16 PROCEEDINGS OF THE AMERICAN ACADEMY. 



were present, and that the errors due to them are beyond the hmits 

 of observational error. 



Nine months after the completion of the experiments, the remaining 

 samples were subjected to a rough analysis by determining the boiling 

 points. If the liquid is pure, the boiling point should remain con- 

 stant until the liquid is completely boiled away. The amount by 

 which the boiling point changes during evaporation can be expected 

 to give at least some clue as to the amount of impurity present. [A 

 substance is considered good enough for most chemical purposes if it all 

 boils within 1°.] Of course this analysis does not give the nature of 

 the impurity present, nor does it give the impurity at the time the 

 experiment was performed. During the nine months the liquids 

 were left tightly corked, but there can be no question that some 

 deterioration had gone on in this time, so that the liquids were all 

 actually better than is indicated by the analyses. The deterioration 

 with time is much greater for some liquids than for others. 



The fractionations were performed by Mr. R. H. Patch at the 

 Chemical Laboratory of Harvard University. The results of his 

 examination are given below. It should be noticed that the tempera- 

 ture readings have not been corrected, so that all they can show is 

 the constancy of the boiling point, not its absolute value. 

 Methyl Alcohol (Kahlbaum). 



93% boils between 64.5° and 65° C. 

 7% " " 65° and 65.8° C. 



Free from ethyl alcohol and acetone, the most likely impurities, 

 and was neutral to litmus. Excellent sample. (It would seem 

 that the impurity was probably water, most of which had 

 probably been absorbed while standing, as the stopper was not 

 perfectly tight). 

 Ethyl Alcohol (Kahlbaum). 



3.8% boils between 77.3° and 77.8° C. 

 96.2% " " 77.8° and 78.0° C. 



The sample showed some suspended inorganic matter, prob- 

 ably iron from the container. (The sample had been standing 

 for the nine months tightly corked with a cork stopper in the 

 tin vessel in which it came from Kahlbaum). With this excep- 

 tion, and admitting the presence of a small quantity of water 

 the sample was a good one. 

 Propyl Alcohol (Kahlbaum). 

 4.% boils below 96.° C. 

 96.% boils between 96.0° and 96.8° C. 

 A good sample. 



