234 Studies of the Processes Operative in Solutions .
XXV.\#151 ; The Influence of Non-Electrolytes on Solubility .
The Nature of the Processes of Dissolution and Precipitation .
By H. E. Armstrong , F.R.S. , and J. Vargas Eyre* Ph. D. ( Received January 1 , \#151 ; Read February 13 , 1913 .
) The subject of " competitive solubility " has been dealt with in two previous communications of this series , Parts II and XI.* In both of these , emphasis was laid on the fact that electrolytes and non-electrolytes are alike active as precipitants of salts from aqueous solutions and it was implied that no " theory " of the process of dissolution which does not take this fact into account can be satisfactory .
The activity of the several precipitants studied was expressed .
in terms of their apparent molecular hydration\#151 ; that is to say , the method of treatment adopted involved the evaluation of the amount of water thrown out of action as solvent water by the precipitant : care was taken , however , to point out that this artifice was introduced merely on the ground of convenience and that the expression " apparent molecular hydration " was not to be interpreted literally .
It was contended , in fact , that the precipitant does not act merely by attracting water to itself but that the condition of the solvent water must also , in some way , be changed by the introduction of the precipitant , especially in the case of a neutral substance such as propylic alcohol .
This was a novel conception , based on views previously brought under the notice of the Society , in 1906 , in a communication on the " Origin of Osmotic Effects " and subsequently in Parts II and VI of these studies .
The conception has since been extended to other phenomena in Parts XIII , XIV , XV , XVIII , XIX and XX of these studies and a large body of evidence has been brought forward to show that solute and solvent are in all cases reciprocally modified .
The experiments now described were instituted mainly in order to ascertain the effects of non-electrolytes on slightly soluble salts , as the hypothesis introduced by Nernst , in 1889 , in explanation of the precipitation of one salt by another , was based more particularly on the results obtained with such salts\#151 ; the hypothesis , namely , that the solubility of a salt is depressed by the presence of another salt if the two have an ion in common .
According to the view put forward by Nernst , the concentration of the " undissociated " part of a salt in a solution saturated with the sarlt is constant , * 'Roy .
Soc. Proc. , J 1907 , A , vol. 79 , p. 564 ; 1910 , vol. 84 , p. 123 .
Studies of the Processes Operative in Solutions .
235 even in the presence of another salt , at each particular temperature ; it is also proportional to the product of the concentrations of the ions of the salt ( the ionic solubility product ) .
If a second salt be introduced which has an ion in common with the salt with which the solution is saturated , a certain proportion of the common ion will be derived from each salt and to maintain the constancy of the solubility product each salt must be re-formed to a certain extent , so that the salt with which the solution was saturated originally is in part precipitated .
As Stieglitz has pointed out , * the hypothesis is based upon premises which are not valid in the case of salts ; moreover , though the results obtained with a few slightly soluble salts may appear to substantiate it , the behaviour of soluble salts is in no way in accordance with such an assumption .
To take a case in point , when hydrogen chloride is added to a solution of sodium chloride , the latter is all but entirely displaced from solution .
An even more striking case is that observed by Etard , f who has shown that potassium chloride is insoluble in a solution saturated with potassium bromide and iodide .
The complete command of the water exercised , in the one case , by hydrogen chloride , in the other , by the two haloids is very remarkable .
Peculiarities such as these did not escape the notice of the older workers , as witness the following statements made by Graham in 1850 in his Bakerian lecture on Diffusion :\#151 ; " In the consideration of solubility attention is generally engrossed entirely by the quantity of salt dissolved .
But it is necessary to apprehend clearly another character of solutions , namely , the degree of force with which the salt is held in solution or the intensity of solvent attraction , f quite irrespective of quantity dissolved .
" In the solutions of two salts which are equally soluble in point of quantity , the intensity of the attraction between the salt and water may be very different .
* " Besides being said to be small or great , the solubility of a substance has therefore to be described as weak or strong .
" The difficulty of evaluating the exact influence one soluble substance will exercise over another in solution must be very great , owing to the existence of peculiarities such as are referred to by Graham .
Indeed , to repeat a statement made in Part II , " It can scarcely be doubted that the forces at work in solutions are too complex in character to be expressed as simple mathematical laws .
The fact that water is itself a complex material , which varies greatly in composition as the conditions are changed , has been left almost wholly , if not entirely , out of account in discussing electrolytic and * ' Journ. Amer .
Chem. Soc. , ' 1908 , vol. 30 , p. 946 .
t 'Ann .
Chim .
Phys. , ' 1894 , ( 7 ) , vol. 3 , p. 275 .
1 This conception is embodied in the term Haftdruclc , proposed by Traube .
Prof. H. E. Armstrong and Dr. J. V. Eyre .
[ Jan. 1 , hydration phenomena ; and far too little attention has been paid also to the existence of salts in solution in various states of molecular aggregation/ ' If it be granted that solvent and solute are reciprocally active in the process of dissolution and that the dissolved substance in an aqueous solution is associated with " water , " it follows of necessity that when a solution saturated with a particular salt is mixed with a second soluble salt which , ex hypothesi , is also hydrated in solution\#151 ; the second salt being of such nature that no interaction of the two can take place\#151 ; some of the salt with which the solution was originally saturated must be precipitated , unless the affinity of this salt for water be so strong that it cannot be overcome by that of the added second salt , as in the case observed by Etard : precipitation must continue up to the point at which the two salts share the solvent in certain characteristic proportions .
But any neutral soluble substance which either combines with water on dissolving or in any other way exercises a dehydrating effect ( cp .
S 9 ) should act in this manner when introduced into the solution and therefore should cause precipitation .
On account of the importance attached by Nernst and others to the behaviour of slightly soluble salts , we have thought it desirable to extend our experiments to the two slightly soluble salts lead chloride and silver acetate and to determine the influence of a number of neutral precipitants not previously studied at the time when the experiments were instituted , though in the interval several accounts have been published by other workers who have had the same problem under consideration .
It may be added that the work now recorded was completed in the spring of 1911 .
The determinations were all made at 25 ' C. The substances used were prepared from materials sold as pure ( cp .
XI , p. 124 ) .
It was not found possible to estimate with certainty the small quantities of lead chloride present in the solutions by gravimetric methods nor was it possible to effect the determination by the usual volumetric method of titration with a solution of silver nitrate in the presence of potassium chromate as indicator , as a precipitate of silver chromate was formed at once ; but by modifying this method , so that all the lead was precipitated as chromate and a slight excess of potassium chromate left to serve as indicator , it was possible to estimate the chloride in the solution by gradually adding a solution of silver nitrate in the ordinary way .
If the titration be carried out in a white porcelain basin , satisfactory results are obtained without difficulty .
This was established by careful experiments with saturated solutions of lead chloride ; the results obtained differed among themselves by less than 1 per cent. 1913 .
] Studies of the Processes Operative in Solutions .
237 When hydrogen chloride was used as precipitant , the amount of acid present in the pipetted portion of the saturated solution was first determined by titration against standardised alkali ; the total chloride present was then determined in the neutral solution by titration with a standardised solution of silver nitrate and the amount of lead chloride present deduced from the two values .
The method used in estimating the amount of silver acetate in the saturated solutions was that known as Pisani 's method , which is based upon the fact that an aqueous solution of iodised starch is decolourised by solutions of silver salts .
A suitable solution for the purpose was prepared by adding to half a litre of hot water from 10 to 12 grm. of soluble starch made into a paste with water ; after boiling the liquid during a few minutes , it was diluted to about 1 litre and mixed with a few drops of an alcoholic solution of iodine , so as to render it ' a deep blue colour .
As it is necessary to use only very dilute solutions of silver salt , the samples of saturated solution were always diluted to 500 c.c. ; the diluted solution was titrated against 500 c.c. of the iodised starch solution after this had been standardised against a solution of silver nitrate of suitable strength .
With practice , very satisfactory results can be obtained by this method .
The extent to which the determinations are in agreement is shown in the following table .
The results of separate experiments with different samples of salt are given in sections I and II , whilst A and B represent those obtained with two samples of the same solution , the difference being that B was withdrawn an hour later than A:\#151 ; .
100 grms. of water at 25 ' C. dissolve Lead chloride .
Silver acetate .
grms. TfA 1*1024 A 1 B 1 -1030 ttJA 1-1026 11 \B 1*1021 grms. T/ A 1-096 \#177 ; tB 1-104 tt/ A 1*114 A\#177 ; 1B 1-117 The results are recorded in the table on p. 238 and are also represented graphically in the diagrams on p. 239 .
The effects on the solubility of potassium chloride of a variety of precipitants not previously studied are also recorded in the table and in the graphs .
Prof. H. E. Armstrong and Dr. J. Y. Eyre .
[ Jan. 1 Precipitant .
Molecular concentration of precipitant per 55 *5 mols .
of water .
Solubility in 1000 grin , of water at 25 ' .
Relative density , d 2\#151 ; 25 Molecular solubility .
Apparent molecular hydration of precipitant .
\#165 ; Lead Chloride .
__ 11 *0276 1*0098 0 *0397 Ethylic alcohol 4 10 *6609 1 *0069 0 *0383 7*39 Glycol " i 11*0352 1 *0116 0 *0397 - 0*15 1 11 *7035 1*0170 0 0421 - 3*41 Acetaldehyde i 10 *7638 1 *0097 0 *0387 + 5*31 ... i 10 *2451 1 *0095 0 *0368 5*25 Paraldehyde i 10*3910 1 *0114 0 *0313 12*83 11 *0210 1 *0098 0 *0396 Paraldehyde Ta 10 *7185 1 *0101 0 *0385 18*28 Glycerol i 11 *3658 1 *0152 0*0409 - 6*95 11 *0321 1 *0104 0 *0397 Propylic alcohol 4 10 *3266 1 *0066 0 *0371 14 *20 1 10 *0262 0 *9984 0 *0324 10 09 Hydrogen chloride ... 4 4 *2849 1 *0058 0 *0154 135 *85 i 3 *6833 1 0098 0 *0132 73*97 Methylic acetanilide It 10 *6567 1*0111 0 *0383 9*85 Lead nitrate 4 14 *2079 1 *0816 0 0510 -63 *91 55 .
To 11*6515 1 *0383 0 *0420 -31 *17 35 To 10 *7144 1 0170 0 *0386 + 79 *92 35 jio 11 1192 1 *0118 0 *0400 -43 *82 Silver Acetate .
10*2350 1 *0081 0 *0613 Glycol 1 9 0145 1 *0154 0*0540 6*62 Glycerol 1 9 *5420 1 *0277 0 *0572 3*76 Propylic alcohol 1 8 *5795 0 *9981 0 *0514 8*98 Paraldehyde Ta 9 *1015 1 *0084 0 *0535 24 *59 \#151 ; 11 *1780 1 *0085 0 *0669 Isobutylic alcohol ... s 8 *1080 0 *9996 0 *0486 20 *33 Propylic alcohol i 10 *1335 1*0059 0 0607 18 *75 Glycol i 11 *1570 1 *0099 0 *0668 0*42 10 *4250 0 *0625 Acetaldehyde 4 10 -3361 \#151 ; .
0 *0619 1*45 Paraldehyde i 9 *5478 .
~~ 0 *0572 18 *20 Potassium Chloride .
367 -700 1 *1820 4 *9315 .
Acetaldehyde 4 375 -119 1 *1791 4 *8968 1 *56 Paraldehyde A 363 -051 1 *1786 4 *8692 1 *42 Glycol i 365 -134 1 *1802 4 *8970 1 *51 l 359 100 1 *1762 4*8160 1*30 Glycerol i 366 -700 1*1830 4 *9194 0*56 Mannitol i 367 -455 1 *1903 4 *9285 0*15 33 A 368 -000 1 *1884 4 *9360 - 0 *22 Molecular proportion of precipitant per 55*5 mol .
prop , water .
1913 .
] Studies of the Processes Operative in Solutions .
Lead Chloride as Solute .
\ \ GLYCOL \ ACETALDEHYDE \ PROPY If \ GLYCEROL \ ETHYUC ALCL HOL \ A A \gt ; METHYL \gt ; ACETANILIDE PARALDEHYDE -to -3 O 5 / O / S 20 25 30 Silver Acetate as Solute .
JSO-BUTYUC A '.COHO/ .
GLYCOL PROPYUC AH ACETALDEHYDi ' Potassium Chloride as Solute .
J/ 4 '/ a \#171 ; . . . .
i * \#171 ; . .
\#171 ; B 1 \#151 ; 1 .
1 . . .
1 i GLUCOSE* CLP .
\#171 ; .
:ol 1 " ETHYL/ \amp ; ALCOHc\ 1 1 { PROPYLIC ALCL .
1 HO* HYDROGEN C .
C HLORiDE I .
S .
S CLY .
TERO , [ / .
' l t .
\#166 ; . . .
MANN/ T\amp ; ~ ''ACETALDEj rrDE PARALDEH X WE -to '5 0 3/ 0 / 5 20 23 30 Apparent molecular hydration of precipitant .
Prof. H. E. Armstrong and Dr. J. V. Eyre .
[ Jan. 1 , 1 .
Taking into account the observations of other workers as well as those brought forward in this and the two previous communications , it is clear that , in principle , no distinction can be drawn between slightly soluble salts and soluble salts .
Owing to the difficulty which attends the determination of solubility in the case of slightly soluble salts , we do not attach weight to the smaller differences observed between the lead and the silver salt , the more as we found it impossible to obtain constant values in the case of the latter , different samples giving different values\#151 ; whether on account of surface changes due to the action of light or to differences in the state of aggregation ( cp .
XI , p. 125 ) .
For example , we are not prepared to regard propylic alcohol and paraldehyde as more active precipitants of the silver than of the lead salt but are inclined to think that the activity of both precipitants , when present in small proportions , may have been overestimated in the case of the silver salt .
2 .
Non-electrolytes and electrolytes alike act as precipitants of salts ; though in both cases substances are to be found which increase solubility instead of depressing it .
3 .
As the activity of non-electrolytes as precipitants is inconsistent with the postulates of the ionic hypothesis , it is obvious that the explanation of the behaviour of precipitants in general must be based upon some broader hypothesis .
4 .
It is necessary , in all cases , to take into account the changes which affect the solvent as well as those that affect the solute ; to neglect consideration of the changes in the solute\#151 ; the almost universal habit of those who postulate the existence of dissociated , separate ions in solutions of electrolytes\#151 ; is an indefensible practice .
5 .
If it be admitted that both solids and liquids are formed by the association of the fundamental molecules characteristic of the gaseous state , dissolution must be regarded as in large measure a process of " depolymerisation " not of the solute alone but also of the solvent .
6 .
It is also necessary , far more than has been customary of late , to take into account the reciprocal changes which attend dissolution , other than those which involve disturbance of molecular complexity , namely , those which involve the formation of new molecular species , whether as the outcome of interactions or of mere association of solvent and solute .
7 .
One of the most striking facts brought out when the evidence now available is considered , is the remarkable antithesis presented by the activity of electrolytes and non-electrolytes as precipitants : whereas , in aqueous solutions , the former are usually the more active the more soluble they are , in the case of non-electrolytes the relationship is of the reverse order .
1913 .
] Studies of the Processes Operative in Solutions .
241 8 .
The simplest case to be considered is that of carbon dioxide ; as the molecules of this gas are unlikely to be present in solution in any other form than as simple molecules such as are represented by the formula C02 , any increase in its solubility cannot well be interpreted otherwise than as due to an increase merely in the number of molecules undergoing some form of hydration which renders them soluble .
Usher 's determinations of the solubility of this gas in presence of various organic substances* are , therefore , of particular value .
From experiments made with semi-normal solutions , he has deduced the results displayed in the-following table , the value given being the number of cubic centimetres of gas dissolved by 1000 grin , of water at 20 ' in presence of the substance quoted :\#151 ; W ater . .
878 8 .
Pyrogallol 894 1 .
Cane sugar . .
797 9 .
w-Propylic alcohol 902 2 .
Mannitol . .
833 10 .
Acetamide 906 3 .
Dextrose . .
841 11 .
Urethane 907 4 .
Glycine . .
864 12 .
Catechol 908 5 .
Carbamide . .
884 13 .
Quinol 928 6 .
Thiocarbamide ... . . .
885 14 .
Antipyrine 935 7 .
Acetic acid . .
893 15 .
Resorcinol 945 9 .
In the case of the three substances first on the list , the observed lowering of the solubility cannot well be regarded otherwise than as a direct dehydration effect , that is to say , as due to a diminution in the amount of free water present owing to the association of a certain amount of the water used with the added solute .
Excluding Nos. 4 , 5 , 6 , 10 , 11 and 14 , which conceivably may exercise an influence as basic compounds , though perhaps scarcely to the extent observed , , the remaining substances apparently all serve to render the water more active as a solvent of carbon dioxide\#151 ; more effective , that is to say , in converting it into a " hydrated " form .
Now the view advocated in previous communications of this series involves the assumption that when neutral substances are dissolved in water they serve to increase the proportion of simpler molecules in the liquid\#151 ; i.e. , molecules of hydrone , OH2\#151 ; and , therefore , to render it a more active agent .
' 10 .
Confining our attention to propylic alcohol , the only one of the substances under consideration of which we have experience , which is an active precipitant of salts\#151 ; more active , in fact , than ethylic alcohol\#151 ; the increase in the solubility of carbon dioxide in its presence , amounting to 24 c.c. , cannot well be ascribed to the direct dissolution of the gas in the alcohol .
* 'Chem .
Soc. Trans. , ' 1910 , p. 66 .
Prof. H. E. Armstrong and Dr. J. V. Eyre .
[ Jan. 1 , Ethylic alcohol depresses the solubility of carbon dioxide in water .
The coefficients found by Findlay and Shen* are as follows ( at 25 ' , 737-747 mm. ) Water ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... 0*825 " containing 2*95 grm. alcohol ( in 100 c.c. ) ... .
0*812 " " 3*01 " " ... .
0*814 " " 8*83 " " ... .
0*786 As ethylic alcohol undoubtedly has a greater affinity for water than propylic , it may be supposed that it is far more completely hydrated and more under the control of the water than is propylic alcohol : consequently , that more molecules of the latter alcohol may be active in promoting dissociation of water molecules in the sense of the expression ( OH2 ) ; c \#151 ; \#166 ; - xOH2 .
On this assumption , a larger proportion of the molecules of carbon dioxide would be subject to hydration in a solution of propylic than in one of ethylic alcohol : hence the increase in solubility .
The argument is of general application .
1L When the results we have obtained with silver acetate , lead chloride and various haloids and those obtained by Eothmundf with silver sulphate , potassium bromate , potassium perchlorate and lithium carbonate\#151 ; four sparingly soluble salts which are more soluble than the two we used but less soluble than the remainder of those we have studied\#151 ; are contrasted , the same substances are seen to act as precipitants and , with certain marked exceptions , the order of their activity is the same .
12 .
In cases in which direct comparison is possible , the less soluble nonelectrolyte is always the more active precipitant : thus the activity of the monhydric alcohols is in the order of increasing molecular weight .
Paracet-aldehyde is more active than acetaldehyde , whether comparison be made of equal weights or of moleeularly similar proportions .
The activity of polyhydric alcohols diminishes as their hydricity and solubility increases .
Thiourea is more active than urea .
13 .
The experiments carried out by Fox and Gauge on the solubility of potassium sulphate^ and by Eothmund on the salts above mentioned are of special interest in this connexion .
The solubility of potassium sulphate is at first unaffected by the presence of cane sugar and is only slightly diminished as the proportion of sugar is increased ; obviously , therefore , as it must be supposed that cane sugar * 'Chem .
Soc. Trans.,5 1911 , p. 1313 .
t ' Zeit .
phys .
Chem.,5 1909 , vol. 69 , p. 523 .
X 'Chem .
Soc. Trans.,5 1910 , p. 377 . .
1913 .
] Studies of the Processes Operative in Solutions .
243 becomes associated with a certain proportion of the water , it must serve from the beginning to promote the solubility of the salt .
Mannitol , glycerol , glycol , pyridine , ethylic alcohol and acetone are active as precipitants from the beginning , in the order mentioned .
Most unfortunately , Rothmund 's experiments were carried out with volume normal solutions : consequently , no two solutions contained the same molecular proportion of precipitant and water , so that the results are not comparable among themselves .
Nevertheless , it is obvious , when strictly neutral precipitants such as the monhydric alcohols are considered , that these maintain their position in order of activity whatever the salt used may be .
The cases in which the precipitants vary in behaviour are significant .
Thus , whilst glucose promotes the dissolution of silver sulphate and of lithium carbonate , it hinders that of potassium bromate .
Similar differences are observed in the case of cane sugar .
Obviously therefore the behaviour of the salts of the " dibasic " acids is different from that of the monobasic : both sulphuric and carbonic acids , however , appear to be in reality monobasic acids upon which a slight extra activity is imposed ; it may well be that an exchange of radicles takes place between such salts and basic substances such as the sugars .
On the other hand , whilst phenol increases the solubility of silver sulphate , it depresses that of potassium bromate and perchlorate\#151 ; but acetic acid acts as a weak depressant in all cases .
It is significant also , from this point of view that acetonitrile promotes the solubility of silver sulphate whilst acting as a precipitant of lithium carbonate : it is well known that nitrogen compounds have a specially strong affinity for silver salts .
# * Similar interpretations may be given of other cases of " irregularity .
" 14 .
Propylic alcohol , which promotes the solubility of carbon dioxide , is a powerful precipitant of salts .
The assumption made previously ( S 8 ) , it may be pointed out here , is a satisfactory explanation of this difference .
In the case of carbon dioxide , an increase of solubility may be explained as the consequence of an increase of the active agent in the solution , that is to say in the number of molecules of hydrone , OH2 .
In the case of salts , however , though an increase in the proportion of molecules of hydrone would render the liquid a better solvent of the salt , it would also favour the tendency of the fundamental molecules of the salt to combine among themselves , as action would set in between hydrone molecules in the solution and those attached to the salt ; in other words , an increase in the number of hydrone molecules wmuld render the solution a dehydrating medium .
vol. lxxxviii.\#151 ; A. s Prof. H. E. Armstrong and Dr. J. V. Eyre .
[ Jan. 1 , 15 .
In the case of salts , an additional force is operative which cannot come into existence in a case such as that of carbon dioxide .
In the solid state , all salts presumably are " polymerised " forms of their fundamental molecules and on entering into solution , such molecules are more or less , if not entirely , dissociated .
The solubility of a salt may be said , in fact , to be determined by the two sets of affinities pictured in the expressions ( EX ) , ^ nBX x(RX ) + y( OH2 ) ^ ( RX)*(OH2V the solubility in each particular case depending on the extent to which the two opposite tendencies come into play .
Both forms of dissociative change shown in the equations would be promoted by those precipitants especially which have but slight affinity for water , as such molecules would be most effective mechanically .
The difference between salts generally and the corresponding acids is probably due to the fact that the acid radicle is more effectively held in check or neutralised by hydrogen than by any metal ; and the differences in salts may be attributed to the different extents to which the metals " saturate " the acid radicles .
On such a hypothesis , it is possible to understand that a metal should have the relatively constant effect which it is known to have in salts .
In the case of a very difficultly soluble salt such as is silver chloride , the affinity of the individual fundamental molecules for one another appears to be very great .
It is not improbable that in so far as such a salt is soluble at all , it enters into solution only in its simplest form represented by the symbol AgCl .
This molecule may be very active ; it may be so fully activated , in fact , as to be completely " dissociated " \#151 ; to use the now conventional term .
The solubility of the salt is slight because the affinity between the fundamental molecules is so special that it prevails almost entirely over the affinity which hydrone has for these molecules .
In the case of soluble salts the balance is far more nearly even .
The fact that salts generally are not precipitable by neutral substances in the inverse order of their solubility is probably to be explained by considerations of this order .
16 .
It will be noticed that in the case of lead chloride and silver acetate the graphs of propylic alcohol and paraldehyde slope backwards\#151 ; in other words these substances are the more active as precipitants the smaller the proportion present .
In all probability , as the amount present is increased , they tend more and more to promote the dissociation of the complex molecules of the salts into simpler more soluble molecules .
17 .
There can be little doubt that the polyhydric alcohols and sugars especially act mainly when not entirely as direct dehydrating agents .
1913 .
] Studies of the Processes Operative in Solutions .
245 In cases in which solubility is increased by such substances there can be little doubt that this is because more soluble substances are formed by the interaction of the substances present .
It is well known that the sugars form compounds with salts.* 18 .
In all cases , a soluble substance must exercise some influence directly as a dehydrating agent ; but the influence it exercises in this way must diminish and the effect it lias on the solvent must increase as the solubility of the substance diminishes ; in so far , however , as the added substance induces the dissociation of complex molecules of the solute , its action will be to increase solubility .
19 .
In fine , complex as the phenomena with which we have to deal in aqueous solutions undoubtedly are , it appears to be possible to interpret them on broad and general grounds by the application of ordinary chemical principles .
Little more is required than to extend to oxygen the conceptions which are accepted in the case of carbon .
20 .
It is only necessary to assume that water is a complex material consisting of a variety of molecular species in proportions which vary with the temperature , some comparable with the polymethylenes , others with hydrols ( cp .
Parts VI , XIX , XXIV ) .
Further , to assume that when substances are dissolved in water they are hydrolated and hydronated in various ways and to various extents and at the same time produce variations in the water ( XVIII ) .
Lastly , to admit that the several forms of " hydrated " compound are not all active chemically and that those which are active are not all equally so\#151 ; as in the case of the polymethylenes .
Such assumptions made , it follows that every variation in the conditions will involve variation not only in the active but also in the inactive constituents of the solution ; in other words , both components of a solution must vary as the conditions are varied .
[ The authors are indebted to Mr. F. W. Jackson for the assistance which he has given them in carrying out their experiments .
] * The greater solubility of lead chloride in presence of lead nitrate cannot well be accounted for otherwise than as the consequence of the formation of a mixed salt , PbCl(N03 ) .

