]\gt ; XVI .
On the Specifc of Constant Volume.\mdash ; Part II .
Carbon Dioxide .
J. JOLY , M.A. , Sc. D. , F.R. , S. Received March 9 , \mdash ; Read Apri126 , 1894 .
THE present paper is occupied with an experimental investigation into the variation of the specffic heat at constant volume of carbon dioxide attending change of absolute density .
The investigation is in continuation of a previous one in which Carbon Dioxide , Air , and Hydrogen were the subjects of a similar enquiry over low ranges of density .
It appeared to me desirable to extend the observations more especially in the case of carbon dioxide , becanse of the extended knowledge we already possess of its isothermaIs , and the fact that its critical temperature is within convenient reach .
Other physical properties of the gas have also received much attention of recent years .
It is also readily procured in a nearly pure state .
The observations recorded in this paper extend , in the one direction , to densities , such that liquid is present at the lower temperature ; and in the other , to a junction with the highest densities of the former paper .
A plotting of the new observations is in satisfactory agreement with the record of the old .
Ib reveals , however , the fact that the linear nature of the variation of the specific heat with ensity , deduced from the former results , is not truly applicable over the new , much more extended range of observation .
For convenience the chart at the end of this paper embraces the former resuIts , and the present paper is extended to include the entire resuIts on the variatio1l of heat with density where the range of temperature , obtaining at each experimen , is approximately the same : that from air temperature to 10 C. Part III .
of this investigation will contain an account of experiments on the riation of the specific heat with perature when the density is kept constant , The division is for convenience of reference .
The value of Professor assistance and advice , and his kindness in giving them to all seeking his help , are so ] known , that it is needless to say that this present research owes much to his assistance .
He is not , indeed , esponsible for mistakes I may have fallen into , but he is to be thanked for saving from many .
The arrangements for carrying out the experiments are essentially the same as ' Phil. Trans A. , vol. 182 , 1891 , 28.12.94 those described in Part I. Some modifications in the structure of the calorimeter , \amp ; c. , must , however , be noticed .
The differential form of the steam calorimeter was used throughout ; the mode of experiment being to compare the calorific capacity of a strong spherical copper vessel , firstly when containing but one atmosphere of the gas , and secondly when charged with the requisite mass of the gas , with that of an idle vessel of closely similar mass and volume .
Each of these comparative observations embraced a series of from three to ten experiments , according to the mass of gas operated upon ; ]arger number being requisite when the mass was small .
The identical calorimeter , described in Part I. , was used in the present experiments up to the etion of the experiments of Table IX .
An improved differential calorimeter was then completed for the requirements of the experiments upon the temperature variation of the specffic heat .
It was accordingly taken into use , and with it the experiments were completed .
As its peculiarities of construction are mainly to render it suitable for the experiments described in Part III .
, its description is deferred to that paper .
To out experiments at the high pressures contemplated , new copper spheres of greater strength than the former ones had to be provided .
The dimensions were also reduced , and , what was most important , the brazed equatorial joint entirely dispensed with .
This was effected by spinning the entire sphere out of the one piece of copper till the vessel was closed down to an orifice of about 2 centims. in diameter .
Into this opening an accurately turned piece of gun-metal was brazed .
Still furbher to diminish the number ofjoints , this piece of brass was so formed that it constituted the valve seat receiving the steel screw-valve , and carried the connecting nozzle used in filling it .
* Several of these vessels were made , designed to have a capacity of about 80 .
centims. , and to be 2 millims. thick in the walk .
Assuming the tensile strength of copper to be grammes per square centimetre , these vessels would possess a bursting strengh of some 300 atmospheres .
The safe limit would be about 100 atmospheres , which pressure was somewhat exceeded on one occasion .
One of these vessels having been chosen for use , its internal and external volume were measured .
The external volume was found to be .
centims. ; the internal , .
Its mass was grammes .
The external volume of the idle sphere was found to be at C. ; and the mass , grams .
The active sphere was now tested by putting into it some 12 grammes of carbon dioxide , and heating it in a current of steam for or 20 minutes .
It became visibly more truly by this treatment , and a -determination of its external volume afforded .
centims. at C. , showing a stretch of .
centims. , probably due ncipally to a more perfect sphericity .
The experiments on the gas were now begun , using a mass of grammes .
Later on , after the experiments of lable , the volumes were again determined .
The external gave cllb .
centims. at C. , which may be considered identical with the previous result ; the ' Part p. 78 .
( Table IX .
) , however , not so concordant with those recorded in Part I. obtain with the lighter vessel then used .
This is not to be wondered at when it remembered that a 1 per cent. ac.curacy in the case of dealing with 4 grammes of gas already a determinaiion closer than one part in 4000 of the total precipitation upon the two vessels in the calorimeter , or one two-thousandth part of that upoIt the vess containing the gas .
The carbon dioxide used in the experiments was obtained from the brewery Messrs. GUINNESS and Co. In the brewery it is removed from the fermenting from a level low down , some 15 or 20 feet below the edge of the vat , which remain filled with the gas .
It is then pu1ified by washing with water and treatment wit permanganate of soda , and compressed into iron bottles .
In use it is best to invel the bottle somewhat , so as to draw from the liquid .
The gas used in the early experiments had a faint alcoholic smell .
Later , gas was supplied to me th kindness of the head engineer , Mr. GEOGHEGAN , which had no perceptible odou Derminations of the amount of impurity were frequently made , by absorption wit alkaline pyrogallol of some 65 .
centims. of the gas over mercury .
The impurit ( air , probably ) was in all cases small ; at worst , about 1 part in 360 by volume .
series of experiments ( Table XII .
) was carried out upon gas prepared in the laborator from pure bicarbonate of soda , and pumped into the sphere with the aid of th mercury pump , described in Part I. This gas contained , according to subsequer determination , an impurity of only one part in 1015 by volume .
These experimenl reveal no discrepancy with those made upon the less pure gas .
It is , indeed , not be expected that the impurity of one in 360 by volume woul produce a perceptibl error .
In filling the gas into the sphere it was passed through an iron drying tub about 1 oentim .
in diameter and.35 centims. in ]ength , filled with asbestos an phosphorus pentoxide .
It was found that the bransfer of the liqnid into the sphe1 was greatly facilitated by cooling the latter with ether poured on muslin around it .
The hrst series of experiments were made upon a charge of grammes , which by determmntion showed an impurity of 1 in 485 by volume .
They are in the followiog table , and the succeeding table gives a calorimetric comparison of th empty spheres applicable to the experiments .
The mode of estimating the pressur obtaining : nd of evaluating the deduction to be made for effects other than that due to the calorific capacity of the gas at constant volume , is given further on , the experimenos have been recorded .
TABLE III .
grammee .
Mean Density .
Mean Pressure mospheres .
TABLE IV .
grammes .
Mean Density .
Mean Pressure Atmospheres .
Correction for spheres Other corrections ) Deduced specific heat OF GASES AT CONSTANT VOLUME .
TABLE V.\mdash ; W grammes .
Mean Density .
Mean Pressure Atmospheres .
Correction for spheres Other corrections Deduced specific heat TABLE VI.\mdash ; W grammes .
Mean Density Mean Pre.ssure pheres .
Correction for spheres Other corrections Deduced specific heat OF GASES AT CONSTANT VOLUME .
TABLE IX.\mdash ; W grammes .
Mean Density .
Mean Pressure Atmospheres .
Correction for spheres Other correcbions Deduced specific heat The unsteadiness of the results with 3 mmes of caused me to carry the reduction of mass no further .
A test of purity applied to the remaining gas a result the same as the former test .
Both samples were from the one bottle .
The following experiments were now made on the empty vessel : TABLE X.\mdash ; Comparison of the Empty Spheres .
These apply to the results in Tables , as well as to the results in Tables XI .
and XII .
Some ( additive ) alterations had been made to the platinum catchwater .
To these I in part ascribe the discrepancies between Tables II .
and X. It is probable , too , that a small quantity of the phosphoric anhydrid had been carried into the sphere in filling it with gas for the experiments of ables I , for it was observable that the mass introduced , according to the weighings made upon insertion , appeared to be , while the addition of all the quantities liberated , gave grammes .
As there was certainly no leakage at any time , the weights obtained upon release of gas were adopted as true weights .
After this , all experiments were carried out in the new form of calorimeter described in Part III .
To connect its results with those obtained in the old apparatus , the next set ( Table XI .
) were made upon gas taken from the same bottle used for the foregoing experiments .
There is satisfactory agreement with the results of Table VII .
TABLE XI.\mdash ; W grammes .
Mean Density Mean Pressure Atmospheres.other corrections Correction fpheres Deduced specific h The next set ( Table XII .
) were carried out on gas prepared from bicarbonate of soda by pure sulphuric acid and pumped into the sphere through drying tubes ( as described in Part I The impurity in this gas was found to be 1 part ( of air l ) in 1105 by volume .
TABLE XII .
grammes .
Mean Density Mean Pressure Atmospheres .
OF GASES AT CONSTANT VOLUME .
The next table contains the foregoing results collected .
I have added , for nvenience .
of reference , the experiments on carbon dioxide contained in Part I. The mean density is designated TABLE XIII.\mdash ; Experiments on Carbon Dioxide .
Low Pressure .
The results collected in Table XIII .
are not obtained without applying to the perimental results the corrections described in Part I. , pp. 85 , et seq. The followdata apply:\mdash ; Internal volume of active vessel cub. centims. at C. Elastic yielding cub. centim. for a pressure of atmospheres .
The next table ( XIV .
) contains the principal data used in the calculations of rrections .
TABLE XIV.\mdash ; Data for Calculation of Corrections .
ines crossing tmals diagonally , from which pintermediate terat bscertained.ressures awere taken separately fhese charts.owest densities , TeSul twhen mestimat opressure s in the tables .
Upon the completion of the foregoing experiments , as the limit of stress to which the vessel might be subjected had not been reached , a fresh series was begun extending to higher lensities .
The sphere was tested with grammes of the gas at for 15 minutes .
The sphere , as already mentioned , further increased in volume .
The experiments then carried out are contained in Table XV .
TABLE XV.\mdash ; Experiments at High Densities .

