586 SCIENCE PROGRESS 



The group of the ketones also shows a practically uniform 

 value in mean of 270. 



These regularities, it must again be stated, are not per- 

 ceived in the case of highly associated molecules, such as water, 

 certain acids, and alcohols, which are known from other 

 phenomena to have an abnormal molecular complexity. In 

 the case of the higher members of the C n H 2w 2 series of acids, 

 this molecular complexity breaks down with increasing mole- 

 cular weight : and it is found that propionic acid (343), butyric 

 acid (343), and iso-butyric acid (341) have the same value of 

 t) . io 5 : y/s. 



Considering that the regularities observed by Arrhenius 

 occur among substances which are chemically unrelated, it 

 would further seem that chemical nature and constitutive influ- 

 ences are, at most, only subordinate agencies in affecting the 

 internal friction of liquids. Or, in other words, chemical 

 nature has no greater influence on the viscosity of liquids than 

 the chemical nature of the gases has upon their relations to the 

 ordinary gaseous laws — a generalisation which needs further 

 evidence before it can be wholly accepted. 



Of late years, as the experimental difficulties have been 

 gradually surmounted, considerable attention has been paid 

 to the influence of high pressures on the viscosity of liquids. 

 Faust has shown from observations on ethyl ether, carbon 

 bisulphide, and ethyl alcohol under pressures up to 3,000 

 atmospheres that viscosity increases very rapidly with pressure, 

 especially at low temperatures, and that the rate of increase 

 is far greater at high than at low pressures. The part played 

 by internal pressure on viscosity has been variously stated 

 by Van der Waals and Tammann. The question has now 

 been re-examined by Arrhenius in the light of Faust's experi- 

 mental results. It is to some extent complicated by the cir- 

 cumstance that one of the substances examined by Faust, 

 namely alcohol, is a highly associated liquid, and, as in the case 

 of all such liquids, its behaviour departs from that of liquids of 

 normal complexity. The constants of Van der Waal's equation 

 would appear to change with temperature, and it is inapplicable 

 to liquids far below their critical points. The quantity a, 

 which represents the molecular attraction, in all probability 

 decreases with increasing temperature. In the case of ethyl 

 ether the viscosity is nearly proportional to the square of the 



