Reviews — Petrographical Provinces, etc. 327 



Some interesting changes in the course of the Irrawacldy near 

 Bhamo are noted. These are apparently due to the cutting back of 

 tributaries, assisted by great differential earth-movements. 



These papers are illustrated by thirteen plates and one geological 

 map on a scale of 4 miles to the inch. 



III. — Petrographical Provinces and the Sequence op Types. 



SOME very important criticisms of the wider generalizations of 

 modern petrology were made by Henry S. Washington in a paper 

 read before the International Geological Congress in 1913 ou 

 " Volcanic Cycles in Sardinia ". 



The Sardinian volcanoes belong to three different volcanic cycles. 

 A careful study of the sequence of the various lava-flows brings out 

 the following points. The sequence as a whole is discontinuous but 

 recurrent. The close of one niagmatic cycle is simultaneous with the 

 ending of one volcanic phase, and the recurrence of a new one 

 coincides with a change in the intensity or character of tlie vulcanicity. 

 In the several successive cycles there is a progressive decrease in 

 differentiation ; thus the earliest cycle comprises rhyolites, trachytes, 

 andesites, and basalts ; while in the latest felspar-basalts are the 

 only lavas which have been extruded. Each cycle closes with the 

 extrusion of felspar-basalt. The sequence here observed is not in 

 accordance with the * law ' of divergence from an average magma. 

 Harker would, however, regard such exceptions as due to the 

 suppression of one of two divergent lines. 



The author points out that the sequence must depend on such 

 factors as the form of the magma basin, the point at which this is 

 tapped, and the character and amount of the volatile constituents 

 wliich escape, so that it may be questioned whether the sequence of 

 types really possesses the great importance usually attributed to 

 it. For most volcanic rejjions it appears that the sequence of types 

 varies somewhat with the character of the regional magma, but nearly 

 always closes with basaltic rocks. The coincidence of the several 

 cycles with' discontinuity of volcanic activity, whi(di is so well sliown 

 in Sardinia, is also observed at Pantelleria and in many other areas 

 of vulcanicity ; on the other hand, there are cases, as at Santorin, 

 where lavas extruded at different periods are practically identical in 

 composition. 



The lavas of Sardinia furnish another of the now numerous 

 instances of the association in the same volcano of the so-called 

 'Atlantic' and 'Pacific' types. It seems impossible to refer all 

 rock types to only two lines of variation. The concentration of a 

 very large number of rock analyses round several types indicates that 

 there has been one general earth -magma. Tlie magmas which 

 characterize different petrographical provinces must have been derived 

 from this general magma by differentiation in all possible directions. 

 Attention must be paid to magnesic, ferrous, and potassic nuigmas, 

 and not only to sodic ones. At present far too much importance is 

 attached to the occurrence of such minerals as a^girite and nephelite, 

 even when present in quite small amount, while the presence, for 

 instance, of potash in biotite is often neglected. 



