B.— CHEMISTRY. 37 



and for esters about 7 (all these being free from co-ordinate links), it is 

 greatly increased by the introduction of a nitro-group 



^° 

 - -W , 



^0 



which contains this link, and is for nitromethane 39 and for nitro- 

 benzene 36. The effect on the boiling point is seen by comparing the 

 alkyl nitrites K— — N=0 with the isomeric nitro-compounds 



R-N^ : 



the latter boil from 50° to 100° higher than the former. Many other 

 examples might be quoted. 



These examples suggest the consideration of associated liquids. As 

 long as we were at liberty to invent new kinds of subsidiary valencies, the 

 existence of association caused no trouble. But now that we claim to 

 have discovered the mechanism of atomic combination, we must identify 

 the link between the molecules of an associated substance with one or 

 other of the forms of link that we have recognised, and it is evident that 

 the co-ordinate link is the form required. We ought therefore to be 

 able to find in every associated substance a donor and an acceptor atom. 

 Such atoms are always found to be present : in the most familiar class of 

 associated compounds, those containing hydroxyl groups, the oxygen 

 atom of this group, with its two pairs of unshared valency electrons, is the 

 donor, and the hydrogen atom, being able as we have seen to increase its 

 valency group from two to four, is the acceptor. We thus* get the 

 possibility of an indefinite degree of polymerisation : — 



vH. /R / 

 R-O^H-O^H-0-> &c. 



That the association does depend on the two atoms of the hydroxyl group 

 is shown by the fact that if we replace either the oxygen by sulphur 

 or the hydrogen by an alkyl group the association disappears : neither 

 the mercaptans nor the ethers are associated. Associated substances 

 possess the properties which we have seen to accompany the co-ordinate 

 link, the high dielectric constant and the low volatility. The latter 

 property is commonly taken to be sufficiently explained by the rise of 

 molecular weight which the association produces, but unless this is much 

 greater than we have any reason to suppose, it will not account for the 

 whole effect. For example, the ethers boil about 60° lower than the 

 corresponding thio-ethers ; hydrogen sulphide boils at -61°, and so 

 unimolecular H 2 should boil about - 120°. If the real formula of water 

 is H 6 3 (and it is very improbable that its average polymerisation is even 

 as great as this at 100°), its true molecular weight is not 18 but 54. This 

 will account for a rise in the boiling point, but not for so large a rise as 

 is actually found. Hydrogen selenide (mol. wt. 81-2) boils at — 42°, 

 and butane (mol. wt. 58) at + 1°. Evidently the polymerised molecules 



