]\gt ; INDEX SLIP .
PEBMAN , P. .
RAMBAY , WILLIAM , and , J.\mdash ; An Attempt to Determine the Adiabatic Relations of Ethyl Oxide .
Phil. Trans. , , 1897 , vol. 189 , pp. 167-188 .
AlISAY WILLIAM , } NNIIS , J. , and BMAN , E. P.\mdash ; An Attempt to Detennine the Adiabatic Relations of Ethyl Oxide .
Phil. Trans. , , 1897 , vol. 189 , pp. 167-188 .
, J. , RNAN , E. P. , and RAMSAY , WILLIAM.\mdash ; An Attempt to Determine the Adiabatic Relations of Ethyl Oxide .
Phil. Trans. , A. , vol. 189 , pp. 167-188 .
Elasticity , adiabatic , of Ethyl Oxide .
Perman , E. P. , Ramsay , WiUiam , and RoseRose-Innes J. Phil. Trans. , , 1897 , vol. 189 , pp. 167-188 .
Ethyl Oxide , adiabatic relations\mdash ; isentropic curves\mdash ; thermal relations .
Perman , B. P. , Ramsay , William , and -nnes , Ilose-Innes J. Phil. Trans , , , 1897 , vol. 189 , pp. 167-188 .
Isentropic Curves of Ethyl Oxide .
Perman , E. P. , Bamsay , William , and Eose-Innes , Innes J. Phil. Trans. , , 1897 , vol. 189 .
pp. 167-188 .
[ 167 ] VII .
An Attempt to Determine the Adiabatic Relations of Ethyl Oxide .
E. P. PERNAN , D.Sc .
, W. RAMSAY , .D .
, .D .
, F.R.S. , J. M.A. , B. Sc. Received November 6 , \mdash ; Read December 10 , 1896 .
THE object of the research described in this Memoir is the determination of the behaviour of ether , in the state of gas , approaching towards the state of liquid , when no heat is communicated to it , so that its is altered adiabatically .
Some experiments have also been made on liquid ether , an account of which we append .
After several proposals , depending for their success on the expansion of the substance , so that it might cool sensible entrance of heat , recourse had to a modification of KUNDT'S process , whereby the velocity of sound in the liquid or gas is ascertained .
The material for research , pure ether , was prepared and dried as already described in a paper by Dr. YOUNG and one of the authors ( ' Phil. Trans , 1887 .
p. 57 , seq It had a constant boiling-point , identical with that already given .
The experiments were carried out by EDGAR PE .
RMAN and WILLIAM RAMSAY during the Session ] 891-1892 .
Various attempts were made to deduce from the results , by a graphical meXhod , the true position of the curves the adiabatic expansion of the gaseous ether , but with small success .
At a later date , J. -INNES examined the figures contained in the tables annexed , and making use of the fact that , without serious error , the isochoric lines , as in the case of isothermal expansion , may be regarded as straight , contributed the mathematical treatment given in Section 5 .
For want of sufficient knowledge of the thermal data for liquid ether , a similar treatment has not been found possible ; but it is the opinion of the authors that the experimental data should be published , inasmuch as they may be found in future to afford a means of determining the adiabatic relations of a liquid , of which the behaviour in the state of gas is now obvious .
I. APPARATUS EMPLOYED .
The pressure apparatus , , was the same as that employed in bhe former research , but important modifications had to be made , so as to render it fit for its present purpose .
31.5.97 The tube containing the , DG , fig. 1 , lay nearly horizontal , instead of , asbefore , vertical .
The gauge-tubes and were , as before , one registerin pressures from somewhat above that of the atmosphere to about 10 atmospheres ; and a second , which to show pressures of about 9 atmospheres , and continued to give reliable determinations up to 50 atmospheres , for its lower portion was expanded so as to form a reservoir .
The gauges were jacketed by running water , the temperature of which was read from a small thermometer placed in a tube bet , ween the two gauges .
The gauges were , as before , filled with dry air , and a correction was applied taken from the results of AMAGAT , on the coefficient of compressibility of air , by which the Fig. 1 .
pressures were made to coincide with the readings of a theoretically perfect gas Count .
Rend .
The low-pressure manometer was calibrated by comparison up pressure of 3 atmospheres with the readings of the height mercury in a open tube ; and the high-pressure gauge was compared with the low-pressure one , and the pressures so controlled .
Experiments made in this manner gave results ilot appreciably differing from those obtained by calculating pressure from the ratios of volumes of air confined in the tubes to that present at the temperature and pressure of filling them .
Both manometers were calibrated by weighing with mercury .
ON THE ADIABATJC RELATIONS OF ETHYL OXIDE .
169 The temperatures , as before , were obtained by jacketing with the vapours of various liqui\amp ; , boiling under known pressures , and are those of an air-thermometer Chem. .
The " " volume-tube\ldquo ; DGI ( fig. 1 ) requires special description ( see also fig. 2 ) .
Tt nsisted of a piece of barometer-tubing , about millims. in internal diameter , and 36 centima long .
To this was sealed a second piece , of 2 millims. diameter and 15 centims. in length .
The end of the narrower tube was sealed , and blown into a small thick-walled bulb , .
A thin rod of glass , I ( CD , fig. 2 ) , 15 centims. long , was next sealed to the bulb so as to form a continuation of the bulb ; and a piece of similar rod having been dropped into the barometer tube , was.attached by sealing to the interior surface of bulb .
By careful fusion the internal and external rods were so manipulated as to be continllolls with each ot.her , the wall of the tube acting as a diaphragm .
By rubbing the external rod with a rug wet with alcohol , it gave a shrill note , and the vibrations were imparted to that portion in the barometer tube , and set in motion the gas contained in the tube .
It was found very difficult to construct such an apparatus which gave satisfactory vibrations .
It appears necessary to have a fairly wide barometer tube , not too thick in the wall , else the vibrations produced are not those of the gas , but of the glass tube .
After many attempts , a tube was at length found , in which no sign of " " glass-vibrations\ldquo ; was noticeable ; when lycopodium dust or powdered silica was introduced , and the rod was rubbed , the dust arranged itself in well-marked heaps , ding through the whole length of the tube .
The difficulty was caused greatly by the necessity of having a tube sufficiently strong to withstand an internal pressure of 40 or 50 atmospheres , and yet capable of owing the wave-length of the contained gas .
This volume tube was graduated in millims. , calibrated by hing with mercury .
It served in all the experiments made .
To regulate the temperature , this experimental tube was surrounded by ajacket of glass , , also horizontally ; and at the end nearest the pressure apparatus a bulb , , was sealed on , in which liquids could be boiled .
It was necessary to ensure an air-tight junction between the external and internal tubes ; hence it was nec ssary to employ indiarubber corks .
But as most of the jacketing vapours used attack and dissolve indiarubber , the indiarubber cork was faced with ordinary cork , which .otected it from the action of the vapour to a great extent .
The other end of the jacket enclosed the small bulb at the end of the experimental tube , and was also provided with an indiarubber cork , faced with ordinary cork ; through these th rod passed tightly , about 10 centims. of its length projecting outside .
A small condenser sometimes surrounded the vertical tube at the further end of the jacket , which led to a manometer , , and pump , so that the liquid in the bulb could be made to boil under any desired pressure , and therefore ( within certain limits ) at any required temperature .
The liquids employed were chlorobenzene , giving a range of tetnperature NDCCCXCVIL 170 DR. E. P. RlIAN , DR. W. RAMSAY , AND MR. J. INNns , between and 13 ; bromobenzene , giving a range to anlXne , of 16 to 18 ; and quinoline , of to .
Observations ere always made at even temperatures , the pressures being adjusted so as to cause the liquid to give off vapour of the temperature , l20o , \amp ; c. , up to The method of filling the tube also requires description .
First , as regards the powder introduced to show the wave-lengths of the vibrating gas .
In air and similar gases , lycopodium dust gives the most distinct figures ; but it is a resinous substance , and hence ib was not possible to place it in contact with ether .
After many trials of magnesia , aIumina , precipitated silica , \amp ; c. , a quantity of somewhat impure silica was prepared , by glass with carbonate of sodium and , and treatment with hydrochloric acid ; this gelatinous silica was not very well washed , but was dried in a impure state , containing a trace of common salt .
It was ignited , and fritted together to rough lumps .
These were powdered in an mortar , and sifted through gauze .
The small ft.agments of microscopic size were approximately spherical , and flowed easily , not adhering to each other .
Pure silica is too dusty ; sand is too large-grained , or , if powdered , too angular ; to obtain the best results the grains should be round and not too small .
A sufficient quantity of such sand was introduced into the experimental tube ; its open end , ( fig. 3 ) , was then sealed to a bulb , , of the form shown in the woodcut ; and the tube at the seal was out into a wide apillary , C. Etber was placed in the bulb , the open end , , was then connected with a water-pump , and the bulb was exhausted .
The ether began to boil and expelled air from the bulb .
By admitting air , ether entered the experimental tube , ; and the bulb was again exhausted and then sealed at E. The amount of air was then reduced to that dissolved in the liquid ether .
The experimental tube was next warmed , so as to boil the ether it contained ; the vapour rushed up , driving pistons of liquid before it , and condensed in the bulb , which was artificially cooled .
After about ten minutes , all possible air had been expelled from the tube into the bulb .
On gently warming.the bulb , and cooling the tube , , liquid ether entered , and the absence of was certain , inasmuch as the ether closed up rapidly and completely , no race of a bubble being left in the tube .
If even a small bubble were left , dissolving in the ether on its way up the sloping tube , the operation was repeated .
The silica during this operation usually found its way into the bulb ( fig. 2 ) at the end of the experimental tube .
The tube was again warmed so as to expel ether .
And here arose a difficulty .
It was essary to guess at the quantity of ether which } required .
Usually , the experiments were adjusted to suit the particular quantity which accidentally remained .
But with a little practice it was found possible to leave in the tube approximately the amount required .
The capillary junction , ( fig. 3 ) , between the experimental tube and the bulb was then sealed , leaving a thin tail of glass .
The silica was then shaken out of the bulb of the experimental tube ( see fig. I , , and fig. 2 , B ) into the tube itself , and distributed along its length us evenly as possible ON THE ADIABATIC RELATIONS OF ETHYL OXIDE .
and the experimental tube was then screwed into the pressure apparatus by means of the iron cap for the purpose .
It was found that if mercury entered the space between the rod , ( fig. 2 ) , and the internal wall of the experimental tube , , it could not be removed by tapping , after the apparatus bad been set up ; and also that it interfered with the vibrations of the rod .
Hence it was necessary to take utions to prevent mercury entering the tube so far as touch the end of the rod .
The experimental tnbe having been placed in position in the pressure apparatus , and the cap tightly on , the temperature was raised to a given point , determined by the amount of ether left in the tube ; the pressure was then raised by turning the screw of the pressure apparatus ( fig. 1 ) , and when the end of the screw came in contact with the drawn-out capillary of the experimental tube , ( fig. 3 ) , which is there shown unsealed , it broke it .
It was necessary so to regulate the temperature of the ether that only a little escaped when the end was broken .
If the pressure in the interior were too great , too much ether escaped ; if too smail , cold mercury entered , and came in contact with the internal rod .
After some practice it was ) possible to arrange the pparatus successfully nearly every DR. E. P. PERMAN , DR. W. RANSAY , AND .
J. NNES , time .
During the breaking of the capiltary point , the whole .
was as to incline the experimental tube .
The pressure was then raised so as to some mercury to enter the tube , ich was the1l placed in a nearly horizontal position .
By tappi1Ig , the silica was made to flow down , : to forni an evenly distributed layer in the experimental tube ; the temperature was then finally adjusted , and after placing the mercury at a given ma.rk on t , he ube , riments were begun .
At the temperature chosen the ether was wholly gaseous .
On the projecting rod , the ether gas vibrated , and the silica arranged itself in heaps , which , when the ether gas was at its great volume , amounted to over thirty in number .
They were , for the most part , regularly distributed .
A millimetre scale etched on a Qlip of mirror , JK ( fig. 1 ) , was laid below the experimental and the distance between the heaps was read .
But instead of reading the actual position of all the heaps an average was taken in the following manner :Suppose the total number of heaps to be 30 ; the 1st and 27th were read ; the 2nd s.nd 28th ; the 3rd 29th ; and the 4th and 30th ; and , dividing by the total number read , four averages were obtained .
These usually agreed with each other to within per cent. ; the mean of all four was taken as the true wave-length .
The middle of each heap of silica was taken as most easily read .
The silica could hardly be said to " " heaps\ldquo ; ; it distributed itself in streaks of varying length across the tube , each set of streaks forming an oval-shaped patch , of which the middle point could be estimated with fair accuracy .
The volume of the gas was then diminished , and another set of readings taken in a similar manner .
The mercury entering the tube coverd up a portio11 of the silica ; but on again increasing volume it flowed away , leaving silica as before .
This , of course , made it difficult , if not impossible , to no accurately the volume occupied by the ether gas ; for the volume of the silica was an uncertain quantity .
But the whole amount of silica was very smaU , certainly not amounting in volume to per cent. of that of the ether .
To compensate for this error , by one in the opposite direction , no correction was made for the increase in volume of the tube on rise of temperature .
The etTor thus introduced is about the same order of magnitude opposite in sign to that involved by neglecting the volume of the silica , and hence the two errors may be taken as compensating each other to some extent .
Moreover , the volume of one gram of ether at nite temperatures and pressures had been determined by Dr. S. YOUNG one of the authors ; and it was necessary to know the actual volume merely for the purpose of deducing the weight .
The volume of the gas was successively decreased , observations made similar to tbose described .
The temperature having been raised by increasing the pressure on the boiling liquid , a fresh set of observations were made .
It was however , whenever the jacketing vapour changed , to refill the tube , because the contraction of the ether , owing to its condensation , brought the mercury into contact with the rod , and filled the space between it and the inner wall of the experimental ON THE ADIABATIC RELATIONS OF ETHYL OXIDE .
tube .
This , of necessity , made the experiments somewhat tedious and difficuIt ; but we ventuoe to think they are as accurate as the method allows .
II .
METHOD or ASCERTAINING THE 0F ETIIER EMPLOYED .
For each set of experiments with a single jacketing the tube had to be specially filled with ether .
This involved a different weight of ether eu.ch time .
lt was necessary to know the weight , in order to calculate the volume occupied by 1 .
of the vapour .
To ascertain the weight , the volume at any given pressure and temperature was ascertained from an isothermal showing from previous experiments , by RAMSAY and , the relations between volume , temperature , and pressure .
From this the weight was calculated at each different volume .
All the results thus obtained were averaged , and the mean results accepted as the weight .
To take an instance : the temperature 14 C. , the actual observed volumes and pressure were found as in the first two columns .
The volume of 1 , conesponding to the pressure in each case , is given in the third column , and is from the Memoir referred to ; and the weight , calculated from the volume of 1 , is to be found in the fourth column .
The mean result of all these determinations was taken as the true weight .
The volumes of 1 .
were then calculated by dividing each observed volume by the mean weight .
The details of one set of observations are given in the following table:\mdash ; Similar sets of observations at other temperatures with the same ether were taken , and from them the mean weight was ascertained .
But it frequently happeued that the results at one temperature did not coincide with those at another .
It must be remembered that the tube lay horizontally , and that on tilting it , after a set of observations had been made , more ether would escape up the tube .
For the hot mercury in passing down gasified ether remaining in that porbion of the experimental tube which was inside the pressure , and the result that the ether was almost invariably increased somewhat in amount at the higher temperature .
Thus , in the above case , calculation of weight at 13 C. , 14 C. , and C. , gave the following numbers as means : the individual resulcs agreeing as closely as those in the preceding table .
Each temperature wastherefore- treated on its merits , and unless no alteration in had been proved to occur on changing temperature , the calculations were made on the basis of measurements at different volumes and pressures at the same tem perature .
III .
DETERMINATION 0F THE FREQUENCY 0 VIBRATING ROD .
Ten sets of experiments were made in order to ascertain accurately the wavelength in air of the note given by the glass rod .
The mean result of each is here given ; the temperature was C. I. II . .
V. VI .
VII .
VIII .
IX .
X. 193 1922 Mean millims. From these results the number of vibrations per second of the rod may be calcu .
lated by the formula The velocity of sound in air was taken as 33253 centims. per second at C. ( ' Encyc .
Brit Article Sound ) .
The wave-length at is millims. Hence per second .
The note was a shrill squeak .
The close agreement of the values for the wave-length quoted above seems to show that the accidental sources of error are here only trifling ; and they may be considered as practically eliminated from the final mean value .
But it is still possible that there may be some constant source of error affecting all the experiments equally ; thus it is well known that the wave-length in gases is affected by the diameter of the tube , and this effect may become quite noticeable if the tube is narrow .
It is not likely that such an effect really existed in the above experiments , for though the tube was rather narrow , the pitch of the note was exceedingly high .
But it was thought as well to make quite certain , and a set of experiments was carried out to test the accuracy of the value of obtained .
( 1 .
) Confirmatory Experiments with Hydrogen .
The velocity of sound in hydrogen has been made the subject of investigation by a large number of experimenters , some of whom nsed tubes sufficiently wide to obviate the note of the vibrator .
Taking the density of air as compared with hydrogen to be , this gives ) per second as the frequency of the vibrator .
( 2 .
) Confirmatory Experiments Argon .
xperiments were made by one of the authors on the velocity of sound in argon ; these experiments were originally underbaken to determine the value of the ratio of the specific heats for argon , but as some observations were carried out the actual tube described above , they can bs used for our present purpose .
The following are the measurements observed.\mdash ; Mean of 40 readings in original tube with argon at C. centims. , Meau of 6 readings in wider tube with argon at C. . . .
centims. . . .
\ldquo ; Mean of 5 readiugs in wider tube with argon at C. centims. , , Mean of 11 readings in wider tube with argon at C. centims. , Mean of 5 readings in wider tnbe with air at C. \ldquo ; Mean of 5 readings in wider tube with air at C. . .
centims. \ldquo ; Mean of 11 readings in wider tube with air at C. centims. From these numbers we easily find that the mean wave-length at in wider tube is for centims. , and for air ; this gives us in the original tube .
( 3 .
) Confirrnatory Experiments with the Phonogrcvph .
An attempt was made to discover the number of vibrations emitted per second by the rod with the aid of a phonograph .
Such an experiment , if it could have been carried out , would have ) valuable as affording a direct determination of the frequency required .
Unfortunately , the sound produced by the vibrator was too feeble to make a distinct impression upon the wax cylinder .
A long tube of the same kind of glass was , therefore , taken , and this produced a much graver and stronger note , capable of leaving a distinct trace upon the th expe ' vie also repeated with a shorter tube .
The following results bKnod : A tuning-fork , of frequency 1024 , was found to leave 232 ] ) ressions in one revolution .
This gives the velocity of sound in this particular kind of glass 263,000 and 262,700 centims. per second for the two tubes respectively .
We take the mean , 262,850 centims. per second , as the velocity of sound in the glass rod , and this nption , though we have no guarantee of its exact truth , is unlikely to introduce any serious error .
We then find that a rod centims. long will emit 18,578 vibrations per second , and this number may be accepted as giving roughly the frequency of the glass rod when performing free vibrations .
In the actual apparatus used , the rod was sealed on to a glass tube , near its middle , and this may well have had the effect upon it of a rider on a tuning-fork , and lowered the tone .
The test in the present case , only a rough one , but it has some value all the same , as it shows that the dust-heaps observed in all these experiments were really due to the vibrations of the fundamental note in air , and not to the higher harmonics in the glass .
It was decided upon the whole to adopt the value of , derived from the experimenbs with air , as there is less manipulation required in this case , and , therefore , less liability to error .
By comparing the experiments with air , hydrogen , and argon , it will be seen that the value of the frequency is uncertain , to an extent of about per cent. ; but it will be shown later on thut the manner in which this constant enters into the equations renders the uncertainty of far less importance than appear .
IV .
CALCULATIONS 0F THE ADIABATIC ELASTICTTY .
Having asccrtained the volume of 1 .
corresponding to each pressure and temperature observed , and having also ' found the corresponding mean wave-lenghs , the velocity of sound in the ether-gas was calculated by the formula\mdash ; ; where V is velocity , the unber of vibrations per second , 17,566 , and the wavelength .
The following results were obtained :\mdash ; ON THE ADIABATIC RELATI OF ETHYL OXIDE .
Temperature 1 Temperature C. Temperature 1 Temperature C. Seventh .
Qninoliue jacket .
Temperature 18 C. Temperature 19 C. Temperature C. Temperature 19 C. elat e ON TEE ADIABATIC RELATIONS OF OXIDI From these observations a set of isothermal curves was constructed , where was plotted against volume ( see fig. 4 ) .
The results are not very concordant ; some DR. E. P. PERMAN , DR. W. RANSAY , AND .
J. ROS -INNES , are erratic , and is to be expected , for it was not always possible so to rub the rod as to produce the required tone .
A rod so placed is capable of many kinds of vibrations , and the particular rate of vibration , though the one most usually produced , may , by some difference in the method , of rubbing , be replaced by another These curves were then further utilized by reading off the values of for equal volumes , 70 , 65 , 60 , and so on , at intervals of 5 down to volume 20 cub. centims. per , and at smaller intervals , viz. , , 15 , 14 , .
centims. per , and so on , down to 4 cub. centims. per grm. at smaller volumes .
From these numbers a second diagram was constructed , representing the values of plotted against temperature , the lines being isochoric .
These curves were observed not greatly to deviate from straight lines\mdash ; they intersected at small volumes : They would probably also intersect at larger volumes if produced .
It lJas not been thought necessary to reproduce this figure as it is virtually contained in fig. 4 .
The value of is , of course , the same as the adiabatic elasticity , and , like the isothermal elasticity , it has the dimensions of a pressure .
It will be remembered that V was measured in centimetres per second , and in cub. centims. per Consequently , the elasticity , as calculated above , must be given in absolute units , or dynes per sq .
centim It was found more convenient to change to pressures measured in millimetres of mercury , and this can be readily done by dividing all the former numbers by V. MATHEMATICAL DISCUSSION 0F PRECEDING RESULTS .
We noticed above that for equal volumes the adiabatic elasticity is roughly a linear function.of the temperature , and this fact seems to hold out some hope of all the experimental results obtained under the power of analysis .
An inspection of the diagram shows that the difference between the isochorals actually drawn and straight lines is no greater than the uncertainty due to experimental errors , so that as far as the direct evidence goes ib is possible that the isochorals are really straight .
Even if it should turn out on subsequent investigation that the isochorals are slightly curved , the treating them as straight lines may be considered as a first approximation to the truth , and as ajustifiable simplification of a very complex problem .
Let us call the adiabatic elasticity measured in millimetres of mercury , then we may put where and are functions of the volume only .
Now , it was shown by one of the authors , in conjunction with Dr. SYDNEY YOUNG , that for constant volumes there is a * The curves actually given in fig. 4 have been calculated by means of the formula on p. 188 ; they are the smoothed version of a set of similar curves drawn to represent the actual observations .
They agree with the observations better than the origmal free-hand curves .
linear relation between the temperature and pressure , so that we may put ' where and are functions of the volume only ( ' Phil. Wag vol. 23 , p. 436 ) .
From these two equations it is possible to eliminate the temperature , and we obtain Thus we see that for constant voIumesE may be expressed as a linear function of ; the next step is evidently to find the value of .
For this purpose the isochoral diagram was employed ; a straight edge was laid so as to pass as near as might be to the drawn points , and the slope of this edge was then determined ; this slope gave the value of To obtain nothing more was needed than to use the values given in RAMSAY and YOUNG'S paper ( loc. cit. , p. .
On calculating the value of for various volumes , it was found to be nearly constant ; such deviations as existed were not arranged in any orderly lanner , but occurred sometimes to one side and sometimes to another , as if fluctuating about a mean value .
It appeared as if might be taken as a fair mean between all the values obtained .
We may therefore put and this leads to the equation , where is a function of only .
In order to determine the value of , the above was written Keeping the volume constant , and ging the temperature , this equation gives values of differing slightly from one another ; the arithmetical mean was chosen as giving the value of for that particular volume .
In this way was determined for a large number of volumes , and it was found that its values could be approximately represented by means of the empirical formula The above formula was selected from a large number that were tried , and it reproduced the found values of fairly well , but it is not suggested as being anything more than an empirical formula .
Unfortunately the Kinetic of Gases is not sufficiently advanced to enable us to form any priori notion as to ON THE ADIABATIC RELATIONS OF ETHYL OXIDF , .
what form the function might be expected to so that we are perforce driven to employ an npirical formula .
With a large amount of time and labour it be to find a formula for simpler in character , and more exact in its results , but it did not seem worth while to undertake the amount of ouble involved in such a search .
The formula can be successfully used for the purposes of interpolation within the field of observation , but any use cf it for extrapolation would be hazardous .
We may now write the equation for as This formula has been obtained as the result of four consecutive approximations , and it is possible that each succeeding approximation has led us further from the .
On the other hand , it is conceivable that the errors of the various approximations have to a large extent neutralised one another .
It was thought therefore to test the above formula by calculating out the isotherma]s , and ccmparing them with the observed measurements .
The result is shown in fig. 4 ; and it is easily seen that the calculated isothermals may be fairly taken as a system of smooth lines passing through the observed points .
From the equation to we obtain .317,700,000 or , .
588,600 238,100 2,3.34,000 Multiply by and we get Integrating Hence , .
The constant is at our disposal , and we may choose it so that the adiabatic shall pass through any assigned point .
By giving a series of diffel.ent values to we obtain a system of adiabatic lines .
We may get a convenient system by giving to successively the values 710,000 , 720,000 , 730,000 , 740,000 , 750,000 , 760,000 ; such a.system is shown in fig. 5 ; the curves have been drawn for a great part of their length so as to exhibit to the eye their general shape , but only that portion of the system.which is intersected by the drawn isothermals has been really investigated experimentally , and can , herefore , be safely trusted .
The adiabatic for passes close to the critical point .
The system ofisothermals and adiabatics intersecting another divides the pIane of inCo a set ofcurvilinear parallelograms , such as that suggested in MAXWELL'S Theory of Heat ' ( p. 156 ) ; and , if our interpretation of the experimental results is correct , it ought to be possible so to select the system of adiabatics that the areas of all the parallelograms are equal .
The following -dynamic investigation win show to what extent this oan be done .
ON THE ADIABATIC RELATIONS OF ETHYL OXIDE .
Let us assume that the equation to the isothermals of a gas can be written in the form where and are functions of the volume only .
Let us also assume that the equation for the adia , batic elasticity can be put in the form where and are functions of the volume only .
Then , by eliminating T. and putting for its value - , we see that we have a linear differential equation for of the first order , the independent variable being .
The integral of such an equation must be of the form where is the constant of integration , while and are functions of only ; we see also that we must have .
This equation determines a system of adiabatics by giving successively different values to and by taking two hbouring values , say and we obtain two adiabatics as near together as we please .
Let Fig. 6 .
now the pair of adiabatics be crossed by two isothermals are temper , tures and , then the area of the infinitely parallelogram formed must be indeof by the Second Law of Tbermo-dynamics .
Let ABCD in fig. 7 denote a magnified picture of this small parallelogram ; then , Fig.,. .
by shearing the figure after the manner of MAXWELL in obtaining his fourth thermodynamical relation ( loc. cit. , p. 166 ) , we obtain DR. E. P. PERMAN , DR. W. RAMSAY , lNNBS , area ABCDincrease of pressure when we pass from lower to higher adi1lbatiet at constant volume .
increase of volume when we pass from higher to lower isothermal along adiabatic .
Now , whatever the variations of pressure volume and temperature , they are subject , if small , to the relation Let us denote differentiation with regard to by accents , then we may put this Also any adiabatic alteration will be subject to the relation Hence , Therefore Now this quantity is to be independent of so long as we keep una.ltered , hence we have function of only .
This can only be the case if expressions are functions of only .
But both these expressions , from their form , can be seon to be functions of only ; henoe they must be numerical constants .
Let , say , ON THE ADIABATIC RELATIONS OF ETHYL OXIDE .
then Substituting for , we may write this so that As long as there is no assumption made about the connection between and this equation simply gives in termq of .
The adiabatics we have drawn , however , were the result of further assuming is equal to a ; and if this condition is introduced , we find that the form of is restricted .
Let us denote by 7 , the constant value of , then , or Put , and we have C. This equation is easily integrated , and we obtain so that where and are constants at our disposal .
If we confine ourselves to volumes which lie between 70 and 4 , the values of may be roughly reproduced by means of the above formula , taking care to give suitable values to and .
But if , in such a formula , we were to attempt to extrapolate to large volumes , the result would be that we should obtain values of totally at variance with experiment .
This shows us once more the need of restricting our formulae entirely to those portions of the field that have been experimentally investigated .
Experiments Liquid Ether .
In experimenting with liquid ether , the apparatus and method of working were similar in every respect to what has already been described .
The volume of 1 .
of the ether was , as before , not directly determined , but was 188 DR. E. P. PERMAN , \amp ; c. , ON THE ADIABATIC ELATION8 OF ETHYL OXIDE .
inferred from a knowledge of temperature and pressure .
It was more easy to obtain well-defined heaps of silica in the liquid than in the gas .
The following results were Pressure . .
of ocity .
Temperature C. Temperature 12 C. Temperature 1 C. Temperature C. Temperature C. Pressure 1eWngath iecond)ityin c Temperature 19 C. Temperature 19 C. Temperature 20 C. Treating these data as before , the isochoric lines for are found to be stlaight .
The investigation was not pushed further , however , as the formulae involved appeared to be even more complicated than in the case of the larger volumes ; but the above experimental data have been given for the sake of anyone who feels interested in the subject .
' Volume not read ; curve nearly horizontal .

