LAND CONNECTION BETWEEN AFRICA AND AMERICA 87 



they are derived from the fusion of pre-existing rocks, sedimen- 

 tary and otherwise, which have been subjected to dynamical stress. 



I have spoken hitherto mostly of the large fragments torn off 

 from the throat of the volcano, but the inclusions in the Nightin- 

 gale rock are more of the nature of fine dust. Such volcanic dust 

 aggregated into beds would be known as a tuff, but it has not been 

 generally recognized till recently that such tuffs may be built up 

 entirely of non-volcanic materials. In the peperino of Italy we 

 have a leucite tuff with many inclusions of non- volcanic rocks, 

 such as limestone, sometimes in large blocks, at others in the form 

 of fine dust. In the cave sandstone of the Drakensberg we have 

 a tuff which is composed entirely of non-volcanic dust, grains of 

 quartz, microcline, plagioclase, tourmaline, epidote, etc., with only 

 an occasional ash-bed. I suspected its origin, when first studying 

 it in Matatiele, from the fact that it was found thickest near the 

 volcanic vents. Its nature, also, showed clearly that it was not an 

 ordinary sediment; for there are beds of this rock 800 feet thick 

 without a trace of stratification. Sometimes there is stratification; 

 either a plane passes slantingly through the mass from top to bot- 

 tom, or it is curved and twisted as if the whole material had been 

 stirred about in a gigantic pot. The structure of the rock was 

 such that I finally decided that it must have been ejected as a liquid 

 mud exactly like the Italian peperino, and that the constituent 

 grains had been blown from the substratum of granite and Archaean 

 schists by explosions. • Von Knebel has described similar tuffs com- 

 posed of disrupted granite from the Ries in Germany.^ 



In 1845 Ehrenberg described a remarkable tuff from the Island 

 of Ascension containing infusoria, which he called a pyrobiolith." 

 Prestwich surmised that these "diatoms" had lived in subterra- 

 nean caverns which the explosion had traversed, ^ but the Chal- 

 lenger Reports state that the organic remains are the siliceous par- 

 ticles of grasses. It seems probable that this tuff was erupted as 

 a boiling liquid mud, which, traversing a plain covered with grass, 

 tore off by rolling action and incorporated grass fragments in the 



1 Zeitschrift der Deutschen Geologischen Gesellschaft, Vol. LV (1903), pp. 236-Q5. 



2 Berichte der Koniglichen Akademie der Wissenschaften (Berlin, 1845), p. 140. 



3 Proceedings of the Royal Society, 1886, No. 246, p. 156. 



