<<

Professional Paper No. 14 Se 'es ' er°SraP ly an(l Mineralogy, 23 E, Chemistry and Physics, 37

DEPARTMENT OF THE INTERIOR UNITED STATES GEOLOGICAL SURVEY CHARLES I). WALCOTT, DIEECTOE

CHEMICAL ANALYSES OF IGNEOUS ROCKS

PUBLISHED FROM 1884 TO 1900

WITH A

CRITICAL DISCUSSION OF THE CHARACTER AND USE OF ANALYSES

HENRY STEPHENS WASHINGTON

W A S III N G T O N GOVERN M K N T l> K I N T I N G OFFICE 11) 0 3 CON TENTS

I ettei of tnnsmittil 5 Prelat e 7 Introduction _ n The cVuracter o£ lock -uialvses 13- General considerations lo. Kepresentatneness K> Character of the mass It, Amount ot material IS Microscopical examination 1<) Accura< y 1Q Possible eirors _>0 Coiiipletenes« 2 > Mam constituents 24 Minor constituents 25 Eating ot anal\ '-cs 28 Accurac\ 2') Internxl evidence ^0 Agre emcnt \\ ith the mode 30 Summation -j^ Fxternal e\idcnce 3g

Methods ^g Indirect e\ idence 3B Completeness =57 Rating adopted in the tables 33 Discussion ol the tables 4 Di \iaion into parts 4^ Part I 44. Part II 4-> Basis ot arrangement ol aualj =es 4f( 4ttempted use ol pptrogriphu sj^tem 4g Adoption ot the quantitative sjstem 47 Quantit iti\e classification of igneous rocks 47 Constiuction of the sjstem 47 Nomenclature of the system 48 Comparison of old and new nomenclatures 49 4 CONTENTS.

Discussion ot the tables Continued Quantitative classification of igneous locks Continued Page Tabular exhibit of divisions and names of quantitative system (5 tables).--...---..... 53 Keasons for anangement according to the quantitative classification ...... 59 Distribution of rocks by new system...... 61 Comment on facts expressed by tables...... 62 Various features of the tables .-...... "...... 64 Geographical arrangement...... 64 Numbering ol analyses ...... 65 Eating of analyses ...... 65 Results ot rating analyses...... 65 Constituents...... 67 Specific gravity ...... 68 Norm ...... 68 Correspondence of norm and mode...... 69 Locality...... 69 Analyst...... 70 Reference --.....-----.-.-...... -...... -...... --.-...-...... 70 Author's! name..-.-.--...... 71 Remarks ...... 71 Correlation of the qualitative and the quantitative systems...... 72 From the qualitative point ot view ...... 73 From the quantitative point of view -...... --...,.--...... --...... --..-...-...-.--...... 76 Meteorites...... -.....-....-..-...... -...... -...... -....-..-...--.-...--...-.--.- 81 Calculation oi center points...... 81 Intioductory ...... -...-.--...... -...-..-...... -----.-.--...... --.-..--.-...... 81 Peisalane...... 84 Perfemane...... -...-.-...... --.....--.-...---.-.-.--...-.-.-.....-. 88 Distribution ot and the average rock ...... 100 Introductory ..-...... -.-.-..-,...... -...... -...... -...... -.-..-..-..-....- ...... 100 The distribution of magmas. -.-...... -.....-....-..-.-....-...-..-.--.-..---...----.... 102 The average lock ...... 106 First method ...... ' 106 Second method ...... 109 List of abbieviations ...... --...... ------.--.....--.....-.. 116 Collection ot rock analyses...... -...... -...... ---.--.--....----..-.. 1-1 Parti...... 121 Part II...... 371 Glossary ...... --...... ---.....---.-.--.---..--.-. 475 Prefatory note to indexes....-...... -...... -..-.....--.....---.--.--.---.-.----.-----.----- 481 Index to text...... 483 Index to new rock names in Part I...... -.-.....-...... -.--...--.....-...-.------486 Index to old rock names m Parts T and II...... -.------.-- .-....----... 489 Index to localities in Parts I and II...... ------.-.-----.------.----- 493 LETTER OF TIUNSMITTAT,

DEPARTMENT OK THE INTERIOR, UNITED STATES GEOLOGICAL SURVEY, ITWm^/Aw, ]). ('., March 20, 1903. SIR: I have the honor to transmit to you a manuscript entitled "Chemical Analyses of Igneous Rocks published from 1884 to 1900, with a critical discussion of the eharae.ter and use of analyses,'' by Henry Stephens Washington, and to recom­ mend that it be published by the Survey as a Professional Paper. This work is pri­ marily a compilation of chemical analyses, and is especially valuable to petrographers and chemists, for it places in one volume material gathered from a great many scat­ tered source's. The critical discussion of the value and use of rock analyses, with comments upon methods of analysis and a review of the bearing of this mass of material upon rock classification, is also of much importance to both petrographers and chemists. The arrangement of the analyses according to the quantitative system for the classification of igneous rocks permits one to compare readily any new analysis with many others of closely allied rocks. A work of this kind is necessarily very expensive when published by a commercial house, and if so issued would be beyond the reach of many who would desire to use it. It is particularly appropriate that the Survey should publish this work, because a very large proportion of the, rock analyses here included have been made in the laboratory of the United States Geological Survey, upon material which is preserved in its petrographic, reference collection. The use of that collection will be greatly facilitated by this work. Very respectfully. WHITMAN CROSS, G-I'O! fig !#t in Charye Section of Petrology. Hon. CHARLES I). WAL.COTT, Director IJ. S. G-eoloyhul PREFACE.

The collection of chemical analyses of igneous rocks here presented was made during a period covering several years, for purposes of ready reference in certain lines of investigation. In view of the gieat value of that standard work of reference, Roth's Tabellen, and of the fact that no such collection of rock analyses has appeared since the last installment of these in 1884," it was thought that an extension and completion of these along the lines of Roth's work, so as to bring them down to the present day, would be welcome to petrographers. This seemed to be the more advisable because the years that have elapsed since 1883 have witnessed a very marked increase in the number, as well as a great improvement in the quality, of chemical analyses of rocks. These years have also shown a steady increase in the importance of chemical analyses as applied to the discussion of petrological problems, and a growing tendency to use them as bases for various systems of classification which have been proposed. The classificatory work on which I have been engaged in collaboration with Cross, Iddings, and Pirsson, the results of which have recently appeared,6 was also a powerful incentive to the formation of the present collection. As has been stated by us, this collection has been appealed to whenever possible, as a test of our proposed system of classification, and has also formed to a large extent the basis of the roots of the magmatic names suggested. The years embraced in the present collection are those from 1884, that of the publication of the last of Roth's Tabellen, to 1900, both inclusive. The former was selected as a starting point because Roth collected nearly all the analyses which appeared in 1883; yet some of those which he overlooked are included in this col­ lection. The year 1900 seemed to be appropriate as the other time limit, partly for the reason that it was the closing year of the last century, and partly because in that year appeared Bulletin 168 of the United States Geological Survey, which is the latest publication embracing all the analyses of rocks made by the chemists of that organi­ zation. At the same time, since the collection is to a large extent illustrative of our

oRoth, J , Beitrage zur Petrographie der plutomschen Gestemo Sitzungsber K Preuss Akad Wiss, , 1884 ft Journal oi Geology, Vol X, 1902, p 555, also, Quantitative Classification of Igneous Rocks, Chicago, 1903 8 PREFACE. proposed S3'srem of classification and nomenclature, some analyses published prior to 1884 or subsequent to 1900 that are of especial interest or importance have been included, as illustrations of the classificatory divisions, or as forming the bases of the magmatic names. The endeavor has been made to present all analyses of igneous rocks, good, bad, and indifferent, that have been published within the time limits selected. For this purpose I have examined the most complete accessible files of all known geological survey publications, as well as those of geological and other scientific societies, the various geological journals, inaugural dissertations, etc. It is of impossible that all analyses published during the time given should have been collected. Some of the less well-known publications of scientific societies are inaccessible in this country, and, indeed, at the tinie the material was collected (1900 and 1901) the later publications of some of the surveys and societies wore, for obvious reasons, not to be had here. But it is believed that the collection is fairly complete, and that it contains at any rate a record of much of the most useful and noteworthy chemical work of the seventeen years which it represents. It is hoped that occasional appendixes will be published later, in which, if possi­ ble, inanj7 of the anatyses here overlooked will appear. For this purpose I shall esteem it a favor to have my atteation called to omissions, and shall be under great obligations to those who will be kind enough to send separates of publications which have, been overlooked in the present collection or which may appear subsequently in journals that are accessible here with difficulty or not at all. It is to be regretted that the current reviews publish so few of the analyses contained in the papers reviewed, for these analyses are often of the greatest interest. In the present trend of petrology they are often of much more importance than the microscopical details, which are frequently cited at length. Although great care has been taken to guard against errors in collecting the analyses, preparing- the manuscript, and reading the proof, mistakes arc almost unavoidable in dealing with the mass of figures represented. I can not venture to hope that they have been entirely eliminated, and can only express my regrets at their existence, and trust that my attention will be called to them a favor which will be very highly appreciated. The collection is strictly confined to analyses'of igneous rocks. It was at one time intended to include also those of rnetamorphic rocks, but the delay which this plan woiild have involved seemed finally to render it inadvisable, though a fairly complete collection of such analyses had also been made in conjunction with those here given. Sedimentary rocks of all kinds have also been omitted, though analyses of volcanic tuffs are included. Analyses of serpentines are likewise excluded, since they are due either to the decomposition of igneous rocks, so that, as explained PREFACE. 9 elsewhere, their analyses would fall under the ban imposed on such altered material, or are derived from either metamorphics or masses, and hence fall outside the scope of the present collection. It was hoped for a time that analyses of meteorites could be given; but owing to the fact that no adequate collection of the analyses of these bodies has yet been made, this would have involved a search through the voluminous and widely scattered lit­ erature published concerning theiii since the commencement of their study. Such a proceeding would have postponed indefinitely the appearance of this work, and therefore, as well as in view of the fact that they form the subject of special studj7, the project was abandoned. I am informed, furthermore, that such a collection is already being- undertaken by another, so that it is to be hoped that this serious lacuna will shortly be filled. It may seem to many petrographers who are acquainted with the chemical analy­ sis of rocks that certain portions of this subject have been gone into with unnecessary detail, or that many truths are stated which arc obvious to any one conversant with the subject. That this is true to a large extent is admitted, but it must be remem­ bered that the greater part of the text which is devoted to this subject has been written with the nonanalytical reader in mind, to whom much of what is here given will undoubtedly prove novel. As will be evident from a perusal of the text and from a study of the analyses here presented, there are far too many petrographers to whom the principles and methods of chemical analysis are as a sealed book, and in their case I can only feel that if I have erred it has been on the side of too little explicitncss or detail. The new terms and names used are those proposed in connection with the recently published quantitative classification, and their meanings will be found in the glossary given on pages 479-483. No apology is offered for their introduction, since in many cases they are the only terms available to express the ideas which it is desired to conve}7 , at least without the use of much longer or circumlocutory phrases, and it was desired as well to take this opportunity of making them more familiar to petrographers. The progress of the work has been greatly facilitated by the kindness of many friends, to whom it is a pleasure to express obligations. I am especially indebted to. the officials of the libraries of Yale and Columbia universities and of the United States Geological Survey, who courteously afforded me the fullest facilities in the examination of serial and other publications. Without their assistance this collection could not have been made satisfactorily complete. I am also indebted to several friends, among whom I may mention especially Profs. J. F. Kemp and J. B. Harrisoii, for several unpublished analyses, which they kindly allowed me to insert in the tables. To my friends Dr. Cross, Professor 10 PREFACE.

Iddings, Professor Pirsson, and Dr. Hillebrand I am grateful for many valuable sug­ gestions and criticisms. Nearly the whole of the manuscript of the text has been .submitted to my-colleagues in the new classification, so that the views here expressed may be regarded, in general, as having their indorsement, and as expressing the ideas of all four. Last, but by no means least, I must express my sense of deep obligation to the Hon. C. D. Walcott for his kindness in allowing the work to appear as a Survey pub­ lication. It is gratifying to feel that this hospitality is in part repaid by the incon­ testable evidence the present volume affords of the preeminent place in rock analysis that is held by the organization of which he is the Director. Finally, I must express my regret at the delay in the appearance of the work, which is published more than two years after the limiting date of the collection. 1 can only plead in extenuation the of other work, especially that in connec­ tion with the new system of classification, and trust that the time which has elapsed has but added to whatever worth the volume may possess. CHEMICAL ANALYSES OF IGNEOUS ROCKS.

By H. S. WASHINGTON.

INTRODUCTION. In the first two or three decades of the last century, when the study of rocks as such was being differentiated from that of and of rock terranes that is, when the science of petrogaphy was in its infancy little attention was paid to their chemical features. It is true that a number had been analyzed, but these were for the most part rocks that were of such a character as to lead the investigator of those early days to consider them as minerals, as was the case with the first described Ihcrzolite and wehrlite. In contradistinction to the individually well-defined rnin- ' erals, rocks were regarded as merely aggregates of minerals, in presumably fortui­ tous combinations, and lacking that definiteness or constancy of composition in one mass or in different masses which would justify their chemical study as a whole. As, however, they became more and more the subjects of special research, beginning with the earliest investigations of Cordier,a a knowledge of their chemical composi­ tion assumed gradually increasing interest. The great importance of this side of the study of rocks was first clearly recognized by Abich,s who pointed out, as early as 1841, the, necessity of a knowledge of their chemical composition for the of such problems as their origin, mode of formation, and connection with the interior of the , as well as the value of a comparison of their analyses as a proper basis for their classification and nomenclature. To him, therefore, is due the credit, of introducing the chemical composition of rocks as a basis for their classifi­ cation; though the good influence of this suggestion for their right understanding was largely nullified by the coincident use of the alone as one of the main factors of classification," an idea which has had a deplorably retarding influence on the development of systematic petrography for many 3rears, and which, even at the present day, holds many systematists firmly in its grasp.

a Of. Cross, Whitman, Jour. Geol., Vol. V, 1002, p. 352. 6 Abich, H.. Natur dor Vulkanische Bildungen, Braunschweig, 1841, p. v. < Cross, Whitman, loc. cit., p. 362. 11 12 CHEMICAL ANALYSES OF IGNEOUS KOCKS.

For about thirty years after Abich's publication chemical analyses occupied a- prominent position in petrographical discussions, being employed largely for the purpose of determining the mineral composition of rocks. Abich, in the work just cited, had adopted as a means to this end the ratio of the of the bases to that of the silica, and was thus enabled to calculate approximately their mineral composi­ tion, as Raimuelsberg and others had also done. A modification of this method, known as the "Oxygen-ratio" (Sauerstoffquo- tient), was introduced by Bischoff," and was long employed for the correlation of rock analyses and for the investigation of the question whether the various oxides were present in rocks in stoichiometric proportions or not. This ratio is obtained, for any given rock, by dividing the amount of the oxygen in all the bases bv that present in the silica, and was supposed to be characteristic of different broad groups. This supposition was eventually found to be unwarranted, as rocks of very diverse chemical or mineralogical characters were shown to have the same oxygen ratio. It is noteworthy that, although he himself pointed out this and other grave objections to the use of this ratio,* Roth gives it in connection with the separate analyses in all his tables, even in the last one, published in 18S4. He explains this by saying 0 that it is done, "only in the absence of a better means of comparison, as well as to make it possible to compare the older and newer analyses." It was also shown by Roth and others that rocks could not be referred to simple chemical formulae; in other words, that the oxides were not present in stoichiometric proportions. This, indeed, is one of the most important of the results arrived at by Koth from a study of the analyses collected by him, and first published in 1861. Shortly after Bischoff's suggestion of an oxygen ratio there appeared

aBischoS, G., Lehrb. Cheni. Geol., Vol. II, 1849, p. 631. rfBnnsen, K., Pogg. Ann., Vol. LXXX1II, 1851, p. 197. 6 Roth, J., Gesteins-Analysen, Berlin, 1861, p. x. e Cf. Zlrkel, Lehrbuch, Vol. I, 1893, pp. 658 et seq. cBoth, J., Biiitr. Petrog. Pluton. Gcst., 1869, p. 87. CRITICAL VALUATION" OF ANALYTICAL WOBK. 13

The microscope poured a Hood of light upon the uiineralogical constitution as well as the texture of rocks, and rendered easy and certain of attainment results which by the older methods, based on anatysis, had been attained only with difficulty or not at all. At the same time the collapse of JBunsen's hypothesis left for the time being no general theoiy of rock formation and genesis in its place, Durocher's liquation hypothesis" and Roth's first suggestion of differentiation* attracting little attention. Petrographers were too busy collecting and assimilating the vast mass of facts discovered through the microscope to be able to devote much of their energies to theoretical studies. With the opening of the last decade of the nineteenth century, however, there arose an interest in the theoretical side of petrology in contradistinction to the descriptive petrography which had hitherto prevailed since 1870. This was started by the work of Lagorio, Teall, Rosenbusch, Brogger, Iddings, and others, and since that time analyses have occupied a more and more prominent place in petrological discussions. They are no longer ornamental adjuncts, but essential parts of most petrographical publications, on which much of the discussion hangs, and from which the most important conclusions are drawn. The crystallographic and optical properties of the constituent minerals and the details of texture are no longer the main subjects of investigation, but are finding their place with the chemistry of rocks and the broad and far-reaching studies based on this. The microscope is sharing the throne with the balance. This is as it should be, since exclusive attention to one aspect tends always to obscure the goal toward which all science is striving the understanding of the " how " and the " why " of things as they are by temporarily setting up objects of interest of subsidiary importance. These subsidiary aims all have their vise in advancing the progress of the science, but it must not be forgotten that they some­ times lead into very devious paths, often of overspecialization, and that it is by keeping our moving forward together along all the lines of approach that we shall most quickly and surely reach our destination.

THE CHARACTER OF RpCK AJSTAT.YSES.

GENERAL C6NSIDERATIONS. With the growing interest in rock analyses it is of the utmost importance that the}7 be used intelligently, not only as to their application in theoretical discussions, but, which is of especial interest here, as regards a just appreciation of their character and applicability to the purpose in view. It is a somewhat surprising and, it must be said, a rather saddening fact that the critical judgment of petrographer.s in general as regards rock analyses seems

"Durother, Ann. des Mines, Vol. XI, 1857, i>. 217. !>Roth, J., Gesteins-Analysen, 1861, p. xix. 14 CHEMICAL ANALYSES OF IGNEOUS BOCKS. to be in abeyance or wholly lacking'. There is among them now, as there seems always to have been to some extent, a tendency to place implicit confidence in the results of analytical work to accept readily whatever figures the analyst may furnish, with scarcely ever an attempt at a critical estimate of the worth of the analysis or a comparison of it with the chemical character of the rock itself as revealed by the microscope. It seems to be taken for granted by nearly all petrog- raphers that the analyst, like the proverbial king, can do no wrong. This applies, not to the personal good faith of the analyst, but to the analytical processes which, possibly because they belong to one of the exact sciences, are for the most part tacitly assumed by the petrographer to be infallible. It is noteworthy that this attitude of miud obtains, not only among beginners, but among the foremost workers in .the science. There are, it is true, some instances of the application of expressed criteria in the selection of analyses in well-known papers, but in all these cases the criteria applied are few, and quite inadequate to the purpose, being confined, for the most part, to the freshness of the rock. In few cases does there seem to be any recognition of the difficulties and uncertainties of analytical work or any practical acquaintance with its methods. It goes without saying that this way of regarding analyses is totally at variance with the best interests of the science. Analyses constitute the basis of much of the investigations and discussions with which petrographers and petrologists must occupy themselves, and it is surely not the part of wisdom to erect elaborate structures on foundations of whose stability and careful workmanship we are not assured. It will be time saved in the end, and the superstructure, however well built in itself, will have a far better chance of withstanding the ravages of time and the attacks of critics if all the rotten and unsound blocks in the foundation are removed and only the best and strongest of material retained. A consequence of this unquestioning confidence in the results of analytical work is that it is often intrusted to a student in chemistry, one with little or no experience in the analysis of rocks, and with no appreciation of the complexities and difficulties inherent in this department of quantitative analysis. That the results furnished by such inexperienced, and to this extent incompetent, analysts should be received with the greatest caution, is a truism to anyone acquainted by actual experience with the difficulties and dangers of rock analysis. But Hillebrand '-' and Pirsson* have expressed so clearly and concisely the views which I hold that I can do no better than refer to them. It is certainly remarkable and significant that in petrography alone of all the sciences the most difficult and intricate work is intrusted to novices, and that their results are accepted by nearly all with the same confidence that is given to the work of an experienced person.

itHillubrand, W. P., Bull. U. S. Geol. Survey No. 148, 1897, p. Hi. SPireson, L. V, Twentieth Ann. Kept. U. S. Geol. Survey, 1900, Pt. Ill, p. 578. CRITICAL VALUATION OF ANALYTICAL WORK. 15

This general laxity among petrographers as regards the quality of rock analyses is only too painfully evident in the present collection. There is to be found an astonishingly large proportion of poor work, much of it of such a character that it would seem that even a tyi'o in analysis would reject it as hopelessly inadequate for use. -Analyses are given with summations over 103 or below 98. Analyses of leucite- basanite exist so low in alkalies that 10 or more per cent .of must be present, even though in the calculation all the potash is assigned to orthoclase and all the soda to albite. Analyses of rocks rich in divine are found which show 30 per cent of alumina and only small amounts of ferrous iron and magnesia. In certain cases of alkalic rocks the alkalies have not been separately determined. Rocks with abundant tegirite contain no ferric iron, and those with sodalite no chlorine, to judge from the figures furnished. Indeed 1 may say that some of the most depressing hours of my life have been those spent in collecting and arranging the "poor" and "bad" analyses here given, especial!y those grouped together in Part II of the collection. They represent a vast amount of misdirected and wasted energy, and, in many cases, have served as foundations for much logical and other­ wise excellent reasoning, which has been almost wholly vitiated by the worthlessness of its basal data. This would often be ludicrous were it not pathetic. The publication and general acceptance of all this worthless analytical matter reveals the weakest side of petrography, and goes far toward showing why, up to the present, no classification of igneous rocks based on purely chemical or quanti­ tatively mineral characters has been proposed, or would be likely to receive the general assent of petrographers. The chief feason for this low standard of criticism on the part of petrographers as regards analysts and analyses would seem to be that, while all of them are neces­ sarily conversant with chemistry from the theoretical side, few have much knowledge of the theory of quantitative analysis or much experience in its methods. The fact is not generally recognized that the complete and adequate analysis of an is one of the most complex and, in some respects, one of the most difficult problems of analytical science, far beyond the capabilities of a novice, and demanding not only chemical knowledge and manipulative skill, but often the exercise of considerable judgment derived from experience in solving the perplexing problems which may present themselves." With the greatly increased importance of chemical analyses at the present time, it can justly be considered that the ability to make an accurate and fairly complete chemical analysis of an igneous rock should form an essential part of the training arid equipment of every petrographer. It is only by such knowledge that one is able to judge at first hand of the true value of an analysis, to see where errors have