98 Journal of the Mitchell Society. [Nov. 



that if a lower mean value were adopted the errors would be 

 shifted toward the end points of the curve. 



The author would here remark that a theory by Crompton 

 to which attention will be called in a following- paper, seems 

 to bear out this suggestion, but also points more strongly to 

 the fact that di-isobutyl does not give the results that similar 

 substances would lead one to expect. 



Chlor-benzene, 240° to 270° inclusive, may easily be due to 



8 P 

 the wrong trend «-=- of the Biot formula at this, its end 



point. 



Brom-benzene at 30° C. The error here is due to the Biot 

 formula, as shown by the fact that at 100° C the constant, 

 from a theoretically calculated density is correct. 



At 160° and 170° C the error may be due to the measure- 

 ment of the vapor density. Prof. Young writes: "The 

 observed volumes of saturated vapor are generally much less 

 accurate at low temperatures than at high, because by the 

 method employed, a given error in reading would have much 

 greater influence at the lower temperature." Also regarding 

 brom-benzene Prof. Young says: "After heating in ordinary 

 daylight at 180°, 190°, 200°, 'brom-benzene slightly acted on 

 by mercury, small quantity of solid being formed, chiefly in 

 form of minute needle-shaped crystals.' In a second series 

 the brom-benzene was 'carefully shielded from the light, no 

 crystals formed.' The volumes were read in the first series 

 only — the quantity of liquid in the second being small — and 

 the results do not seem to me so satisfactory as with the 

 other substances examined." 



Iodo-benzene at 190° to 210° C, inclusive, may be due to 

 measurement of vapor density. Dr. Young writes: "In the 

 case of iodo-benzene similar crystals (to those with brom-ben- 

 zene) were found, although the substance was shielded from 

 daylight. All readings had to be taken by gas light and are 

 therefore less accurate than usual." 



The tendency towards continually increasing values of the 



