2534 TYRER : ADIABATIC AND ISOTHERMAL COMPRESSIBILITIES OFCCXXXV1.-Adiabatic and Isothe~~mal Compresstbilztiesof Liquids between One and Two Atmospheres'P rc'ssure.By DANIEL TYRER.THE present paper constitutes an extension of a previous one (T.,1913, 103, 1675) on the compressibility of liquids, in which a newand precise method for the determination of compressibilities atlow pressures was described.Since the publication of the previous results, i t was discoyeredthatt a small correction to the observed values had been overlooked.This was due to neglecting the small change in pressure on theliquid caused by the change of level of the liquid in the verticalcapillary tube (see the previous paper). Although this pressure isvery small compared with the total external pressure change, itintroduces a correction greater than the general experimental error,and must therefore be taken into account.Fortunately, thiscorrection can be accurately calculated from the observed data,and, on the average, causes the results to be raised by about 0.5per cent. All the previously obtained results have accordinglybeen corrected, and the corrected values are contained in the tablesin this paper. The necessity of this correction is to be regardedas a disadvantage of the piezometer. It has also been noticed thatfor rather viscous liquids, such as aniline, the capillary tuberequires a long time to drain, and for easily vaporised liquids,such as ether, slight errors are introduced by the evaporation ofa little liquid in the capillary tube during an experiment.Toavoid these disadvantages and sources of error, the piezometer wasmodified, as is fully described below. In order to test the1 accuracyof the previous resulta (corrected as explained above), a few ofthe determinations were repeated, in each case with the new piezo-meter. As will be seen from the following tables, the differencesbetween the old and new results are, in general, very small, whichgives considerable confidence in the validity and precision of themethod.In order to determine the isothermal compressibility from theadiabatic value, use is made of the following thermodynamicequation :P = a + ~(2y.Jb c,where p and a are the isothermal and adiabatic compressibilitieLIQUIDS BETWEEN ONE AND TWO ATMOSPHERES’ PRESSURE.2535respectively, v the specific volume, J the mechanical equivalent ofheat, C, the specific heat a t constant pressure, and T is theabsolute temperature.I n the previous work, in order to obtain values of dv/dt and ofC,, the results of other investigators were relied on. I n the caseof C,, errors, even moderately large, have a comparatively smalleffect on the accuracy of 8, but in the case of dvldt the effect oferrors is considerable. It was found that in many cases the valuesof dv/dt which had been calculated from the specific-volume dataof various investigators contained considerable errors, and i t wastherefore necessary to make a series of accurate specific-volumedeterminations for each liquid, from which accurate values ofdv/dt could be calculated.The practical part of the work istherefore divided into two parts, namely, tlie determination ofadiabatic compressibilities a t different temperatures and the deter-mination of specific volumes.,4ppratus: The New Form of Pieaometer.The construction of the piezometer can best be understood byreferring to the diagram. The liquid to be investigated is con-tained in the inner vessel, A , of ordinary soda-glass, filled com-pletely up to and between the two taps, T and T,, and the mercurythread in the horizontal capillary tube, B. The two tubes C andD are attached by stout rubber tubing to a small air-pump andmanometer. On increasing the pressure, the mercury thread isdepressed in the graduated tube, B.When the temperature isconstant, the pressure is released, and the-change in position ofthe mercury thread in B is noted. The filling of the piezometeris effected by first exhausting i t of air, by attaching the sidetubeE to a strong pump, and then allowing the liquid to enter throughthe tap T. When almost filled, the capillary tube B is dried, themercury thread allowed to enter, and then the last few C.C. of airexpelled from the apparatus by warming. It is emptied by invert-ing the instrument and attaching the tube E t o a suitable ex-hausted receptacle.For further details of the rest of apparatus, and the manner ofworking, the previous paper must be consulted. No correctionbeyond that of the compressibility of the glass is necessary.Forthe compressibility of the glass, the result of Amagat (2.18 x 10-6)has been taken as correct.The volume of the piezoniebr used was about 450 c.c., and thediameter of the graduated capillary tube depended on the com-pressibility of the liquid in the apparatus, but was such as t o giv2536 TYRER : ADIABATlC AND ISOTHERMAT, COMPRESSTBILITIES OFa change of reading of about 6 to 10 cm. for a pressure change ofabout one atmosphere.The average error of the adiabatic compressibility determinationsdoes not appear to bo greater than 0-1 per cent. Probably thegreatest constant error lies in the correction for the compressibilityof the glass, but this must be comparatively small, f o r it was shownin the previous communication t h a t results obtained with a copperpiezometer, which requires a much smaller correction than glass,are in good agreement with the results obtained with a glasspiezometer.Determination of Values of dvldt.I n order to be able to calculate accurate values for the functiondv/dt, very accurate specific-volume data-carried out, to the fifthdecimal place a t least-are necessary.The dilatoineter is notcapable of giving such a degree of accuracy, and hence tlie longermethod of the pyknometer had t o be used. The ordinary SprengeLIQUIDS BETWEEN ONE AND TWO ATMOSPHERES' PRESSURE. 2537form of pyknometer is liable to considerable error for volatileliquids, owing to evaporation in the two capillary tubes, and forviscous liquids there is always an appreciable quantity of liquidwhich clings to the sides of the unfilled part of the capillary taube.A new form of pyknometer was therefore devised which has onlyone capillary tube instead of two.On the capillary tube is etcheda fine mark, and tho open end is widened out and provided witha ground-glass stopper. The instrument had an approximatevolume of 70 C.C. The empty part of the instrument above themark on the capillary tube can be thoroughly freed of adheringliquid. The filling and emptying of the pyknomet'er is effected byattaching to i t a small dropping funnel provided with a side-tubeand tap, through which the instrument is exhausted. The liquidis then run in, and fills the pyknometer completely. The liquiddoes not come into contact with any rubber connexions during thefilling, and is always kept in contact with dry air only.The fill-ing under exhaustion also serves to free the liquid from dissolvedair.As the capillary stem of the instrument may be made to anydegree of fineness, the adjustment of the volume may be made t oa very superlative degree of accuracy. The weighing of the pykno-meter was made to 0.1 milligram, and all weights were reduced toa vacuum. The thermostat used consisted of a large 40-litre water-bath provided with a motor-driven stirrer and a thermoregulator ;the temperature remained constant to less than O*0lo. It may besaid, therefore, that practically all error lay in the temperaturereading. For this purpose, a series of finely graduated thermo-meters was used, capable of being read with accuracy t o less thanO*0lo, which had been standardised by comparison with the normalhydrogen thermometer.I n addition, they were compared withanother set of thermometers, and the fixed points (melting andboiling points) were tested, and the melting point of sodiumsulphate (Na,S0,,10H20, 32.38O : Richards and Wells, Zeitsch.plhiysikal. Chem., 1890, 26, 690) was carefully determined. A t nopoint was the correction greater than 0'02O. The error of tempera-ture readings was probably not, on the average, greater than O*0lo.Now for liquids the value of duidt of which is less than 0.001,an error in the temperature reading of O*0lo causes an error in thespecific volume for a liquid of average density which affects thesixth decimal place only.Hence it is reasonable and logical t ocalculate the specific volumes to six decimal places. A t the highertemperatures the degree of accuracy would be rather less than this,on account of 3 somewhat greater temperature error2538 TYRER : ADIABATIC AND ISOTHERMAL COMPRESSIBTLITIES OFThe volume of the pyknometer was accurately determined a tdifferent temperatures by weighing it filled with boiled, distilledwater. These calibrations were also repeated, using distilledmercury in place of the water.* The calibrations were repeated a tvarious times to see whether the volume of the pyknometer wasakering with time, but no change was observed.The Pure Liquids.They were generallyfractionally distilled or frozen, and, where possible, they were driedover phosphoric oxide.The fractionation was continued until aliquid of constant boiling point was obtained, and the density wasunchanged by further distillation. I n the following table aregiven the constants (boiling points and densities a t 0.) of the pureliquids used, where thaw had come within the scope of accuratemeasurement. (Only boiling points below looo were accuratelymeasured.) The corrections of the boiling points to normalpressure were made by calculating the values of d t / d p from theClapeyron-Clausius latent heat relation, which is a much moresatisfactory method than the use of tables of experimental valuesof dt f d p , a function extremely difficult to measure with accuracy.All the liquids used were highly purified.Liquid.Carbon disulphide.........Ethyl acetate.. .............Ethylene chloride .........Chloroform ..................Toluene .....................Aniline ........................Nitro benzene ...............wXylene .....................Ethyl bromide ............Ethyl iodide ...............Benzene .....................Ether ........................Methyl alcohol ............Ethyl alcohol ...............Chlorobenzene ............Density at 0".1.293040.924681-282481 -52 0490.884121.038931.20323 at 20"0.881 5 11.498211.980380.88946 at 10"0.736390.810400.806451- 12780Boiling pointat 760 mm.46-26'77-1583.4561.21 --38-4072-5280.2834.6064.7278.32IAdiabatic Compressibilities and Specific Volumes.I n the following tables are given for each liquid the experi-mental results of the adiabatic compressibilities. For the liquidswhich had been previously investigated, the values corrected, asexplained in the introduction, are given, together with a few* The specific volumes of water given by Thiesen, Schecl, and Diesselhorst(Landolt-l3ornstein, " Tabellen ") were used, and for mercury the results of Chappuis(ibid. )LIQUIDS BETWEEN ONE AND TWO ATMOSPHERES' PRESSURE.2539supplementary values obtained by me'ans of the new form of piezo-meter.For each liquid, also, the experimental specific-volume data aregiven, together with the constants in the' equation Vt= 7, + at + bt2.It was exceedingly laborious to find an equation containing fourterms on the right-hand side to fit the experimental results for thewhole temperature range, and it was better to find two equationsof three terms each covering a range of not more than 40°.I n the following table, to expresses the temperature in degreescentigrade, a is the adiabatic compressibility, and v the specificvolume.Carbon Disulphide.to.0.012.7712-5820.3029-2 133.3940.0519.8327.1834-02to.0.014.1420.9629-4236-7744.8054-310.09.3 124.3135.2049-89a x 10'.to. V.Old rt sults corrected.52.95 0.0 0.77337057.94 11.66 0.78385757.74 16.97 0.788 7 6860.61 23.74 0.79515463.9 1 30.72 0*80188865.49 39.07 0.8 1 0 14469.13New results.60.4363.3566.36VZ 0.7733 70 + 0.00088 18 t f 0.05 15 10 L'.(7hloro f orm.a x 10".Old results corrected.59-3065.9 168.9974.0078-2083.1 190.6558.5462-6570.7577.3587.25New results.to.0.010.3 115.9425-0732-0640.3547.4054.60?'.0.6550970.6634080.6681310-6759420.6820880-68 96440.6962780.703199~t=0.655097+0.0,79260 1+0-0,1576 L'.(0-30").~t=0*680294 +0~~~88323(~-30)+0~051945(t-30)'J. (30"-60")2540 TYRER : ADIABATIC AND ISOTHERMAL COMPRESSIBILITIES OFToluene.to.12.4216.0127.1038.7450.2660.1668-5684.1 10.012-4147.1057-7267.6090.002 1.5036.4329-4247.27a x 108.Old results corrected.Glass piezometer.64.1965.8571.1677-3283.8190.8497-32Copper piezometer.109.858-7464.1082.1089.0096.88115.3New results.68.0076.0572-1182.46to.0.014.9821.1725.7830.7440.0050.1373-0579.7379.2499.172'.1-13 10651.1488601.162369191686891.1686891.18072 11-19 1481.224941.235291,234431.26660Z J ~ = 1.131065 + 0.001 1630 t + 0*051959 I '.l't= 1.19131 + 0.0013918 (t - 50') + 0.052834 (1.-_ 50')).(0-40').Benzene.1'.9.5520.3830.5236-3842.7649.5449.6765-4016.1521.9131-334 1.3050-2962.9715.9632.0354-5864.00a x lo6.Old results corrected.Glass piezometer.61.4066-7672.3575-6479.8084-9284.6297.5863-9967.8672.8 I79-4285.5995.20New results.04.1673.2488-5696.45Copper piezometer.2'.11.9214.4017-9024-4247.9558-3362.3072.06T.1.126871.130231.135061.144021.178221.194021.200251.21589~ l t = 1.124278 +0~0013508(t- 10)+0*O5l860(t- 10)".t't= 1-1 8129 + 0*0015038(t - 50) + O*O53086(t - 50)'.[5&80°]LIQUIDS BETWEEN ONE AND TWO ATMOSFHEREY' PRESSURE. 254 1Ether.LO.0.011.1915.5425.3030.6914.6321.1127.6210.9118.5424.9329.50to.0.015-5228-4135-2843.6461.9062-0672.310.012.6417.8227.5039.6252-6560.0272.50a x 10". to.114.3 0.0129.4 13.29139.6 23.90149-5 28-00158.3Copper piezometer.134.5142-8151.3New resulfs.128.24138.62149.05167.73Old results corrected. .Glass piezometer.t't = 1.35793 + 0.0020514t + O*O5542t2.Ethyl Alcohol.a x lo6.Old results corrected.83.9294.45102.04105.99112.36118.9128-5139.6New results.83.8191.3794-22100-7109.31120.05126-69139.95to.0.014-2123.9839.1646.3254.0162-7172-072'.1.357 931-386151.410041.419632.'.1.2399981.2586981.2720431.2936901.304591.3 16381.330471.346311.239998 + 0.0012909S + 0.0,1767t2.q't= 1.29494 f 0.0014749(/ - 40) + 0*053958(t - 40).[40-'70°].VOL. cv2542 TYRER : ADIABATIC AND ISOTHERMAL COMPRESSIBILITIES OFCarbon Tetrac?doride.to. a x loti. to.Old results corrected.Glam piezometer.0.0 63.38 0:012-43 69.33 ' 37-4120-77 73-24 37.9029.09 78-38 46-8938.29 83-88 53.7647-10 90.70 62.6367.72 99.28 72-4367.63 106.7416.16 70.7238-69 84.2360.10 101.650.0 63.0416.27 71.2524.08 75.5438-97 84-8246.85 89-3053.25 95.68Copper piezometer.27-10 76-98New results.t+ = 0.612869 t 0*0:371724f + 0*051227t2.~t=0*643530+0*0,~81438(t- 40)+0*051507(t - 40)'.[40-70'1.U.0.6128690.6257280-6418150.6492430.6560210.6627 150-67 1525Chlorob enzene.to. a x 106. to.Old remlts corrected.0.0 49.19 0.013.40 54.04 24-3224.24 57.54 38.8935.63 61.83 47-9443.97 65.37 60.6352.79 69-59 73.5462-02 74.3671.67 78.8780.47 83-30.0 49-3823-78 57.3941.26 64.09New results.V t = 0.886685 t 0.0384104t + 0*0,9506tS.t*t=0*921848+0.0391697(t - 40)+0*051265(t -7'.0.8866850.9077010.9208300.9293360-9416350- 96428240)'. [40-80°].Methyl -41cohol.to.a x lo6- to.0- 0 88-94 0.013-65 98.06 11-3316-37 99- 18 24.4221-37 102.93 41.6629.65 109.04 55-2030.80 109-70 58-3124.33 105.6439-09 115.9343.69 120.12Vt= 1.233821 + 0.0014089 t + 0*052240 t2.vt = 1~278204+0~0015261 ( t - 30)+ 0.053740 ( t -2'.1.23382 11.2500721.2695531.296391.318991-3243LIQUIDS BETWEEN ONE AND TWO ATMOSPHEUES' PRESSURE. 2,543to.0.010.5619.1728-2331.98Ethyl Bromide.a x 1Oa. to. 0.72-99 0.0 0.66746379.45 5.17 0.67204685-74 &75 0.67529390.80 17-58 0.68356296.53 25.42 0-69109831-54 0.697345=0.667463+0*0,87466 1+0.0,2307 1'.Aniline.to. a x lo6. to.0.0 32.94 0.010.66 34.54 5-2520.42 36.35 21.8930-38 38-25 30.6339.57 40.25 50.1249.69 42-76 41-6560.04 45.08 62-3573.56 48-88 79.0485.70 52.43 98.9585.50 52-10~t=0*962534+0.0:379697 t+O*0~8005 t2.?'t = 1.004384 + 0.03869 53 (l - 50) + 0.051 288 (1 -Ethylene Chloride.21.0- 9625340.9667590.9803731 * 0044 890.997 1171.0153191.030741.050030-98771350)'.[50-lOO"].to. a x lo6. to.0.0 48.17 0.010.81 51.95 5.5519.31 55-31 15.4120-5 1 55-53 2 1-0330-52 60.08 31-9825.45 57-72 40- 1327.25 58.56 50.7439.15 64-16 59.7750.23 69-93 74.9859.79 75-6773.50 85-02t?t=0.779738 + 0.0386779 t + 0*051459 t2.~t = 0.8 1 6 7 8 5 + 0.039 84 2 ( 1 - 4 0) + 0*051 9 43( t -2'.0.7797380.78457 10.7934570.7986380.8089620.8169120-8275790.8370430.85359 140)?. [40-80"].Acetic Acid.to. a x 10". to. 2).18.98 75.10 19.36 0.95 180129-55 81-36 24-97 0-95768539-51 87-19 29.54 0.96238839.37 86.73 34-26 0.9 6 7 3 749-23 93-57 39.57 0.97301660.80 102.03 49-69 0.9839377-44 114.00 60.04 0.9955173-48 1.0107879.11 1.0175599-07 1.04250[ 6O-1OO0].7't= 0.95246 1 + 0.0010266 (t-2O)+0.O51212 (t-20)'.~t ~0.99546 + 0.001 110 (t -- 60) + 0*0,2405 (t - 60)2.8 c 2544 TYRER : ADIABATIC AND ISOTHERMAL COMPRESSIBILITJES OFto.14.6721-1030.7740.0050.1360.1573.0782.00Nitro b e nz e n e .a x lo6.to.37-33 12- 9538.91 2 1-2541-10 30-7343-26 39-1045.43 49.9948-41 68.5352.2654- 92~t=0*826343+0*0367665 (t- 13) + 0.06162 (t- 13)'.Ethyl Zodide.to. a x loo. to.0.0 59.7 1 0.012-17 65.0 I 10-3323.86 70.86 19.1540.85 80.28 39-3252-10 88.08 24-2062.60 95.83 46-4553.8062.72vV~ = 0.604927 + 0.0366522t + 0*051072t2.W t = 0622849 + 0.036295(1- 30) + 0*061114( tm-Xylene.to. a x 10';.to.0.0 57-27 0.015.10 63.44 18-832 1.38 66.49 22-6230.92 71-16 ( 9 ) 29.6232.73 71-67 39-3740.69 75.82 60.5449-72 80.76 69-2861.60 88.00 75-9278-83 97.21 78.8698-86vt = 1.1344 17 + 0*001092 6t + 0-05146 lt' ,~ ' t = 1~192698+0~0012402(~- 50)+0.0,2002(t-Ethyl A ce t at e .to. a x lo6. to.0.0 70.30 0.019-28 83-02 18-2221.89 84.81 27-1830.62 91.50 34.0534- 13 94.93 41.0440.47 100.34 52-2350.14 109-39 59-9362.35 122.9 73.2162.90 123.6~ t = 1~081456+0~0013700t+0.053282t2.Vt= 1.14154 +0*0216247(t- 40) + 0.054028( 1-V.0.8263090.8319840.8385380.8444020.8522090.858420[ 13-60°].V.0.5049270.5 108 8 10.5161440.5288080.5 19 1720.533 5 1 20.5384640.6446230)2.[30-601.2' *1.13441 71.1555001-1598771.168081.1796941.1933671.204381.226271.2 301 21.25867-50)'. [50-100°].21.1.0814561.107511.1211171.1319091.1432101.1619701- 17 5521.1999440)'. [40- 8 0 O . LIQUIDS BETWEEK ONE AND TWO ATMOSPHERES’ PRESSURE. 2545to.2.406.4514.1219.3724.4235-7546-9563.5175.3386.7590.22a x lo6.Old resultscorrected.50.1049-0547-1846.0045.2 144.0843-1942.8042.6742.8 142.88Water.to.9.7111.5110-9814.2319.3820.3824.9824-2929-4634.6238.9047-9560.6573.5084.40a x 10’.New results.48.3548-0348.1347.3 146-2 146.0945-3445-3744.5643.7643.6943-0442.55 ’42.5743-00Isothermal Compressibilities and Values of dv/dt.Values of d v l d t were determined by differentiating the equationsgiven in the foregoing tables. For the wider temperature rangeswhere two equations were necessary, i t was found that there wasusually a slight break in the continuity of the d v l d t values atthe intermediate point.This break was removed by plotting theresults, and then drawing a smoothed curve, from which values ofd v / d t were read. I n addition, the accuracy of the results waschecked by calculating them in another way. The mean value ofd u / d t between each pair of succeeding points was determined bysubtracting the specific volumes and dividing by the temperaturedifference.The result was taken to refer t o the mean ternpe_rature.Then, by plotting a curve of all the values thus obtained, resultswere obtaineld a t regular temperature intervals, which, in general,agreed excellently with the values obtained from the equations.From the values of d v / d t and the adiabatic compressibilities,values of the isothermal compressibility P have been calculated byaid of the thermodynamic equation given in the introduction. Aknowledge of the specific heat a t constant pressure (C,) is alsonecessary. As already explained, comparatively large errors in thelatter quantity affect the calcula€ed values of fi but slightly, andso no special determinations of specific heats have been made, butresults obtained by other investigators have been relied on.Forbenzene and carbon tetrachloride, the specific heat determinationsof Mills and McRae ( J . Physical Chem., 1910, 14, 797; 1911,15, 54) have been used in the calculations. Schiff’s results(Annalen, 1886, 234, 300 ; Zeitsch. physikal. Chem., 1887, 1, 376)were employed in the cases of toluene, nz-xylene, ethyl acetate2546 TITREII : SDIABATIC AND ISUTHERMAL COMPRESSIJITLITIES OFand clilorobenzene ; Regnault's results (" Relations des ExpQri-ences" and Mem. de Z'Accrd., 1862, 26, 262) for carbon disulphide,ether, ethylene chloride, ethyl iodide, chloroform, and ethylalcohol. I n the case of aniline, Griffith's results (Phil. Mag., 1595,[v], 39, 47, 143) were used. I n a few other cases, the values ofthe specific heats employed in the calculations are the mean valuesof several observers.These are given in the tables.B enz ene.to.01020304050607080to.010203040506070to.01020304050to.010203035a x 105.56.0061-0066.3272-0078.2685-1292.99101.9111 1.5t l l p t .0.00 13 1613511387142414641509156116231694Car6 o n Tetrachloride.a x 10'.63.2668.1473.2878.8585-3292.90101.20109.60&./(it.0*00072007417764278848147843287439071Garb o 7% Disulphide.63.21 0*000881866.79 912060-50 9422(34.52 972469.08 1002674.6 1033a x 106. dvJdt.Ether.a x 10'- dl+lt.114.30 0.00205 1127-05 2159140.85 2268158.7 2376169.0 2430u x lo';.8 1.9588.4595.65103.1511 1-41120-51130.03143.16156.5p x 106.91-0398-31105.96114.34123.94134.97147.16:159.81B x 106.81.4487-5293.84100.55107.92116-3k? x 106.152.97170.31188.9721 1.8244.1,IQUIDS BETWEEN ONE AND TWO ATMOSPHERES' PRESSURE.2547Ethylene Chloride.to*01020304050607080to.0102030405060708090to.02020304050607080to.010203040506070to.010203040506070a x 106.48-2361.6055.4259.8064.5969.7875.8082.5389.95a x lo6.32.8934.4 736.2538.2040.3242-6345-1847.8350.5753.45tlvldt.0.00086788970926095 6098750.00 102 2106011001143Aniline.LI'L'ICEI.0-00079808120827084288595877589679180942096 90Chl or0 b enze ne.R x lo6.clr;/clt.49.40 0.000841052.66 860056-03 878359.69 898563.73 920068-10 943872.88 968378.00 993883.50 0~001019a x 106.58.7563.0367.5072.4578-1084.1290-8098-16Toluene.dvldt .0-0011631200123912811328137914331490Ethyl Iodide.a x 106.59-7264.1068.7373.8779.7386.4593.80101-75dddf.0.00056525868608 1629665166735695871868 x 10"70.0575.1580.7386.9793.75101.20109.73119.17129.60B x 106.41-3143.5745-8648-3451-0453.9557.1560.5164-0767.87B X 106.67.0271.1275-2379.8285.0290.4296.28102.64109.29B x 106.79.3484-9490.8097-28104.70112.76121.57131.23B x 106.85-5992-4099.53107.29115.86125.40135.72146.82548 T Y R ~ K .: ADIABATIC AND ISOTHERMAL COMPRESSIUILITIES OFE thy1 Ace tute.to.010203040506070to.01020304050601".01020304050607080to.01020304050607075to.0102030405060708090100a x 10'.70.3076.4383-4191.2299.86109.25120.1133.4dvldt.0.0013701434150015671635170817841864C hl oro f or Tn .a x lo6.68-6462.9868.3174-1080.4887-3394.70a x 10'.57-2761-4065.6970.3575-4880.9387.0493-40100~00dvldt.0-0007940823086388863921895980-0010005m- X y 1 m e .d L'jlLt.0.00109251121114911791210124312791316135GE thy1 Alcohol.a x 106.83-8589-5295-65102.20109.62117.82126.65137-04142-95a x lo6.50.7546-1544.5243.6043-0242-7042.6042-7643.0543.3548.38dYltU.0.00 12881326136514131470164216241719I825ll'ntc~.dv/dt.+ 0*0,88 + 0-032073043804555265926557207 82- 0.046813 x 10,;.96.29104.98114-72125.3137-0149.8164-2181.4B x 1 0 .85.9092-94101.16110.04119.96130.74142-46B x 106.75-3980.4785.6991.3897.67103.9011 1-56119.24127-30P x 106.99.95106.28113.07120.61129.12138.78149.34161.8168.9P x 109,50-7848-4346-4545-2044-6944.6244.8945.4446.3 147.4348.6LIQUIDS BETWEEN ONE AND TWO ATMOSPHERES' PRESSURE.2549Ethyl Bromide.to. a x 10'. dvldt. Cp.* /3+ 10;.0 72.99 0,0008747 0.210 109.0510 79.05 9208 0.213 119.3420 86-45 9669 0.216 131.1830 94.84 0.00 10129 0.219 144.2040 103.70 1059 0.223 157.3* Cnlculated front rosults of Regnanlt (Zoc. tit.) and Battelli ( A t t i R. Accnd.Lincci, 1907, [v], 16, i. 243).Acetic Acid.to. a x lo6. dvldt. C,. * B x 106.15 72-80 0.001 01 9 0.480 88.7120 75.73 1029 0.485 91.9830 81-55 1048 0.494 98.4840 87.50 1068 0.504 105.1050 94-10 1093 0.514 112-5560 101.18 1122 0.523 120-66108.65 1157 0.533 129.3780 70 116.30 1200 0.542 138.57* From results of Schifl', Tiniof&ev, Ludekiog aiid others. (See Laridolt-Bornstein, " Tabellen,")Methyl Alcohol.to.a x lo6. daldt. Up.* B x 106.0 88.95 0.001409 0.570 107.5910 95-30 145 1 0.588 114.9420 101.95 1495 0.606 122.730 109.18 1543 0.625 131.040 11 7.02 1599 0.643 140.360 125.48 1666 - 1* From results of Regnaclt (Eoc. &.), Kopp (Ann. Phys. Chem., 1848, [ii], 75, 981,Timofh, (Compt. rend., 1891, 112, 1261) and Walker and Henderson (Trms. Itoy.Soc. Camdcc, 1902, [ii], 8, 105).Nitrobenzene.to. a x 10'. d'L'IClt. Up.* B x 101;.0 36-40 0- 0006730 0.338 44- 7010 38-61 6852 0.345 47.3020 40.90 6975 0-362 49.9830 43-22 7097 0.358 52.7940 45-70 7219 0.365 55-5850 48-37 7342 - -60 50.28 - - -* Prom results of Regnault (Zoc. c i t . ) and Sclilamp (Ann. Phys. Chent., 1896, [iii],58, $59).Uiscussioit of Results.As has been stated, the genesal order of error in the results forthe adiabatic compressibility is about 0.1 per cent.This is, how-ever, independent' of any possible error in the accepted value(Amagat's value) for the compressibility of glass. In any case2550 TYRER : ADIABATIC AND TSOTHEICMAL COMPRESSIBILlTIES OFeven supposing that this value contains, say, a 10 per cent. error,then the consequent error in the value of the adiabatic compressi-bility is only 0.3 per cent. in an average case. It is evident, there-fore, that very little error can arise from this source. Theaccuracy of the calculated values of P depends chiefly on theaccuracy of the values of a. Errors in the determinations ofdv/dt and C, have little effect on the value of fl.For example,a 5 per cent. elrror in the value of Cp introduces only an error of1 to 1.5 per cent. in the value of P.It must be remarked that whilst the observed values of theadiabatic compressibility refer t o a mean pressure of about 1.5atmospheres, those of dv/dt and of C, refer to the normal atmo-spheric pressure. The effect of pressure on the value of a is sosmall, however, that the results may be considered as all referringt o the atmospheric pressure, without any appreciable error beingmade. I n none of the experiments was observed a change of awith a change of pressure of about half an atmosphere. It maybe claime?, then, that the values of the compressibility obtainedby this method for low pressures are far more accurate than resultsobtained by the direct method, which yields very discordant resultson account of the evolution of a very small but important quantityof heat during compression.The effect of this evolution of heaton the results obtained by the direct method will be easily appreciated from what is explained in the note a t the end of this paper.In every case except water the value1 of both the adiabaticand isothermal compressibility increases with rise of temperature,the increase being the greater for the isothermal compressibility.For water, both compressibilities show minimum values.The theoretical application of the results obtained in this workare reserved for a future paper.Note O?L some Previous L)eter)ninatiom of the Compessibilitiesof Liquids at Low Pressures.On comparing the compressibility measurements a t low pressureswith some results of early investigators, it' was found that resultswhich were considered by their authors to be isothermal are reallyadiabatic.On studying the work of Quincke ( A I ? ~ .Phys. Chem., 1883, [iii],19, 401), i t was found that, judging from his method of deter-mination, his results could not possibly be isothermal, but wereundoubtedly adiabatic. The same was found to bs the case forthe determinations of Grassi (Ann. Chdm. Phys., 1851, [iii], 31,437), of Amaury and Deschamps (Compt. rend., 1869, 68, 1564)LIQUIDS BETWEEN ONE AND TWO ATMOSPHERES’ PRESSURE. 2551and of Collodon and Sturm (Ann. C&m. Phys., 182T, [ii], 36,113), and of a few other investigators. It is important to pointout the true nature of the results of these investigators, becausethey have always been recorded in tables of physical constants andproperties of liquids as isothermal compressibilities (see, forexample, Landolt-Bornstein, ‘‘ Tabellen,” 191 2).Their results differed very considerably from the isothermal com-pressibility determinations of later investigators, a matter whichappears to have caused some surprise, although the reason of thediscrepancy was apparently never discovered.When a liquid is compressed by a small pressure-say, by oneatmosphere-there occurs a small rise in temperature amountingt o a few thousandths of a degree, yet sufficient, if neglected, t ocause a difference of 10 to 40 per cent.in the observed isothermalcompressibility, and, as can be easily imagined, the elimination ofsuch a small change of temperature is exceedingly difficult.It can be easily understood, therefore, that early investigators,ignorant of this small change of temperature, and using therino-meters scarcely sensitive enough to detect it, would obtain, notthe isothermal compressibilities which they were attempting t omeasure, but really adiabatic compressibilities, or, more correctlyin the majority of cases, results which lay between the adiabaticand isothermal values.Quincke’s method of investigation was very similar in principleto that described above. He operated a t very low pressures, thehighest he used being, in fact, little more than 50 cm.of mercury.The latent heat liberated in the compression would be so smallas to be undetectable, and i t is quite impossible that any appreci-able quantity of this heat should have disappeared during anexperimentl. Moreover, he mentions that he worked as quickly aspossible, as then better results were obtained. Quincke himselfdoes not appear t o have considered the possibility of the liberationof heat during the compression. I n the table below, Quincke’sresults are compared with t’he adiabatic and isothermal valuesrecorded in this paper. It -will be Seen that there is quite a closeagreement between Quincke’s results and the adiabatic valueswhich leaves no room for doubt that Quincke’s values are reallyadiabatic.Grassi (Zoc.c i t . ) used a similar form of piezometer t o that ofQuincke, but he worked a t higher pressures, up to 8 or 9 atmo-spheres. A t the higher pressures, the heat produced in the com-pression would be so appreciable thatl much of i t would be lostduring the time of an experiment, but at the) lower pressures th2552 TYRER : COMPRESSIBILITIES OF LIQUIDS, ETC.compression would be adiabatic, or approximately so. Now, as theisothermal compressibility is much greater than the adiabatic,Grassi found, as a consequence, that the compressibility of a liquidincreases with the pressure, whereas it really decreases. Forinstance, it will be seen in the table that for ether, alcohol, andchloroform the compressibility, according to Grassi, is greater a tthe higher pressures.It will be noticed that where a comparison with the presentauthor's results a t one to two atmospheres is possible, Grassi'sresults agree quite well with the adiabatic values.Jt may be con-cluded, therefore, that for the lower pressures Grassi's results areapproximately adiabatic, but a t the higher pressures they liebetween the adiabatic and ths isothermal values.Amaury and Deschamps (loc. cit.) made compressibility measurements between 1 and 10 atmospheres' pressure. For a change ofpressure of 10 atlmospheres there would be quite an appreciablechange of temperature, but as they took minute readings of thevolume change and eliminated what they considered to be acci-dental changes of temperature by plotting the readings against thetime, and extrapolating to the zero point on the time ordinate,their results would be approximately adiabatic. The pressures fortheir experiments being higher than correspond with the author'sresults, i t is to be expected that their values will be somewhatlower than the new adiabatic values.From a study of the early experiments of Collodon and Sturm(Ann. Chim. Phys., 1827, [ii], 36, 113, 225; Ann. Yhys. Chem.,1828, 12, 39), it would appear that their results were largelyaffected by the negligence of the latent heat of compression, andthat, their values are, as it were, partly adiabatic and partly iso-thermal.Water 0.0" 1-1-5 50.30 60.75 50.780.0 1-105 53.93 53.21 81.44disulphide \ 17.0 1-1.5 63.78 59.35 92-301 6.0 1-1-6 59.70 59.00 85.951-1-5 82.82 83.85 99-950.0 1-1.5 116.57 114.30 152.97Benzene \ 16.78 1-1-7 66.10 64.60 93.11!::1 1-1.7 97.45 94-08 111.2( 14-32 1-1-5 134.23 132.6 178.1 EtherQuincke(loc. ci't.NEWBERY: ELECTROMOTIVE FORCES IN ALCOHOL. PART V. 25530.0"Water I Grassi(loe. G i t . )( Ethyl alcohol { 1::;Ether oh amp8,(loc. cit.) Carbon disulphide 14-0 =-- iPressurefound tohaveno influenceon the com-pressibility.1-7-821-3-411-1-581-8-361-2-301-9-401-1-671-8-971-1-261-1-301-1.9250.349.948-0131.0111.0140.0163.082.885.390.499.162.864.876.31-10 83-51-10 91.11-10 109.01-10 128.01-10 63.550.7549.7748.20-114.30132.3087-9591.3562.2304-00----83-8592.50114.30132.3058-2850.7 849-7748.25-152.97177.6104.6108.391.795.0----99.95109-5152.97177.690- 1The rwults of other investigators were carried out a t sufficientlyhigh pressures to ensure the complete elimination of the heat ofcompression.The practical part of the work was carried out in the PhysicalChemistry laboratories of the University of Geneva.THE UNIVERSITY,MANCRESTER