of Homologous Compounds. 15 



Associating Substances. 



There can be no doubt that water, the alcohols, and the 

 aliphatic acids must be classed as associating substances, the 

 molecular association occurring in all cases in the liquid state 

 and with the acids in the state of vapour also. The researches 

 o£ Ramsay and Shields on the surface energy of liquids 

 indicate also that the lower aliphatic cyanides, nitro-com- 

 pounds. and ketones exhibit molecular association in a smaller 

 degree than the hydroxyl compounds* . 



The data for associating compounds are given in Table VIII. 



(P- 16). 



There are many irregularities in this Table, chiefly in the 

 data for the alcohols and acids. Among the differences 

 tor the normal primary alcohols there is one negative value, 

 — 2 o, 0, but this is preceded by a very high positive value, 

 + 6°35 ; the most probable explanation of this irregularitv 

 seems to be that the observed boiling-point of octjd alcohol 

 is too low, perhaps 3° or 4°, but of course there may be 

 smaller errors also in the boiling-points of other alcohols. 



The very high value, -j-5 0, 75, in the differences for the acids 

 is not >o easily explained, for the values on either side of it 

 are not markedly low ; it appears as though there were several 

 small errors accumulating in this difference. Possibly the 

 observed boiling-points of ^-valeric, caproic, and oenanthylic 

 acid are all too high, and those of caprylic, pelargonic, and 

 capric acid too low. 



The boiling-points of the iso-acids, with the exception 

 probably of those of isobutyric and isovaleric acids are too 

 uncertain to be included in the table. Taking the boiling- 

 points of these two acids as 427° and 448° abs., the observed 

 and calculated values of A would be 21°'0 and 22°'75, dif- 

 ference 4-l 0, 75. 



As regards the secondary alcohols, there is a negative 

 difference, — 4°'5, preceded and followed by large positive 

 differences, +8°'2 and 4 7 0, 6. These irregularities seem to 

 indicate serious errors in the observed boiling-points of the 

 three highest alcohols examined, that of heptyl alcohol being 

 almost certainly too high, but it is probable that the error in 

 the mean difference may be small. 



The boiling-points of isobutyl and iso-amyl alcohol are, 

 no doubt, nearly correct, and the difference 2°*05, between 

 the calculated and observed values of A is small, as in other 

 cases where the variable radical contains an iso-group. On 



* See also Guye, Archives des Sciences physiques ct naturelles, 1894 [3], 

 xxxi. pp. -58 k 403. 



