THE NATURAL HISTORY OF IGNEOUS ROCKS. 23 



them with the Lower Palaeozoic eruptions, as apparently 

 Bertrand would do, they assume a different significance. In 

 the former view they afford, as Geikie has remarked, a 

 striking parallel with the great system of Tertiary dykes in 

 the Inner Hebrides, the south of Scotland, and the north 

 of Ireland. These dykes, it may be observed, are basic 

 dykes, and in well-known cases, such as the lava-plains of 

 the Snake River and the " Deccan traps," which seem to 

 point to the tranquil welling out of large quantities of lava, 

 these lavas are basaltic. Something of the same tendency 

 may be observed in some large basalt-volcanoes of the 

 ordinary type, such as those of Hawaii. Extravasation on 

 a " plateau " scale is, however, by no means confined to 

 basic lavas ; witness, for instance, the highly acid rhyolites 

 of the Yellowstone Park. In explosive eruptions of the 

 most violent kind, on the other hand, the material seems to 

 be in many cases of intermediate (andesitic) composition, 

 Krakatau affording a memorable example. Other inter- 

 mediate lavas, such as the often-described trachytes of 

 Auvergne, occur in dome-like masses, as if when extruded 

 they had been too viscid to flow away from the orifices that 

 gave vent to them. These and other well-known facts, 

 pointing apparently to some connection between what we 

 may term the " habit " of lavas and their petrographical 

 nature, belong rather to the branch of the subject to which 

 we now proceed. 



Coming then to more strictly petrographical considera- 

 tions, the first thing to observe is that there exists a very 

 sharply defined geographical distribution of groups of igneous 

 rocks differing from one another in important chemical and 

 mineralogical characteristics. To appreciate this we must 

 recoQfnise a broad two-fold division of the rocks as a whole. 

 We are accustomed to speak of acid, intermediate, and 

 basic rocks, etc., tacitly assuming that the rocks fall roughly 

 into a linear series, such that, as the silica-percentage varies 

 from one extreme to the other, the percentages of the other 

 constituents vary as functions thereof. Though this is true 

 to a certain extent, the variations actually shown by analyses 

 of igneous rocks indicate a much higher degree of com- 



