34 G. K. GILBERT DOMES AND DOME STRUCTURE 



through a process of peeling. One layer at a time is carried away, and 

 the surface at each stage coincides approximately with one of the 

 partings. 



Whatever the cause of the dilatation producing the partings, they are 

 formed in succession from without inward. For each one the deter- 

 mining strains are themselves conditioned not only by the form of the 

 outer surface, but by the form of the last made parting. Parallelism is 

 not perfect but approximate, and the departures from strict parallelism 

 are of such nature as to reduce or omit angles and other features of 

 irregularity. The inner partings reflect only the general features of the 

 external sculpture. As peeling progresses and the zone of competent 

 strain moves inward, the outer surfaces are successively more and more 

 simple in contour, and the newly developed partings are endowed with 

 still greater simplicity. 



Opposed to the rounding process is corrasion. The attrition of a 

 detritus-armed stream or glacier saws through the rock plates with little 

 regard for the presence or absence of partings. By so doing it creates 

 discordant elements of topography and modifies the conditions under 

 which the expansive strains are developed. In the Sierra the effects of 

 glacial corrasion are at present conspicuous. By the corrasion of the 

 Tenaya trough the base of Half dome was sapped, so that a part was 

 sheared off by gravity, producing a vertical flat face (figure 1), in which 

 the structureless nucleus was exposed. In this face the " dome struct- 

 ure " was developed, but, being conditioned by a plane outer surface, 

 the new partings are plane (except at the edges), and thus simulate 

 ordinary plane joints. 



