Ohsidii mites. 44 1 



The artifii'iul origin of obsidianites is shown to be chemically 

 impossible. 



Analyses of rocks from the various places named as possible 

 sources of the obsidianites, are compared with the analyses of 

 the obsidianites and it is shown that there is nothing; in com- 

 mon between them. 



It is pointed out that there is an apparent provincial dis- 

 tribution of obsidianites, and if this is proved to be correct, it is 

 shown that a cosmic origin is the only possible one. 



Analyses of billitonites are given and classified, and shown 

 to be geuetically connected with the obsidianites. 



Six analyses of moldavites, when classified, are found to have 

 few representatives among terrestrial rocks, and the argument 

 in favour of a cosmic origin for them is strengthened. 



APPENDIX. 



The following additional information has come to hand since 

 the above paper w^as read : — 



yew Zealand. — In answer to a letter inquiring about the 

 occurrence of obsidianites in New Zealand, Dr. Marshall wrote : 

 — •'■ I think I can say without any qualification that there is no 

 record whatever of the occurrence of such objects in New 

 Zealand. Of course you are aware that obsidian, as a rock, 

 occurs at many localities, notably at Mayor Island, Rotorua, and 

 near Whangaroa, but even in these districts I have seen no 

 obsidian bombs, to say nothing of obsidianites." 



Queensland. — Mr. Dunstan. Government Geologist of Queens- 

 land, informed me that they had often inquired about the occur- 

 rence of obsidian bcAnbs in Queensland, but could get no 

 specimens, and, further, had not heard of any being found. 



Xew South Wales. — Mr. Card called my attention to the 

 record of obsidianites in the Records of the Geological Survey 

 of New South Wales, Vol. VII., Pt. III., p. 218. Four specimens 

 are figured. At the same time he told me that Mr. Mingaye 

 was not satisfied with the analysis of the Uralla obsidianite, as 

 he hiid very little material to work on, and intended analvsing 

 another specimen from the same locality. 



