150 



University of California. 



I vol. 3 



diameters evidently increase in proportion as the original 

 temperature was high and the rate of cooling slow, a rule which 

 appears to hold generally in the development of columnar 

 structure in lavas. 



The short fragments of these columns brought to the labora- 

 tory are of hexagonal or pentagonal cross-section, the prism 

 faces of each column being generally in part slightly concave, 

 in part slightly convex; in one pentagonal column, for example, 

 two adjacent faces are convex, and the other three concave. 

 When a column is broken transversely at any point, a peculiar 

 phenomenon is observed; one of the surfaces of fracture always 

 has in the centre a circular, button-like protuberance, the other, 

 of course, a corresponding saucer-like depression. The direction 

 of convexity is always the same in a given column, but its 

 relation to the dyke was not noted. This phenomenon is signifi- 

 cantly analagous to the transverse "ball and socket" jointing 

 often developed in the columns of basalt. 



The material of these tuff-columns is somewhat harder and 

 more compact than the rock of the same beds beyond the zone of 

 contact-modification. Under the microscope, however, no con- 

 siderable modification of the internal structure is to be observed. 



From the sandy and gravelly beds locally found near the top 

 of the John Day section two types of rock are represented in our 

 collection. 



One is a fine-grained conglomerate, composed of well rounded 

 bits of compact buff tuff like that most characteristic of the 

 Upper John Day. 



The other is a sandstone varying in texture from coarse to 

 moderately fine, and of a deep green color. The grains macro- 

 scopically may be seen to be mainly feldspar, but small fragments 

 of rock are also visible. The color is due to the abundant 

 cement, which consists mainly of a compact deep green sub- 

 stance, though the bright cleavage faces of a zeolite are 

 occasionally seen. 



Under the microscope the grains are found to be sub-angular 

 in form. The feldspars, when determinable, are found to be 

 plagioelase with the average composition of andesiue; although 

 orthoelase may be present its presence was not demonstrated. 



