68 



SCIENCE PROGRESS 



39, 23) ; it will suffice for present purposes to say that they very 

 closely resemble in general appearance the parts NQRS of Fig. 6. 

 The dispersion at the minimum seems, therefore, to be positive 

 just as it is at the maximum, whence it appears that the disper- 

 sion may maintain its character over a range of temperature 

 which includes both a maximum and a minimum. But although 

 intersection apparently does not occur between the maximum at 

 H and the minimum at S (Fig. 6), it might be that the T-R 

 curves would cut each other on the farther side of the minimum, 

 as is indicated in Fig. 5. This question, as well as the previous 



Fig. 5. 



one, was therefore investigated by an examination of the T-R 

 curves in the case of homogeneous zsobut}^ diacetyltartrate, 

 which exhibits, like the di-trichloracetyl derivative, a minimum 

 at medium temperatures — a minimum which presumably corres- 

 ponds to that found in the curves for ethyl tartrate dissolved in 

 quinoline, since the dispersion in that region in both cases is of 

 the same character. Examination of the rotations of this ester 

 from — 20° to 4- 200° for six colours of light {J.C.S., 191 6, 109, 

 1 1 55) revealed a distinct minimum in each of the curves, which, 

 like those for eth^d tartrate dissolved in quinoline, show no 

 sign of intersection before the minimum is reached, and it is 

 also clear that after passing the minimum value they again 



