A SPHERULITIC DOLERlTb] FROM VRYHEID, NATAL. o8o 



is a diagrammatic sketch, to actual size, of a portion of the 

 weathered surface of the lower sill, shown in PI. XXVIII. 



It will be observed that groups of radial lines start from 

 numerous centres distributed over the surface, and that in 

 genei-al the different groups are bounded by lines which are 

 sensibly straight, so that the whole surface is divided up into 

 a number of polygonal areas. The avei'age diameter of these 

 areas is 4 cm., but in some places they are as much as 8 cm. 

 across. The several centres are by no means equidistant 

 from one another, but, wherever two groups of radial lines 

 meet in a straight line, this line is midway between the 

 centres of the groups. The irregular distribution of the 

 centres thus gives rise to a great variety of shape in the 

 polygonal areas. On the specimen from the lower sill some 

 of these bounding lines are seen to be curved : such lines are 

 always nearer to one of the centres than to the other, and are 

 concave towards the nearer centre. The radial structure is 

 developed on all weathered surfaces of the specimens, and is 

 not confined to any one plane. Patches apparently free from 

 radial structure (represented by the dotted areas in the figure), 

 are still bounded by sharp lines, and can be shown to be 

 cross-sections of rays starting from centres in the body of the 

 rock. 



The structure is shown equally well on pieces of both sills, 

 but in the upper one the centres tend to be farther apart. 

 In one case two centres areas much as 10 cm. apart, and here 

 the line dividing the two areas is very indistinct ; the rays 

 seem to diminish in strength as they recede from the centre 

 and at 3'5 cm. from it they are only faintly marked. 



In addition to the radial lines brought out by weathering, 

 incipient cracks are often seen starting from the centres. 

 These are usually three in number, and are equally inclined 

 to each other, but they rarely extend for more than 0*5 cm. 

 from the centre. These cracks were well shown in a piece of 

 the rock which Dr. Warren had cut through and polished; they 

 evidently occur throughout the rock, and are intimately 

 connected with the radial arranofement. 



