September 7, 191 1] 



NATURE 



3 J 9 



SECTION C. 



GEOLOGY. 



Opening Address by Alfred Harker, M.A., F.R.S., 



President of the Section. 



Some Aspects of Modem Petrology. 



In accordance with the custom which permits the 

 occupant of this chair to open the proceedings with observa- 

 tions on some selected subject, 1 wish to invite your attention 

 in certain points concerning the genetic relations of igneous 

 tocks. The considerations which I shall have to lay before 

 you will be in some measure tentative and incomplete; and 

 indeed, apart from personal shortcomings, this character 

 must necessarily attach to any discussion of the subject 

 which 1 have chosen. For petrology is at the present time 

 in a slate of transition — the transition, namely, from a 

 merely descriptive to an inductive science — and at such a 

 time wide differences of opinion are inevitable. If I should 

 seem to do less than justice to some views which I do not 

 share, I hope this fault will be attributed to the limitations 

 of time and space, not ft my intention of abusing the brief 

 authority with which I find myself invested. 



The application of microscopical and special optical 

 methods, initiated some fifty years ago by Dr. Sorby, gave 

 a powerful impetus to the study of the mineral constitution 

 and minute structure of rocks, and has largely determined 

 the course of petrological research since that epoch. For 

 Sorby himself observation was a means to an end. His 

 pteresl was in the conclusions which he was thus enabled 

 to reach relative to the conditions under which the rocks 

 were formed, and his contributions to this problem will 

 always rank among the classics of geology. The great 

 Bnajoriti of his followers, however, have been content to 

 record and compare the results of observation without 

 pushing their inquiries farther ; and indeed the name 

 "petrography," often applied to this line of research, cor- 

 rectly denotes its purely desi riptive nature. A very large 

 body of fact's has now been brought together, and may be 

 found, collated and systematised by a master-hand, in' the 

 monumental work of Rosenbusch. Beyond their intrinsic 

 interest, the results thus placed on record must be of the 

 hiyh.;vt value as furnishing one of the bases upon which may 

 eventually be erected a coherent science of igneous rocks 

 and igneous activity. 



In earnest of this promise, recent years have witnessed a 

 very marked revival of interest in what we must call at 

 present the more speculative aspects of petrology. This 

 Manifests itself on the side of the petrographer in a growing 

 disposition to seek a rational interpretation of his observa- 

 tions in the light of known physical principles, and on the 

 Side of the field geologist in a more constant regard for the 

 distribution, mutual associations, and mode of occurrence 

 of igneous rocks. I will add, as another hopeful sign of 

 the times, a decided rapprochement between the laboratory 

 and the field, too often treated in practice as distinct 

 departments. 



As regards the former, the movement which I have noticed 

 is merely a return to the standpoint of Sorby, the father of 

 modern petrology. It is true indeed that, before his time, 

 the problem of the origin of igneous rocks had engaged the 

 ingenuity of Scrope and Darwin, of Bunsen and Durocher, 

 and many others ; and the bold speculations of the heroic 

 days of geology have justly exercised a lasting influence. 

 Ihe petrologist of to-day, however, has at his command a 

 much ampler range of information than was possessed by 

 his predecessors. In addition to the rich store of petro- 

 graphical data already mentioned, he can press into service 

 on the one hand the results of physical chemistry and on the 

 other much additional knowledge which has been gathered 

 concerning the structure of the earth's crust and the distri- 

 bution of various rock-types, both in space and in time. 

 Either of these branches of the subject would furnish 

 material for a much longer address than my assurance could 

 venture or your complacence would endure. I have chosen 

 ill.' geographical aspect of petrology; but, before proceeding 

 to this, I will say a few words concerning the experimental 

 side. r 



Data from the Experimental Side. 



'I hat the modern developments of physical chemistry 

 starting from the phase rule of Willard' Gibbs, must in 

 theory furnish all that is necessarv to elucidate the 

 NO. 2184, VOL. 87] 



crystallisation of igneous rock-magmas, has long been per- 

 ceived by some petrologists. This recognition "is in itself 

 an advance. Natural rock-magmas, however, are far more 

 complex solutions than those which chemists have employed 

 in working out their laws, and the problem in its entirety 

 is of a kind almost to daunt inquiry. Despite the courageous 

 attempt made by Prof. Vogt, whose enthusiastic lead has 

 done so much to inspire interest in the subject, it seems 

 clear that the application of the laws of chemistrv to the 

 particular class of cases with which the petrologist is con- 

 erned demands as a prerequisite a large amount of experi- 

 mental work in the laboratory. The high melting-points of 

 the rock-forming minerals, their extreme viscosity, and 

 other specific properties render such work extremely difficult 

 and laborious. That most of the practical difficulties have 

 now been overcome is due in the first place to Dr. A. L. 

 Day and his colleagues of the Geophysical Laboratory at 

 Washington, who have thus opened out what is virtually a 

 new field of investigation. The methods of high tempera- 

 ture measurement have been perfected and the thermometric 

 scale standardised up to 1550 C., thus embracing the whole 

 range of rock-formation. Calorimetric measurements have 

 been so far improved that it is now possible, for instance, 

 to determine specific heats, even in the highest part of this 

 range, with an accuracy ten times greater than has hitherto 

 been usual at ordinary temperatures. Incidentally there has 

 1" ' n, in the hands of Mr. F. E. Wright, a notable enlarge- 

 ment of the scope of ordinary petrographical methods, since 

 it has been found necessary to devise special means of 

 measuring with precision the crystallographic and optical 

 constants of very minute crystals. 



The American chemists have already determined the tem- 

 perature-range of stability of numerous rock-forming 

 minerals. Beginning with the simpler cases and working 

 always with chemically pure material, they have established 

 quantitatively the mutual relations of the various possible 

 forms in a number of two-component systems and in one of 

 three components. So far as these instances go, the mutual 

 lowering of melting-points in a silicate-magma is now a 

 matter of precise measurement, and it is no longer inferred, 

 but demonstrated, that the order of crystallisation of the 

 minerals depends upon their relative proportions in the 

 magma. The perfect isomorphism of the plagioclase felspars 

 has been finally established, and a certain degree of solid 

 solution between quite different minerals has furnished the 

 explanation of some apparent anomalies, such, for instance, 

 as the variable composition of the mineral pyrrhotite. As 

 a single illustration of how these investigations in the 

 laboratory provide the working petrologist with new instru- 

 ments of research, I will cite the conception of a geological 

 temperature-scale, the fixed points on which are given by the 

 temperature-limits of stability of various minerals. It is 

 often possible, for example, to ascertain whether quartz in 

 a given rock has crystallised above or below 575 C, this 

 being the inversion-point between the a- and /3- forms of the 

 mineral. At about 8oo° there is another inversion-point, 

 above which quartz is no longer stable, but gives place to 

 cristobalite. In like manner we know that wcTllastonite in a 

 rock must have crystallised below 1190 , pyrites below 

 aid so for other cases. We may confidently hope that, 

 with the aid of such data, we shall soon be enabled, by 

 simple inspection, to lay down in degrees the temperature- 

 range of crystallisation of a given igneous rock. 



I here are now several laboratories where high-tempera- 

 ture research, of the rigorous order indicated, is being carried 

 out ; but the work is peculiarly arduous, and results come 

 slowly. Some branches of the inquiry, notably those in- 

 volving high pressures, and again the investigation of 

 ms into which volatile components enter, are as yet 

 virtually untouched. For these reasons it would be pre- 

 mature to hazard at this stage any more detailed forecast of 

 the services to be rendered to petrology by synthetic experi- 

 ment. I will accordingly leave this attractive subject, and 

 pass on from the laboratory to the field. 



Geographical Distribution of Igneous Rocks. 



Here the existing situation is very different. Instead of 



following out definite lines already laid down, we are con- 



d in reducing to order a great mass of discrete facts 



di iwn from many sources. The facts which enter into 



consideration are those touching the distribution of various 



