2032 PRIDEAUX : THE VAPOUR PRESSURES AND MOLECULARCCX I I I. -The Vapour P m w u r e s and MolecularVolumes of the Mercuric Halides and the Rela-tions between Atomic Volumes of Elements Beforeand After Combination.By EDMUND BRYDGES RUDHALL PRIDEAUX.THE present investigation has the aim of comparing the volumes ofliquid elements with those of their liquid compounds. As com-parison temperatures the boiling points under atmospheric or otherequal pressures have been retained. The results of Young andothers show that at vapour pressures below one atmosphere it is amatter of indifference whether equal pressures or correspondingpressures are chosen for the above comparison. The procedure offinding atomic volumes at the boiling point is too well known toneed description, It has been departed from in several particulars.(1) No attempt has been made to tabulate difference of molecularvolume, the ratios only being compared.(2) No experimentally inaccessible atomic volumes, such as thoseof carbon and hydrogen, have entered into the calculation.(3) The effect of structure on the molecular volume has beenas far as possible eliminated by considering only simple compounds,in which there is not more than one multivalent element.The term atomic or molecular volume at any pressure is usedthroughout to denote the total volume occupied by the gram-atomor gram-molecule liquid at such a temperature that its,vapourpressure is that specified. Under these conditions the ratio betweenthe specific volumes at, for example, 760 and 200 mm.pressure isa constant for liquids which are not associated or dissociated.For example, -___- v‘‘760) is for C6E6 1.050, PC1, 1,050,VD4 2 w ’ )C,H,Br2 1.048, 0, 1.048, HC1 1.049. Instances could be multiplied.Associated and monatomic liquids : H,O 1.030, CH,*CO,H 1,041,Hg 1.012, A 1.014 (760-400), the normal value for this pressureinterval being 1.028.These relations may be connected in the following way wit.htheories involving the conception of (‘ co-volume.” The term isused throughout as generally to mean the volume through whichthe molecules have motion relatively to one another, sometimescalled “ free ether,” it5 distinguished from the “ bound ether ” orinteratomic space, which, together with the volumes of the atomsthemselves, makes up the (‘ b ” of van der Waals’ equation(Clausius) which no other atom can penetrateVOLUMES OF THE MERCURIC HALIDES.2033Now, if we imagine an ideal and normal liquid expanding fromthe condition in which the co-volume is zero (which, as will be shown,is probably not in most cases -273O), the vapour pressure willincrease at the same time, and since the pressure is a function of thetemperature and the increase of cevolume also a, function of thetemperature, the increase of cevolume may be expressed as afunction of the pressure.Thus if JTp and Vo are the volumes of liquid at “ p ” and zeropressure :4(p)Vo being therefore the co-volume at= Yo[ 1 + &)I*p.”For another liquid, also expanding between (‘ o ” and “ p ” :A t a higher pressure, p l , the volumes become Tr0[1 + +(p,)] andV:[ 1 + +’(p,)] respectively.But as shown above, in the case of normal liquids :that isTherefore + ( p ) = +‘(p) and + = +I.these liquids.to the lowest, and eventually, zera pre-sure :(1) The increase of co-volume is the same function of “ p ” for allAssuming that the same law of expansion holds down(2) The co-volumes ‘vot+(p) and Fvo+(p) at any pressure areproportional to the actual volumes of the molecules Vor and Vo.(3) The ratios between the volumes of the liquids at equal vapourpressures are equal to the ratios between the actual volumes of themolecules.As soon as the vapour becomes so dense that the specific molecularattraction begins to have an effect in that phase, the nature of therelation between increments of vapour pressure and volume willprobably change, and +(p) take another fwm.By taking this intoaccount, Mills (J. Physical Chem., 1902, 6, 209; 1904, 8, 383, 593)Q has deduced the formula -- = constant for a series of normal x- yoliquids at different temperatures- (q =inner heat of vaporisation of1 gram liquid ; d and D = densities of liquid and vapour). This equ*tion holds up to the critical temperature. It can be shown that attemperatures so low that JD may be left out of account compare2034 PRIDEAUX : THE VAPOUR PRESSURES AND MOLECULARwith 7 2 (considerabIy below the boiling point under atmosphericpressure), Mills' equation necessarily leads t o the regularity men-tioned above.For consider two liquids, A and B, at two pressures,P and P, and let Q be the total molecular latent heat of vaporisationand €2 the gas constant. To the values of P, T, Q and d for thet,wo liquids assign the symbols:P' T A Y"B Q'A Q'B d'A d ' ~P TA 25 Q" QB d~ dg.Then for each liquid, by Mills' formula:Multiply by 5!"/T' :KT' &'/T&Id';i -r . . . . . . . , . , (2).Divide equation (2) for A by (2) for B :But for normal liquids :Alsowhere " c " is very small, not greater than 0*0005 for the mostdissimilar liquids : thereforeThis relation can only be deduced from Mills' formula a t tem-peratures at which the density of the vapour may be neglected.A recalculation of the constant (Mills, J . Pltysical Chem., 1909,13, 512) has proved that the formula holds with the greatestaccuracy in this region.The assumption, then, that +(p) remains constant down t o thelowest pressure is in accordance with the above theory of molecularattraction, and with the facts of expansion down to the lowestvapour pressure at which liquid volumes have been compared.The volumes at zero pressure could be found by extrapolationif the pressures were known with sufficient accuracy.The laws ofexpansion of liquids lead to an ideal state of zero co-volume for aperfect liquid, just as the laws of expansion of gases lead to thestate of an ideal gas at the absolute zero.In the case of liquids, however, it does not seem probable thaVOLUMES OF THE MERCURIC HALIDES. 2035the zero of pressure should occur at the same temperature in eachcase. The necessity for a special zero in each case has already beenfelt in explaining the deviations from the reduced equation ofcondition (Young, Stoichiometry, p.237).I f it were possible, then, to obtain the liquid volumes of non-associated elements and compare them with the volumes of liquidnon-associated compounds all at the same pressure, the ra€ios ofthese volumes would correspond with the ratios of the actualvolumes of the molecules, the expansion or contraction on com-bination being thus discoverable.Owing, however, to the scarcity of data, a direct comparison ofthis sort is in few cases possible, and where it can be made, mostof the expansions are found t o be not quite normal, The degreeto which these irregularities influence the ratios will appear fromthe experimental data.EXPERIMENTAL.Materials.-The mercuric chloride was resublimed in a current ofdry chlorine.Several analyses were made both of the freshlyprepared salt and that which had been boiled for some time in theair, or heated in the dilatometer. The mercury was determined asmercuric sulphide, and the chlorine as silver chloride, and theresults were satisfactory.The mercuric bromide prepared to arder by Hopkin and Williamsgave satisfactory results on analysis, and was employed withoutfurther treatment in some cases. A sample redistilled with a littlebromine gave identical results on analysis and in the dilatometer.The mercuric iodide was also Hopkin and Williams’ preparation.It was analysed by electrodeposition on a silver-plated platinumbasin cathode, and gave theoretical results.It was afterwards re-distilled to remove traces of a non-volatile impurity, whichapparently did not affect the analysis, but made the liquid surfacedifficult to locate in the dilatometer.DiZatometers.-These were of fused silica, and were supplied bythe Silica Syndicate. They were graduated as required by a,diamond fixed in place of the needle on a, divider, and calibratedwith mercury in the usual way. The bulbs were cylindrical, ofabout 2 C.C. capacity. The stems were graduated in mm., and eachcm. length held from 0.02 to 0.03 C.C.Thermometers.-Three nitrogen-filled mercury thermometers wereused :(1) Reading up to 600° in 2O; standardised at the NationalPhysical Laboratory2036 PRTDEAUX : THE VAPOUR PRESSURES AND MOLECULAR(2) To 600° in 2O, and (3) to 500° in lo; these were standardisedThermostats.-For temperatures up to 260° a bath of paraffinwax was used, and for these as well as higher temperatures thevapours of liquids boiling under various pressures were employed.These have been tabulated by Landolt and Bornstein from themeasurements of Ramsay and Young as follows :by comparison with (1).260-280" ........................ Monobromonaphthaleiie.280-302" .......................Diphenylamine.360" ........................ Mercury.For the interval 31 0-339O, anthracene, boiling under diminishedpressure, was used instead of the vapour of mercury.The paraffin wax thermostat consisted of a large beaker ofresistance glass jacketed with asbestos and heated by a gas flamefrom below.Additional heat was supplied by an electrically heatedframe of iron wire in the liquid, which was stirred by a brass fanwheel at the bottom, and automatically regulated by a glasscylinder holding about 150 C.C. of air, which extended to the bottomof the liquid, and communicated by means of a capillary tubewith a mercury gas regulator. The temperature could by thismeans easily be kept constant to a few tenths of a degree. Thethermometers, etc., were not directly in contact with the liquid,but were protected by tubes of combustion glass.The vapour-bath first used had somewhat the form of it Liebig'scondenser. To a vertical glass tube open at each end was sealedan outer tube above and below.The vapour of the boiling liquidfilled the space between the tubes. The liquid was contained intwo side-bulbs joined to the lower end of the outer tube at anangle of about 45O. It was heated by burners and also electricallyby platinum spirals. The electrical heating was used mainly toprevent bumping a t low pressures. The lower part of the apparatuswas protected by an asbestos box packed with magnesia. In orderto save time, this was electrically heated by nickel wire. The upperpart was shielded by movable rings of asbestos covered with felt.It was found that this form had several disadvantages when usedfor high boiling liquids-the radiation is large compared to theevaporating surface, and the dead space between the internal sealand junction of sidetubes adds considerably to the time requiredto send the ring of condensing vapour a sufficient distance up thetube.For the purpose in hand, however, the apparatus had thegreat advantage of allowing the introduction of dilatometers frombelow, so that they could be heated from the top downwards, thusavoiding the sometimes troublesome distillation of drops of liquidto the upper part of the tubeVOLUMES OF TEE MERCURIC HALTDES. 2037In the form of vapour-bath afterwards used, the dilatometer witsintroduced from above into a glass tube about 2 cm. in diameter,which was heated by the vapour of a, liquid boiling in a muchlarger bulb, the lower end of the inner tube being protected fromthe radiation of the superheated liquid by the well-known methodrecommended for thermometers in Young's Fractional Distillation.The vapour space was connected through a reservoir of aboutYapour pressures of mercuric cldoridc, bromide, nitd i d Ic.~ 220200 1801009080i 060 2sB 50 .-s$E2u40 530 '201010 litres capacity and a drying tower to a water exhaust pump orpressure pump as the case might be. The pressures were read ona mercury manometer attached to a wooden metre scale.Vapow Pressures.-These were regulated and read by means ofthe air reservoir and mercury manometer already mentioned.Theliquids were boiled under various pressures in a combustion glasstube. The sides of the tube were protected from overheating byhorizontal pieces of asbestos board, and from cooling radiation b2038 PRIDEAUX : THE VAFOUR PRESSURES AND MOLECULARslip rings of asbestos covered with felt.The liquid was in everycase boiled at such a rate that the thermometer column was com-pletely immersed in the vapour. The bulb of the thermometer wascovered with asbestos.The results are tabulated below, and also shown in the figure.The experimental points are indicated by X. A few vapourpressures at lower temperatures by Wiedemann, Stelzner, andNiederschulte (Ber. deut. physikal. Gees., 1905, 7, 159) are indicatedas follows: mercuric chloride by circles, mercuric bromide bycrosses, and mercuric iodide by circles. Most of these were deducedfrom the loss in weight when a, measured quantity of air waspassed over the solid at the temperature in question, and some byobserving the pressure and temperature when the compounds weresuddenly sublimed into a cooled tube.The curve so obtainedought not, of course, to be continuous with that of the vapourpressures of the liquids, but the observations in the neighbourhoodof the melting point are not numerous enough for any discontinuityto be observed. The three observations recorded (by the authorsquoted above) at higher temperatures do not agree with the presentresults, probably because the method they employed was not sosuitable for higher temperatures. The pressures used for the sub-sequent calculations are those read from the manometer at 16-0°.The following are the vapour pressures (to the nearest millimetre)corresponding with the corrected temperatures.The index numbersto the right of the pressures refer to different series of experiments.Mercuric Chloride.P.(min.) to. P. to.844, 309.0 751, 303.0836, 308'2 741, 302.5822, 307.5 730, 302'7820, 308-1 724, 301.5816, 306.8 719, 301.9810, 307'5 711, 301'4803, 306.7 705, 300'5800, 307'0 702, 300.8798, 305'9 690, 300'0790, 306'4 665, 297.5781, 305.1 627, 294.8780, 305.8 612, 294*5778, 304'9 587, 292-0770, 305.0 582, 292-3766, 304'8 548, 289.0759, 304.3 544, 289.4757, 303.9 530, 287'9754, 303.6 505, 286.1Mercuric Bromide.P.(nlm.) to. P. to947, 331.0 729, 318.3897, 328'4 728, 317.8847,, 325'4 719, 317'3819, 323 9 712, 317.2819, 324.0 705, 315.8807, 323.3 705, 315.8800, 322.8 702, 316.6789, 321.8 692, 315.9779, 321.3 641, 312.2778, 321.4 6G9, 309.2772, 320.9 602, 309.5770, 320.9 562, 306.37G6, 320'7 519, 302 5760, 320.3 485, 298.5755, 320.0 410, 290.7753, 319.8 409, 292.0750, 319.5 370, 287.5749, 319.6 331, 283'0740, 319'0 252, 271.57381 318'8 225, 266'08365 324 9 7194 317.87884 322.1 680, 315.0758, 319'8 4474 296.0Mercuric Iodide.P.(mm.) to.P. t".861, 360.5 730, 351.0829, 358.3 720, 350.8821, 358.0 720, 350.5819, 357'9 711, 350-1809, 357.1 703, 349.5801, 856.8 701, 349.3799, 356'4 699, 349.3790, 355.9 646, 345.2783, 355.3 598, 341.3780, 355.1 551, 337'0770, 354.5 500, 332.0760, 353.7 453, 327.5760, 353.5 404, 321.9758, 355.5 357, 316 0750, 353.0 310, 309.5246, 352.5 270, 303.7i43, 352'5 232, 297.5740, 352.VOLUMES OF THE MERCURfC HALIDES.2039A calculation of the value of heat of vaporisation divided byabsolute temperature from the above data was undertaken, sincethis constant might be expected to throw some light on the questionas to how far these compounds partake of the nature of fused salts.The values of &/T for fused salts ought Lo be unusually high inview of the high degree of association which is usually attributedto t'hem on other grounds (Bottomley, Trans., 1903, 83, 1421;Lorenz and Kaufler, Ber., 1908, 41, 3727). The vapour-pressurecurve was first investigated by Ramsay and Young's method.Theabsolute temperatures a t which the compounds attain certain vapourpressures are tabulated below, and compared with the correspondingtemperature for a standard liquid (fluorobenzene) :TABLE I.I.P. (THgCl,).240300530 561.0700 573.0aoo 579.5860 583.090011.T( HgBr,).542-0551-5576.5589 0596.0599.8601%r i r .T(Hg1,).572.0581.5608.5622.5629.5€33 -5636'0IV. L/IV II./IV. III./IV,TC,H,F.325.5 1.665 1.7573315 1.664 1.765317.0 1.617 1'662 1.754356.0 1.610 1.655 1.749360.5 1-608 1'653 1.747393.0 1.606 1'651 1-746364.0 1.650 1-745From the above results a value of the constant was obtained inthe equation :in which TA are the temperatures of mercury halides, and 5'"~ thoseof fluorobenzene at the same pressures.'( c " for HgCI, = 0*00050.,, ,, HgBr, = 090037.The values of dP/dT were then found by a graphical method ata series of temperatures as tabulated below, and from these theheat of vaporisation Q and &IT from the formula:,, ,, HgI, = 0.00025.1*985T TE calories.Q = m( d2')Q and &IT at 760 mm.Q(ca1.) &ITHgCI,. ....................... 13910 24 *1HgBr, ..................... 14200 23 23HgI, ........................ 14700 23.5It has been shown by Nernst that the normal value of Q/Z' a2040 PRIDEAUX : THE VAFOUR PRESSURES AND MOLECULAR760 mm. increases slightly with the boiling point of the liquid inquestion according to the formula :Q/ T = 9.510g.T - O.O07T,and for a liquid boiling at 304-354O :& / T = 22.19 normally.There appears, then, t o be a certain amount of association, less,however, than in the case of water and alcohoIs, for which &/T =26.The association evidently diminishes in the order :Hg Cl ,-+Hg Br,+HgI,.Specific Gravities of the Lipuida.Procedure.-After the right weight (10-15 grams) had beenintroduced into the dilatometer, this was evacuated, sealed, andafter a preliminary heating (great care is necessary t o avoid burstingthe dilatometer) introduced into one of the thermostats andkept at each temperature until the volume remained constant.After the experiments were completed, the dilatometers wereopened and the contents boiled out, the weight of substance beingthus checked.In calculating the volumes corresponding with eachscale division of the dilatometers, the expansion of the silica was nottaken into account, since it proved to be outside the range ofaccuracy aimed at.Thus the greatest range of temperatures was from 255O to 357O.The linear expansion of silica is :0.449 x 10-6 or 0.59 x 10-6.Taking the larger figure, 3a- 1.8 x 10-6, and the total expansionof unit volume contained by a silica bulb at room temperature is4-59 x 10-4 at 255O, and 6-43 x 10-9 at 357O.The measured volumes have therefore to be increased, and thespecific gravities decreased by about 1 in 2000.This small correction has no effect on the relative specific volumes,but has been introduced into the absolute molecular volumes.Thefollowing specific gravities are calculated directly from the experi-mental data, the numbers to the left referring to different seriesof experiments in different thermostats with different quantities ofcompound and different thermometersVOTAUMES OF THE MERCURIC HALIDES. 2041t'.( 1 ) 281.0(2) 287 7290.7300'6( 7 ) 240.0244.0251.0261 -5(2) 2415248.0252'0259.0(3) 250-0(1) 254.5260.5266 0270 0276.0281 .O(2) 259.5267.5575.0D.43984'3764.3694.3485.1125'1105.0785'0465.1045.0825.0il5.0485.0i65.2365-2185.2075-1905 1695.1575 '2265.2085.li7Mercuric Chloride.t".D.(3) 290-7 4.377301.6 4.357301 'G 4.348( 4 ) 291.0 4'377301.5 4'357Mercuric Bromide.(3) 253.0258'0260 0260 '0272 0278'0286.5293.5(4) 354.55.0665.0525'0445'0625 -0445.0074-9904-9644.938Mercuric Iodide.(2) 281.0(3) 275.0279'0288.0286 *O292'0294.5301'5( 4 ) 282.55.1605 - l i 75.1665.1545.1415,1255'1135.0905.156to.(5) 311.0318.0328.0336 .O(6) 357.0(4) 302.0( 5 ) 258 52805292.0301 -5(6) 313.0320 -0329.5339.0(4) 2945301.5(5) 311.5323 0329.5337.5339 *o(6) 339.0( i ) 356.0D.4.3324.3164.2934.2764.2384 '9085-0544'9764.9445.0544'8744'8444.8114.7845.1195 0905.0755'0395.0144-9854'9784.9674'915The alteration of density with temperature can be expressed asMercuric Chloride, 280-335O :Mercuric Bromide, 240-340° :Mercuric Iodide, 255-355O :follows :D, = 4 400 - 0.0032 18 (to - 280).Dt=5*116 -0.00338 (to-240).Dt=5 238 - 0.00322 (to - 255).Densities and Vapoui Pressures of the Elements Compared withthose of Compounds.The data taken from the tables require no comment.The densities of liquid iodine are those of Billet from Constantsof Nature (Smithsonian). They appear somewhat irregular, whichis due probably to the dark colour of the vapour at the highertemperatures.Special weight has been given to the values at thelower temperatures, and that found at the boiling point byDrugman and Ramsay (Trans., 1900, 77, 1228). The columnVOL. XCVlI. G 2042 PRIDEAUX : THE VAPOUR PRESSURES AND MOLECULARheaded (( pressure " refers as throughout to that of the saturatedvapour :TABLE 11.D emi t y.Pressure. Cl,.Br,. I,. Hg. HgCI,. HgBr, HgI,.200 1.617 3.109 3'888 12.901 5.037 5'124400 1.585 3.050 3.795 12'824 4.944 5-026560 1.572 3'018 3.749 12'785 4.380 4-890 4'970760 1.558 2.985 3'705 12.747 4.348 4.840 4.920860 1.550 2.975 3-685 12'738 4'336 4.821 4.900The expansions over certain intervals of pressure are comparedwith the normal below. Since the volumes are accurate t o 0.2 percent., the expansion, for example, between 560 and 860 mm. ofmercury is 0-010 & 0.002 in the case of mercuric chloride, and thisis distinctly below the normal value. Thus for the pressure intervalnamed, the value of the ratio 0(560)/D(860) is: Normal 1.023,Hg 1.003, C1, 1.014, Br, 1.014, I, 1.017, HgC12 1.010, HgBr, 1.014,HgI, 1.014.The cause of these abnormalities is probably association as affect-ing not so much the liquid volumes as the vapour pressures, andhence the temperatures of comparison [see van't Hoff, Lectures,111, p.27 (Lehfeldt)].From the numerical values of the expansion the degree ofassociation would appear t o diminish in the order:C1, -+ Br, -+ I,; HgC1, --+ HgBr, -+ HgI,.Now, comparing the expansions of the compounds with those oftheir constituent elements, it may be seen that the magnitude ofthe abnormality is such as would be produced if the expansion wereadditively composed of the expansions of the elemenb.Yk(860) and 2 Y~l(86O) Thus the ratiosare, for HgC1, and Hg+ 2C1 .................. 1.010 and 1.012HgBr, ,, Hg+2Br ..................1.014 ,, 1.012Hgf, ,, Hg + 21 ..................... 1-014 ,, 1.015V ~ ( 5 6 0 ) ZV"(560)While the magnitude of the possible volume error (14-20 percent. on the expansion) does not permit of a certain conclusion onthis point, yet the evidence for such a relation is strengthened byan examination of the few other compounds for which data, areavailable.Thus, comparing the ratios V M and SV, at 760 and 200 mm.:PCI, ..................... 1.050 P + 3Cl .................. 1.050fC1 ..................... 1'041 I + 01 ..................... 1'043The importance of this for the question in hand rests in thVOLUMES OF THE MERCURIC HALIDES. 2043increased constancy it gives to the ratio VH/SV under variouspressures.In the first part of the paper reasons were given why the ratioshould be quite constant in the case of normal elements and com-pounds far enough removed from their critical points.It now appears that a constant ratio JT& J ~ A can also be definedfor the abnormally expanding elements and compounds considered,as may be seen from table 111.The variations of the ratios are within the limits of error to beexpected from at least one of the volume measurements involved.TABLE 111.100 JToEunzes of the Elements become on CondbinationP=860 nim.760 560101.7 -109.6 -HgC12,Hg+2Cl .................. 101.9 101.8HgBr, 107.4 107.3 - .............................. 107 *5HgI, ................................. 109 '5 109.5P=760 mm. 560 400 200PC13--+P 4- 3Cl .................. 105.6 105.2 104.9 105.0PBr3 ............................... 107'8 - - 107.5PCI, .................................96 ' 5PRr, ................................. 101 *9 - - 101.6NOTE.-The vapour pressures and atomic volumes of the phosphoriis halidesand iodine chloride (referred to on p. 2042) are taken from the tables and Trans.,1907, 91, 1711 ; 1909, 95, 445. The temperature corresponding with 200 mni. forphosphorus tribromide has been calculated by Ramsay and Young's method(comparison liquid, phosphorus trichloride).The expansion on combination therefore increases with increasingatomic weight for the mercuric halides as for the other compoundsquoted (except phosphorus pentachloride, in which case the con-traction on combination diminishes). According to the first partof this paper, the changes of volume on combination areapproximately equal to the difference between the sums of theinteratomic volumes of elementary molecules and the interatomicvolume of the compound molecule. If this is correct the physicalinterpretation to be put upon the results is that, for example, theinteratomic volume of a C1, molecule is less than that of HgC1,(Hg being monatomic), and that in the other cases (except PC1,)the sums of the interatomic volumes of the elementary moleculesare less than the interatomic volume of the compound molecule.Now, comparing together two similar combinations, such as mercuricchloride and bromide, the relative increase of interatomic volumeon combinat,ion is greater in the latter case, indicating a, smallermolecular attraction for the atoms of mercuric bromide than forthose of mercurio chloride. In the cap88 of phosphorus halides, also,- I -6 s 2044 FRANK : CONTRIBUTIONS TO OUR KNOWLEDGEthe bromide combination exhibits a smaller molecular attractionas judged by these volume relations than the chloride. Thus therelative affinities are in the few cases investigated in the sa.me orderi ~ s the same affinities deduced from other considerations.I desire t o express my thanks to the Royal Society for it grant inaid of this research, and to Professor Donnan for the facilitiesafforded at the Muspratt Laboratory and the interest he has takenin the work.THE MUSPRATT LABORATORY,THE UXIVERSITY, LIVER POO L