598 CLARENCE N. FENNER 



the violent explosions, the frothy, semi-liquid pumice and the 

 dense, rigid andesites reacted differently and were forcibly torn 

 apart. The forms of many of the fragments of andesite are such 

 as to suggest fracture by the explosions. If the forces acting upon 

 them were of such magnitude as to produce fracture, it hardly 

 seems surprising that they became separated from the pumice. 



Pieces of obsidian, of ordinary maximum dimensions of three 

 to four inches, and of angular shape, were found everywhere in 

 the Katmai ash-fall within the area of coarse ejecta. Many 

 specimens were collected for study, and they furnish interesting 

 information. It is difficult to conceive any origin for them other 

 than the lava pool that gave rise to the pumice; and the presence 

 within them of pieces of pumice (which would inevitably float on 

 the lava) and their general nature suggest that they represent 

 chilled crusts on the surface of the lava. They contain quantities 

 of inclusions of various kinds: sediments, andesites, other obsidian, 

 pumice, and separate crystals. If it be granted that they are 

 fairly representative portions of the lava, the inclusions they con- 

 tain "frozen in" in all stages of disintegration are of great instruct- 

 ive value. 



Another material present in the ash-fall is evidently closely 

 akin to this conglomeratic obsidian. The matrix is semi-pumiceous 

 to glassy, and the numerous inclusions are of the same sort as 

 those in the obsidian. It is interpreted as a surface scum formed 

 over areas of seething lava. The inclusions in this also have been 

 caught in various stages of disintegration. 



From the evidence presented by these specimens, the absorption 

 of xenoHths by the magma seems to have taken place in several 

 somewhat different ways. The chemical composition of the frag- 

 ments and their porosity were probably important factors in the 

 matter. In some instances peripheral solution, especially of the 

 groundmass, seems to have been the principal process, or well- 

 defined tongues of lava may cut off portions and allow them to 

 float away. Probably a more usual form of attack is one involving 

 an intimate penetration of the whole mass. Several factors may 

 have been involved in this: first, that portion of the groundmass 

 of the rock that consolidated last of all and forms a binding-material 



