in which 



1876.] Specific Volumes of Liquids. 285 



Due precaution was taken to ensure regular ebullition and to prevent over- 

 heating. In correcting the indications of the thermometer for the cooled 

 portion of the column, I have made use of a special series of observations 

 to obtain the value of I in the well-known expression 



t— the observed temperature on the thermometer, 



£'=the mean temperature of the cooled column, 



w=the length of column, measured in degrees, possessing the 



temperature t\ 

 3 = a constant, usually taken as -000154, i. e. the value of the 

 apparent expansion of mercury in glass for 1°, 



since it is evident, as pointed out by Holtzmann and by Wiillner and 

 Landolt*, that by taking 3= -000154 the results are over-corrected when 

 n is small. The barometric observations are corrected and reduced by 

 the aid of Schumacher's Tables, The corrected boiling-points are reduced 

 to the uniform pressure of 760 millims. by means of the formula 



0-0375°(760-A), 



in which h is the actual height of the barometer at the time of observation. 

 As is well known, this expression does not afford absolutely accurate 

 results, since the relation of the boiling-point to pressure differs with 

 each liquid ; nevertheless the results are more nearly comparable by adopt- 

 ing it than by neglecting it altogether. In one or two cases I have re- 

 duced the indications by the aid of data derived from vapour-tension 

 observations when these have been to hand, the difference between the 

 results thus afforded and those obtained by means of the formula above 

 given has never exceeded 0°-05 C. ; hence we may assume that for the ordi- 

 nary range of barometric pressure the formula is generally applicable. 



The specific gravities of the various liquids were taken at the tempe- 

 rature of melting ice, and are compared with water at 4°. The weighings 

 were made by the method of vibrations, and are reduced to a vacuum* 

 The results thus obtained, and also those calculated for the boiling-points, 

 express the weights in grams of 1 cubic centimetre of the several liquids 

 at these temperatures. The values given for the specific volumes indicate 

 therefore the volume in cubic centimetres of equivalent weights in grams 

 of the respective liquids at their boiling-points. 



I. Phosphoryl Trichloride. 



Prepared by heating phosphorus pentachloride with phosphorus pent- 

 oxide. The product commenced to boil at 107°, the greater portion 

 distilling at 107 o, 5 (uncor.), under a barometric pressure of 755-2 millims. 

 Since this boiling-point is about 2°-5 lower than that usually assigned to 



* Ann. der. Chem. u. Pharm. Suppl. 1867-68, p. 140. 



