2i8 THE GEOLOGY OF BRITISH EAST AFRICA part hi 



is not inconsistent with the ordinary type of volcanic action. 

 Moreover, beds of volcanic tuff and ashes are abundant, and all 

 the evidence points to the former occurrence of a vast number 

 of small craters scattered over the area. Later on, when 

 marching across Laikipia, we passed over another lava plain of 

 more recent date, and there found dozens of the denuded stumps 

 of small craters. That craters may also have existed on the 

 Snake River plains is not improbable, for several facts sug- 

 gested that these were older than I had inferred from previous 

 descriptions. 



Hence, as the study of these African lava plains revealed 

 abundance of fragmentary materials due to explosive eruptions, 

 but none of the fissures, I was gradually driven to abandon 

 Richthofen's theory in favour of one that seems better to agree 

 with the conditions seen in the field. This may be explained 

 by reference to the accompanying diagram (Fig. 8). The 

 upper figure represents an ideal section across an ordinary 

 volcanic crater in a region of folded and contorted rocks. In 

 such a case lines of weakness will traverse the country along 

 the axes of the folds. If beneath the area there be a mass 

 of rock, either molten or ready to melt when the pressure 

 is relieved, then this will force its way to the surface through a 

 vent, and by a series of explosive eruptions build up a volcanic 

 cone. According to the fissure-eruption hypothesis, the strain of 

 this subjacent lava is supposed to tear open a crack or fissure 

 instead of a single short vent ; and, owing to the great size 

 of the aperture, the rock quietly wells forth without explosions. 



According to the hypothesis suggested in the lower of the 

 three figures, the lava sheets may be produced without the 

 assumption of fissures. Wherever these great lava plains are 

 known, they occur on high plateaux composed of rocks which 

 either retain their originally horizontal position, or are of remark- 

 able uniformity in composition. If such a plateau occurs over 

 a great subterranean lava reservoir, then, as the contents of this 

 expand, the rocks above will be subjected to tension in all 

 directions. The plateau will therefore be traversed by a double 

 series of lines of weakness crossing one another like a network. 

 The intersections of these lines will be points of equal weakness, 

 and when the pressure from below is sufficient to force an 

 opening to the surface by one of these, many others will give 



