70 PROCEEDINGS OF THE AMERICAN ACADEMY 



determination, — so desirable when dealing with mixtures, like 

 those of the hydrocarbons from petroleum, that cannot be made to 

 yield, by the common process, liquids having a constant boiling- 

 point, nor, properly considered, a very near approach to one, but 

 only a pretty long " boiling-space." 



In the present experiment the readings of the thermometer were 

 as follows : — 



Observed temperature of the liquid at commencement of distillation, 29°.5 



8 minutes later, 29°. 7 



10 " 29°.9 



21 " 30°. 1 



5 " 30°.3 



At the latter temperature the liquid had distilled almost to dry- 

 ness; and at 31°, six minutes later, not a drop of liquid remained 

 in the retort. The duration of the experiment was 50 minutes, 

 and the range of distilling temperature 1°.5, or 0°.8, if we omit 

 the last, and doubtless least reliable, observation. As the ther- 

 mometer was doubtless directly influenced by the superheated re- 

 tort-bottom during the distillation of the last few drops of liquid, 

 the last observation is considered of no value except as furnishing 

 additional evidence of the high degree of purity of the substance, 

 i. e. for so volatile a body of this class, obtained by fractional dis- 

 tillation. That it should be otherwise disregarded is shown by 

 the fact that the bulk of the liquid passed over with so slight a 

 variation of temperature, viz. only 0°.4, from 29°. 7 to 30°. 1, dur- 

 ing the long interval of 31 minutes. The average of the first five 

 observations, or 29°. 9, is therefore taken as the observed boiling- 

 point; which, corrected for pressure and the upper mercurial col- 

 umn, becomes 30°. 1, being almost identical with the boiling-point 

 that Frank land originally found for this body. The latter circum- 

 stance, with other corroborative evidence, such as the fact that my 

 analyses gave almost invariably a somewhat larger percentage of 

 hydrogen for the members of my 1st series than for the correspond- 

 ing terms of the 2d (which may prove, however, merely accidental), 

 has led me, while somewhat in doubt on this point, to regard this 

 body and its homologues of the 1st series as the true hydrides of 

 the alcohol radicals, or marsh-gas series; and the 2d series, pro- 

 visionally, as the iso-hydrides of these radicals, — a term which 

 may be appropriately retained if further examination shall furnish 

 no stronger evidence of difference in constitution. This view is 



