16 A. L. Du Toit—Pipe- Amygdaloids. 
Melaphyre of Darmstadt by Kuppers,! Klem,? Wittich,? and others 
under the names of ‘‘separation-bodies, contraction-cylinders, and 
bubble-trains.”” The lava being rather weathered, the cylinders break 
out from the melaphyre, but sometimes the continuity of the material 
still exists in places. Like the Barkly West examples these 
structures are only found in the lower part of the upper of two 
lava-flows, the rock above passing into a normal amygdaloid. 
Wittich explains them on the supposition that the upper portion 
of the second flow became chilled rather rapidly, so that the bubbles 
formed from the vapour given out in the lower portion were unable 
to escape. Kuppers, on the other hand, considers them as due to 
contraction during solidification. Neither of these views seems to 
account satisfactorily for the Darmstadt examples, nor will either 
of them hold in the case of the Barkly West specimens, occurring 
as they do in a rock which is otherwise devoid of amygdales. 
I think there can be little doubt that the gas which filled the 
vesicles in these structures has not come from the lava-flow itself, 
but has been derived from cavities and fissures in the rock underlying. 
The lava being in most cases too mobile for the successful production 
of pipe-amygdales, the bubbles have ascended from the bottom in 
continuous rows and thus formed ‘bubble-trains.’ Vertical sheets 
of amygdaloidal rock in the diabase probably indicate that the gas has 
been emitted from a long fissure, while its evolution over a wider area 
would account for the masses of vesicular material of irregular outline 
and several square feet in section occasionally seen. It is perhaps 
even possible that with a great irruption of gas from below, a lava-flow 
that would otherwise have been compact might in great part become 
converted into amygdaloidal rock. 
The gas may have consisted either of the air entangled in the lower 
lava-flow or of steam formed by the vapourising of water contained in 
it or else overlying it. In the case of the Ventersdorp lavas there 
is not sufficient evidence to show whether they were subaqueous 
in character, although certain interbedded quartzites point in that 
direction. The Stormberg lavas, however, contain towards their base 
intercalations of sandstone, shale, and volcanic ash, and the earlier 
flows have undoubtedly been erupted subaqueously. 
Professor Schwarz‘ has attempted an explanation of the pipe- 
amygdales based upon Wittich’s views, but I think that there is more | 
likelihood of their having been formed by the steam generated in the 
flowing of the molten rock over moist surfaces; this is the theory 
adopted to explain the Indian examples.’ 
As Cohen pointed out, the condition essential to the formation of 
these elongated amygdales was that the lava should be sufficiently 
viscid. In thin sections of the amygdaloids the vesicular cavities 
rarely show evidence of having originated prior to the crystallization 
1 Kuppers: Centralblatt fiir Min. Geol. u. Pal., 1901, pp. 481,609; 1902, p. 521; 
1903, p. 409. 
2 Klem: ibid., 1903, p. 217. 
3 Wittich: Tschermak’s Min. u. Petrogr. Mitt., vol. xxi (1902), p. 185. 
* Schwarz: Ann. Rep. Geol. Commission for 1902, p. 57. 
5 Medlicott & Blanford: ‘‘ Geology of India,’’ 1893, p. 261. 
