AND SPHEROIDAL STRUCTURE. 



147 



horizontal curving joints with their convexities upward ; so that the 

 whole is divided into a series of plano-convex, meniscoid, or 

 concavo-convex blocks, in length perhaps about 2 to 4 feet, and in 

 thickness 4 to 12 inches (fig. 6). This structure continues with but 

 slight indications of any tendency to vertical master joints for some 

 35 feet ; and then the basalt for the remaining space (perhaps about 

 4 yards) becomes rather suddenly rudely columnar*. How much 

 has been removed by denudation from the top of the flow it 

 is impossible to say. At the end of the spur there is evidence 

 that its surface was uneven and a ridge already defined when 

 the basalt flowed; for the tabular structure curves in both the 

 longitudinal and transverse sections, so as to remain symmetrical with 

 the surface of the ground. It is therefore evident that it is con- 

 nected with the form of the surface of the rock. Close to the 

 station of Eougeac (railway from Le Puy to Arvant) I saw a similar 

 structure in a mass of basalt. A very rough sketch of this in- 

 teresting section (jotted down from the railway -carriage window) 

 is given below (fig. 7). 



Fig. 7. — Section at Eougeac . 



A B C D 



A. Basalt. B. Stratified volcanic ash. C. Aggloineratic volcanic ash. 

 D. Basalt, with curvitabular structure. 



Here also the structure is affected by the form of the bounding 

 surface ; but in this case the convexities of the curves are turned 

 downward. 



(d) Cup-and-ball Structure. — The next structure to be noticed 

 is that form of cross-jointing of columnar rock which goes by the 

 name of cup-and-ball structure. Usually the columns are divided 

 across, at variable distances, by plane cross joints at right angles to 

 their axes. These sometimes, as in the Gross Oelberg and Weilberg 

 (Siebengebirge), and at Murat (De'partement de Cantal, Auvergne), 

 are as much as 15 or 20 feet apart ; but commonly they are not more 

 than a yard, and sometimes less. Occasionally, however, the joint- 

 surface is not a plane, but more or less curved, so that the convex 

 extremity of one segment fits into the concave extremity of the next. 

 The form of this curve varies : sometimes its curvature is but slight, 

 and it is continued down to the sides of the prism ; at others the 

 curvature is more marked, the outline of the ball being more nearly 

 circular, and is not continued to the sides of the prism, a flat space 



* Mr. Scrope (p. 106) describes this basalt as separated at some points into 

 very regular prisms of five or six sides, which exfoliate by decomposition in 

 slaty lamina; at right angles to their axes. I do not remember to have seen 

 this structure conspicuous in the part which I examined. 



