96 Dr. H. Herwig's Investigations on the Conformity 



of the measurements), when the chloroform was not yet protected 

 by the yellow glass, that a small trace had been already decom- 

 posed. This, however, could not make a greater difference in the 

 weight than 1*5 milligrm. The examination of bisulphide of car- 

 bon, which will be subsequently described, gave a similar result, 

 where the traces of the sulphur which might be separated during 

 the boiling out and sealing of the bursting bulb also could not 

 have produced the difference of weight necessary in order to bring 

 the vapour-densities actually found into accordance with the theo- 

 retical ones. That in both these cases no error can lurk in the 

 method which would induce the differences may be indubitably 

 recognized from this, viz. that at each temperature the final 

 vapour- densities for the most various volumes are, within the 

 limits of errors of observation, exactly proportional to the final 

 vapour-densities at all other temperatures. Besides, in general 

 the experimental determinations of the vapour-densities do not 

 rigorously lead to the theoretical densities. Even though many 

 of the old determinations could not give any exact results because 

 no attention was paid to the question whether the vapours were 

 sufficiently far from their condensation, yet deviations from these 

 causes must always give only an increase in the vapour-density 

 over the theoretical values, while a converse course of determina- 

 tions would furnish equally important smaller values. 



Now, as to the relation holding for vapour of chloroform cor- 



py 



responding to that found for alcohol, viz. =c*/a + t,l 



first of all conjectured that, even if the like holds here also, 

 the constant c might perhaps be different from that found to be 

 valid for chloroform, in such sort that the temperature at which 

 the perfectly saturated vapour follows Mariotte's law might, 

 for chloroform, lie as much under 9°*5 as the boiling-point of 

 chloroform under atmospheric pressure lies under the boiliug- 

 point of alcohol. Meanwhile the first set of experiments showed 

 decisively that this was not the case; on the other hand, the 

 surprising result presented itself, that in the admittedly valid 

 PV 



formula = c Va + t the constant c had the same value as for 



alcohol. In what follows I give the Table of v i calculated from 

 the specified relation with c="0595, and at the same time, as for 

 alcohol, the extreme limits of v x and c Y which are consistent with 

 the observations. In this case I have sought to observe more 

 accurately the exact point of retreat of the vapour from the state 

 of maximum tension. I must remark that at the last tempera- 

 ture (6i° - 8) the apparatus unfortunately did not sustain the 

 compression which was necessary in order to arrive at the state 

 of maximum tension. The only observations that I could make 



