204 
fall towards the earth, and either the act of coming in contact with the- 
roeks or the differences in temperature would shatter the frail glass bubbles, 
but the bleb which had formed at the bottom would remain. Chips would: 
\ 
\ 
\ 
\ 
\ 
\ 
\ 
\ 
\ 
\ 
\ 
1 
I 
I 
/ 
/ 
/ 
/ 
/ 
/ 
/ 
/ 
/ 
/ 
Fig. 45. 
Section of a hollow obsidian button showing its relation to the original bubble 
(dotted circle). 
fly off on changes of temperature just as .they do from slag or unannealed 
glass, and this has taken place. The buttons are the blebs of single 
bubbles, hence their symmetry. Around the edges of these the rim, when 
the glass bubbles break away, is often plainly visible. The examples con¬ 
tracted in the middle are the blebs of double bubbles, hence their 
peculiar form, and around the edge of these the rim where the bubble was 
firmly joined to them is apparent in some cases. The stopper-like bodies are 
blebs of a heavier class, and around their edges the line where the bubbles 
joined them appears to have chipped off. In the case of the remarkable 
hollow sphere previously referred to, it is the bleb of a large glass bubble, 
and the step around the periphery marks the. line where the bubble joined 
this hollow bleb. The dispersal in greater abundance over some tracts 
would indicate the prevalent direction of the wind as well as the direction 
from which they were derived. These bubbles were undoubtedly blown 
in volcanoes. Whether these volcanoes are the points of eruption so com¬ 
mon over the southern portion of Victoria or not. has yet to be determined. 
The analyses made of them show over 70 pe.r cent.; of silica, and in this 
respect they differ widely from the basic lava flows, but the bubbles may 
have preceded the lavas. Examples are found that are waterworn in the 
alluvial drifts of Victoria, and that are sandworn at Coolgardie, Western- 
