of Homologous Compounds. 7 



In many cases determinations by different observers are by 

 no means concordant, and some of the boiling-points quoted 

 are doubtless inaccurate. The selection of material in such 

 an investigation a< this is. indeed, a difficult matter, for if all 

 observed boiling-points, without exception, were included, 

 the errors would in some cases be so large as to obliterate all 

 signs of regularity ; while, on the other hand, the exclusion 

 of particular results is a dangerous and somewhat invidious 

 process. The general plan which has been adopted is as 

 follows : — 



(1) All boiling-points which I have myself observed are 

 included. 



(2) When an experimenter has made a series of determi- 

 nations, most of which are undoubtedly accurate, the whole 

 series has been included ; thus, all the observations of 

 Schumann and of Gartenmeister on the esters and those of 

 Dobriner on the ethers have been accepted, while the 

 boiling-points of the higher paraffins, as determined by 

 Krafft. have all been employed in calculating the constants 

 for the general formula. 



(3) When there are two or more fairly concordant obser- 

 vations of the boiling-point of a given liquid, and there is no 

 apparent reason for preferring any one of them, the mean 

 has been taken; but a determination by a worker of known 

 accuracy has always been adopted in preference to others. 



(4) When there are two or more discordant observations 

 they have all been rejected unless there is reason to consider 

 any one of them to be accurate. 



(5) If in any series the number of trustworthy observa- 

 tions is sufficient to leave no doubt as to the approximate 

 boiling-point of a given member, and the calculated value 

 differs widely from this, it has been rejected. Thus the 

 differences between the calculated and observed values of A 

 for the higher ethers are in most cases small, the mean of 

 those given in the tables being +0 o, 87. But the observed 

 boiling-point of ethyl hexyl ether (134° to 137°= 135°'5) is 

 31 "'1 lower than that of ethyl heptyl ether as determined by 

 Dobriner, while the calculated value of A for the temperature 

 135°'5 is only 2T°'0. The rise of 3 during the distillation of 

 ethyl hexyl oxide also shows that that ether was imperfectly 

 pnrified, and it therefore seemed justifiable to omit it from 

 the table. 



On the other hand, the boiling-point of di-octyl ether, 

 having been observed by Dobriner, has been included, 

 although it appears to be 2° or possibly 3° too low. 



