrometer method , six Clark 's , say O.G.S. Possibly entirely surface action .
Exposed to damp air , resistance diminished in our apparatus 1000 fold .
Specific resistance by galvanometer method with a voltage of 291 volts greater than C , G , S , from to , At a temperature of , the specific I18 PROFESSOR R. THRELFALL AND MR. J. H. D. BREARLEY resistance , after the current was on for , say , ten minutes , is C. It was probably a case of discontinuous conduction , and the resistance increased about fourfold during an hour 's application of the voltage .
The voltage not reversed on the sulphur during these experiments , as such a course complicates matters enormously .
Film of October 28 , 1892 .
As the experiments just described broke down at a temperature of between and 10 C. , it was necessary to make them over again , in order to find how soluble sulphur behaves when heated up to the melting-point .
For this purpose the same sulphur was re-melted at a maximum temperature of 14 C. , and as there , in our experience , absolutely no probability of our being able to anneal the film and then re-heat it to the melting-point without disturbing its contact with the , we decided to measure its resistance as the temperature fell .
The film was prepared on October 28 ; it was annealed during the afternoon , and the following observations taken during the night when tram traffic was suspended and externat disturbance less to be feared .
The resistances were such that the odd galvanometer was most conveniently ] oyed ; it was read by a lamp and scale .
The temperature fell from .
to in forty minutes .
The resistance was so low that only one Clark cell ( the large master cell ) was used .
During an hour the temperature was kept between and 13 5 , and the resistance gradually increased , so that the deflexion fell off from 396 to 354 divisions in half-an-hour .
On revel:sing the battery at a rather lower temperature , 13 85 , the deflexion at once increased to 548 , and in the course of about twenty minutes fell to 423 .
These results are typical .
The film probably contained some insoluble sulphur at this temperature ; when it was examined afterwards , several naIyses showed a mean content per ceut .
of insoluble sulphur and a melting-point of from to 11 66 C. , but this is no criterion of the actual composition of the film at this stage .
The corrected area was square oentilns .
, and the thickness was centin ] .
We will not trouble the reader with such a full account as in the previous case , for the proced .
ur was exactly the same ; the table , therefore , be rather fuller .
At a mean temperature of the mean specific resis with the battery on both ways was C.G.S. , which is not much less than the resistance after an infinitely .
long application of voltage .
The capacity at this temperature could only be taken very approximately , to the leak ; the result was a specific inductive capacity of .
We believe this is not far from the truth , but lay no stress on it .
We will give the full set of observations of resistance at the next tempelature examined , C. , partly to illustrate the effect of , time and reversal , and partly to enable anyone to form an opinion as to the actual effects obtained .
The differences : are really greater than they appear , for the ratio of twsuccessive swings with this galvanometer was , as deduced from twelve experiments with elongations of ON THE LECTRICAL PROPERTI ] OF PURE SULPHUR .
same amplitude as occur in the table .
The corresponding logarithmic decrement is about .
The galvanometer sensitiveness is per scale division ( ELLIOTTS , at about 1 metre ) of double elongation .
We have to thank Mr. POLLOCK for assistance in making all the observations with this film .
Time , 8 hours 55 minutes .
Charged for three minutes , terminal of battery to the upper plate .
Clark .
Temperature C. Galvanometer reversed .
D. deflection 338 .
, , , , \ldquo ; .
, , 318 , , , .
, , .
, , , , .
, , 306 C. , , , , B-A .
, , 310 hours 14 minutes .
Battery reversed .
Temperature C. Galvanometer reversed .
D. deflection 414 B-A .
, 392 , 12 C. Time 9 mins .
Galv .
rev . .
, , 392 B-A .
, , 372 , \ldquo ; A-B .
, , 363 , , , 9 25 , , , , B-A .
, , 348 , Mean of all deflections Mean Specific Resistance at say 12 5 .
The capacity could not be taken .
An attempt was made to detect residual polarization by the galvanometer\mdash ; but without effect\mdash ; the voltage to be expected being so small as to compare with the thermoelectric effects unavoidable with one part of the circuit so much hotter than the rest .
We shall return to this .
Observations showing an increasing resistance were made at various points down to C. , at which point the sulphur evidently began to crystallize , for the resistance increased rapidly that we observed best by noting the rate at which the light spot moved over the scale towards the zero .
The resistance increased meanwhile about twelvefold .
Nothing is to be got by noting the rate of ohange of resistance with change of temperature , unless we had actually had a rapid thermometer between the plates .
Our thermometer was merely in the imuer box with the sulphur , and did not indicate the temperature-variation of the sulphur to be expected while it was freezing .
The summary of the changes that occur is as follows ( see table for numerical details ) .
( 1 .
) At the resistance is considerably increased .
( 2 .
) Between and the rapid change begins .
( 3 .
) Between and the velocity of the change increases .
Fig. 3 .
When the deflection had fallen to 12 divisions , it remained constant .
rIhese results probably chiefly refer to the time of crystallization of the sulphur , which , as everyone knows ( GERNEZ , .
cit is a process requiring a good deal of time .
The ON THE ELECTRICAL PROPERTIES OF PURE SULPHUR .
sulphur was again heated and the resistance diminished steadily , til ] at the spobegan to move at a .te of 10 divisions per second , and went on faster and ffiter , till after three minutes it had gone off the scale .
cooling , the resistance was undoubtedly permanently reduced\mdash ; say tenfold ; it increased , as before , down to a temperature of C. : when the sulphur crystallized and the resistance suddenly increased .
A curve been drawn ( fig. 3 ) showing the variation of resistance with temperature , the observations being for the most part made with the voltage each way ; at the points of rapid change this was not possible .
The hysteresis of diagram is partly due to the time required during crystallization .
The rate of variation of temperature at important points was never more than one in five minutes .
The permanent reduction of specific resistance by heating was observed again with another film at a later date , and appears to be a natural phenomenon .
.\mdash ; Showing Resistance of the Film of Octoher 28th , 1892 , at Different Temperatures .
Temperatures .
Specific resistance in C.G.S. units .
C. falling rising .
C. falling rising .
3 .
Above this the resistance fell as shown on the culve till about when the sulphur melted .
The two following relations are taken from the curve:\mdash ; At C. rising On cooling again at . .
At C. rising On cooling at . .
( uncertain ) change very rapid * We provisionally attribute the fall of resistance to a change in amount of insoluble sulphur present , probably , almost certainly , an increase .
It may be objected that we ought have melted our film at a lower temperature , and thus prevented the formation of insoluble sulphur at all .
The reply to this is , that unhappily one cannot adjust the plates properly unless one has some time for the operation , and the cooling surface of the plates is great that sufficient time is not afforded if the sulphur is nearer to its point to begin with .
Film of November 9th , 1892 .
It was now necessary to examine the behaviour of a film rapidly cooled so as to form a mixture of soluble and insoluble sulphur .
Our first attempt failed , owing to * Further research shows that this provisional conclusion must be modified .
water getting on the film , the second attempt succeeded .
The sulphur was heated to a temperature of about to and kept there for fifteen minutes ; the cooling was effected as rapidly as possible by water , consistently with not allowing any water to touch the sulphur .
Some portions of the suJphur wele cooled more rapidly than others , and consequently the percentage of insoluble sulphur present varied in samples taken from different parts of the dish .
The film itself probably contained about per cent. of insoluble sulphur , but this estimate is based on analysis of sulphur eut awa from the sides of the plates close to the film .
The film itself was heated in the course of the experimental work , which rendered any conclusion from the analysis in its final state inapplicable .
The melting-point of the sulphur , after it had been between the plates as the subject of experiment , was from to C. The area of the film ( corrected for bubbles , \amp ; c. ) sq .
centims. ; the thickness of the film , by spherometer , was02718 centim. As the specific inductive capacity appeared very different in this case from the values found previously for soluble sulphur , and as the thickness of the film is the most critical measurement in determining the value , we will give the means of four groups of ten settings , each on pieces taken from different parts of the film frotll which the thickness was deduced .
They are Mean , 335985 inch .
Ten settings on glass .
Mean , 346711 .
Difference , 010726 inch centim. We consider that the greatest possible error of any group of settings cannot amount to more than 4 per cent. , or the film is clearly not quite regular .
A measurement made by focussing with a Zeiss microscope gave about centim. , but it is only of as a rough check .
It will be safe to consider the mean thickness as correct , within , say , 10 per cent. , at least .
The sulphur ( as soon as the plates were sufficiently cooled by placing the outer dish in water and putting wet filter paper on the upper plate ) was set up in the oil bath with drying material .
A rough test at a temperatule of C. , voltage made with the galvanometer , gave\mdash ; Specific C.G.S , ON THE ELECTRICAL PROPERTIES OF PURE SULPHUR .
No marked difference was observed between the conductivity when the battery was first on , and five minutes afterwards , but the resistance slowly increased as usual .
The capacity was measured by the DE SAUTY method , using the small storage cells with a voltage of volts .
The instantaneous balance was reached with between 1070 and 1072 out in one arm , and out in the other .
The resulting capacity was C.G.S. Applying a correction of \mdash ; 2 per cent. for the key and per cent. for the value of the standard , or , in all , \mdash ; 1 per cent. , this becomes 00212 , say .
The corrected value of the specific inductive capacity is .
We can only explain this large value by that the film measurement is at fault , or that the specific inductive capacity of mixed sulphur is really high .
The film measurements are given in order to aJlow the reader to form his own opinion on this point .
The sulphur was kept in the drying box for 24 at a temperature of about C. The specific resigtance C.G.S. So it has not changed.mucb by continued dryin ; the battery was not reversed , and the mean deflection was 57 divisions .
After the battery had been applied for about one hour the double deflection fell to 17 divisions .
The voltage was 264 volts .
The current having become fairly constant , it was thought advisable to investigate whether the deflection was due to surface conductivity .
From previous experiments it is known that the sulphur surface conductivity , after a day 's drying , will not make the apparent specific resistance less than of the order of C.G.S. , but this was with highly soluble sulphur .
It is not likely that the admixture of insolub]e sulphur will increase the surface conductivity , but it is possible that it may do so , and in that case we should expect that exposure damp air would reduce the apparent resistance .
The lid was therefore taken off the , and the plates freely exposed to the air for ten minutes .
Dry bulb , F. ; wet bulb , F. No change having been detected , a current of air which had been passed through a reversed wash-bottle with the water at C. , was allowed to play for about five minutes on the sulphur between the upper plate and the dish .
The deflexion or reversal of the galvanometer remained precisely the same as bsfore .
The conclusion is that the conductivity is probably almost wholly due to the sulphur , and that the latter is not affected by damp air to anything like the same extent as glass is .
This is confirmed by some observations to be described , in which , when the sulphur plates are heated in a closed box with drying material to C. , the conductivity is increased , whereas , if it were due to surface action , the opposite effect would be expected .
The sensitiveness of the galvanometer ( new ) , when these experiments were made , was about ampere per micrometer scale division , and the period was eleven seconds .
This .
was the steadiest day we ever had , the galvanometer could .
have been read to 2 division perfectly , had it been necessary .
The next day was almost equally steady , and we had a good example uf the effect of conductivity , which was further shown by the galvanometer needle kicking slightly ffom time to time .
The following table will indicate the effect we mean ; we ' oftemr observed it , but on this occasion the magnetic conditions were so steady that we } to be quite certain that the effect is not to be expIained by irregularity .
When taking the sensitiveness of the instrument , within a few minutes , not a of such an effect was to be observed , and the insulation of the wire , \amp ; c. , was found to : be perfect,.as far as our methods would show .
TABLE showing Irregular Conductivity of Sulphur Film containing a Mixture , of Soluble and Insoluble Date , Nov. 11 , 1892 .
Voltage , Sensitiveness , amperes .
Temperature of air , sulphur , C. After the voltage has been applied long enough for the sulphur to take up a steady state ( say 20 minutes ) , double tion , 6 divisions .
B-A , , , .
, 18 A-B , , , 6 B-A , , , , , 6 B-A , , , , , 16 A-B , , , , , 33 The capacity was found to be exactly the same as before , and the .
platesi regarded as a condenser gave a bright spark on discharge .
This irregular conduction ( though not always so well-marked ) was generally observed , and may be considered as a natural phenomenon .
The sulphur was then tested for residual charge in the manner formerly described .
Sensitiveness of electrometer was 41 divisions per Clark cell : the insulation of the apparatus was so good that no corrections had to be applied for leakage .
Afffi charging for five minutes with a voltage of volts , the residual charge gave .
deflection of divisions in one minute , and 11 divisions after two or three more .
If we take 10 divisions as a basis of calculation , the residual charge is 147 per cent. of the initial charge , or between three and four tinles as much as we got before .
But the conduotivity is , say , a thousand times greater , which seems to show that there is no accounting for the residual effect by considering the ductivity alone .
If we adopt MAXWELL'S calculation as a basis , then the tion .
from layer to layer is greater in the case under considerathan in that .
formerly dealt with .
The conductivity of tlra film great ON THE ELECTRICAL PROPERTIES OF PURE SULPHUR .
enough as a whole to account for any serious diminution of the residual charge , as was ascertained by observing how the residual charge was held .
The voltage of the residual charge was only about volt in this case , and less in the former case\mdash ; so that it is conceivably a case of galvanic ization only .
Against this , it may be urged that the residual charge increases with the time\mdash ; strong evidence that the polarization is dielectric , not galvanic .
In order further to test this conclusion 103 volts , instead of , was applied for the same time ( five minutes ) , and the residual charge , after one minute , was found to be 187 per cent. of the initial charge .
A good many very concordant observations were used in obtaining these values , and there is no doubt about the residue being larger in proportion with the smaller voltage .
This result is not in accordance with the results obtained by BOUTY for mica , which was , at all events , partly owin to the charge creeping over the mica edges .
In our case , we cannot assert that the small residual charge observed by us was not due to the charge creeping over the sulphur surface .
Since the residual is so small that we are unable to discover its origin by experiment , there is not much to be gained by speculation .
We may , however , say that this film was probably a great deal more wanting in homogeneity than the perfectly soluble film , and that we expected a much larger residual charge in consequence .
That such was not ined , may either indicate that the possible explanation suggested by MAXWELL does not apply in this case , and that we are actually in presence of the new kind of polarization he suggests , or that though the want of homogeneity is considerable , the variation of the ratio of specific inductive capacity to specific resistance , from point to point , is only small .
As there is every reason to expect that the constitution of a mixed film will tinually change , owing to instability of the insoluble sulphur and the gradual tendency of monoclinic to pass into octahedral sulphur , the capacity was taken in order to detect any change that might have taken place .
This was seven days after the film was first made .
The capacity of the condenser was found to have changed somewhat , more than the change of temperature would be likely to account for , temperature for both condensers , corresponding to a change of only C. The now was such that the balance changed from 1071 to between 1089 and 1090 with 40 Clark This leads to a value of of as against obtained before .
There is no doubt about the film having undergone some change .
The was , therefore , measured ( battery both ways ) and found to be C.G.S. , with a voltage of volts and a galvanomt , ter sensitiveness of ampere per micrometer division .
resistance has , therefore , increased also .
There is no doubt that a real change has taken place , for the increase of specific inductive capacity precludes the possibility of the plates havin separated .
We attribute this to the conversion of the prismatic sulphur into octahelral , the only change to be expected , so that the specific inductive capacity is higher for the latter than the former .
The ffim was then heated and observations made at to C. The resistance was found to have diminished , the conduction was unsteady and the usual effects were observed .
With battery both ways after the steady state was reached the conduction was found to have increased between and in the ratio of to ( the respective gations deduced from 24 observations ) .
The value of the resistance at C. is C.G.S. The value of the specific inductive capacity at C. was apparently , and at could not be distinguished from this ; the temperature of the standard did not change meanwhile .
At , C. the stance was much decreased and the conduction unsteady , the effect on reyersal was not so strongIy marked .
Thus : Temperature , 10 5 C. Battery Double elongation 472 Ditto , reversed 465 Battery Double elongation ( at once ) 444 Ditto , reversed 465 The results of these experiments are collected in the following tables:\mdash ; TABLE X.\mdash ; Showing Variation of Specific Inductive Capacity of Film of November 9 , 1892 .
phur In the above table no correc are applied for increase in the linear dimensions .
of the electrodes .
If this be done it will be found that the constancy of the capacity as the temperature rises indicates a small positive temperature coefficient .
in the specific capacity which may , however , be due to a partiaI annealing .
This is dealt with later when we had command of a better standard .
ON THE RICAL PROPERTIES OF PURH ] TABLE XI.\mdash ; Resistance of Film of Mixed Sulphur of November 9th , The resistance results may be summed up as follows:\mdash ; ( 1 .
) mixed sulphur film is at least 1000 times more conducting than the purely crystalline film .
( 2 .
) The resistance changes from day to day , independently of the surfa action , and is always less when the voltage- is applied , ( b ) when it is reversed ( at ordinary temperature ) .
( 3 .
) The resistance is discontinuous , particularly when the battery is first applied .
( 4 .
) The resistance decreases nearly two hundredfold between and 10 C. ( 5 .
) Dampness cannot increase the surface action beyond a certain point ; thus with our it could not bring the apparent specific resistance down below C.G.S. Film of November 18 th , 1892 .
An attempt was made to increase the amount of insoluble sulphur present by cooling the plates more rapidly .
With object in view the sulphur and plates were heated up to C. , and kept at that temperature for half-an-hour .
During the heating ( which was ried out in an improvised air bath ) a stream of pure carbon dioxide , dried by calcium chloride only , was allowed to flow into the air bath .
In spite of this some of the sulphur caught fire .
The plates were cooled by placing the dish on a block of ice , and covering the upper plate with a zinc tray filled with ice .
By this means the metal plates , and with them the sulphur , were reduced to a temperature of nearly C. in than five minutes .
The cooling at the commencement was very rapid , and no water got on to the film or into any part of the apparatus .
Several analyses were made of the material from the dish , but they varied between 5 per cent. and 7 per cent. only ; rather to our surprise .
We believe , 128 PROFESSOR R. THRELFALL AND MR. J. H. D. BREARbEY however , that the film was more quickly cooled the sulphur round its edges , in which case the film probably contained upwards of 7 per cent. of insoluble sulphur .
The colour of the sulphur in the dish was anything but satisfactory\mdash ; it was appreciably darker than it ought to have been\mdash ; especially immediately after cooling .
Some samples were cut out with difficulty from round the plates , and on evaporation from platinum left an appreciable stain .
It was thought that possibly the carbon dioxide atmosphere in which the sulphur was heated had led to contamirlation under the action of the slightly burning sulphur .
A number of experiments were therefore made by setting fire to sulphur in partial acid atmospheres , but no trace of contamination could thus be brought about .
We consider the dark colour as probably due to the action of a certain amount of dust which must have collected on the sulphur during the manipulations\mdash ; some of the same sulphur having been used in a previous experiment .
The results obtained with this film were similar to those with the last film , but with the peculiarities exaggerated .
The plates were Ultimately caused to come apart by raising the temperature , and it was then found that though the film was not as strongly coloured as the sulphur round the edges , it was distinctly darker than it ought to have been .
It was much broken up by air bubbles , and not at all regular in thickness .
The upper plate was slightly displaced .
The mean corrected area may be taken at 155 .
centirns .
The mean thickness at centi The melting-point of the Phur was taken as usual in several experiments in which the temperature of the bath before the sulphur was introduced was raised to within a fraction of a degree of the melting-point , so as to avoid annealing .
The constancy of the results showed that this was satisfactorily attained .
The suIphur used to find th melting-point was cut the sides of the plates as close as possible to them .
rrhe temperature at which melting commenced was ; it was complate at .
The initial temperature of the bath varied from to The capacity taken when the film was first dried in the thermostat , gave a correc ed value for the specific inductive capacity of the sulphur of about .
The resistance was tested after drying with phosphorus pentoxide for 20 minutes , and aftercharging with 258 volts for five minutes .
The deflection of the galvanometer was about 75 divisions ; on reversing the battery and placing the galvanometer in circuit , after about two minutes , the deflection on reversal was too great to observe , but was estimated at about 270 divisions .
After keeping the battery on way for 15 minutes , and again reversing the galvanometer , the deflection amoumted to 126 divisions .
Aftel *A subsequent examination of the measul.ements showed great differences between the thickness *in different parts of the film , and the plates were afterwards discovered to have lost their perfect The absolute values , both of the specific resistance and specific inductive capaCity this fflm , therefore , be completely relied upon .
It is probable that the of too and from the entioned .
ON THE ELECTRICAL .PROPERTIES OF PURE SULPHUR .
10 minutes the battery was again reversed , and the deflection could now be read .
Successive elongations on reversal of the galvanometer were , and as before the conduction was discontinuous , the galvanometer needle bemg thrown violent , lie in the direction of increased conduction every few minutes , and then quickly back .
The mean effect both ways reduced to a specific esista n of The film was then left for six days in the box with the phosphorus pentoxide , so as to get thoroughly dry , in case it was not so wheu the above results were obtained .
The effect of time on the resistance was noted as follows:\mdash ; Temperature of sulphur , C. , 312 volts .
On first applying the voltage battery to insulated plate the deflection corresponded to a resistance of ohms .
After three utes this increased to ohms ; fter twenty-five minutes to ohms .
The battery was then reversed , the initial effect died out in about three and during ten minutes ' observations were made so as to fix a mean value for the resistance , with the previous observations .
Up to three minutes the resistance was ohms .
From three to ten minutes mea , n resistance waes ) It is not easy to get very accurate values , because the change is always on , and the discontinuous conduction makes it necessary to have several observations at each point .
To test whether OHM'S was obeyed , we applied a voltage of 155 volts , and took the resistance both ways , up to three minutes , reversing several times first .
The meau value was , instead of , when we used the higher voltage .
Thig result was probably affected both by the fact that the large voltage vigs applied first , and by the influence of the time acbion , though the latter was elihtinated so far as possible .
We do not think that the deviation from OHM'S law rved oan be explained away by any of these effects .
A substance , whose eon\amp ; ictivity iyiay be a function of the quantity of electricity which has passed throogh it , ean hardly be expected to obey OHM'S law .
On Novetfiber 29 , after the had been allowed to rest , the resistance was , by reversing the battery ( 312 volts ) , and allowing the voltage to remain on for three minutes each way .
The two sets of deflections were nearly equal , showing that a steady state had been reached .
The resulting value was Resistance ohms .
After one hour and twenty minutes this increased ( with battery on one way only ) to .
the battery ( 158 volts ) was then applied the same way , and the resulting value was .
or OHM'S law is not obeyed , though the steady state has been reached .
The battery having been on one way for about two hours , it was thought a good opportunity for us to measure the effect of reversal , but , on this being done , the conduction was too great to be observed for several minubes ; after , say , five minutes , it was still six times greater than before the battery was reversed .
The full voltage was employed , and the resistance went on increasing , the current falling towards its former value for ten minutes , when observation had to be suspended .
The efl'ect of keeping the exposed $to the action of the battery seemed so definite , and at same time so curious , that it , was thought worth while repeating the observations .
This was done on December 1 , 1892 , at a temperature of and a voltage of volts .
After a preliminary , to see that all was right , the positive end of the battery was put to the top sulphur plate and left , thus connected for one hour ; there was then a mean throw on reversal through the galvanoIneter of 28 divisions ; the current had become practically .
the observations were : Throw \mdash ; . . .
\ldquo ; . . .
30 divisions The telegraph pole in circuit only gave an elongatio of 2 divisions .
The mean e]ongation due to current through the sulphur is therefore 26 divisions .
On reversing the battery , and observing , as quickly as possible\mdash ; within about two minutes\mdash ; the first two elongations were in the mean 92 divisions ; after five minutes ( about ) this fell to 71 divisions ; after tbirty minutes to 27 divisions .
On being again reversed , the first elongations were 81 divisions ; after six minutes 75 ; and after thirty minutes 26 divisions .
On again reversing the first effect was only 43 divisions , so : the sulphur behaves as if it got fatigued , and refused after a time to show the effect .
If we suppose that the conduction is the expression of the breaking-up and re-arrangement of molecular groupings , such change being from more complex to more simple , this is the we should expect when the groups capable of giving way under the applied electric stress became weeded out : .
In these experiments of December 1 the smallest specifio resistance observed was C.G.S. , and the greatest The galvarlometer sensitiveness was ampere per micrometer division of double elongation .
Effect of of Temperature , December 2nd , 1892.\mdash ; Temperature of .
sulphur , C. Normal resistance ( i.e. , after three minutes charging each way ) was ON THE ELEOTRICAL PROPERTIES OF PURE SULPHUR .
ohms , and the corresponding specific resistance was C.G.S. Voltage , 805 volts .
The temperature was raised slowly to C. and kept constant at this temperature .
The resistance decreased very rapidly , and was observed to be before the spot went off the scale and the voltage had to be reduced .
This is a reduction of , say , one hundred and sixty fol The behaviour of the film was now curious .
The temperature being allowed to rise from C. to lo C. in thirt , minutes , the conductivity went on increasing so fast that the battery voltage required to be continually reduced at , and after thirty minutes the conductivity had increased to one hundred times ( say ) its value at This increase of conductivity did not take place gradually , but by jumps\mdash ; the galvanometer needle was thrown further and further out by sharp jerks which continually tailed off only to recommence .
This effect went on till the temperature of 10 C. was reached , in fifteen minutes , after this the temperature was allowed to fall in order to discover whether the change in the sulphur would reverse .
This it did immediately\mdash ; the light spot coming back as markedly as it had previously moved off .
An attempt was then made to reverse the change again , and the temperature was slowly raised to , but the resistance still went on increasing .
This suggested that the plates were coming apart , and on examinatiou when all cold this was found to be the case , for , though no sign of cracking was to be seen , the capacity was fou.nd to vary when a tinfoil coated jar placed on the upper plate was filled with about. .
of mercury .
The plates were therefore dismounted , and the upper one was found to be only held in position by the sulphur round its edges .
We were pleased to find the capacity test so sensitive and positive .
We ]ooked carefully too see if there was any indication of a spark having passed , but could find no reason for supposing that this had taken place ; we do not think we could have missed it if.such had been the case .
The owing table will sum up the resistance results .
1.32 PROFESSOR R. THRELFALL AND MR. J. H. D. BREARLBY Resistance a containing 7 per cent. ?
of Insolubl .
Sulphur .
nresi tecific.esista cesAf hfter 4utes Besista cattery baysB rafter safter 3abontry opphed aminutes chargeth , Film of th , 1893 .
It has been already stated that some uncertainty attached to the constitution of the film just considered , and this was increased when , on examining the electrodes by the true plane , it was found that one of them had not preserved its perfect flatness .
A.nother circumstance ) to induce us to re-examine mixed films .
For this purpose the plates were again adjusted for ffitness , and a proper zinc dish constructed to carry ice on the upper plate .
In order to obtain further security against surface action we reverted to our old plan of erecting a quartz rod fence along the sulphur surface between the dish and the upper plate .
The sulphur employed was some of PROPERTIES OF PURE PHUR .
the hanc .
sulphur given to us by .
CHANCE .
This melted and to get rid of )articles of dirt , and was then twice ] this sulphur we made a fibm on April 17th , 1893 , and put it in the box while still warm with phosphorus pentoxide and sulphuric acid .
The data are\mdash ; corrected area , sq .
centims. ( including edge correction ) ; thickness , 022686 centim. It consisted of soluble prismatic sulphur ( probably without admixture of octahedral sulphur ) , with 5 per cent. of insoluble sulphur .
Colour , clear lemon-yellow .
When the plates were separated greater part of the film was still transparent , but it became opaque on rubbing or attempting to detach it from the plates .
We think it probable that since the film was much broken and torn up in forcing the plates apart , that it was in the transparent form during the experiments , at all events before it was heated .
In this case , it must be considered to have consisted entirely of monoclinic and amorphous sulphur .
An elaborate investigation of the properties was made but , as nothing new turned up , the table will afford sufficient information .
The resistance is uniformly higher than iu the case of the film of similar composition of November 9th , 1892 .
We attribute this to the fact that , having now a great deal more material , we probably succeeded with the purification rather better than before .
Of course the of burning two hundred grammes weight was carried out , and resulted in such an almost inappreciable residue that we must have succeeded better than was to be expected in keeping the dust off during the burning , or rather boilinoa *Sulphur is conveniently filtered by means of glass wool in.a zinc funnel .
A bit of platinum wire , folded several times , should be placed below the funnel to catch any shreds of glass that may cpme throngh .
This precautiou is very necessary .
Chance sulphur treated like this has a horrible smell of gas lime when it is broken up after cooling , which shows thab it reqlires to be distilled if sure results to be obtained .
ON THH EELECTRICA PPROPERTIE OF PURE SULPHUR . .
5 @ 5 fla 5 PROFESSOR R. THRELFALL AND MR. J. H. D. ARLEr The question of surface conductivity is finally settled by this film ; got just the same results whether the quartz rods were in or not , or whether we kept the film in the box or exposed it ) damp air , or even breathed on the surface of the sulphur between the dish and the upper plate .
Using the excellent condenser of Messrs. CLARK , MUIRHE-tD , and Co. , we were enabled to make the capacity comparisons with greater accuracy , and are thus in a position to assign an approximate to the temperature coefficient of the specific inductive capacity\mdash ; which turns out to be positive , and of the approximate value per degree Centigrade .
The increase in specific resistance suggests that perhaps even now we are using a material of sufficient purity .
In favour of this view we may state the film was not quite evenly coloured , some portions appearing of a browner shade than others .
On the other hand know of no reason to suspect any impurity , even the action of dust having been carefully guarded again and the sulphur itself being admirably pure .
It was perhaps , however , insufficiently exhausted of gas .
On the other hand , we know that when nnealed the resistance of sulphur as we used it is much higher , and it is difficult to credit a supposed impurity with the property of itself undergoing a change of constitution at the same temperature as that at which sulphur anmeals .
The observations show a very definite reduction of resistance as following a rise of temperature to C. , and persists to some extent on cooling .
We noticed this before .
We know from a special experiment , in which some finely-powdered sulphur was heated to C. for an hour , that the percentage content of insoluble sulphur is thereby diminished .
For instance , in a very careful expel.iment made on pure , finelypowdered sulphur in quantities of about ten grams , it appeared that the percentage content of insoluble sulphur was before annealing , and after annealing for several hours .
A clitical examination showed that was entirely outside the nits of any possible experimental error .
The arlnealing at C. had been previously observed .
We provisionally attribute the decrease of resistance to this partial annealing , which therefore tends to show that the conductivity of mixed sulphur has a maximum value somewhere between the per cent. and 5 per cent , content of insoluble sulphur .
The explanation of the temperature curve is , however , not by any means obvious , although it seems safe to draw the following couclusion , assuming that the temperature lag is inappreciable .
The conclusion is that there must be at least two causes in operation influencing the resistance .
One of these is probably the temperature per se , and the other the rate of transformation of insoluble into soluble sulphur , which is probably a complex phenomenon .
The complexity may possibly arise from itself not being direct , the amorphous sulphur being first converted into a less stable form .
There is also evidence , which will be given later , that at least two varieties of insoluble sulphur coexist a rapidly cooled film .
On referring to the table it will be evident that all the phenomena similar to those formerly observed , but on a smaller scale , owing to the conduction being less .
All the effects of time , of reversal of voltage , \amp ; c. .
observed as before .
We ON THIB ELECTRICAL PROPERTIES OF SULPHUR .
append a short table giving the relation between temperature and resistance , by of pplement to the general table of results .
TABLE XV.\mdash ; Voltage , 280 volts .
Fig. 4 .
Specific Inductive Capacity.\mdash ; Five pieces of film , lred as usual , the following values for the thickness in inches:\mdash ; ( 1 ) 009022 ( 2 ) 009917 ( 3 ) 008689 ( 4 ) 008058 Centre ( 5 ) 008978 Mean0089316 inch centim. These meas ( lrements were made with very great care and represent the greatest accuracy the method is capable of ; .individual measurements of the same bit of film not differing by more than 00008 inch .
The differences are to be partly attributed to partial crushing or crumbling of the film during measurement , though there is clearly also a want of uniformity in the thickness of the film itself In obtaining the capacity by the DE method , a balance was first obtained when the film was in the box , the lid was taken off and a new balance obtained ; the plates were taken out and supported on their ebonite legs on the laboratory table at some distance from anything else .
Removing the lid of the box changed the resistance required for balance from to , and placing on the table reduced this to 2510 .
The connections and keys by themselves require 47 ohms in the adjustable arm .
The other was always 100,000 ohms .
We see from the small effect produced by lifting the plates out of the box which was to earth , that the free capacity may be neglected- compared to the part of the capacity which is due to the dielectric action of the sulphur .
We have used the 2463/ 105 in calculating the capacity , allowing for the capacity of the keys , \amp ; c. The standard required no correction at all , it was 1 microfarad .
The resulting capacity is therefore ofnrad as stared in the summary .
sulphur condenser did not leak sufficiently to affect the galvanometer by more than one division with 300 volts , and hence with 40 Clarks may be considered to insulate perfectly .
The mica standard , however , leaked quite appreciably , so that its resistance was only about ohms .
The capacity was measured again at , and the balance was found to be unchanged .
In this case the sulphur condenser also leaked , i. e. , its resistcmce was about oh1ns , as against , say when it was cold .
In order to find whether these leaks ( which are of course exceedingly small ) produced any measurable result , we made use of the formula given by GLAZEBROOK , ' Phil. Mag. , vol. 1 , p. 376-377 , 1881 .
This is where and are resistances out in the arms , and and the corresponding insula- tion resistances of condensers .
We found that the correcting factor was about ON THE ELECTRICAL PROPERTIES OF PURE SULPHUR .
in one case , and in the other .
Consequently our results are quite independent of leakage .
Absorption , unless it be instantaneous , is guarded against by the point at which the kick vanishes .
Some doubt having arisen as to the correctness of the formuIa obtained by GLAZEBROOK , Mr. A. POLLOCK was kincl enough to look into the matter for us .
His concJusion is that instead of the ] quoted we ought to have being the galvanometer resistance , the other ]tters retaining their meanings .
This does not lead to any appreciable correction of our values .
Now , when the sulphurcondenser is heated , the diffexence in the coefficients of expansion of sulphur and must lead either to the film separating from the aluminium or to both it and the aluminium being thrown into a state of strain .
We know that the plates remained firmly attached to the film , from a subsequent examition , and , consequently , we are driven to admit that heating strains both the film and the plates .
Now the coefficients of linear expansion as obtained from Sir W. THOM'Collected Papers ' ( vol. 3 , p. 209 ) , which numbers were taken from " " ants of Nature are for sulphur , say , 000063 , and for aluminium , 000022 .
To be quite satisfactory , of course , the linear expansion of particular kind of sulphur ought to be observed , but the values given ( mean of those of FIZH : AU and KOPP ) are probably not very far wrong .
It is probable that when the aluminium are heated they a , re quit , free to expand laterally and also vertically ; in fact , the gold dish is so weak and soft , it may be set on one side , and the round the aluminium plates will , if anything , tend to drag them outwards .
We may assume , therefore , that when the condenser is heated the aluminium plates increase in area , and are subject to some little stress , while the sulphur film increases freely in thickness , but is prevented , by sticking to the plates , from expanding beyond their limits .
film being thin and weak , compared to the plates , nearly the whole of the resulting strain is borne by it .
Consequently , any change in condenser capacity observed on heating will be due either to the direct effect of temperature change in altering dimensions or dielectric constant , or to the stresses thereby set up .
We will assume the former alternative and see to what results it leads .
The facts are*thab when the sulphur is heated , say from C. to the mica standard and resistance boxes being at constant temperature , no change in the balance is observed ; for instance , it remains steady a microfarad within Qile figure in the last place .
The temperature coefficient of specific inductive capacity , which would be required to compensate for the tion of capacity , computed from the variation of the various dimensions with temperature , is This , of course , must be further investigated we can regard it as a physical constant , but we may take it as pretty certain that there is either a } ) erature coefficient , or a strain coefficient , or both , or that the specific inductive capacity depends on the composition of the film , for the sulphur is partially annealed at C. , , therefore , possibly at C. It was a distinct oversight not to have tested this at the time by observing the capacity again when the sulphur was cold , and ultimately ] to another experiment being undertaken for the special purpose of setting this question at rest .
The result of this experiment was to prove that the specific inductive capacity of films , containing rather less insoluble sulphur than this , is disJctly greater the more insoluble sulphur they contain .
Hence the temperature coefficient must be greater than the above if it is at all affected by the annealing , for the annealing would almost certainly have reduced the capaciby .
While we used the old standard condenser the absorption was so great as to make the observations for capacity insufficiently exact for the present purpose .
SECTION VI .
The Determination of Inductive of rieties of Sulphur by the Jfethod of Weighing .
In the prosecution of these studies by one of us and Mr. J. B. ALLEN , some experiments were made the specific inductive capacity of sulphur by the method of weighing .
In these experiments a good many sources of electrification suitable for the purpose were investigated , and a good many causes of error detected and rectified .
The results , however , only referred to one plate , i.e. , to one containing per cent. of insoluble sulph u and even these results were ot in such good agreement as could be wished .
The best way of preparing plates of sulphur had however , been discovered , and we felt sure that by gradually improvi ng the apparabus , we sbould ultimately succeed in obtaining reliable results .
The great merit of the weighing method , however , from our point of view , is that it makes the resulting value of the specific inductive capacity absolutely independent of any electrical measurements whatever .
In the experiments previously described , we relied on the constancy of the voltage of Clark cells , on standardized resistance coils , and on certificated condensers .
It was now felt that an absolute determination of one of the constants ought to be made by a perfectly independent method , in order to supply as complete a check as possible on the ) revious work .
No method could have met our ends more perfectly than the method of weighing , for the result ON THE ECTRIC ARTIES OF PURE SULPHUR .
remains entirely independent of the absolute value of anything ( as a measurement of , of course , should do ) ; all that is required is a fairly acculate knowledge of the atios of the weights employed , and of two distances which can be measured with fair accuracy .
The drawback to the method is that it is not as susceptib]e of being carried to a high degree of accuracy as some more directly electrical methods .
Both measurements oflength were reduced to the same standard , by fine callipers of ELLIOTT BROS .
, whose abQolute indications are immaterial , but which we believe from other comparisons to be practically correct .
The comparison of the gilded brass and platinum weights with each other was carefully carried out on a fine balance , and resulted in showing that the weights were so well adjusted that none of the corrections observed can affect our results\mdash ; the differences being much less than the error of weiglning against the electric forces .
As we spent several months in perfecting our apparatus , after the experiments of Mr. ALLEN and one of us were finished , we decided to illustrate our description with a small drawing to scale ( Plate which we hope may be of assistance to other people using the method , and which will enable us to shorten our description considerably .
A is the zinc suspended disc .
The lower surface was scraped true to a surface plate after the three pillars had been } to its upper surface .
The diameter of this plate was centims. , and it weighed ( uncorrected ) grams .
Each of the pillars V , V , V were bored and tapped so as to carry three screws\mdash ; Iying in a plane parallel to the plate .
These screws are necessary to alIow of the adjustment of three fine German-silver wires , by which the plate is supported .
It is necessary to have these screws well made if accurate and convenient adj nstment is sought after .
The point of junction of the three wires is attached to a single fine German-silver wire , which passes to below the balance case , where it is hung from a frame originally made for the purpose of weighing quartz rods in water , in order to determine their coefficient of expansion .
This frame ] be arrested just like the pan of an ordinary balance .
The rough adjustment of the length of the wire was made by a minute box-wood shackle .
The balance stood on a state shelf , and the rest of the apparatus on a slate table below it , and we may here remark that one of the first desiderata is steadiness and solidity in all the parts of the apparatus .
fine spiral of wire connects the plate to the guard-ring .
is a guard-ring of zinc ; the top and bottom surfaces bein turned parallel to each other in a lathe .
The inner diameter is 13 centims. and the outer diameter oentims .
, leaving the ring with a radial breadth of centims. The clearance on each side of the suspended plate is , therefore , 175 centim. We could have done with less clearance than this , but an examination of the theory showed that no error at all comparable with those of measuring and weighing would be introduced by an popular space of these dimensions .
The guard ring is supported by three strong legs , , made of zinc , } provided with good and jamb nuts , , D. The ring is put to earth by a soldered connection , which ultimately is soldered to a water-pipe .
The high potential plate is a large glass disc , centims. in diameter , and optically flat .
It was originally silvered , but the silvering having gradual .
tarnished by contact with sulphur , and its conductivity having become uIlreliable , it was replaced by some good thin tin-foil , fastened down with thin starch paste , and put on so well that no appreciable deviation from flatness could be detected .
It was attached by a clamp and wire , to the potential end of the transformer .
The thickness of the plate , centims. , gets rid of any fear as to flexure producing an appreciable effect .
This plate was supported by three ebonite cones , , 2 centims. in diameter at the thick part , and centims. long .
These cones were got to insulate well , they were supported by a wooden stand , ular in shape , and carried in turn on three levelling screws , working through brass fittings , and bearing on brass caps supported .
by three of the double glass and paraffin insulators , , we have ad occasion to ref'er to so often before .
All this part of the apparatus requires to be well and substantially made , or it tends to tilt over a little .
The top of the glass disc was cen tims .
above the slate bench .
wire connecting the tin-foil to the transformer was short and thick , but it was found that no alteration was made by twisting it into a spiral , so that no " " resonance\ldquo ; effects are to be feared .
Screens of tin-foil pasted on to cardboard completely surrounded all parts of the apparatus .
The first screen stood on the guard ring and enclosed the back of the sus- pended plate , the wire to pass through a small hole .
Outside the whole affair other and larger screen .
Both these screens were cylindrical and properly shaped , and were made in two parts joined by tin-foil flaps .
balance was also completely encased in a screen of tin-foil and gauze , allowing the pointer to be observed with ease .
The panets of balance case were separately screened by gauze to allow of the adjusting of weights , \amp ; c. The elementa1y mistake of having bad screen connections was not made .
Screens are very necessary to prevent air currents , as well as to hinder electrostatic action .
The balance was in a room with a south aspect at the opposite end of our building to the room occupied by machinery .
Balance.\mdash ; This was a rather short arm balance , which was formerly used for weighing quartz , \amp ; c. It has a high range of sensitiveness , though this was diminished to a convenient amount in order to secure other advantages .
Jt furnished with an adjustable screw stop at , by Mr. ALLEN .
The balance will carry 1 kilogram , so that there is no danger of warping .
Electro-Dynamic Action.--This was tested by insulating the glass plate , the other being connected to on one side , and to the terminal of the transformer on the other .
No trace of any action could be detected , , either nothing at all or else something incomparably smaller than we could approach io the experiments on weighing .
ON THE ELECTRICAL SULPHUR .
Auxiliary Condenser.\mdash ; This is necessary to swamp the effect of changes in capacity produced by putting in or taking out the sulphur .
We used ten " " gallon\ldquo ; Leyden jars in parallel with the guard-ring condenser .
Transformer.\mdash ; It appeared from the experiments of one of us and Mr. , that of all the methods of excitation tried , nothing was so steady as a large coil used as a transformer , the alternator being excited by a small current only , and driven by rather loose belting from a well loaded gas engine which has a short period of irregularity , The alternator has a very massive field-magnet system which rotates , and this arts sufficiently well as a fly-wheel to smooth most of the irregularities of the gas engine , when these are short in period .
Some irregularities , however , remain and impose the limit to the accuracy attainable .
It was found best to have a weak field on the alternator , and use all the armature coils , instead of strongly exciting it and using a armature coils in series with resistances .
This was contrary to expectation .
The induction coil could also be used with its break , or.with a clock-work break .
Mr. ALLEN found , however , that these were not so steady , and the same remark applies to a number of attempts made to use a WInSHURST machine driven at a constant speed , and having one permanently to earth , and the other earthed by a high resisbance cuum tube , a bit of caustic potash in it. .
An irregular excitation will very quickly spoil the knife edges of the balance .
The ffequenoy of the alternator was about 60 , and we got better results with this than the frequency was less .
In order to keep a check on the voltage , the latter was measured continuously by a KELVIN static voltmeter reading to 12,000 volt , .
About 5000 to 6000 was the voltage generally employed , Adjustments.\mdash ; The apparatus having been provisionally set up , and all rough adjustments made , the following procedure was adopted .
The glass plate was elled as accurately as possible by an ordinary laboratory spirit-level , reading perhaps to something like of arc .
The guard-ring was then similarly levelled at a suitable distance , from to 2 centims. in our experiments , from the lower plate .
This was then checked by measuring the distance between the top of the tin-foil and the bottom of the guard-ring , in several places by a wedge reading to 1 millim. directly. .
The adjustment being found correct , the hanging plate was adjusted to the guard.ring by eye , the under surface of the ring flat and level , and the suspended plate also being flat .
This can be done with accuracy , just as in sighting a barometer .
Of course the preliminary adjustment was good enough to let the plate hang concentrically in the guard-ring .
The adjustment must be made with the plate counterpoised , and is rather tedious , but presents difficulties that may not be overcome by patience .
The lower plate then excited in order to if the suspension stretched appreciably , but this found not to be the case .
The screens were then placed in position , and the experiment was ready for observation .
: Swlph.ur Plates.\mdash ; These we re made in two ways , and we consider that too much care cannot be exercised in this part .
the work .
" " Insoluble\ldquo ; Plates.\mdash ; A light sheet zinc mould was prepared , capable of containing a plate about two inches greater in diameter than the one to be employed ( which was made usually of rather larger diameter than the guard-ring ) , and say four or five times as thick .
The sulphur , previously filtered as described , through glass wool and platinum gauze , was melted and raise to a sufficient temperature by placing the zinc mould on a hot flat plate , It was noticed that after the filtered sulphur was cold , when it was being broken up for the final melting , it still possessed a strong sulphury smell , which in this case suggested gas lime .
This shows that there is an advantage in distilling the sulphur , although the amount of impurity present may be imperceptible on boiling the filtered sulphur in a polished platinum dish .
However , the impurity seems to be got rid of by ) elting in the dish , for the smell as not detected on breaking up the plates the close of the experiments .
The mould having been filled with sulphur at the proper temperature , i.e. , after it had become very viscous , above 17 C. , the whole affair was plunged into cold water and agitated so as to cool it as quickly as possible .
When the plate was sufficiently , i.e. , atter some hours , the mould and sulphur were mounted on a lathe , and the bottom of the mould cut completely away , till only a flat surface of sulphur was left .
This was then reversed on the face plate , and the free surface turned down til ] we reached a homogeneous part of the sulphur .
or beyond that if desirable .
A round cake of the desired size was bhen cut out .
The first plates were then scraped to a surface plate ; but it soon appeared that the natural warping of the plates rendered this an unfruitful labour .
Consequently our later practice was to grind the plates down on a sheet of glass-paper , glued to a flat ( really flat ) iron plate , till they were true to a straight edge .
lt was necessary to experiment at once before the plates warped too much .
Soluble Plates.\mdash ; These can ot be handled so freely as plates consisting partly of soluble and partly of insoluble sulphur , however we found the following method gave just as perfect results .
The mould consisted of a flat plate of " " opal\ldquo ; glass , which is very smooth and generally better than sheet glass .
A ring of plaster of Paris was cast on such a plate and dried , then previousl annealed sulphur was carefully heated to just above the melting-point and poured into the ] levelled by he glass plate .
The sulphur was caused to cool as fast convenient , in order to prevent the formation of internal cayities .
The plate , when cold , left the glass and plaster easily , and it was , of course , much larger than requisite .
In order to prepare the upper surface , the plate was held flat on the bed of a BROWN and SRARPE milling machine , and milled at a good speed till it was homogeneous and of the proper .
This gives just as good a flat surface , or rather better , than any of our lathes , and leaves a very fine appearance .
Owing to the rapidity of cooling , any ON THE ELECTRICAL PROPERTIES OF PURE SULPHUR .
impurity from the plaster of Paris is localized to the edge of the )late , i.e. , is about half an inch outside the guard-ring , where it can do no harm during the weighing .
\mdash ; One of our plates warped in a very regular manner , so as to become concave on one side and convex on the other ; the curvature looked spherical and was assumed to be so .
As we did not know what error this might lead to , and still did not wish to lose the plate\mdash ; on which a good deal of time had been measured the curvature by a spherometer , and Mr. POLLOCK kindly investigated for us the proper formula .
As , however , it involved a difficult integration , it was considered advisable t.o calculate the effect on the value of of some extreme assumption as to the thickness of the , plate .
For this purpose it was supposed that the plate increased in thickness by the same amount all over , as at the boundary of the suspended plate .
The radius of curvature was found to be for the concave side of the plate , centims. ; and for the nvex side , 134177 centims. ; and the thickness of the plate was centim. The diameter of the plate was taken at 5 inches centims. In a particular experiment it was found that the value of , when the plate considered as having the maximum effective thickness of 94407 centim. instead of 943 centim. , would be , instead of when the measured thickness was taken .
Since this difference is outside the experimental limits of weighing or measuring , and is itself probably so far in excess of the actual correction , even if the lines of force are appreciably " " refracted we decided to ignore the effect of the curvature .
Method of erimenting.\mdash ; The experiment consists in observing the attraction in grams weight on the suspended disc , both when the sulphur plate is in and out .
To begin with , everything must be carefully dusted , to get rid of shreds of all kiffis , and the hanging plate counterpoised\mdash ; it weighs , as has been stated , about grams .
The current is then turned on and observed to be steady by the voltmeter\mdash ; it never really gets steady , the voltmeter oscillating through about one division , corresponding to , say , " " permanent\ldquo ; variation of half a division or 100 volts in 5000 , or say , 2 per cent. ; this would cause the force to vary by 4 per cent.\mdash ; this is an extreme limit to the fluctuations .
Weights are added until the pointer of the balance is observed just to incline to leave the z.ero\mdash ; of course , as soon as it does leave it the force diminishes and the balance tilts over .
The poin having been ascertained as nearly as ible , the rider is run down to increase the weight by 005 gram , and if this produces a decided tilt , the observation is taken as exact .
This means that the force can llot be measured to less than the weight of 5 milligrams owing to fluctuations of potential difference .
The forces to be observed from about 5 gram weight without the sulphur to 2 grams with it , and these forces appear in the final result as a ratio which enters to the power of one half .
The uncertainties of weighing , therefore , to about005 gram in one gram , or , say , 5 per cent. , and affect the value of to about25 per cent. IIDCCCXCVL\mdash ; A. The attraction , without the sulphur , having been observed , the attraction is again measured with the coil commutator reversed , in oase there is any outstanding electrification ; in general this produces no effect .
Thero was no difference in any of the experiments leading to our values .
The next step is to earth the high potential plate and carefully dust every part of the apparatus , including the sulphur plate ; the connecting wire is dusted by running a flame along it .
The sulphur pIate is then inserted , and an observation made of the value of the counterpoise , in order to find if there is any free electrification .
If there is , the plate is taken out and the electrification removed by a flame\mdash ; a risky proceeding , and one we never indulged in unless it was absolutely necessary .
T.he process of finding the balance was then repeated as .
When a sufficient number of observations , both with the plate in and out , had been made , the screens were removed , itnd the distance between the guard-ring and the glass plate remeasured .
There was never a difference of more than millim. At the end of the series of experiments all the plates available were experimented on in this manner one after the other , so as to get good comparative results .
The apparatus was then dismounted , and the distance between the fixed and suspended changed , after which the adjustments were re-made and the whole series gone through again at the new distance , with a view to eliminating accidental errors of setting , weighing , and measuring .
The sulphur plate must not be less than five or six millims. from the suspended plate , or the free electrification its upper surface has a sufficient P.D. from that of the earth to cause brush discharges .
Consequently we always examined some plate or other in the dark to see if this did occur .
If it does , of course , the result is that we get too low readings and a too low value of K. On the other hand , the nearer the top of the sulphur plate is to the suspended plate the better , for the greater the forces to be observed .
The " " art\ldquo ; of the experiment , we should say , lies in adjusting the voltage till brushing just does not occur .
This is also provided for by placing the coil terminals so near together that a discharge occurs there before it occurs bebween the sulphur and the earthed part of the apparatus .
It is also exceedingly ecessary to remove all dust and sulphur crumbs between each set of weighings .
On thinking the matter over in the light of our present experience , we have an idea that there might be advantages in using a liquid of high electric strength instead of air\mdash ; in fact , to proceed as QUINCKE did in determining the dielectric constants of liquids by weighing , in order to get the advantage of the large forces be was able to Analyses of the sulphur plates were made by breaking them up , and taking samples from the active part .
This was also necessary in order to find out whether there were any cavities in the plates .
In only one plate was a cavity discovered , but happily it was under the guard-ring almost entirely , and was very insignificant .
ON THE ELECTRICAL PROPERTIES PURE SULPHUR .
The partly insoluble plates were annealed in some cases in order that the difference of .
specific inductive capacity might be observed in one and the same plate .
The annealing produced both buckling and cracki n ; the latter does not matter , and the former has been dealt with .
In cases where the plates were annealed the analysis was made on samples collected during the turning or milling of the final surfaces .
The formula of calculation is best given in the following form : where and are the forces observed , is the distance between the fixed and suspended plates , is the thickness of the sulphur , is the quantity we are in search of .
The results will be sufficiently evident from the following table , which includes unsatisfactory values as well as satisfactory ones .
The former are included with a view of showing how we gradually nproved our .
as we elinxinated various souroes of error\mdash ; bad conductivity of silver coating , irregularity of driving , effect of brushes due to imperfect dusting , or too high voltage , \amp ; c. , \amp ; c. THE ELEOTRICAL }RTIES OF PURE SULPHUR .
The net result of the measurements detailed in Table XVI .
is as follows:\mdash ; Temperature , C. For mixed ontaining 8 per cent. of insolubl sulphur 1 .
, , , mpletely.solubl sulphur ( aged monoclinic ?
) . . . .
The value of for the soluble sulphur is the mean of values obtained on two different samples .
It appear from these values that the specific inductive capacity rises as the percentage of insoluble sulphur increases .
Experiments yet to be made on a oak of sulphur containing a very large amount of insoluble .
The results of the experiments described in the foregoing investigation may be summarized as follows:\mdash ; ( 1 .
) The specific resistance of perfectly soluble sulphur , either simply molloclinic , or agedlinic , which has lost its raphic properties and become opaque , but which practically preserves its old melting-point , is very high , certainly above C.G.S. units .
( 2 .
) An admixture of say 5 per cent. of amorphous sulphur reduces thi , great specific resistance to say .
units .
( 3 .
) The conductivity , such as it is , is marked by some peculiarities ; it to be discontinuous , resembling in this the behaviour of surface films and of .
bad conductors ( glass rods and ebonite ) generally .
( 4 .
) The conductivity is always much greater for some time after the battery is reversed , say for three or four minutes .
( 5 .
) The " " residual\ldquo ; charge , when the sulphur is regarded as a dielectric , is small .
( 6 .
) The conduction does not obey OHM'S law .
The specific conductivity is greater , the greater the voltage .
( 7 .
) The specific inductive capacity of mixed sulphur is greater than of pure soluble sulphur .
( 8 .
) The conductivity of mixed films increases enormously while they are annealing , and even perfectly crystalline sulphur shows traces of conductivity at temperatures near the melting-point .
( 9 .
) Melted sulphur conducts comparatively well , say 1000 times as well as a cold mixed film .
NOTE ADDED APRIL 1 , 1895 .
Experiments .March and April , 1894 .
A possible explanation of the very divergent values of , as exhibited by different odifications of sulphur , may be sought in the corresponding variations of density .
To test this point , an examination was made of the density of the sulphur employed in the experiments for by weighing .
The accuracy of the determination of density extends to the figure at least .
The following table will show that the variation of specific inductive capacity is not to be explained by supposing that it is merely proportional to density .
TABLE showing Relation between the Specific Inductive Capacity and the Density of the Forms of Sulphur .
DESCRIPTION OF PLATE 1 .
New High Grade Gatvanometer .
Fig. 1 .
Elevation .
A. Fine adjustment in azimuth for control magnets .
B. Coarse for raising or lowering same .
C. Fine , 1 ) .
Brass ffame or bridge carrying control system .
E. Iron slabs forming magnetic screen .
F. Wooden stand carrying iron slabs .
G. Brass ring carrying , supported by three levelling screws on lead pillars , H. H. Lead pillars carrying G. I. Ebonite slab insulating and supporting coil system .
J. Glass and paraffin insulators supporting ebonite slab I. K. Concrete slab supporting the whole apparatus .
L. Sandstone pillar supporting K. control magnets . .
Bottom control nets .
O. Wooden carriage for Fig. 2 .
Section .
A\mdash ; O. Same as in fig. 1 .
Q. Position of millboard case enclosing the coils , \amp ; Details of coils and mounting ( small letters ) . .
Ebonite terminals . .
Brass frames carrying coils .
and .
Brass base . .
Brass cell , in which the whole of the ooils and magnet system are free to turn . .
Levelling screws .
Fig. Coil support in plan and section . .
Ebonite terminals . .
Brass frame . .
Quartz rods in coil . .
Brackets carrying quartz rods . .
Carriage fastened to but adjustable to magnet system by loaving a play in the holes . .
Magnet formin part of system . .
Wire forming coil . .
Paraffin mounting . .
Ebonite mountin .
Glass front supporting a pair of ebonite lars , Threlfall Brearley .
Fig. 1 .
COlL SUPPORT PLAII .
Trans. , 1896 , , Plate 1 .
Fig. 2 .
Fig. 3 .
from of DESCRIPTION OF Fig. 1 .
Thermostat .
A. Section .
B. Plan . .
Outer case of brass . .
Inner case of brass silvered inside . .
Heavy mineral oil filling annular space between and weight to keep in position . .
Ebonite blocks to support .
Dish with drying material .
bonite cones supporting and insulating glass plate .
Innermost brass box , with lid , silvered inside and outside .
Holes in lid for and electrodes .
Space between filled with air .
cover . .
Glass plate supporting ] dish and sulphur . .
Aluminium plates forming armatures of condenser ( see fig. 5 , Plate 3 ) . .
Zinc dish supporting an inner gold dish ( see fig. 5 ) . .
Electrodes from upper and lower minium plates . .
Standard thermometer . .
Ordinary high range thermometer . .
Brick .
Quartz sticks in sulphur . .
Surplus sulphur .
Threlfall , 1896 , Fig. 1 .
B. PLAN . . . .
OF PLATE 3 .
Fig. 2 .
Four-pole ebonite key . .
Wooden base . .
Ebonite rods with mercury cups in upper ends .
H'ig .
3 .
Reversing key , \amp ; c. .
Wooden base . .
Ebonite rods and mercury cup in upper ends . .
Rods of fused quartz mounted in ebonite ends , and with mercury cups . .
Tin-foil wrapped on ebonite handle . .
Discharging point , only used when determining residual oharge to connect , thus instantaneously discharging condenser .
For resistance measurements the legs at opposite corners were metallically connected as usual in a commutator .
Fig. 4 .
Micrometer telescope used with new galvanometer . .
Object glass . .
Graduated micrometer head . .
Scale divided on glass , replacing usual comb . .
Movable spider line . .
Eye-piece . .
Brass mounting , carries , and , and is capable of a universal motion ( not shower Fig. 5 .
Gold dish and aluminium plates . .
Aluminium handle used to lift when making film .
This is removed and replaced by electrode when film is in position . .
Aluminium plates between which the sulphur film is formed . .
Outer zinc dish . .
Inncr gold dish supported by .
Surplus sulphur .
rods sulphur .
NOTE .
\mdash ; Since the sketch was made , a zinc dish has been fixed to the upper aluminium plate .
It is scraped true on the bottom and joint sealed with sulphur .
The dish can then be filled with pounded ice , when making mixed sulphur films .
'hrdfall Brearley .
Phil. Iran , .
1896 , , Plate 3 .
MlCROMf Scdle ?
' A Fig. 5 .
, 1896 , DESCRIPTION OF PLATE 5 .
A. Suspended zinc plate attached to balance arm .
Connected to earth by a flexible spiral of No. 40 wire , B.W.G. B. Guard-ring of zinc with parallel faces .
, C. Zinc legs attached to by brass screws .
, D. Brass levelling screws with -nuts , leyelling guard-ring .
E. Glass plate covered with a sheet of tin-foil , which the high potential armature of the condenser .
F. Tin-foil surface attached to flat glass plate E. G. Cake of sulphur .
H. Connection from to the induction coil .
I. Metallic connection from legs of guard-ring to outer tin-foil case and earth .
, J. Ebonite pillars supporting glass plate .
K. Wooden base supporting , and E. , L. Levelling screws by which the tin-foil surface may be brought level and parallel to A. , M. Glass and paraffin ating stands supporting the three levelling screws , , L. N. Millboard box coated inside and out with tin-foil , and protecting the back of the suspended disc from induction .
P. Large millboard box coated with tin-foil , and surrounding all the apparatus .
It is put to earth , as is N. Q. Metallic casing surrounding the balance .
S. denser in parallel with secondary terminals to swamp the capacity of the sulphur condenser .
T. High potential Kelvin voltmeter .
X. High potential terminal of secondary . .
Low potential terminal of secondary , put to earth on water main .
The same earth connection is made throughout .
l'hrelfall and Brearley .
Phil. , I896 , , Plate 5 .

