280 I'. K. WRHiirr UEhllDlAiS FKOM ICELAND 



(3) The matter of internal strain noted above is significant for two 

 reasons: (a) its directive influence on tlie etching solutions, and (h) tlie 

 light which it throws on the former history of the fragment under exami- 

 nation. In the case of a large mass of obsidian, the internal strains set 

 up on very slow cooling are virtually compensated at the period of their 

 formation, so that the chilled product is a remarkably well annealed mass 

 of glass far superior in this respect to the best optical glass. 



Strain birefringence in the large fragments of the obsidian (specimen 

 88438) is hardly detectable even in the largest splinters, which are suffi- 

 ciently transparent for observation. Around the radial spherulites (speci- 

 mens 88433 and 88433) no strain birefringence could be detected in the 

 adjacent glass in the thin section. This proves a very perfect state of 

 annealing. On the other hand, the small etched fragments of obsidian 

 in specimen 88435, which were found near the pond noted above and at 

 some distance from the main obsidian mass, are in a state of severe strain. 

 Small splinters of these fragments show gray interference colors and 

 uneven distribution of the regions of differential compression and ten- 

 sion. The strain in these fragments is apparently even greater than that 

 produced on heating a small splinter of the annealed obsidian in a Bun- 

 seu burner to a temperature of 1,000° or 1,200° ('. and then quenching it 

 in water. This proves clearly that the fragments are not simply frag- 

 ments of the annealed obsidian mass which have been broken off and 

 transported to their position and there etched, but that they were cooled 

 very rapidly from a high temperature to relatively low temperatures. 

 Tlie natural inference is that they are shattered ejectamenta of the ob- 

 sidian magma after the manner of the bombs of less siliceous map-mas, 

 or that they represent fragments of the outer chilled crust of the obsidian 

 magma. The distribution of the strain phenomena indicates the first 

 rather than tlic second inference. On tlie assumption that these frag- 

 ments represent bombs, the irregular rupture shrinkage cracks are readily 

 explained. A study of the types of volcanic bombs ejected by rhyolitic 

 magmas and of the distribution of strain in them would be of interest in 

 this connection. It is also important to note that on heating the obsidian 

 in the Bunsen burner the obsidian tends to evolve gas bubbles and thus 

 to become pumiceous. The tendency toward pumiceous development in 

 some of the etched specimens has been noted above. 



THE MOLDAVITES 



In view of the great similarity between the etched surfaces of obsidian 

 fragments at Hrafntinnuhryggur and those of the tektites, especially the 

 nioldavites of Bohemia, which have been considered to be of extraterres- 



