]\gt ; III .
On the Specifc of at .\mdash ; Part I. Air , Carbon Dioxide , Hydrogen .
By J. , Assistcmt to the Professor of Civil Engineering , Trinity College , Dublin .
Communicated by Professor FITZGERALD , M.A. , F.R.S. , F.T. Received September 2 , \mdash ; Read November 20 , 1890 .
[ PLATE 6 .
] Introduction .
THIS paper is occupied with a consideration .
from a purely experimental point of view , of the specific heats of three gases : ( 1 ) of air , chosen as being typical of the sensibly perfect gas , and one the properties of which have entered more into thermodynamical considerations than those of any other fluid ; ( 2 ) of carbon dioxide , as typical of the imperfect gas , in the case of which there is also the advantage of an already extensive knowledge of its other properties ; ( 3 ) of hydrogen , as a gas whose behaviour , as regards BOYLE'S law , suggests that it occupies a peculiar position among perfect gases .
The specific heats of these gases are , in this first notice , treated over pressures ranging approximately from 7 to 25 atmospheres .
The range of temperature is not sensibly varied , the lower limit being that of the prevailing atmospheric temperature , upper that of steam at ordinary .
The question of the relation between the specific heats of gases and their temperatures I hope to make the subject of a future notice .
For advic , and assistance in all matters uf difficulty , more especially in the ratio of the nature of the corrections to be applied to the experiments , I ve to thank Professor FITZGERALD .
Where my work touched on science I had the friendly assistance of Professor EMERSON REYNOLDS , and must tbank him for his readiness to help me .
In this connection , my thanks are also due to Mr. E. WERNER .
The labour of reducing the results of the experiments much ]ightened by the kind help of Mr. H. H. DIXON , of Trinity College .
The Board of Trinity College , in providing a suitable room for my experiments , have rendered it possible for me to attain a far higher degree of consistency in my NDCCCXCI .
16.4.91 results than was previously possible .
Some only of the experiments recorded in this paper shared this advantage ; their results lead me to hope that future work will possess a higher degree of juracy .
The Government Grant Committee have enabled me to meet the expenses ding the construction of apparatus .
To the personal care which Messrs. YEATES and SON have devoted to the construction of the apparatus the practicability of the experiments may be said to be due .
It is sufficient , as testimony of this , to refer to the qnalities displayed by the vessels they constructed for holding the gases .
On the Possibility of Accnrate Determinations .
In two previous communications made to the Royal Society , have inquired into the accuracy of the method of as a means of investigating the specific heats of substances .
I claimed for it a high degree both of truth and sensibility , comparing favourably with , if not transcending , that of all other methods .
I might also have claimed for it the quality of elasticity in an important degree .
With a delicate balance , the smallest mass may be dealt with .
Increasing the power of the balance and the dimensions of the calorimeter , it is easy to deal with large masses or with bulky vessels impossible of treatment by other methods .
This is the method employed in the following investigations on the specific heats of gases .
For its principles and for its mode of employment specially reference to gases , I must refer to my paper on the Steam Calorimeter ( loc. in which a plate will be found illustrating the large differential calorimeter used in my experiments .
The spherical copper vessels , one of which experiment will contain the gas under a .
or less degree of pressure , are shown suspended in the calorimeter .
The essential detail of construction , whereby the freedom of the wires suspending the vessel from the balance is secured , is figured in the adjoining plate given in that paper .
There will be no need of further description of the apparatus here .
I assume here , on my former proofs , the reliability of the method , and pass on to the question of the degree of accuracy possible of attainment in this application of it .
The possibility of such experiments has always been a question of degree .
For with the amount of stress to which any suitable material for enclosing the gas may be safely subjected , the weight of the vessel must many times traoscend that of the gas , so that , in spite of the rather high specific heat to be ascribed to the latter , the calorific capacity of the gas must remain only a fraction of that of the containing vessel .
Nor by increasing the size of the vessel is the disproportion ished .
For , the vessel spherical , the rate of stress on a great circle for a given pressure will increase as the diameter ; the weight of the vessel as the square of the diameter , * ' On the Method of Condensation in Calorimetry\ldquo ; ' Roy .
Soc. Proc vol. 41 , p. 352 , et seq. ) ; and ' On the Steam Calorilneter\ldquo ; ( ' Roy .
Soc. Proc vol. 47 , p. 218 , et seq OF GASES AT CONSTANT VOLUME .
75 if the kness be maintained constant .
But , assuming in each case the spherical to be charged to the limiting safe stress , an increase in the diameter necessitates an increase in the thickness of the material proportional to the increase of stress , that is , proportionaI to the crease of diameter .
, finally , the weight of a vessel which is always at its limiting safe stress increases the cube of the diameter ; and this is the rate of increase of its volume or of the weight of the contained gas .
An inferior limit to the size of vessel is , however , fixed in the case of the present experiments , by he sensibility of the ordinary chemical balance in estimating the small weight of steam which is condensed due to the thermal capacity of the contained gas ; a superior limit by the increased risk on a chance explosion and the inconvenience of providing large quantities of steam for filling the calorimeter rapidly at the commencement of an experiment .
After various trials and experiments on larger vessels , two spheres of copper having an external diameter of about .
were designed .
One , " " A to stand a test pressure of 500 lbs. on the square inch and a working pressure up to .
The other .
" " to stand a test of 1000 lbs. and a working pressure of .
A smaller margin of safety than this was not considered advisable as , at the temperature of the steam , a decreased resistan of the vessel was to be expected .
The thickness of the walls of these vessels would be , by calcuIation , closely one-half and one millimetre respectively .
And to these dimen.sions the spheres were as accurately as possible constructed .
Now if we take the case of sphere and supposing it filled with air to a pressure of some 22 atmospheres at , rising to 30 atmospheres at or 450 lbs. per square inch , the following proportion obtains between the weight of steam thrown down by the vessel and the weight of steam precipitated by the contained and its catchwater weigh nearly 100 grms. Its specific heat may be taken as .
Over the range of the condensation of is therefore just grms. The air in weighs grms. Its condensation is .
Thus one-bhirteenth of the precipitation is due to the air .
But as we must be content to use the sphere at pressures lower than this but still too high for sphere , it is necessary to regard the ratio as descendin as low as one to eighteen .
Forestalling speculations on the possibility of dealing with this one-thirteenth , say , to one per cent. of its value\mdash ; or to of the whole condensation , I give the results of an entire series of six experiments in which this holds , in fact the series from which\mdash ; taken at random\mdash ; I have borrowed the foregoing figures .
The mass of air , grms. , affords over the mean range of a precipitation of .
There is a want of balance the thermal capacities of the vessels , making the observed precipitation .
These are the mean results of the six experiments .
To compare the experiments among themselves I divide the weights of precipitation observed in each case by the range of temperature for that experiment .
The six numbers thus obtained represent the weight of steam thrown MR. J. JOLY ] SPECIFIC HEATS down in each experiment per degree of range .
The consistency of the experiments is thus truly represented by these numbers .
They are : \mdash ; 0.018070 Mean precipitation divided by mean range ( See among the experiments , Air VIII .
) The departure from the mean in one case only amounts to so much as half per cent. .
From these figures it is evident that close measurements are practicable .
the mode in which these measurements are effected , I may in the first place refer to a former statement of my own that by this method of calorimetry observations of the thermal capacity may be carried to the one-thousandth part .
This was based on experiments on minerals where surface conditions were varied by breaking up the specimens between successive observations .
Now , although it is probable that this degree of accuracy would not obtain in the case of successive experiments on large objects of low density , effected in the ordinary way , on the other hand the effect of the simultaneous differential arrangement is to limit or banish several sources of error .
A.ny source of error common to the two spheres and to the gas is not accumulated on the gas .
The gas simply bears its own share .
Thus a mis-reading of temperature affects the result for the gas simply by the amount of the error , and will not , as in the case of successive differential experiments , by falsifying the specific heat of a mass thirteen times as effective , thermally , as the gas , throw thirteen times the error upon the gas when the allowance for the thernlal effect of the sphere subsequently comes to be made .
With this class of error are eliminated to a great extent effects of radiation and effects due to buoyancy .
Theoretically , indeed , in the use of the aneous differential method , there is reason to consider the entire sensibility of the method directed to the evaluation of the effects of the gas as if isolated from its containing vessel .
Practically this will not be fully lttained .
But it appears evident from all considerations that this application of the method to the calorimetry of gases is not straining the method ; and the expectation is , I think , justified that in the use of this method the old difficulties will be found to be overcome .
In the tables of experiments , given further on , discrepancies in many cases occur .
But these cannot all be laid to the charge of the calorimet.ry .
In the case of hydrogen ' ' On the Method of Condensation .
cit. , p. 361 , OF GASES AT CONSTANT VOLUML .
77 nearly all these discrepancies are very certainly due to varying amounts of impurities in the samples of gas .
Again the earlier experiments on air were made under difficulties which it would be needless to enumerate .
The principal one , that of insufficient supply of steam , arising from an inadequate gas main ; was not overcome till after the experiments on carbon dioxide were concluded .
I did not think it necessary to repeat these , for their agreement inter se was fairly satisfactory , but in the case of the far more difficult problem , the variation of the specific heat of air , I did repeat , under the improved conditions .
The results , from the present point of view , go to show the degree of precision attainable in the careful application of the method .
I allude to experiments VIII.-XI .
, which are seen plotted on the diagram of experiments on air , Plate 6 .
The method is delicate in its results .
It is , I think , capable of dealing with some subtle questions in calorimetry , but , as in other methods of measurement , when the higher resources of the method are called upon , care and judgment are needed to guard against the entry of error .
The Vessel for Holding the The earlier attempts at effecting the measurements of the specific heat were confined chiefly to efforts at overcoming two principal ficulties : construction of a vessel sufficiently light , strong , and staunch to contain the gas , and when this was accomplished , to secure the filling of it with dry gas at the requisite degree of pressure .
As regards the first difficulty it was soon found that vessels constructed of more than one piece were only reIiable wheu joined with very carefully joints .
Several explosions , which but for utions might have been disastrous , occurred owing to various forms of soft and hard soldered , overlap joints tearing open in the steam .
The ordinary form of stop-cock for admitting the gas proved useless .
At high pressures it was impossible to avoid leakage .
A screw valve was accordingly designed .
This has proved very satisfactory , and may be so contrived as add hardly more than one gramme to the mass of the sphere .
Finally , copper was found to be so satisfactory in all respects that its use was decided upon .
It is surprisingly tough , not scattering on explosion , but tearing open , and admits of being spun into very thin hemispheres .
So far as thermal capacity is concerned , there is not much choice , the densities of metals being about inversely as their specific heats .
Spheres A and are built up of hemispheres meeting within a narrow belt of their own thickness .
This belt is widened at one point , and here the screw valve is soldered in its place .
For this soft soldering is permissible and convenient .
The belt is brazed in with the utmost care to secure a deep and staunch joint .
The screw valve is shown in section in the cut ( fig. 1 ) .
The valve is opened by screwing back the central pin , which is of hard tempered steel , terminating at one end in a burnished conical point , MR. J. JOLY ON ]CIFIC HEATS which crushes into the soft brass when the valve is closed .
The other end of the piu is fitted with a little eye for suspending the whole from the wire in the calorimeter .
The sphere can be attached to the compression pump by the side tubulure .
While being filled there will be leakage along the thread of the valve pin , if the thread be 1lot choked by strong wax or shellac run in while the valve is warmed with a spirit flame .
To avoid this trouble , and to st , ill further lighten the valve , I devised the form ( fig. 2 ) .
In this the valve , as introduced into the calorimeter , terminates with the cross-hatched part of the figure .
The remainder is a fitting for attaching it to the pump .
The central spindle is surrounded by a stuffing box , and consists of a forked key for gripping and turning the valve pin .
A longitudinal slot in the latter , extending just to the cone , allows of the free entry or exit of gas when the cone is lifted from its bed .
The sphere before use is attached to a small hydraulic hand-pump carrying a metallic gauge , and tested to the requisite pressure by pumping water into it .
In this process the form of the vessel visibly becomes more spherical ; the internal volume may increase nearly three per cent. Leakages are now detected and marked .
Subsequently these are staunched with soft solder , applied while a vacuum is maintained in the sphere .
The sphere will now probably be quite gas-tight , but it must be further tested when filled with gas by immersing it in water , and observing for some time if any bubbles grow upon its surface .
The catchwater of platinum foil with which each sphere is provided , is arranged to be easily removed .
It is , in fact , hooked on to two little wire loops soldered to the sides of the sphere .
The idle sphere , or counterpoise , is not , of course , tested .
It is only necessary to make sure that it is perfectly air-tight .
It is closed when full of air at normal pressure , and is as similar to the active sphere as regards material , weight , and volume , as possible .
The equality of volume should not be rted from to an extent exceeding three or four cubic centimetres .
Their thermal values are adjusted by inserting calculated weights of copper wire into the idle sphere .
It is easy to secure that this in the first instance shall be the lighter of the vessels .
More than an approximate equality in the capacities of the vessels is rdly requisite .
OF GASES AT CONSTANT VOLUME .
79 Differential experiments on the thermal behaviour of the spheres are in any case requisite at intervals during a research .
The Storage of the The gas\mdash ; the earlier experiments were on air\mdash ; was at first pumped into the sphere by a single acting vertical-cylinder pump , of the kind formerly much in use for liquefying carbonic acid gas .
It was found that the gas , although supplied dry to the pump , not only took up moisture from the oil lubricating the piston , but took up the oil itself Two series of experiments , effected in the single calorimeter by the successive differential method , were carried out a- this time on air .
A filtering tube had been inserted between pump and sphere , but the odour of oil from the air when being released proved that this had failed to clean the air .
The experience with the single calorimeter , too , showed that some improvement in dealing with the bulky vessels was desirable .
The calculations for the error of buoyancy involving density of the steam , of the air , and of the hygrometric state of the latter , as well as the thermal expansion of the sphere , were risk and laborious .
And so the differential calorimeter was devised .
A larger and more effective filtering tube was next attached above the pump , and the experiments appearing in a preliminary note*were carried out .
It was subsequently found that the dryness of air so stored was open to doubt , and it was resolved to seek some different means of toring the gas .
The apparatus at present in use was accordingly designed .
This apparatus consists of a receiving apparatus and .
an oil pump .
The compression pump formerJy in use was converted into an oil pump by adding above the cylinder suitable valves and an expansion chamber , after the manner of the ordinary force-pump .
The pump is driven by a flywheel turned by hand , and can pump oil up to pressures of 35 or 40 atmospheres .
This pump is attached to the receiving apparatus ( fig. 3 ) by the two tubes leading to the right of the figure .
The receiving appatatus consists of two strong glass vessels\mdash ; one placed at a level higher than the other .
These are 11 mm. thick in the walls .
The internal volume of the upper one is 335 , of the lower one 344 .
It is desirable that there should be this of volume between them .
These vessels , which are about 5 .
in internal diameter , have , at the lowest estimate , assuming the glass to be well annealed , a bursting strength of some 180 atmospheres .
The glass receivers are carried in cast iron frames , furnished each with two slowthreaded screws , placed with their axes truly in line .
The extremities of these screws are rounded , and bear against cast iron stoppers closing the ends of the receivers .
In this way the uniform bearing of the stoppers against the ends of the receivers is secured .
Leather washers , soaked in wax , are interposed next the ] .
With this * Roy .
Soc. Proc vol. 45 , p. 83 .
arrangement , by screwing down the upper milled-headed screws , and without oesort to the more powerful screws below , perfect immunity from leakage at the preesures is attained .
Fig. 3 , All the connections are effected through the stoppers , as is seen in the fi rece verse communicate one with another by a bent wrought-iron tube , oonnectin th gure .
The two undermost stoppers .
The stopper of the lower receiver has a single OF GASES AT CONSTANT VOLUME .
nection with an upright wrought-iron tube , terminating in a three-way stop-cock .
Through this the lower receiver is put in connection with the high-pressure tube ( of lead ) leading to the oil or through a glass tube with the little oil tank , placed above , and only partially seen in the figure protruding from behind the apparatus .
Or the high-pressure tube from the pump may by this stop-cock be placed in connection with the oil-tank directly , the receiver being then closed .
The lower stopper of this receiver is fitted with a wrought-iron stop-cock , commanding the filling or emptying of the receiver with mercury .
The second and uppermost receiver has two stop-cock c , onnections above ; one , to the right , with the train of drying and purifying tubes through which the gas to be pumped is drawn before it enters this receiver ; the other , to the lefb , leads to a high-pressure wrought-iron drying tube , of narrow bore ( 1 cm .
) , at the further extremity of which is the screw fitting for taking the tubulure on the valve of the copper sphere .
is taken to keep the passages which are drilled in this uppermost stopper of the higher receiver , fine in bore , about 1 mm. The method of applying the apparatus is as follows :Clean dry mercury is let into the lower receiver .
When this is just filled the mercury rises in the lower neck of the upper receiver .
The tank must now be supposed to be filled with clean Rangoon oil , and the three-way cock connecting the pump and the receiver .
On now working the pump oil is drawn from the tank through a glass tube ( this is the higher right hand tube in the figure ) , and returned by the pump into the receiver .
The mercury in this is displaced into the upper vessel , ling the gas in it into the dryin -tube and sphere .
When the higher receiver is quite filled with mercury , owing to the greater content of the lower vessel , some three or four centimetres ' depth of mercury remains in the latter .
Closing now the left hand stop-cock of the upper receiver , setting the three-way cock to permit the oil in the lower one to flow into the tank , and finally opening the stop-cock connecting the upper receiver with the train of drving-tubes , the flows back quietly by its own weight into the lower receiver , and gas is as quietly drawn into the upper one , to be again expelled , by the same operations , into the sphere .
The volume of the sphere being some 160 .
and of the drying-tube 31 c.c. , one operation of the pump displaces sufficient gas to raise the pressure in the sphere by atmospheres .
In the knowledge of this I have dispensed with a pressure gau ge in the use of this pump .
But a gauge of small capacity attached between the pump and the high-pressure drying-tube is undoubtedly desirable , in case of leakage or of a mistake in reckoning the number of operations .
In the use of this pump the air only comes in contact with the mercury and glass of the upper receiver .
After the pump has been in use some time , however , the tnercury , judging from a certain dimming of the glass at the lower part of the receiver , appears to become to some extent greasy .
This is much retarded by screwing a short tubulure of soft wood into the central ation of the under stopper of the lower NDCCCXCI .
vessel , extending some 2 .
up into the mercury , which has all to flow out through this tubulure .
This greasiness only does harm in so far as it may communicate moisture to the gas .
No volatile hydrocarbon is taken up .
To investigate this question a careful experiment was kindly carried out by Mr. WERNER in the chemical laboratory .
A quantity of litres of air compressed into the sphere , and which before entering the pump had been passed through a potash bulb to free it to a great extent ffom the normal of the atmosphere , was very slowly allowed to escape through a heated combustion tube 40 .
in length , filled with platinised asbestos , in order to burn any ocarbons into carbon dioxide .
The issuing gas was bubbled slowly through baryta water .
At the end of the operation a hardly appreciable opalescence ( soluble in hydrochloric acid ) was visible .
As a small proportion of the normal in air , escaping the potash bulb , would have caused he slight precipitation observed , there are no grounds for suspecting any appreciable impurit .
contracted in the receiver .
The air compressed in this apparatus is , also , I may observe , quite free from odour .
The final drying of the compressed air is , however , necessary .
Before the addition of the drying-tube ( which is filled with dried asbestos and phosphoric anhydride ) , an experiment was made by passing grms. of air ( which had been dried through three large -tubes of phosphoric anhydride , and compressed directly from the receiver into the sphere ) through a carefully counterpoised drying-tube , which through- out the operation was maintained suspended upon the balance .
An increase of milligrammes in the weight of the tube was obtained .
As some 2 milligrammes would , on receiyed principles , saturate the volume occupied by the air in the sphere , this error could not be suffered to remain .
The tube was therefore added .
The experiment was subsequently repeated , grms. of air being passed through the carefully equilibrated drying-tube .
The result was absolutely nil .
Later , the experiment was again repeated on grms. of air , The result in this case was unexpected .
There was a slight , but appreciable , loss of weight .
This can only be explained in the more complete drying of a minute plug of cotton wool placed in the tubulure of the drying-tube to hinder the free entry of air while the weighing was being effected .
It may be assumed from these observations , I think , that the various gases compressed in this pump , if delivered pure into the receiver , derived no impurity in being transferred to the copper sphere ; and it will only be necessary to dwell in each case on the means taken for generating and purifying the gases .
Procedure in Making the Experiments .
The sphere having been repeatedly ' washed out\ldquo ; with the gas to be dealt with , in a dry , pure state , is hung from the wire in the calorimeter against its counterpoise , the idle.sphere , and the difference in weight observed .
In the process of " " washing out\ldquo ; the sphere is both heated ( by a spirit flame ) and exhausted between each OF GASES AT CONSTANT VOLUME .
83 admission of the pure gas .
During the first equilibration it is filled with the pure gas at normal pressure .
It is now screwed uP to the pump and the requisite quantity of gas pumped into it .
It is , hen returned to the calorimeter , and when all is cold the added weight of gas\mdash ; subject to a ] correction\mdash ; found by difference .
An experiment cannot be performed till the thermometer is stationary\mdash ; meaning thereby a state in which the motion of the thermometer cannot be perceived in observations made at intervals of some ten minutes .
All openings in the calorimeter are now closed and the balance set in active vibration to , the air in the calorimeter stirred , and thus secure greater ormity of temperature .
I have lately in order to attain the same end used a small stirriu .
apparatus consisting of a little fan driven round by clockwork .
This mounted at the extremity of a stopper which enters the calorimeter by the wide opening at the back , through the steam is admitted .
It is withdrawn at the moment of turning the current of steam into the calorimeter .
The motion of the fan keeps the air in the calorimeter in active motion , while all openings are closed against the ingress of external air .
While steam is being got up in the boiler the thermometer is observed at intervals .
It will generally be found to rise a little .
The final reading before admitting steam is taken as that of the spheres .
The variation will be small , seldom more than of a degree .
The observation of the behaviour of the spheres in the steam takes about five minutes .
If loose weights have been inserted in the idle sphere it will be found that its final assumption of the steam temperature will probably be delayed beyond that of the filled sphere .
The inequality in the rate of heating reveals itself in the gradual acquisition of weight of the idle sphere .
This effect was remarkably conspicuous in the experiments upon result depending , doubtless , on the high conductivity of that gas .
* In the course of minutes the rate of increase of the sphere is diminished so as to be no longer perceptible , but the balance is observed till the fifth minute closes .
This supposes no increment The increment generally indicates a weak burner , or leaky joints , so that the steam is free from -or a preponderance of the radiation of one sphere over the other .
The first evil is met by the use of a powerful close-fitting joints and long preliminary clearing ( 7 to 10 minutes ) of the boiler and steam ways with the live steam .
The second evil is generally caused by one of the spheres ning greasy from handling filled other .
seldom touched ) , and is met by freeing the surface from grease by washing well with dilute ammonia .
A.lthough occa , sionally troubled in my earlier by this evil , I have now for many * It might be suggested that this dilference in the rate of heating of the spheres would give rise to an elror in estimating the initial temperature , except in the case where this was actually .
I may serve , however , that a comparison of a number of the results of observations in the cases where was rising with the oases where was falling revealed no settled difference in ono ection .
The rate oi change of is very slow .
months , using the calorimeter from twice to four times in the day , withIout intermission of a day , only observed it on two or three occasions .
It can , indeed , be eliminated\mdash ; very closely\mdash ; by calculations based on continued observation.of its value On the Steam Calorimeter .
cit but this necessitates the retention of the vessels in the steam an excessive time , which is to be avoided as the stress on the active sphere is then often very great .
The upper limit of temperature is most conveniently observed by a thermometer which is inserted in the calorimeter , in place of the low-range one , just before steam is let in .
When the experiment is concluded the spheres are simply dried in a soft clean cloth and laid aside to cool .
The heat remaining in them completes the drying of their surfaces very thoroughly without further trouble .
When all is again cold , the spheres are restored to the calorimeter and an observation made on their equilibration .
This is an assurance against leakage having occurred during the experiment .
Absolute return to the former equilibration is , perhaps , exceptional , but the departure from it is generally small , some two or threetenths of a milligramme , and will as often appear to indicate loss from the one sphere as from the other .
When the of experiments deemed requisite is completed a careful equilibration of the spheres is made and some or all of the gas let out .
If it is desired to make a further investigation upon this of gas at a lower pressure , the desired amount is retained by suffering the escape of gas to occur slowly while the sphere is humg from the balance , the weights being so adjusted that there be equilibrium when the proper amount of gas remains in the sphere .
after a series of investigations , by the addition of all the weighed quantities which have been suffered to escape , the weight of gas put in in the first instance is found .
This value , as be , is often found to differ by a small quantity from the first determined value of the weight .
The erence is generally negligible , except in the case of hydrogen , where the total weight is small .
In this case , as the second determination of is more reliable than the first , it is accepted .
In the first determination there is evidently some chance of the vessel contracting some impurity upon its surface in the process of filling .
The thermal comparison of the spheres at normal pressures is now effected .
I have generally thought it inadvisable to increase the weight of gas dealt with by exhausting the active sphere .
In the process of exhaustion there was risk of spoiling the weighings by contamination of the surface of the sphere .
There was also the of a back-leakage of air into the sphere the process ; injurious , either by the fact that it was undried or as an impurity if other gas was being dealt with .
The number of observations made on the thermal equilibrium of the " " empty\ldquo ; spheres is some five or six generally .
In the earlier experiments these observations were more frequent , as trifling aIterations on the valve or for leakage , \amp ; c. , were then often necessitated .
OF GASES AT CONSTANT VOLUME .
85 Corrections to be Apphed to the Experime , Results .
The direct results of experiment are subject to some corrections befcre they can be a , pplied to the calculation of the specific heat of the gas .
The conditions of experiment involve that the two spherical sels have their specific heats compared when each is filled with gas at the pressure of one atmosphere , and again when a further weight ( grammes ) of gas is compressed into one of them .
It is necessary to inquire whether the presence of the added quantity of gas does not give rise to any thermal effect other than that due simply to its capacity heat at constant volume between the limits and 1 .
Correction for Thermal Expansion of the Vessel.\mdash ; The voiume of the sphere increases during the experiment by the thermal expansion of the copper .
The gas in the sphere is not , in fact , maintained at absolutely constant volume .
The thermal effect of this on the added quantity of gas , , is evidently that which is to be considered .
The initial pressure due to is affected negabively by the increased volume of the sphere , positively by the rise of temperature .
Work , , is done by the expansion of the gas through the increased volunle of the sphere , and the thermal equivalent of this must be deducted from the experimental result .
CaJling the initial volume and the initial pressure , at any moment during the rise of temperature .
and , where is the coeflicient of of pressure at constant volume , and the coefficient of thermal expansion of copper .
From this a complete formula for the correction is found\mdash ; But for this an approximation may be substituted .
A good approximation is found to be\mdash ; The correction is subtractive .
A numerical estimate of its value is given further on .
2 .
Effect of the Elastic Distension of the Vessel.\mdash ; In considering the last correction it was assumed , for the sake of simplicity , that the copper sphere is so rigid as not to yield to the increased pressure exerted by the gas as it rises in temperature .
obtain exact data in calculating this correction , experiments were made on the sbrinkage of the spheres when a known weight of gas was liberated .
These experiments consisted in carefully weighing the sphere in water with a sinker before and after releasing a certain weight of air .
These are given later , with data connected with the spheres .
If the increase of volume by distension is estiinated for the change of pressure , the usual equation for work of expansion gives the amount of the correction when reduced to its heat measure as a weight of steam precipitated .
The correction is subtractive .
3 .
Effect of the Material of the Sphero.\mdash ; The increased pressure of the gas in the sphere has the effect of distending the copper sphere , as just seen ; this , according to well known principles , results in a cooling of the copper .
Experiments have been made by JOULE , EDLUND , and others , on the amount of this cooling in wires exposed to suddenly-applied traction .
The cooling of the copper in the present case is made good by the condensation of steam .
An estimation , however , of the amount of this correction in an extreme case showed that it was too small to merit consideration .
It is not , therefore , further considered .
4 .
Correction for Buoyancy due to Distension in the Steam.\mdash ; The distension in the calorimeter has also to be considered , as occasioning another correction .
The increased displacement in the steam will tend to decrease the apparent weight of the sphere , and so dirninish the amount observed as the weight of steam precipitated due to the gas .
The amount of the correction is the increase of volume due to rise of pressure multiplied by the absolute density of steam at the temperature .
It is additive .
5 .
Correction for Buoyancy due to Distension at .\mdash ; In that series of experiments in which the one sphere is filled with gas at high pressure , the air-to-steam displace .
meant effect of the filled sphere is greater than it is in the series of experiments in which both spheres are at the same pressure .
The distension due to grammes of gas at is multiplied into the relative density of air and steam .
As the correction results in a greater apparent increased weight of the filled sphere on the admission of steam into the calorimeter , the correction is subtractive .
6 .
Correction for Unequal Thermal of the Spheres.\mdash ; This correction has been already alluded to .
It is effected on a series of experiments the mean results of which are taken , giving a correction grammes of steam over a range .
In applying it the amount of is adjusted for the mean temperature range of the series of experiments under consideration .
It must be effected before the next correction .
7 .
Reduction of the Welght of Precipitation to Vacuo.\mdash ; By multiplying the observed weight of precipitation less into the absolute density of steam , this correction is effected .
It is additive .
The analogous correction for the displacement in air of the weights added to the balance to equilibrate the itated steam is unnecessary , if the weights are , as usual , of platinum or gold .
The total result of these corrections is to reduce the direct measurement by about per cent. In estimating the relative importance of these corrections a few figures will be of service .
I take the case of the series Air II .
, in which the stronger sphere was used .
OF GASES AT CONSTANT , 87 The data Volume of sphere at and 760 mm. internal pressure c.c. ( by experiment ) .
Distension due to grms. of air at C. ( atmospheres ) ( by experiment ) .
In the series Air II .
:\mdash ; Mean , Mean , Weight of air W. has first to be found , by a subtractive correction for temperature ( of using the coefficient of expansion , applied to the measured volume at , and by an additive correction for distension effected on a first rough estimate of pressure at spheres in this case .
This last correction amounts to .
Finally therefore ralue V is found to be In estimating , the weight of a volume of air , at the temperature and at the pressure 760 mm. is divided into W. This giyes the approximate pressure , due to , as atmospheres .
The thermal expansion over the range , is calculated to be From these values the correction ( 1 ) is first calculated .
This reduced to its thermal equivalent grammes of steam precipitated is found to amount to a correction of To estimate the second correction , that for the effect of elastic distension uf the sphere , the pressure at the temperature is found by the coefficient of increase of pressure , ( REGNAULT ) .
It is found to be and the distension due to the increase is , by calculaton , .
Taking as and as , the correction is calculated and found to be .
of steam .
The buoyant effect of the distension in the steam ( 4 ) is found to .
; and the correction for .
the air-to-steam displacement effect of the distension due to is In this series , Air II .
, the weight of precipitation due to the unequal thermal values of the spheres is , on four experiments , found to be over the range .
AB this excess is on the idle sphere it is a correction .
Calculating the proportional correction over the range and adding this to the observed precipitation , the correction for displacement of the precipitation in the steam is with sufficient accuracy applied to the result .
It is .
The final result is a correction of , which is taken as 1 milligrm .
, as the balance reads reliably no closer than to the fourth place of decimals .
The true weight of precipitation to be ascribed to the calorific capacity of the air enclosed at constant volume , can now be arrived at .
Having calculated thus the corrections for one series of experiments effected in a particular vessel , the labour of future calculations be much lightened by basing the new corrections on the old , proportionately to pressure and range of temperature .
R. J. JOLY ON IFIC HEATS In the calculation of the corrections in the of carbon dioxide , which departs very sensibly from BOYLE'S law , the estimation of pressure is based on the results of the experiments of ANDREWS ( ' Roy .
Soc. Proc vol. 24 , p. 458 ) on the departure of this gas from conformity with BOYLE'S law , and on its coefficient of expansion at constant volume at various pressures and temperatures .
The last data are not as full as would be desirable .
8 .
A correction on the apparent weight of gas compressed into the sphere is necessary in the case of hydrogen where the total weight of gas is small .
The correction arises from the distension of the sphere due to the pressure , and is equal to the amount of distension multiplied into the density of the air at the time the weighing is cted .
Whether it is positive or negative will depend upon whether the weight has been determined upon the occasion of filling the sphere or emptying it .
It is negligible in the case of a gas so dense as air .
Thus in the numerical example I have given the distension .
affords a correction of .
This , in a weight of gas of grms. , is outside the value to five figures of the specific heat .
In the case of hydrogen the same distension would arise for a weight of but one-third of a gramme .
The correction has therefore been applied in all cases to the weights of hydrogen .
Calculation of the Density and Pressure .
On Plate 6 will be found charts giving a graphic representation of the results of the experiments .
In these the mean result of each series is taken as ordinate , the mean absolute density ( mass of one cubic centimetre ) of the gas as abscissa .
The mean absolute is the ratio , where is the total weight of gas enclosed in the sphere , V the mean volume during experiment .
The total weight of gas is obtained with sufficient accuracy by adding to the measured weight of gas the ated weight of a volume of the gas equal to the volume of the sphere , and at the temperature and pressure ( approximately ) obtaining in the " " empty\ldquo ; sphere at the time when the weight of gas was being determined .
The volume is obtained fro1n the standard observations on the sphere , corrected for thermal expansion to the temperature ( i.e. , to the mean temperature of the experiment ) , and the distension due to the total weight of enclosed gas at the mean temperature of experiment.as before .
This is a troublesome estimation , as it has to be independently calculateq for each series .
Simultaneously with calculating the last constant we arrive easily at an estimate of the actual or total pressure in the sphere at the temperature , that is , the pressure due to the total weight of gas in the sphere , us above , at the actual mean volume , as above , and at the mean temperature of experiment .
The numbers obtained the ssures in the sphere can only be considered as approximate .
This OF GASES AT CONSTANT VOLUME .
89 more .
applies to the case of carbon dioxide , for which gas I used , however , as far as possible the determinations of ANDREWS , as in the calculations ( ante ) of pressure and density for the evaluation of the errors .
Particulars of the Spheres used to contain the and other particulars .
Sphere A.\mdash ; Weight grms. , .
]ighter than its counterpoise ; internal volume .
at .
This sphere appears to have increased permanently in volume from the effects of the experiments .
The above is from a recent observation , and , although no comparably accurate observation was made before the experiments , a reliable observation of the difference between its external volume and the external volume of its counterpoise , made before the experiments , afforded .
; one made recently gave , an increase of in external volume .
Its shrinkage was determined recently in two experiments .
In one the release of grms. of air at afforded a shrinkage of .
; in the second experiment the air released weighed , and the shrinkage was .
Temperature of the air .
These results agree closely , and a mean result is accepted .
Sphere B.\mdash ; Weight without catchwater grms. ; internal volume , measured June , 1890 , at ; measured May , 1889 ( before the experiments recorded in this paper ) at .
There has therefore been no permanent set in the distension of this sphere .
Its volume differs from that of its counterpoise by .
; its weight by 6 grms. Its distension has been measured by two experiments ; one , in October , 1889 , affordin c.c. for a pressure of atmospheres , or a weight of air at .
A second experiment ( June , 1890 ) showed a shrinkage of for the release of grms. of air at .
These agree very closely ; the first is taken for calculating the The Thermometer.\mdash ; The thermometer used in obtaining the initial temperatures throughout , is one of NEGRETTI and ZAMBRA'S , divided openly to tenths of a degree , and standardised at Kew .
At the close of the experiments it was compared with a recently standardised thermometer , when no appreciable error was detected .
The upper limit of temperature was in the earlier experiments obtained from observation of a fine barometer corrected for capacity by a movable scale .
Subsequently a standardised thermometer of NEGRETTI and ZAMBRA'S , reading to tenths , was used .
Comparing its indications with the readings derived from the barometer , no appreciable difference was found to obtain .
Value ascribed to the Heat of Steam.\mdash ; The latent heat of steam is assumed to have the values determined by REGNAULT , and is derived from his formula , for each experiment .
This constant has been investigated by , BERTHELOT total heat ) , FAVRE and SILBERMANN .
These agree closely , but all are a little lower than REGNAULT'S results , and would tend to lower the results of the experiments by a very slight .
( inappreciable amount .
The Experiments on Air .
The results of the first experiments*on air effected in the differential calorimeter , were to ascribe to it a somewhat higher specific heat than subsequent experiments have .
The air in these experiments was compressed in contact with oil , and hence moisture and impurities in the air are probably the source of difference .
These experiments , too , appeared to indicate invariability of the specific heat over a considerable range of pressure .
Subsequent experiments , which from the very fact of the existence of the discrepancy have been multiplied in number , disprove this result .
There is an undoubted diminution of the specific heat with decreasing pressure or density .
A plausible explanation of the first result exists in the very certain presence of moisture in the gas , taken in conjunction with the course of procedure adopt , ed. The lower pressures were obtained in fact by releasing a certain quantity of the gas between the first and second series of experiments .
Now if moisture be present , the effect of this procedure is to concentrate the impurity at the lower pressure , for on the moment of release there is an intense chilling of the vessel , which will have the effect of precipitating any moisture in suspension in the gas .
A small thing will conceal the decrement of thermal capacity , as will be judged from the succeeding experiments .
I may point out here that over and above the quantity of moisture ( calculated to be some 2 milligrammes ) necessary to saturate the volume of 160 .
of air at the normal temperature , , the effect of moisture is to cause very serious error .
Water , in the form of a liquid , within the sphere must , in fact , be conceived as adding its latent heat of vaporisation to intensify the error .
For every milligramme of water evaporating within , there will be thrown down externally a corresponding milligramme of steam .
I have already described the results of observations which showed that even with the use of a much safer apparatus for effecting the compression of the gas , a considerable weight of water was , in some way , taken up by the gas in its passage through the pump ; so that it was found essential to re-dry the gas immediately before its entrance into the sphere .
The presence of moisture in the preliminary experiments is therefore very probable .
Not at first fully recognising the probability oferror in the first results , it was some time before I recognised the possibility of a variation of the specific heat with pressure .
The ordinary contention that from REGNAULT'S experiments on air at constant pressure the conclusion is legitimate that at constant volume there is no variation of specific heat , was so ( apparently ) forcible , that any other result was deemed improbable .
In truth , however , the assumption of perfectness for air is at variance with the facts , and it is probable that the variation of specific heat is an outcome of its imperfectness .
Later on , the experiments on carbon dioxide , a typically * ' On the Specific Heats of Gases at Constant Volume ( Preliminary Note OF AT CONSTANT VOLUME .
91 imperfect gas , revealed a much more marked variation of specific heat with density , which by analogy tends to confirm the results on air .
A still more remarkable evidence is afforded by the behaviour of hydrogen , as will be seen later .
The first two series of experiments on air succeeding those published , effected on separate samples of air , brought the specific heat by some 2 per cent. ( The results in the Preliminary Note are not corrected for errors of expansion , \amp ; c and revealed a variation following the density .
With the improved conditions under which my experiments are now carried on , the more recent results are again somewhat at variance with these two , the first effected with the use of the improved pump .
I think it best , in the case of each gas , to give the experiments in order of date .
To effect the drying and purifying of the air in these and the succeeding experiments , it is passed through a train of -tubes , consisting of a tube containing pumice soaked in strong caustic potash , a tube containing pumice soaked in strong sulphuric acid , and then through three tubes packed with dried asbestos fibres , separated by layers of phosphoric anhydride .
These tubes are closed with ground stoppers , are some 2 cms in diameter , and afford an effective length of 40 .
each .
Finally , between receiver and sphere , the high-pressure drying tube , having a length of 40 , and a capacity of 31 , also containing phosphoric anhydride , removes what moisture may be taken up in the receiver .
Regarding the mode of tabulating the results of the experiments , it should be mentioned for clearness , that the column headed gives the total apparent weight in grammes of precipitation observed in each experimenb .
The amounts of the corrections , in grammes , to be applied in each case , are given below the ] umns , as well as the mean of the column adjusted for all corrections .
This is designated .
On this number , on ( the weight of gas ) , on the mean of the column ( the latent heat of steam ) , and on the mean range , the final result is calculated , which is given at the foot of the table .
The number given as the specific heat is the specific heat compared weight for weight with water .
The last column , ) , is added for the purpose of affording a means of judging of the agreement of the experiments inter se .
MR. J. JOLY ON THhl SPECIFIC HEATS AIR I.\mdash ; W grammes .
Mean Absolute Density , .
Mean Pressure , Atmospheres .
Rate of correction for spheres for .
Sphere B. Corrections for thermal expansion , \amp ; c. Deduced specific heat At the conclusion of this series , the air enclosed was passed out through a weighed drying tube , and assurance of the freedom ( sensibly ) from moisture of the air thereby obtained , as previously described .
Three experim'ents on the spheres at normal pressures were made after the release of the air .
COBRECTION for Spheres ( Air 1 Mean range , 8652 .
Mean In the table the sign of indicates whether the excess precipitation is on the active or on the idle sphere ; when in the former it is marked .
- , being then subtractive as a correction .
OF AT CONSTANT VOLUME .
93 AIR II .
grammes .
Mean Absolute Density , .
Mean Pressure , Atmospheres .
Rate of correction spheres for .
Sphere B. Corrections for thermal expansion , \amp ; c. Deduced specific heat Some alterations been made to the spheres , a fresh set of observations on their state of thermal equilibrium was necessitated for obtaining the correction for the last experiments .
CORRECTION for Spheres ( Air II Mean range Mean These two series of experiments will be found plotted the highest points on the Air Diagram , Plate 6 .
They show a fall of specific heat with diminishing density , but from after results it appears the rate of decrement is somewhat exaggerated .
In subsequent experiments the number of observations in each series was increased .
An attempt was now made to fix points in the line of specific heat at lower pressures .
The sphere A was taken into use for the purpose .
The number obtained in series III .
for the specific heat agrees fairly with latel observations .
The succeeding series , IV .
and V. , are still more in harmony with the best observations made so1ne months later .
Although , on the small masses of air used in these series , very definite conclusions as to the exact position of the true line cannot be drawn ; yet , from the number of experiments made and their agreement among themselves , it is probable that they are a fair approximation to the truth .
A ance at the diagram shows that what is very probably the true line passes fairly well through the mean of IV .
and V. In the series III .
, IV .
, and V. , the air was separately compressed for each series , and in the case of series III .
it was examined for the presence of moisture in the manner before described .
AIR III .
grammes .
Mean Absolute Density , .
Mean Pressure , Atmospheres .
Rate of correction for spheres for .
Sphere A. Corrections for thermal expansion , \amp ; c. Deduced specific heat for Spheres ( Air III Mean range Mean OF GASES AT VOLUMII .
AIR IV.\mdash ; W grammes .
Mean Absolute Density , .
Mean Pressure , Atmospheres .
Rate of correction for spheres for .
Sphere A. Corrections for thermal expansion , \amp ; c. Deduced specific heat CORRECTION for Spheres ( Air IV .
) .
Mean range Mean AIR gramme .
Mean Absolute Density , .
Mean Pressure , 681 Atmospheres .
\mdash ; \mdash ; Rate of correction for spheres for .
Sphere A. Corrections for thermal expansion , \amp ; c. Deduced specific heat CORREOTION for Spheres ( Air V Mean range Mean Some months after the completion of the foregoing observations , when under more favourable conditions ( see Introduction ) the results ofexperiments were visibly improved consistency , and more experience had been obtained , I resumed the observations on air .
The series VI .
, VII .
, VIII .
, and IX .
were carried out .
In these the sphere alone was used .
VI .
and VII .
are on the one sample of air ; so are VIII .
and IX .
That is , the air for the low density series was in each case provided from that used at higher pressure by suffering part of the latter to escape .
As great precautions were taken OF GASES AT CONSTANT VOLUME .
97 ( by the use of freshIy filled tubes , \amp ; c. ) to secure effective drying of the air , this procedure was probably not productive of error ( see p. 90 ) .
In the interval between the series VII .
and , experiments on the equilibrium of the vessels were made .
AIR grammes .
Mean Absolute Density , .
Meau Pressure , Atmospheres .
Rate of correction for spheres for .
Sphere B. Corrections for thermal expansion , \amp ; c. Deduced specific heat AIR VII . .
Mean Absolute DeIJsity , .
Mean Pressure , Atmospheres .
CORRECTION for Spheres ( Air VI .
, , VIII .
, IX Mean range Mean AIR VIII .
grammes .
Mean Absolute Density , 0.027944 .
Mean Pressure , Atmospheres .
OF GASES AT CONSTANT AIR grammes .
Mean Absolute Density , Mean Pressure , Atmospheres .
Rate of correction for spheres for .
Sphere B. Correction for thermal expansion , \amp ; c. Deduced specific heat omitting the first two experiments as and therefore , probably containing some error or inaccuracy , the results for stand thus : .
From these it is , I think , well established that the mean specific heat at the mean pressure of 15 atmospheres has the value to three figures very closely of Taking the means of the last four\mdash ; by far the most reliable results\mdash ; the specific heat of air at constant volume to four fgures , for the mean density mean pressure atmospheres .
The close agreement of the experimental numbers is a strong reason for accepting these last four results as reliable .
If it be asked how far this agrees with theory , it may be answered that theory has more than one value to ascribe to the specific heat , for the data with theory has to deal are not invariable .
Taking the value of the specific heat at constant pressure as ( REGNAULT ) , and the value of ( CLAUSIUS , Mechanical Theory of Heat ' ) as , the number is obtained ; assuming to be 0.23s9 ( WIEDEMANN ) , the number becomes .
But further , in the fraction the exact value of ( the mechanical equivalent of heat ) is still in question .
CLAUSIUS assumes 42355 , but a recent discussion of the results of ROWLAND* and others appears to indicate more nearly a value This brings the fraction to ; which , applied to REGNAULT'S result as the value of applied to WIEDEMANN'S result ( the most reliable ) , gives as .
These are supposed to be the values at 760 mm. pressure .
Basing on the value of and WIEDEMANN'S result for , the number is found for .
Finally it may be calculated on the values , and only , and if the is taken as is found to be Of the above results that value which appears most directly based on the most reliable experimental results assigns as the value of at the pressure of one atmosphere .
Now , assuming that the specific heat is alinear fimcbion of the pressure , experimental data , VI .
, VII .
, VIII .
, and IX .
, afford value 0 for at the pressure of one atmosphere .
On the reality of the variation of the specific heat with change of pressure I may serve that the free of error are almost unanimous in showing such a variation .
The coincidence ( I might say the identity ) of the slope of VI .
and .
with slope of VIII .
to IX .
is remarkable .
It may be objected that the difference of the actual numbers is very small .
I find that the " " probable error\ldquo ; of the specific heat ( so far as the series of values of indicates ) in the case of IX .
, in the case of series VIII .
If the coincidence of error on the lower result with a \mdash ; error on the higher result for the specific heat be assumed , the numbers just coincide .
The lower becomes , the higher .
But the same coincidence has to be assumed for the series VI .
and .
There is , further , the analogy with carbon dioxide , which case there is no room for doubt as to the variability of the specific heat .
In the case of hydrogen it will subsequently be seen there is also very certain variability of the specific heat , but in the opposite direction ; that is , while in the case of air and carbon dioxide the specific heat increases with increasing density , in the ca , se of hydrogen it diminishes .
Now this for all three gases agrees with their behaviour as regards at the pressures covered by the experiments .
It is well known that in the case of air and carbon dioxide the imperfectness of the gases is revealed by departures from BoYLE'S law , small in the case , large in the other , but in the opposite direction to that which is apparent in the case of hydrogen , whose deviation from the law is , as in the case of air , small .
These facts , taken together , are , I think , a strong confirmation of the reality of the * ' ' On the Mechanical Equivalent of Heat ' Amer .
Acad. Proc 1880 .
Thermodynamics , ' by DE VonsoN WOOD , New York , 1889 , p. et seq. OF GASES AT CONSTANT VOLUME .
variation bserved in the experiments on air .
For if there is no corresponding reality to the variation shown by nearly all the experiments on air , it is necessary to suppose some source of error varying in the one direction with the pressure .
Considered now in connection with the other two gases it is necessary to suppose this error in intensity for the more imperfect gas and eversed in for the superperfect gas .
It appears therefore that : The specific heat of air at volume is constant under variation of pressure , but increases with increase of pressure , the limits of pressure 10 to tmospheres ( which the experiments sufficiently cover ) at the mean of If we assume that the function connecting density and specific heat over the range of observation is a linear one , the uation . . . . . . .
( 1 ) affords the specific heat as a function of the absolute , in agreement with the experiments VI .
to IX .
The Experi , ments on Carbon Dioxide .
The gas used in these experiments was generated by the action of " " pure\ldquo ; dilute sulphuric acid on bicarbonate of soda .
It was stored in a large glass holder head of about 50 .
of water .
From the holder it passed through two -tubes containing sulphuric acid on pumice , and three -tubes wibh phosphoric dride packed with asbestos .
Through these the receiver drew the gas as requlred , and passed it on into the sphere through the high-pressure dr.ying-tube filled with phosphoric anhydride .
I give the results in the order in which they were obtained .
It is my intention shortly to endeavour to trace the specific heat of this gas more minutely throughout , and up to its state as a liquid .
The spheres for this purpose ave indeed already been completed .
CARBON Dioxide I.\mdash ; W grammes .
Mean Absolute Deneity , 0:019950 .
Mean Pressure , Atmospheres .
CORRECTIONS for Spheres ( Carbon Dioxide I.-III .
) .
This series of experiments on th and III .
same on spheres also applies to the ensuin two series , II .
samp of gas is used in II .
and III .
CARBON Dioxide III .
grammes .
Mean Absolute Density , Mean Pressure , Abmospheres .
Rate of correction for spheres for .
Spbere A used .
Correction for thermal expa , nsion , \amp ; c. Deduced specific heat \mdash ; CORRECTION for Spheres ( Carbon Dioxide IV.-VI .
) .
Mean range Mean OF GASES AT CONSTANT VOLUME .
CARBON Dioxide V.\mdash ; W grammes .
Mean Absolute Density , Mean Pressure , Atmospheres .
Rate of correction for spheres for .
Sphere used .
Correction for thermal expansion , \amp ; c. Deduced spccific heat CARBON Dioxide grammes .
Mean Absolute Density , Mean Pressure , Atmospheres .
Rate of correction for spheres for .
Sphere used .
Correction for thermal expansion , \amp ; c. Deduced specific heat The foregoing experiments are plot , ted in the diagram for carbon dioxide , Plate 6 .
On the purity of the gas I may observe that , in tho course of compressing the gas for one of the experiments , I drew off some of the gas and tested its purity by absorption wibh caustic potash .
A very bubble of air ( presumably ) remained over .
It was impossible , in the tube used , to ascertain its volume ; but as this was the impurity in 80 , and as from the approximation of specific heats and , indeed , of densities , a small impurity would not affect the result , I did not think it necessary to delay over its exact estimation .
IDCCCXCI .
, on the diagram , the mean of Experiments I. and II .
with the mean of Experiments .
and V. gives a line probably expressing closely the true locus of the specific heat .
From this line the equation for the specific heat as a function of the density is found to be . . . . .
( 2 ) Taking again the means of Experimeuts I. and II .
, IV .
and , as affording reliable results , which seems justifiable , the specific heat of carbon dioxide at constant volume is 6988 at the pressure atmospheres and density , and at the pressure atmospheres and density ; both at the mean temperature C. These nbers differ by so considerable an amount as to render it most that the experiments express other than a real variation of the specific heat of carbon dioxide .
And the conclusion seems justified that the specific heat of carbon dioxide is constant , but increases very with increase of density throughout the of pressure 10 to 25 atmospheres at the mean temperature of I add for convenience a table of the experiments .
EXPERIMENTS on Carbon Dioxide .
The specific heat at the pressure of one atmosphere is by calculation .
The number obtained by CLAUSIUS from the experiments of REGNAULT is .
But apparently CLAUSIUS ' result applies to REGNAULT'S experiments in over the range to .
If CLAUSIUS ' formula*be applied to REGNAULT'S results over the range to , a result departing from the experimental value of in the opposite direction is obtained , i.e. , 0.157 .
The amendment of CLAUSIUS ' formula for the value of raises the last number by a trifling amount .
The cause of the of these results is not apparent .
It is suggestive that in the case of the more perfect gases , air and hydrogerl , the deduction of the one specific heat from the other is attended by no such wide discrepancy .
* Theory of Heat , ' p. 55 .
London , 1879 .
OF GASES AT CONSTANT VOLUME .
The xperiments on Hydrogen .
The difficulties in the of obtaining this gas in a sufficiently pure state proved so great , that what can only be considered as results have been obtained from the experiments up to the present .
It is true thab as the procedure of generating and storing the gas was refined upon , results closely agreeing with the theoretical value of the specific heat were obtained , but always so far differing from each other as to show so difference in the purity of the samples of gas dealt with .
It has , hence , been impossible to assign with certainty a definite experimental value to the specific heat .
The earlier experiments are rejected as having been effected on very impure hydrogen , affording values entirely too low .
A very small impurity by volume will seriously affect the result of experiment .
The error due to an admixture with another gas is , in fact , very nearly proportional to the percentage of impurity by weight .
The hydrogen was generated from granulated distilled zinc , under the action of dilute " " pure\ldquo ; sulphuric acid .
It was purified , in the case of the immediately succeeding experiments , by passing it through a tube containing pumice and a solution of silver nitrate , and then dried by passage through a tube containing sulphuric acid on pumice , and three tubes of phosphoric anhydride , passing a fourth tube of the same nature between receiver and sphere .
From the generating bottle it passed into a glass gas-holder under a head of about half a metre of water for the supply of the pump .
At the conclusion of the Series I. a portion was allowed to escape and a second series of experiments made on the residue .
The gas released from the sphere was in part passed into a eudiometer having a capacity of.50 .
and burned with an excess of electrolytically prepar.ed oxygen .
The residual oxygen was then treated first with caustic potash ( which revealed no carbon dioxide ) and then with pyrogallic acid .
small quantity , estimated later ( by weighing a bead of mercury occupying the same volume in the tube ) as , remained .
The error by volume was estimated as 1 in 350 , or , if the impurity was nitrogen , 1 in 25 by weight .
This will be nearly the maximum error to be ascribed to the thermal result .
The eudiometric experiment , however , is not satisfactory , as there is , I am informed , a liability to the evolution of carbon monoxide from the alkaline pyrogallol .
It was , indeed , frequently found , in subsequent tests of like nature , that it was possible to obtain a second ignition on passing a spark when the absorption had reached a certain stage and apparently ceased .
1 decided in consequence to discontinue these tests , and had hoped to substitute determinations of density as tests of purity in future experiments , taking Lord RAyLEIGH'S recent researches for comparison in judging of the deglee of purity .
The method I had intended to use for the determination of density only necessitated the use of a large volumenometer into which the .
from the ) might be passed , the weight of the gas being determined by successive weighings of the sphere .
This method is described in a forthcoming number of the 'Proceedings of the Royal Dublin Society .
* It surmounts or reduces many of the sourees of error arising in the application of REGNAULT'S method .
However , so far as the present experiments are concerned , I regret to say I have not been able to apply it , as the construction of the volumenometer was not completed in time .
Some preliminary experiments on the application of this method appear to show that it is capable of a hi gh degree of accuracy .
The degree of purity of the gas in the experimenfs from .
onwards must only be inferred from the increasing precautions taken to guard against the entrance of impurities .
That higher degrees of purity were obtained as the ents advanced may , indeed , partly be inferred from the results of he experiments themselves by the observed rise in the deduced specific heats .
My subsequent attempts to test for impurities by absorption were so unsatisfactory and contradictory as to be valueless .
I.\mdash ; W gramme .
Mean Absolute Density , Mean PresSure , 2912 Atmospheres .
Rate of correction for spheres for .
Spbere B. Correction for thermal expansion , \amp ; c. \mdash ; Deduced specffic heat * .
Dublin Soc. Proc vol. 6 , p. 534 .
OF GASES AT CONSTANT yOLUME .
HYDROGEN II .
gramme .
Mean Absolute Density , Mean Pressure , Atmospheres .
Rate of correction for spheres for .
Sphere B. Correction for thermal expansion , \amp ; c. Deduced specific heat HyDROGEN III .
gramme .
Mean Absolute Density , Mean Pressure , Atmospheres .
HyDROGEN I gramme .
Mean Absolute Density , Mean Pressure , Atmospheres .
Rate of correction for spheres for .
Sphere B. Correction for thermal expansion , \amp ; c. Deduced specific heat CORRECTION for Spheres ( Hydrogen I.-VII .
) .
Mean range Mean As seen by the estimation of the nitrogen present in the gas used in I. and II .
, maximum error of 5 per cent. may affect these results .
But this is very probably excessive .
For the succeeding series , .
and IV .
, further precautions were taken in dealing with the hydrogen .
Glass tubes were substituted for rubber as far as possible , the tubes being brought end to end and connected by a thick rubber , varnished externally .
A test for the presence of sulphuretted hydrogen , by passing the gas into a tube containing a filter paper soaked in acetate of lead , revealed no trace of that impurity .
The next two series , V. and VI .
, are upon the one sample of gas .
VI . .
1007 gramme .
Mean Absolute Density , Mean Pressure , Atmospheres .
The Series .
and .
show a fall in the specific heat with increase of density .
The specific heat , too , is raised above that obtained in the preceding series .
Undoubtedly purer hydrogen had been obtained .
What greatly conduced to this an addition to the arrangements whereby the connection of the gas-holder with the generat ng bottle could be maintained throughout the operation of pumping .
Prevlously the coupling with the bottle had to be transferred to the high-pressure receiver .
The new arrangement providing for a supply of gas not limited to the capacity of the gasholder , enabled a much greater quantity to be used as a wash-out in the first instance .
The difficulty of obtaining concordant observations on different samples of hydrogen having become very apparent , I resolved to continue the system of making two observations upon the one sample of gas .
It appeared to me , indeed , that a int of fsr higher interest than the actual value to be ascribed to the specific heat of the gas was the question to the nature of variation with change of pressure .
In the very next series , VII .
, however , my intention of duplicating the experiments was frustrated .
In the gas , in fact , at the conclusion of the series , I detected an odour of nitrous fumes , which induced me to seek the source of the impurity and start a fresh series .
The source of the impurity turned out to be the tube containing the nitrate of silver , which was probably becoming affected by acid carried by the gas .
The tube was suppressed , and a tube containing copper potassium tartrate , carried on pumice , substituted .
The use of this tube entirely removed the odour of which the gas quitting the generating bottle had to some extent , and , in fact , the gas passed on quite free from ] It is improbable that VII .
is affected appreciably by the amount of oxide of nitrogen present .
A very small quantity would have sufficed to impart the odour .
HYDROGEN VII .
gramme .
Mean Absolute Density , Mean Pressure , Atmospheres .
The nexc two series , VIII .
and IX .
, are on the one sample of gas .
This would appear to have been the purest sample of gas prepared .
The specific heat has risen quite to the theoretic value , and the fall of specific heat with increasing density is well marked .
OF GASES AT CONSTANT VOLUIIE .
1 .
13 Series X. , XI .
and XII .
are all on the one sample .
The extremes X. and XII .
agree with the testimony of all the latter experiments as to the nature of the variation of the specific heat with density , but the central series is cordant .
HYDROGEN VIII .
gramme .
Mean Absolute Density , Mean Pressure , Atmospheres .
Rate of correction for sphercs for .
Sphere B. Correction for thermal expansion , \amp ; c. Rate of correction for sphercs for .
Sphere B. Correction for thermal expansion , \amp ; c. Rate of correction for sphercs for .
Sphere B. Correction for thermal expansion , \amp ; c. Rate of correction for sphercs for .
Sphere B. Correction for thermal expansion , \amp ; c. Rate of correction for sphercs for .
Sphere B. Correction for thermal expansion , \amp ; c. Rate of correction for sphercs for .
Sphere B. Correction for thermal expansion , \amp ; c. Rate of correction for sphercs for .
Sphere B. Correction for thermal expansion , \amp ; c. Rate of correction for sphercs for .
Sphere B. Correction for thermal expansion , \amp ; c. Rate of correction for sphercs for .
Sphere B. Correction for thermal expansion , \amp ; c. Rate of correction for sphercs for .
Sphere B. Correction for thermal expansion , \amp ; c. Rate of correction for sphercs for .
Sphere B. Correction for thermal expansion , \amp ; c. Rate of correction for sphercs for .
Sphere B. Correction for thermal expansion , \amp ; c. Rate of correction for sphercs for .
Sphere B. Correction for thermal expansion , \amp ; c. Rate of correction for sphercs for .
Sphere B. Correction for thermal expansion , \amp ; c. Rate of correction for sphercs for .
Sphere B. Correction for thermal expansion , \amp ; c. Rate of correction for sphercs for .
Sphere B. Correction for thermal expansion , \amp ; c. Rate of correction for sphercs for .
Sphere B. Correction for thermal expansion , \amp ; c. Rate of correction for sphercs for .
Sphere B. Correction for thermal expansion , \amp ; c. Rate of correction for sphercs for .
Sphere B. Correction for thermal expansion , \amp ; c. Rate of correction for sphercs for .
Sphere B. Correction for thermal expansion , \amp ; c. Rate of correction for sphercs for .
Sphere B. Correction for thermal expansion , \amp ; c. Rate of correction for sphercs for .
Sphere B. Correction for thermal expansion , \amp ; c. Rate of correction for sphercs for .
Sphere B. Correction for thermal expansion , \amp ; c. Rate of correction for sphercs for .
Sphere B. Correction for thermal expansion , \amp ; c. Rate of correction for sphercs for .
Sphere B. Correction for thermal expansion , \amp ; c. Rate of correction for sphercs for .
Sphere B. Correction for thermal expansion , \amp ; c. Rate of correction for sphercs for .
Sphere B. Correction for thermal expansion , \amp ; c. Rate of correction for sphercs for .
Sphere B. Correction for thermal expansion , \amp ; c. Rate of correction for sphercs for .
Sphere B. Correction for thermal expansion , \amp ; c. Rate of correction for sphercs for .
Sphere B. Correction for thermal expansion , \amp ; c. Rate of correction for sphercs for .
Sphere B. Correction for thermal expansion , \amp ; c. Rate of correction for sphercs for .
Sphere B. Correction for thermal expansion , \amp ; c. Rate of correction for sphercs for .
Sphere B. Correction for thermal expansion , \amp ; c. Rate of correction for sphercs for .
Sphere B. Correction for thermal expansion , \amp ; c. Rate of correction for sphercs for .
Sphere B. Correction for thermal expansion , \amp ; c. Rate of correction for sphercs for .
Sphere B. Correction for thermal expansion , \amp ; c. Rate of correction for sphercs for .
Sphere B. Correction for thermal expansion , \amp ; c. Rate of correction for sphercs for .
Sphere B. Correction for thermal expansion , \amp ; c. Rate of correction for sphercs for .
Sphere B. Correction for thermal expansion , \amp ; c. Rate of correction for sphercs for .
Sphere B. Correction for thermal expansion , \amp ; c. Rate of correction for sphercs for .
Sphere B. Correction for thermal expansion , \amp ; c. Rate of correction for sphercs for .
Sphere B. Correction for thermal expansion , \amp ; c. Rate of correction for sphercs for .
Sphere B. Correction for thermal expansion , \amp ; c. Rate of correction for sphercs for .
Sphere B. Correction for thermal expansion , \amp ; c. Rate of correction for sphercs for .
Sphere B. Correction for thermal expansion , \amp ; c. Rate of correction for sphercs for .
Sphere B. Correction for thermal expansion , \amp ; c. Rate of correction for sphercs for .
Sphere B. Correction for thermal expansion , \amp ; c. Rate of correction for sphercs for .
Sphere B. Correction for thermal expansion , \amp ; c. Rate of correction for sphercs for .
Sphere B. Correction for thermal expansion , \amp ; c. Rate of correction for sphercs for .
Sphere B. Correction for thermal expansion , \amp ; c. Rate of correction for sphercs for .
Sphere B. Correction for thermal expansion , \amp ; c. Rate of correction for sphercs for .
Sphere B. Correction for thermal expansion , \amp ; c. Rate of correction for sphercs for .
Sphere B. Correction for thermal expansion , \amp ; c. Rate of correction for sphercs for .
Sphere B. Correction for thermal expansion , \amp ; c. Rate of correction for sphercs for .
Sphere B. Correction for thermal expansion , \amp ; c. Rate of correction for sphercs for .
Sphere B. Correction for thermal expansion , \amp ; c. Rate of correction for sphercs for .
Sphere B. Correction for thermal expansion , \amp ; c. Rate of correction for sphercs for .
Sphere B. Correction for thermal expansion , \amp ; c. Rate of correction for sphercs for .
Sphere B. Correction for thermal expansion , \amp ; c. Rate of correction for sphercs for .
Sphere B. Correction for thermal expansion , \amp ; c. Rate of correction for sphercs for .
Sphere B. Correction for thermal expansion , \amp ; c. Rate of correction for sphercs for .
Sphere B. Correction for thermal expansion , \amp ; c. Rate of correction for sphercs for .
Sphere B. Correction for thermal expansion , \amp ; c. Rate of correction for sphercs for .
Sphere B. Correction for thermal expansion , \amp ; c. Rate of correction for sphercs for .
Sphere B. Correction for thermal expansion , \amp ; c. Rate of correction for sphercs for .
Sphere B. Correction for thermal expansion , \amp ; c. Rate of correction for sphercs for .
Sphere B. Correction for thermal expansion , \amp ; c. Rate of correction for sphercs for .
Sphere B. Correction for thermal expansion , \amp ; c. Rate of correction for sphercs for .
Sphere B. Correction for thermal expansion , \amp ; c. Rate of correction for sphercs for .
Sphere B. Correction for thermal expansion , \amp ; c. Rate of correction for sphercs for .
Sphere B. Correction for thermal expansion , \amp ; c. Rate of correction for sphercs for .
Sphere B. Correction for thermal expansion , \amp ; c. Rate of correction for sphercs for .
Sphere B. Correction for thermal expansion , \amp ; c. Rate of correction for sphercs for .
Sphere B. Correction for thermal expansion , \amp ; c. Rate of correction for sphercs for .
Sphere B. Correction for thermal expansion , \amp ; c. Rate of correction for sphercs for .
Sphere B. Correction for thermal expansion , \amp ; c. Rate of correction for sphercs for .
Sphere B. Correction for thermal expansion , \amp ; c. Rate of correction for sphercs for .
Sphere B. Correction for thermal expansion , \amp ; c. Rate of correction for sphercs for .
Sphere B. Correction for thermal expansion , \amp ; c. Rate of correction for sphercs for .
Sphere B. Correction for thermal expansion , \amp ; c. Rate of correction for sphercs for .
Sphere B. Correction for thermal expansion , \amp ; c. Rate of correction for sphercs for .
Sphere B. Correction for thermal expansion , \amp ; c. Rate of correction for sphercs for .
Sphere B. Correction for thermal expansion , \amp ; c. Rate of correction for sphercs for .
Sphere B. Correction for thermal expansion , \amp ; c. Deduced specific heat HYDROGEN IX .
gramme .
Mean Absolute Density , Mean Pressure , Atmospheres .
No. No. No. No. No. No. No. No. No. No. No. No. No. No. No. No. No. No. No. No. No. No. No. No. No. No. No. No. No. No. No. No. No. No. .
12017 Rate of correction for spheres for .
Sphcre B. Correction for thermal sion , \amp ; c. Deduced cific heat HYDROGEN X.\mdash ; W gramme .
Mean Absolute Density , 000l9@74 .
Mean Pressure , Atmospheres .
HYDROGEN XI.\mdash ; W gramme .
Mean Absolute Density , Mean Pressure , Atmospheres .
Rate of correction for spheres for here B. Correction for thermal expansion , \amp ; c. Deduced specific heat OF GASES AT CONSTANT VOLUME .
HYDROGEN XII .
gramme .
Mean Absolute Density , Mean Pressure , Atmospheres .
Rate of correction for spberes for .
Sphere B. Correction for thermal expansion , \amp ; c. , Deduced specific heat CORRECTION for Spheres ( Hydrogen VIII.-XII .
) .
Mean range Mean The diagram for Hydrogen , Plate 6 , is a sufficient resum6 of the foregoing experiments .
It will be understood that the lines throughout the diagram connect experiments effected on gas of the same purity .
The two lines low down on the diagram depict the first four series in which there was very certainly a greater degree of impurity than in subsequent series .
The impurity consisting almost certainly of air , its effect will be to conceal the true variaand XIL , agree as regards the negative variation of the specific heat with increased density .
X. and XI .
are discordant .
Nearly all the experiments on the purer gas being at one upon this poin , it is , I think , allowable to assume that these experiments prove that the specific heat of hydrogen is not constant , but varies with the density , diminishing with of density , and vice versed .
As already remarked , this seems in harmony with the nature of the departure of this gas from BoYLE'slaw .
The theoretic value assigned by CLAUSIUS to the specific heat of hydrogen is Experiment VIII .
is just this number , IX .
rises above it .
The rest lie less or more below it .
The principal result of these first experiments on gases is the broad fact of the departure of the specific heat at constant volume from constancy .
Some such result might have been , perhaps , anticipated from the known imperfectness of all gases .
But the mvariability of the two specific heats is in text books most generally insisted on .
It is probable that a gas whose molecules were without action upon one anotherwould show invariability of specific heat .
As it is , it appears that the mutual lence of the molecules of some gases being of a positive nature , that of others ( hydrogen molecules at least ) may be of a negative nature .
But the subject is only touched upon in this first paper .
Many questions remain to answered .
Is the specific heat dependent on temperature limits If the perfectness of a gas depends upon its temperature , how will this affect the nature of the relation between specific heat and density ?
Will the rate of variation alter ?
Presumably it will .
It is not improbable that at very high temperatures the slope of the line for air ( or even for carbon dioxide ) , after approaching more and more nearly to parallelism with the axis of density , will take up the position of a perfect gas , and even assume the negative inclination of the -perfect gas .
ilarly with rise of temperature the slope of the line for hydrogen may become lntensified ; or , if we suppose the range .
to be sinking , the oppos te ay be the course of events , hydrogen ultimately falling into agreement with the behaviour of the other gases .
There may be for each gas atemperature , such that for it the gas behaves in all respects as a truly perfect gas , in which neither a positive nor a negative latent heat would become apparent at .
high densities .
But , again , it may be that a truly perfect gas not possible nature , that is , with molecules of the one kind .
For OF GASES AT CONSTANT VOLUME .
is evident that a gas undistinguishable in its thermal behaviour from a perfect gas could be compounded out of a mixture , in the proper proportions , of say air and hydrogen .
For dealing , at , to some extent , with the question of the ence of temperature on the specific beat , arrangements are bein made , and I am in hopes that reliable quantitative measurements will be found attainable .
Other questions\mdash ; as that of the mutual influence , thermally , of different intermixed gases , will suggest , probably , as fields of research in which the present method may find application . .
Trams .
1891 .
A.Plata 6 .
, hffi

