294 Mr. J. B. Hannay .
Observations , \amp ; c. ( continued ) .
Office of the Chief Signal Officer .
Daily Bulletin of Weather Reports .
June to August , 1877 .
4to .
Washington 1879-80 .
Annual Report of the Chief Signal Officer , 1879 .
8vo .
mgton 1880 .
The Qffice Office of the Surgeon-General U.S. Army .
Index-Catalogue of the Library .
Vol. II .
4to .
Washington 1881 .
The Office West Point : U.S. Military Academy .
U.S. Geographical Surveys west of the 100th Meridian .
Yols .
II-VI .
4to .
Washington 18 75'80* The Academy .
Adams ( A. Leith ) , F.R.S. , G. H. Kinalian , and R. J. Ussher .
Explorations in the Bone Cave of Ballynamintra , near Cappao-h county Waterford .
4to .
Dublin 1881 .
The Authors ' Campin ( Francis ) The relative merits of Vacuum and Air Pressure Brakes .
8vo .
Leeds 1881 .
The Author Samt-Lager ( Dr. ) Nouvelles Remarques sir la Nomenclature Bo-tanique .
8vo .
Paris 1881 .
The Author .
Struve ( Hermann ) Fresnel 's Interferenzerscheinungen theoretisch und experimentell bearbeitet .
8vo .
1881 .
The Author .
Todd ( D. P. ) Report on the Total Solar Eclipse of 1878 .
4to .
Washington 1880 .
On the Use of the Electric Telegraph during Total Solar Eclipses .
8vo .
Observations of the Transit of Mercury , 1878 .
May 5-6 .
8vo .
And three Reprints from the American Journal of Science .
The Author .
" On the Limit of the Liquid State .
" By J. B. Hannay , f .R.S.E .
, \amp ; c. Communicated by Professor G. G. Stokes , LL. I ) .
, D.C.L. , \amp ; c. , Sec. R.S. Received February 22 , 1881 .
Read March 10 , 1881 .
The uncertainty which characterises our knowledge of the true condition of a fluid immediately above and below the critical temperature , induced me to enter into a full examination of various fluids , with the object of gaining accurate definitions of the liquid and gaseous states , as well as to arrive a$ a true conception of the state of matter to which the term vapour can be applied .
In a former paper , which the Royal Society has honoured me by publishing , * experiments were detailed which seemed to show that the liquid state terminated at the critical temperature , and that no amount of pressure would suffice at .
any higher temperature to render the fluid capable of exhibiting ' " *911 the State of Fluids at their Critical Temperatures .
" " Proc. Roy .
Soc. , " vol. 30 , p. 484 .
On the Limit of the Liquid State .
surface tension or capillarity ; in fact , that the state of a fluid above that temperature coincided with the properties we call gaseous .
The paper concluded , " The difference between the liquid and the gaseous states is not then entirely dependent upon the length of the mean free path .
but also upon the mean velocity of the molecule .
" That is to say , we may compress a gas ( when a few degrees above the critical temperature ) to a less volume than it might occupy as a liquid , and it will still remain gaseous .
In the following paper , therefore , the term liquid will be applied only to such bodies as exhibit surface tension , either as capillarity or by forming a permanent limiting surface when in contact with a vapour or gas .
The term gas will be applied to that state of a fluid which precludes its being reduced to a liquid by pressure alone , in other words , to any fluid above its critical temperature .
The term vapour will be applied , as has already been done by Andrews , to fluids which can be reduced to liquid by pressure alone , that is , to any aeriform fluid at a temperature lower than the critical .
Thus carbon dioxide is a vapour at ordinary temperature , but is a gas at temperatures over 31 ' .
A further distinction of gas and vapour lies in the fact that , on increasing the pressure , the volume of a gas goes on diminishing in a regular way , whereas there is a part of the curve representing pressure and volume of vapour where the cuive is asymptotic , that is , where the vapour is in contact with its liquid .
In the following paper reasons will be shown for believing that the gaseous state depends entirely upon the mean velocity and not upon the mean free path of the molecule at all .
The difference between vapour and liquid , on the other hand , is entirely one of the length ot the mean free path .
The methods of experiments used were similar to those detailed in the paper above referred to , but a larger apparatus was employed , so that the results might be more distinctly visible .
It was soon noticed that the readings of pressures of manometers varied a little with the diameter of the tube employed , the smallest bore giving the highest reading , and this was the case to such an extent as to cause an error of two atmospheres in / 0 , and abou* five in 100 ; the higher the pressure the greater the difference .
1 he wide tube was about 0'8 millim. in diameter , the smaller OT millim. How , whether this error was caused by the hydrogen condensing against the glass , and being thus lost as a manometric substance , oi whether it was caused by the hydrogen being dissolved in the film of moisture which may be supposed to adhere to the interior of the tube , has not yet been determined .
It has been shown by Professors Living and Dewar that the moisture adhering to glass is not driven off till nearly a red heat is reached , and we may be sure that tin capillarity of the smaller tube would cause it to retain moisture more eagerly than the larger one .
" Whatever was the cause , it was almost invariably found that manometers with small bores gave higher readMr .
J. B. Harm ay .
Fig. 1 .
On the Limit of the Liquid ines than those with wide bores .
They were dried by passing dried hydrogen through them lor over two hours and keeping them warm " " in orteto obtain readings which would always be near the truth , and be more independent of accidental errors , an apparatus was constructed with two manometer tubes , and the manometers were made of tubes as wide as was consistent with the strain they were destined to bear .
The apparatus as used is shown in fig. 1 , where the two manometers are shown fixed in the two upright branches while the pressure screw is at the right hand , and the working tube at the left .
The air-bath has been drawn as though it were transparent , to show the internal arrangement .
The working tube is recurved , so that the liquid to be experimented upon is contained between the mercury and the sealed top of the tube .
The air-bath consists of two cylindrical baths with holes in the lids for passing the working tube through , and an outside cover which keeps the heat from the lamp from being too quickly radiated .
The internal baths are supported by one of Fletcher 's solid flame burners , and the bottoms are covered by a layer of non-conducting cement .
The outside cover has openings at the top for the escape of the burnt gases , and its top is covered with a thick layer of asbestos wool , to prevent cooling .
The whole of the baths and cover were made of iron , as the high temperature used caused copper to scale heavily .
Two thermometers were used in the bath , one on each side of the working tube ; and at first each thermometer had a little one fixed to it for temperature corrections , but it was subsequently found that one liuug b6tw66n the two gave quite as much accuracy .
The two thermometers used were of soda-glass with cylindrical bore , and registered the same temperatures to within 0''5 , ^ being chosen from a number .
They were heated and cooled from 0 to 0 over seventy times , and one sent to Kew , where it was compared with the standard up to 100 , and the stem calibrated and the coi i ections up to 350 calculated .
The zero points of both thermometers rose from 0o,2 to 2'"2 during the preliminary heating and cooling .
The changes were determined daily .
The temperatures given in this papei may , therefore , be considered practically correct .
Two small thermometers were fixed to the manometers for temperature corrections .
After trying several stands I found that shown in fig. 1 to be the most convenient and steady ; it is simply a large block of wood with a groove cut in it , in which the tube lies , the two upright arms preventing any movement of the apparatus .
Ihe packing of the joints has been described before , and I would only add that I find it better to face the India-rubber packing with leather by fixing a piece of fine soft leather to the face of the plug with india-rubber solution .
Mr. J. B. Hannay .
r1'7 18 1 l0p *\#171 ; " .
f The " \#187 ; I \#187 ; *ed was A of an inch , and it soon cat the leather facings ; the second was T\gt ; \#165 ; \gt ; and it was found to last much longer ; and now working with a half-inch screw it has not required repacking for three months , although in constant use ' The dimensions of the apparatus as used are as follows Length of horizontal tube , 24 inches ; height of vertical branches , 8 inches ; caps finches long by If inches diameter ; screw , \#163 ; inch ; external diameter of tube , I3 inches ; internal diameter , f inch ; length of manometers , -s- inches to 26 inches ; external diameter , \#163 ; inch to S inch ; internal diameter , from ^ \#153 ; ch up to TV inch .
Small bath , 5 inches high by 4 inches diameter ; larger bath , 7 inches by 6 inches ; external cover 13 inches high by 9 inches diameter .
In each of the baths and in the cover two vertical slits were cut and fitted with mica windows , and a light placed behind allowed an observer to see clearly what occurred 1 he measurements are given in English standards , as engineers who construct such apparatus always use that method of measurement .
As Amagat has shown that hydrogen is the only gas which follows Boyle s law at high pressure , that gas was always used as the mano-metnc substance , and was carefully purified and dried before use .
The drying was done by passing it through five ( J-tubes with pumice-stone and strong sulphuric acid , and then through two ( J-tubes with phosphoric anhydride .
The manometers used were always 04 millim. m internal diameter , as narrower manometers always gave higher readings .
In determining the pressure of alcohol at its critical temperature , the difference of pressure indicated by different manometers puzzled me at first , especially as there was no difference in temperature , but upon determining the diameters of the manometer tubes it was found that the highest pressures were registered by the smallest bores .
The pressure of alcohol at its critical point as registered by the different manometers was as follows :\#151 ; Temperature ( theory ) .
232'14 232 -07 232 -12 231 -99 Diameter in millims. 0 -142 0-272 A 0 -480 B 0 -628 Pressure in atmospheres .
These numbers are the means of thirty measurements in each case .
The manometers A and B were used in most of the first portion of the work , but as they both broke subsequently , they were replaced by two otheis , A and B , and these again by A !
' and B^ .
Wdien a pressure of On the Limit of the Liquid State .
29^ over 300 atmospheres is required , these wide manometers are very apt to hurst , so that for high pressures a narrower tube must be used .
The first work undertaken was to ascertain without doubt the critical point of pure anhydrous ethyl alcohol , and this was done as follows \#151 ; The alcohol sold as absolute by the makers was fractionated , and the middle third taken .
This was placed in a retort with freshly burnt lime , and an inverted condenser adapted to it .
After it had been boiling for some time , the end of the condepser was fitted with a drying tube of calcium chloride , to prevent moisture from entering .
The cohobation was continued for a week , and the alcohol then distilled off .
The first fifth was rejected , as was also the last .
The receiver was a small flat-bottomed flask , which is shown fitted up for use ( after it was filled with alcohol ) in the front of the drawing .
It was arranged as a wash-bottle , having the tube for the entrance of air connected with a small vitriol tower , and an india-rubber ball , fitted with valves , to apply pressure .
The exit tube was adapted to the experimental tube by a piece of india-rubber tubing , through which was forced a piece of capillary tubing .
When the apparatus was to be used in experiment the arrangements were made as follows The cap with the pressure-screw was first fitted on next the experimental tube , with its point sealed up , and the whole filled up to the top of the manometer branches with mercury .
The manometers were now placed in position and screwed tight .
The apparatus was then tilted so as to raise the point of the experimental tube , keeping it , however , above the level of the lower ends of the manometers , and the point then broken off .
If the point were below the level of the manometers some gas might escape .
The wash-bottle arrangement is then fitted to the experimental tube and the ball compressed .
Alcohol is driven over and escapes by the capillary tube , and this is continued till the inside of the tube has been well washed and all impurities removed .
The capillary tube is then withdrawn , when the small puncture in the india-rubber at once closes itself .
The screw of the pressure apparatus is then retreated , and when sufficient alcohol is made to enter the apparatus , the joint is undone and the whole wash-bottle arrangement placed under a bell-jar over oil of vitriol for use another time , the india-rubber tube being clipped .
The screw is then further retreated to leave a small air-space over the alcohol , which is then boiled and the point sealed , and the tube placed in the air-bath .
A mercury regulator , such as I have described elsewhere , was sometimes used when the temperature was required to be constant for long .
The following tables contain some of the series of observations on alcohol , and are given to show the numbers obtained when the work is done with every care .
The alcohol used was different in each case , so that slight variations in the averages may be due to differences in the 300 Mr. J. B. Hanna liquid used .
The numbers for pressures are arbitrary scale readings , and are reduced to actual pressures at the end of the tables .
Table No. I. Critical Temperatures and Pressures of Alcohol .
T and T ' are the two thermometers in the bath , one on each side of the experimental tube .
t and t ' two thermometers for correction of T and T ' .
P and P ' are the readings of the two manometers A and B. t " and t " ' two thermometers for correction of manometers .
T. T ' .
t. t ' .
P. P ' .
t " .
233 234 72 84 187-2 224-5 16 -5 17 232 233 75 89 187 -8 224-7 17 16 -8 232 234 74 85 185 -9 223 -4 18 17-5 235 235 75 80 188 -5 225 17 16 -9 230 230 82 75 185 -9 223 -3 18 18 -2 232 -5 235 -5 90 85 192 -4 225 -7 15 15 233 234 -5 88 92 191 -6 225 -5 15 15 -5 232 230 -5 85 87 186 -2 224 -3 15 15 232 230 90 85 187 -8 224-7 16 16 232 230 92 85 187 -8 224 -8 16-5 16-2 231 234 85 67 187 -8 224-7 16 -2 16 -5 235 234 -5 69 85 181 5 223 -5 17 -5 17 234 234 92 72 187-2 224-5 17 -3 17 232 -5 232 5 90 84 188 -5 224 -8 18 -2 18 236 237 86 90 188 224 -7 16 -5 16 -2 233 232 80 84 186 -9 224 15 -5 15 -3 235 236 92 78 187 -2 224-5 14 -5 15 231 232 84 90 185 -2 224 -1 15 -5 15 -8 230 231 82 84 193 -6 225 -9 16 -8 16-5 230 -5 232 94 82 184 225 -8 16 -2 16 230 -8 229 -5 85 80 187 224 -5 16 16 231 -5 234 82 85 186 -1 224 -3 15 15 233 -5 232 -5 90 .87 185 -4 224 -2 17 16 -8 234 234 90 89 196 -7 226 -5 17 -3 17 -2 233 230 85 85 196 -3 226 -4 16 -8 16-6 233 234 62 72 189 -4 225 20 -5 20 233 232 65 60 389-6 225 18 -7 19 232 231 86 68 188 -5 224 -8 19 -2 19 229 231 72 65 186 -8 224 -4 18 -5 18 -5 230 231 -5 60 65 186-6 224 -4 17 -5 17 229 230 -5 68 65 186 224 -2 16 16 231 230 69 69 187-2 224 -5 15 -5 35 On the Limit of the Liquid State .
301 T. 233 231 230 229 230 231 231 -5 232 T ' .
t. t ' .
P. P ' .
t " .
t " ' .
231 230 231 5 229 230 '5 231 231 5 74 181 5 223 -5 14 -9 15-5 4 o 72 73 181 7 223 6 16 -2 17 -2 4 Ll 70 84 185 '2 224 -3 15 15 -5 09 80 ... 184 -8 224 -2 16 -2 17 -2 *j \lt ; \#151 ; 87 85 187 '8 224 -8 18 -5 16-7 75 82 187 7 225 17 -2 18-5 85 90 185 -2 224 -8 19 16 00 80 80 182 -7 224 2 15 15-9 232 5 Portion of thermometer scale exposed , 80 ' to 232 ' .
Average T 231 971 Qorrected average 235 ' '67 .
" T ' 232 ' '21 / Height of mercury in manometer above tube P =0 91 atmos .
" " \#187 ; " P , =1 '14 " Average P jpressure in atmos ' Corrected jg''*4}67-56 atmos .
Probable error of mean temperature 0 19 .
1 } " pressure 0'13 .
Table II .
Critical Point of Alcohol continued .
T. T ' .
t. Manometers A and B\ t ' .
P. P ' .
r. f ' " .
231 -5 232 82 78 288 143 19 18 -5 232 231 80 82 287 140 20 19 *5 233 233 78 75 289 145 19 20 233 -2 233 -5 79 76 288 144 20 20 233 233 80 85 287 140 21 21 232 232 -5 81 79 290 148 20 20 233 233 82 81 287 140 20 19 231 229 79 84 290 147 19 18 -5 231 229 80 72 288 143 16 17 231 229 79 78 2S7 140 16 16 -5 230 230 78 85 287 140 17-5 16 231 -5 231 '82 92 286-5 140 17 17 231 231 -5 81 88 287 140 -5 16 -5 16.-5 231 -5 232 85 77 288 144 15 15 -5 231 -7 231 -7 84 80 287 141 16 16 232 232 -5 75 81 289 143 17 16 5 232 232 78 78 290 148 18 -5 18 232 -2 232 79 79 290 149 18-5 18 VOL. XXXIII .
Y Mr. J. B. Hannay .
T. T ' .
t. t ' .
P. P ' .
t " .
t'". .
231 -8 231 -5 80 85 288 147 18 18 231 -8 232 82 79 287 142 17 17 231 -5 232 -2 81 80 288 144 16 16 232 -2 232 -2 80 80 288 -5 145 16 -5 16 232 232 80 81 287 142 16 -5 17 232 2 232 -5 81 80 287 140 17 17 232 232 85 81 285 138 16 -5 18 232 -2 232 -2 76 80 287 142 15 -5 16 232 232 77 78 285 .
138 15 15 231 -8 232 -2 75 76 288 145 14 -5 14 -5 232 232 80 79 287 142 15 -5 15 232 232 78 80 287 148 15 15 Portion of thermometer exposed , 82 ' to 232 ' .
Average of T 231'-87 ) n _ T ' 231'-85 J Uorrected mean temperature , 235'-43 .
Height of mercury in manometer above experimental tube__ Average P 287*66 , f 66-qn t " P ' 142*17 J Corrected mean Pressure | j. " * " j 66-88 .
Probable error of mean temperature 0o,16 .
\#187 ; \#187 ; pressure 0"09 .
Table III .
Critical Temperature and Pressure of Alcohol .
Manometers A " and B " .
T. T\ t. f. P. P ' .
t " .
t " 232 232 -5 82 78 212 122 -8 16 18 230 232 75 82 212 -5 123 -5 17 17 232 232 *2 78 78 213 125 19 18 232 -2 232 -5 80 80 211 121 20 19 232 *2 232 -5 82 81 210 -5 120 20 -5 20 -5 232 232 -5 84 83 211 121 20 20 233 232 80 80 211 -5 122 19 19 232 232 -2 80 79 212 123 18 -5 18 231 -5 232 -8 80 76 212 -2 123 -4 18 -5 17 232 231 -5 80 81 212 1 123 -2 18 -6 19 232 -5 232 85 83 211 -9 123 19 19 231 -5 231 -8 82 82 212 121 18 17 232 -5 231 -5 89 87 212 -3 123-7 17-5 16 232 232 72 75 211 -6 121 -8 16 -5 16 -5 231 *5 232 -5 74 75 212 -2 122 -7 14 -2 14 T. 230 231*5 232 -5 232 -5 231 -5 T ' .
231 -8 232 -5 On the Limit of the Liquid State .
303 t. 70 t ' .
76 P. 212 -5 P ' .
124 t " .
14 -5 t ' " .
15 77 213 125 16 16 -5 4 O *70 79 211 120 -5 16 16 4 J 77 80 210 -5 120 15 -5 15 -5 4 4 *7K 80 211 -8 121 15 -5 16 4 O qo 80 211 -7 121 15 -5 16 Ou S3 82 211 -3 123 16 15 oo 84 HI 83 212 1 123 16 16 80 ... .
212 0 122 -8 16 16 O A QK 86 212 -2 123 15 -5 16 -5 OO 80 80 70 80 211 8 122 15-5 15 -5 81 212 -5 122 -9 17 17 -5 78 212 -6 123 5 19 18 4 O 75 80 212 -2 123 15 -5 15 4 O 78 80 211 121 15 15 4 O 75 74 212 -2 122 -8 16 15 4 O 78 78 212 -5 123 -9 18 17 Portion of thermometer exposed , 80 ' to 232 ' .
Average T 231c,88 | Qorrected mean temperature 235''39 .
T'231'-77j Height of mercury in manometers over experimental tube\#151 ; Average P 211-92 1 Qorrec^e(j mean pressure { ^ qo 1 36'78 atmos .
p ' 122-48 / I- 66-82 J Probable error of mean temperature 0o,10 .
J } " pressure 0-06 .
We see from the foregoing tables that the mean of over a hundred experiments gives a critical point for alcohol of 235 47 under a pressure of 67'07 atmospheres .
The reason why so many experiments were done was because the first two or three series did not agree well , and it was only after some experience was gained at the work that good results were obtained .
I hare no doubt that by further refining of the methods better results would be obtained , but I do not think the above numbers would require material alteration .
Having now fixed the critical temperature and pressure of alcohol under its own vapour , the next work consisted in determining the critical temperature of the same liquid under greater pressure .
When any greater pressure than the critical is used , the tube is filled with a homogeneous fluid , the two states of a fluid being impossible under such a pressure .
The critical state cannot , therefore , be observed under such conditions , as all the phenomena by which the liquid state can be recognised are dependent upon the observation of a limiting surface having a certain t 2 304 Mr. J. B. Hannay .
contractile power , and such power cannot be observed except the liquid have a free surface\#151 ; that is , a surface bounded by another fluid with which it is not miscible .
It was found that all liquids such as water , hydrocarbons , ethers , \amp ; c. , however immiscible they may be at ordinary temperatures , mix freely or act upon one another long before the critical point of one of them is reached ; therefore liquids will not serve to furnish a free surface .
A gas , therefore , is the only substance which will bear any pressure without becoming miscible ; and if it is insoluble in the liquid ( and all liquids have some gases insoluble in them ) we are provided with a substance .to overlie the liquid which will allow of a limiting surface being seen at any pressure .
This was the method used .
A quantity of pure dry hydrogen was placed over the alcohol , and pressure applied , and the effect of rise of temperature observed .
It was seen that when the temperature rose to the critical point the line dividing the alcohol from the mixture of alcohol vapour and hydrogen became quickly indistinct , and was replaced by a broad mark , indicating a gradual change in the refractive index of the fluid when passing the place where the liquid surface had been , showing that diffusion was taking place .
On lowering the temperature before much diffusion had taken place , the point where the liquid surface had been became dim just as the temperature passed below the critical temperature and the sharp limiting surface became re-established .
The pressure was then increased by decreasing the volume of hydrogen , and the experiment repeated many times at the new pressure .
The temperature must be raised much more slowly for these observations than for the simple observation of the critical point with alcohol alone , as in the latter case the upper and lower portions of the fluid become of the same density , instantly obliterating the line of the meniscus ; but when hydrogen is above the alcohol the line remains although the alcohol be gaseous , until it is obliterated or broadened by diffusion .
Thus , on raising such a mixture to the critical temperature , it is necessary to keep the temperature steady , to ascertain whether or not diffusion will take place .
If a rise of temperature is going on , the thermometers will register a higher temperature than that at whi h diffusion began .
In this way the following series of observations were carried out:\#151 ; Table IY .
Alcohol with Hydrogen .
Manometers A " and C. T. T ' .
t. t\ P. P ' .
r. 230 232 80 75 220 288 17 17 229 229 75 73 221 5 291 20 18 231 230 -5 78 70 219 287 21 19 On the Limit of the Liquid State .
T. 234-5 234 232 233 232 232 -2 231 -5 231 -5 233 231 -5 232 232 233 -5 232 232 -5 231 231 *5 232 231 -5 229 229 -5 230 231 231 230 229 230 228 230 T ' .
229 5 228 -5 231 232 -5 232 -5 232 -5 232 5 232 5 OQO t. 77 t ' .
72 . . .
P. 219 P ' .
287 t " .
18 r. 16 4 / 80 219 -5 287 19 15 / u 7Q 82 219 -5 287 17 15 4 O oo 80 217 285 -5 16 16 78 80 220 -5 288 15 15 4 O 7Q 74 219 -8 287 14 14 4 V 62 fiQ 74 215 5 282 -5 15 5 15 69 218 -8 286 16 16 74 74 220 287 -5 17 17 232 -2 231 5 233 5 232 -5 233 232 231 5 232 -5 233 232 -5 228 -5 229 -5 229 232 231 230 230 229 228 220 79 78 219-5 288 16 15 4 u 7Q 79 217-5 286 17 17 qk 82 217-5 285 -5 15 -5 16 o o 87 85 212 275 16 16 O 4 89 83 219 285 17 18 80 76 78 218 284 17 17 76 219 -5 287 15 15 4 U 76 74 219-8 287 16 15 4 v 75 80 2211 288 -2 17 17 4 O 75 78 223 283 19 18 4 O 79 80 218-8 286 -5 17 17 4 i/ 80 80 77 78 80 218 -9 285 -5 16 16 80 219 -2 287 16 16 78 80 218-6 218 -5 286 -3 286 -2 16 17 16 17 4 O 79 79 219 -2 286 -9 15 -5 15 5 4 76 77 219 -3 287 18 17 4 v 84 80 80 82 218 -8 286 -5 17 17 80 218-7 286 -5 17 16 79 218-8 286 -6 16 16 Portion of thermometer scale exposed 82 to 232 .
Average T 231'-301 Corrected mean 234''78 .
T ' 231'16 J f p =0-93 atmos .
Height of mercury in manometers over tube \lt ; p/ __Q.gy ^ Average P ^'^Jpressure in atmog Corrected|\#174 ; ^}82'53 atmos .
Probable error of mean temperature 0 *17 .
" " pressure 0*08 .
From this it would appear that the critical temperature , or , at least , that temperature at which the meniscus disappears , is lowered slightly Mr. J. B. Iiannay .
thJ pointdr'gen ^ Furtlier exPeriments were conducted to settle Table V. Alcohol with Hydrogen , Thermometer D. T. 234 T ' .
228 t. 78 t ' .
77 P. 290 t. 232 230 82 80 291 AO 1 \#163 ; 232 232 75 73 290 AO 1 8 231 -5 231 80 79 291 AO 229 231 5 82 84 291 AO 1 7 230 231 77 78 292 A / IQ 231 232 72 75 288 AO 1 *7 232 231 79 78 290 A / 15 -5 l 230 230 81 81 289 231 231 82 82 ... 288 AO 15 -5 1 7 231 231 80 79 288 229 229 75 75 292 If !
228 -5 228 80 82 291 AO 17 Average T 230o,261 ~ T ' 230o,84 J ^orrec*e\lt ; i mean 2340,1'4 Average P 29007 .
Corrected pressure 183'7 atmos .
Probable error of mean temperature 0o,34 .
\#187 ; j ?
pressure 0'19 .
Table VI .
Alcohol and Hydrogen .
Manometers A " and B " .
T. 232 T ' .
231 t. 68 t ' .
72 P. 222 P\ 1 71 r. i ^ r. 1 K 231 -5 232 72 72 222 *4 170 -2 172 171 AO AO 1 6 232 -5 233 84 80 223 AO 1 7 AO 1 7 233 232 -5 92 90 222 A t lfi A ( 16 231 231 92 92 221 *8 JL I JL 169 *5 171 AO AO 18 229 230 90 90 ... 222 AO 17 5 15 -5 16 AO 17-5 15 5 16 234 235 88 90 222 -2 171 -2 171 -4 231 232 75 74 223 On the Limit of the Liquid .
230 -5 230 -5 231 232 t. 232 -5 80 233 -5 77 232 231 -5 231 5 232 231 *5 231 231 232 232 233 234 233 -5 234 232 -5 232 -5 232 232 232 232 232 232 232 232 -5 232 2 232 *7 232 -8 232 .2 t ' .
p. P ' .
t t"\ 222 -2 170 *3 15 15*5 222 *2 170 -2 15 15 222 -4 170 *2 16 *5 16 222 170 *5 17 17 222 *2 170 18 18 *5 221 *5 170 *2 18 *5 18 5 221 *8 169 *4 19 19 221 *9 169 *8 16 17 223 170 16 16 222 *5 171 *3 15 15 *5 222 7 170 -6 16 -5 222 -8 170 -9 17 222 -4 171 16 222 -8 170 5 16 171 15 171 15 171 -8 15 -5 170 7 15 5 170 17 231 *5 232 77 78 222*8 170 -5 17 17 231 7 231 5 77 75 222-4 170 *9 16 17 231 *5 231 80 79 222-5 170 -4 15 -5 16 232 232 81 .
82 223 170 -6 15 15 233 233 80 80 223 1 170 -8 16-5 15 -5 233 5 233 *5 79 79 223 171 9 17 16 231 231 *5 78 78 222-4 171 -7 18 17 -5 230 *5 230 -5 80 81 222-2 170 -5 18 18 230 -5 231 81 80 222 170 -2 17 17 232 232 85 85 222-7 170 -8 18 17-5 232 5 233 80 80 222-5 170 -5 17 17 232 231 5 75 78 222-6 171 -3 16 17 234 233 79 78 222-8 171 -5 18 18 231 *5 232 80 78 222-4 172 17 -5 17 Average T 232'*161 Corrected mean 235'-68 ?
) T ' 232'*09 J Average P 222-441 Corrected meail 12272 atmos .
" P ' 17074 J Probable error of mean temperature 0'*09 .
, , " pressure 0*10Mr .
J. B. Hannay .
Table VII .
Alcohol with Hydrogen .
T. T ' .
t. t ' .
P. P ' .
t " .
t'". .
230 231 75 76 226 191 16 16 230 '5 231 80 82 227 192 17 17 231 '5 232 82 80 226 '6 191 -4 15 15 232 232 92 90 226 -4 191 -6 15 15 233 231 85 85 226 -4 191 -2 15 15 232 5 233 90 92 226 '5 191 15 15 5 232 231 82 80 226 -2 192 -2 15 5 16 229 230 '5 78 77 227 191 *8 16 16 230 '5 230 76 76 226 -8 192 16 16 '5231 '5 231 '5 77 77 227 191 -4 15 '5 15 -5 232 230 82 82 226 -2 191 16 16 231 '8 232 80 82 226 191 2 17 17 230 -4 232 73 72 226 '4 191 -4 18 17 5 232 -5 231 '3 72 75 ... 226 -6 191 3 17 17 231 232 80 80 226 '9 192 1 15 15 Average T 231 ' *351 " T'231'-61J Corrected mean 235o,04 .
Average P 226*581 " P'191'55/ Corrected mean 178'80 atmos .
Probable error of mean temperature 0''18 .
\#187 ; " pressure 0'07 .
Prom these tables we see clearly that the critical temperature is not materially altered by a very large increase of pressure ; in fact , in the last case the pressure is nearly three times as great , and yet we have only a lowering of the critical temperature by about one degree .
It was found , however , that as the pressure was increased the solubility of the hydrogen in the alcohol also increased , so that at high pressures a very considerable lowering of the critical temperature takes place .
When the two fluids have thoroughly mixed at a temperature over the critical point , the passage of the mixture through the critical temperature downwards is not attended with immediate liquefaction ; in fact , this does not take place till a temperature 10 ' lower is reached , the hydrogen preventing the alcohol from assuming the liquid condition .
At a pressure of 250 atmospheres the meniscus disappeared , or rather became broad at 225 ' ; but diffusion did not take place completely ; the surface seemed to be destroyed , but the action did not go deeper , while at 300 atmospheres the meniscus was lost at 220 ' .
This is plainly owing to the action of the compressed hydrogen , and could we have a gas quite insoluble in the liquid , this lowering would not take On the fjimit of th Liquid State .
309 place .
It seems clear then that the temperature at which the peima-nent surface of a fluid ( which constitutes liquidness ) disappears is not altered by increase of pressure , and this is equivalent to saying that the critical point is the termination of an isothermal line which marks the limit of the liquid state .
It next remained to be seen whether any other insoluble gas would act in the same manner as hydrogen .
It must be remembered that hydrogen is furthest removed from the liquid condition and the least dense body known , and the nearer the density of the superincumbent gas approaches to the density of the liquid , the greater effect will it have upon the critical temperature .
To test this , a quantity of nitrogen was placed over alcohol , and the experiment conducted similarly to those with hydrogen .
Table VIII .
Alcohol with Nitrogen .
Manometer C. T. T ' .
t. f. P. t " 231 5 232 78 79 274 16 232 231 5 77 77 267 16 231 -5 231 5 78 78 262 15 5 232 5 232 80 81 262 15 232 231 -5 82 80 270 15 231 5 232 80 80 270 15 232 -5 232 81 81 276 16 230 231 80 82 276 17 229 230 80 78 276 18 231 -5 232 75 74 278 17 231 -5 231 -5 76 78 277 16 232 *5 232 78 80 273 15 Average T 231'-501 Corrected mean temperature 235'T1 .
" T ' 231'*58 / Average P 27P75 .
Corrected mean pressure 82'35 atmos .
Probable error of mean temperature 0''22 .
" " pressure 0T9 .
These numbers plainly show that the meniscus of alcohol disappears at the same temperature , whether under tne pressure of its own vapour or at a pressure of eighty atmospheres with nitrogen , affording further proof that the liquid state terminates at the critical temperature .
Another method still remained to be tried , that of measuring the capillary height of a liquid under various pressures and temperatures .
The method used at first was to fix a small piece of capillary tubing Mr. J. B. Hannay .
into the interior of the working tube by melting a small piece of sili cate of soda , and causing the tube to adhere ; but this was found to be disadvantageous , as the tube almost invariably burst where the silicate was adhering The method latterly used woe simply to make an obtuse-angled bend on the tube , so that the piece of capillary tube could not pass beyond this , but became wedged .
The capillary heieht was measured by a cathetometer in the usual manner .
The following table gives the results:\#151 ; Table IX .
Capillary Height of Alcohol .
Manometers A " and B " .
T. T ' .
t. t ' .
10-5 10-5 . . . .
19 19 16 16 34 34 16 15 *5 54 53 -8 15 -5 \gt ; 15 *5 67 67 15 15 81 80-5 16 16 90 90-2 17 17 108 108 16 16 122 122 18 17 127 127 18 18 137 136*5 20*5 20 145 145 23 23 155 154 *2 27 27 166 5 166 *5 35 34 172 173 42 40 186 186 53 50 192 193 60 60 195 195 60 61 202 202 65 66 205 205*5 70 72 209 210 70 70 214 214 72 75 .220 220*5 75 75 225 224*5 78 78 229 230*5 79 80 .
112 137 *5 165 185 189 *5 27 199 52 204 83 206 90 208 *5 107 210 115 211 118 * 212 123 t " .
16 5 16 *5 16-5 ... 16 15-5 ... 15 Cap. Ht .
564 557 540 513 487 468 450 428 402 388 374 358 334 300 264 238 211 194 179 161 141 127 84 49 8 From this table curves Hos .
I , fig. 2 , and IX , fig. 3 , have been drawn , while Table X gives the corrected values for capillary height under the pressure of alcohol vapour .
The numbers expressing the pressure of the vapour at temperatures below those at which the manometers used began to register are taken from Regnault 's observations .
On the Limit of tlio Liquid State .
Fig. 2 .
Cap. Ht .
T. Mr. J. B. Hannay .
Table X. Curves I , Fig. 2 , and IX , Fig. 3 .
P. in atmos .
Cap. Ht .
235 'i ... .
66 *7 0 40 99 198 269 321 365 401 436 467 501 530 555 579 230 64-9 220 58-7 200 37-5 180 18-1 160 10-9 140 7-4 120 millims. 3219 100 1694 80 812 60 350 40 134 20 44 0 13 As examples have now been given of the numbers obtained in all the different methods employed , and the probable error calculated , the tables of original numbers will be omitted and only the corrected values given , and where several observations have been made the probable error will be given .
The following table gives the capillary heights of alcohol at various temperatures under compressed hydrogen , giving a much higher pressure than the vapour alone .
Table XI .
Curves II , Fig. 2 , and X , Fig. 3 .
Alcohol Capillarity under Pressure .
T. P in atmos .
Cap. Ht .
230 -3 163-5 o 220 125-0 63 200 106-0 159 180 79-0 234 160 53-5 286 140 37-0 333 120 27-8 372 100 24-3 409 80 22 1 442 60 18-3 473 40 16-7 503 20 15-0 531 0 14-2 557 On the Limit of the Liquul State .
O 1 o OlO Table XII .
Curves III , Fig- 2 , and XI , Fig. 3 .
Capillary Height of Alcohol under High Pressure .
T. P in atmos .
Cap. Ht .
4)01 236-8 0 Ala1* O . . .
99D 224 -9 26 AlAj . . .
on 182-2 129 ZUU . . .
1 8A 145-3 207 iov ... 1 AA 136 264 1DU . . .
1 AA 128 1 312 . . .
1 9A 114 354 1 \amp ; \j . . . .
1 AA 100-1 391 lUu . . .
on 86-8 428 oU . . .
AA 79 458 DU . . .
AA 69 1 489 . . .
9A 66 517 L\J . . .
0 64-2 544 From these tables we see that the capillary height of the liquid is lowered by a gas under pressure impinging on its surface .
Thus at ( 36'7 atmospheres ( the critical pressure of alcohol ) , the capillary height falls to zero at 235'4 , at 163'5 atmospheres of pressure zero is reached at 230o,3 ; while at 236'8 atmospheres capillarity disappears at 224 '6 .
It is curious to note that , although the capillary action had ceased at these temperatures , the liquids had not assumed the gaseous state , as Tables IV , V , VI , VII , and VIII show that in no case up to a pressure of 183 atmospheres did the alcohol diffuse into the hydrogen at a temperature below 234 ' .
As capillarity is entirely a surface phenomenon , the surface tension of a liquid seems to be weakened by the impinging of a gas under pressure upon its surface , and this we might expect to be the case , as we can imagine a constant disturbance of the surface of the liquid , owing to the high velocity of the hydrogen molecules striking it ; whereas , not being soluble to any extent , few hydrogen molecules penetrate to disturb the liquidness of the interior .
It would thus appear that , under such conditions , capillarity is not a true measure ol the liquidness or cohesion of a fluid , and were the pressure high enough the surface of a liquid might be made to disappear , while its interior was in a truly liquid condition .
This question can be most readily settled by passing a liquid , whose surface tension has thus been caused to disappear , through a capillary tube , and .observing whether increase of temperature diminishes the time of flow , for the resistance of a fluid is decreased by increase of temperature , while that of a gas is increased .
The experimental realisation of such a test is difficult , but apparatus is being at present constructed for the trial .
314 -Mr .
J. B. Hammy .
In drawing out the above tables ( which may exhibit slight irregularities ) reconrse was made to a large number of quite separate observations , as full series are often difficult to obtain , owing to some failure in the apparatus when it lias been in use for some time at hio-h pressures .
Hitherto only one liquid\#151 ; alcohol\#151 ; had been used in these experiments , so it was determined to try the same experiments with other liquids , and those chosen were carbon disulphide , carbon tetrachloride , and methyl alcohol .
The carbon disulphide was digested over sodium for some time and distilled off pure quick-lime .
This gives a liquid having no offensive odour and quite colourless .
It was distilled into an apparatus similar to that used for alcohol , and preserved under a bell-jar over oil of vitriol .
Four sets of observations were done in order to determine the critical temperature and pressure accurately ; in all 163 experiments .
The results are as follows :\#151 ; Mean temperature corrected ... ... 277'*68 .
Probable error of mean ... ... . .
0O-16 .
Mean pressure corrected ... ... ... 78*14 atmos .
Probable error ... ... ... ... . .
0*07 The apparatus was now arranged so that the critical temperature could be observed under pressure with hydrogen with the following results:\#151 ; Mean temperature corrected ... ... 274 ' *93 .
Probable error ... ... ... ... ... 0'*09 .
Mean pressure corrected ... ... . .
171*54 atmos .
Probable error ... ... ... ... ... 0*07 " For these numbers sixty-two experiments were done .
Nitrogen was then substituted for the hydrogen , and the experiments conducted as before .
Forty-one determinations were made to obtain the following means:\#151 ; Mean temperature corrected ... ... 273'*12 .
Probable error ... ... ... ... ... 0'*19 .
Mean pressure corrected ... ... . .
141*45 atmos .
Probable error ... ... ... ... ... .
0*16 Here we see that while the pressure of the nitrogen on the carbon disulphide is much lower , the temperature at which the meniscus disappear# is also lower .
This is likely owing to the greater solubility of the nitrogen in the liquid , as the density makes it approach much nearer to the density of the disulphide than hydrogen .
A third series was conducted with hydrogen , using , however , a much lower pressure:\#151 ; On the Limit of the Liquid State .
Mean temperature corrected ... ... 277'"55 .
Probable error ... ... ... ... ... 0o,14 .
Mean pressure corrected ... ... ... 95-86 atmos .
Probable error ... ... ... ... ... 0'09 " Here we see only a very faint lowering of the critical point , only 0'2 of a degree , although the pressure has been increased twenty atmospheres ?
The capillary height of the disulphide under various conditions was next determined .
The numbers gave as follows:\#151 ; Table XIII .
Curves IY , Fig. 2 , and XII , Fig. 3 .
Capillarity of Carbon Disulphide .
T. P in atmos .
Cap. 277 78-1 0 260 581 45 940 43-8 74 220 33 7 114 200 25'5 155 180 19 6 197 160 14-8 236 140 10-9 277 120 7 *6 316 100 4 *3 359 80 2-8 397 60 1-6 437 40 0-8 476 20 517 0 558 Table XIY .
Curves Y , Fig. 2 , and XIY , Fig. 3 .
Capillarity of Carbon Disulphide with Hydrogen .
T. P in atmos .
Cap. Ht .
260 172-1 0 260 167 2 9 240 160 46 220 153 -5 84 200 146-3 122 180 138-2 164 Mr. J. B. Hannay .
T. P in atmos .
Cap. Ht .
160 131 *5 ono 140 124-8 QA9 120 117-7 O \gt ; 7Q 100 110-9 80 ... 103 -8 qkq 60 96-5 oqj .
40 90*2 AAiL 20 83-2 0 78 ... .
515 Table XV .
Curves VI , Fig. 2 , and XIII , Fig. 3 .
Carbon Disulphide with Nitrogen .
T. P in atmos .
Cap. Ht .
273 131 o 260 240 96-1 fil 220 100 200 141 180 197 160 140 278 120 317 100 ... 358 80 305 60 436 40 33 3 478 20 51 7 0 . . .
28 1 555 On examining these tables , and the curves which graphically represent them , we see that here also the capillary action of the liquid is weakened by a gas impinging upon its surface , even at low temperatures , and again we see that the capillarity is reduced to zero before the liquid is really gaseous , showing that like alcohol the surface tension of carbon disulphide is destroyed by the activity of the molecules of the gas overlying it .
The curve for carbon disulphide and hydrogen , XIV , fig. 3 , should not be really a straight line , but a number of accidents happened during these experiments , and the curve is made up from many readings from different manometers , that for the On the Limit of the Liquid State .
middle part of the curve evidently reading a little low .
The whole result agrees well with what was shown from alcohol .
The next body examined was methyl alcohol , a sample of which was carefully purified in the same manner as the ethyl alcohol , until its boiling point was constant .
It was then distilled off quick-lime into the wash-bottle arrangement for use .
It gave as follows : Mean temperature corrected ... ... 232'*76 .
Probable error ... ... ... ... ... . . .
0o,21 .
Mean pressure corrected ... ... . .
72'85 atmos .
Probable error ... ... ... ... ... .
0T2 \#187 ; These means were taken from three series , each of thirty experiments .
The same experiments were then carried out with methyl alcohol and hydrogen and nitrogen , as in the case of ethyl alcohol and carbon disulphide , yielding the following results from twenty-two experiments:\#151 ; Mean temperature corrected ... ... 230o,14 .
Probable error ... ... ... ... ... 0 *09 .
Mean pressure corrected ... ... . .
128*60 atmos .
Probable error ... ... ... ... ... .
0*12 " Here we have , as before , a slight depression of the critical temperature .
Experiments were then tried with higher pressure .
Forty-seven gave the following means:\#151 ; Mean temperature corrected ... ... 227 *92 .
Probable error ... ... ... ... ... . . .
0 *10 .
Mean pressure corrected ... ... . .
191*40 atoms .
Probable error ... ... ... ... ... .
0*07 " A still higher pressure was then applied .
The means of eighteen experiments were\#151 ; Mean temperature corrected ... ... 226 *82 .
Probable error ... ... ... ... ... .
0'*26 .
Mean pressure corrected ... ... . .
262*00 atmos .
Probable error .
... ... ... . .
0*09 " We have a further depression of the critical point , and as the point was determined with great difficulty we have an increase of the probable error .
However , the result confirms the other experiments .
The same mode of experiment was then carried out with methyl alcohol and hydrogen and nitrogen as with ethyl alcohol , with the following results .
VOL. xxxiii .
z Mr. J. B. Hannay .
Table XVI .
Curves VII , Fig. 2 , and XV , Fig. 3 .
Capillarity of Methyl Alcohol .
T. P. Cap. Ht .
232 7 72-8 ... .
o 230 ... . .
64-8 ... .
5 220 ... ... . .
51 2 ... ... . .
52 200 ... ... .
34-8 ... ... . .
133 180 ... .
23*2 ... .
202 160 16-4 257 140 11-9 309 120 7 1 359 100 ... ... . .
4-2 ... ... .
402 80 ... ... . .
1-9 ... ... . .
441 60 477 40 513 20 545 0 ... . .
577 Table XVII .
Curves VIII , Fig. 2 , and XVI , Fig. 3 .
Capillarity of Methyl Alcohol under Pressure with Hydrogen .
T. P. Cap. Ht .
224 123 4 0 220 119-7 27 200 104-2 106 180 90-5 179 160 79-2 ... ... . .
237 140 69 1 287 120 613 336 100 55-0 283 80 48-1 423 60 42-6 458 .40 38*4 493 20 33 8 527 0 29-9 ... ... . .
557 In these experiments we see again that increase of pressure never increases the liquidness of the fluid , and never enables it to remain liquid at a temperature above the critical point .
On the Limit of the Liquid State .
319 The last liquid examined was carbon .
tetrachloride , and this was very carefully dried and purified by fractionation and distillation off quick-lime , the purified liquid being kept with the same precautions as were used in the other cases .
It was seen , however , that the tetrachloride acted upon the mercury , forming a white crystalline body which crystallised out as the liquid cooled .
It appeared to be mercuric chloride , as it dissolved in water ; but whether the crystals were pure mercuric chloride or a compound of that body with some other chloride of carbon there was not sufficient obtained to determine .
The critical temperature and pressure were determined with twenty different samples , using one quantity only for two or three readings , and the following numbers were obtained :\#151 ; Mean temperature corrected ... ... .
282'*51 .
Probable error ... ... * ... ... ... 0'*38 .
Mean pressure corrected ... ... ... 57*57 atmos .
Probable error ... ... ... ... ... .
0*14 " On attempting to obtain the critical temperature under pressure it was found that the hydrogen at once dissolved under pressure , and not only dissolved , but formed a compound with the tetrachloride .
Some curious observations were made on the relation of pressure to chemical combination with this mixture .
It was found that for a given temperature and pressure only a certain amount of combination would take place , leaving the excess of hydrogen overlying the tetrachloride quite free .
'If then more pressure were applied , the manometer would jump up say five atmospheres , and then gradually fall about four or four and a half atmospheres , and again become stable , and this would take place each time , a large portion of hydrogen disappearing for a small permanent rise of pressure .
A sudden rise of temperature had somewhat the same effect , but as the temperature could not be varied so suddenly the effect was not so obvious .
Several bodies were formed by the action of the hydrogen , the action being capable of being pushed so far as to form chloroform .
Nitrogen was used as a pressure substance , and it answered well .
The following numbers were obtained from twenty-seven experiments : Mean temperature corrected ... . .
277 '56 .
Probable error ... ... ... ... . .
0'*29 .
Mean pressure corrected ... ... ... 142*88 atmos .
Probable error ... ... ... ..* . . .
0*13 " Prom this it will be seen that by increasing the pressure to nearly three times the normal , a fall of five degrees in the critical temperature has taken place .
This was no doubt due to the surface tension being 320 Mr. J. B. Hannay .
destroyed by the action of the gaseous molecules , so another series was tried at a lower pressure .
Forty-two experiments were done Mean temperature corrected ... ... 282'-50 .
Probable error ... ... ... ... ... 0o,12 .
Mean pressure corrected ... ... ... ... 83-18 atmos .
Probable error ... ... ... ... ... ... . .
0'38 Here we see that a large increase of pressure has not altered the critical temperature at all , as was before seen to be the case with alcohol .
A series of capillary measurements was commenced with this liquid , but a series of accidents interfered with the work , and only twelve reliable readings were obtained , and no time has been at my disposal since to finish the work ; but sufficient evidence of the course of capillary action has already been gained from the other liquids to draw conclusions as to the liquid state and its limit .
Three curves have been drawn to show the depression of the critical temperature with increase of pressure , and these lines have been continued down the curve of vapour pressure to show the break at the critical point .
This will be clearly shown in curves Nos. XVII , XVIII , and XIX , fig. 4 .
The consideration of these results yields a novel mode of looking at the states of matter which I have illustrated in fig. 5 .
From this it appears we might classify matter under four states ; first , the gaseous which exists from the highest temperatures down to an isothermal passing through the critical point and depending entirely upon temperature or molecular velocity ; second , the vaporous , bounded upon the upper side by the gaseous state and on the lower by absolute zero , and depending entirely upon the length of the mean free path , because shortening of the mean free path alters the state ; third , the liquid bounded upon the upper side by the gaseous state , and on the lower by the solid or absolute zero ; fourth , the solid On the Limit of the Liquid State .
Figk 5 .
whose condition is also determined by both pressure and temperature .
The gaseous state is the only one which is not affected by pressure alone , or in which the molecular velocity is so high that the collisions cause a rebound of sufficient energy to prevent grouping .
Another distinction between the gaseous and vaporous states is that the former is capable of acting as a solvent of solids , whereas the latter is not .
The two conclusions arrived at from this work are\#151 ; 1st .
The liquid state has a limit which is an isothermal passing through the critical point .
2nd .
The vaporous state can be clearly defined as a distinct state of matter .
To the original distinction between these two states given by Andrews\#151 ; namely , that of condensibility\#151 ; I have added another , that of solvent power .
A vapour over a liquid holding a coloured solid in solution is colourless , but on passing the critical temperature the whole becomes coloured .
In some cases , however , the solid is deposited and redissolved as the temperature rises , showing that the more perfectly gaseous the greater the solvent power .
Andrews 's distinction compels us to travel along an isotherm , mine requires high pressure ; both are thus arbitrary , requiring given conditions , but this is the case with many of the other distinctions used in science .
My thanks are due to my assistant , Mr. Ewing McConechy , for his assiduous aid during the above described investigation .
VOL. XXXIII .

