1 46 SCIENCE PROGRESS 



C05 per cent., was probably three times as great, and even after 

 allowing for this the two series of observations differ at 25 per 

 cent, to the extent of 0*0004, in spite of the fact that the experi- 

 mental error was estimated at 0*0000063 1 only. Greater interest 

 attaches, however, to the fact that from 5 per cent, to 18 percent, 

 there is an exact linear relationship between the density and 

 concentration, the maximum error amounting to only two units 

 in the fifth place of decimals ; it is therefore evident that 

 Pickering's break at 12*1 per cent, must have been due either to 

 unsuspected experimental errors similar to those referred to 

 above, or to the inability of the bent lath to follow the changes 

 of curvature in the derived curve which he selected for 

 examination. 



If produced to zero-concentration the linear part of the 

 density curve would cut the axis of density at 0*99939 instead 

 of at 0*99866 — the actual density of water at 18 . There is, 

 therefore, considerable bending in the region from 5 to o per 

 cent., which includes Pickering's break at NaOH . 53H2O. That 

 this sudden twist is due to the depolymerisation of the complex 

 water-molecules by the caustic soda, and not to the formation 

 of a new hydrate, is abundantly proved by a study of the 

 density-temperature curves (Bousfield and Lowry, loc. cit 

 pp. 274-80). The maximum density, which forms the chief 

 evidence of the presence of the ice molecules, is lowered to 

 o° by the addition of o*6 per cent. NaOH, and to — io° by 

 2 per cent. NaOH only; at 12 per cent. NaOH the solution 

 has become so far simplified that the effect of temperature can 

 be represented by a quadratic instead of a cubic equation. The 

 density-temperature curves for the concentrated solutions are 

 practically straight, but on dilution steadily approximate to the 

 complex curve which represents the expansion of water. It is 

 therefore impossible to avoid the conclusion that the com- 

 plexities of the dilute solutions are due to the inherent 

 properties of the solvent and not to those of the solute. 



The new interpretation of Pickering's breaks, which is 

 described above, differs considerably from that which he him- 

 self advocated, but not more widely than his breaks differ 

 from those which were sought and found by Mendeleef. The 

 latter looked for sudden changes of direction and curvature 

 similar to those which actually occur in the case of the 

 hydrate S0 3 . H 2 ; such changes can only occur in the case 



