September 7, 191 1] 



NATURE 



described by Dr. Teall and more recently by Dr. Shand ; 

 while Dr. Flett has recognised alkaline rocks of more than 

 one age in Cornwall and Devon, and Mr. Tyrrell is engaged 

 in studying another interesting province, of Permian age, in 



Thai the distinction between the alkaline and the calcic 

 rocks embodies some principle of real and fundamental 

 significance becomes very apparent when we look at the 

 geographical distribution of the two branches. Taking 

 what the German petrographers call the " younger " igneous 

 rocks, i.e., those belonging to the latest system of igneous 

 activity, we find it possible to map out the active parts of 

 the earth's crust into great continuous regions of alkaline 

 rocks on the one hand and of calcic on the other. An 

 alkaline region comprises numerous petrographical pro- 

 vinces, which may differ notably from one another, but 

 agree in being all of alkaline facies. In like manner a 

 common calcic facies unites other provinces, which collec- 

 tively make up a continuous calcic region. Concerning the 

 igneous rocks of earlier periods our knowledge is less 

 complete, but, so far as it goes, it points to the same general 

 conclusions. 



These considerations enable us to simplify at the outset 

 the problem before us. If we would seek the meaning and 

 origin of petrographical provinces, we must inquire in the 

 first place how igneous rocks as a whole come to group 

 themselves under two great categories, which, at any one 

 period of igneous activity, are found in separate regions of 

 the earth's crust. The fact that a given district may form 

 part of a calcic province at one period and of an alkaline 

 one at another, precludes the hypothesis that the composition 

 of igneous rocks depends in any degree upon peculiarities 

 inherent from the beginning in the subjacent crust. The 

 same objection applies with scarcely less force to various 

 conflicting suggestions based on an assumed absorption or 

 " assimilation " of sedimentary rocks by igneous magmas. 

 Thus Jensen supposes the alkaline rocks to be derived bv 

 the assimilation or fusion of alkaline sediments at great 

 depths. Daly propounds the more elaborate, and on a first 

 view paradoxical, theory that alkaline have been derived 

 from calcic magmas as a consequence of the absorption of 

 limestone. These geologists agree in regarding the alkaline 

 rocks as relatively unimportant in their actual development 

 and in some sense abnormal in 'their origin. For Suess, on 

 the other hand, it is the calcic rocks which owe their dis- 

 tinctive characters to an absorption of sedimentarv material, 

 enriching the magma in lime and magnesia. Apart from 

 difficulties of the physical and chemical kind, all such 

 theories fail to satisfy, in that they ignore the separation of 

 the two branches of igneous rocks in different regions of 

 the globe, each of which includes sediments of every kind. 

 What then is the real significance of this regional separa- 

 tion ? The obvious way of approaching the question is to 

 inquire first whether the alkaline and calcic regions of the 

 globe present any notable differences of a kind other than 

 petrographical. 



Relation between Tectonic and Petrographical Facies. 

 The close connection between igneous activity and dis- 

 placements of the earth's crust has been traced bv Suess, 

 Lossen, Bertrand, de Lapparent, and others, and is a fact 

 sufficiently well recognised. We have here, indeed, two 

 different ways of relieving unequal stresses in the crust, 

 and it is not surprising that they show a broad general 

 coincidence both in space and in time. We can, however, 

 go farther. Not only the distribution of igneous rocks in 

 general, but the distribution of different kinds of rocks, is 

 seen to stand in unmistakable relation to the leading 

 tectonic features of the globe. It is very noticeable that 

 petrographical provinces, and in particular provinces 

 belonging to opposite branches, are often divided by 

 important orographic lines. This is illustrated by the 

 Cordilleran chain in both North and South America, and 

 again by some of the principal arcs of the Alpine system 

 in Europe. If, now, we examine the actual distribution 

 more closely, in the light of Suess's analysis of the con- 

 tinents and oceanic basins, we perceive another relation 

 still more significant. It is that, as regards the younger 

 igneous rocks, the main alkaline and calcic regions corre- 

 spond with the areas characterised by the Atlantic and 

 Pacific types of coast-line respectively ; I briefly drew 



NO. 2184, VOL. 87] 



attention to this correspondence in 189D, and a few years 

 later Prof. Becke, of Vienna, arrived independently at the 

 same generalisation. Recalling the two classes of crust- 

 movements discriminated by Suess, he says it appears that 

 the alkaline rocks are typically associated with subsidence 

 due to radial contraction of the globe, and the calcic rocks 

 with folding due to lateral compression. The greater part 

 of Becke 's memoir is devoted to a comparison of the two 

 branches in respect of chemical composition ; but here, I 

 think, he has been misled by taking as representative of 

 the whole alkaline " Sij>j>e " or tribe the rocks of one 

 small and peculiar province, that of the Bohemian Mittel- 

 gebirge. Some petrologists have followed Becke in adopt- 

 ing the terms Atlantic and Pacific as names, or at least 

 synonyms, for the two branches of igneous rocks. Others, 

 perhaps with some justice, deprecate the use of the same 

 terms in a petrographical as well as a tectonic sense, so 

 long as the implied relationship is still a matter of dis- 

 cussion. 



I would point out in passing that the association of the 

 alkaline rocks with areas of subsidence helps to explain 

 the relatively small part which they play in the visible 

 portion of the earth's surface. We may not unreasonably 

 conjecture, for instance, that the volcanic islands scattered 

 sparingly over the face of the Atlantic Ocean, from the 

 Azores to Tristan d'Acunha, are merely fragments of a 

 very extensive tract of alkaline rocks now submerged. 



The generalisation associated with the name of Becke, 

 in so far as it may ultimately commend itself to general 1 

 acceptance, must have an important bearing on the problem 

 of the origin of petrographical differences. The time is 

 not ripe for any dogmatic pronouncement, but I will venture 

 to indicate briefly the general trend of the inferences to be 

 drawn. It seems clear that only a trivial effect at most 

 can be allowed to original and permanent heterogeneity of 

 the earth's crust, or to such accidents as the absorption by 

 an igneous magma of a limited amount of the country- 

 rock. The division between alkaline and calcic regions, 

 and the separation of distinct provinces within such 

 regions, point rather to the same general cause which, at 

 a later stage, produced the diversity of rock-types within 

 a single province, that is, to magmatic differentiation. 

 Here, however, the differentiation postulated must be on 

 a very wide scale, and must take effect in the horizontal 

 direction. Its close connection with crust-movements 

 clearly indicates differential stress as an essential element 

 in the process. The actual mechanism can be at present 

 only a matter of speculation, but I think the clue will be 

 found in such observations as those of Mr. Barrow on the 

 pegmatites of the Scottish Highlands. Conceive an exten- 

 sive tract to be underlain by a zone which is neither solid 

 nor liquid, but composed of crystals with an interstitial 

 fluid magma. If this be subjected to different pressures in 

 different parts of its horizontal extent, its uniformity will 

 necessarily be disturbed, the fluid portion being squeezed 

 out at places of higher pressure and driven to places of 

 lower pressure. The precise nature of the differentiation 

 thus set up will depend on the relative compositions of the 

 crystalline and fluid portions, and the subject could not be 

 very profitably discussed without fuller knowledge concern- 

 ing the order of crystallisation in rock-magmas. Whether 

 or not the explanation be ultimately found in this direc- 

 tion, the relation between the two tectonic types and the 

 two branches of igneous rocks must, I think, find a place 

 in the final solution of the problem. 



I intimated at the outset that my remarks would not be 

 confined to matters already settled and indisputable. It 

 will be easily understood that some statements which I 

 have made, for the sake of clearness, without qualification 

 are subject to exceptions, and exceptions have, indeed, been 

 urged by critics whose opinions are entitled to respect. 

 The most uncompromising of these critics, Dr. Whitman 

 Cross, has laid it down that : " Only generalisations with- 

 out known exceptions in experience can be applied to the 

 construction of a system that may be called natural." I 

 hold, on the contrary, that such a science as Geology can 

 be advanced only by the inductive method, which implies 

 provisional hypotheses and successive approximations to the 

 truth. A generalisation which brings together a mass of 

 scattered observations, and endows them with meaning, is 

 not invalidated by the discovery of exceptions. These 



