HOLECUUBS AND ATOMS 323 



properties are determined by the composition of the molecules, and by the 

 properties of the elements forming them. Thus the density of solids 



determine the osmotic pressure, as in the experiment described in Chap. I. Note 19. 

 De Vries found in vegetable cells a convenient means of determining such (isotonic) 

 solutions, which produce indentical osmotic pressures. A thin slice of a coloured plant 

 tis^in f,, r instance, of Tradescantia discolor is moistened under the microscope with 

 the solution to be tested. If its osmotic pressure be equal, or less than that of the 

 liquid contained in the cells, no visible change occurs ; but if the liquid taken be endued 

 with a greater osmotic pressure than the cellular sap, then the water will pass from 

 the cell, and the coloured matter of the cell will shrink away from the envelope, and 

 this process is easily observed under the microscope. Knowing, then, the osmotic 

 pressure for any one substance for instance, for sugar with different strengths of solu- 

 tions, it is possible to find the osmotic pressure of all other substances investigated, 

 because it is shown by direct experiment that the osmotic pressure increases in pro- 

 portion to the strength of the solution. Thus having fixed on any one substance for 

 instance, sugar and on one of its solutions, we may see the substance of the results 

 which have been attained. 



If (as on p. 64) a one per cent, solution of sugar be taken, then according to the 

 experiments made by Pfeiffer (1877) its osmotic pressure = 53 '5 centimetres at 14. 

 According to the formula of sugar, C^H^On, its molecular weight M = 342, and as the 

 weight of a cubic centimetre of a one per cent, solution of sugar 1'003 gram, therefore 

 the weight of sugar in a cubic centimetre of the solution, or s in the preceding formula 

 (p. 320 : 6255 s (278 + J) = Mp), is equal to O'OIOOS gram, and therefore, according to this 

 formula (as M = 342 and t = 14), p = 52'6 centimetres. This shows that if the sugar were, 

 instead of being in solution, in a state of vapour, then in following the law of Avogadro- 

 Gerhardt it would produce a pressure equal to the osmotic pressure. This deduction 

 (whose sense is at present not clear) forms the substance of Van't Hoff's doctrine 

 (Chap. I. Note 19), when i = I (Chap. I. Note 49). 



Consequently, the molecular weight determines the osmotic pressure (and together 

 with it the vapour tension and temperature of freezing, according to Note 49, Chap. I.), 

 and therefore the molecular weight itself may be determined by the osmotic pressure as 

 well as by the vapour density. 



But so simple a relation only exists for dilute solutions of substances like sugar, 

 which do not conduct an electric current, for which *' = !. For salts and acids which 

 conduct a current, this factor varies up to = 4 (Chap. I. Note 49). Arrhenius explains 

 this phenomenon by supposing (partially after Hittorf and Clausius) that a portion of 

 such substances in solutions, and especially in dilute solutions, occurs in a state of dis- 

 sociation, and owing to this the number of molecules is multiplied (Chap. I. Note 45). 

 As the conceptions of this order have been as yet but very little developed, and as in 

 regarding solutions from this point of view, which is warmly supported by Ostwald, the 

 water or solvent in general, which certainly plays an important part in solutions, and 

 especially dilute ones, is entirely lost sight of, I consider it premature at present to ex- 

 plain the theory of Arrhenius, but think that it contains the seeds for further develop- 

 ment and for its being merged into a fuller theory of solutions. 



For the matter now under our consideration we need not see in the fact of the 

 variability of i any hindrance to employing (a) the determination of the osmotic pressure, 

 (b) the freezing point of a solvent (the so-called Raoult's method), and (c) the variation 

 of the vapour density as means for finding the molecular weight of a substance in solu- 

 tion, not only in the ordinary cases when * = 1, but even in those cases where i~> 1. 

 These methods have already proved useful for solving the question of the molecular 

 weight in many particular instances among the hydrocarbons, and some portion of their 

 application to inorganic compounds will be mentioned in the further course of this work. 



It may not be superfluous to remark that the osmotic pressure in the cells of organisms 

 attains several atmospheres, and probably forms one of the causes determining the 



Y 2 



