704 SCIENCE PROGRESS 



in the direction of the z axis strings are found formed of 

 large and small spheres placed alternately in contact. A 

 considerable number of derivatives crystallise in a form corre- 

 sponding to one or other of the three theoretically possible 

 crystalline modifications of benzene, as will be shown later. 



The first two alternative arrangements of crystalline benzene 

 are conveniently distinguished according to their geometrical 

 derivation. Barlow and Pope show that the arrangement first 

 described corresponding to orthorhombic benzene can be 

 derived from the hexagonal close-packed assemblage of equal 

 spheres. This arrangement is said to possess hexagonal 

 marshalling. The arrangement of rhombohedral form can in 

 a similar manner be derived from the cubic close-packed 

 assemblage of equal spheres ; the structure is in this case said 

 to possess rhombohedral marshalling. 



The third variety of arrangement is also derived from 

 the cubic closest-packed arrangement but in a quite distinct 

 manner; it involves a considerable distortion of the benzene 

 unit not however amounting to the remarshalling of its atoms. 



The scientific worker is ever on the alert to discover 

 hidden relationships between different objects, whatever may 

 be his sphere of work. To the chemist or the physicist the 

 properties of a substance are not of such great interest in 

 themselves but there is a great fascination in comparing them 

 with those of some other substance, whether related to it or 

 not. In the same way, although it is a great achievement to 

 have solved the problem of the internal structure of crystalline 

 benzene, it is of far more absorbing interest to trace this 

 crystalline form or the basis of it through a whole series of 

 benzene substitution products. Work of this kind, the way 

 having been pointed out by Barlow and Pope, has been 

 carried on by other observers. 



It may be inquired, how are the structural relationships 

 between benzene and its substitution products to be recognised ? 

 In the first place, we know that the dimensions of the 

 hexagonal mode of marshalling benzene units and of the 

 pseudo-cubic mode are both practically given by x '.y \ z ■=^ 

 3"ioi : 3"48o : 2780. Now it is conceivable that in a substitution 

 product only one or two of these dimensions should be 

 changed and consequently in the dimensions of the derived 

 assemblage we might expect to find one or two of these 



