484 E. MALLET ON THE MECHANISM OF 



the joints lay, and whether the structure differed at different parts, 

 being amorphous and jointed into great blocks or having a more or 

 less slaty or shivered structure conferred upon it. 



I gave particular attention to the prevalent directions in which 

 elongated air-bubbles were found imbedded in each dyke. 



It is a fact now familiar to most geologists that the longest 

 axis of air-bubbles occurring in lava or other igneous rocks is 

 turned more or less exactly in the direction in which movement or 

 flow has taken place in the mass while in fusion, just as in most 

 igneous granites which contain large crystals of felspar or others of 

 a flattened form, the larger dimensions of these crystals indicate 

 the line of flow of the granite while in fusion. This is prominently 

 seen in the granite of Dartmoor in those quarries from which a 

 large portion of the pavements of the city of Exeter have been ob- 

 tained. It is obvious, therefore, that in a dyke the longest axes of 

 these air-bubbles should indicate the direction in which its molten 

 material has been forced into the cleft. If the injection had been 

 from below the lowest point of the dyke, and the flow into the cleft 

 had been therefore approximately vertical, as was perhaps the case 

 in many of the subaerial dykes of geologists, and, so far as I can 

 comprehend the not very clearly expressed views of Mr. Scrope, 

 also in volcanic dykes (Volcanoes, 1862), then it is obvious that 

 in these dykes of Somma the prevalent direction of the longest 

 axes of these bubbles should be approximately vertical. If, on the 

 other hand, the injection of these dykes has been lateral, and 

 though perhaps inclined upwards, still, on the whole, horizontal, as 

 being derived from the molten matter, more or less filling a crater, 

 from which the dykes diverged, then the longest axes of the air- 

 bubbles should be found nearly coincident with the plane of the 

 dyke, and if not horizontal, not very divergent by inclination 

 slightly upwards, or, with much less probability, still more slightly 

 downwards. Should the prevalent direction of the longest axes 

 prove transverse to the plane of the dyke, it would cast considerable 

 doubt upon the fact of the dyke having been filled by injected matter 

 at all — just as in lava streams which have flowed and consolidated on 

 gentle slopes, although, owing to intestine movements of the still 

 viscous mass, bubbles with their longest axes vertical are occasionally 

 found, the greatest preponderance of these axes is in the direction of 

 the flow, and were we to find the preponderant direction to be 

 vertical, it would cast much doubt upon whether there had been an 

 approximately horizontal flow at all. 



I also observed the characteristics of each dyke as indicated by 

 its included minerals. On this last point I may at once remark 

 that although it is generally and correctly stated by authors that the 

 distinguishing characteristic between the ancient lavas of Somma, 

 and the modern ones of Vesuvius is the great prevalence of leucite 

 in the former, the statement that the dykes of Somma are all of 

 leucitic lava is by no means correct. Leucite is, no doubt, the charac- 

 teristic mineral included in the larger number of these dykes ; but 

 pyroxene is also found in most of them, and in some is very preva- 



