RESEARCHES ON THE VOLATILE HYDROCARBONS. 161 
shall slightly incline towards the body of the retort. If necessary, some pieces of ice, 
which will adhere firmly to the cloth, may be laid along the neck to insure complete 
condensation of the vapors during ebullition. While the retort is in this position, 
ebullition is continued for considerable time, until it ceases to have any effect on the 
height of the mercury in the thermometer. The lamp being now removed for the 
moment, the neck of the retort is turned down, and quickly connected with a Liebig 
condenser. The lamp being now replaced, the distillation is commenced. So soon as 
the mercury in the thermometer shall have become constant, which will now occupy 
but a few seconds, the temperatures by the retort thermometer and the side ther- 
mometer are carefully noted, and also the time at which these observations are made. 
During the distillation, which is continued nearly to dryness, the readings of the ther- 
mometers and of the watch are noted at regular intervals, or so often as any appre- 
ciable variation of the retort thermometer shall have taken place. The average of the 
several observations, or of those corresponding to the longer intervals of time, apply- 
ing the corrections for atmospheric pressure and for the upper column of mercury, 
according to Kopp,* is taken for the true boiling-point. I have generally obtained 
the hydrocarbons so pure that the whole quantity operated upon would distil within 
the range of 1° of temperature, and not unfrequently within 0°.5. In a few. cases, 
however, when the quantity of material at command would not permit of the attain- 
ment of so high a degree of purity, the distillation would range over two or three 
degrees ; in such cases I have generally taken the average of the temperatures corre- 
sponding to the longest interval of time, as probably representing more nearly the 
true boiling-point of the body. In stating my results, however, I shall give the limits 
of temperature within which the distillation was effected. The thermometers em- 
ployed in the determinations were the best that I could obtain from Fastré of Paris ; 
for the temperatures below 100° the instrument used was calibrated, and the scale 
divided into fifths of a degree. The determinations above 100° were all made with 
one thermometer. 
The method just described differs in some respects from that of Kopp. He objects 
to the practice of taking boiling-points with the thermometer bulb immersed in the 
liquid,+ on the ground that the thermometer in this condition hardly ever indicates a 
* Poggendorff’s Annalen, 1847, LXXII. 38. 
+ “Liisst man die Kugel des Thermometers in die siedende Fliissigkeit tauchen, so zeigt fast nie das Instru- 
ment eine constante Temperatur, sondern das Ende des Quecksilberfadens ist in stetter hiipfender oder 
zittender Bewegung ; der auf diese Art gefundene Siedepunkt kann nur mehrere Grade hoher liegen, als der, 
welcher gefunden wird, wenn sich die Kugel des Thermometers in dem Dampf der siedendén Fiiissigkeit 
befindet.” 
VOL. IX. 25 
