Obsidian ites. 445 



as fatal to an explanation of their prodmtion by volcanic ajiency. 

 This has occasioned Mr. E. J. Dunn to put forward an hypothesis 

 as to their mode of distribution from volcanoes.' Mr. Dunn su*;- 

 gests that the Obsidianites are the " blebs of obsidian bubl)les.'' 

 Apart from the «reneral ariruments against a terrestrial origin 

 which have been put by Walcott, Suess, and other writers, there 

 are certain others special to Mr. Dium's hypothesis which appear 

 to me to render it quite untenable. 



(1) Tlie forms of the obsidian buttons are not. with a few 

 possible exceptions, those which a liquid drop assumes when 

 hanging from a bubble. The formation of the frequently occur- 

 ring '■ dumb-bell " type by the union of two separate bubbles 

 is quite inadmissiljle. If the l)ubbles, and a fortiori the blebs, 

 were perfectly liquid, the two latter would certainly coalesce to 

 a. drop of circular horizontal section : if they were not perfectly 

 liquid we should expect signs of discontinuity at the junction : 

 such do not occur. Mr. Dunn considers that the form figured by 

 him in Fig. 45 of his paper was attached to the original support- 

 ing bub])le around the projecting rim. There are, however, no 

 signs of fracture around the rim such as we would expect had 

 the obsidianite broken away from the parent bubble, nor have 

 I been able to find such signs in any one of the numerous other 

 specimens with rims which I have carefully examined. The 

 specimen figured by Mr. Dunn is one of very nnustial type, and 

 the connnoner forms will not afi'ord even its frail support to 

 his ingenious hypothesis. For instance, in Professor Spencers 

 Central Australian collection are many specimens without the 

 bubble-suggesting rim at all. 



(2) A more conclusive objection is the following: — 



The pressure within a liquid Inibble is determined by the total 

 curvature of its inner surface (and, of course, the surface-tensicjii 

 of the liquid). It is evident, without exact investigation, that 

 one part of the interior cannot be convex while another is concave. 

 Hut as both upper and lower surfaces of the obsidianites are in- 

 variably convex, it is obviotis that the attachment of the "bleb" 

 to the bubble in the way imagined 1)y Mr. Dunn is a physical 

 impossibility. 



1 Rec. Geol. Survey of ^■i^tOl■i.1, vol. ii., pt. I\'. 



