SMITH. 



EXPANSION OF ETHER AND ALCOHOL. 



455 



Pagliani, because my observations did not cover that region. This 

 curve, after passing its well-known minimum, is seen to become steeper 

 as the amount of alcohol in the mixture is increased ; and it is seen at 

 once that the compressibility of absolute alcohol decreases rapidly with 

 the introduction of a small amount of water. This decrease in the 

 compressibility would mean an increase in a. It is therefore probable 

 that the values of a given in Table XIII, which are for alcohol contain- 

 ing 0.5 per cent water, may be as much as 3 per cent greater than the 

 corresponding values for absolute alcohol. Allowing for this possible 

 discrepancy, we may still consider the values here found as being of 

 interest in connection with those obtained from the data of Amagat. 



20 



40 CO 



Curve II. 



80 



100 



We may perhaps see most clearly the contribution which these recent 

 values make to our knowledge of the van der Waals a in alcohol by 

 inserting them in their proper places in the table given by Professor 

 Hall. As given there, they have been corrected for the possible error 

 due to the impurity of the alcohol. That is, I have reduced by .3 per 

 cent the values of a given in Table XIII, after making an estimate of 

 the value at 80° from the value at 78^, assuming at the same time that 

 Amagat used absolute alcohol. The values of a calculated from my 

 data are enclosed in brackets in order to distinguish them from the val- 

 ues obtained from Amagat's data. The numbers enclosed in parenthe- 



