FEAGME]S[TAL VOLCANIC ROCKS— CONGLOMERATES. 75 



limited in weight to a few ounces and show a sorting or selection of sizes. 

 But in most cases the sizes have a much wider range. 



Geologists have been in the habit of distinguishing two classes of the 

 coarser fragmental beds. First, volcanic conglomerates ; second, volcanic 

 agglomerates or breccias. The conglomerates contain fragments more or 

 less rounded by attrition, which is held to be an indication that they have 

 been gathered together and arranged by the action of the water. The 

 breccias contain fragments which are angular and are presumed to have 

 been showered down around the vents from which they are supposed to 

 have been projected. Beds corresponding to both classes are abundant in 

 the High Plateaus and of very great thickness and area. But I am dis- 

 posed to accept the conclusion that they have all had a similar origin, and 

 that the projection of fragments from active vents and their descent in a 

 mitraille has had very little to do with their accumulation. As a rule, nearly 

 all of the fragments show comparatively little abrasion. Some, indeed, are 

 considerably worn ; most of them are very little rounded at the angles of 

 fracture, and a great pi'oportion are in a condition in which it is difficult to 

 say whether they have been abraded slightly or not at all ; for when 

 detached from the matrix the surfaces are corroded by some action which 

 may have been weathering pi-ior to their final burial or the solvent action 

 of percolating water after their burial and prior to the consolidation of the 

 stratum. None of the fragments exhibit the sharp edges formed by fresh 

 surfaces of fracture. Thus, while well rounded fragments (like those of 

 glacial drift or stream gravel) are uncommon, it is not certain that any 

 notable proportion have been absolutely free from attrition. The average 

 amount of attrition is generally small — far less than in conglomerates 

 usually occurring in a regular system of fossiliferous or stratified rocks. 

 No sharp distinction can be drawn between those beds of which the 

 included fragments exhibit a considerable amount of abrasion and those in 

 which no abrasion can be clearly proven. There is every degree of this 

 action and every shade of transition Thus it becomes impracticable to 

 draw any line here between conglomerates and breccias. 



It has seemed to me that the small amount of abrasion in the con- 

 glomerate fragments is susceptible of a partial explanation. The well- 



