74 



NATURE 



[March 17, 1921 



In this the arrows indicate the direction of an 

 imagined plane or intra-molecular interface i separa- 



C— N 

 ting the hydrophile groups • • , which are con- 



O H 

 solute with water (in virtue of residual affinities tend- 

 ing to complete the amino- and carboxyl groups), jrom 

 the hydrophobe or hydrocarbon groups — CHR. Not 

 only in one and the same protem molecule, but also 

 to a variable extent between molecules, we may admit 

 that this primary orientation leads to mutual attrac- 

 tion between water-soluble and water-insoluble groups 

 respectively. Without any actual cleavage of the 

 molecule, we have orientation and a stratichemical 

 field of force which is of a similar character, in 

 essence, to crystallisation, but results in incomplete 

 instead of complete equilibrium. The hydrocarbon or 

 lipoid atom groups will approach the fluid on the 

 solid state according to molecular weights and con- 

 stitution ; hence the system may be likened, in one 

 aspect, to a sub-molecular emulsion, the lipoid groups 

 tending to form interconnected sheets of atom-groups 

 necessarily permeable to water and water solutes, 

 although mechanically developing a stress resisting 

 rupture in virtue of the fields of attraction and repul- 

 sion induced. The micelles are the smallest pluri- 

 molecular units thus built up. 



Applied to soaps, we have similarly a mutual attrac- 

 tion and solution of the hydrocarbon portions of the 

 fatty acid radicles, without cleavage from the water- 

 soluble portion, which dissolves and ionises. The pas- 

 sage of ionised micelles "through the open network 

 of the gel as freely as through the sol " (Laing and 

 McBain, loc. cit., p. 15 19) appears quite consistent 

 with the hypothesis now suggested. Further, the 

 form of the micellar aggregates — strings, sheets, 

 networks of molecules — w-ill, on this view, be a func- 

 tion of the original molecular constitution operating 

 through intra-molecular orientation, and modified by 

 lonisation and tautomerism, where these occur. The 

 quasi-solubility in water of sodium, etc., soaps, being 

 associated with ionisation, passes to insolubility in 

 water with the non-ionising calcium, aluminium, iron, 

 etc., soaps, when the solubility of the fatty acid por- 

 tion (or hydrocarbon group) becomes dominant, and 

 soap sols in non-aqueous solvents result. The stiffening 

 to gels here with increased concentration and lowered 

 temperature mav be due to orientation of both the 

 hydrocarbon and the metallic residues respectively, of 

 the latter either directly or as oxides, these being solids 

 at such temperatures. 



In general, it is submitted that the present hypo- 

 thesis gives a more generalised basis of explanation 

 of what McBain regards as "not vet explained " (loc. 

 cit., D. 1518), viz. "the stable existence of any col- 

 loidal aggregate." S. E. Sheppard. 



Research Laboratorv. Eastman Kodak Co., 

 Rochester, N.Y., January 18. 



Through the kindness of the Editor I have been 

 given an opportunity of commenting upon the specu- 

 lations advanced by Dr. S. E. Sheppard in the fore- 

 going interesting letter. 



It is evident from reference to his paper in the 

 Journal of Industrial and Engineering Chemistry that 

 Dr. Sheppard is tempted to divert^e from the views 

 of nearly all who have studied the properties of sus- 

 pensoid colloids, such as fine suspensions of particles 

 of gold in water, and to regard these as being merely 

 pseudo-colloids. In his opinion, a jellv made from 

 gelatin, protein, starch, or soap would be the typical 

 colloid, thus reverting to Graham's conception that 



NO. 2681, VOL. 107] 



it is the substance, and not the physical state of sub- 

 division, that makes a colloid. 



It is impossible to exclude ionisation hypotheses 

 from colloid chemistry now that it has been demon- 

 strated that soaps in colloidal form are excellent con- 

 ductors. At the same time we are quite clear that 

 a theory of gels cannot depend upon ionisation pheno- 

 mena, since gels occur in non-aqueous solvents which 

 possess no measurable conductivity. 



It is difficult to understand exactly what is meant 

 by some of the technical terms used or coined without 

 definition; but apparently Dr. Sheppard 's conception 

 of a stable colloid is a substance which contains atoms 

 or atomic groupings, commonly found in chemicals 

 which are insoluble in the solvent under discussion. 

 For example, in aqueous sodium palmitate the long 

 paraffin chain is regarded as being in itself insoluble 

 in water, in contradistinction to the sodium atoms 

 and ions. This is considered to result in a tendency 

 for these hydrocarbon chains throughout the solution 

 to become linked to each other through the residual 

 affinity of the paraffin part of the molecule, to form 

 sheets of molecular network co-extensive with the 

 solution. The sodium end of each molecule is 

 regarded as "dissolved" and subject to ionisation. 



This conception is sufficiently elastic to conform 

 to many of the facts, but surely such a word as "dis- 

 solved " loses its significance when applied to a 

 solution in which both "undissolved" and "dis- 

 solved " parts of the molecule are present in a state 

 of molecular subdivision. Thus, in the case of an 

 aqueous solution in which gold is present in the form 

 of single atoms — a case which has been very nearly 

 realised — the gold would not be regarded as dissolved 

 in the water, since gold and water are "hetero- 

 chemical." The modern or current conception is cer- 

 tainly that this Would be a true solution of gold, 

 although highly supersaturated. It is evident that 

 the difference is one of words, and not of scientific 

 fact. 



Again, it would be difficult to explain on Dr. Shep- 

 pard's conception the existence of gels such as that 

 of rubber in benzene, in which surely every part of 

 the hydrocarbon must be considered potentially soluble 

 or " consolute " with benzene. Further, on what 

 chemical grounds could one predict the formation of 

 a g:el of cadmium in alcohdl? 



I cannot but feel that even this conception of con- 

 tinuous open molecular network as constituting the 

 typical colloid still leaves unexplained the stable exist- 

 ence of the colloidal aggregates of sols as distin- 

 guished from gels. In the case of an ordinary soap 

 solution or sol, for example, perfect reversible equili- 

 brium prevails, and vet the soap does not exist as 

 "a continuous, semi-rigid framework," nor yet as 

 single, independent molecules — that is, as crystalloid — 

 since when in the latter condition it exhibits familiar 

 crystalloidal properties such as osmotic activity. 

 Hence our conclusion is that the soap is largely in the 

 form of particles, each an aggregation of large 

 numbers of molecules. Miss Laing found that there 

 is a very ready change from sol to gel without altera- 

 tion of either conductivity or osmotic activity. We 

 seem forced, therefore, to conclude that the gel is 

 built UD from the same colloidal particles as the sol. 



Similarly, in Svedberg's example of cadmium or 

 cadmium oxide in alcohol, which at rest forms a jelly 

 but on stirring reverts to a fluid sol, the colloidal 

 particles of the sol must undoubtedly be those of the 

 gel also. In this case the individual colloidal particles 

 are presumablv crystalline, in analogy with the experi- 

 mental results recently obtained in Sherrer's X-ray 

 investigations. 



