TRANSACTIONS OF SECTION B. 615 



The following Papers were read : — 



2. On SchuUer's Yelloxo Modification of Arsenic} 

 By Professor H. McLeod, F.li.S. 



3. On the Electrolysis of Glass. 

 By Professor W. C. Roberts- Austen, F.R.S. 



4. On the Relations between the Viscosity of Liquids and their Chemical 

 Nature.^ By Dr. T. E. Thorpe, F.R.S., and J. W. Eodger. 



During the past four years the authors have been making observations on the 

 viscosity of liquids with the view of establishing relationships between this property 

 on the one hand and molecular weight and molecular structure on the other. They 

 have employed the tube-method of measuring the viscosity coefficieut, and in the 

 case of each liquid the observations extend over the temperature range between 0° 

 and the ordinary boiling-point. More than eighty liquids have now been 

 examined. 



For liquids which do not appear to contain molecular aggregates, at any tem- 

 perature, the following conclusions may be drawn : — 



1. In homologous series the viscosity coefficient is greater the greater the mole- 

 cular weight. 



2. The coefficient of a normal compound is greater than that of the isomeric 

 iso-compound. 



3. The coefficient of an allvl compound is intermediate to those of the corre- 

 sponding normal propyl and "iso-propyl compounds, and, in general, constitution 

 exerts a regular eft'ect on the viscosity coefficient. 



Liquids which appear to contain complex molecules in certain cases do not obey 

 these rules. Formic and acetic acids are exceptions to Rule 1. The alcohols do 

 not conform to Rules 2 and 3. In general, the effect of temperature upon viscosity 

 is much greater for complex than for simple liquids. In both classes of liquids the 

 behaviour of the initial members of several homologous series does not accord wjth 

 that of higher homologues. 



In attempting to quantitatively connect viscosity with chemical nature, the 

 authors have used two magnitudes— the molecular viscosity and the molecular vis- 

 cosity-work — which may be derived from the viscosity coefficient. If »/ be the 

 viscosity coefficient and v be the molecular volume, the molecular viscosity is t/I'S, 

 or the product of j; and the molecular area. The molecular viscosity- work is jjy, 

 the product of r] and the molecular volume. The values of these rnagnitudes 

 have been examined at three different series of temperatures of comparison— viz., 

 the ordinary boiling-points, the corresponding temperatures of 0°-6, and tempera- 

 tures of equal slope or points on the viscosity curves at which temperature is 

 exercising the same effect upon the viscosity of each liquid. On ascribing 

 definite partial values to the atoms and the diiierent modes of atom-linkage — the 

 iso-grouping, double linkage, the ring-grouping, &c. — it has been found possible to 

 calculate the viscosity magnitudes of the great majority of the simple liquids. 

 The results obtained at equal slope are by far the most precise, but even here the 

 alcohols, water, and, to a less extent, the acids are anomalous, doubtless on account 

 of the influence of molecular complexity. A strong point in favour of the new- 

 method of using equal slope as a condition of comparison is that the stoichiometric 

 relationsliips obtained at any one value of the slope appear to be general, and thus 

 to be independent of the particular value of the slope at which the comparisons 

 ai"e made. This conclusion applies even to complex liquids like water and the acids, 

 but not to the alcohols, which of all the liquids examined exhibit, at all of the 

 systems of comparison, the most exceptional behaviour. 



' Published in the Chemical News, Ixx., p. 139, Sept. 2], 1894. 

 2 Published in full in the Phil. Trans., 1894. 



