546 INFECTION AND RESISTANCE 



excessive use of such classifications is likely to narrow rather than 

 broaden our conception of the whole subject matter of the field. It 

 would seem that the most stimulating point of view is to be reached 

 from the acceptance of the suggestion of Wolfgang Ostwald, that the 

 chief problem of colloid chemistry at the present time lies in deter- 

 mining the influences of the degree of dispersion upon the physical 

 and chemical properties of all liquid solutions, mixtures, suspensions, 

 or what-not. If this point of view be taken it follows that the form 

 and size of the particles in a disperse system are a matter of the first 

 importance. 



If the degree of dispersion in a given system be not too great the 

 form of the particles may be readily observed under the microscope. 

 Such evidence shows that the spherical form predominates enor- 

 mously over all others, although under carefully controlled conditions 

 ovoid forms may appear, as in the case of gelatin and agar. These 

 ovoid forms are taken by von Weimarn (loc. cit.) and others, as evi- 

 dence of directive forces, and hence of incipient crystallization. If 

 the system be treated in such a way as to decrease the dispersion, as, 

 for example, if a reagent be added which tends to flocculate the col- 

 loid, but not in sufficient quantity to produce actual precipitation, 

 the decrease in dispersion may take place in two quite different 

 ways : first, the size of the particles may increase, as in the case of 

 oil emulsions ; second, the particles do not coalesce but become at- 

 tached together in chains and groups which, in many cases, resemble 

 bunches of grapes. This sort of aggregation may go so far as to pro- 

 duce web-like structures. The jellying of gelatin has been shown to 

 be due to the development of such web structures. Glue shows much 

 less tendency in this direction, and if some acetic acid be added, as 

 in the preparation of commercial liquid glues, this web formation is 

 almost entirely absent, and the adhesive qualities are at the same 

 time greatly improved. It seems quite certain that both of the above 

 modes of aggregation are possible in one and the same system at 

 different stages in its condensation. Thus highly dispersed copper 

 sulphid becomes aggregated in its first stages of condensation by an 

 actual increase in the size of the spherical particles. After these 

 reach a certain fairly definite size further aggregation takes place 

 by the grouping together of these spheres. It is generally recognized 

 that all grouped and webbed structures are secondary. 



A large number of very important investigations have been di- 

 rected toward the determination of the size of disperse particles 

 throughout the greatest variation in dimensions. The fact first 

 noted by Graham, that substances in colloidal solution show a very 

 small, and frequently almost negligible, rate of dialysis, points di- 

 rectly to the supposition that the particles in such solutions are in a 

 far less dispersed state than in solutions of crystalloid substances. 

 The rate of dialysis is directly determined by the rate of diffusion. 



