44 G Kerr Grant : 



(3) In order that a spherical bubble of glass, vacuous, let 

 us suppose, within, should float in air, its thickness must not be 

 more than a certain fraction, approximately .00017, of its 

 radius. In order that it should not collapse under the air- 

 pressure its thickness, on the other hand, must be not less than 

 another definite fraction, approximately .00024, of its radius. 

 It is impossible, therefore, for a glass-bubble, vacuous within 

 and strong enough to withstand air-pressure, to float in air. 

 This conclusion I have been able to verify experimentally. 



Tlie case is obviously worse for the obsidian bubbles with 

 their heavy blebs attached, of Mr. Dunn's hypothesis ; unless, 

 indeed, Mr. Dunn imagines them to have been blown with 

 hj^drogeji or helium ! 



Although other objections may be brought against the 

 " Bubble-theory " of the origin of obsidianites, the above are, I 

 think, suflicient to show its extreme improbability; and since no 

 other plausible explanation of their mode of distribution has 

 yet been advanced by the advocates of the terrestrial theory, 

 we are driven to explain their occuyrence by means of a 

 meteoritic hypothesis. 



It has been objected that on this hypothesis such objects 

 would be found scattered over the whole land-surface of the 

 earth, and not confined to three comparatively small areas. 

 This objection would be at least equally valid against any vol- 

 canic theory of origin, unless it could be established that the 

 volcanoes of those parts of the world where obsidianites occur 

 bear a character distinct from those of the remainder. I am not 

 aware that anyone has attempted to show this. 



Nor is the objection at all dangerous to the meteoritic theory. 

 The virtual identity of chemical and physical properties in all 

 obsidianites, as also In the Moldavites and Billitonites, strongly 

 suggests, not merely a similar, but the same parentage ; pro- 

 duction, in fact, not by a long-continued succession of meteoric 

 falls, but in a single meteor shower. Tliis hypothesis woidd well 

 account for the confinement of each species to a relatively small 

 and well-defined area on the earth's surface ; and to push it a 

 step further it may be suggested as a possibility that the three 

 swarms above-named have been produced in successive returns 

 of the same meteor-shower. 



