]\gt ; ; Mr. E. H. Griffiths .
Value of [ Feb. 1 Il .
: ' The Value of Mechanical Equivalent of Heat , deduced from some Experiments performed with the view of establishing the Relation between the Electrical and Mechanical Units , together with an Investigation into the Capacity for Heat of Water at different ares.\ldquo ; By E. H. GRIFFITHS , M.A. , Assistant Lecturer , Sidney Sussex College , Cambridge .
Communicated by B. T. GLAZEBROOK , F.R.S. Received January 19 , 1893 .
( Abstract .
) The paper of which this communication is an ab.stract gives the particulars of an investigation which was commenced in the year 1887 , and extended to the close of 1892 .
The object of the enquiry is sufficiently indicated on the title page .
The values of the mechanical equivalent obtained by Joule in his later determinabions differ amongst themselves by as much as 1 per cent. , and differences amongst the results obtained by succeeding observers are , with the exception of Rowland 's in 1880 , still greater .
The harmony amongst the values obtained by Rowland is marvellous ; but , since his manner of tion was the same tlJroughout whole series of his experiments , his conclusions stand in need of confirmation by different methods of observation .
Obher observers who have attempted to obtain the value of the mechanical equivalent , by means of the work done by an tric current , have been hampered by constant perplcxities as to the of the electrical nnifs adopted .
The science of ical measurements has now arrived at such stage that its units may be regarded as sllfficiently established , , therefore , the time seems particnlarly appropriate for an enquiry into the relation between those units and the mechanical ones .
The difficulties of such an investigation are , of , as is shown by the discrepancies amongst the results obtained by those observers who , in recent years , have adopted electrical methods .
cause of inaccuracy has been preseltt in all determinations I have examined , , the increase in temperature of the conductor above the temperature of the medium in which it was placed , and the consequent undetermined alteration in its resistance .
Rowland writes as follows : " " There can be no doubt that expcriments depending on the heating of a vire give oo a value of the equivalent , seeing that the temperature of the wire during mus6 always * B. A. Report , ' 1892 .
Academy , ' June , 1879 , p. 153 .
1893 .
] , anical Equivalenf of Heat .
be higher than that of the water surrounding it , and hence more heat will be generated than there should be A short account is given iu this summary of the manner in which this difficulty bas been overcome , and I think it will be seen that this objection to the electrical method of investigation is now removed .
I have defined the unit as the quantity of .
heat required to raise unit mass of water through C. the air thermometer at C. , and so much confusion has al'ison from ambiguity as to the value of the unit , as defined , I have given reasons in support of the suggestion that this nition should be generally adopted .
Throughout the of this enquiry I have been ably assisted by Mr. G. M. Clark , B.A. , Sidney , Cambridge , and this communication should , in justice , be regarded as a joint contribution .
rFhe valne of an tion of this kind depends , in a reat measure , on the attention given to matbers of detail .
It is , therefore , impossible , in a short abstract , to produce the evidence on which resuIts are based , and we content ourselves with a brief outline of the method adopted and the conclusions arrived at , .
without attempt ' to tify those conclusions .
If a alorimeter is suspended in a chamber , the walls of which are ab a cons .
ant , we can , by observations over , small range across that outside , deduce the rate of rise due to the mechanical work done in the calorimeter , when the supply of heat is derived from irriDg only .
By epeating the observations in a similar manner whose mean ture differs from that of the Yotnding walls , we obtain change in temperature due to the combined effects of the radiation , ' .
, onduction , and at all points of our whole ange of temperature .
As the success of the method depends ( 1 ) on ) possibility of maintaining the exterior temperature unchanged , and on the regularity of the supply of heat due to the , we briefly indicate our method of securing those conditions .
1 .
The calorimeter* suspended within an air-tight stecl chamber .
The walls and fioor of this chamber were double , and the space between them filled with mercnl'y .
whole structu in a tank containing about 20 gallons of water , and ] ) rted in such a manner that there were about 3 inches of wate above and beneath it .
The mercury was connected by a tube * was of cylmdrical form , and suspcJlded by thrce ) It was made of " " gilding metal , ' ' which both and cxtcrnally as ( .ove a considerable thic.kness of gold .
All )faccs within were gilded .
8 Mr. E. H. Griffiths .
Value of Feb. 16 , gas ]egulato of a novel form , which conbrolled the supply of gas to a large number of jets .
Above those jets was placed a that silvel .
tube , through which tap water was continually flowing into the tank , all parts of which were tained at an equaI temperature by the rapid otation of a large screw .
'rhus , the calorimeter may be regarded as suspeuded within a chambe ] ' placed in the bulb of a huge thermometer \mdash ; the mercury in that bulb weighing 70 .
A clIange of C. in the temperature of the tank )vater caused the mercury in the tubes of the regulating apparatus to rise about 300 mm. Special arrangements were made by which it was possible to set the apparatus so that the walls surrounding the calorimeter ] be maintained for any length of time at any required temperature , from that of the tap water summer about winter up to C. or C. We know by observation that the temperature of ) steel chamber ( when once adjusted ) did not vary by C. , and we believe the variations were much less .
2 .
We experienced difficulty in devising a suitable form of ; and we attribute the failure of our earlier experiments to defects in the ordinary forms .
We find it impossible , withont a lengthy description , to give a clear idea of the stirrer ultimately adopted .
We can only state here that it was completely immersed when the depth of the water exceeded 1 cm .
, thab its bearings otltside the steel chamber , and that the water was thrown from the bottom to the lid of the calorimeter .
More than 100 experiments performed ( many of ) ]asting several hours ) in order to mill the of when the calorimeter contained different masses of water .
The harmony amongst the resuIts was These experiments proved that over our range of temperature , was a linear function of , and Newton 's law of cooling appeared to hold strictly true over a rallge of C. below to C. above the of the surrounding , i.e. , from C. to C. ; and our experimental results were of such a nature that a very small departure would have been rent .
We found that with our form of stirrer where was th rate of revolution , and some constant .
This relation ]leld true for all values of between 26 and 34 secolld , and , as our experiments we proposed to a rate of nearly as possible 30 revoIutions per second , we were able to make the necessary correction for small deviations from the lormal rate .
In order to diminish the irregularities in the motor , a special \mdash ; rise in temperature per 1 second due to the stirring .
) gain or loss temperature per 1 second due to radiation , \amp ; c. , when C. If the temperatures were reckolled on the mercury thermometer scale , curl.ature would be considerable .
1893 .
] the Mechanical Equivalent of Heat , form of pressure regulator was constructed , and fouud to be fairly , the variations in during all experiment being general ] very small .
The pressure in the space between the calorimeter and the walls of the steel chamber was reduced , as a rule , to between and 1.0 mm. * If is the capacity for heat of the calorimeter and its contents , quantity of heat lost or gained per second by radiation , \amp ; c. , per unit difference of temperature , and provided the is unaltered , the value of should be constant atever the weight of water .
It was not until the close of our work that we were able to obtain the value of M. Iv then found that the value of varied greatly with changes of pressure , and our resuIts ( although not necessary for the purposes of our investigation ) are interesting , since they bear out Bottomley 's conclusion that the ] is a sudden decrease in the loss by radiation when the pressnre falls below mm. We extract from our paper the following table:\mdash ; Table XXVI .
essurein m The critical point of the curve deduced the above table occurs at a higher pressure , and the bend is somewhat sharper than is the case with the curve given by Bottomley .
If is the rate of rise due to the non-electrical supply , and that due to the electrical then , . . . . . . . . . . . .
( 1 ) .
pressures were ascertained by a McLeod 's gauge .
'Phil .
Trans 1887 , A. Mr. E. H. Griffiths .
Value of [ Feb. 16 , We have described the manner in which we determined the last of this equation , and thus , by direct observation of , we were able to obtain the valus of and . . . . . . . . . . . . . . . . . . .
( 2 ) , where is the resistance of the coil , and the capacity for heat of the calori.meter and its contents at a temperature We have now to indicate the methods by which we ascertained the of the energy supplied by the current .
E.\mdash ; The extremities of the CA , CB will ( when the galvanometer shows no deflection ) be maintained a difference of potential equal to that due to the ] at C. Let be an adjustable resistance placed in the circuit which communicates with ( the sto cells ) , and let be the resisbance of the calorimeter coil AB .
Whatever variations may take place in and in the potential difference at , it is vays possible ided the adjustment of is sufficiently under control ) to , by close attention to the indications of the galvanometer , the points and at a constant difference of potential .
A special rheochord was designed of which it was found poseible , in spite of variations in the resisance of AB , to maintain the potential difference unctIanged thought e experiment , and we believe that in no case did the variations exceed 1 , / 10,000 of the mean difference of potential during each experiment .
The Clark cells ( which were placed at C ) were constructed by us ccording to the directions of Messrs. GIazebrook and Skinner , and I , on several occasions , been directly compared with the Cavendish , and , indirectly , with the Berlin , standard .
Their differences from th 1893 .
] fh Mechanical of Heat , standard are small ] , and their mean E.M.F. at C. voIts .
* R'.\mdash ; Had it been possible to maintain a constant value for , it would have greatly simplified the experimental work and the .
In the year 1890 we devoted mnch time to the examination of the various copper-manganese-nickel alloys , and we performed a series of determinations with a coil whose coefficient was practically zero .
The reasons which led us to reject these alloys and adopt a platinum wire will , we think , be found sufficient .
The value of was first determined by a dial-box ( legal ohms ) constructed by Messrs. Elliott .
Mr. Glazebrook has been so kind as to perform a complete standardisation of this box by means of the B.A. standards .
The resulting corrections have been applied , and the values of expressed in true ohms , as defined by the 'B.A .
Report , ' 1892 .
If is the resistance of the coil when at the standard temperature , then , where is the temperature coefficient of th wire and is the excess of its temperature above the temperature of the calorimeter .
It is difficult describe in a few seutences the manner in which we determined the value of but the following explanation may serve to indicate the method of procedure .
Supposs , and to be the arms of a Wheatstone 's bridge of which is the coil .
Let the arms and be equal , not only in resistance but in mass and dimensions , and let when the coil is at a certain temperature , the reading of the ermometer in the calorimeter being .
Let be built up of a mass of metal a small temperature coefficient and a cooling hurface .
S If the bridge balanced when a certain ) is passed through it , the balance will be destroyed if the temperature of the coil be raised as the current is increased , for the increase in temperature of may be neglected , and and will remain equal , however values slter , they are by equal currents and their capacities for heat are the same .
Equilibrium can , howevel , be restored by cooling the calorimeter to a certain temperature , * A full descl'iption of these cells will be found ) Messrs. Glazeblook and Skinner 's paper Phil. Trans .
) .
ticulars of this box have been given in a previous paper ( ' Phil. Trans 1891 , p. 44 ) .
wirc had a thin coating of amber varnish , and the insnlation appeared to be sufficient .
In order to test , a series of obsel.vations of were taken when the lorimetc was filled with pure pentane .
The increase in did not in 22,000 .
S The mass of German silver used by us in the arm weighed severn !
pounds and contained about 1800 feet of single vire in triple and double slrands .
12 Mr. E. H. Griffiths .
The Value of [ Feb. 16 , say , We then know that has resumed its former value , and is therefore at the temperature ; thus the value of By changing , the potential difference of the ends of the coil , we can find values of corresponding to values of E. By applying to the ends of the coils the potential balance previously described , the potential difference at the extremities of can , by increasing the Fi .
M.F. at { he ends of .
the , be raised to that of 1 , 2 , \amp ; c. , Clark cells .
The following table gives the resuIts obtained by this method .
During our observations , the rate of stirring was the usual one , but we also investigated the effect of small in the rate .
The last column gives the values deduced from the parabola ( where is the number of Clark cells ) , as we found it convenient to express the differences in temperature by the COIl.esponding differences in R. A difference of between the last two columns corresponds to a change of 1 in 8600 in R. The correction is a most important one , and the neglect of it by previous investigators ciently accounts for their failure to obtain satisfactory results by observation of the heat developed in a wire by electric current .
Table IX .
, Time.\mdash ; An electrical clock with a seconds pcndulum 1VRS used as our standard for time .
It was carefully with a chronometer by Dent .
A was contl by this clock .
The rate of the clock was a losing until August 21 , but after that date its error was less than 1/ 25,000 , and uo eution was necessary .
balanoe sensitive to a clJange of 1/ 100,000 of the Ieast mass measured by us lllld a set of ] 's weights*were used in our determination of ) * These weights vere 1standar by Messrs. Oertling in August , 1892 .
1893 .
] the Me jhanical Eivalent of Heat , .
13 nperature.\mdash ; At the time of writing ( December , 1892 ) our hermometry is based on measurements made by platinum thermometers .
We propose to make , at an early date , a direct comparison of our standard .
with the air thermometer by means of the apparatus described by .
Callendar , has been so kind as to promise his assistance .
In a previous paper the details of a careful comparison of the platinum and the air thermometer have been .
It was then shown that the platinum-air difference curve gave values of at all temperaturss from to .
accurate within C. of the real value of , and that discrepancies appeared to have an experimental origin .
The experience of two years ' work with platinum ) ermometers bas but increased our confidence in them .
Should a direct comparison with the air thermometer modify our determinations of differences of temperature ( and it is only differences which are important to us ) , our results will be modified accordin .
This will not , however , necessitate repetition of the experimental work , as the corrections involved will be of a numerical order only .
A cury thermometer by Hic , ks ( labelled ) was standardised by direct comparison with different platinum thermometers , the observations being taken under conditions to those prevalent our " " and the temperatl ) corresponding to the readings given in column 1 , Table XVII , were thus ascertained .
We ound it necessary to calibrate the mercury thermometer not only for irregularities the bore , but for certain f , trictly recurrent changes in " " lag we have found to be invan:iably associated with a rising mercury thermometer .
We are unable to briefly describe he method adopted , and we confine ourselves to stating that it was based on observations of time .
The results of this investigation , which extended over several montbs , prove that a calibration of th : ordinary kind is sufficient if accurate observations have to be obtained with a rising mercury thermometel .
As an tion we give the following table , which shows the difference in the value of certain ranges on thermometel ' when steady and when rising at the normal rate of our experiments .
* ' Roy .
Soc. Proc January , 1891 .
' Phil. Trans 1891 , , p. 155 . .
Mag , 1892 .
S The stem of is thermometer graduated in mm. : about 40 mm. C. Mr. E. H. Griffiths .
The Vatue of [ Feb. 16 , Table XVI We have now indicated how the various qu$lntities in equation ( 2 ) were determined , the exception of and ; we can therefore deduce from cquation ( 2 ) { he time ( T ) of rising C. at any of our range when and is the potential dilference of one Clark cell .
We thus get If be the cvht of water , and the water equivalent of the calorimeter at the standard , and if and be the temperalure coefficients of their specific , then * [ Note , February 18 , ] \mdash ; On , I received a thernlometer ( by M. Tonnelot ) which had been under by .
Gnillaume for previous two .
I have made adircct ison of this thermometer ith Em at three , observing all the precautions by .
Guillaume , and the results arc as follows : \mdash ; ' ' diffcl .
by C. in actu.tion at C. , ) as to of is of ented a rison , and ations lve b , hoxs over , to indicate that thernnometric errors small .
1893 .
] the uivalent of Heaf , hence . .
, . . .
( 4 ) .
By observations with different weights of water , and and observing and , the corresponding times , we obtain by subction .
( 5 ) .
Hellae when ( i.e. , at the standard temperature ) we can find without first the values of , or the water squivalent of the imeter , and by repeating the observations over different ranges we can find without previously ; or , having obtained , we can find and , and then by equation ( 4 ) deduce value of from a experiment .
We have adopted both methods as a check upon the calculations , which involve much arithmetic .
The latter method is the more convenient , as it enables us to the results of separate experiments , bub it canI ) be applied until the values of , ltnd have previously been obtained by ations on two different at two temperatures .
he table shows a few of the results given in Table XL of our paper .
We have divided our experiments into Series I and II , and we have given reasons why more weight should be attached to tlJe latter series .
We here give a summary of the of deduced from Series II .
By " " group\ldquo ; we denote all experiments cooducted with the same weight of ater , and in case a group contains experiments performed with diiferent values of ( where is the number of cells which determine the potential ence ) .
As was in some cases changed from 2 to 6 , the rate of production of heat was eased 9 times .
The agreement amongst the results of experiments performed with different values of is not shown in the portion of the table here given , but it is close , and aifords a satisfactory proof of the accm.acy of the values to and , and the validity of the method employed to ascertain the actual temperature of the coil .
The number of experiments pel'formed in each is shown by the under the heading " " mean The extent of ) irrcgularities is clearly indicated by this table .
smooth curve " " was in each case so drawn that the some of positive and negative areas between it and experimental ( given by the numbers in the columns headed " " mean was zcro . .
E. H. Griffiths . ?
Value of [ Feb. 16 , Table \mdash ; Values of ( results of Series IT only ) .
curv .
The values of at integral values of the temperature over our .
were read off from the smooth curves .
We give the values at , and C. Table XLI .
VaJues of at , and C. From this table we obtain the following results:\mdash ; Specific heat of water at in terms of its spccific 'neat at , deduced from colunlns 4 and . . . . . . . . . . . . . . . . . . . . . . . . . .
Ditto and 5 Ditto from columns 5 and Mcan . . . . . . . . . . . . . . . . . .
1893 .
] the anical Equivalent of Heat , Hence , adopting C. as the standard temperature , the SPECIFIC HEAT OF WATER Also by means of equation ( 15 ) we get the following of : Columns 4 and 6 . . . . . . . .
' lfean. . . . . . . .
This value of , as previously pointed out ( equation 5 ) , is entirely of the value assigned to the water equivalent of the calor meter .
And we find the water equivalent of the calorimeter at C. in terms of water at C. .
The water equivalent of the calorimeter at C. in terms of water at C. grams .
Hence water equivalent We can now find the capacity for heat of the calorimeter and contents for any weight of water at , and C. , and deduce the value of from each group separately .
The Groups and are experiments performed on and grams respectively .
rfable XLIII.\mdash ; Values of J. We have in the above table given the values resulting the calculation at different temperatures , for the limit of our experimental errors is thus clearly indicated , since the values of ought ( in the absence of experimental errors ) to be identical ab all temperatures .
The close agreement between the values from different groups , and from ths same group at different temperatures , is a satisfactory proof of the accuracy of our determination of the water *Over the range to C. All weights are reduced to vacuo .
18 yalue of the Mechanical quivalent of Heat , .
[ Feb. 16 , equivalents of the calorimeter , and of the changes in it and in the capacity for heat of the water .
If we reject Group ( and we have already stated that it has little value ) , the resuIts are practically identical .
Hence ( the thermometry depending on comparisons with platinum thermometers ) if we asstlme 1 .
The unit of resistance as defined in the 'B.A .
Report , ' 1892 ; 2 .
That the E.M.F. of the Cavendish Standard Clark cell at C. voIts ; * 3 .
That the thermal unit quantity of heat required to raise 1 gram of water through C. at C. , the most probable value of This , by reduction , gives the following:\mdash ; kilogramme-metres in latitude oi Greenwicb .
-lbs .
per tbermal unit C. in latitude of Greenwich .
-lbs .
per thermal unit F. in latitude of Green{vich .
The of this abstract is already undnly great , and we wilI , therefore , not enter on any discussion of the results beyond remarking that if we express Rowland 's value of in terms of our thermal unit we exceed his value by 1 part in 930 , and we exceed the mean of Joule 's determination by 1 part in .
The difference between Rowland 's value of the temperature coeffiof the specific heat of water and ours would , however , cause both his our values of to be identical if expressed in terms of a thermal unit at C. *If assume the E.M.F. of our Clark cells to be the same as that of the Cavendish standard ( and we are inclined to think we have over-cstiJnated the diiferencs ) , we get The ) obtained by us in Rowland obtained the mean value of Joule 's determinations by assigning values to diiferent experiments , and the above comparison refcrs to the numbers thus obtamed .
If , however , we attach equal weight to all Joule 's results , as given by mean exceeds our value by 1 4280 , assuming our expression for the temperature coelIicient of the specific of water .

