An , Hh iia Mh Me i Hh AAU 4 otis PO Hn LT Aaa ik (W"} hi by A Ly ih ip La Wee eth Tuts ¥ yh ah uy Pn wae RR ? 4 Thi yw a) Ut A ai CALE Vitae 8 KT eat \ RENKIN WALA Ca vod SAM ’ Aft) Sa we a feat on aA f on 1h i yi yO iy VA RN ‘) WAY Hh + COTE TS ACM NT BT i Wek Wat Pha A PS ae ‘ii yi Wt Me Mey ‘§ bait 7 ls pote sith » Nuria fier aman i ch Yh nbs Hein Mey, i shoe sath g AAA ah " ne aK Aa ih i 1 a nie hi igh i ehh Ky Waa, f a iy " Sat ane Ait “ UY Net i oe |e He: wit as tS i vi (4 ‘a Wit Bienes tio we i we ¥ iil i We veh ree. y oS ht: ci po Be ah vi duties an Ne ae oe 0 mk : SOONG ve tae sone a nie BA wi Nis ie a) cent Hh i ry mT Petty ist 62 if ate stays Bi » sul dait Man ine ah { by Hae rwey tod ahi beagles chp #e, wh Lape iy Py s dala ele Mul oa fi Vit uns ae ay aa ) a: as igh me Ege tA Punnett aaa re ii ie rae Way Rent eg i oi ide ty vy tod hte v jr Cie ih ee We oe eit ts ri MN fist Ute ae ti ie sh oh ce eth iG hoy me Rtas MN ) a 3) ae es ; 2S aoe 7 THE UNIVERSITY OF CHICAGO PRESS CHICAGO, ILLINOIS * THE CAMBRIDGE UNIVERSITY PRESS LONDON AND EDINBURGH THE MARUZEN-KABUSHIKI-KAISHA TOKYO, OSAKA, KYOTO, FUKUOKA, SENDAI THE MISSION BOOK COMPANY SHANGHAI THE Nie.‘ JOURNAL OF GEOLOGY A Semi-Quarterly Magazine of Geology and Related Sciences EDITED BY THOMAS C. CHAMBERLIN AND ROLLIN D. SALISBURY With the Active Collaboration of *SAMUEL W. WILLISTON ALBERT JOHANNSEN Vertebrate Paleontology Petrology STUART WELLER ROLLIN T. CHAMBERLIN Invertebrate Paleontology Dynamic Geology ALBERT D. BROKAW, Economic Geology ASSOCIATE EDITORS SIR ARCHIBALD GEIKIE, Great Britain *HENRY S. WILLIAMS, Cornell University CHARLES BARROIS, France JOSEPH P. IDDINGS, Washington, D.C. ALBRECHT PENCK, Germany fe JOHN C. BRANNER, Leland Stanford Junior Uni- HANS REUSCH, Norway versity GERARD DEGEER, Sweden RICHARD A. F. PENROSE, Jr., Philadelphia, Pa. T. W. EDGEWORTH DAVID, Australia WILLIAM H. HOBBS, University of Michigan BAILEY WILLIS, Leland Stanford Junior FRANK D. ADAMS, McGill University University CHARLES K. LEITH, University of Wisconsin” *GROVE K. GILBERT, Washington, D.C. WALLACE W. ATWOOD, Harvard University CHARLES D. WALCOTT, Smithsonian WILLIAM H. EMMONS, University of Minnesota Institution ARTHUR L. DAY, Carnegie Institution *Deceased VOLUME XXVI JANUARY-DECEMBER, 10918 THE UNIVERSITY OF CHICAGO PRESS CHICAGO, ILLINOIS ebruary, March, M ie November / ‘ fi a) Composed and Printed By The University of Chicago Press _ Chicago, Illinois, U.S.A. ; ‘ \ ng ei CONTENTS OF VOLUME XXVTJ NUMBER I Low-ANGLE FAULTING. R. T. Chamberlin and W. Z. Miller. THE Hart MOUNTAIN OVERTHRUST AND ASSOCIATED STRUCTURES IN Park County, Wyomine. C.L. Dake . THE ORIGIN OF VEINLETS IN THE SILURIAN AND DEVONIAN STRATA OF CENTRAL NEw York. Stephen Taber TRANSPORTATION OF DEBRIS BY ICEBERGS. O. D. von Engeln PETROLOGICAL ABSTRACTS AND ReEviEws. Albert Johannsen. REVIEWS . ; : Fa NUMBER II GENESIS OF THE ALKALINE Rocks. Reginald A. Daly CONDITIONS OF DEPOSITION ON THE CONTINENTAL SHELF AND SLOPE. C. A. Cotton THE NORTHWARD EXTENSION OF THE PHYSIOGRAPHIC DIVISIONS OF THE Unitep States. I. W.N. Thayer PETROLOGICAL ABSTRACTS AND REVIEWS. Albert Johannsen. REVIEWS NUMBER III DIASTROPHISM AND THE FORMATIVE Processes. IX. A SPECIFIC Mope oF SELF-PROMOTION OF PERIODIC DIASTROPHISM. T. C. Chamberlin CorAL REEFS AND SUBMARINE Banks. W. M. Davis SANTO DOoMINGAN PALEONTOLOGICAL EXPLORATIONS. Carlotta J. Maury Bock FAULTING IN THE KLAMATH LAKES REGION. Douglas Wilson Johnson 97 vi CONTENTS OF VOLUME XXVI THE NORTHWARD EXTENSION OF THE PHYSIOGRAPHIC DIVISIONS OF THE Unitep States. II. W.N. Thayer . THE GEOLOGY OF THE KILDEER Mountains, NortH Dakota. ‘Terence T. Quirke PETROLOGICAL ABSTRACTS AND REvIEws. Albert Johannsen. REVIEWS NUMBER IV CoRAL REEFS AND SUBMARINE Banks. W. M. Davis PERMO-CARBONIFEROUS GLACIAL DEPOSITS OF SouTH AMERICA. A. P. Coleman THE SUBPROVINCIAL LIMITATIONS OF PRE-CAMBRIAN NOMENCLATURE IN THE ST. LAWRENCE Basin. M. E. Wilson CORRELATION OF THE EARLY SILURIAN ROCKS IN THE Hupson BAy Recion. T. E. Savage . NOTES ON SEDIMENTATION IN THE MACKENZIE RIVER Basin. E. M. Kindle NOTES ON THE MISSISSIPPIAN CHERT OF THE ST. Louts AREA. Donald C. Barton . EDITORIAL PETROLOGICAL ABSTRACTS AND REviEws. Albert Johannsen. RECENT PUBLICATIONS i NUMBER V as CoRAL REEFS AND SUBMARINE BAnkKs. W. M. Davis THE IRON-FORMATION ON BELCHER ISLANDS, Hupson BAy, WITH SPECIAL REFERENCE TO ITS ORIGIN AND Its ASsocIATED ALGAL LimesTones. E. S. Moore INTERNAL STRUCTURES OF IGNEOUS ROCKS; THEIR SIGNIFICANCE AND ORIGIN; WITH SPECIAL REFERENCE TO THE DULUTH GABBRO. Frank F. Grout THE HABITAT OF THE SAUROPOD Dinosaurs. Charles C. Mook PETROLOGICAL ABSTRACTS AND REvrews. Albert Johannsen. REVIEWS PAGE 237 255 272 283 289 310 325 334 341 SiS Sli CONTENTS OF VOLUME XXVI NUMBER VI Two-PHASE CONVECTION IN IcGNEous MAcmaAs. Frank F. Grout PERMO-CARBONIFEROUS CONDITIONS VERSUS PERMO-CARBONIFEROUS Time. E. C. Case NOTES ON THE GEOLOGY OF EASTERN GUATEMALA AND NORTHWESTERN SPANISH HonpURAS. Sidney Powers . NOTES ON A COLLECTION OF ROCKS FROM HONDURAS, CENTRAL AMERICA. Wilbur G. Foye Lorss-DEPOSITING WINDS IN LoutsIANA. F. V. Emerson VoLUME CHANGES IN METAMORPHISM. Waldemar Lindgren DESCRIPTIONS OF SOME NEw SPECIES OF DEVONIAN FossiLs. Clinton R. Stauffer ON THE NATURE AND ORIGIN OF THE STYLOLITIC STRUCTURE IN TEN- NESSEE MARBLE. C.H. Gordon . THE VALLEY City GRABEN, UtaH. C. L. Dake REVIEWS RECENT PUBLICATIONS NUMBER VII Tue ECOLOGICAL SIGNIFICANCE OF THE EAGLE CREEK FLORA OF THE CotumBIA RIVER GorcE. Ralph W. Chaney On OOLITES AND SPHERULITES. Walter H. Bucher RHYTHMIC BANDING OF MANGANESE DIOXIDE IN RHYOLITE TUFF. W. A. Tarr THE RELATION OF THE Fort SCOTT FORMATION TO THE BOONE CHERT IN SOUTHEASTERN KANSAS AND NORTHEASTERN OKLAHOMA. Walter R. Berger A Form oF MULTIPLE Rock Diacrams. Frank F. Grout A Tyrer or IGNEOUS DIFFERENTIATION. Frank F. Grout Post-GLacIAL MOLLUSCA FROM THE Maris OF CENTRAL ILLINOIS. Frank Collins Baker RECENT PUBLICATIONS 574 576 ond 593 610 Vill CONTENTS OF VOLUME XXVI NUMBER VIII SAMUEL WENDELL WILLISTON. Henry Fairfield Osborn CHARLES RICHARD VAN Hise. T. C. Chamberlin HENRY SHALER WILLIAMS. Stuart Weller WORLD-ORGANIZATION AFTER THE WoORLD-WAR—AN OMNINATIONAL CONFEDERATION. T.C. Chamberlin . A GEOLOGICAL RECONNAISSANCE IN Hartt. A Conrtripution} To ANTILLEAN GEOLOGY. * William F. Jones PAGE 673 690 608 7OI 728 NUMBER 1 7 Ce eo JOURNAL of GEOLOGY cates A SEMI-QUARTERLY SRY MOR EDITED By ae THOMAS C. CHAMBERLIN AND ROLLIN D. SALISBURY With the Active Collaboration of . ‘SAMUEL W. WILLISTON, Mettebrate Paleontology ALBERT JOHANNSEN, Petrology ei _ STUART WELLER, Invertebrate Paleontology ROLLIN T. CHAMBERLIN, Dynamic Geology oy ALBERT D. BROKAW, Economic Geology s ASSOCIATE EDITORS SIR ARCHIBALD GEIKIE, Great Britain JOSEPH P. IDDINGS, Washington, D.C. CHARLES BARROIS, France ' JOHN C. BRANN ER, Leland Stanford Junior University ALBRECHT PENCK, Germany RICHARD A. F. PENROSE, Jr., Philadelphia, Pa. a ae oni AGEN WILLIAM H. HOBBS, University of Michigan x FRANK D. ADAMS, McGill University T. W. EDGEWORTH DAVID, Australia _ BAILEY WILLIS, Leland Stanford Junior University CHARLES K. LEITH, University of Wisconsin GROVE K. GILBERT, Washington, D.C. WALLACE W. ATWOOD, Harvard University CHARLES D. WALCOTT, Smithsonian Institution WILLIAM H. EMMONS, University of Minnesota HENRY S. WILLIAMS, Cornell University ARTHUR L. DAY, Carnegie Institution ‘ JANUARY-FEBRUARY 1918 LOW-ANGLE FAULTING - - - = - R. T. CHAMBERLIN AND W. Z. MILLER T THE HART MOUNTAIN OVERTHRUST AND ASSOCIATED STRUCTURES IN PARK COUNTY, WYOMING - - 5 = - - - - - =e Cob DAK BIW 46 THE ORIGIN OF VEINLETS IN THE SILURIAN AND DEVONIAN STRATA OF CEN- TRAL NEW YORK - - = - - - - - STEPHEN TABER 56 ) ’ 5 ’ TRANSPORTATION OF DEBRIS BY ICEBERGS --0..D. von ENGELN 74 4 ’ ' i PETROLOGICAL ABSTRACTS AND REVIEWS ALBERT JOHANNSEN 82 HEINE cat een a AE aries cae TUR i A aI UR EE FELIS! Pe OME RSr hy OF CHICAGO PRESS CHICAGO, ILLINOIS, U.S.A. THE CAMBRIDGE UNIVERSITY PRESS, Lonpon anp EpINBURGH THE MARUZEN-KABUSHIKI-KAISHA, Tokyo, Osaka, Kyoto, Fukuoka, SENDAI THE MISSION BOOK COMPANY, SHANGHAI mea cusets sy EDITED By ene ee me a iene al THOMAS C. CHAMBERLIN AND ROLLIN D. SALISBURY _ With the ‘Active Gallsbeatiod of SAMUEL W. WILLISTON ___ ALBERT JOHANNSEN _ ee, _ Vertebrate Palenaealonys Deny ie _ Petrology Nini pins) 2: STUART WELLER . at ROLLIN T. CHAMBERLIN __ ie uae : ! Invertebrate Paleontology : Dynamic Geology | : mae ALBERT D. BROKAW. idan sid Economic Geology The qeemicl of Geology i is published by the University of Chicago at the ‘University Press! semi- quarterly, on or about the following dates: February 1, March 15, May 1, June 15, August r, September 15, _ November 1, December 15. | The subscription price is $4.00 per year; the price of single copies is 65 cents. Orders for service of less than a half-year will be charged at the single-copy rate. {[ Postage is" prepaid by the publishers on all orders from the United States, Mexico, Cuba, Porto Rico, Panama Canal Zone, Republic of Panama, Hawaiian Islands, Philippine Islands, Guam, Samoan Islands, Shanghai. {| Postage is charged extra as follows: For Canada, 30 cents on annual subscriptions (total $4.30), on | single copies 4 cents (total 69 cents); for all other countries in the Postal Union, 53 cents on annual sub- scriptions (total $4.53), on single copies 11 cents (total 76 cents.) ] Patrons are requested to make all remittances payable to The University of Chicago Press in postal or express money orders or bank drafts. The following are authorized to quote the prices indicated: For the British Empire: The Cambridge University Press, Fetter Lane, London, E.C. 4. Vong subscriptions, including postage, 18s. 8d. each; single copies, including postage, 3s. 2d. each. For Japan and Korea: The Maruzen-Kabushiki-Kaisha, 11 to 16 Nihonbashi Tori Sanchome, Tokyo, Japan. Yearly subscriptions, including postage, Yen 9.00 each; single copies, including postage, Yen 1.50 each. For China: The Mission Book Company, 18 Peking Road, Shanghai. Yearly subscriptions $4.00; single copies 65 cents, or their equivalents in Chinese money. Postage extra, if pera direct outside of Shanghai, on yearly subscriptions 53 cents, on single copies 11 cents. Claims for missing numbers should be made within the month following the regular month of publi- cation. The publishers expect to supply missing numbers free only when losses have been sustained in transit, and when the reserve stock will permit. Deliveries are not guaranteed within the war zones of Europe, nor in other countries where transportation is disturbed by conditions incident to war. Business correspondence should be addressed to The University of Chicago Press, Chicago, Ill. Communications for the editors and manuscripts should be addressed to the Editors of THE JournaL or GEroLoey, the University of Chicago, Chicago, Ill. Contributors are requested to write scientific and proper names with particular care and in citations to follow the form shown in the pages of the Journal. Separates, if desired, must be ordered in advance of publication. Separates of leading articles will be printed as ordered, of which 50 (without covers) will be furnished gratis, the actual cost of the remainder (and covers, if desired) to be paid for by the author. The actual cost may vary, and will depend upon the amount of work in remaking the pages into forms, presswork, paper, binding, etc. Separates con- taining half-tones may be expected to cost somewhat more, the increase depending upon the number of cuts and the amount of work required upon them. . Entered as second-class matter, March 20, 1893, at the Post-office at Chicago, IIl., under the Act of March 3, 1879 VOLUME XXVI NUMBER 1 THE fowrRNAL OF GEOLOGY JANOARY PEBROARY rors LOW-ANGLE FAULTING R. T. CHAMBERLIN AND W. Z. MILLER University of Chicago . CONTENTS INTRODUCTION Prevalence of Low-Angle Overthrust An Unsolved Problem PREVIOUS INVESTIGATIONS METHODS OF PRESENT EXPERIMENTAL INVESTIGATION Apparatus Materials RELATION OF FAULTS TO STRESS AND STRAIN Factors Wuitcu LOWER ANGLE OF FAULTING Effect of Normal Component of Stress Effect of Heterogeneity of Material Possible Influence of Length and Shape Length Shape Effect of Rotational Strain Rotational Strains Developed by Bedding Experimentation: A Single Strong Layer amid Much Weaker Material Several Strong Layers between Very Much Weaker Ones Less Difference in Competency General Discussion Rotational Strain in Homogeneous Material Piling up of Material a Possible Factor How Rotational Strain Develops Fracture Lessening the Resistance Above Greater Resistance and Drag Below Effect of Weighting Résumé R. T. CHAMBERLIN AND W. Z. MILLER bo INTRODUCTION In the literature of structural geology it is commonly stated that rigid materials subjected to non-rotational strain tend to fracture along planes which are inclined approximately 45° to the direction of applied force. This conclusion has been developed partly from a mathematical analysis of stress and strain relations and partly from results observed in the familiar practice of crush- ing cubes of building stone to determine their strength. That 45° is the angle at which rigid materials normally fracture under direct compressive stress appears to be very generally accepted. This angle, therefore, has come to be regarded by structural geologists as the theoretical angle at which thrust faulting, under ordinary con- ditions, should occur. But if the actual angles of dip of a large number of thrust- fault planes in the earth be tabulated and averaged, it is found that the mean inclination is less than 45° from the horizontal. Accord- ing to Leith an average compiled from folios of the United States Geological Survey gives a dip of 36° for planes of thrust faults and 78° for planes of normal faults.‘ An inspection of numerous cross-sections from various other countries gives results in fair agreement with these figures. The average dip angle of thrust- fault planes, as they occur in nature, is considerably less than 45°. While the most prevalent type of thrust-fault plane, that of the — ordinary reverse fault, dips somewhat less steeply than 45°, it still does not depart widely from that governing angle. Neverthe- less, in notable variation from this, field studies in the last few years have brought to the attention of geologists impressive evidence of. the prevalence and the great importance of what may well be called a different genus of fault, namely, the great low-angle overthrust. Its generic characteristics are the very low inclination of its fault plane and the extraordinary horizontal displacement often attained. Such low-angle overthrust faulting has been well described, as it is strikingly shown in the Northwest Highlands of Scotland, where the Moine, Ben More, Glencoul, and other remarkable thrusts form 1C. K. Leith, Structural Geology, 1913, p. 55. ———s LOW-ANGLE FAULTING classic examples of the genus (Fig. 1). In the extreme north SE. of Sutherland the various rock groups overlying the Moine thrust plane can be shown to have been driven westward for a distance of ten miles.2. Horizontal shiftings of comparable magnitude occurred along the Ben More, Glencoul, and other planes of thrusting which le beneath the Moine thrust and add to the remarkable nature of the phenomena. ‘Though since thrown 7) into gentle folds, in many places it is clear that these planes of slippage were originally not far from the horizontal. In some other por- tions of the British Isles analogous phenomena have been observed. Similarly, in Scandinavia the very intense Caledonian deforma- tion manifested itself in horizontal overthrusting of astonishing magni- tude. The vertical displacement is slight, but the horizontal slip is measured in tens of kilometers. Allt’ Ealag AS Section from Elphin to Allt Ealag (about 6 miles in length). = EN Gj aa ‘pp Tp z Cromalt Tz a ae CEA IN g TNS AE Md, 2 SS Pollan 2 < WS NG SRR WW T a Re) = =~ 5 QS S&S = 6 LAS s 1G. 1.—The Scottish Highland type of overthrust. (T;) near Allt Ealag on the right is the Moine overthrust. the section is that of the Ben More thrust. Tk >— ——_ 1B. N. Peach, John Horne, W. Gunn, C. T. Clough, and L. W. Hinxman, ‘‘The Geological Structure of the Northwest High- lands of Scotland,’ Mem. Geol. Surv. of Great Britain, 1907, pp. 463-504. 2 John Horne, 7bid., p. 460. 3A. E. Térnebohm, ‘‘Grunddragen af det Centrala Skandinaviens Bergbygegnad. Kongl,” Svenska Vet. Akad. Handl., Bd. 28, No. 5 (1896), pp. 190-95 and PI. IV; P. J. Holmquist, ‘‘ Bidrag till diskussionen om den skandinaviska fjaillkedjans tektonik,” Geol. Foren. Férhandl, XXUIL (1901), 55-71. 3 Elphin The thrust plane From ‘Report on the Recent Work of the Geo- The gently folded thrust plane (72) which runs nearly the whole length of Two other major thrusts cut through the faulted slices near the left end of the section. 4 I logical Survey in the North-West Highlands of Scotland, Based on the Field-Notes and Maps of Messrs. B. N. Peach, J. Horne, W. Gunn, C. T. Clough, L. Hinxman and H. M. Cadell,” Quart. Jour. Geol. Soc., XLIV (1888), 426, Fig. 10. Note the relation of the overthrusts, or major faults, to the minor reverse faults. 4 R. T. CHAMBERLIN AND W. Z. MILLER In the southern Appalachian Mountains the Rome and Carters- ville overthrusts run parallel to each other for over 200 miles. They are thought by Hayes to show horizontal displacements of not less than 4 miles and 11 miles respectively, and possibly much more. The inclination of the fault planes is here frequently as low as 5°; it is rarely more than 25°.1 The steeper portions of the plane as now seen are largely the result of subsequent warping. Farther north, in Tennessee, a possible continuation of the Cartersville thrust is the Buffalo Mountain fault which, according to Keith, was a low- angle overthrust whose original displacement along the shear plane was at least 20 miles.? Subsequent folding and faulting have so disturbed this fault plane that its original inclination cannot be very closely determined. More to the north, the earlier Taconic revolution also developed low-angle overthrusts. Of these may be noted the Great Western fault of eastern New York,’ the St. Lawrence and Champlain fault, which runs from Vermont to the city of Quebec and beyond,’ and possibly the Cowansville overthrust of Missisquoi and Brome counties, Quebec, though the age of the last has not been closely determined as yet. In any case the measured horizontal displace- ment of the last is 11 miles, and it is thought likely that the actual displacement was much greater.’ It is a nearly horizontal over- thrust, whose plane is very close to the present surface, and along which the Georgian slates on the east have been shoved over the Trenton slates and limestones of the Farnham series to the west. The Rocky Mountains of Montana and Alberta are bordered on their eastern front, throughout at least 350 miles of their extent, tC. W. Hayes, “‘The Overthrust Faults of the Southern Appalachians,” Bull. . Geol. Soc. Amer., II (1891), 141-54. 2 Arthur Keith, U.S. Geol. Surv. Geol. Atlas, Roan Mountain (Tenn.), Folio 151, 1907, P. 9- 3 James D. Dana, Manual of Geology (4th ed.), 1895, p. 528; S. W. Ford, ‘“Obser- vations upon the Great Fault in the Vicinity of Schodack Landing, Rensellaer County, N.Y.,” Am. Jour. Sci., XXIX (1885), 16-19. 4G. A. Young, ‘“‘The Geology and Petrography of Mount Yamaska, Province of Quebec,” Geol. Surv. Can. Ann. Rept., XVI (1906), 9. 5 Robert Harvie, “‘Brome and Missisquoi Counties, Quebec,” Sim. Repi., Geol. Surv. Can., 1914, pp. 98-99. i LOW-ANGLE FAULTING 5 by great overthrusts whose planes dip in under the mountains at low angles. McConnell has estimated that on the South Fork of the Short River in Alberta the horizontal displacement of the Cambrian strata—which here rest upon the Cretaceous—has been about 7 miles, while the vertical displacement amounts approxi- mately to 15,000 feet." In the Glacier National Park of Montana, Willis found the Proterozoic strata which make up the outermost range (here called the Lewis Range) overthrust at least 7 miles upon the Cretaceous of the foothills. The dip of the thrust plane, as determined by Willis by graphic construction, ranges from 3° to 7° 45’.2. More recently Campbell has been able to show that where the Great Northern Railroad crosses the range this great mass of strata has been shoved at least 15 miles northeastward along the Lewis thrust plane, and were the original position of the mountain mass known the distance might prove to be much greater. At the International Boundary the northward continuation of the Lewis thrust has been termed the Waterton Lake thrust by Daly. The known extent of the bodily movement here represented is about 8 miles, as measured on the perpendicular to the line tangent to Chief Mountain and the outermost mountains of the Clarke Range. But the actual movement, according to Daly, has probably been ro miles or more, and may be as much as 4o miles, for “it is not impossible that the entire Clarke Range (the equiva- lent of the Livingston Range of Willis) in this region represents a gigantic block loosened from its ancient foundations, like the Mount Wilson or Chief Mountain massifs, and bodily forced over the Cretaceous or Carboniferous formations.’’4 The Willard thrust discovered by Blackwelder in the Wasatch Mountains of Utah has a displacement, so far as exposed, of about 4 miles, though this is probably but a small fraction of its total R. G. McConnell, Geol. Surv. Can., II (1886), Part D, p. 33. 2 Bailey Willis, ‘‘Stratigraphy and Structure, Lewis and Livingston Ranges, Montana,” Bull. Geol. Soc. Amer., XIII (1902), 331-43. 3M. R. Campbell, ‘‘The Glacier National Park,’ Bull. 600, U.S. Geol. Surv., IQI4, p. 12. ° 4R. A. Daly, “Geology of the North American Cordillera at the Forty-Ninth Parallel,” Mem. 38, Geol. Surv. Can., Part I (1912), p. ot. 6 R. T. CHAMBERLIN AND W. Z. MILLER displacement. Though the fault plane locally has a dip as high as 50° owing to later warping, it averages about 15°." The Bannock overthrust, recently described by Richards and Mansfield, when traced through southeastern Idaho and Utah along its course, now made sinuous by erosion, has a length of approxi- mately 270 miles, and involves a horizontal displacement of not less than 12 miles. The thrust plane itself is a gently undulating surface nowhere steeply inclined, sometimes dipping to the east and sometimes to the west. If this slight plication be the result of subsequent folding, the shear plane must originally have been very nearly horizontal. In eastern Wyoming the Absaroka and Darby faults are really of the overthrust variety, although what remains of these planes shows a higher angle of inclination than most of the preceding. The fault plane of the Darby thrust is, in general, not far from parallel to the bedding of the overthrust sheet. East of Yellow- stone National Park the Hart Mountain overthrust is thought by Dake to show a displacement of not less than 22 miles, making no allowance for recession of the eastern front by erosion. Assuming average thickness for the beds involved, the vertical displacement is over 6,000 feet. In the Alps, so long and carefully studied, some of the most remarkable structures known to geologists are still in process of being worked out. As yet there is lack of perfect accord as to some of the features of their interpretation. They have commonly been interpreted as extraordinary and wonderfully drawn-out overfolds (nappes de recouvrement). Among certain geologists there has developed a disposition to substitute, in interpretation, over- thrust sheets of the Scottish Highland type® for these extreme 1 Eliot Blackwelder, ‘“New Light on the Geology of the Wasatch Mountains, Utah,” Bull. Geol. Soc. Amer., XXI (1910), 517-42. 2R. W. Richards and G. R. Mansfield, ‘‘The Bannock Overthrust, a Major Fault in Southeastern Idaho and Northeastern Utah,” Jour. Geol., XX (1912), 681-709. 3 Alfred R. Schultz, ‘‘Geology and Geography of a Portion of Lincoln County, Wyoming,” Bull. 543, U.S. Geol. Surv., 1914, pp. 84-87, and structure sections. 4C. L. Dake, ‘‘The Hart Mountain Overthrust and Associated Structures in Park County, Wyoming,” Jour. Geol., XX VI, No. 1 (1918), p. 50. 5 Bailey Willis, ‘‘Report on an Investigation of the Geological Structure of the Alps,” Smithsonian Misc. Coll., LVI (1912), No. 31, pp. 1-13; also James Geikie, Mountains, Their Origin, Growth, and Decay, 1913, pp. 116-17. ps gy cmaieon atid som ail lla, amass mimmmaalcaalihl caiarear LOW-ANGLE FAULTING 7 overfolds. If this be the true explanation, it would add to our list this remarkable structure of the Alps as a most pronounced and complicated case of low-angle faulting. Similar structures have been reported from Spain, Euboea, the Balkans, and the island of Timor; in the last case an extensive sheet of shallow water strata, ranging in age from Triassic to Eocene, has been thrust over what appear to be deep-sea deposits of nearly the same age.* Detailed studies elsewhere—practically the world over, indeed— are bringing to light overthrust faults of great displacement along gently inclined planes. This sort of faulting seems, therefore, to constitute a phenomenon of a definite, independent type. It seems to belong to a genus of its own, distinct from the ordinary reverse fault, though the two are no doubt connected by composite types that bind them together. The common reverse fault is defined by displacement along planes neighboring 45° or a little less, and is confined to more limited movement on these planes, while the great overthrusts slide along planes that approach horizontality and involve displacements of astonishing magnitude. ‘Though each great low-angle overthrust is commonly attended by a retinue of reverse faults of lesser magnitude—a fact which suggests that there may be a kinship between them—nevertheless an inspection of any good section, as in the Scottish Highlands, shows a radical difference between the two types. Some distinctive feature seems to be added to simple straight compression to form the low-angle overthrusts. PREVIOUS INVESTIGATIONS Willis has divided thrust faults into four classes, the break- thrust, stretch-thrust, shear-thrust, and erosion-thrust. Of these the shear-thrust and the erosion-thrust are low-angled overthrusts, while the other two classes belong to the more common group of reverse faults. The shear-thrust is a class to cover the conspicuous Scottish Highland type, while the erosion-thrust covers a special case of alternate competent and incompetent strata in which the upper competent formation carrying the thrust is first removed 1G. A. F. Molengraaff, ‘‘ Folded Mountain Chains, Overthrust Sheets, and Block Faulted Mountains in the East Indian Archipelago,’ Compte Rendu, Congrés Géol. Int. (Toronto, 1913), pp. 689-702. 8 R. T. CHAMBERLIN AND W. Z. MILLER from the crest of a broad anticline by erosion. When subjected to further lateral thrust, the upper beds on one limb of this anti- cline, encountering little resistance in front, ride forward over the subaérial surface. Related to this form of erosion-thrust is another special type described by Hayes from the southern Appalachians. Thus the field of the low-angle fault is not an unexplored one, since explanations have been offered for certain special types of overthrusts. The very definite explanation for the Rome and Cartersville overthrusts of the Southern Appalachians was worked out by Hayes as early as 1891.2 The key of this explanation was suggested by the massive and peculiarly competent dolomite formations which alternate with weaker shale layers. In this = 3 Rigidity = eh =— : =—= 3 = ===) Medium ===> Sea Ze ee ES Minimum E “eth Maximum —— ——— SZ <=. > Aa SSS +=<4 Minimum 7 ———— if Fic. 2.—A theoretical section to represent the position of the fault plane (PP’) in the Rome and Cartersville thrusts. From Hayes. type of deformation the strata are thought to have first flexed into a : pair of gentle anticlinal bends some notable distance apart. Between the flexures the strata remained essentially undisturbed. Finally a break occurred near the crest of one of the anticlines, and the thick, competent formations sheared more or less horizontally along a slippage plane which followed closely the bedding of the weak shales (Fig. 2). The erosion-thrusts of Willis and the special form so clearly described by Hayes are dependent upon appropriate rock strata and structure, and thus these explanations, while they fit admirably the © conditions in the southern Appalachians for which they were devised, do not apply to various other overthrusts where the necessary stratigraphic conditions do not obtain. They thus con- stitute a particular type due to special conditions. They do not apply to the very remarkable overthrusts of the Scottish High- « Bailey Willis, ‘‘Mechanics of Appalachian Structure,’ U.S. Geol. Surv., 1 30h Ann. Rept., Part II (1893), pp. 222-23. 2C. W. Hayes, loc. cit., II (1891), 141-54. LOW-ANGLE FAULTING 9 lands, where the thrust planes cut through very heterogeneous assemblages of rock material. Different principles apparently control the latter. It was with a view to obtaining light upon the mechanism of the Scotch overthrusts that Cadell, in 1888, even earlier than Hayes, undertook his experimental researches which since have become classic. In these instructive researches Cadell made use of a pressure box, one side of which could be thrust forward by means of a powerful screw. In this box he built up a succession of layers of plaster of Paris, interstratified with layers of sand, to imitate beds in the earth. After the plaster had set into rigid strata, lateral pressure was applied by means of the screw which moved the pressure block. In this manner, as the final outcome of many trials, he succeeded in imitating rather closely the peculiar imbricate and overthrust structure which the members of the Scottish survey were deciphering from the greatly disturbed terranes of the North- west Highlands.* | Those of Cadell’s conclusions which relate to overthrusting may be quoted: 1. Horizontal pressure applied at one point is not propagated far forward into a mass of strata. 2. The compressed mass tends to find relief along a series of gently inclined thrust planes, which dip toward the side from which pressure is exerted. 3. After a certain amount of heaping up along a series of minor thrust planes, the heaped-up mass tends to rise and ride forward bodily along major thrust planes. 4. Thrust planes and reversed faults are not necessarily developed from split overfolds, but often originate at once on application of horizontal pressure. 5. A thrust plane below may pass into an anticline above, and never reach the surface. 6. A major thrust plane above may, and probably always does, originate in a fold below. 7. A thrust plane may branch into smaller thrust planes, or pass into an overfold along the strike. 8. The front portion of a mass of rock being pushed along a thrust plane tends to bow forward and roll under the back portion. 9. The more rigid the rock the better will the phenomena of thrusting be exhibited. *H. M. Cadell, “Experimental Researches in Mountain Building,” Trans. Roy. Soc. Edinburgh, XXXV (1890), 337-57. 10 R. T. CHAMBERLIN AND W. Z. MILLER The result of Cadell’s experimentation was to produce a concrete picture of the manner in which the complex structure of the North- west Highlands may have developed. As pressure was gradually applied, the artificially prepared strata were first sliced into separate blocks by ordinary reverse slice faults which dipped in the direction from which the pressure was applied. A piling up of the sliced blocks followed. After sufficient piling up had occurred, a low-angle major thrust plane broke through the piled-up mass of slices, and the whole overlying mass rode forward bodily upon this gently inclined plane which Cadell termed the ‘‘sole” (Fig. 3). This behavior would seem to suggest that the heaping up of material LA Witttren.. °*"-~ LO d ttt, ld BLOCK AB =e BLOCK “de Vs Fic. 10.—Drawing of experiment which shows how the inclination of the fracture plane may vary greatly in crossing beds of different competency. Because of the operation of a rotational strain following the fracturing of the brittle competent layers, there resulted nearly horizontal shearing through the weak, clayey layer. The strain ellipses are drawn upon the fracture line to illustrate the variable nature of the strain. _ Strong layers (.S) composed of equal parts of plaster and clay; weak layers (W) of one part plaster and two parts clay. Fig. ro. When pressure was first applied in this experiment, the stronger layers carried most of the thrust, while the softer layers yielded and accommodated themselves so far as was necessary by compacting. With increasing strain the upper strong layer frac- tured at an angle averaging 20°, and this plane of fracture was projected through the overlying clay. With this fracturing a strong rotational strain developed below. This caused almost horizontal shearing through the soft clayey layer, where the pre- vailing angle of the fault plane is found to be less than 5°, and as a result of these shearing stresses the strong plaster layer below was faulted at 12°. Bedding, therefore, where there is sufficient difference in the relative competency of the strata, may be an important factor in 28 R. T. CHAMBERLIN .AND W. Z. MILLER determining low-angle faulting. As shown in Fig. to, in which the ~ ellipsoids representing the axes of strain are drawn upon the beds, the lowering of the angle in this way seems to be the result of shearing stresses and rotational strain. A difference in competency is thus one means of developing rotational strain, and the type of faulting described above comes under the category of rotational strain thus produced. The difference in competency may not be solely because of a difference in the kind of rock, but it may result. also from a very unequal distribution of bedding planes which are planes of weakness. It is to be noted that an abundance of strongly marked, closely spaced bedding planes, in addition to making the competency of the formation less with respect to more massive adjacent formations, also makes splitting parallel to the bedding of the bedded rocks much easier than breaking across the bedding. With less resistance offered in that direction, fault planes crossing very thin bedded shales will be lowered to a certain extent toward parallelism with the bedding. The more bedding planes and the more pronounced they are, the lower the angle of faulting. General discussion.—The Lewis overthrust in the Glacier National Park of Montana, according to the well-known explana- tion of Willis, who would classify it as an erosion thrust, appears to be a case of low-angle faulting controlled by bedding.’ The fault plane, where observed, is located in the Proterozoic Altyn limestone, whose bedding it appears to parallel closely. Above the fracture plane, for the most part, are rigid, brittle, competent strata. What lies below the fault plane is not known, since the oldest formation in the vicinity is the Altyn limestone just above the fault plane. This has been overthrust upon the Cretaceous. In the explanation given by Willis, the sequence of events is, first, gentle folding by which there was developed a low, unsymmetrical anticline whose gentler west limb had a nearly constant westerly dip. Erosion then removed the crest of the fold, thus leaving the west limb a thick sheet of competent strata, unweakened by secondary flexures t Bailey Willis, ‘““Mechanics of Appalachian Structure,’ U.S. Geol. Surv., 13th Ann. Rept., Part II (1893), p. 223 and Pl. LIV, Figs. 6 and 7; ‘‘Stratigraphy and Structure, Lewis and Livingston Ranges, Montana,” Bull. Geol. Soc. Amer., XIII (1902), 331-43. LOW-ANGLE FAULTING 29 and in a position to carry thrusts from the west. Because of the erosion of the crest of the anticline, support from the east limb had been to a considerable extent removed, and frontal resistance to a thrust from the west greatly reduced. With resistance in front lessened and resistance beneath unchanged, or very much less diminished, lateral thrusts developed shearing stresses which caused the overthrust. This type would, therefore, be an overthrust due to rotational strain fostered by the special attitude of the strata and especially by the lessening of the resistance to a forward move- ment of the upper layers because of preceding erosion. ‘The shear- ing then took place along a bedding plane as a line of weakness. The pretty structural explanation of the southern Appalachian overthrusts offered by Hayes was entirely dependent for its work- ing qualities upon appropriate stratigraphic formations of widely different competency. Similarly, though to perhaps lesser degree, the erosion-thrust of Willis is dependent upon appropriate stratig- raphy and antecedent history. Admitting that each of these explanations fits the particular case, or type of cases, for which it was devised (which was probably all that the authors intended), it is clear that an explanation on either of these lines cannot fit the type of overthrust which is so wonderfully displayed in the Scottish Highlands. In these remarkable dislocations the low- angle overthrusting did not occur until after the continuity of bedding over the overthrust area had been completely interrupted and displaced by repeated slice faults at the ordinary angle of 40° to 45°. The Scottish overthrusts did not follow any one weak formation, as did the overthrusts in the southern Appalachians, but cut straight through the various rocks of many previously faulted blocks. It is clear that a more general raison d’étre for low-angle faulting must be sought. 2. Rotational Strain in Homogeneous Material Piling up of material a possible factor.—One of the most char- acteristic features of the Caledonian diastrophism which produced the faulted structure of the Scottish Highlands was the development of a remarkable imbricate structure prior to breaking along the great thrust planes. It seems to be well established, both from the field 30 R. T. CHAMBERLIN AND W. Z. MILLER evidence and from Cadell’s experiments, that the order of events was first slice faulting of the ordinary 45° angle type, and that, after a mass of slices had piled up in this manner, a low-angle thrust plane broke through the mass of slices and the whole mass above rode bodily forward on this plane as a ‘“‘sole.’” Similarly it will be observed that the well-defined Rome over- thrust in the southern Appalachians occurred in the midst of a Fic. 11.—Slice faulting, developing an imbricate structure. The bottom layer was composed of pure clay; the next above of mixed sand and clay; the third was a thin layer of sand; and the heavy competent layer at the top was made of plaster two parts, sand one part and clay one part. The brittle top layer arched up and frac- tured as the first slice fault developed. Each successive slice fault broke out below and in front of its predecessor. series of ordinary reverse faults. Directly south of the town of Rome, Georgia, there are mapped six large reverse faults just to the east of the line of the great overthrust. They are thus in the mass which traveled westward with the overthrust.? Similar slice faults in series, though they cannot be traced continuously into the par- «J. Horne, ‘‘The Geological Structure of the Northwest Highlands of Scotland,” Mem. Geol. Surv. of Great Britain, 1907, pp. 471-76; H. M. Cadell, of. cit., pp. 347-48. 2C. W. Hayes, U.S. Geol. Surv. Geol. Atlas, Rome, Ga., Folio 78, 1902. Structure Section Sheet. LOW-ANGLE FAULTING 31 ticular breaks near Rome, are especially numerous in the same relation to the overthrust throughout the southwest portion of the quadrangle. Somewhat analogous relations are to be noted else- where. Directly in front of the Lewis thrust in Montana there is represented on the structure sections a series of slice faults formed as if in preparation for another overthrust which presumably, if the deformation had been carried further, would have broken through lower than the Lewis slip and to the east of it.* Fic. 12.—Deformation of specially shaped mass. Slice faulting resulted. In this particular experiment the pressure blocks were not held rigidly in place by con- trolling flanges, but were free to rise or become tilted. A relationship between a piling up of rock masses and the development of the low-angle overthrust has therefore been suggested. It might at first seem possible that lateral thrusting applied upon the piled-up mass, thus bringing forces to bear in a higher plane than would be the case if there were no piling up, would, on the lever-arm principle, develop a rotational strain which would cause fracturing at a lowered angle. To test this question experimentally, there was molded in the box a homo- geneous mixture of clay and plaster, which was high adjoining both pressure blocks and low in the middle. The rectangular 1 Eugene Stebinger, ‘‘Geology and Coal Resources of Northern Teton County, Montana,” Bull. 621, U.S. Geol. Surv., 1916, Pl. XV. Be R. T. CHAMBERLIN AND W. Z. MILLER outlines of the prepared block, by exaggerating any case of piling up likely in nature, ought not to fail to reproduce the low angles, if such be due to piled-up material acting in this way. The results are shown in Figs. 12 and 13. ‘There was first slice faulting on the right-hand side, from which the pressure came. Each successive fault broke below the previous one as the mass was compressed more and more. This is similar to the experience of Cadell. As compression went on, the planes of the earlier faults became Fic. 13.—Same as Fig. 12. After further compression distorted and obscured by the later deformation. At length a low- angle fracture broke through from the left, or resistance block side. This was no doubt determined by lines of least resistance due to weakening by the previous fracturing. To test the matter further a mold of pure paraffine was pre- pared in essentially the same shape. The paraffine had the advan- tage of being more nearly homogeneous than the clay-plaster combination. In the two trials made, fracturing proceeded directly across the elbow at approximately 45° (Fig. 14). Lest the right- angled elbow might play an unsuspected part in determining the angle of splitting, paraffine was molded into a block having the shape shown in Fig. 15. When pressure was applied the block faulted at the farther end. It faulted at the farther end because LOW-ANGLE FAULTING Be the force per square inch was greater there (owing to the smaller cross-section over which it was distributed) than at the other end _near the pressure block where it was distributed over a larger area of cross-section. Rupture occurred where the intensity of stress was greatest, even though it was farthest removed from the pressure block. The fault averaged 42° for its whole length. The angle shows that it was caused by a non-rotational strain. The same was true also of the two previous tests. The shape of the block, at least to the extent of the variations tried in these experi- ments, apparently does not change the nature of the strain. But perhaps, after all, only a non-rotational strain could develop under RESISTING PRESSURE Frc. 14.—Block similar in shape to that shown in Figs. 12 and 13, but composed of paraffine. A 45° fracture developed. the conditions of these experiments, since the pressure block is guided rigidly forward by the controlling flanges of the machine and so cannot turn. But one may conclude, nevertheless, that a piled-up mass having a higher standing cross-section to be pushed forward, does not, of itself, add a rotational element to the strain when laterally compressed, nor, so far as this principle is concerned, does it lower the angle of fracture. How rotational strain develops fracture-—To show how a rota- tional strain will deform such a block as was used in the experiment just described, another block of paraffine was cast in the same mold and subjected to a rotational strain in the following manner. As before, the pressure was applied from the same long side, but instead of being applied against the whole surface of that side it was applied ‘only to the upper half of it. The resisting block, as before, but- tressed the whole of the shorter left-hand side. With the opposing forces acting horizontally at quite different elevations, a rotational couple was developed. As the strain slowly increased the paraffine 34 R. T. CHAMBERLIN AND W. Z. MILLER near the pressure block first yielded somewhat by plastic deforma- tion. Then, as a result of the shearing stresses, it started to break along a very low angle of fracture near the bottom of the block (Fig. 16, break A). The experiment was stopped at this point and the block removed for study. After a rest of a few days the deformed block was again placed in the crushing machine and Fic. 15.—Deformation of specially shaped paraffine block under non-rotational strain. Pressure applied uniformly upon right-hand face. Fracture averages 42°. pressure applied as before. But instead of further splitting along the old line of breakage near the bottom of the block, an entirely new break occurred at a much higher level (Fig. 16, break B). This new fracture extended completely across the block. Though irregu- lar in detail, its general direction was very close to horizontal. To verify these results, a new block of paraffine was cast in the same mold and pressure was again applied in the same way. The result again was breakage along a nearly horizontal shearing plane (Fig. 16, break C). In breaking out at the surface, however, the fault plane, LOW-ANGLE FAULTING 35 in both cases, turned upward, producing a considerably steeper angle in the immediate vicinity of the surface. Experiments therefore show that the effect of a strong rotational strain, even in homogeneous material, is to produce shearing and complete rupture, essentially parallel to the direction of the applied force. If the thrusting be in a horizontal direction, the plane of rupture will approach horizontality. In these experiments it was noted that the low-angle shearing required fewer turns of the screw and thus the application of less force than the 45° fracture from Fic. 16.—Deformation of paraffine blocks (same mold as block in Fig. 15) under rotational strain. Pressure applied only to the upper half of the right-hand face. The fracturing, though irregular, was not far from horizontal. non-rotational strain. This merely bears out the well-known fact that the resistance of materials to shearing stresses is much less than to direct compressive stress. Hence the disposition to shear if con- ditions allow. Lessening the resistance above-—Deformation by rotational strain may thus be developed in homogeneous material by sufficiently increasing the effective differential stress in the upper portion of the mass with respect to that in the lower portion. It is the greater unbalanced pressure in one portion over another which is effective. This unbalancing of pressure may be accomplished in several ways. Within the earth it may be produced either by increasing the lateral 26 R. T. CHAMBERLIN AND W. Z. MILLER thrusting in the upper portion or by lessening its resistance, while the lower portion remains unchanged. It may also be developed by diminishing the thrust below or by increasing the resistance in the lower part, while conditions in the upper portion remain essentially the same. Or it may be accomplished by some combination of these. The first process facilitates deformation in the upper part; the second retards deformation in the lower part. Whatever affects the ratio influences the character of the deformation. The greater the difference developed the greater the shearing tendency. The intensity of thrusts and the resistance at different horizons in the earth should therefore be a vital factor in determining shearing. At the present time the location and intensity of lateral thrusts in the earth are so imperfectly understood that a treatment of that topic is reserved for further information. Rather more, however, is known concerning the resistance offered. Factors which either lessen the resistance above, or increase it below, may play a part in overthrust faulting. The resistance above may be diminished in several ways. The erosion-thrust of Willis and Hayes already discussed is a clear- cut illustration of how it may be accomplished in heterogeneous materials. Resistance above is here reduced by erosion which removes the heavy, competent upper layers from the crest of an anticline. The resistance of the remnants of the upper layers to forward movement having been sufficiently diminished in this way, this more movable portion shears nearly horizontally along a bedding plane as a line of weakness (see Fig. 2). Shearing along bedding planes and the control of overthrusts by differences in the competency of the beds are related phenomena. In homogeneous material the resistance above is lessened by other means. The experiments of Cadell indicated that before the low-angle overthrust occurred there was first slice faulting and the piling up of slices. Slice faulting to a remarkable extent was asso- ciated with the Scottish overthrusts and to a certain extent with those in the southern Appalachians. If the mere piling up of materials, as such, does not introduce a rotational element to the strain and so lower the angle of fracture, nevertheless the repeated slice faulting and moving of fault blocks do have an effect upon LOW-ANGLE FAULTING 37 the resistance of the faulted strip. The slicing and secondary shattering would seem to weaken the superficial sheet which has suffered the faulting. It may be perhaps that the superficial shell, freer to move as a general mass than it was before slicing, while the lower, deeper levels have not been equally affected by what has taken place, now finds it easiest to slide bodily forward over the less movable lower portion. If this be true, it would make the rupturing by the preparatory slice faulting far more important in the develop- ment of low-angle overthrusts than the piling up of material. Greater resistance and drag below.—With horizontally directed compressive stresses in operation rotational strains would also tend to be produced by the co-operation of any factor which increased the resistance of the deeper portion of the rock mass involved, while the resistance of the more superficial portion to such stresses remained the same. The far-reaching experimental studies of Dr. Adams and his colleagues have shown that,on account of the increasing rigidity of the rocks due to cubical compression from the weight of overburden, resistance to deformation in the earth should increase with increasing depth below the surface.’ It is concluded that with ‘increasing depth greater and greater stress differences are required to deform the rocks. From this principle it would seem to be a legitimate deduction that, for a lateral thrust of given magnitude, rock deformation should take place more readily near the surface of the earth than at a greater depth beneath the surface, and that in any case (barring the effects of local heating, or liquefaction) deformation should become less with depth, unless the magnitude of the stress differences which cause the thrusting increases as rapidly with increasing depth as does the resistance offered by the rocks.’ t Frank D. Adams, ‘‘An Experimental Contribution to the Question of the Depth of the Zone of Flow in the Earth’s Crust,” Jowr. Geol., XX (1912), 97-118. Since this was written, the principle of increasing resistance to deformation with depth below the surface of the earth has been strongly affirmed by Adams and Bancroft as the result of further experimental researches. (See Frank D. Adams and J. Austen Bancroft, ‘‘On the Amount of Internal Friction Developed in Rocks during Deformation, and on the Relative Plasticity of Different Types of Rocks,” Jour. Geol., XXV [1917], 5907-637. Also Louis Vessot King, ‘‘On the Mathematical Theory of Internal Friction and Limiting Strength of Rocks under Conditions of Stress Exist- ing in the Interior of the Earth,” Jowr. Geol., XXV [1917], 638-58.) 28 R. T. CHAMBERLIN AND W. Z. MILLER In general, so far as these principles hold, there should be a tendency, strong or feeble according to the quantitative factors, for surficial shearing over a less movable portion below. Rotational strains thus brought into being might conceivably in some cases be a primary cause of low-angle shearing, or in other instances might co-operate as a secondary factor with other more important causes in producing a similar result. Many glaciers, notably those in North Greenland, have devel- oped in places horizontal shearing planes which are often grouped into distinct zones.t. The englacial drift is definitely arranged along these planes of movement, and this.in turn influences the rate of melting on the steep edge of the glacier, so that these planes of shear have, in many instances, become very conspicuous. ‘These lines of débris are especially prominent in the lee of an embossment of rock over which the glacier has just passed. In most cases the shearing planes may be interpreted as due to the greater resistance of the rock knobs below. They seem to be further developed by the increased load of débris in the lower part of the glacier and by drag on the bottom beneath the moving mass. The rotational strain thus engendered causes nearly horizontal slippage of the upper portion over the lower. Where formed in the lee of an embossment of rock another factor enters to increase the rotational element. The rock mass protects the lower portion of the glacier from much of the push which the upper portion is receiving.’ Thus while the upper portion of the ice is free to move forward, not only is the resistance of the lower portion of the ice to forward motion increased, but at the same time the actual thrusting to which that portion is subjected is diminished. C. E. Decker has described various minor folds and small thrust faults, mostly of Quaternary age, which affect the strata close to the surface in northeastern Ohio and northwestern Penn- sylvania.2. The fault planes of these thrusts are commonly inclined at low angles (Fig. 17). If their proximity to the surface is of real tT. C. Chamberlin, ‘Glacial Studies in Greenland,” Bull. Geol. Soc. Amer., VI (1894), 203-10. 2T. C. Chamberlin, op. cit., pp. 207-8. 3 Charles E. Decker, unpublished manuscript. LOW-ANGLE FAULTING 30 significance, it would suggest a strong tendency of layers near the surface to shear over less movable layers below. E. EFFECT OF WEIGHTING Although the piling up of faulted slices does not of itself cause the development of rotational strain, when the mass is subjected to horizontal compression, it may indirectly bring about that result by weakening the resistance of the upper portion owing to the Fic. 17.—Thrust fault in Chagrin shales. On Paine Creek, 6 miles east of Paines- ville, Ohio. Fault plane dips 15°S.W. Throw 1 ft.,; heave 2 ft. 11 in. Charles E. Decker. preliminary fracturing. If, in addition, the mass piled up is of sufficient magnitude, it may theoretically affect the result in another way owing to the fact that the additional weight of the piled-up mass adds a new force at right angles to the horizontal thrust. Figure 18 will illustrate the behavior of this force. In this diagram the horizontal thrust was taken to be three times the vertical force due to gravity. The resultant of these two forces will be inclined downward 18° 26’ from the horizontal. Frac- turing as the result of these two forces will be determined by the direction of this resultant of forces. As this is inclined 18° 26’ downward from the horizontal, faulting, even though it should take 40 R. T. CHAMBERLIN AND W. Z. MILLER place at an angle as high as 45° upward from the resultant of the forces, would still be only 26° 34’ from the horizontal. The relative magnitude of the horizontal thrusting force and the weight of the heaped-up mass determines how much the angle of faulting, under the given conditions, will be diminished from 45°. If the weight of the load gave a force equal to half that of the lat- eral thrust, the angle would be lowered because of this factor to the extent of 26° 34’, thus making it 18° 26’ from the horizontal. If the vertical force due to the extra load amounted to one-fourth the horizontal thrust, the angle of faulting would be lowered approxi- mately 14° 2’, leaving it 30° 58’ from the horizontal. The resulting — angle for the various stress ratios may readily be calculated. i> Lu, 39404 WOILYSA HORIZONTAL THRUST Fic. 18.—Diagram to illustrate the position of a fault plane inclined 45° to the resultant of forces. The horizontal thrust is here taken to be three times the vertical force. Result: the fault plane will be inclined 26° 34’ from the horizontal. The effect of adding load, and hence additional force acting downward, is to subject the material under thrust to increased cubical compression. According to the principles so strikingly worked out by Adams and his colleagues, the effect of this should be to increase the internal resistance of the material and thus necessitate a much greater stress difference to initiate deformation than would be required without the additional load. Greater stress difference necessitates much greater lateral thrusts. As a result faulting may be hindered or even prevented altogether until much greater thrusts are developed. It may also, in conse- quence, be caused to take place elsewhere, as, for example, some distance beyond the edge of the loaded area. In our experiments with weighting the faulting most frequently appeared at the surface close to the border of the weighted portion, the fault plane dipping LOW-ANGLE FAULTING AI under the heavily burdened portion (see Fig. tg). But in these experiments the loads were relatively light. A load light in proportion to the horizontal stress will thus influence the angle of fracture, depending upon the ratio of vertical and horizontal stresses. A load very great in proportion to the horizontal stress will prevent faulting altogether within the loaded area. The influence of the load upon the angle of thrusting will therefore reach a maximum value somewhere between a load which - Fic. 19.—Effect of local weighting in locating the position of faults. The material to be faulted was clay stiffened with plaster; the added load was damp sand. In experiments of this sort the fracture plane most frequently appeared at the surface close to the edge of the piled-up overburden. is light and a load which is heavy relative to the horizontal stress. What the proper ratio for the maximum effect will be cannot well be determined until more is known of the limiting strength of rocks under stress.‘ Some idea, however, may be gained possibly by a rough inspection of the factors involved. The stress difference necessary to cause faulting at a given depth in the earth would need to be sufficient ‘to exceed the sum of the crushing strength of the given rock at the surface, plus the weight of overburden which must be lifted, plus again the increased strength of the material ‘Louis V. King, “‘On the Limiting Strength of Rocks under Conditions of Stress Existing in the Earth’s Interior,” Jowr. Geol., XX (1912), 119-38. 42 R. T. CHAMBERLIN AND W. Z. MILLER resulting from the compression under the load. Suppose there were 10,000 feet of rock piled up above the plane along which the fault is to occur. Assuming a specific gravity of 2.7 for the rock, the pressure resulting from this column will amount to about 11,760 lbs. per square inch. If the stress along the axis of greatest stress, which is here horizontal, be taken to be three times this, it would need to be 35,280 lbs. per square inch. As the axis of least stress is vertical, the stress difference would amount to 23,520 lbs. per square inch. To cause faulting, this stress difference must equal the crushing strength of the rock under surface conditions augmented by the increased strength of the material induced by the hydrostatic pressure or cubical compression. This increased strength because of the depth is considerable, but pending more work of the type carried on by Adams and his colleagues this is not easily evaluated." | However, this stress difference clearly would not be sufficient at this depth to deform the stronger rocks, like granite, and probably — not rocks of average strength, though very likely it would be sufficient to deform the weaker rocks. Under these conditions, if the horizontal thrust were less than three times the vertically acting force, the stress difference would be proportionately still less effective in deformation. A ratio of more than three to one would, on the contrary, be more effective. A ratio of thrust to the weight of not less than three to one would seem to be required for extensive faulting through rock formations of average strength under a load ranging up to 10,000 feet of rock. This might lower the angle of faulting by 18° or less. But this reduction in angle from 45° falls far short of developing the approximately horizontal slippage planes of the great overthrusts. With loads greater than 10,000 feet of rock, the resistance of the underlying rock is still further increased. While the increase in resistance probably does not mount up in direct proportion to the increase in balanced pressure, nevertheless for any thicknesses of rock likely to be piled up by diastrophic agencies there probably would not be a very radical change in the ratio of axes of stress necessary for faulting. At ‘More data are now available. See Frank D. Adams and J. Austen Bancroft, Jour. Geol., SXV (1917), 597-637; also Louis Vessot King, ibid., XXV (1917), 638-58. LOW-ANGLE FAULTING 43 best only a part of the lowering of the angle from 45° can be explained in this way. If the low angle of the great overthrusts were solely a matter of load steadily accumulated by piling up slice fault blocks, then each successive slice fault should break through at a progressively lower angle. ‘There should be a complete gradation from the first-formed fault near 45° to the final overthrust approaching horizontality. While some progressive lowering of the angle of the successive slice faults is to be noted in some Scottish Highland sections and elsewhere, nevertheless there appears to be a great final jump from the minor slice faults to the great horizontal overthrust. F. RESUME The great overthrusts which are now coming to be recognized as a prevalent and commanding type of mountain structure are the result of conditions differing considerably from those which produce ordinary reverse faults. The distinguishing features of the over- thrusts are the extremely low angle, which often approaches hori- zontality, and the very great displacement along the plane of slippage. The great displacement is made easier by the gentle slope of the fault plane. The low angle of the fault plane is the net result of the operation of several factors. Among the factors which will lower the angle of faulting from the theoretical 45° may be listed the following: t. The normal or direct stress which, along planes inclined 45° to the line of application of the force, has an intensity as great as that of the tangential stress. It acts as a frictional resistance to shearing by the tangential stress. The lower the angle of the fracture plane, the less will be the frictional resistance due to the normal component of the stress. Hence the tendency to fracture at angles below 45°. 2. Rotational strain, which will lower one of the planes of no distortion (shearing plane) from 45° in pure non-rotational strain to o° in the extreme case of rotational strain. Rotational strains may be developed from horizontal compressive stresses: (a) in homogeneous material: (1) by any factors which will increase the intensity of the tangential stress in the upper portion of the mass 44 R. T. CHAMBERLIN AND W. Z. MILLER undergoing thrusting with respect to that in the lower portion; (2) by any factors which will lessen the resistance in the surficial portion without proportionately changing that below; and (3) by any factors which will increase the resistance of the deeper portion of the zone subject to thrusting while the upper portion remains ‘freer to yield; (6) in heterogeneous material by bedding, or similar structures, which present differences in competency of the right sort and thus call into operation some of the foregoing factors. 3. Preliminary piling up of material in the first stages of deformation, thus increasing the load and the vertically acting gravitative force. The combination of the horizontal thrusting force and the vertical gravitative force gives a resultant which is inclined downward from the horizontal. Even should faulting take place in a plane 45° from this resultant, it would still be inclined less than 45° from the horizontal. 4. Possible minor factors, as heterogeneity of material, length of deformed mass with respect to its other dimensions (after analogy of long column), shape of deformed mass, etc. To these factors, operating in various combinations according to the individual peculiarities of each particular case, are attributed the low-angle fault planes of the great overthrusts. THE HART MOUNTAIN OVERTHRUST AND ASSOCIATED STRUCTURES IN PARK COUNTY, WYOMING: ¢. L. DAKE Tulsa, Oklahoma INTRODUCTION Field work during the summer of 1916 brought to light what is believed to be one of the most interesting major thrusts yet described in the northern Rocky Mountains. So far as the writer is aware, the true nature of this fault has not heretofore been described, although Fisher? refers to it in discussing the structure of Hart Mountain. The area studied embraces a narrow strip of territory lying between the area covered by Fisher’s Big Horn Basin report, just mentioned, and the Absaroka quadrangle on the west. STRATIGRAPHY The stratigraphy of the region is essentially the same as that given by Fisher in his above-mentioned paper describing the area adjacent on the east. The divisions of the Cretaceous adopted in mapping are those used by Lupton, since they represent more detailed work than was done by Fisher. Hewett‘ has also described the stratigraphic column in some detail in the region immediately east of the area mapped:-by the writer. The follow- ing table of formations is largely compiled from the three reports mentioned above. Detailed description of the various strati- graphic units will not be given, except in the case of the so-called Fort Union(?) regarding the age of which there may be some question. t Published by permission of the Wyoming State Geologist. 2 C. A. Fisher, U.S. Geol. Survey, Prof. Paper No. 53, p. 37. 3 Lupton, ‘‘Oil and Gas near Basin, Wyoming,” U.S. Geol. Survey, Bulletin 621L. 4 Hewett, ‘‘The Shoshone River Section,” U.S. Geol. Survey, Bulletin 541, pp. 89- 113. 4 OL 46 Cu DAKE Fort Union( ?).—Since there is some uncertainty as to the age of the formation described under this name, and since upon this formation more than on any other depends the dating of the fault in question, it seems worth while to give a detailed description of its relations and lithologic character. TABLE I TABLE OF FORMATIONS . 98 Thickne System Formation Characteristics aa "Eee Quatennany eee |e een Glacial till and terrace gravels ? Mertianyer ioral tetcruce cate spen tens Andesitic breccias and lavas ? Cretaceous or Tertiary....| Fort Union(?) | Buff to yellow sandstones, conglomerates, 300-++ red and gray shales Codyeee een se Gray to black shale, sandy near top 2,000 Bromblersears nie Gray sandstones and shales, with ben-| 500 tonite Cretaceous... .|{_ Thermopolis -and Mowry..| Gray and intensely black shales, with| 900 sandstone and bentonite (Cloverly........| Gray cross-bedded sandstone and shales} 110 Jurassic or Cretaceous..| Morrison...... Variegated red, gray, maroon (etc.), 500 shale, and sandstone jurassic Sundance...... Greenish-gray shales and sandstones and| 500 thin fossiliferous limestones Permo-Trias...| Chugwater.....| Red sandstones and shales with gypsum| 750 imalbateaerc see Massive gray limestone. 100 Tensleep...... Massive gray sandstone and brown 100 Carboniferous.. quartzite Amsden....... Red shale and red to gray limestone 150 Madison.......| Massive gray limestone, conglomeratic] 1,000 at the base in places Ordoviciana | Bighormesn se Massive gray limestone 300 Cambrian.....| Deadwood.....| Conglomerates, sandstone, and limestone 800 iBre-Cambriany| sae eerie Red granite, gneisses, and schists ? It consists of an unmeasured thickness of alternating beds of yellow sandstone with red and white or gray clays. The sandstones vary from buff to bright yellow, and occur in several beds from 2 to 20 feet thick. They are cross-bedded on a large scale, and contain many concretions of brown sandstone varying from a few inches up to 10 feet or more in diameter. The concretions are harder than the matrix and weather out in large numbers, occurring abundantly over the surface of the ground. At many places the sandstones are finely conglomeratic, the pebbles averaging between one-fourth THE HART MOUNTAIN OVERTHRUST 47 and one-half inchin diameter. Red granite, basalt, quartzite, sand- stone, black chert, brown chert, and shale make up the bulk of the pebbles. The shale members of the formation are dominantly gray, but contain many red layers. Thin lignitic seams were noted, but no leaves were found in a sufficient state of preservation to permit identification. One thin seam of black coal was found in the formation. At one point the beds were seen to rest on the Cody with slight angular unconformity, although at several other points they par- take of the folding of the older formations. Though no fossils were found in these beds, the abundance of the red clays, the pebbly character, and the slight angular uncon- formity at the base all seem to favor correlation with what Hewett has called the Fort Union. It is true that Hewett has made no mention of the large concretions which are so abundant, and the writer noted similar ones from the basal Laramie of Fisher (Hewett’s Gebo), not far to the east of the area mapped. At the same time similar concretions were noted, however, at a much higher horizon in Fisher’s Laramie, in what is believed to represent Hewett’s Fort Union. Fisher describes such concretions as occurring in the Laramie, but does not indicate the exact horizon. Along the North Fork of Shoshone River these beds trace continuously into what Hague’ has mapped as Pierre and Fox | Hills. At the same point, however, Hewett? calls them ‘‘ Tertiary sandstones and shales probably of Wasatch age.” Structural reasons will be given later for believing that they are earlier than Wasatch. STRUCTURE The major thrust.—The main plane of fracture occurs at or near the base of the Madison (Mississippian) limestone, which has been thrust out over beds varying in age from Madison to Fort Union( ?). At the southernmost point, where the fault was located, the lime- stones rest on sandstones of Fort Union(?) age, but toward the north the stratigraphic throw decreases until at the northernmost ™ Hague, Absaroka Folio, U.S. Geol. Survey, Folio 52. 2 Hewett, ‘Sulphur Deposits in Park County, Wyoming,” U.S. Geol. Survey, Bulletin 450, p. 478. 48 Cue a DAKE extremity the Madison rests on the Chugwater Red Beds. For the most part the Mississippian limestones are more resistant than MAP OF HART MOUNTAIN OVERTHRUST PARK COUNTY WYOMING PLZ > a | Shoshoge N ge ey, ve CO= Dectncod ard Bighor. AI = Modisen P =Amsocn, Tersleep & Lirbhar RP = Chegweter Red Beds, Scole:— Ove mch= Er ght miles the beds on which they rest, and this condition results in a pro- nounced Madison escarpment along most of the course of the fault. THE HART MOUNTAIN OVERTHRUST 49 This escarpment is a very striking feature of the scenery along the Cody Road to Yellowstone Park, just west of the Shoshone Reservoir. Traced westward up the valley the fault on either side passes beneath the Tertiary volcanics and is lost. It passes in a bold escarpment around the east end of the divide between the North and South forks of Shoshone River and extends for several miles up the north side of the latter valley, where it is again buried beneath the Tertiary breccias. It does not reappear on the south side of the valley of the South Fork except as an isolated peak of Madison resting on Fort Union(?) at the east end of Carter Moun- tain in Sec. 36, T. 51 N., R. 103 W. Several miles south of this point, however, abundant Madison bowlders are noted along the slopes on the east and southeast of Carter Mountain, indicating that the fault block of Madison probably occurs buried beneath the lavas of that mountain, or even possibly outcropping in small overlooked exposures along the lava scarp of that divide. On the north side of the North Fork Valley the scarp makes a sharp re-entrant where Trout Creek cuts through the faulted block. In Sec. 6, T. 52 N., R. 103 W., the scarp bends abruptly northwestward and extends for a long distance along the west side of Rattlesnake valley, overlapped at several points by Tertiary volcanics, beneath which it passes at the head of the valley. In this distance the Madison rests on successively older beds, from Cody shale to Chugwater “‘Red Beds.” The trace of the fault could not be found on the east side of Rattlesnake valley, from which the block of faulted Madison has probably been largely removed by _later erosion. It is suspected that fragments of the block still remain on the top of Rattlesnake Mountain anticline, but if so they probably rest, Madison on Madison, and have not been detected. Along the divide between Rattlesnake valley, Pat O’Harra valley, and Dead Indian valley the trace of the fault plane could not be found, either because wholly removed by erosion or because Madi- son was faulted on Madison and not detected. In Sec. 3, T. 54N., R. 104 W., the fault plane again emerges from beneath a small patch of Tertiary breccia and traces easily northwestward, to the point where the wagon road crosses Dead Indian Ridge, a distance 50 CP DAKE of four or five miles. Throughout this distance the Madison rests on the Red Beds. In Sec. 22, T. 55 N., R. 104 W., two prominent hills of Chugwater are capped with isolated patches of Madison. In Sec. 16, T. 55 N., R. 104 W., the trace of the fault plane is lost, probably because it passes wholly into the Madison, where it is not easily detected. This view is supported by the fact that the Madison in this region appears to be excessively thick, as though repeated. Not far from the center of T. 54 N., R. 102 W., occurs an isolated peak known as Hart. Mountain. It consists of a cap of several hundred feet of Madison limestone, entirely surrounded by late Cretaceous and Tertiary sediments. It has been described‘ as due to a circular fault. Because this mass of Madison is entirely isolated in outcrop, it is not possible to demonstrate the continuity of the major thrust, just described, to this point. But the suppo- sition hardly admits of doubt that Hart Mountain constitutes a portion of the large fault block so widely exposed to the west, especially in view of the similarity of stratigraphic units involved. The extreme irregularity shown by the fault trace is largely due to erosion, in part to later deformation, since the fault plane dips at various but low angles at various points. The north and south extent of the fault has been proved for over 25 miles in a straight line and for more than double that distance measured along the sinuosities of its course. Exclusive of the Hart Mountain outlier the easternmost and westernmost exposures are separated by a distance of 7 miles; including Hart Mountain, by about 16 miles. At the westernmost exposure the fault passes beneath the Tertiary andesites and is lost. At this point the Madison rests on the Fort Union (?), which in turn . can be traced without break at least 6 miles farther west. If the movement was from the west eastward, as will be shown later, the fault plane must pass at least this far west, hidden below the lava, but cut through by erosion before the lava was poured out. If this is the case, the amount of displacement must have been not less than 22 miles, making no allowance for recession of the eastern front by erosion. Using average figures for the thickness of the beds involved, the vertical displacement is over 6,000 feet. tC, A. Fisher, Joc. cit. THE HART MOUNTAIN OVERTHRUST 51 There is nothing sufficiently regular about the dip of the fault plane to indicate the direction of movement. The fact, however, that the Cretaceous and Tertiary rocks in the Bighorn Basin to the east are continuous and but slightly disturbed for nearly too miles indicates plainly enough that the faulted block did not come from the east. The thick cap of volcanics to the west makes it impos- sible to expect any evidence in that direction, but in spite of that is seems clear enough that this block moved from west to east. The South Fork thrust.—Along the lower valley of the South Fork of Shoshone River a second thrust fault is exposed on both sides of the valley, below the Hart Mountain thrust already described. As far as this fault could be traced, Sundance was found resting on Cody or Fort Union(?), and a section from the river level to the top of the ridge on the south side of the valley reveals the following situation: Top Madison -2e. Major thrust Fort Union( ?) Cody Frontier Thermopolis and Mowry Cloverly Morrison Sundance a.) «Minor thrust Cody Bottom On the north side the section is as follows: Top Madison ee Vajyorthrust Frontier Thermopolis and Mowry Cloverly Morrison Sundance . . . . Minor thrust Cody Bottom 52 Cl DAKE This fault seems to pass wholly into Cody shales both to the north and south and cannot be traced more than 6 or 8 miles. It probably shows again, in Sec. 11, T. 52 N., R. 104 W., on the north side of the North Fork, where abundant Sundance fossils are found on Cody shale slopes, just at the base of the main Madison scarp. It has a horizontal displacement approximating 10 miles, and a vertical movement of about 3,000 feet, or nearly 10,000 feet for the two faults combined. Near the west line of T. 51 N., R. 103 W., the trace of this fault, which lies about S. 45 W., on both sides of the valley is suddenly lost in an area of intense brecciation, which is believed to mark the site of a transverse fault. The transverse fault seems to have shifted the trace of the thrust northwest about a mile, and west of the point of disturbance the thrust can be traced only on the south side of the valley. On the north side it is probably buried beneath Tertiary lavas. It has not been possible to prove the identity of the thrust planes east and west of the transverse area of disturbance, but the similarity of stratigraphic relationships — [Sundance on Cody and Fort Union( ?)| seems to indicate the possi- bility of the foregoing explanation. The thrust plane appears to have been sharply folded along an axis lying about northeast and southwest, parallel to the trend of the South Fork Valley. At one point where a deep gorge cuts the axis of a sharply overturned anticline, not far south of Ishowooa Post-Office, yellow sandstones are exposed in a very small area beneath typical Sundance beds. The sandstones carry no fossils, but are similar in appearance to the Fort Union(?), and if of Fort Union(?) age the exposure represents a ‘‘Window”’ or “ Fen- ster,’ such as has been described by several writers, in connection with major thrusts elsewhere. Beartooth fault zone.—Along the eastern edge of the Beartooth Plateau, from the Clark Fork to the Montana line, is a zone of thrust faults, probably related to the same forces producing the faults already described though not continuous with them. The southern extremity of this zone lies about 8 miles north and 4 miles east of the northernmost point to which the Hart Mountain thrust was traced. This group of faults, all of which are associated with THE HART MOUNTAIN OVERTHRUST 53 overturned folds, in places carries the Pre-Cambrian granite out over the “Red Beds.”’ The fault planes could not actually be observed, but are undoubtedly much steeper than the plane of the Hart Mountain thrust. The amount of horizontal displacement was not determined. Mechanics of the faulting.—While no theoretical discussion of the mechanics of these great faults is proposed, it seems worth while to present some observations which may ultimately help to throw light on the problem: 1. The fault contact is practically everywhere concealed by talus from the Madison cliffs, but at several places it could be located within a few feet, and the zone of crush breccia is notably thin at most points. 2. The great limestone block above the fault plane is little folded. But while it presents the general aspect of a nearly flat- lying horizon, locally the dips are high, as a result of numerous small normal faults which appear to have been the result of the settling of the great block after the thrust ceased. 3. The soft shales below the major thrust plane, while much crumpled at places, are nearly horizontal and almost undisturbed over considerable areas where exposed by erosion several hundred feet lower than the major thrust. HISTORY The history of the faulting in this region depends, for its correct solution, on the careful determination of the age of the beds herein called Fort Union(?) and on a knowledge of the relation of these beds to the faulted block of Madison limestone. The second part of the problem is comparatively simple, and will be discussed first. For long distances these beds occur at the foot of the Madison scarp, and there was little question from the very first that they passed beneath the Madison block. In view, however, of the statement of Hewett that these beds are probably of Wasatch age, it was thought possible that the sandstone might have been de- posited, after the faulting and erosion, against the foot of the limestone scarp, since it was nowhere possible to find an actual contact of the Madison resting directly on the Fort Union(?). 54 CED AGE Ti, however, the sandstones were laid down against the base of these high cliffs, they should contain a coarse and abundant angular limestone conglomerate, whereas a careful search nowhere revealed any Madison limestone in any of the sandstone of the formation. This constitutes abundant evidence that the Fort Union(?) actu- ally passes beneath the fault block and does not lap against its foot. As to the second problem, the exact age and equivalence of the beds called Fort Union(?), a less definite conclusion is possible. They are younger than the Cody, upon which they rest with slight angular unconformity at places. The fact that they are involved in the major faulting and folding makes it probable that they are not Wasatch, as Hewett has suggested, since beds of that age are known to cover similar major faults in Idaho.t The only other formations with which it seems at all possible to correlate them are Hewett’s Gebo (Fisher’s basal Laramie) or Hewett’s Fort Union( ?) (Fisher’s upper Laramie), and to the writer the evidence seems in favor of the latter conclusion. If these beds are the equivalent of Hewett’s Fort Union, it still remains to determine whether | they represent the equivalent of the original Fort Union and whether they are very late Cretaceous or early Tertiary, problems with which this paper has nothing to do. These faults involving the Fort Union(?) pass at many points beneath the Andesite, which Hague, in the Absaroka Folio, has called the Early Basic Breccia and which he considers to be of early Neocene age. This would date the faulting as taking place after the sedimentation of the Fort Union(?) and before the Neocene, probably in very early Tertiary time, since following the faulting long erosion had trenched the region deeply and in places completely cut away the fault block, before the Basic Breccia was laid down. CORRELATION WITH OTHER FAULTS Richards and Mansfield? have presented a concise statement of the available information regarding major thrusts in the northern Rocky Mountains, and hazard a possibility that these may consti- tR. W. Richards and G. R. Mansfield, ‘‘The Bannock Overthrust,” Jo. Geol., XX (1912), 704. 2 Op. cit. THE HART MOUNTAIN OVERTHRUST 55 tute parts of one major thrust, carved into isolated portions by erosion. As they suggest, however, such correlation awaits further and more careful study, particularly as to the dating of the faults. To the information they have gathered may be added the more recent work of Haynes* on the “Lombard Overthrust” in Mon- tana. The present paper is presented as a further contribution to the subject, and, while conclusions as to correlation are still premature, the writer believes it to be quite improbable that these various faults will ultimately be found to be a part of one great over- thrust. It seems much more likely that they represent numerous ‘“‘Decken”’ or rock sheets, the one driven over the edge of the next after the manner described by Geikie? in discussions of Alp ne structure. Two such rock sheets, one above the other, are exposed along the South Fork of Shoshone River, in the area here described. t Jour. Geol., XXIV (1916), 269. 2 Geikie, Mountains, Their Origin, Growth, and Decay. THE ORIGIN OF VEINLETS IN THE SILURIAN AND DEVONIAN STRATA OF CENTRAL NEW YORK* STEPHEN TABER University of South Carolina, Columbia, South Carolina CONTENTS INTRODUCTION STRATIGRAPHIC FEATURES STRUCTURAL FEATURES Types OF VEINS SOURCE OF THE VEIN MINERALS DESCRIPTION OF THE FIBROUS VEINS ORIGIN OF THE FIBROUS VEINS DESCRIPTION OF THE NON-FIBROUS VEINS ORIGIN OF THE NON-FIBROUS VEINS NATURE OF ForcES THAT SEPARATED THE VEIN WALLS SUMMARY AND CONCLUSIONS INTRODUCTION Although the origin of metalliferous veins has long been of interest to the geologist and mining engineer, very few facts have been definitely established concerning the mechanics of vein forma- tion. Direct investigation of the subject is difficult because of the complexity of the processes involved and because only the final results are available for examination. The evidence that may have existed during the early stages of vein growth has commonly been obliterated by alterations due to vein-forming solutions or to secondary changes. Most metalliferous veins are found in regions of dynamic metamorphism and where igneous processes have been active. Consequently these veins, as a rule, furnish little or no evidence relative to the mechanics of their origin. The study of small barren veins in regions of unaltered sedimentary rocks has been largely neglected because they are of no commercial impor- tance, and yet such veins often furnish more positive evidence con- « Presented in abstract at the Albany meeting of the Geological Society of America, December, 1016. 56 VEINLETS IN THE SILURIAN AND DEVONIAN 57 cerning the mechanics of vein formation than is to be found in the larger and more complex ones. It was for this reason that the present investigation was undertaken. Many small veins are present in the Silurian and Devonian formations of central New York. ‘These rocks are well exposed in the numerous limestone and gypsum quarries of Cayuga and Onondaga counties. In the summer of 1916 the writer visited all of the quarries now being worked in these counties and nearly all of those that are idle, but most of the data used in this paper were obtained in the extensive quarries found in the vicinity of Union Springs. STRATIGRAPHIC FEATURES The rock formations outcropping in the region studied are listed below: Devonian Skaneateles shale Cardiff shale Marcellus shale Onondaga limestone Oriskany sandstone and conglomerate Silurian Manlius limestone ~Roundout limestone Cobleskill limestone Bertie waterlime Camillus shale (Syracuse salt) Vernon shale ) Salina formation Rock salt in the form of lens-shaped beds is present at many places immediately below the Camillus shale, but it has been re- moved in solution wherever the covering is less than about 1,000 feet thick, and therefore is never found near the outcrops of the strata." The Camillus shale contains intercalated beds of impure mag- nesian limestone and of gypsum. The limestone layers are more abundant in the upper part of the shale and probably represent * D. H. Newland and Henry Leighton, “‘Gypsum Deposits of New York,” N.Y. State Museum Bull. 143, 1910, p. 21. 58 STEPHEN TABER transitional stages toward the Bertie waterlime. The gypsum is highly argillaceous and in places grades into gypsiferous shales. Partings of shale, ranging in thickness from a fraction of a centi- meter up to several meters, are usually present, dividing the gypsum into several beds. These beds thin out and disappear, so that their number and thickness vary greatly in different districts. In many sections they are entirely absent. Gypsum may be found in small quantities all the way from the bottom to the top of the Camillus shale, but usually most of it is near the top. The Onondaga limestone is commercially the most important of the limestone beds, and therefore there are many quarries located all along its outcrop; but the Cobleskill and Manlius limestones are also being quarried at several places. ‘The lower layers of the Onondaga limestone, locally known as ‘“‘gray limestone,” have a well-developed crystalline texture similar to that of marble. The limestone forming the upper portion is bluish gray in color, dense, fine-grained, and contains numerous nodular concretions of chert or hornstone. Microscopic examination shows that the “blue lime- stone’’ consists essentially of irregular grains of calcite and small crystals of pyrite, while rhombic crystals of calcite may occasionally be distinguished. The chert varies in color from light bluish gray to almost black, and on freshly fractured surfaces is often difficult to distinguish by color or texture from the inclosing limestone (see Fig. 6). It is irregularly distributed, occurring often in small isolated nodules, though more commonly the nodules are arranged in well-defined rows or layers, and in places these layers of disconnected nodules pass by gradation into more or less continuous and uniform layers or bedded veins, which may be 3 cm. or more in width and extend for distances of many meters. The chert masses have evidently been formed through replacement of the limestone, for some of them contain fossils in which the details of structure are perfectly preserved. On weathered surfaces the chert, because of greater resistance, stands out in sharp relief. Microscopically the chert is cryptocrystalline, and the boundary between limestone and chert is not sharply defined. In passing from limestone to chert there is a gradual though rapid decrease in calcite with a correspond- VEINLETS IN THE SILURIAN AND DEVONIAN 59 ing increase in silica, but all of the chert examined contains numer- ous inclusions of calcite in the form of rhombohedral crystals (0.05 mm. and less in diameter), somewhat larger than the similar rhombs in the limestone. STRUCTURAL FEATURES The rock strata have been disturbed only slightly since their emergence from the sea. In general, the dip is toward the south at an average inclination of 7 to ro m. per kilometer, but in a few places, because of gentle folding, there is locally considerable variation from this average. Jointing is well developed throughout the area, and is probably due chiefly to the adjustment of strains resulting from folding and tilting. Appreciable openings are not found along these joints except near the surface, where, under favorable circumstances, they have been widened by the solvent action of descending surface water, and in such instances little or no deposition is to be observed on their walls. The fracturing of the rock strata seems to have resulted from compressive forces which would tend to prevent the formation of open fissures. ‘The joints cut the veins of the region, and are therefore, in part at least, of later origin. A thrust fault with displacement of a few centimeters is exposed in the Backus quarry, two miles north of Union Springs, and here the drag of the rock strata on both sides of the fault plane indicates that the displacement was accompanied by sufficient pressure to keep the fracture closed. A narrow vein of selenite follows this fault. Hopkins has described several thrust faults in the vicinity of Syracuse, the displacements ranging from a few centimeters to a little over a meter." Certain local disturbances of the rock strata, not noticeable in the overlying formations, may be observed in the Cobleskill lime- stone and the upper beds of the Salina. In places these strata have been pushed upward in such a way as to form low domelike eleva- tions on which the joints sometimes have a radial arrangement. A group of six or more domes may be found a kilometer southeast tT. C. Hopkins, ‘‘The Geology of the Syracuse Quadrangle,” N.Y. State Museum Bull. 177, 1914, p. 20. 60 STEPHEN TABER of Aurelius Station. They are strung out in a general north and south line near the bottom of a hill slope, and at the foot of the hill, close to the base of the domes, there are several large springs with deposits of calcareous tufa below them. A small quarry has been opened in one of the larger domes, which has a diameter of about 50 m. and height of 4 m. Hartnagel* thinks that these domes are due to an increase in volume of the underlying beds, because of the formation of gypsum from anhydrite; but the present writer has found no evidence supporting this view. The shape of the domes, their location, and their general associations are such as to suggest that they have been formed in the same way as the salt and gypsum domes of Louisiana and elsewhere, which have been described and explained by Harris.’ No open fissures, except where joints had been widened at the sur- face by weathering, and no veins were observed in any of the domes. Open spaces of appreciable size are infrequent except in the upper beds of the Salina. The Bertie waterlime contains numerous small cavities attributed by Vanuxem to the solution of salt, since ~ they sometimes exhibit the hopper-shaped outlines of halite crystals. The intercalated layers of magnesian limestone in the Camillus shale usually show the same porous structure and hopper-shaped casts. These cavities are frequently lined with a calcareous deposit. Small cavities, caused by the partial solution of fossils, are occa- sionally found in some of the limestones, and these openings are often lined with calcite, chalcedony, or crystals of quartz. Open fissures of mechanical origin were found in only one locality, in Camillus shale exposed by a cut on the Lehigh Valley Railroad about 100 miles west of Cayuga Junction. The cracks are Icm. or more in width, and are partly filled with a calcareous deposit having the appearance of finely banded travertine or onyx marble, the layers of which are tinted various shades of light yellow and reddish brown. The material is similar in every way to the deposits lining cavities in the Bertie waterlime and to layers in tC. A. Hartnagel, ‘‘Preliminary Observations on the Cobleskill (‘Coralline’) Limestone of New York,” V.Y. State Museum Bull. 69, 1903, p. 1135- 2G. D. Harris, ‘‘The Geological Occurrence of Rock Salt in Louisiana and East Texas,” Econ. Geol., IV (1909), 12-34. ing ei tie VEINLETS IN THE SILURIAN AND DEVONIAN 61 some of the calcareous tufa now forming in places on the surface. These fissures are possibly due to the solution and removal of under- lying salt beds or perhaps to other superficial disturbances, since the deposits were evidently formed in the belt of weathering. TYPES OF VEINS Structurally the veins are of two different types: one fibrous, the other more or less coarsely crystalline and non-fibrous. The former are composed either of gypsum or calcite, the crystal fibers extending transverse to the strike of the veins, which run in all directions, but are generally parallel to the bedding. These veins are lenticular and continue for short distances only. The non- fibrous veins usually consist of gypsum or calcite, but the calcite veins sometimes contain accessory quartz and pyrite. They are more persistent and more uniform in width than the fibrous veins, and most of them are vertical or steeply inclined. The evidence indicates that each type had a different mode of formation. SOURCE OF THE VEIN MINERALS In the veins under consideration there can be no question as to the source of the vein minerals, for it is evident that they have been derived from the neighboring rocks. The veinlets found in the gypsum-bearing strata of the Salina are composed of gypsum, while those occurring in the limestones, waterlimes, and calcareous shales consist essentially of calcite. It is not the purpose of the writer, however, to imply that the vein minerals found in other and larger veins have usually had a similar source. DESCRIPTION OF THE FIBROUS VEINS The fibrous (satin spar) veins are larger and more abundant in the gypsum-bearing strata, probably because of the greater solubility of gypsum as compared with calcite. As a rule they are less than 3 cm. in width and from 20 to 50 cm. in length, but in places they have a width of over 10 cm. and extend for distances of many meters. Most of the veins are highly lenticular in form; where a vein thins out it may be replaced by another a little to one side, so that the ends overlap. Veins frequently split into two or 62 STEPHEN TABER more branches, but the intersection of veins is extremely rare. When numerous they are commonly grouped to form linked-vein systems, as in Fig. 1. The vein fibers are usually normal to the inclosing walls, occasionally they are oblique, and very rarely they are curved or abruptly bent. In some veins most of the fibers apparently extend from wall to wall without a break, while in others there is a well-defined central parting frequently marked by the presence of inclusions of the wall rock. Small vugs are found in a << VAAN SU WOT os, x ON iro Mm IT Lore LUT Srey ae ap, zs Fic. 1.—Veins of fibrous gypsum exposed in walls of quarry near Union Springs New York. few veins, and these are lined with gypsum crystals of normal habit. The veins of fibrous calcite are similar to those of fibrous gypsum except that ordinarily they are smaller and not so numerous. In both gypsum and calcite veins the fibrous structure is as highly developed in the larger veins as in those that are smaller, the diam- eter of the fibers apparently being independent of the size of the veins. However, the diameter of the crystal fibers does vary markedly with any change in the texture of the wall rock. In the fine-grained limestones and shales the fibers commonly have a diameter of 0.05 mm. and less, while in the Onondaga ‘“‘gray lime- stone,’’ with its coarsely crystalline texture, the diameters are as VEINLETS IN THE SILURIAN AND DEVONIAN 63 great as 2 mm., and the fibrous structure is hardly noticeable in the narrower veins. Where veins of fibrous calcite in the Onondaga “blue limestone” pass through chert nodules, there is a sharp change in texture, the veins becoming coarsely crystalline and non- fibrous within the chert. This sudden change in texture is easily noticeable in veins that are less than a millimeter in width, when they are examined in thin sections under the microscope. Where the veins are non-fibrous, the individual crystals usually have their longer dimensional axes parallel rather than transverse to the strike of the veins; and, especially in the smaller veins, most of the crystals extend from wall to wall. In the larger veins these crystals have maximum diameters of over 5 cm., while in the fibrous portion of the same veins the fiber crystals are uniformly 0.1 mm. or less in diameter. The larger crystals of calcite frequently show warped cleavages, and under the microscope undulatory extinction is common in these crystals and also in those of fibrous form. The fibrous crystals are very irregular in cross-section, since the prisms are not bounded by plane surfaces as is often true of the crystals found in the non-fibrous portions of the veins. Vugs lined with calcite crystals of normal habit (simple rhom- bohedrons with some scalenohedrons) are occasionally present in the fibrous portion of the veins where the walls are of limestone, but they are more abundant where the veins are coarsely crystalline and have chert walls. The walls of the veins are sharply defined, and inclusions of the wall rock, limestone as well as chert, are com- mon. When one wall of a vein contains angles or other irregular- ities, there are corresponding irregularities in the opposite wall, such that the two surfaces would fit closely together if placed in contact. ORIGIN OF THE FIBROUS VEINS In previous papers’ the writer has cited evidence tending to prove that cross-fiber veins of the asbestiform minerals could not have been formed through any process of replacement or of recrys- tallization 7m situ and that they were not deposited in open fissures. «Stephen Taber, ‘“‘The Origin of Veins of the Asbestiform Minerals,” Proc. Nat. Acad. Sci., II (1916), 659-64; and ‘‘The Genesis of Asbestos and Asbestiform Minerals,” Bull. Am. Inst. Min. Eng. No. 119, 1916, pp. 1973-08. 64 STEPHEN TABER Most of the objections raised against these theories of vein forma- tion are equally applicable in the case of the veins of fibrous calcite and gypsum; and, in the descriptions given above, much confirma- tory evidence may be found. All of the structural features char- _ acteristic of these veins have been duplicated in fibrous veins grown in the laboratory where their origin and growth could be observed in detail.t ‘In view of all the facts obtained from field investiga- tions and laboratory experiments, the conclusion is inevitable that the veins of fibrous calcite and gypsum have been formed through a process of lateral secretion, the growing veins making room for themselves by pushing apart the inclosing walls, and that the fibrous structure is due to the circumstance that the material for crystal growth was accessible in only one direction. Calcite and gypsum are not normally fibrous, and wherever they have developed this structure it is due to the physical conditions which have prevented crystal growth, except in one direction. Merrill has described fibrous incrustations of gypsum forming on the walls of caves, and notes that the growing crystals not infrequently force off pieces of the limestone of considerable size.” Laboratory experiments and field investigations indicate that the essential conditions for the growth of fibrous minerals, such as calcite and gypsum, are: (1) the growing crystals must be in contact at their base with a supersaturated solution; and (2) the so- lution must be supplied through closely spaced capillary or subcapil- lary openings in the surface of the wall rock. In the fine-grained limestones and shales the constituent particles are relatively small, and therefore the open spaces which are chiefly subcapillary in size are closely spaced; but in the crystalline ‘‘gray limestone” with its coarser texture these openings while no larger are necessarily more widely spaced. This explains the coarse texture of the fibrous veins occurring in the “gray limestone.” The coarsely crystalline non-fibrous structure of veins where they pass through chert masses is due to the relative impermeability of the chert which has here «Taber, ‘“The Origin of Veins of the Asbestiform Minerals,” Proc. Nat. Acad. Sci., II (1916), 659-64; and ““The Genesis of Asbestos and Asbestiform Minerals,” Bull. Am. Inst. Min. Eng. No. 119, 1916, pp. 1973-08. 2G. P. Merrill, ““On the Formation of Stalactites and Gypsum Incrustations in Caves,” Proc. U.S. Nat. Mus., XVII (1894), 81. VEINLETS IN THE SILURIAN AND DEVONIAN 65 prevented the addition of new material directly through the walls, thus forcing it to reach the growing crystals by diffusing between the walls. Vugs result from a deficiency of material necessary for growth because of insufficient concentration or because of relative inaccessibility. The latter probably explains the greater abun- dance of vugs between chert walls. DESCRIPTION OF THE NON-FIBROUS VEINS The non-fibrous veins range up to 5 cm. or more in width and in some instances are exposed for distances of 15 or 20 m. along the strike. Where they pinch out and disappear, they are sometimes replaced by others a few centimeters to one side or farther along the line of strike. Such vein systems may be traced for over 50 m. The veins show no appreciable change in appearance where they pass from one rock to another of different texture or composition. A vein exposed in the limestone quarry near Farleys can be traced upward through the argillaceous Manlius limestone, 20cm. of Oriskany conglomerate, and into the Onondaga limestone, yet at no place is any variation in its appearance perceptible. The vein walls are sharply defined, and fracture usually takes place more readily along the contact between vein and wall rock than in other directions. The opposite walls of a vein are parallel even when they are very irregular, and they would therefore fit intimately together if placed in contact (see Fig. 2). Angular fragments of the wall rock are occasionally present in the veins; and in many instances, by making parallel sections, it is possible to prove that they are in contact neither with other fragments nor with the walls. Most of the fragments show no evidence of rota- tion although they have been displaced through distances of 2 cm. or more (see Fig. 3). In places a fragment adhering to both walls of a vein appears to have been separated into several fragments by continued vein growth, as in Fig. 4. Some veins have a banded structure with coarsely crystalline non-fibrous calcite in the center and a band of fibrous calcite along each wall. This is probably due to two stages of vein growth, as is indicated by the veins sketched in Fig. 2. Other veins are roughly banded, with pyrite along the walls and calcite in the center (see 66 STEPHEN TABER Fig. 5), but in such cases the pyrite was deposited subsequent to the deposition of most of the calcite, and well-formed cubes and pyritohedrons may be found replacing impartially vein calcite and wall rock. Small seams of pyrite in places cut directly across the veins. Fey C ED, Scale 0 Ye. Inch ee Fic. 2.—Calcite veins in limestone. Coarsely crystalline calcite (C) and fibrous calcite (F). ORIGIN OF THE NON-FIBROUS VEINS The facts cited above preclude the theory that these veins are due to recrystallization of country rock in situ or that they could have been formed through replacement; and the presence of de- tached inclusions of wall rock argues against the hypothesis that the veins were deposited in open fissures. If the veins were formed as a result of fissure filling, deposition of vein matter must have begun on the walls and continued inward until the opposite sides VEINLETS IN THE SILURIAN AND DEVONIAN 67 met, thus forming a suture line near the center, but there is no evidence that such a suture was ever present in any of the veins under consideration. Large calcité crystals commonly extend without interruption from wall to wall, and in one vein a well- formed crystal of quartz, with a pyramid at each end of the prism, Fic. 3.—Calcite vein in limestone showing angular inclusion of the wall rock. Two-thirds natural size. was found extending across almost the entire width of the vein (see Fig. 5). The best evidence bearing on the origin of the veins is, perhaps, furnished by certain chert nodules containing veinlets of calcite, ranging up to 2 or 3 mm. in width, which do not extend into the inclosing limestone (see Fig. 6). The force separating the chert walls was applied so gradually that any stresses set up in the 68 STEPHEN TABER limestone were adjusted by recrystallization, in the same way that slabs of marble or limestone may be slowly deformed under forces acting through a long period of time. This process probably also explains the curving walls of the lenticular veins. The facts here listed are difficult or impossible of explanation under any of the hitherto generally accepted theories of vein forma- tion. They are, however, easily explained on the hypothesis that Fic. 4.—Calcite veins in limestone showing inclusions of the wall rock. Two- thirds natural size. the vein-forming solutions entered along fractures, bedding planes, or other planes of weakness, where the openings were chiefly capil- lary or subcapillary in size; and that the separation of the vein minerals from solution was accompanied by the development of a force sufficient in magnitude to push apart the walls, and thus gradually make room for the growing veins. Circulation of solu- tion through such narrow openings must necessarily be extremely slow, and under these conditions diffusion through the solution becomes an important factor in supplying additional material to the growing crystals. VEINLETS IN THE SILURIAN AND DEVONIAN 69 Where veins pass through chert masses, most of the calcite crystals extend from wall to wall, and are oriented with their longer dimensional axes parallel rather than transverse to the vein walls— a fact that is likewise true of many non-fibrous veins in limestone and shale. Since this may be observed in the largest as well as the smallest veins, it means that the average number of vein crystals Tian a SOS EROS EE Sovessososbe OSSeS Gosboonooue 49.00 30.87 47.82 46.47 45-34 45.61 ADI Ore a aes 1.30 1.50 0.67 1.33 1.30 1.96 INA) rig ane tenet eras 15.70 13.58 13.56 15.97 16.59 14.35 ete reese ransuocn costae 5333 6.71 4.73 5-97 5.83 6.17 He On ate ae ania: O87 6.43 4.54 4.27 4.76 4.03 VETO ee eee Sik 0.31 0.21 tr. 0.01 0.01 0.19 Nig Oneness as eee 6.17 10.46 7.49 5.87 5-43 6.05 CaO Re Re ech een 8.05 12.306 8.01 10.54 Ir.64 9.49 IN Oa oui aioe Qa ir 3.85 4.37 1.69 2.93 plea 110) Scpelea reece eae TG 1.87 Bo2a 4.83 4.55 3.69 15 LAO) aay arava easier ae 1.62 2), Oy BBG 2 3D 1.12 2.60 1B Oe ie a eee eae 0.45 0.94 1.10 0.73 0.50 0.74 100.00 | 100.00 99.79 | 100.00 | 100.00 | 100.00 * Includes 0.29 per cent CO:. Perhaps fuller statement of the hypothesis might annul some of the writer’s doubts which have been raised by these and other fundamental facts of magmatic differentiation, but at present he fails to see that the mechanism works as it should if the pure- fractionation theory were correct. Admission of some assimilation by magmas.—Bowen states (pp. 84, 89): As a matter of fact, plain evidence is found in the field that magmas do assimilate, especially when they occur in the large bodies commonly termed batholthss eis... It may well be, also, that some melilite rocks are formed by the crystallization of a syntectic magma formed by the solution of lime- Stones... If the absorption of any considerable amount of limestone by a magma can be admitted, it may be expected to have a very unusual effect upon themagma. 2...) . The taking of silica from feldspar molecules by the lime and the consequent production of feldspathoid molecules might well be supposed a reasonable possibility. Some alkaline rocks may, perhaps, be so generated. Yet Bowen believes ‘‘that normally the alkaline rocks enter into an eruptive sequence as the products of differentiation solely.” Only a very small addition of foreign lime (1 to 10 per cent) to a large mass of basaltic magma would be necessary to produce 32 REGINALD A. DALY the total amount of desilication observed in any nephelitic or leucitic body yet discovered. Considering the fluxing power of limestone or dolomite on silicate melts, such moderate, local assimi- lation is surely not improbable. More:siliceous sediments or even gneissic or granitic rocks may be simultaneously dissolved; but, on account of its character as a flux, a carbonate rock is likely to be absorbed in greater volume. Hence it does not follow that the solution of “‘an occasional bed of limestone in a terrane consisting principally of siliceous gneisses . . . . would entail simultaneous absorption of a much greater quantity of relatively siliceous material”’ (Bowen, p. 63). The relative amounts absorbed must really depend on a number of factors, including the contact relations of each layer of the country rocks to the invading magma. Meaning of melilitic. rocks.—Bowen’s admission regarding the melilite rocks is of particular significance when one remembers. the exceeding intimacy of alkali-rich rocks with melilite basalt, nephelite-melilite basalt, and alndite.* Résumé.—The prominent difficulties with Bowen’s theory may now be summarized. The sinking or rising of crystals in magmas is a true cause of diversity in igneous rocks, but it is not the only important cause; perhaps it is much less important than the separation of liquid phases. Apart from the possible development of ‘liquid immis- cibility”’ in an initially homogeneous magma, separation in the liquid phase is to be expected: (1) if the magma at the time of emplacement in a large chamber were heterogeneous; (2) if for any reason gases are concentrated locally within the magmatic body; and (3) if the assimilation of country rocks, or of their volatile constituents alone, takes place. The field evidences for im- portant assimilation are not consonant with the pure-fractionation theory. Bowen’s theory seems to imply a greater facility of differ- entiation for basaltic and other magmas than they actually possess; most sills and laccoliths are not visibly differentiated, even though they show features implying fluency after injection. The theory gives no good explanation of the comparatively abrupt transition « See the writer’s Igneous Rocks and Their Origin, p. 436, and Appendix D. GENESIS OF THE ALKALINE ROCKS A between the syenite and shonkinite of the Square Butte laccolith, nor of the apparent homogeneity of that shonkinite. It fails to take account of the lack of ultra-femic phases in the Pigeon Point and other sills which display notable salic differentiates. Its con- sequence, that all monomineralic rocks except certain sulphides are crystal “‘rafts,”’ may fit the case of anorthosite, but not other cases. Other special objections are found in the chemical nature of quartz diabase, basanite, and the ‘‘alkaline basalts.” The absence of foyaitic phases in, or apophysal from, most large bodies of granite, granodiorite, and quartz diorite does not agree with the theory as developed. Similarly quartzose lavas, expected on the theory, are not found in many volcanic piles containing trachyte or phonolite with basalt. The lack of quartzose lavas in the enormous and therefore long-lived volcanoes distributed over the main ocean floor is a fact not easily explained by the theory. Bowen has given the clearest, most detailed argument for the significance of fractional crystallization which has yet been pub- lished. His general theory takes cognizance of: the proved genetic association of alkaline rocks with subalkaline magmas, especially the basaltic; the small size of alkaline bodies; their richness in gases and rare elements; the consequent effects regarding grain and ‘variability of composition; the roof positions of alkaline differ- entiates in many sills, laccoliths, and batholiths; and the con- centration of alkalies. Nevertheless, all of these facts are explicable also on the syntectic-differentiation hypothesis, which does not encounter certain difficulties facing the theory of pure fractional crystallization. GENERAL CONCLUSION Thus the study of the geological publications issued since the completion of the manuscript of the writer’s Igneous Rocks and Their Origin has led him to renewed faith in the general explanation there advanced for most of the alkaline rocks. Several expert field observers have sympathetically entertained the hypothesis of control by the syntexis of basic sediments charged with volatile matter. Evidence for the derivation of alkaline rocks from sub- alkaline magmas has been still further accumulated. Some authors have expressed a degree of confidence that basalt is the only primary 134 REGINALD A. DALY magma eruptible since the end of pre-Cambrian time. Foye has made an extraordinarily important contribution in showing the vast quantity of alkaline solutions which may emanate from granite invading a limestone terrane. Like Allan-he has given new, good evidence of the influence of gravity in the separation of magmatic phases. Foye corroborates the decision of Adams and Barlow as to the syntexis of limestone and granite in a complex now famous for its nephelite syenites. Quensel’s discovery of an analogy to Alno in the Almunge district and of the peculiar abundance of vesuvianite in the Almunge canadite is likewise to be particularly - recorded. The arguments presented by Cross, Marshall, Richards, and others against the sediment-syntectic explanation of the alkaline rocks are seen to be inconclusive. Shand finds the syntexis of limestone a partial explanation of undersaturation in igneous rocks. Smyth’s view concerning the alkaline series is not acceptable, on the ground that he fails to show cause for the local and exceptional assembling of alkaline elements from subalkaline magmas, assumed by him to be purely juvenile. Experiments by Ross and other students of the commercial-potash problem show the power of lime to volatilize the alkalies from feldspathic or clay mixtures, even at comparatively low tem- perature. Certain aspects of Bowen’s comprehensive theory have been studied. Serious doubt adheres to some of his fundamental postu- lates, summoned to explain the descent of alkali-rich rocks from subalkaline magma. On the other hand, many features of his excellent paper must receive hearty commendation from all thinking petrologists; in masterly fashion he has indicated many new, im- portant lines of thought and research. CONDITIONS OF DEPOSITION ON THE CONTINENTAL SHELF AND SLOPE Cc. A. COTTON Victoria University College, Wellington, New Zealand CONTENTS INTRODUCTION OPINIONS AS TO THE MODE OF FORMATION OF THE CONTINENTAL SHELF THE PRESENT-Day SHELF LARGELY CONSTRUCTIONAL The Hypothesis That Marine Erosion Is Alone Responsible for the For- mation of the Continental Shelf The Hypothesis That Subaérial Planation Followed by Submergence Explains the Continental Shelf The Hypothesis That the Continental Shelf Is Formed as a Result of Deposition THE STRUCTURE OF THE BUILT SHELF With the Shore Line Fixed in Position With the Shore Line Advancing With the Shore Line Retreating With the Shore Line Alternately Retreating and Advancing LITHOLOGICAL CHARACTER OF THE BEDS STRUCTURE OF A SHELF BUILT DURING POSITIVE MOVEMENT STRUCTURE OF A SHELF BUILT DURING NEGATIVE MOVEMENT INTRODUCTION Undoubtedly deposition of sediment has gone on around the margins of the continents in all ages, forming thick accumulations which have encroached upon the ocean basins. To what extent such marginal accumulations have been later uplifted and thus added to the continents may be a matter for difference of opinion; but, though in some parts of the world the sedimentary rocks represent mainly deposits in interior epicontinental seas, there are other parts—the New Zealand area, for example—in which the nature of the rocks indicates that the sediments of which they are 135 136 CA. COTTON formed accumulated marginally to, rather than upon, one of the’ continental protuberances. The subject of marginal sedimentation is closely connected with considerations as to the form and origin of the continental shelf. While this is recognized by geologists and implied by many in their writings it is not usually explicitly stated. OPINIONS AS TO THE MODE OF FORMATION OF THE CONTINENTAL SHELF There isa very striking contrast normally present between the gentle slope of the continental shelf and the relatively steep descent, known as the continental slope, from the edge of the shelf to the depths of the ocean. In textbooks this is pointed out, but an explanation of it is generally avoided. Lake,’ however, has included a discussion of the problem. In Lake’s textbook and in a recent paper by Gardiner’ several hypotheses bearing on the subject are formulated. From both the reader receives the impression that, if a hypothesis of glacial deposition be put aside as of local application, there remain in the field to account for the continental shelf in low latitudes two rival hypotheses or groups of hypotheses behind each of which there is an equal weight of authority—namely, a hypothesis of erosion with or without subsidence of the eroded surface, and a hypothesis of accumulation according to which the shelf has grown owing to deposition of sediment. It is clear, however, that both these writers favor the hypothesis of accumulation and regard the shelf as for the most part a built feature. The acceptance of this view is the basis of many geological writings, such, for example, as an article by Chamberlin on “The Ulterior Basis of Time Divisions . . . . ,’’3 in which it is taken as axiomatic that the shelf is formed partly by cutting but mainly by deposition. The same point of view is implied by the use of such terms as ‘‘continental delta”? by Gulliver,’ signifying conti- =P. Lake, Physical Geography (Cambridge, 1915). 2 J. Stanley Gardiner, “Submarine Slopes,” Geog. Jour., XLV (1915), 202-10. 3T. C. Chamberlin, Jour. Geol., VI (1898), 449-62; also Editorial, pp. 424-26. 4. P. Gulliver, ‘‘Shoreline Topography,” Proc. Amer. Acad. Arts and Sct., XXXIV (1890), 176. DEPOSITION ON CONTINENTAL SHELF AND SLOPE § 137 nental shelf, “built terrace” by Gilbert and others, and ‘‘topset,”’ “foreset,” and “bottom-set beds”? by Chamberlin,’ who thus establishes the analogy of the continental shelf with deltas. Some clear statements have been made as to the origin of that portion of the continental shelf fringing eastern North America. Willis writes as follows: The plateau is composed of sands which are indeed fine near the eastern edge, yet are distinctly granular and incoherent. But soundings on the steep slope beyond the 1oo-fathom line have brought up very fine silt from the bank of which that slope is the surface, and this silt passes at its foot into globigerina ooze. The zone of transition from clean sand to silt is as sharp as the edge of the slope and is coincident with it. It is evident that the suspended mud which escapes beyond the estuaries and sounds of the littoral is swept out until the undertow expands over the edge of the escarpment, and is diffused in deep water; there the silt forms a great bank 10,000 feet high, with a slope of 3 to 8 degrees, which has grown seaward during geological ages, and continues to expand as erosion continues on the land. The structure of this deposit can only be inferred, but it is worthy of consideration. The surface of accumulation, to which bedding planes are probably parallel, is inclined at a considerable angle, and traverse the bank from top to bottom obliquely to the vertical thickness. The direction of the growth is outward, not upward. The conditions of deposition are similar to those of a delta advancing into fresh water, and the structure of the deposits is probably similar to that shown by Gilbert for a fresh water delta.? The generalization by the same writer that ‘‘the ocean basins are now somewhat overfull . . . . not large enough to hold all the waters, which therefore extend over the margins of the continents,’ does not necessarily contradict the foregoing; but it omits to state that, though the growth may now be outward only, during the postulated overflow of the oceanic waters the shelf must have maintained itself by upward growth. Barrell, also, writes as follows: Ocean waves are known to have a perceptible effect to a depth of about too fathoms, planing away the shore and the higher parts of the bottom, carrying the products of fluviatile and marine erosion outward to deep water. tT. C. Chamberlin, ‘‘ Diastrophism and the Formative Processes. VI. Foreset Beds and Slope Deposits,” Jour. Geol., XXII (1914), 268-74. 2 B. Willis, ‘‘ Conditions of Sedimentary Deposition,” Jour. Geol., I (1893), 497-08. 3 B. Willis, ‘Principles of Paleogeography,”’ Science, N.S., XX XI (1910), 241-60 (see p. 244). 138 nea) C. A. COTTON The waves move material along the bottom and prevent the settling of the finest silt until the limit of wave action is reached. Beyond that limit the bulk of the material is rapidly deposited from suspension. In protected situations this depth becomes less and in many places is not over fifty fathoms. There is thus built outwards around the continents a subaqueous terrace, its top gently sloping to a depth of too fathoms or less, its front much steeper in comparison, and giving sharpness to the continental margin." Unqualified support of any hypothesis which assigns the for- mation of the shelf entirely to erosion is rarely met with. In an explanation of the shelf given by Mill,? however, marine erosion with stationary sea-level is given first place, and even in an Ameri- can textbook submergence of a plain “‘worn to low relief” is the i only explicit explanation given. That this statement does not really express this author’s view is shown, however, by the following passage, which appears on another page in an explanation of coastal plains (p. 508): ‘‘Off the eastern coast of United States there is a level sea-bottom plain, known as the continental shelf. ... . it there should be an uplift of 600 feet, this very level plain would be added to the continent. .... It would be underlain by un- consolidated sediments.” The following clear statement on the subject by Gilbert and Brigham is also worthy of attention as an emphatic refusal of sup- port to the hypothesis of submergence of a plain of erosion: It will be remembered that the resemblance of Chesapeake Bay to a branching river was explained by saying that the Coastal Plain had sunk down so as to let the sea flow into the Susquehanna Valley. Because we now point out that the plain is an old sea-bed which has risen, it must not be thought that one fact contradicts the other. Both changes have taken place, but at different times. After the plain had been formed under the sea it was lifted so high that rivers dug deep valleys across it; then it was lowered part way, to the present height.4 A somewhat similar statement from Chamberlin and Salisbury may be quoted: ‘‘Almost nowhere does the real edge of the con- «J. Barrell, “The Upper Devonian Delta of the Appalachian Geosyncline,” Am. Jour. Sci., XXXVI-XXXVII (1913-14), 249. 2H. R. Mill, The Realm of Nature (2d ed.; London, 1913), p. 219. 3R. S. Tarr, College Physiography (New York, 1914), p. 641. 4 Gilbert and Brigham, An Introduction to Physical Geography (New York, 1902), p. 153. DEPOSITION ON CONTINENTAL SHELF AND SLOPE 139 tinent appear above the ocean... . . A continuous shelf almost universally borders the continents..... The continent was recently so deformed that these shelves were out of water,” and, on a later page: “‘The continental shelf . . . . may be supposed to have been built out upon the border of the sea-basin by progressive sedimentation.’ The meaning of all these writers is clearly that a continental shelf—a constructional feature, built of marine sediments—has been uplifted and dissected and then resubmerged. In textbooks the statement is frequently made that the true boundary of a continent is situated at the outer edge of the conti- nental shelf—the “continental edge’ of Murray—rather than at the shore line.? At first sight this reads like a negation of the view that the shelf is mainly a built feature of recent growth. Such state- ments seem, however, generally intended merely to convey the information that, when mean slopes of the earth’s surface are plotted as a hypsographic curve, there is a sharp change of average slope at the depth of the edge of the continental shelf but none at sea-level. THE PRESENT-DAY SHELF LARGELY CONSTRUCTIONAL Since ‘‘rival”’ hypotheses of erosion, marine and subaérial (fol- lowed by subsidence), and deposition have been put forward by Lake, Gardiner, and others, it is necessary to examine critically the consequences of each. The hypothesis that marine erosion is alone responsible for the formation of the continental shelf —Undoubtedly wave action is capable of eroding, not only at the shore line, but at considerable depths, wherever the ratio of the waste supply to the transporting power of the sea is sufficiently low to permit parts of the rock bottom to be swept clean, and where there is at the same time sufficient movement of the water to move particles of appreciable t Chamberlin and Salisbury, Geology, III (New York, 1906), 521, 526. 2 The shelf is described, for example, as ‘“‘submerged parts of the continental pro- tuberance”’ (R. D. Salisbury, Physiography, p. 22), and it is stated that “‘the area between the actual shore and the 1oo-fathom is regarded as belonging to the con- tinents, though at present overflowed by the sea’’ (Hatch and Rastall, The Petrology of the Sedimentary Rocks [London, 1913], p. 11). 140 Car COMHON size across the exposed rock. ‘This is an accepted principle. Wave motion extends, it is believed, to a depth of about 100 fathoms in the open ocean; it has been shown to be sufficiently strong to move gravel at a depth of 36 fathoms.’ There is thus no reason to believe that, zf the bottom were kept swept clear of sediment, erosion could not take place—though, obviously, the slowness of erosion would increase with depth—as far seaward as the edge of the shelf in 100 fathoms, at which depth there would be an abrupt change of slope from that of the shelf to that of the initial sea floor. This is the gist of the hypothesis as put forward by Mill and by Lake. The postulate of submergence following the cutting of the platform introduced by Gardiner is unnecessary. The question arises whether clean sweeping of the cut platform such as is postulated in the preceding paragraph is possible; and it would appear that such a state of affairs must be extremely © uncommon if it ever occurs. In the ordinary case deposition of the waste produced by the cutting of the platform, together with the supply from the neighboring land, must take place seaward of the area in which erosion is taking place. It is indeed obvious, as has often been pointed out, that the observed continental slope must be the surface of the last-formed layer of this sediment and repre- sents more or less accurately the slope at which it came to rest. It seems highly probable that earlier-formed layers came to rest at the same inclination, and that some portion of the shelf is every- where composed of such inclined layers. The combined processes of erosion and deposition have been shown by Davis? and by Fenneman: to produce a “graded profile”’ or “profile of equilibrium” from the shore to the edge of the shelf, which is a fairly even slope though somewhat concave near shore and convex seaward. Fig. 1 illustrates the growth of a shelf by this combination of erosion and deposition. It will be seen that, 1A. R. Hunt, “Formation of Ripple Marks,” Proc. Roy. Soc., XXXIV (1882), I-13 (see p. 10). . 2;W. M. Davis, ‘“‘The Outline of Cape Cod,” Geographical Essays (Boston, 1909), Pp. 690-724 (see pp. 700-703). 3 N. M. Fenneman, ‘“‘The Profile of Equilibrium of the Subaqueous Shore Terrace,” Jour. Geol., X (1902), I-32. DEPOSITION ON CONTINENTAL SHELF AND SLOPE 141 while, as the shore line recedes, erosion of the sea bottom must go on near shore, as the shelf front advances the graded profile is maintained farther seaward by deposition. The stream of waste in transit seaward will generally, on account of its slow movement, form a thick layer on the outer part of the shelf precluding the possibility of erosion of the deeper layers of strong currents along the shelf, so that the built portion of the shelf may be narrow compared with the cut portion. In sucha case erosion must extend into relatively deep water, parts of the eroded sediment or of the rock floor. It is, however, conceivable that in exceptional cases much or all of the waste may be swept away by Fic. 1.—Diagram to illustrate the formation of a continental shelf by marine erosion and deposition. floor of bedrock being occasionally swept clean. In this way are perhaps to be explained some recorded occurrences of hard bottom far out on the continental shelf. The modification of this hypothesis which requires subsidence following marine planation has nothing to recommend it. Fig. 2a is a copy of the figure used by Gardiner in illustration of the hypothesis, modified by the addition of a thick layer of sediment (coarsely stippled) on the top of the cut platform, which must be added to restore a normal profile. The profile of a later stage of a shelf developed by cutting and building on this initial form is also added (lightly stippled). The hypothesis that subaérial planation followed by submergence explains the continental shelf —In considering this possible mode of shelf formation care must be taken to exclude cases of temporary emergence of a shelf already formed. References to portions of the shelf thus temporarily uplifted, partially dissected, and again submerged have been already quoted; and oscillations of the I42 CRA CORON: continental borders, if not of all land surfaces, are known to be of common occurrence. Rapid submergence of a peneplain would undoubtedly result in the formation of a sea bottom of small relief with a general seaward slope; but it is impossible to imagine a peneplain the level surface of which would, in the period preceding submergence, end abruptly at the shore line with a sharp transition to a steep slope into the adjacent ocean basin. Even if such a state of affairs were possible, it would be necessary, in order to account for the formation of the Fic. 2.—Diagrams copied from those used by Gardiner to illustrate theories of shelf formation, with additions showing the development by deposition of a profile of equilibrium. continental shelf in this manner, to postulate a submergence of all the continental margins to the same extent, about 600 feet, and in most cases to disregard entirely the evidence afforded by the geo- morphology of coastal lands as to their erosional and deformational history. A submerged peneplain could not be expected to provide, ready made, a graded subaqueous shore profile. Upon an initial surface so formed wave action would immediately come into opera- tion, cutting in some places and depositing in others; and so, even if the initial form of any part of the continental shelf may have been a submerged peneplain, the sequential form which the shelf exhibits today must be ascribed to the work of waves. The development of a peneplain, moreover, necessitates a very long period of erosion, during which the land mass on which it 1s DEPOSITION ON CONTINENTAL SHELF AND SLOPE 143 cut remains essentially stationary in level, and during which an enormous quantity of waste is carried to the sea. Such waste will be deposited seaward as a continental shelf of unusually great breadth, and this, when submergence of the peneplain takes place, will subside also. Since, presumably, the edge of the shelf before subsidence will be situated at or just within the 1oo-fathom line, after subsidence has taken place, when the width of the plain will have been added to that of the former shelf, the composite shelf so formed will extend into water much deeper than roo fathoms. Upon such a shelf a thick layer of sediment would require to be deposited to reduce the depth at its edge to 100 fathoms; and so it appears evident that submergence of a peneplain without the subsequent co-operation of wave action on an extensive scale is not competent to produce a continental shelf of typical form. The profiles of the shelf in the diagrams given by Gardiner to illustrate the ‘‘formation of the continental shelf by the submer- gence of an abrasion platform” and “of a land plain” are quite unlike those found in nature. This is perhaps intentional, as that author does not appear to favor either hypothesis. Such initial profiles would be very readily modified by wave and current action, according to the principles deduced by Fenneman. Two stages of such modification are shown in Fig. 2b as an addition to Gardiner’s diagram of ‘‘submergence of a land plain.” The deposit added in an early stage of the modification is coarsely stippled, and that in a later stage lightly stippled. The hypothesis that the continental shelf is formed as a result of deposition.—The conditions of delta-formation by streams bearing a load of coarse waste into lakes are now well understood chiefly owing to the work of Gilbert, and an extension of the principles established by Gilbert to cover the case of rivers supplying large quantities of finer waste to the ocean has more recently been made by Barrell.?, It is but a step from the consideration of the 1G. K. Gilbert, ““The Topographic Features of Lake Shores,” U.S. Geol. Surv., 5th Ann. Rept., 1885, pp. 69-123. 2J. Barrell, ‘Criteria for the Recognition of Ancient Delta Deposits,’’ Bull. Geol. Soc. Am., XXIII (1912), 377-446. 144 C. A. COTTON submarine portions of deltas, which have been shown by Barrell’ to cover often a larger area than the subaérial portions and to form large seaward protuberances of the continental shelf, to the con- sideration of the shelf as a whole. Off all coasts deposition of a greater or less amount of sediment is always in progress, not only off the mouths of rivers, but along the whole length of the coast lines, and there can be a difference only in degree between such deposition and that by which the seaward, submarine portions of deltas are built forward. In the shallow water near shore conditions may be entirely different from those on the shallow-water portions of deltas; for here, as has already been shown, erosion may be going on, either deposition or erosion being in progress according to the state of balance between waste supply and transportation. Erosion near shore, however, will, generally speaking, merely have the effect of pushing the zone _ of deposition farther seaward. From observation of the form and structure of the small deltas laid bare and dissected owing to the lowering of the level of the water in lakes it is known that waste, after being discharged from a river, is transported outward from the shore owing to agitation and forward movement of the water, and traverses the upper, gently sloping surface of the mass of sediment already deposited (‘‘subaqueous plain,’’ Barrell) until, reaching the edge, it slips over into deeper and stiller water and comes to rest at the angle of repose on the more steeply sloping front of the mass (“‘foreset slope,” Barrell). Similarly, in the case of larger deltas of finer waste being built into the ocean, the waste is transported seaward across a subaqueous plain which is essentially a part of the con- tinental shelf, extending outward to the depth at which the bottom is no longer sensibly stirred by wave action (‘‘wave base,” Gulli- ver’). This depth in the open ocean is clearly indicated by the t Op. cit., Fig. 1, p. 388; ‘“‘The Strength of the Earth’s Crust,” Jour. Geol., XXII (1914), 39-42. 2F. P. Gulliver, “Shoreline Topography,” Proc. Amer. Acad. Arts and Sci., XXXIV (1809), 176-77. As defined by Gulliver wave-base, the limiting plane toward which marine erosion will tend to lower the wave-cut platform, is ‘‘the depth to which the maximum wave action is possible.”” The term has been redefined by Fenneman as the depth ‘“‘at which wave action ceases to stir the sediments” (N. M. Fenneman, ‘‘Lakes of Southeastern Wisconsin,” Wis. Geol. and Nat. Hist. Surv., Bull. No. 8, 1902, p. 25). DEPOSITION ON CONTINENTAL SHELF AND SLOPE 145 -“mud-line” of Murray and Renard, which, according to Murray, is situated at a depth of about 100 fathoms.’ On the slope below this depth the finest particles of sediment “‘come permanently to rest on the bottom.’”’ Beyond the mud-line is the steeper foreset slope on which even the finest waste can come permanently to rest; but the foreset slope is much less steep and the transition to it from the subaqueous plain much more gradual than in the case of the. small deltas of coarse material in lakes. For one thing water- logged mud will not remain at rest on any but very gentle slopes, and, as Barrell points out, “‘where, as in the case of large rivers, the detritus is mostly fine in texture, the foreset beds are built largely by material settling from suspension.’”’ The foreset beds, in which sediment swept outward along the bottom is mixed with what has settled from suspension, grade into the bottom-set beds or pelagic deposits built entirely of material settling from suspension. In the case of portions of the continental shelf remote from the mouths of large rivers the mode of accumulation must be essentially the same as in deltas; but there will generally be present in the waste supplied from the land a much smaller proportion of the fine mud particles resulting from subaérial weathering. No doubt much of the material broken by wave attack on the shore line and supplied by smaller rivers becomes very finely comminuted in its long passage across the shelf; but the absence of a large body of mud sufficiently fine to be carried far seaward in suspension is reflected in a sharper transition from the continental shelf to the continental slope, and in a steeper inclination of the latter than is found in delta fronts. Thus the structure of the continental shelf may be regarded as presenting generally a closer resemblance to that of the sub- aqueous portions of small lake deltas than is shown by the corre- sponding portions of the deltas of great rivers. The materials of which the shelf is being built forward are known from samples taken by oceanographers from the last layer added. ‘These deposits on the continental slope fall into the “terrigenous’’ division of ‘‘deep-sea deposits, beyond 100 fathoms”’ in Murray and Renard’s classification of marine deposits,? and the t Sir John Murray, The Ocean (Home University Library), (London), p. 202. 2*‘Tyeep-Sea Deposits,” Report of the Scientific Results of the Exploring Voyage of H.M.S. “Challenger,” 1873-76.(London, 1891); Murray, The Ocean, p. 201. 146 CA CORTON most common types bordering continental coasts are “‘blue mud” and ‘‘green mud.”? Examination of the samples shows that the muds have been subjected to chemical changes on the sea bottom, but that they retain their original characters sufficiently to show that the material has been derived from the land. As a rule they are heterogeneous from the admixture of larger or smaller rock and shell fragments. ... . Rock fragments and mineral particles may make up as much as 75 per cent in some cases, the most characteristic species being quartz; the usual proportion of mineral particles is about one-fourth of the whole deposit. Amorphous clayey and muddy matters are always abun- dant, the average percentage being about 60, generally increasing in amount with greater distance from the land. Deltas and a more or less continuous shelf of typical form may be produced artificially on a small scale in a laboratory experiment, the apparatus for which has been described,” but, while such experi- ments serve admirably their purpose of illustration, it is obvious that quantitative results bearing on the relation of the depth of the edge of the shelf to wave-base would be difficult to obtain. As Davis remarks, “‘In most problems of geology and geography experiments have rather an illustrative than a demonstrative value.”’ It has often been pointed out that the edge of the continental shelf is situated everywhere (off exposed coasts) at a constant depth of about too fathoms, that is to say, at the greatest depth to which the water is stirred by wave action. Asa matter of fact the 100-fathom line is situated generally at about the center of a convex curve to which the gentle slope of the continental shelf is tangent and which passes into the more steeply inclined surface of the con- tinental slope. The slope begins to steepen usually from a depth of about 70 fathoms, at which depth it would seem that the effects of wave action are becoming extremely feeble. This constant depth of the “‘continental edge,” together with the fact that a shelf borders, with rare exceptions, all the coasts of the world whatever their origin, whether by regional uplift or subsi- dence, by warping or by faulting, is of great significance (see Fig. 3). * Murray, The Ocean, p. 203. 2R.S. Tarr and O. D. von Engeln, ‘“‘Representation of Land Forms in the Physi- ography Laboratory,” Journal of Geography, VIL (1908), 73-85. i47 DEPOSITION ON CONTINENTAL SHELF AND SLOPE (oeIq) FYs jequautquos ay} SuUIMOYS pj1om ay} Jo de[—'f “ory NO1LIafONd S.MOLVINAW NO GTaYOM AHL a. 148 CLAS CORTON It indicates that the present-day shelf has taken form since sea and land assumed their present relative levels,* and therefore that the molding of the shelf is geologically a very recent event; for most coastal lands afford evidence of considerable movements either of uplift or of subsidence having taken place at a not very distant date. Nansen,” reasoning from the same data, comes to a very different conclusion. While he recognizes the practical uniformity of level of the edge of the continental shelf, he argues from the assumption that the shelf is an ancient feature and makes the remarkable deduc- tion that, though great changes of level have taken place, which in at least some cases he regards as movements of the land rather than of the ocean, the continental margins have everywhere returned to their ancient level. A general idea of the width of the continental shelf is given by Fig. 3. More precise data may be obtained from an inspection of ocean charts upon which the 10o-fathom submarine contour line is or may be drawn, or more conveniently from the numerous maps in Stieler’s atlas which show the 200-meter line. Variation in width depends on several factors; no doubt the most important is the depth of water into which the shelf has grown outward since the latest movement of the strand. If the vertical movement has been small, as, for example, around the British Isles, the present shelf is a modified older shelf, added to at the margin, and therefore broad. Another important factor must be the presence or absence of abundant waste from the land, which will be largely determined by the presence or absence of large rivers in the vicinity. Still another must be the time that has elapsed since the latest important movement of the strand. Variations in the width of the shelf do not affect the question of its essential continuity, which, as pointed out above, prove its capacity for rapid growth and renewal. It is obvious that the «T. C. Chamberlin says of it: ‘“The terrace is as universal (at least in its initial stages) as the sea border and is a necessary consequence of the relations of sea and land”’ (Jour. Geol., VI [1898], 526). 2 F. Nansen, ‘‘Oscillations of Shorelines,’ Geog. Jour., XXVI (1905), 604-9. DEPOSITION ON CONTINENTAL SHELF AND SLOPE 149 shelf bordering a recently uplifted coast which has been cut back to a line of cliffs by energetic wave action must, at its landward edge, be a wave-cut platform; but even in this case farther seaward the waste derived from this retrogradation of the shore together with that brought down by rivers while cliff recession was In progress must have been deposited, forming a seaward extension of the shelf in the manner described on an earlier page. Off uplifted coasts the shore lines of which have not retreated an appreciable distance from the initial position, the shelf must be almost entirely constructional, and the same is true in the case of depressed coasts the initial embayed outline of which has not yet been much modified by marine erosion. THE STRUCTURE OF THE BUILT SHELF The principle is now well understood that, just as in a river the ratio of load to transporting power determines whether degradation or aggradation shall take place, so along a shore line retrogradation or progradation of the shore occurs according as the waves breaking on the shore are underloaded and hungry or are overloaded with waste from another source. With the shore line fixed in position.—It will simplify the dis- cussion of the structure of the built shelf if an ideal case is considered first, in which the load and the transporting power of waves and along-shore currents remain exactly balanced during the building of the shelf, so that the shore line neither retreats nor advances. Fig. 4a is an ideal section of a shelf built under such conditions. The front of the shelf being situated at the level of wave-base during all the stages of growth, it is clear that for each foreset bed there will be a relatively very thin topset bed, the two being differ- ent facies of the same stratum, that the topset beds will thicken seaward, and that successive topset beds will approach more and more nearly to perfect horizontality. With the shore line advancing.—Fig. 4b represents the case in which the waves as they reach the shore are overloaded throughout the whole period during which the shelf is being built. The coast will be continuously prograded, and upon the growing strip of strand plain additional sand from the beach will be piled to form 150 C. A. COTTON dunes. ‘The structure of the resulting shelf will be similar to that with stationary shore line with the exception that each (theoretical) foreset bed will have a subaérial as well as a subaqueous portion. In an actual case the stratification will be very irregular in the shoreward portion of the shelf, as there will be much cross-bedding in both the shallow-water and the subaérial deposits. A very similar result is produced where the shore Jine advances seaward owing to the growth of an alluvial plain formed by con- fluent deltas, as, for example, along the coast of the Canterbury Plains, New Zealand. The structure of the whole mass of deposits, Fic. 4.—Ideal sections across the continental shelf; a, the shore line fixed in position; }, the shore line advancing; c, the shore line retreating. including the alluvium of which the plain is built and the shelf which fringes it, will be similar to that of the delta of a single river as described by Barrell. With the shore line retreating.—In Fig. 4c the commoner case is represented in which the waves reaching the shore have power to erode and where, therefore, retrogradation of the shore takes place (see also Fig. 1). The shelf will consist of two portions, a cut platform and a built platform, and it is obvious that the ratio of the width of the cut to that of the built platform may vary widely. As has been pointed out on an earlier page, it is theoretically possible that, in exceptional cases, there may be no built platform, the material broken by waves together with that coming from the land being swept away by along-shore currents and the shelf being entirely or almost entirely a cut platform such as that of the DEPOSITION ON CONTINENTAL SHELF AND SLOPE Tse Norwegian coast as interpreted by Nansen. On the other hand, owing to nearness of a large source of supply of waste from the land, the shelf may differ but little from that shown in Fig. 4a, there being but little retrogradation of the shore and the shelf consisting almost entirely of a built platform. Davis, in considering the development of the graded shore profile at Cape Cod,’ comes to the conclusion that ‘‘the critical point, where marine action changes from degrading the near-shore bottom to aggrading the off-shore bottom, migrates seaward.” Clearly, however, in the general case, after a continental shelf has been developed, the direction of migration of this critical point will depend upon the relative rates of seaward growth of the shelf and landward retreat of the shore line. When the rate of growth of the built platform is rapid compared with the rate of retreat of the shore line the critical point in the profile will generally migrate landward. In this case no part of the built platform will be sub- sequently eroded and each topset bed will overlap the preceding one, lying upon and protecting from further erosion a strip of the cut platform. If, on the other hand, owing to rapid shore recession, the critical point migrates seaward, some of the earlier-formed topset beds will be obliquely truncated by the cut platform. While the former case is perhaps the commoner in nature, it is the latter which lends itself to diagrammatic representation, owing to the mechanical difficulty of drawing a broad shelf on a narrow page. With the shore line alternately retreating and advancing.—The shore-line features of parts of the coast of New Zealand, notably in eastern Marlborough and western Wellington, indicate that, owing to some disturbance of the balance between load and trans- porting power of waves and along-shore currents, retrogradation and progradation have occurred alternately. Where such alter- nation is taking place, somewhat complex structures in the topset beds will result. During each period of shore retreat the previously formed topset beds near shore will be eroded, and during each advance fresh topset beds will be laid down unconformably on an eroded surface. tW. M. Davis, Geographical Essays (Boston, 1909), pp. 702-3. 152 C. A. COTTON LITHOLOGICAL CHARACTER OF THE BEDS The topset beds deposited during stillstand will constitute such a small proportion of the whole mass of sediment that geologically they will be of slight importance in uplifted shelf deposits. Ocean- ographic data indicate that the materials of which they are com- posed vary widely, ranging from gravel and sand to mud, with or without shells. The bulk of the deposit will consist of foreset beds built out over a thinner series of bottom-set beds in which extremely fine particles that have remained a long time in suspension will be present mixed with organic remains. Farther seaward these will grade into pure pelagic deposits. TABLE I A=Composite analysis of green and blue muds B=Composite analysis of 78 shales A B Si Os ee waceyae esate: 57-05 58.38 TO ieee oN a at! it By 0.65 AE Osea rae: 09422 15.47 1A Oe 28 ocho oun eee 5.07 4.03 18S) O) ais! ale ee aes 2.30 2.46 Vir OAR SMR ie Conia Fy ea Nie giles POEL a oe dN ed Qe eo hs dete cena Ege 27, 2.45 CaO nee ena: 2.04 Boue Ba @ rae iesrel austen ane arts 0.06 0.05 TKO ae aay gi mh) Was 3.25 IN EIA OG ey vcilaiaailenee are 1.05 1.31 12210 ra als Sanco Oecle ae 0.21 0.17 SOTO pea ee een ue mete narnrec tile tale eee ae 0.65 SIRS LD Cae a aes fovea Mir hl ea eee Ns oe ale COS ere sg steal osil lector sis ae ous one 2.64 Cares is ae ea 1.69 0.81 (organic) BEL OS Sas At) sacri ibsley Tait] 5.02 Other constituents... . OST QO Ai Mla anos A ieeca a aerate 99.9964 100.46 Leaving aside volcanic and coral muds, which require special conditions for their formation, we find that, as shown by samples obtained from the continental slopes, blue muds and green sands and muds are the most common materials of the foreset beds, the former where the supply of waste is ample and the latter where the supply is more meager. Both of these are well represented ‘ DEPOSITION ON CONTINENTAL SHELF AND SLOPE 153 among the sedimentary rocks known. to geologists, the former constituting shale and mudstone and graduating into marl, and the latter being represented by greensand and less pure glauconitic rocks. The samples obtained from the continental slope show well- marked stratification., As Murray remarks: ‘The analogues of the now-forming terrigenous deposits are to be found in all geologi- cal periods.’” In this connection an instructive comparison can be made of the chemical analysis of a composite sample of 4 “‘green muds” and 48 “blue muds”’ from various parts of the continental slopes with that of a composite sample of 78 shales taken as giving the average composition of the argillaceous sedimentary rocks, but probably not by any means all of foreset origin. The analyses are given by Clarke. STRUCTURE OF A SHELF BUILT DURING POSITIVE MOVEMENT Without going into the question of the possible causes of changes in the relative levels of sea and land the fact may be accepted that accumulation of sediment has in the past been very commonly accompanied by positive movement. No doubt positive move- ment is in some cases a rise in sea-level and in others a sinking of the floor upon which sediments are being laid down, and in still other cases a combination of both of these. It is convenient, how- ever, in an investigation of the probable structure of shelf deposits, to regard all changes in the relation of the shelf to sea-level as ver- tical movement of the shelf rather than of sea-level. The latter will affect the relation of the sea to the neighboring land also, that is to say, it will be regional in its effects, and obviously precisely similar effects will be brought about by regional subsidence. By regarding the sea-level as fixed and the shelf as subsiding, however, we are able to investigate the effects of differential as well as regional changes of level. To begin with, it may be supposed that a shelf of moderate dimensions has been built forward during a period of stillstand «Sir John Murray, The Ocean, p. 213. 2 [bid., p. 234. 3 F. W. Clarke, ‘“‘The Data of Geochemistry,” U.S. Geol. Surv., Bull. 616, 1916, p. 514 and p. 28. 154 CA. COTTON throughout which there has been a steady stream of sediment from the land. If now this already formed shelf be supposed to subside very slowly, as the shelf sinks, the stream of waste supplied by rivers from the adjacent land may be reduced in volume, may remain sensibly constant, or may be increased in volume, according as the land subsides also, remains stationary, or is uplifted. Let us suppose, however, that the stream of waste is sufficiently con- tinuous and the movement sufficiently slow to allow the shelf to be maintained during subsidence. Obviously maintenance on a sinking floor of a shelf with its outer edge at a fixed depth (approxi- mately 100 fathoms) below sea-level involves deposition of material on the top of the shelf as well as, or instead of, in front of it. Less of the waste will be available for deposition as foreset beds than in the case of a shelf being built during a period of stillstand. The shelf will be built upward of material arrested in transit seaward | owing. to its being lowered below the depth at which wave action stirs it sufficiently to keep it in motion. This accumulation of topset beds may attain an enormous thickness, as has been shown by Barrell to be the case in deltas, the thickness depending only upon the amount of subsidence during which a continuous supply of waste is kept up. Assuming that during subsidence the supply of waste is always sufficiently abundant to insure the continuity of deposition of top- set beds over at least a portion of the area, we may investigate the conditions of deposition of topset beds, foreset beds, and pelagic deposits, understanding by the latter term deposits formed either in moderatly deep or relatively shallow water where the supply of bottom waste fails, the deposit consisting of organic and minute inorganic particles settling from suspension (‘‘flotation beds,” Chamberlin). The form of a single stratum deposited in a given time during stillstand will be that shown in Fig. 5a, that is to say, the bulk of the terrigenous sediment will be in the foreset portion, the edge of the shelf advancing seaward a considerable distance during the given time. With continuous subsidence in progress, however, the formation of thick topset beds may use up so much of the waste that there is little or none left to build the edge of the shelf forward. DEPOSITION ON CONTINENTAL SHELF AND SLOPE 155 If there is still a small advance there will be a terrigenous foreset bed, but it will be relatively thin (Fig. 50). It is only a step from this case to one in which there will be no terrigenous foreset bed, the waste being just sufficient or even insufficient to build a topset bed for the full width of the built shelf, the edge of the shelf in the latter case retreating landward. In no case, however, will the con- tinuity of a stratum be broken, for, if the foreset bed fails, the pelagic deposits will extend up the continental slope to join the topset bed (Fig. 5c). During long-continued subsidence, the duration of which may be comparable with the length of a geological period, the movement Pelagic Pelagic LE Fic. 5.—a, Variation in thickness of a stratum deposited during stillstand; b, a stratum deposited during subsidence with abundant supply of waste; c, a statum deposited during subsidence with restricted supply of waste. is probably never absolutely uniform and continuous. The geo- logical record affords evidence to the contrary, and what is known from physiographic evidence of the more recent movements points in the same direction. Oscillations being left out of account, con- sideration may be given to the consequences of fluctuations in the rate of subsidence. Clearly, with a constant rate of supply of waste, alternating periods of extremely slow and relatively rapid subsidence may determine alternating advances and retreats of the edge of the shelf. It is obvious also that, with slow subsidence at a uniform rate in progress, fluctuations in the supply of waste may produce similar results. Increase in the supply of waste may result from differential elevation of a portion of the neighboring land mass, after which a falling off will take place as a result of regional subsidence or peneplanation of the land, to be followed by a further increase when differential elevation is renewed, and so on. 156 C. A. COTTON The effects of fluctuations in the rate of subsidence of the sea floor and in the rate of supply of waste may be considered together. When the ratio of waste supply to rate of subsidence is sufficiently large the front of the shelf will advance (Fig. 6, B); and when this ratio is sufficiently low the front will retreat (Fig.'6, C or D). Notable fluctuations in the value of the ratio will determine alternating periods of advance and retreat, and these alternations will produce interstratification of beds of very varied lithological character in the shelf deposits which are not without parallel in the known rocks. In the landward portions of the shelf intercalations of subaérially deposited material may be present in the dominantly marine top- set beds, as Barrell has shown to be the case in those portions of Fic. 6.—Diagram illustrating the growth of a shelf during a period of subsidence. A, edge of the shelf with sea-level SL*; B,C, D, possible positions of the edge of the shelf with sea-level SL?. the shelf which are the submarine portions of deltas.t. An instruc- tive example of such alternation has recently been described by Stebinger.? Farther seaward the mass of sediment deposited during sub- sidence may be for a considerable distance composed of marine topset material throughout. The texture of these topset beds will depend very largely upon the relief and also upon the nature of the rocks of the neighboring land; and it may be expected to vary ver- tically, fine sediments marking periods when the land has been reduced to low relief by erosion and coarser sediment marking periods of renewed differential elevation. The topset material will tJ. Barrell, ‘Relative Geological Importance of Continental, Literal and Marine Sedimentation,” Jour. Geol., XIV (1906), 353-54, Fig. 10, p. 445; ‘‘Criteria for the Recognition of Ancient Delta Deposits,” Bull. Geol. Soc. Am., XXIII (1912), 309, Fig. 4. 2. Stebinger, ‘The Montana Group of Northwestern Montana,” U.S. Geol. Surv., Prof. Paper g0G, 1914, pp. 61-68. DEPOSITION ON CONTINENTAL SHELF AND SLOPE 157 not, in general, be subjected to very prolonged chemical and mechanical disintegration on the shelf before being buried. When sandy it may contain, but little altered, grains of the less stable minerals of the shore rocks. If these rocks are predominantly igneous the topset sands when consolidated may form arkose or greywacke, similar to that forming the great bulk of the strata in the mountain ranges of New Zealand. Of these rocks, commonly ascribed to the Maitai System, Marshall writes: The material of which all these rocks is composed has been derived from plutonic masses, for they are composed of grains of quartz, feldspar, and horn- blende or augite..... The great thickness of the sediments shows that the area was one of deposition for a considerable time, though the general coarse- ness of the material shows that the deposition was relatively rapid, and took place on a coast-line. Presumably the coast-line fringed a large continental area, from the surface of which rivers carried large quantities of sand.! Preservation of the grains of decomposable minerals, however, cannot safely be taken as certain proof of topset origin of the beds in which they occur, for it is conceivable that, with a very abun- dant supply of terrigenous material during a period of little or no subsidence, similar material might be buried in foreset beds. Barrell? has indicated the origin of alternations of sandy and muddy layers of small thickness as a result of topset deposition. The sediment on the outer, deeper part of the continental shelf, near the maximum depth at which the bottom is ever stirred by wave action, is affected only by the waves produced by exception- ally severe storms, such as occur only once in a number of years. During the interval between two such storms an unsorted mixture of mud and fine sand accumulates. When a storm occurs, the gentle stirring of the bottom which it produces causes the finer particles of the superficial layer to go temporarily into suspension, the larger grains remaining as a layer of clean washed sand. After the storm, subsidence being continually in progress, another layer of sandy mud is laid down above the sand, and by the time the next great storm occurs the sand layer and the deeper part of the tP. Marshall, Geology of New Zealand (Wellington: Government Printer, 1912), pp. 184-85. 2 J. Barrell, “Criteria for the Recognition of Ancient Delta Deposits,” Bull. Geol: Soc. Am., XXIII (1912), 428. 158 C. A. COTTON ee overlying mud layer have been so deeply buried that they are secure from further disturbance. The sand of a superficial layer is, however, again washed clean, and so the process goes on until there are innumerable alternating sand and mud layers throughout a great thickness of strata. Toward the outer part of the mass of sediment intercalations of foreset beds and pelagic deposits may be present in the topset beds owing to advances and retreats of the edge of the shelf correspond- ing to fluctuations in the ratio of waste supply to rate of subsidence, the possibility of which has been pointed out on an earlier page. In Fig. 7, which represents diagrammatically a section of a shelf Fic. 7.—Alternation of lithological types in shelf deposits resulting from fluctua- tion in the ratio of supply of waste to rate of subsidence. Topset beds, black; foreset beds, white; pelagic beds, with cross-lines. The top of the shelf at successive stages is shown by the white lines. The vertical scale is exaggerated about ten times. the front of which has alternately retreated and advanced during upward growth, A is the front of a shelf built forward during a period of stillstand preceding the subsidence, B, D, and F are positions of the front after episodes of small ratio of waste supply to rate of subsidence, and C, EF, and G are positions of the front after episodes during which this ratio has had a large value. In order to simplify the diagram the lateral transitions from one type of sediment and from one slope to another are represented as per- fectly sharp, but it must be borne in mind that in nature these transitions are gradual. An inspection of this diagram makes it clear that above A there will be in the region represented by the middle part of the diagram the following succession: ‘DEPOSITION ON CONTINENTAL SHELF AND SLOPE 159 TOM kopset DedSuacra ace Coarse- to fine-grained sandstone, arkose, and grey- wacke, with interbedded shale or argillite. o- Foreset beds. -.22-- More or less calcareous bluish mudstone or green- sand. SrakelagicsbedS:ee- asm: Argillaceous to pure limestone, perhaps glauconitic, or somewhat sandy. 4 Topset beds.......- As above. 6 Foreset beds: ... 2 .: As above. 5) Pelagic beds4. 42. As above. 4 Topset beds. 3. Foreset beds. 2 Pelagic beds. 1 Topset beds. Farther seaward there will be no intercalations of topset deposits, but an alternation of limestone bands with bluish mudstone, Fic. 8.—Enlargement of the portion ABC of Fig. 7 marl, or greensand. In this way may perhaps be explained the intercalation of the Amuri limestone of New Zealand, between mudstone and marl in Marlborough, and between greensand and glauconitic sandy limestone passing upward into marl in North Canterbury, and also the intercalation of the Oamaru limestone between beds of greensand in Otago, New Zealand. The intercalations of foreset beds in the pelagic deposits will become more calcareous and thinner and will finally die out sea- ward. Obviously, under certain conditions of restricted waste supply this may occur in moderately shallow water, the proportion of foreset beds in the whole mass of sediment being very small. In Fig. 7, which is designed to represent the vertical distribution of lithological types, the continuity of strata is not apparent, but in Fig. 8, which is an enlargement of the portion ABC of Fig. 7, the 160 (Co Ale COIN ON continuity of strata is shown. It will be seen that inthe section AB, built during the period of retreat of the shelf front from A to B, the topset beds will pass laterally into pelagic deposits, and that in the section BC, built during the period of advance from B to C, the first-formed strata will pass somewhat rapidly through foreset beds into pelagic deposits. In the later-formed strata of the section BC, however, the foreset portions will assume greater importance, the transition to pelagic deposits taking place in the deeper water farther seaward. STRUCTURE OF A SHELF BUILT DURING NEGATIVE MOVEMENT Negative movement generally involves uplift of the adjacent lands as well as of the sea floor. So erosion will as a rule be revived and the supply of waste increased. Also as the sea retreats the former surface of the shelf will be subject to subaérial and marine erosion, producing a further supply of waste. Negative movement will therefore be generally accompanied by heavy sedimentation on the continental slope. It is conceivable, therefore, that the shelf may grow seaward with sufficient rapidity to maintain its edge at the usual depth throughout a period of rather rapid movement. During such move- ment the width of the shelf (measured from the ever-changing shore line) may diminish, remain constant, or even increase. In the first case the topset slope will be steepened by submarine erosion, and in the last case it will become less steep owing to deposition of topset beds. With constant width will go constant slope, with neither deposition nor erosion, the erosion that goes on at the shore line affecting only the emerging land. From the foregoing it appears that, while topset beds are not necessarily absent, they can be only very thin, and they will be largely removed by erosion as the shelf emerges. The bulk of the deposit forms foreset beds, and the material of these will be of somewhat coarse texture. THE NORTHWARD EXTENSION OF THE PHYSIO- GRAPHIC DIVISIONS OF THE UNITED STATES |W. N. THAYER Consulting Geologist, Cincinnati, Ohio PART I INTRODUCTION Preliminary to a study of the economic bearing of the physiog- raphy of North America the writer found it desirable to inquire into the extension of the generally recognized physiographic divisions of the United States southward into Mexico and northward into Canada and Alaska. ‘The results of the study of Mexican physiog- raphy were published in the Journal of Geology early in 1916.1 The present paper embodies the results of an investigation of the north- ward extent of these divisions. The plan of this paper is to discuss briefly the physiographic divisions of the United States which touch our northern border, and to compare with them the adjacent territory north of the International Boundary by reference to surface features, boundaries, structure, and physiographic history, and by this means to show that the divisions of the United States have northern extensions that project them far into Canada and in some places into Alaska. This paper is avowedly one of correlation, and no effort has been made to give detailed descriptions of the Canadian or Alaskan areas. Such work must be left to the future, as there are still large expanses of territory that have never been fully explored, much less studied, with care sufficient to allow an accurate classification of surface features or the drawing of permanent boundaries. The generalizations advanced in this paper will, of course, be subject to change as our knowledge of the north country increases. Fenneman’s classification has been used wherever it could be adapted to the continental scope of this paper. The writer’s own tW.N. Thayer, Jour. Geol., XXIV (1916), 61-94. 161 162 W. N. THAYER experience and field study have contributed in a small way to the text, but he has also drawn liberally on the published work of others. He is particularly indebted to Dr. N. M. Fenneman for advice and criticism. THE COAST RANGES SECTION OF THE PACIFIC BORDER PROVINCE The term ‘‘Coast Ranges”? may be used with perfect freedom when discussing topographic features within the United States, because in both popular and scientific thought the mountains desig- nated by the term are quite definitely delimited. Freedom in the use of the term is restricted beyond the International Boundary, however, for in Canada there is also a ‘“‘Coast Range” in no way related to the “Coast Ranges” of the United States, and the name is very definitely fixed in the language of the people as well as in scientific usage. It becomes necessary, therefore, to distinguish these features by appropriate terms that shall leave no room for ambiguity. This will be done in the present paper by the expedi- ent of using the plural form, “Coast Ranges,’’ for those mountains both in the United States and in Canada which face the open ocean and the singular form, “Coast Range,” for that Canadian member of the Pacific System which is separated from the open ocean by numerous mountainous coastal islands. The two features have but little in common and differ widely in their records of physiographic history. The mountains that border the coast of the United States from the Sierra de Los Angeles to the straits of San Juan de Fuca ‘“‘are neither a single range nor alike in character and history, but they are for the most part contiguous and may be treated as a single general province.’* A similar characterization may be made of the mountains of Vancouver and Queen Charlotte islands, the Alexander Archipelago, the St. Elias group, the Kenai Peninsula, and Kodiak Island, and for the same reasons they may be considered as an extension of the Coast Ranges of the United States and as belonging to the same province. There is some objection to this broad view, particularly because a large part of the region north of the forty-ninth parallel has not tN. M. Fenneman, Ann. Assoc. Am. Geog., IV, 133. PHYVSIOGRAPHIC EXTENSION OF UNITED STATES 163 been studied in detail, but it is supported in a general way by numerous geologists who have studied portions of the northwest coast of North America. Dawson,’ Ransome,’ and Clapp’ tenta- tively place the mountains of Vancouver and Queen Charlotte islands in the province with the Coast Ranges, and Willis and Smith* are definitely committed to the idea. Bancroft’ says that the system of the Coast Ranges is continued northward along the coast of Alaska in the mountainous islands of the Alexander Archipelago, and that these islands express the probable continuity of a range that formerly bridged all gaps existing between the Olympic Mountains of Washington and the St. Elias group of Alaska. Brooks® continues the system northward, connecting the Alexander Archipelago with the St. Elias group and that group, through the Chugach Mountains, with the mountains of the Kenai Peninsula and Kodiak Island. The four more or less dissimilar divisions of the Coast Ranges in the United States have this in common—they represent dissected peneplains. Topographically they consist of individual ridges, and, excepting the Klamath Mountains, they follow the contour of the coast with considerable parallelism between them. The tops of the ridges are generally flat and the upland has a rolling, mature character, with peaks rising here and there above the general level. These peaks are residuals or monadnocks that resisted erosion during the first cycle. Diastrophism and vulcanism have con- tributed something to the present topography, but, though locally prominent factors, they have been generally subordinate to erosion. The drainage of the Coast Ranges is a reliable index to the causes that have produced the present topography. The larger westward-flowing streams, that is, those which cut across the ranges to empty their waters directly into the Pacific, are without excep- tion antecedent. They have preserved the courses which they tG. M. Dawson, Bull. Geol. Soc. Am., XII, 61. 2 F, L. Ransome in Problems of American Geology (Yale University Press), p. 359. 3C. H. Clapp, Geol. Survey Canada, Guide Book No. 8, Part III, p. 280. 4 Bailey Willis and G. O. Smith, U.S. Geol. Survey, Folio 54. 5 J. A. Bancroft, Geol. Survey Canada, Mem. 23, p. 18. 5A. H. Brooks, U.S. Geol. Survey, Prof. Paper 45, pp. 27-42. 164 W. N. THAYER formerly had while the old surface was being subdued. The Sacra- mento and Columbia rivers are prominent examples of this type. TABLE TO EXPLAIN THE ACCOMPANYING MAP Major divisions, the strongly characterized parts of the continent, are separated by heavy lines, and are distinguished by Arabic numerals; Provinces are separated by light lines, and are distinguished by letters; sections are distinguished by Roman numerals. Major Divisions 1. Laurentian Plateau (a) (b) (c) (d) Province Section Laurentian Plateau proper Superior Highlands Adirondack Mountains Unnamed 2. Atlantic Plain (a) (0) Continental shelf Coastalsblainmeer eee . Atlantic section . Gulf section _ . Yucatan section 3. Appalachian Highlands Piedmont region Blue Ridge Mountains Appalachian Valley St. Lawrence Valley Appalachian Plateaus New England region 4. Interior Plains Interior Low Plateaus Central Lowland........ Great Plains Wyoming Basin Texas Hill region Canadian Great Plains Anatuvuk Plateau I. East. Lake section . West. Lake section . Driftless area . Till plains 5. Interior Highlands Ozark Plateau Ouachita region......... I. Arkansas Valley II. Ouachita Mountains 6. Rocky Mountain System Southern Rockies Boundary group Mackenzie Mountains Endicott Mountains 7. Intermontane Plateaus Columbia Plateau Colorado Plateau Basin-and-Range province Interior Plateaus Yukon Plateau 8. Pacific Mountain System (f) Sonoran Desert (g) Anahuac Desert Plateau (hk) Sierra Madre I. Sierra Nevada (a) Pacific Mountains....... II. Cascade Mountains III. Coast Range of British Columbia IV. Coast Range of Alaska Pacific Border province... Coast Ranges........... I. California trough . Puget trough . Copper River basin I. California-Oregon-WashingtonRanges . Alaskan Ranges 9. Southern Mexican High- lands (a) (b) Volcanic province Sierra del Sur province to. Isthmian Lowlands =~ 8 LS Tehuantepecan province PHYSIOGRAPHIC EXTENSION OF UNITED STATES 165 The streams now occupying the longitudinal valleys have in part inherited courses from the earlier cycle, but they generally follow structural lines or have had their courses determined by folding 70 80 100 20 = SS SS T] = += = - W.M: Davts~ 289 PERMO-CARBONIFEROUS GLACIAL DEPOSITS OF SOUTH AMERICA A. P. CoreMaN 310 THE SUBPROVINCIAL LIMITATIONS OF PRE-CAMBRIAN NOMENCLATURE IN THE Cr EAWRENGH BASIN ©. gua pete ne ee ee NL Ex WabsoNes 325 CORRELATION OF THE EARLY SILURIAN ROCKS IN THE HUDSON BAY REGION T. E. SAVAGE 334 NOTES ON SEDIMENTATION IN THE MACKENZIE RIVER BASIN’ - E. M. Kinvte NOTES ON THE MISSISSIPPIAN CHERT OF THE ST. LOUIS AREA Downatp C. Barton EDITORIAL Ls aaepiaist Oy S SGN aR ey ATIC | pees UME Det cee tga SUN SuPer cna Mantel lia PETROLOGICAL ABSTRACTS AND REVIEWS -. - - - | - ALBERT JOHANNSEN RCN TEU ERMC AMIONG suegmermme ena en iyay Vee uel tenn tyres Neck cur ey OV | wD W WH & SI on LS) (rhe UNIVER iY OC CHICAGO PRESS CHICAGO; TELINOIS, US 7A. THE CAMBRIDGE UNIVERSITY PRESS, Lonpon anp EpINBURGH THE MARUZEN-KABUSHIKI-KAISHA, Tokyo, Osaka, Kyoto, Fukuoka, SENDAI THE MISSION BOOK COMPANY, SHANGHAI % thee Schad? Sen Ae wk a ey aA ae EDITED BY THOMAS C. CHAMBERLIN AND ROLLIN D. SALISBURY ie With the Active Collaboration of - SAMUEL W. WILLISTON ALBERT JOHANNSEN ts yy: 2 Vertebrate Paleontology — Petrology ee oR oe teajaniegh ‘ , STUART WELLER ROLLIN T. ‘(CHAMBERLIN © re etl a SAT ee Invertebrate Paleontology : _ Dynamic Geology ALBERT D. BROKAW : 2 aeti et pes Economic Geology eae ping aan The Journal of Geology is published by the University of Chicago at the University Press semi- quarterly, on or about the following dates: February 1, March 15, May 1, June 15, August 1, September 15, 2 November 1, December 15. | The subscription price is $4.00 per year; the price of single copies is" 65 cents. Orders for service of less than a half-year will be charged at the single-copy rate. { Postage is prepaid by the publishers on all orders from the United States, Mexico, Cuba, Porto Rico, Panama Canal Zone, Republic of Panama, Hawaiian Islands, Philippine Islands, Guam, Samoan Islands, Shanghai. se {| Postage is charged extra as follows: For Canada, 30 cents on annual subscriptions (total $4.30), on single copies 4 cents (total 69 cents); for all other countries in the Postal Union, 53 cents on annual sub- scriptions (total $4.53), on single copies 11 cents (total 76 cents.) {| Patrons are requested to make all remittances payable to The University of Chicago Press in postal or express money orders or —_ drafts. | The following are authorized to quote the prices indicated: For the British Empire: The Cambridge University Press, Fetter Lane, London, E.C. 4. Yearly subscriptions, including postage, 18s. 8d. each; single copies, including postage, 3s. 2d. each. For Japan and Korea: The Maruzen-Kabushiki-Kaisha, 11 to 16 Nihonbashi Tori Sanchome, Tokyo, Japan. Yearly subscriptions, including postage, Yen ‘93 oo each; single copies, ee postage, Yen 1.50 each. For China: The Mission Book Company, 18 Peking ia Shanghai. Yearly Peer ton’ $ $4.00; single copies 65 cents, or their equivalents in Chinese money. Postage extra, if mailed direct outside of Shanghai, on yearly subscriptions 53 cents, on single copies 11 cents. Claims for missing numbers should be made within the month following the regular month of publi- cation. The publishers expect to supply missing numbers free only when losses have been sustained in transit, and when the reserve stock will permit. Deliveries are not guaranteed within the war zones of Europe, nor in other countries where transportation is disturbed by conditions incident to war. Business correspondence should be addressed to The University of Chicago Press, Chicago, Il. Communications for the editors and manuscripts should be addressed to the Editors of THE JOURNAL OF GEOLOGY, the University of Chicago, Chicago, Ill. Contributors are requested to write scientific and proper names with particular care and in citations to follow the form shown in the pages of the Journal. é Separates, if desired, must be ordered in advance of publication. Separates of leading articles will be printed as ordered, of which 50 (without covers) will be furnished gratis, the actual cost of the remainder (and covers, if desired) to be paid for by the author. The actual cost may vary, and will depend upon the amount of work in remaking the pages into forms, presswork, paper, binding, etc. Separates con- taining half-tones may be expected to cost somewhat more, the increase depending upon the number of cuts and the amount of work required apes them. Entered as second-class matter, March 20, 1893, at the Post-office at Chicago, IIl., under the Act of March 3, 1879. VOLUME XXVI NUMBER 4 THE JOURNAL OF GEOLOGY MAY-JUNE 1918 CORAL REEFS AND SUBMARINE BANKS W. M. DAVIS Cambridge, Massachusetts PART II Fundamental postulates of the glacial-control theory.—It we now return to the consideration of the principal postulate of the glacial-control theory, it should be noted that the features appealed to in evidence of long-enduring ocean-bottom stability are not directly concerned with the deep ocean floor itself, for the ocean floor is inaccessible to geological observation; nor do they concern the geological structure of islands with elevated reefs which give a decipherable record, for nearly all of these islands seem to have had elevations and subsidences during their long history. The glacial- control theory is chiefly concerned with atolls and submarine banks, the history of which is not directly decipherable; and the features which are taken to demand long-enduring crustal stability are, as already noted, the nearly level floors and the similar and moderate depths of atoll and barrier-reef lagoons and of submarine banks, although they are all covered with unconsolidated calcareous deposits of recent deposition and unknown thickness. In seeking to account for these well-substantiated submarine features, the glacial-control theory advances reasons for thinking that the nearly level lagoon deposits must be of small thickness and 289 290 W. M. DAVIS must therefore rest on nearly level rock platforms at a moderate and fairly uniform depth in all the coral seas. It is next argued that the production of the inferred rock platforms was preceded by subaérial erosion and marine abrasion during a long preglacial period of “‘general crustal stability in the coral sea,”’ whereby lofty volcanic islands were worn down to insular lowlands. Then, in view of the plausible inference that reef-making corals were killed during the glacial period, it is further supposed, as previously outlined, that the chilled and lowered ocean was free to cut away the dead reefs and to reduce the worn-down islands to flat platforms. Finally, as the ocean rose and warmed in postglacial time, the existing reefs are believed to have grown up around the platform margins, while unconsolidated calcareous deposits were strewn evenly over the platform surface; or, if reefs did not grow up, the platforms remained as submarine banks. The ingenuity with which the successive conditions and pro- cesses are enchained is certainly admirable, but the fundamental assumptions do not seem to be fully established, and the conclusions reached are not unescapable. The links of the chain of argument must be separately examined. The plausible inference that reef- making corals were killed during the glacial period will be con- sidered in the next two sections. Coral reefs not destroyed during the glacial period. ee is certainly plausible to suppose that reef-building corals may have been killed by the lowering of oceanic temperatures during the glacial epochs of the glacial period, but no sufficient tests of the correctness of this supposition have been brought forward. As far as I have been able to analyze the question, the reefs were not so completely divested of protecting organisms, except along the margin of the coral zone, as to expose them and their inclosed islands to abrasion, for the consequences of such abrasion are not found in the expectable form of spur-end cliffs (KN, Fig. 3b) on the central islands of fringing or of close-set barrier reefs, as I have elsewhere pointed out ;* hence this essential element of the glacial-control theory lacks 1“ A Shaler Memorial Study of Coral Reefs,” Amer. Jour. Sci., XL (1015), 223-71; see pp. 236-46. ‘‘Problems Associated with the Study of Coral Reefs,” Sci. Monthly, II (1916), 559-77. CORAL REEFS AND SUBMARINE BANKS 291 support. The previously stated arguments which lead to this conclusion need not be repeated here, but additional evidence to the same end may be presented from Hawaii, Tahiti, and Murea. Evidence from Hawai, Tahiti, and Murea.—Let it first be noted that Hawaii, Tahiti, and Murea are three of four islands—the other is Rotuma, north of Fiji—which Daly instances as having “submarine benches so narrow as to prove the extraordinary power of fresh lavas to resist the Pleistocene breakers” (182). Rotuma I have not seen, but Gardiner’s account of it’ suggests that the narrowness or absence of an abraded bench around it may be due rather to the recency of its latest lava flows than to their resistance. Hawaii also was not reached during my voyage of 1914, but Bran- ner’s account of it? and the several excellent topographic sheets recently published by the United States Geological Survey of its northern and northeastern coast give good reason for thinking that the general absence of a cliff-backed bench around its shores is due to the recency of the eruptions that have covered most of its slopes with lava flows in which valleys are not yet eroded; while the occurrence of deep consequent valleys and strongly clift intervalley spur ends in the Hamakua district of the northeast coast gives equally good reason for regarding that part of the island as much older than the rest—old enough in fact to have suffered submature dissection and mature abrasion, but at a time when the island stood several hundred feet higher than now, as Branner has so well shown. By constructing longitudinal sections and cross-sections of the great valleys from the topographical maps, I have inferred that the amount of submergence since the valleys were cut down and the cliffs were cut back is some 800 feet. But whether submerged or not, the island of Hawaii ought not to be instanced in proof of the belief that the absence of spur-end cliffs on the central islands of barrier reefs elsewhere in the Pacific is due to the resistance of their lavas: to my reading Hawaii supports the other side of the tJ. S. Gardiner, “‘The Coral Reefs of Funafuti, Rotuma, and Fiji... . ,” Proc. Cambr. Phil. Soc., UX (1898), 417-503. 2J. C. Branner, ‘‘Notes on the Geology of the Hawaiian Islands,” Amer. Jour. Sci., XVI (1903), 301-16. 292 W. M. DAVIS discussion, namely, that volcanic islands which are indented by drowned-valley embayments half a mile or a mile wide must, like northeastern Hawaii, have had strong cliffs cut in their spur ends if they were exposed to abrasion during the considerable period required for the erosion of the now drowned valleys; and as such cliffs are prevailingly absent on reef-encircled islands the islands must have been protected from abrasion—that is, their reefs must have been clothed with growing organisms of some sort—while the now drowned valleys were eroded. Tahiti gives even stronger evidence to the same end. This island consists of a larger and a smaller volcanic cone, joined in an isthmus. The slopes of the cones are submaturely or maturely dissected by deep, steep-sided, radial valleys, the lower ends of which have been embayed by submergence after they were eroded; but the embayments are now, with few exceptions, filled with deltas which have expanded outside of their embayments so as to form a continuous alluvial belt around much of the shore line; it was this “broad belt of low land at the foot of the mountains” which Darwin properly interpreted as indicating “‘a long stationary period” for Tahiti. It is, however, not the radial valleys but the spur-end cliffs that are the most significant feature of Tahiti in the present connection. Agassiz gave a good account of them. Some of the cliffs rise 500 or 1,000 feet above present sea-level on the more exposed coasts; but the northwestern or leeward corner of the island is without visible cliffs for a short distance. Their origin must have been contemporaneous with the erosion of the valleys; hence the base of the cliffs and the platform that must have been abraded at a small depth below sea-level when the cliffs were cut back must be now, like the distal portion of the valleys, submerged. The depth of submergence, inferred from the cross- section of some of the larger valleys, may well be 500 or 600 feet; the lagoon, much aggraded, is naturally of less depth. It is immaterial for the moment whether the opportunity for cliff-cutting around Tahiti was given while the ocean was chilled and lowered in the glacial period, or whether it was given while the island stood higher and was, like Reunion, reefless because a sheet of shore detritus prevented coral growth, as I believe is the more CORAL REEFS AND SUBMARINE BANKS 203 probable explanation.* The lesson to be learned here is that the time which sufficed for the erosion of the steep-sided valleys of Tahiti sufficed also for the cutting of its great cliffs; hence, far from proving the power of fresh lavas to resist the breakers, Tahiti proves that, resistant as its lavas may be—and they certainly look very resistant where ledges are exposed in the cliff faces and on the valley sides—the waves of the trade-wind sea can strongly abrade the coastal slope of a volcanic island in the same period of time that is needed for the erosion of submature or mature valleys; and that the spur-end cliffs thus formed may be of such height that they will still rise hundreds of feet above sea-level after the cliff base and the platform in front of it and the valley ends between the clift spur ends are submerged hundreds of feet below sea-level. It there- fore still seems to me reasonable to be guided by the conclusion elsewhere set forth that the general absence of spur-end cliffs in the central islands of close-set barrier reefs and especially in islands like Rarotonga, that have only a fringing reef, contradicts the assump- tion that reef-building corals were killed and that the islands that they normally protect were exposed to abrasion during the glacial period. In order to avoid misunderstanding, let it be added that many spur ends of reef-encircled islands are nipped off in low, freshly cut bluffs,-5 to 30 feet in height, and that these low bluffs are fronted by wave-swept rock platforms of small breadth visible at low tide; hence these bluffs and platforms must be ascribed to wave work of the lagoon waters now or recently in progress at present sea-level. Further, let it be noted in passing that the subsidence inferred for Hawaii must have taken place while its gigantic young cones were in process of formation by eruption; and that a subsidence of Tahiti during its eruptive growth is also suggested, but less defi- nitely, by its association with its more submerged neighbors. These two instances deserve consideration as antidotes to the pre- vailing idea that volcanic eruption is necessarily associated with upheaval. Itis eminently possible, if not probable, that the reverse relation may often obtain, and still more possible that subsidence may set in shortly after eruption ceases. t “Clift Islands in the Coral Seas,”’ Proc. Nat. Acad. Sci., II (1916), 284-88. 204 W. M. DAVIS Murea, near Tahiti, may be briefly considered. Its dissection is far more advanced than that of Tahiti; its slopes are less steep, its embayments are much larger and wider; several spur ends of its northwest or leeward coast are rather strongly truncated in sloping facets resembling mature cliffs, but around most of its circuit the spurs descend gradually to sea-level; the barrier reef here is compara- tively close-set. Surely, if the great cliffs of Tahiti were cut under the conditions assumed in the glacial-control theory, great cliffs should have been cut at the same time around Murea, especially on its southern and eastern sides; the absence of such cliffs indicates that the sea has not had access to the island shore, or, in other words, that the encircling reef of Murea has long protected the island from abrasion. The absence of cliffs cannot be explained under the glacial-control theory by postulating an exceptional resistance for the lavas of Murea; for if the total duration of lowered sea-level during the glacial period were long enough for small streams to deepen their valleys and for the slow processes of weathering to widen the deepened valleys to the form now seen in the drowned- valley embayments in spite of the exceptional resistance postulated for the island lavas, then the waves of the lowered sea working through the same duration of time should all the more have cut great spur-end cliffs; for there can be no question that the waves of the trade-wind sea are much more powerful agents of island sculpture than the streams of short valleys and the weather changes on the valley sides. The same statement may be made for Raro- tonga in the Cook group; it is elaborately dissected by wide valleys, but its embayments are replaced by alluvial plains. Its spur ends are not clift. The Society Islands, therefore, do not support—indeed, they strongly contradict—the consequences expectable from the glacial- control theory as to the abrasion of preglacial reefs; and I believe that they likewise give no support to, if they do not contradict, the fundamental postulate of the “long period of nearly perfect stability for the general ocean floor” on which that theory is constructed; for the depth of submergence indicated by the form of the mountain slopes that inclose the embayed valleys of these islands demands a greater submergence than the glacial-control theory provides. CORAL REEFS AND SUBMARINE BANKS 295 Furthermore, the zodlogical evidence provided by Crampton’s recent elaborate investigations is so strongly confirmatory of the subsidence which earlier investigators had inferred from similar but briefer studies that the postulate of stability for these islands in particular needs revision. Crampton’s conclusion is: ‘‘The occur- rence of related forms [of land snails] in Tahiti, Raiatea, and Moorea means that in former times these islands were connected by land; that the common ancestral stock ranged over the whole land mass, and that its local products differentiated into the distinct species after the process of subsidence had isolated the mountains now forming the separate islands.’ No truncated volcanoes are known in the coral seas.—If atolls have been formed by the processes of the glacial-control theory, then after an atoll is sufficiently uplifted and dissected an abraded volcanic platform, around the margin of which the atoll reef was built up, should become visible. Most uplifted atolls are either not raised enough or not eroded enough to reveal a foundation 40 fathoms or 240 feet beneath their reef level. In eastern Fiji, however, a number of limestone islands, that I believe were formed — as atolls, have been sufficiently uplifted and dissected to reveal their volcanic platform if it ever existed. In no case is it laid bare; hence, as far as these examples go they discountenance the theory. I have given details of these islands, taken chiefly from reports by Gardiner and Agassiz, in another paper.? There are, on the other hand, anumber of uplifted fringing and barrier reefs in Fiji, and also I believe in the New Hebrides, the Solomon Islands, the Philip- pines, and elsewhere, which rest, as above noted, unconformably on the previously eroded slopes of volcanic or other foundations, and in which the length of the eroded slope is so great—and I believe it may be added, the volume of erosion accomplished in shaping the slope is so large—that neither the length of the slope nor the volume of the erosion can be reasonably explained as the work of subaérial destructive processes while the sea-level was lowered only 40 1H. E. Crampton, ‘“‘Studies in the Variation, Distribution, and Evolution of the Genus Partula.... ,” Carnegie Institution of Washington, 1916, p. 206. 2“The Structure of High-Standing Atolls,” Proc. Nat. Acad. Sci., IIL (1917), 473-79- 206 W. M. DAVIS fathoms in the glacial epochs. All such reefs appear to me to demand subsidence for their explanation; but as they all occur in the Western Pacific, and not in the central area where only atolls prevail, they do not bear directly on the atoll problem. All that can be said of such unconformable reefs and of the uplifted and dissected atolls of Fiji is that their evidence is highly favorable to Darwin’s theory, and that it is in some degree irrelevant to the origin of open-ocean atolls, which are the main subject of the glacial- control theory. . As far as my reading goes only three sections have been pub- lished in which the foundation of an actual coral reef is represented as a truncated volcanic mass. One section is of the island of Mango, Fiji, as interpreted by E. C. Andrews; the truncated sur- face is represented as covered by a now elevated coral reef into and over which later volcanic rocks are erupted; the section is reproduced as if authentic in de Margerie’s translation? of Suess’s Antlitz der Ende; but, as the surface of truncation 1s drawn below present sea-level and as the accompanying text gives no sufficient evidence of its existence, it must be regarded as hypothetical. My brief visit to the island did not enable me to examine its structure closely; but nothing that I saw gave any support to the theory of its having suffered truncation before its elevated reef was formed, nor did it appear to me that the elevated reef is older than the volcanic rocks that are associated with it. If the island really has been completely truncated, it constitutes a remarkable exception to the rule prevailing in Fiji, where the other volcanic islands have not been cut back enough to form strong shore cliffs. A second section of a truncated volcanic island is to be found in Pirsson’s account of the recent boring at Bermuda,’ where volcanic rocks were reached at a depth of 245 feet below sea-level and pene- trated to a depth of 1,278 feet. It appears to me regrettable that a single boring of this kind should be accepted as giving sufficient «—. C. Andrews, ‘‘The General Geology of the Fiji Islands .... ,” Bull. Mus. Comp. Zool., XX XVIII (1900), 1-50. 2 La Face de la Terre, III (1913), 1061. 3L. V. Pirsson, ‘‘Geology of Bermuda Island; the Igneous Platform,’ Amer. Jour. Sci., XXXVIII (1914), 189-206. CORAL REEFS AND SUBMARINE BANKS “2077 ground for drawing the volcanic cone with a broadly truncated surface ‘“‘cut away by the action of the sea” at the depth where igneous rocks were first encountered; almost any other form would accord as well with the recorded facts; indeed, in view of the strong variations of magnetic force on the island surface,’ it is eminently probable that the buried volcanic surface is uneven; if so, it is also probable that the depth of 245 feet is not the minimum depth of the volcanic foundation; for if the surface is uneven it is not likely that a single boring, located without any knowledge of the under- structure, would reach the culminating point. The-third example of a truncated volcanic island is Mangaia in the Cook group, which is briefly described by Marshall as having ‘a, well-developed marine erosion surface forming the summit of the island 650 feet above sea level. An alluvial flat... . sepa- rates the high volcanic land from a ring of coral 125 feet above sea level’... . 200 of 300 yards from the volcanic land.”? ~ This example, being visible, is better attested than the other two. It would seem to represent a volcanic cone that was completely trun- cated by abrasion before any reef defended it, and then elevated; but the barrier-reef ring now surrounding it may well have been formed during a later depression before a later elevation. _It is not without careful consideration that I have been con- strained to reject the assumption that reef-building organisms were so completely killed during the glacial period as to leave the reefs an easy prey to the waves. The assumption is, as already noted, certainly a plausible one at first hearing, and it merits careful examination; but as the result of the best examination that I have been able to devise it proves to be erroneous. Thus a problem is laid before zoologists: If coral reefs are today limited by tempera- ture conditions and if the ocean were significantly cooled during the glacial period, why were not the reef-building organisms then killed? Perhaps the organisms were killed and the reefs were cut away on islands near the borders of the coral zone, as will be tJ. F. Cole, ““Magnetic Declination and Latitude in the Bermudas,” Terrestr. Magnetism, XIII (1908), 49-56. 2P. Marshall, “Coral Reefs of the Cook and Society Islands,” Proc. Austral. Assoc. Adv. Sci., XIII, 1912, 140-45. 208 |W. M. DAVIS further considered later. Perhaps the corals were very generally killed, but the nullipores were not; if so, the nullipores, although unable to construct a reef alone, might cover and protect the exposed flanks of an already constructed reef for a geologically brief epoch until the corals could again establish themselves upon it. The floors of atoll lagoons—Whatever solution the question of the survival of reef-building corals during the glacial period may eventually receive, it may be set aside for the present and return may be made to the main facts upon which the glacial-control theory is built, with the question, Is it really impossible to explain the smoothness of atoll-lagoon floors and of submarine banks with- out prolonged abrasion of still-standing islands? This seems to me no impossibility. Those smooth surfaces are not the result of abrasion, even if an abraded surface exists beneath them; they are the result of the even distribution of organic sediments by agencies now in operation, whatever the shape of the foundation that the sediments rest upon, as will appear from the following consideration. As to atolls, it is true that the waters of their lagoons are generally placid, but it is also true that at times of storm they are agitated sufficiently to become turbid by stirring up the bottom sediments. Gardiner’s testimony on this point, based on observa- tions in the Maldives, is important: ~ It is only in a few protected situations, where the depth is as great as 40 fathoms or more, that the lagoon bottom appears not to be churned up by the currents and waves. In heavy weather the lagoon water is almost milky, and floating surface nets [for zodlogical collecting ?] are almost useless on account of the enormous amount of mud in suspension. The total amount of mud that passes out of the lagoon in the water is enormous.? Hence I cannot accept Daly’s statements that ‘‘the lagoon floor . is little or not at all disturbed by any waves or currents generated in the lagoon itself,” and that ‘‘the filling and smoothing out of the hypothetical ‘moat’’’ about a subsiding island is little aided by the mud from the reef. “The coarser detritus’? washed in from the reef flat often does “‘form a well-defined terrace slowly tJ. S. Gardiner, “‘The Origin of Coral Reefs... . ,” Amer. Jour. Sct., XVI (1903), 203-13; see p. 2I0. CORAL REEFS AND SUBMARINE BANKS 299 growing inward from the reef,” as may be seen in many lagoons, for example, those of Tahiti and Raiatea in the Society Islands. Furthermore, as a large quantity of muddy sediment is formed on the reef flat by disintegrating agencies, organic and inorganic, acting on the blocks and scraps of coral rock washed in from the exterior reef face, and as the interior terrace of white granular detritus is free from fine silt, it follows that an important share of fine detritus from the reef must reach the lagoon floor; and this reef silt, as well as the fine organic sediments formed on the lagoon floor or in the lagoon waters, is distributed by the waves and currents therein generated, as Gardiner states. Darwin had earlier reached a fair understanding of this problem: “The greater part of the bottom in most lagoons is formed of sedi- ments; large spaces have exactly the same depth, or the depth varies so insensibly that it is evident that no other means, excepting aqueous deposition, could have levelled the surface so equally” (26). In my own limited experience I saw the waters of two barrier-reef lagoons rolling in heavy waves under strong winds— once in Fiji, once in the broad lagoon of the Great Barrier reef off the Queensland coast—and on both occasions the water was gray with suspended sediments. It seems evident that under such conditions the finer sediments of lagoon floors will be lifted chiefly from the shallower parts and will settle in about the same amount everywhere; and the continuation of such changes will tend to produce and maintain a fairly smooth surface of sedimentation, such as actually exists. Another view has been announced. After assigning a small value to the general distribution of lagoon sediments by the waves of gales and storms, and a large value to the distribution of sedi- ments by currents driven under steady winds, Daly has recently reached the conclusion that lagoon floors thus aggraded during subsidence should not be level, but should be deeper to windward and shallower to leeward; then, on examining charts of atolls and verifying the general levelness of lagoon floors, he concludes that subsidence has not taken place, but that the reefs have grown up and the lagoons have been evenly aggraded in postglacial time because their sediments have been deposited on smoothly abraded 300 W. M. DAVIS still-standing platforms.t The conclusion does not seem tenable because the smooth floors of a good number of medium-sized atoll lagoons in the trade-wind belts have moderate depths, such as 20 or 25 fathoms, and must therefore, even under the glacial- 3 control theory, have been aggraded by 15 or 20 fathoms if their abraded platform lies at a depth of 40 fathoms. Moreover, the fact, more fully stated in the next section, that small atolls have on the average shallower lagoons than large atolls proves that a good share of their sediments is washed in from their inclosing reefs. If aggradation by material thus partly supplied from the margin, partly from locally formed organic detritus, and distributed by trade-wind currents in the lagoons ought to produce a slanting surface of deposition, these lagoon floors should not be level; their levelness therefore contradicts the supposed necessity of slanting aggradation and confirms the theory of equable distribution and even aggradation by the lagoon waters. Hence it may be concluded that lagoon floors tend to become and to remain nearly level, what- ever form the foundation of their inclosing reefs may have had, and whatever the thickness of their sediments may be; and, conversely, that the form of a lagoon floor gives no indication of the form of its buried foundation. I am therefore constrained to think that the general levelness of atoll-lagoon floors is no sufficient reason for the existence of a level rock platform at a moderate depth beneath. The depth of lagoon floors —li, then, the smoothness of lagoon floors is no sufficient proof of the existence of a smoothly abraded rock floor beneath them, we may next inquire as to evidence for the existence of such a rock floor that is found in the similar depth of atoll lagoons and of submarine banks. The depths are not all alike. As to atolls, Daly has shown that on the average the smaller ones are the shallower, and from this he draws the accept- able conclusion that “the smaller the platform the higher was the proportion of reef débris in the veneer, and the more rapidly has the lagoon area been shallowed”’ (183), and again that “‘the filling of the lagoon [by inwashed sediments] is in indirect proportion to tR. A. Daly, ““A New Test of the Subsidence Theory of Coral Reefs,” Proc. Nat. Acad. Sci., II (1916), 664-70. CORAL REEFS AND SUBMARINE BANKS 301 the width of the platform [atoll]”’ (192);' but as this conclusion would follow equally well whether the reefs have grown up with the rising ocean around a still standing abraded platform, or in a stationary ocean around a sinking and submountainous foundation, no ground for choice between the two theories under discussion is here provided. A closer scrutiny of the figures, however, reveals considerable differences of lagoon depth in atolls of about the same size, and this seems more consistent with the variable conditions offered by the theory of intermittent subsidence than with the strictly uniform conditions assumed under the theory of glacial control. For example, among 12 atolls listed in Daly’s table as from 21 to 30 kilometers in diameter, the smallest value of maximum lagoon depth is little more than half the value of the largest maximum; and this smallest maximum is in an atoll the diameter of which is greater than the one which has the largest maximum. Ringgold atoll, in Fiji, is given as having a maximum lagoon depth that is more than twice as great as the maximum depth in North Argo atoll, of the same group, though both have the same moderate diameter of 10 miles. Among wide barrier reefs the one adjoining New Caledonia on the northeast has a maximum depth less than half that of the lagoon of similar breadth on the northwest of Viti Levu, in Fiji. Moreover, some lagoons reach unusual depths. In the large lagoon, just mentioned, northwest of Viti Levu, the largest island of the Fiji group, the lagoon deepens and the inclosing reef is submerged, as the distance from the island increases, in such a way as strongly to suggest recent tilting; a sounding of 59 fathoms is the maximum there recorded, but the outermost part of the lagoon has not been measured. The large lagoon of the Exploring Isles in the eastern part of Fiji deepens eastward to 80 or go fathoms, and this again suggests tilting, as Agassiz noted. Cases of this kind are as significant as they are exceptional. But it should be noted that many atolls of the Central Pacific are, according to the latest charts of all sources available in the Hydrographic Office at Washington, incompletely surveyed; further exploration is needed t The misprint “‘direct”’ in the original article is here changed to “‘indirect”’ with the author’s approval. 302 W. M. DAVIS before their testimony can be used. In any case, even in atolls of similar diameter, the depths of the lagoons vary so much that they alone cannot be taken as proving that the atol! reefs have been built up with the rise of the ocean from stationary platforms of uniform depth. The fact that the maximum depth of atoll lagoons seldom exceeds 40 fathoms does, however, suggest the existence of some control that has prevented the occurrence of greater depths; but this control may be found elsewhere than in the abrasion of plat- forms at a uniform depth below sea-level. For example, if the subsidence of an atoll or a barrier-reef island is relatively rapid, the reef will be somewhat submerged, the inwash of detritus from the reef will be active, and the increase of lagoon depth will be retarded; if subsidence is, on the other hand, relatively slow, the reef will be maintained at sea-level and will broaden its surface, sand islands will be formed along its edge, and inwash of detritus into the lagoon will practically cease; hence, in spite of slow subsidence the lagoon will not be rapidly shoaled. Thus, unless sub- sidence be unusually rapid, there appears to be a series of spontane- ous reactions which tend to prevent lagoon depths from varying by large measures. Wherever unusually rapid subsidence occurs, the atoll would be drowned and converted into a submarine bank. The scarcity of such banks in the Pacific, as far as it is now explored, suggests very strongly that subsidence has rarely been unusually rapid; and slow, equable, or intermittent subsidence, being a near approach to long-continued stability, does not appear to be particularly incredible. But, however that may be, the known depths of atoll lagoons can be explained as well by the theory of intermittent subsidence as by the glacial-control theory. The depth of submarine banks will be discussed in a later section. The volume of existing reefs—The glacial-control theory sug- gests that a very uniform upgrowth of reefs should take place in the uniformly rising postglacial ocean, and that all existing reefs should therefore be of similar surface breadth and of similar volume. Daly finds this to be the case; he says that “‘the widths as well as the heights of the existing barrier and atoll reefs are of the proper CORAL REEFS AND SUBMARINE BANKS 303 size, if these calcareous rims originated on the platforms in post- glacial time” (219). As to the height of reefs we know little, because the depth of their base is undetermined. Nothing is gained by assuming their base to be 40 or 50 fathoms below present sea-level, for the resulting uniformity of height has no verification. However that may be, the visible differences in the height of reefs above their lagoon floors and the breadth of reefs at sea-level seem to me too great to support the above conclusion. Yap, in the west- ern part of the Caroline group, Western Pacific, and. Rodriguez, in the Southern Indian Ocean, have broad reef plains, a mile or more in width, attached to the central islands; that of Rodriguez is 4 miles wide on the southwest side of the island. Borabora, in the Society Islands, has a barrier reef and reef flat half a mile or a mile wide, and a comparatively shallow lagoon; Mbengha, in Fiji, has a reef and reef flat of similar width around the southern side of its lagoon; the reef around Nairai, in central Fiji, is a half or a quarter of a mile wide; Budd reef, in northeastern Fiji, is narrow, generally less than a quarter-mile across, and its lagoon is 46 fathoms deep; yet near by is the long and irregular Ringgold atoll, in which the reef is half a mile or a mile wide, though the lagoon is of similar depth to that of Budd reef; Tahiti has a narrower reef, often discontinuous; Fauro, in the Solomon Islands, is fringed with a narrow sea-level reef and surrounded by a submerged bank, 70 fathoms deep in places, on which a very imperfect reef rim is found; Palawan, the southernmost member of the Philippines, is elab- orately embayed along its western coast, where the headlands are neither clift nor fringed with sea-level reefs, but are fronted by a broad submerged platform, varying in depth along its length with maximum of 60 fathoms. The Marquesas Islands, nearer the equator than the numerous Paumotu atolls, have no sea-level reeis; their headlands are strongly clift, and a submerged bank extends around them. Various submarine banks have well-defined reef rims that fail to reach the surface, or that rise very discontinuously to the surface, as will be specified below; and some submarine banks are flat and rimless. Differences of these kinds in reef volume are more consistent with the unlike conditions introduced by intermittent subsidence, varying as to rate, amount, place, and 304 W. M. DAVIS date, than with the uniform conditions demanded by the glacial- control theory. The exterior profile of coral reefs.—There is another feature of coral reefs to which Vaughan has called attention as indicating the existence of a submerged platform previous to, and independent of, the formation of the present reefs; this is the continuation of the shallow lagoon floor, not only where it is inclosed by a reef, but also through uninclosed sectors of its area where a marginal reef is wanting. A good number of examples of this kind are known in Fiji; but it is significant that the breach in the reef is in practically all cases on the leeward side. It is further significant that the uninclosed sector, which may like the rest of the lagoon floor have a general depth of 20 or 25 fathoms, slopes gradually to a depth of about 40 fathoms at its free margin and then pitches down with a steep descent to deep water; and it is still further significant that this change of declivity occurs at essentially the same depth as that at which the gentle exterior slope of the reef itself changes to a steep pitch. Instead of regarding these features as indicating the existence of a submerged platform of earlier and independent origin on which the reef was afterward built up, I am inclined to interpret both of them as resulting from aggradation by the transporting agencies of the ocean waters with respect to present ocean-level; in short, as small insular “shelves”? of rapid development with respect to the present level of the ocean and therefore as correspond- ing to great continental shelves of long-continued development with respect to the average relation of land and sea through modern geological periods. The change of slope outside of a continuous reef will be first considered. Darwin was, I believe, the first to suggest this view as to the origin of the exterior profile of coral reefs in his account of Keeling atoll. ‘As the external slope of the reef is the same round the whole of the atoll and round many other atolls, the angle of inclination must result from an adaptation between the growing powers of the coral and the force of the breakers and their action on the loose sediments”? (74); and he later added, ‘Considering the manner in which the beds of clean coral .... graduated into a sandy slope, it appears very probable that the depth at which reef- CORAL REEFS AND SUBMARINE BANKS 305 building polypifers can exist is partly determined by the ex- tent of inclined surface which the currents of the sea and the recoiling waves have the power to keep free from sedi- ments” (84), thus foreshadowing a view that is now generally accepted. A good number of other observers have interpreted the gradual slope and the steep pitch outside of a reef in the same way. Thus we read in the “Challenger’’ report regarding the reef at Tahiti, ‘““The whole of the space from the edge of the reef to a depth of 35 fathoms was covered with a most luxuriant growth of corals, with the exception of one or two small spaces where there was white coral sand”; the steeper pitch to greater depths was covered by coral blocks, “‘ which have been torn away from the ledge between the edge of the reef and 35 fathoms during storms, or by overhanging masses which have fallen by their own weight. In this way a talus has been formed on which the corals living down to 35 fathoms have found a foundation on which to build further seawards, for this [upper] slope is the great growing surface of the reef.”* It may be noted that the depth here given for abundant living corals is unusually great, and that Agassiz nearly thirty years later found a smaller proportion of growing corals and a larger proportion of dead corals, coral fragments, and coral sand on the same slope; hence the population of the slope presumably varies in relation to the master-storms of decades and centuries. But the important matter to note here in the present connection is that the outer slope of the reef, like the reef itself and the prograded ‘“‘belt of low land at the foot of the mountains”’ of Tahiti, represents an adjust- ment of aggrading processes and aggraded forms with respect to present sea-level, by whatever changes the present relation of land and sea-level have been brought about. Gardiner, whose studies of reef slopes are both intensive and extensive, says of the Fiji reefs, ‘“The section outside all is nearly the same, a gentle slope to about 40 fathoms and then a sudden steep.’’ t Narrative of the Cruise of H.M.S. ‘Challenger,’ I, Part 2 (London, 1888), pp. 779, 781. 2**The Coral Reefs of Funafuti, Rotuma, and Fiji... . ,” Proc. Cambr. Phil. Soc., IX (1898), 417-503; see p. 445. 306 W. M. DAVIS The same experienced observer later made a more specific state-_ ment regarding the Maldives: From the outer edge of an encircling reef flat there is generally to seaward a gradual slope to 30-50 fathoms in 200-400 yards, succeeded by the steep. . ... This slope is essentially the growing area, being covered almost com- pletely by living organisms. ... . The outwash of detritus, largely due to undercurrents [or to general agitation by wave and current action, with the result that the finer sediments are chased about until they finally settle in deep water outside the reef where they will not be again disturbed ?], causes a raining down of coral masses and sand over the edge of the steep, carrying it out and allowing the extension of the whole outer side as a fairy ring.t The growth of a reef outward, like a fairy ring, is again mentioned by the same author in his elaborate report on the Maldives. A later statement by the same author is, “The steep . . . . is built up by masses of coral rock from the reef above, its angle represent- ing that at which such material comes to rest in sea-water.”3 It is interesting to add that the change of slope on the exterior of a reef occurs at the same depth as that at which, according to Daly, ‘‘the charts of the world show the break of slope”’ (199) on continental shelves. Lagoon floors of discontinuous reefs.—As to the free border of an uninclosed lagoon-floor sector, where the inclosing reef is wanting as above mentioned: It is easy to conceive that the lagoon floor there represents a more or less aggraded portion of a pre-existent platform, elsewhere inclosed by a superposed reef; and it is not difficult to explain the origin of the platform by abrasion according to the glacial-control theory as stated by Daly, or to leave it un- specified, except to say with Vaughan that it is independent of the reefs which are growing upon it. But it is also easy to conceive that the lagoon floor of today is the more or less aggraded lagoon floor of earlier days; that the lagoon floor of earlier days may have been deepened during times of rapid subsidence which caused the t “The Origin of Coral Reefs .... ,” Amer. Jour. Sci., XVI (1903), 203-13; See py 211. 2The Fauna and Geography of the Maldive and Laccadive Archipelagoes (Cam- bridge), I (1903); IL (1906); see I, 175, 182, 183, 317. 3 “The Indian Ocean,”’ Geogr. Jour., XXVIII (1906), 313-32, 454-55; see p. 455. Also ‘‘Submarine Slopes,” ibid., XLV (1915), 202-10. CORAL REEFS AND SUBMARINE BANKS 307 upgrowth of a narrow reef, or shoaled during stationary periods and periods of slow subsidence when the reef was broadened; and that the shoaling may have gone so far as to convert a narrow young reef and a deep lagoon into a mature reef plain,‘ if subsidence were long enough in abeyance. It is furthermore easily conceivable that prevalent subsidence may have been for a time neutralized in a phase of no submergence, while the ocean-level was falling as a glacial epoch came on, and then accelerated for a time in a phase of rapid submergence while the ocean-level was rising as a glacial epoch passed off; that a mature reef plain, formed during the neutralized phase of no submergence as a glacial epoch came on, might not be completely rimmed around by a new reef which grew up during the following phase of accelerated submergence as the glacial epoch passed off; and that in such case a failure of growth might reasonably enough take place on the border of the leeward sector of the reef plain, where the quantity of fine sediment shifted about by the waves might very likely prevent or retard coral growth. All of these mental schemes—mere figments of the imagination— are, as Is said above, easily conceived by anyone who cares deliber- ately to consider the coral-reef problem; the difficulty in the prob- lem lies elsewhere, namely, in the discovery of tests by which one of the schemes may be shown to represent better than any other the processes of the invisible past and thus to offer the best expla- nation for the invisible as well as the visible structures of the present. The best explanation that I have been able to reach is as follows. First, regarding the exterior profile of continuous reefs: In view of the evidence already given, from which it appears that the pro- duction of platforms by abrasion during the glacial period is improb- able if not impossible, and in view of the fair accordance between the depths at the outer margin of the uninclosed sector of a lagoon floor and the depth on the outer face of a reef where the change takes place from a gentle slope to a steep pitch, I am persuaded that the change of slope at 40 fathoms is not an inheritance from a time when that part of the reef lay at or close to the surface of the ocean, but, as already stated, a consequence of adjustment between t “The Great Barrier Reef of Australia,” Amer. Jour. Sci., XLIV (1917), 339-50; see p. 3406. 308 W. M. DAVIS the detritus to be transported and the agents of transportation with respect to present sea-level. Like the reef itself, the two elements of its exterior profile—namely, the gentler slope down to 40 fathoms and the steeper pitch below—have been brought by organic growth and by inorganic processes into a normal relation to sea-level. Secondly, regarding the free border of an uninclosed lagoon-floor sector of atolls and barrier reefs: In view of the alternate retarda- tion and acceleration of submergence. by the combination of prevalent subsidence with the periodic changes of ocean-level during the glacial period—for this is, in my mind, the chief value of the glacial-control theory—I am strongly inclined to regard any “platforms” that may exist, now more or less aggraded, beneath incompletely or completely inclosed lagoon floors as nothing more than the surface of earlier reefs normally broadened while sub- mergence was so slow that narrow reefs were transformed into mature reef plains. ‘Thus interpreted the present reefs are merely new, still young, and relatively narrow growths above their mature predecessors; narrow, because they have been developed while submergence has been accelerated. This inclination of opinion has been strengthened by an examination of charts of submarine banks both within and without the coral seas, and that aspect of the problem must next be examined. Let it be noted, however, that if the explanation above suggested for the change in the exterior slope of a reef at 40 fathoms depth be correct, then Daly’s conclusion that ‘‘the present atoll, barrier, and fringing reefs . . . . have been developed nearly or quite in the same interval of time” cannot be supported by the agreement of “their sectional areas and their volumes, as measured, in each case, above the break of slope at the platform on which the crown- ing reef stands” (233). A further conclusion is also vitiated, namely, that inasmuch as “the surface outcrops and volumes of the greater barrier and atoll reefs, measured from the levels of the lagoon floors, are respectively nearly equivalent in the Pacific and Indian oceans,” therefore “the earth’s crust must have sunk at a nearly uniform rate throughout the enormous area described . if these reefs were formed by subsidence” (233, 234). CORAL REEFS AND SUBMARINE BANKS 300 Neither the break of the exterior slope nor the level of the lagoon floor of existing reefs is, as far as I can make out, the record of a former lower sea-level: both appear to have been brought into relation with present sea-level by processes now acting. It cer- tainly seems as reasonable to explain the exterior profile of coral reefs in this way as to explain “the break of slope on the [conti- nental| shelves . . . . near the 40-fathom line”’ as a result of wave and current action at the present stand of the ocean surface, and therefore since the present stand was assumed. [To be continued | HE CSLO CARBONIFEROUS GLACIAL DEPOSITS OF SOUTH AMERICA A. P. COLEMAN University of Toronto INTRODUCTION The proofs of tremendous glaciation at the end of the Carbon- iferous have been given in great detail from South Africa and Aus- tralia and less fully from India; but comparatively little has been reported regarding Permo-Carboniferous glaciation in South America. Having seen something of the tillites and glaciated rock surfaces of typical localities in the first three regions, it seemed desirable to visit the less-known glacial deposits of South America. During the past summer this has been accomplished, and it is proposed to give in this paper a brief account of what was observed. The probability that certain Carboniferous or Permian bowlder conglomerates with a matrix of shale in southern Brazil were of glacial origin was recognized by Orville Derby as early as 1888;7 but no striated stones were found, and the evidence seemed scarcely suf- ficient to establish the point. In 1907 and 1908 I. C. White and David White made it almost certain that the widespread bowlder conglomerates were glacial, the latter showing that the accompany- ing flora, as collected during I. C. White’s examination of the coal deposits of southern Brazil, was identical with floras in a similar relation in South Africa and India.? Still the final proof, the finding of striated stones, was lacking. In 1908 J. B. Woodworth studied the southern Brazilian tillites, finding striated stones plentiful and making their glacial origin absolutely certain. His report on the field work accomplished, with ‘Orville Derby, “‘Spuren einer carbonischen Eiszeit in Siid Amerika,’ Newes Jahrb. fiir Min., etc., Band II (1888), 175. 2, David White, ‘‘Permo-Carboniferous Climatic Changes in South America,” Jour. Geol., XV (1907), 615-33; and I. C. White, in his Relatorio Final on the Brazilian Coal Fields, 1908, pp. 11-110. 310 PERMO-CARBONIFEROUS GLACIAL DEPOSITS 2k the aid of two young Brazilian geologists, Eusebio Oliviera and Juvenal Pacheca, is admirable and has served as guide in a portion of my own work." In the following year H. Bross confirmed his conclusions, finding striated stones near Itararé.? Up to r911r Permo-Carboniferous tillite had been reported from southern Brazil only; but in that year T. G. Halle described tillite of the same age from the Falkland Islands, which may be looked on as belonging to the South American region. In rorz2 tillite with polished and striated stones, including bowlders up to 1.5 meters in diameter, was described by C. Guillemain, from Uruguay;! and in 1913 J. Keidel announced the finding of tillite of this age in southern and western Argentina in a paper read before the Twelfth Geological Congress in Toronto; while in 1916 he gave an excellent and full account of the tillite at Sierra de la Ventana in southern Argentina.® From this outline of the literature on the Permo-Carboniferous glaciation in South America it will be seen that our knowledge of the distribution of the tillite is rapidly growing, extending now to three countries, and that this glacial formation is widely spread in southern Brazil and perhaps also in Argentina. In Brazil the excellent field work of Oliviera and Pacheca is constantly extending the known area of glaciation, so that in this respect South America may soon rival South Africa and Australia and surpass India. It was planned to visit localities for tillite in several parts of Brazil and at Sierra de la Ventana in Argentina, and these plans were carried out successfully, largely through the help and advice of Pacheca in Sao Paulo and of Keidel at Buenos Aires. I am specially indebted to Dr. Pacheca, who has mapped the tillite in detail in the state of Sao Paulo, and who showed me a number of tJ. B. Woodworth, Bull. Mus. Comp. Zool., Harvard, LVI, No. 1, Geol. Series, X. Geol. Exped. to Brazil and Chile. 2 H. Bross, “Glacial Spuren in Parana,” Cent. Bl. fiir Min., etc., 1909, pp. 558-61. 3T. G. Halle, Geol. Mag., N.S., V, 264-65; also Bull. Geol. Inst., Univ. Uppsala, XI, 144-56. 4C. Guillemain, Neues Jahrb. fiir Min., etc., Beilk Band XXXIII, 208-64. 5 J. Keidel, Comptes Rendus, pp. 676-80; also in La Geologia de las Sierras de la Provincia de Buenos Ayres, Tomo XI, No. 3, Anales del Ministerio de Agricultura of the Argentine Republic, 1916. Bye A. P. COLEMAN typical outcrops in the field. Through his courtesy and kindness I had an admirable introduction to the Paleozoic bowlder clays of Brazil. f TILLITES IN THE STATE OF S O PAULO In tropical Brazil, with its moist, hot climate, weathering goes to great depths, and even in the southern parts of the country, where the climate is warm temperate, the products of decay mantle most of the surface; so that fresh outcrops of rock are seldom to be found under natural conditions. On this account the most satis- factory points for geological work are cuttings along the railways or depressions where roads ascend hills and the wheels of vehicles have worn their way downward. To an observer fresh from parts of North America where the Pleistocene glaciation has left clean surfaces of almost unchanged rock this is most disconcerting, and it takes a little time to adjust one’s self to the new conditions. The working out of field relations is greatly hampered by the products of weathering, which usually hide even the weathered rock and may — accumulate to considerable depths on slopes and in valleys. Under Dr. Pacheca’s guidance a number of railway and road sections were visited to the north of the city of Sao Paulo, the first region being near the thriving city of Campinas. In railway cuttings two or three kilometers from the city, tillite rests upon gneiss, probably of Archean age, two mounds of the gneiss rising with rounded forms suggesting roches moutonnées; but unfortu- nately weathering has gone so far at the contact as to destroy any smoothed or striated surface that may have existed in the beginning. According to Dr. Pacheca this is almost the only example known in Brazil of tillite resting on what must have been a scoured surface of solid rock. The tillite is weathered and scarcely harder than certain Pleistocene tills, though its yellow or red or chocolate-brown color is unlike the customary bluish gray of North American bowlder clay. Striated stones were not found here, though the general appearance of the rock was that of tillite, some of the more resistant stones inclosed‘in it, such as quartzite and granite, still showing subangular forms and smoothed surfaces. Overlying the tillite are shaly or sandy beds, distinctly stratified and including some- PERMO-CARBONIFEROUS GLACIAL DEPOSITS 313 times a thin layer of conglomerate, the series being closely related to the tillite and sometimes interstratified with it, as in many sections of Pleistocene glacial deposits. The next point visited was near Capivary, reached by a narrow-gauge railway from Campinas, where a delightful walk of 18 kilometers disclosed excellent sections of tillite containing typi- cally shaped and sometimes striated stones (Fig. 1) including Fic. 1.—From Capivary, Sao Paulo, Brazil bowlders of sandstone, quartzite, conglomerate, and granite occasionally reaching a diameter of one meter. The coarse con- glomerate forming some bowlders quite suggests the Huronian tillite of Cobalt, Ontario, and their source in Brazil is unknown. The tillite has a thickness of three or four meters and rests on soft, stratified sandstone or sandy shale sometimes cross-bedded. In one place it is covered by a sheet of trap mostly weathered into the bright-red ‘“‘terra roxa’’ which forms the best soil for coffee and sugar plantations. A visit was made to a section near Villa Raffard, about four kilometers southwest of Capivary, to see some large bowlders of 314 A. P. COLEMAN the ancient conglomerate, including pebbles of jasper, another feature suggesting the Huronian rocks of Ontario. These con- glomerate bowlders have also been mentioned and figured by Woodworth. In several places shales accompanying the tillite have been “thrown into folds a few meters in dimensions, either by ice-thrust at the time of glaciation or by later compression. Over the tillite in places a brown fine-grained unstratified material, pierced by worm or root holes, stands up as low cliffs and suggests an ancient loess. At one point on the railway near Capivary (kilometers 160-61) shales beneath the tillite appear to have been crumpled and then truncated, showing an unconformity between the tillite and the soft sediments beneath; and some soft sandstone bowl- ders in the tillite are quite like an irregularly bedded sand- stone frequently found in the same position, suggesting the same relationship. | The fresh tillite is often solid enough to make vertical faces in cuttings and shows spheroidal shapes when weathering has begun. Ultimately the rain breaks it down into slippery clay, gray or yellowish or red in color, strongly suggesting a rain-crumbled Pleistocene till. Near the small station Elias Fausto, Dr. Pacheca has found tillite inclosing very large granite bowlders, one partly disclosed measuring 3X3 X2 meters, and looking like one of the Iowan bowlders of the Western states. My last excursion under Dr. Pacheca’s guidance was to Limeira, 50 kilometers northwest of Campinas, where a round of 21 kilo- meters was made over country roads giving an opportunity to see typical bowlder clay extending over many square kilometers of gently rolling country. The tillite is usually chocolate colored and in places reaches a thickness of 25 meters, while in other places it has been cut through by the stream valleys. In the steep walls of a sunken road leading out of the town plenty of well-striated stones were found, and except for the prevailing red color and a little greater hardness the outcrops reproduce perfectly the features of a region of bowlder clay in North America. It was hard to believe that the rock was as old as the Permian. PERMO-CARBONIFEROUS GLACIAL DEPOSITS Bin5 Limeira is about in latitude 223°, a degree within the tropics, so that in Brazil, as in India and Australia, the ancient ice-sheet reached much nearer to the equator than any Pleistocene ice-sheet. It is estimated by Drs. Florence and Pacheca, of the Geological Survey of the State of Sao Paulo, that the outcrops of tillite extend from northeast to southwest for 500 kilometers, with a width of from 50 to roo kilometers; and it must be added that the tillite follows the gentle northwestward dip of the rocks of the region and probably extends far beneath the Triassic beds in that direction. It is evident that one is dealing with deposits formed by a great ice-sheet spread out over a peneplained surface and not with the results of mountain glaciers. TILLITE IN STATES SOUTHWEST OF SAO PAULO After the admirable introduction to the study of Brazilian glacial deposits provided by the kindness of Dr. Pacheca there was little difficulty in recognizing the characteristic appearance of the tillite, and on the journey by rail from Sao Paulo to Montevideo in Uruguay some of the localities described by Woodworth were visited, the first just beyond the southwestern boundary of the state of Sao Paulo between Itararé and Sengens. In railway cuts near Sengens, Woodworth had found striated stones and large bowlders of sandstone; and a walk along the railway between the two stations proved extremely interesting. Following the crooked narrow-gauge railway from Sengens northeast toward Itararé tillite is seen for eleven kilometers (from km. 241 to km. 230) resting usually on sandstone, occasionally with a hummocky surface and in one case with a suggestion of furrowing in a direction from southeast to northwest or vice versa. The sandstone is still soft, and when the tillite was deposited may have been softer, so that large blocks could easily be lifted and inclosed in the glacial mate- rials. In addition to these masses of local rock there are quite large bowlders of shale and of granite, and a multitude of smaller stones, many of a harder sandstone than the underlying rock, and a few of quartzite. The tillite varies in thickness, sometimes reaching ten meters. Parts of it near kilometer 241 have been more or less t See Woodworth’s Report, p. 62 and Pls. xxi and xxii. 316 A. P. COLEMAN pushed and crumpled, and not far to the northeast is the great fault and escarpment mentioned by Woodworth. The best display of tillite is about at kilometer 235, where the smaller stones are very frequently striated, more so than in any other till I have seen, whether Pleistocene or older. Many of the glaciated stones show not only “soles” but well-defined facets, as if they had been firmly held till a face was ground flat and then adjusted at another angle, resulting in another flat face. These facets sometimes come together sharply. In early days similar faceted stones from the Permo-Carboniferous tillite of India attracted attention. It would seem as if the Permo-Carboniferous ice-sheets held their imbedded stones more firmly than those of the Pleistocene. Why? Were their bases colder or was there a greater thickness of ice, giving a stronger pressure ? As may be seen from the train, tillite extends several kilometers on the route southwest; but the next stop was made at Ponta Grossa, midway across the state of Parana, where I. C. White had described outcrops of glacial conglomerate.t On the side of the ridge on which the town is built, cuttings, made for streets and for drainage, disclose reddish, sandy glacial deposits containing sub- angular stones of various kinds, a few of which were found to be striated. A fairly good section is seen also on a road leading into the country. Above the tillite there is a sheet of trap weathering into a very red soil, and beneath it sandstone followed by black shale from which Devonian fossils are reported. A visit was made also to Serinha, 70 or 80 kilometers to the southeast, where Woodworth suspected an older tillite. Typical bowlder clay is passed between Palmeira and Nova Restingua and may be seen at Porto Amazonas. There is a rapid descent from Palmeira to Serinha, which is in a deep river-valley at the base of sandstone cliffs. The tillite here takes the form of blue or yellow shale, readily weathering to clay, containing subangular stones, chiefly sandstone, quartzite, and granite. No striated stones were found, but the bed looks like a glacial deposit. It is overlain by 200 feet of firm sandstone resembling the rock found beneath the tillite at higher levels. Beyond this fact no clue to its age was 17. C. White, Relatorio Final on the Brazilian Coal Fields, 1908, p. 51. PERMO-CARBONIFEROUS GLACIAL DEPOSITS Sh i847) observed. The whole series of rocks, including the two tills, seems to lie nearly horizontal, doubtless with a gentle dip north- westward following the regular trend of the stratification in south- ern Brazil. The tillite at Serinha looks no older than that described before, and may represent merely a Carboniferous forerunner of the more important glaciation to follow. Southeast of Ponta Grossa the railway lies too far west to give opportunities of observing the glacial deposits, passing over trap- sheets, Triassic sandstones, etc.; but I. C. White’s account of the bowlder conglomerates associated with a low grade of coal and Permian plants in the state of Santa Catharina, e.g., at Orleans, shows that tillite continues to latitude 28°.1 His map of the Tubarao series, which includes the Orleans glacial conglomerate, extends the tillite to the southern end of Brazil, in Rio Grande do Sul, though his account does not specially mention bowlder con- glomerates as having been observed in that part of the country. Guillemain, by finding tillite with striated stones at Fraile Muerto in northern Uruguay, not far from the boundary of Brazil, as noted in the introduction, continues the region of glaciation still farther to the south. Including the 500 kilometers reported from Sao Paulo this gives a length of about 1,500 kilometers from north- east to southwest, running in latitude from 223° to about 323°. The tillite has not yet been found to outcrop continuously for this long distance, but the known localities are sufficiently numerous to make its continuity highly probable. Its known width is estimated at from 50 to too kilometers in Sao Paulo, but it is unknown how far it extends beneath the Triassic sediments and trap-sheets to the northwest. TILLITE IN SIERRA DE LA VENTANA There is a long gap between the Permo-Carboniferous deposits of Brazil and northern Uruguay and the nearest outcrops of tillite discovered in Argentina, which are in the Sierra de la Ventana not far from Bahia Blanca. Dr. J. Keidel, chief of the Geological Section of the Argentine Survey, was good enough to plan an excursion to this locality for me. A rail journey of 537 kilometers tT. C. White, Relatorio Final on the Brazilian Coal Fields, 1908, pp. 11-13 and 51. 318 A. P. COLEMAN _ southwest from Buenos Aires brings one to the small station among the hills, after passing a vast stretch of prairie-like pampas with few or no outcrops of rock. ‘The Sierra rises as rocky ridges with deep valleys between, one of them followed by the river Sauce Grande and others by its tributaries. The railway crosses the river just south of the station and follows up the valley of a small stream in the Arroyo Negro. The best exposures of tillite are found in the railway cuttings along the Arroyo within seven kilometers of Sierra de la Ventana, and these will be described first. The unweathered tillite is dark, bluish gray and entirely different in appearance from the usually red or brown and much-decayed tillite of Brazil. The rock is hard and shows some slaty cleavage, and the stones scattered through it are often a little squeezed or broken and slightly step-faulted. The weathered tillite is greenish or yellowish and crumbles somewhat readily, setting free the inclosed stones, but from the unweathered rock it is difficult to extract them unbroken. The fresh tillite is very like that from some outcrops of the Dwyka in South Africa, where the rock has under- gone squeezing and distortion in mountain-building operations; and it closely resembles the Huronian tillite of Cobalt and might easily be taken for it in hand specimens. The pebbles and bowlders inclosed include several species of rocks, granites and hard sandstones being commonest. They are seldom more than half a meter in diameter and have the character- istic shapes of glaciated stones. A considerable number have well- striated surfaces and are typical products of ice action. In some of the cuttings cross-bedded quartzite and more or less water-formed conglomerate occur also, apparently interbedded with the tillite; and in several places quartzite overlies the tillite con- formably. The base of the tillite was not seen in the railway cuttings, and a search was made for it to the north, where a small stream flows toward the Sauce Grande, but in vain. On this stream the tillite has been squeezed into schist conglomerate with a marked cleavage, reminding one of the Temiscaming and Doré conglom- erates of Ontario. A search still farther north showed no solid rock for several kilometers until the base of the northern range PERMO-CARBONIFEROUS GLACIAL DEPOSITS 319 of hills was reached, where quartzite, mica schist, and slate were encountered. Sections were examined a few kilometers up the river from the station and several fresh-looking outcrops of tillite were found at the water’s edge. Ascending the slopes from such outcrops one finds weathered tillite for a few hundred yards, then a cliff of tillite, followed by a covered belt where only quartzite pebbles can be seen for a height of about 15 or 20 meters. A second cliff of tillite reaches 85 meters above the river and is followed by quartzite to the top of the ridge. The lower bed of quartzite seems to be inter- glacial, corresponding to the band of quartzite and water-formed conglomerate seen in the railway cuttings. A section a kilometer or two down the Sauce Grande shows no base to the tillite, which has a thickness of 90 meters, as determined by aneroid, and is covered by quartzite including a band of tillite. None of the sections was entirely satisfactory, since on the gentler slopes the sclid rock is more or less hidden; but the thickness of the glacial beds seems to be not less than 60 meters and may be much more than that. An excellent account of the glacial deposits of Sierra- de la Ventana is given by Keidel in La Geologia de las Sierras de la Provincia de Buenos Ayres (1916), as mentioned in the introduction to this paper; and the statement is made that the origin of a number of the inclosed bowlders is unknown. Keidel puts stress on the resemblance of these deposits to the Dwyka, but gives no proofs of their age except that they are later than the Devonian, as shown by the inclusion of pebbles of limestone with Devonian fossils. The hard and somewhat metamorphosed character of the rock, which seems to suggest a greater age, is to be accounted for by the action of orogenic forces. One of Keidel’s plates represents the tillite as somewhat folded in a way that would add to the apparent thickness of the bed, but in my own field work no clear evidence of folding was seen, though compressive action was evident. TILLITE NEAR SAN JUAN IN WESTERN ARGENTINA Following a plan suggested by Dr. Keidel an excursion was made to exposures of tillite in western Argentina somewhat south of 320 A. P. COLEMAN San Juan. The nearest point to the outcrops on the railway between San Juan and Mendoza is at Paradero, kilometer 489. The railway traverses a desert country covered with sand and stones with isolated hills of rock not far to the east and the loftier Chico de Zonda, a range of foothills of the Andes, about eight kilometers to the west, as shown on Stappenbeck’s geological map of the region. Walking westward over the desert from the railway there is a gentle rise for two or three kilometers, followed by low ridges between profound ravines, apparently cut by temporary streams due to cloud-bursts in the mountains. At about five kilometers west there are steeply tilted red shales dipping westward, followed by hills of a green, basic eruptive, greatly weathered, and then high cliffs of gray Immestone. In the latter rock, fragments of crinoids and a syphon of orthoceras were found. It is indicated on the map as Silurian. A little to the south of this section, where a narrow valley pene- trates rugged hills, a greenish-gray shaly or slaty rock occurs, crumbling to fine débris on the surface, and including one or two bands of dark-brown pebbles and larger stones (Fig. 2). Most of the stones are fairly well rounded, as if rolled on a beach or in a river, and many have been ‘broken and recemented. Frequently they have been broken again where they lie on the surface, probably by alternations of heat and cold. A number of these stones are striated, often on more than one face. The largest seen was half a meter or somewhat less in diameter and was strongly scored. The stones are mainly basic eruptives, quartzite or limestone, the last too much attacked to show marks of glaciation. These stones appear to have been imbedded in the weathered, shaly rock, and in a ravine near by a few isolated ones are found still inclosed. The series appears to be tilted, but the dip and the limits of the bowlder bed could not be sharply determined, and in places two bowlder beds occur separated by a few meters of shale. These outcrops of loose, striated stones were followed for nearly a kilometer in a southerly direction, running parallel to the strike of the rocks in the foothills. Somewhat to the southwest, where the narrow valley is steep- walled and approaches the cliffs, a side ravine disclosed an abso- PERMO-CARBONIFEROUS GLACIAL DEPOSITS 22m lutely different section, in which a bowlder conglomerate rudely stratified in parts rises as a ridge about 30 meters high. This is of a kamelike character and includes sand, gravel, and stones of all sizes up to a meter in diameter. They are often rounded, but may be of various shapes and consist of many kinds of rocks—granite, FIG. 2. From tillite south of San Juan, Argentina gneiss, quartzite, vein quartz, sandstone, and limestone having been observed. Striated stones seem rare, only one poorly marked one having been found. It may be remarked, however, that in Pleistocene kames also it is unusual to find distinctly striated stones. Beneath the kamelike bed there are two or three meters of sandstone, and across a wide valley to the south a cliff shows six or seven meters of the conglomerate underlying, apparently 322 A. P. COLEMAN conformably, a hundred meters or more of sandstone with a west- ward dip. The two types of deposit just described are as different as pos- sible, though both seem to be glacial, but I was unable to determine how they are related to one another, since there has been folding, faulting, and squeezing during the formation of the mountain range, rendering the relationships complicated. Before leaving Buenos Aires, Dr. Keidel had referred to two tillites, a lower and an upper, corresponding probably to the two deposits just described. He mentioned also that near the lower tillite Talchir plants had been found, and some distance farther south Kharbari plants, giving-a clue to the age of the deposits. He has also found tillite to the north of San Juan, reaching in one place latitude 28°, and has discovered a striated surface of Devonian limestone beneath the tillite. Specimens of the tillite and of the glaciated surface are to be seen in the Museum of the Survey on Calle Maypu in Buenos Aires. His account of the very interesting glacial deposits in the western foothills of the Andes must be awaited for details as to their general features and relationships, but I am able to confirm his statements as to the glacial character of the beds so far as seen by myself. CONCLUSIONS From the descriptions given it will be seen that there are three widely separated regions of known Permo-Carboniferous glaciation in South America, the deposits differing much in appearance and lithological character, but all showing plainly the effects of ice action. The Brazilian tillites are the most widely distributed and the least changed. ‘They occur along the dissected edge of a table- land rising several hundred meters above sea-level and dip gently inland beneath sandstones and trap-sheets of the early Mesozoic. One or two diamond-drill cores prove that the tillite extends for 50 kilometers or more beneath the Triassic beds, but how much farther they go in that direction is unknown. There can be no doubt that they once reached farther seaward, so that the original ice-covered area must have been much greater than the present known area of tillite. As marine fossils have been found by Oliviera interbedded with the tillite on the Rio Negro in the state PERMO-CARBONIFEROUS GLACIAL DEPOSITS ROR of Parana,’ we may conclude that the region was at that time not a tableland but a comparatively low plain. In any case the whole character of the widespread, almost flat sheet of tillite in southern Brazil is such as must have resulted from ice action of the conti- nental type. Mountain glaciers could not have provided so extensive, uniform, and relatively horizontal a deposit as the Brazilian geologists have found. From what center the ice spread out is not known, though the numerous bowlders of granite and gneiss suggest a motion inland from the belt of Archean along the Atlantic coast. In that case the ice-sheet must have extended far toward the southeast, perhaps beyond the present’edge of the continent. However, there are granites and gneisses farther west, and outcrops of these rocks which existed in Carboniferous times may lie buried under later deposits toward the west or north. The bowlders of ancient conglomerate containing jasper may some day be traced to their source, giving evidence of the direction in which the ice moved. As to the tillites of the Rio Sauce Grande and of the belt along the foothills of the Andes near San Juan, the areas known to be covered by them are so small that local mountain glaciation might perhaps account for them; though the fact that tillite of the same age occurs on the Falkland Islands and that a great ice-sheet covering many thousands of square miles reached sea-level a few hundred miles to the north or east suggests that a very large part of South America must have been ice covered. It is not unlikely that the areas of ice action coalesced to form a single great sheet T,300 miles or more in diameter and covering hundreds of thousands of square miles, something comparable to the vast continental ice- sheets of Europe or North America in the Pleistocene. The northern edge of this ice-sheet reached at least one degree into the tropics in Brazil; and this occurred, not on high mountains, but on comparatively low ground, as shown on a former page. Recent advances in the study of the South American Permo- Carboniferous glacial deposits bring that continent into the same rank as South Africa and Australia with respect to the area then covered by ice, while India has been much surpassed. The mag- nitude of the geological problem involved is growing from year to * Woodworth’s Report, p. 20. 324 A. P. COLEMAN year, and the difficulty of accounting for such tremendous climatic changes is by no means lessening. The fact that the most extensive ice-sheets were in the Southern Hemisphere and that India only in the Northern Hemisphere shows important glaciation at the end of the Carboniferous forms one of the puzzling features of the problem. The idea that a change in the position of the poles could account for Permo-Carboniferous ice-sheets has been completely set aside by the discoveries in South America, since with the South Pole planted in the middle of the Indian Ocean southern Brazil would have been within the tropics. The theory of glaciation due to elevation is disproved also by the evidence from Australia and South America, showing that the _ice-sheets reached sea-level; and in any case it is inconceivable that such vast areas could all be elevated the necessary thousands of feet at the same time. Even if they were sufficiently elevated to give the required temperature, a large enough supply of moisture could hardly be arranged for on the greatly enlarged continents which this implies. The high tableland of the Andes is arid or semi- arid at present, and even the loftier peaks usually show little snow and few and small glaciers. It is evident that elevation alone will not account for the millions of square miles of mevé and ice-fields which must have covered much of India, Africa, Australia, and South America. The most satisfactory theory is that of refrigeration due to changes in the earth’s atmosphere; but even this fails to explain why Europe, Northern Asia, and North America should have been so little affected when great regions in other parts of the world were powerfully glaciated. One would expect that the change of climate affecting the tropics in India, Australia, and South America, and probably also in Africa would have been felt everywhere. It may be, however, that while the refrigeration was universal the supply of moisture necessary to form glaciers was lacking in Northern Asia, Europe, and North America. They may have had no ice- sheets for the same reason that Siberia was left uncovered with ice in the Pleistocene Ice Age; because the position of the open seas and the direction of the atmospheric circulation made them rela- ‘ tively dry regions with little snowfall. THE SUBPROVINCIAL LIMITATIONS OF PRE-CAMBRIAN NOMENCLATURE IN THE ST. LAWRENCE BASIN M. E. WILSON: Canadian Geological Survey, Ottawa, Canada INTRODUCTION The detailed geological work carried on in recent years through- out the southern part of the Canadian pre-Cambrian shield has shown that the geological succession in the ancient terranes of this territory is regionally less uniform and includes a greater number of rock series than was formerly supposed. Moreover, it has become evident that the widespread correlations implied by the use of the same nomenclature nearly everywhere throughout this great pre- Cambrian province assumes much more with regard to the regional succession in these ancient rocks than is actually known. Although it is not possible generally to demonstrate with mathematical conclusiveness that geological formations occurring in different localities are equivalent, nevertheless the premature use of the same name for formations, the correlation of which is open to question, or the continued use of the same name for forma- tions after it has become evident that their correlation is in doubt, is misleading, and is an obstacle rather than an aid in geologi- cal investigation. Hypothetical correlations of groups of rocks occurring in widely separated districts may serve for comparison or as a stimulus to investigation, but all the advantages of such tentative correlations may be attained by using a general ter- minology (Proterozoic, Archaeozoic, etc.) and thereby avoiding the definite correlations implied in the use of names of local origin. In the pre-Cambrian province which occupies the northern part of the St. Lawrence River basin there are four geographically and geologically separate subprovinces: (1) the region northwest of ‘Published with the permission of the Director of the Geological Survey of Canada. 326 M. E. WILSON Lake Superior, (2) the region south of Lake Superior, (3) the region extending northeastward from Lake Superior and Lake Huron to Lake Timiskaming and Lake Mistassini, and (4) eastern Ontario and the lower St. Lawrence, with which might be included the Adirondack region. With the possible exception of some of the late pre-Cambrian series occuring in the Lake Superior and the Timis- kaming subprovinces, the evidence upon which the rocks of these separate regions can be correlated is exceedingly meager, and for the present, at least, the only logical course would seem to be to build up a separate nomenclature in these various subprovinces by using those names already defined in these localities, supplemented by such local new names as become necessary from time to time as geological investigation is continued. OBJECTIONS TO AN INTER-SUBPROVINCIAL NOMENCLATURE The widespread correlations implied in the use of a common nomenclature throughout all the pre-Cambrian subprovinces of the St. Lawrence basin has been based on the assumption that the succession of formations within the various subprovinces has been worked out to practical completeness, and on the application of certain principles by which the correlation of the various formations in these widely separated areas is presumed to be established. The purpose of the following discussion is to point out that the assump- tion that our knowledge of the succession of formations in any of the subprovinces is complete is open to question and that the principles by which pre-Cambrian rocks are generally correlated are in part inapplicable and as a whole quite inadequate for the establishment of a pre-Cambrian nomenclature embracing all the territory in the St. Lawrence basin in which pre-Cambrian rocks occur. OUR KNOWLEDGE OF THE SUCCESSION OF FORMATIONS IN THE SUBPROVINCES INCOMPLETE The numerous regional classifications of the pre-Cambrian rocks of the St. Lawrence basin which have appeared from time to time in recent years, and the use of such terms as Keewatin, Lauren- tian, and Huronian nearly everywhere throughout this great pre- LIMITATIONS OF PRE-CAMBRIAN NOMENCLATURE 327 Cambrian province and at points hundreds of miles from those in which these names were originally defined, would seem to imply that our knowledge of the succession of formations within this vast territory was much more complete than is actually the case. Only a very small part of the territory in the St. Lawrence basin in which pre-Cambrian rocks occur has been actually mapped in detail, and even in those localities which have been mapped in considerable detail and have been regarded in the past as type areas the suc- cession of formations formerly supposed to be present has in many cases been considerably modified by more recent investigation. THE PRINCIPLES OF, PRE-CAMBRIAN CORRELATION INAPPLICABLE OR INADEQUATE Continuity or approximate continuity of outcrop.—The principle of continuity, or approximate continuity, of outcrop is the most conclusive of all the means by which the relationship of rocks can be determined. But it is inapplicable to the correlation of the various rock series occurring in the different pre-Cambrian sub- provinces for the reason that these are geographically and in part geologically separate from one another. Between the Timiskaming and the Grenville subprovinces there intervenes an extended belt of banded gneisses; between the Timiskaming and the western subprovinces there are the little-known wooded pre-Cambrian highlands on the north and overlapping Paleozoic sediments on the south; and between the northwestern and the southwestern sub- provinces lie the waters of Lake Superior. If, therefore, a common nomenclature be employed for all the pre-Cambrian subprovinces of the St. Lawrence basin, this nomenclature must be based on other less conclusive principles of correlation. Lithological similarity —This criterion has been widely applied in the correlation of pre-Cambrian formations, although it is in reality of very limited application; for the pre-Cambrian rocks of 1 A. C. Lawson, Geol. Surv. Can., Mem. 28, 1912, and Mem. 40, 1913, p. 4. W. G. Miller and C. W. Knight, Aun. Rep. Ont. Bureau of Mines, XXII, Part 2, 1914. R. C. Allen and L. P. Barrett, Jour. Geol., XXIII (1915), 680. W. H. Collins, Geol. Surv. Can., Sum. Rept., 1916, p. 183. C. R. Leith and R. C. Allen, Jour. Geol., XXIII, 703. 328 M. E. WILSON the Canadian shield, both sedimentary and igneous, are for the most part common types which might be deposited or intruded or extruded in any epoch of earth history. It has been largely on the © basis of this principle that the name Keewatin, first applied by Lawson to the metamorphosed basal volcanic complex occurring in the region northwest of Lake Superior, was extended, first to the Timiskaming region and later to eastern Ontario, a district nearly 1,000 miles distant from the locality in which the term was originally defined; yet volcanic rocks of this character are among the most common in the earth’s crust. They are represented at some point in nearly all the pre-Cambrian series of the St. Lawrence basin; are likewise abundant in later formations throughout the world, as in Great Britain, where they occur at numerous horizons ranging in age from the early Palaeozoic to the Tertiary; and are in process of formation at one or more points on the earth’s surface today. Asa consequence of this unscientific method of correlation the name Keewatin, although presumed to represent a definite formation, in reality is now applied in the Canadian pre-Cambrian subprovinces to any highly metamorphosed volcanic rock without regard to age. Similar stratigraphical succession of beds.—The larger part of the pre-Cambrian surface rocks of the region under consideration are volcanic flows or clastic sediments, in which a regular sequence of strata is uncommon, and this criterion is therefore inapplicable except to the late pre-Cambrian rocks. It has been especially useful in the mapping of the Huronian series in the Timiskaming subprovince, the Lower Marquette in the region south of Lake Superior, and the Animikie sediments in the region northwest of Lake Superior. Similar serial succession.—The widespread correlation implied in the nomenclature applied to the pre-Cambrian rocks of the St. Lawrence basin has been based to a considerable extent on this principle, although the apparent similarity in the serial succession may very frequently be explained in other ways. The principal objection to the use of this criterion is that it neglects to consider the possibility of overlap. Sedimentary rock series are not gener- ally deposited continuously or uniformly over immense areas, and LIMITATIONS OF PRE-CAMBRIAN NOMENCLATURE 329 where they have been deposited they are very commonly swept away in part by later erosion, before succeeding series are laid down. Moreover, the pre-Cambrian surface rocks are to a large extent volcanic flows or land sediments, and on this account are much more discontinuous than sediments of marine origin. Mode of origin of formations.—This criterion is of limited appli- cation; for sediments originating in the same way may be deposited during different geological epochs, and likewise sediments originat- ing in different ways may be deposited contemporaneously in adjoining localities. It might be of value in the correlation of certain uncommon types, such as glacial deposits, which generally occur only at long intervals in geological time. Relationship to batholithic intrusions.—The relationship of the pre-Cambrian surface rocks to the great epochs of batholithic invasion is of great assistance in correlation and may possibly eventually prove to be the most important of all the criteria used in the classification of the pre-Cambrian rock series into major groups; for geological investigation throughout the world has shown that batholithic intrusions are an accompaniment of mountain- building movements in the earth’s crust, and are thus directly related to the great regional changes in rock structure, to regional metamorphism, and to the uplifts which give rise to the great erosion intervals which form the dividing lines between the great pre-Cambrian terranes. Some of the applications and limitations of batholithic invasion in rock correlation are included in the following: Batholithic massifs are co-extensive with the mountains they underlie, and since mountains are generally extensive and linear the massif should also be extensive and linear. The extent of the outcrop of the massif will depend, of course, on the extent to which unroofing has been carried. In the Rocky Mountains, for example, unroofing has apparently only begun; in the Coast Range batholith of the Pacific Coast, on the other hand, unroofing is almost com- plete; and in some of the pre-Cambrian batholiths of the Canadian shield not only is the unroofing largely completed, but the batholith has also been reduced to base level. If two batholithic massifs have been intruded in a given region the younger may displace the first. Hence the conspicuous 330 M. E. WILSON structural features of the region would be those of the younger massif, and all evidence of the former presence of the older batholith might be destroyed except for such remnants as happened to remain in association with the roof rocks in the geosynclinal belts. Mountain-building periods and hence also periods of batholithic intrusion occur at long intervals separated by erosion periods and the development of peneplains. ‘The rocks in the vicinity of an intrusive batholithic mass are generally highly folded and metamorphosed; hence, if in a given area in the vicinity of a batholith flat-lying rocks occur which have not been greatly metamorphosed, it may be inferred that they have not been intruded by the batholith. : Batholiths are composite, and their intrusions continue during long intervals of time so that their various parts are only approxi- mately of the same age. ; Batholithic rocks are lithologically so similar that it is generally impossible to distinguish between batholiths of different ages except by means of their relationships to other rocks of which the age is known. Recently A. C. Lawson has contributed an interesting paper to the discussion on the “‘Correlation of the pre-Cambrian Rocks of the Region of the Great Lakes,” in which he formulates the hypothesis that throughout the region extending from the Adiron- dacks to northwestern Ontario there were in pre-Cambrian time “two and only two periods in which great granitic batholiths were developed in the earth’s crust.” On the basis of this hypothesis he correlates all the pre-Cambrian rocks occurring in the territory to which his hypothesis is applied.t. This hypothesis, if true, would undoubtedly greatly simplify the problems of pre-Cambrian nomenclature and correlation in the region under discussion; but an examination of the hypothesis from either a deductive or an inductive standpoint seems to indicate that it is an unwarranted assumption. The principal fact on which Lawson’s hypothesis of two and only two periods of granitic batholithic intrusion was based was that at the time the hypothesis was formulated only two periods 1 University of California Publications, X (1916), 1-10. LIMITATIONS OF PRE-CAMBRIAN NOMENCLATURE 331 of batholithic intrusion had been recognized in most of the pre- Cambrian subprovinces of the St. Lawrence basin. There is a very apparent reason, however, why two and only two batholithic intrusions can be recognized in a single locality, namely, that if a third batholith were intruded in a district where two batholiths were already present the evidence of the former presence of one or other of the older batholiths would probably disappear.’ If it be assumed that batholithic intrusions represent the interior portions of mountain chains it is obvious that the prolonged erosion, which generally follows an orogenic uplift, must inevitably result in the stripping off of the roof rocks from the underlying massif and the replacement of surface rocks by plutonic types in the district where the uplift has occurred; also that successive crustal movements of the orogenic type in the same or adjoining localities must eventually bring about the disappearance of all trace of rocks originally present in such zones of disturbance. It is probable that within the base-leveled pre-Cambrian complex which underlies the larger part of the Canadian shield evidence of the presence of more than two separate periods of batholithic intrusion would not generally survive in a single locality. I, however, the succession of formations can be determined over an extended area, as where less metamorphosed late pre-Cambrian sediments occur, the number of batholithic intrusions which can be recognized might be increased. Thus, as a result of the more extended areal geological studies of recent years, evidence is accu- mulating that at least three definite periods of batholithic invasion are represented in several of the pre-Cambrian subprovinces of the St. Lawrence basin. The folded and metamorphosed pre-Cambrian rocks occurring along the southern margin of the Canadian shield have in the main a northeasterly structural trend; likewise the granitic batholiths, so far as their areal distribution has been determined, are distributed in northeasterly trending zones; thus the region (approximately 1,000 miles in length) extending from the Adirondacks to the Lake of the Woods, to which Lawson’s hypothesis has been applied, lies almost transverse to the regional trend of pre-Cambrian folding, tA.C. Lane, Am. Jour. Sct., XLIII (1917), 42. 332 M. E. WILSON mountain building, and batholithic invasion. Moreover, mountain systems throughout the world are generally narrow and linear and, where zones of crustal disturbance composed of several mountain systems, such as the cordillera of North America, occur, the systems composing the zone are generally of varying age. Hence, if granitic massifs represent the interior of mountain systems exposed by denudation, it is more probable that the northeasterly trending pre-Cambrian batholithic zones of the St. Lawrence basin, instead of belonging to two and only two periods of batholithic intrusion, in reality represent several periods of batholithic development. Relationship to igneous intrusions other than batholith—Igneous intrusions other than batholiths, especially if they are composed of unique rock types, can likewise be employed for purposes of correlation, but generally only within a single subprovince. The principle has been used for inter-subprovincial correlation in the case of the late pre-Cambrian diabase intrusions, however, all of which have been generally regarded as Keweenawan in age. Folding and metamorphism.—Since folding and metamorphism are accompaniments of mountain building and batholithic invasion, these criteria are in reality included under the head, “‘ Relationship to Batholithic Intrusions.”’ It can be generally inferred that in the same district those rocks which are most highly folded and metamorphosed are the oldest in age. This does not follow in the case of widely separated regions, however; for it has been found that rocks which are flat-lying and slightly metamorphosed in one district may be highly folded, metamorphosed, and intruded by granite batholiths in another locality. CONCLUSION From the preceding discussion the following conclusions may be inferred: that the regional succession of formations within the pre-Cambrian of the St. Lawrence basin has not yet been sufficiently worked out for the establishment of a definite nomenclature applicable to the whole of this territory; that, since the St. Lawrence basin extends for approximately 1,000 miles in a direction transverse to the trend of pre-Cambrian mountain building and batholithic intrusion, it is theoretically improbable that the LIMITATIONS OF PRE-CAMBRIAN NOMENCLATURE 333 geological history of the various parts of this territory in pre- Cambrian time would be the same; that the geological succession of formations so far determined to be present in these various parts indicates that their history has not been the same; that the pre-Cambrian formations occurring in the St. Lawrence basin fall naturally into a number of separate subprovinces; and that it is advisable for the present that a separate nomenclature be employed in each of these subprovinces. The principal advantages that might be attributed to the widespread correlations implied in the present nomenclature applied to the pre-Cambrian rocks of the St. Lawrence basin are that they serve as tentative hypotheses for the investigator in the field, and as summaries of existing knowledge for the science as a whole. But it is doubtful whether tentative hypotheses have a place in geological nomenclature, and furthermore, for the investigator familiar with the facts in the field, tentative correlation tables indicating the most probable relationships of the formations occurring in the various subprovinces can afford all the advantages of a regional nomenclature, and for those’ not familiar with the facts in the field a general classification from which names of local origin are excluded could be used. Such a classification might be less definite, but it would be scientific, since it would express what is actually known, or what is at least generally accepted, by those familiar with the facts in the field rather than tentative hypotheses. CORRELATION OF THE EARLY SILURIAN ROCKS IN THE HUDSON BAY REGION? T. E. SAVAGE University of Illinois, Urbana, Illinois The oldest rocks of Silurian age known in the Hudson Bay region are present in the banks of Nelson River, about forty-five miles above tide-water. The best exposure is about four miles below the outcrop of Richmond strata at the lower Limestone Rapids of the river, where a vertical ledge outcrops to a height of 28 feet above low water. The rocks are nearly horizontal or gently undulating, and consist of yellowish-brown, rather fine-grained dolomite, in layers 4 to 1o inches thick. Masses of this dolomite form a pavement along the banks of the river at intervals for several miles below the main exposure, indicating that the river is actively cutting into these strata in places east of their actual outcrop. The fossils in this dolomite appear to be restricted to a narrow zone in the lower part of the bed. The most abundant are molds and casts, mostly of the ventral valve, of shells of the species described by Whiteaves as Conchidium decussatum from the basal Silurian strata at the Grand Rapids of Saskatchewan River. These shells are in places so crowded together as to make up the greater part of the rock layers, just as they occur at the Grand Rapids outcrop, where they are also restricted to a narrow zone. The shells of this species found in the Nelson River region show a wide variation in the ratio of their length and width, in the degree of convexity or galeation of the ventral valve, and in the develop- ment of the mesial fold on the ventral valve. Some of the partially exfoliated shells even show a distinct mesial sinus extending from the beak over the umbonal region of the ventral valve, which becomes obsolete or is transformed into a mesial fold in the middle «The strata discussed in this paper probably fall within the later half of the Oswegan series of the New York Classification. 334 EARLY SILURIAN ROCKS IN HUDSON BAY REGION 335 and anterior portions of the valve. Similar variations occur in the shells of this species found at the Grand Rapids of the Sas- katchewan River, described by Kindle* as follows: Conchidium decussatum belongs to a group of shells in which the specific characters are very plastic. .... The ventral valve shows three well-marked types of contour, viz.: (1) Strongly convex with a more or less clearly defined median ridge extending from the umbonal region to the front of the shell. (2) Very convex with tumid umbonal region rounding regularly from the median region to the lateral and anterior margins without trace of median ridge. (3) Strongly convex in median and anterior region with or without median ridge, but with a broad shallow sinus extending from the beak across _ the umbonal region. These three types of contour make striking contrasts when individuals in which they are best developed are compared; but the intermediate forms, in which neither the presence of ridge or sinus nor their entire absence can be positively stated, make difficult any attempt to dis- criminate them as distinct varieties. It is noteworthy that the shells showing a mesial sinus in the umbonal region of the ventral valve are young forms, and the writer is convinced that the more striking differences shown in the ventral valve of this species represent different growth stages in the individuals; the youthful stages show a mesial sinus from the beak across the umbonal region or farther anteriorly, while in the old stages the mesial sinus has disappeared and a distinct mesial fold is frequently developed. . The above-mentioned characters are the principal ones on which Twenhofel founded the genus Vzrgzana, and it seems cer- tain that the species formerly known as Conchidium decussatum really belongs to the genus Virgiana. Through the kindness of Dr. Kindle a comparison was made of the shells of this species from Nelson River with those from the Saskatchewan region in order to make sure of the identity of the species from the two localities. The shells from the Grand Rapids locality also show unmistakably the characters of the genus Virgiana’? to which this species is here referred. « E. M. Kindle, ‘‘ Notes on the Geology and Paleontology of the Lower Saskatche- wan River Valley,’ Geol. Surv. of Canada, Mus. Bull. No. 21 (Geol. Series No. 30), October 14, 1915, p. 10. 2Specimens of these shells were also sent to Dr. Twenhofel, the author of the genus, who agreed with the writer that they were true Virgianas. B20 LE SAVAGE The variation presented in the ventral valve of this species is similar to that shown in the shells described by the writer as Virgiana barrandei var. mayvillensis, and V. barrandet var. major from the Mayville limestone in Wisconsin. At the time those varieties were described the only other known representatives of this genus were Virgiana barrandei and a variety of that species occurring in the Becsie River (earliest Silurian) formation of Anticosti Island, from which it was thought that they might have been derived. However, the Virgzana shells from Wisconsin are now known to be more closely related to Virgzana decussata than to the Anticosti forms. In recognition of this relationship it is ~ here proposed to elevate the varieties Virgiana barrandei var. mayvillensis and Virgiana barrandet var. major to the rank of species. The former differs from Virgiana decussata in having somewhat fewer and coarser radiating plications, less numerous concentric markings, and usually is relatively wider in the anterior part of the shell. Virgiana major is a larger shell than V. decussaia, and generally has a much more strongly developed keel-like median ridge on the ventral valve. Regarding the age of the strata containing these shells in the Grand Rapids region Kindle’ says: Close comparison between the faunas of the Grand Rapids section and those of eastern Silurian sections, owing to the dearth of common species, is difficult. The dominance in the lowest (Silurian) fauna of this section of such a genus as Conchidium, however, makes it probable that the base of the section represents a Silurian horizon not earlier than the Clinton, and probably of early Niagaran age. This argument is no longer applicable, since instead of belonging to the middle Silurian genus Conchidium, the species in question belongs to the genus Virgzana, which is an early Silurian genus known elsewhere only from strata of pre-Niagaran (Alexandrian) age. The early Silurian age of the strata containing Virgiana decussata in the Grand Rapids of the Saskatchewan and the Hudson Bay regions can be shown by their relations to associated strata in dif- ferent areas. In the Grand Rapids region the layers containing Virgiana decussata are succeeded by strata which contain the fossils tk. M. Kindle, op. cit., p. 9. EARLY SILURIAN ROCKS IN HUDSON BAY REGION | 337 Pterimea occidentalis, Isochilina grandis var. latimarginata, and Leperditia hisingeri var. fabulina. In the Hudson Bay region a zone a few feet above the horizon of Virgiana decussata furnished shells of Camarotoechia ? winiskensis, Pterinea occidentalis, Isochilina grandis var. latimarginata, and Leperditia hisingeri var. fabulina. In the northern peninsula of Michigan‘ early Silurian strata con- taining Camarotoechia? winiskensis, Isochilina grandis var. lati- marginata, and Leperditia hisingeri var. fabulina overlie the strata containing Virgiana mayvillensis, which is a near relative of Virgiana decussata. In eastern Wisconsin Virgiana mayvillensis occurs in the uppermost layers of the Mayville limestone above which there is a stratigraphic break, the horizon of Camarotoechia ? winiskensis, Isochilina grandis var. latimarginata and Leperditia hisingert var. fabulina, present farther east in northern Michigan, having been removed by erosion. However, there is no doubt that the strata which in northern Michigan contain Virgiana mayvillensis correspond in age to those containing the same species in the upper part of Mayville limestone in Wisconsin, as they are clearly a north- eastward continuation of thesame beds. The relations of the strata containing Virgiana to the overlying and underlying beds in the regions above described are shown in the columnar sections in Fig. r. In Wisconsin there was found in the quarry near Peebles a zone only a few feet below the horizon of Virgiana mayvillensis and apparently conformable with it, which yielded such characteristic Edgewood species of fossils as Dalmanella edgewoodensis, Rhyncho- nella ? janea, Rhynchotreta parva, and Atrypa putilla. The position of Virgiana mayvillensis in Wisconsin in the upper part of the May- ville limestone, which at a slightly lower level contains a character- istic Edgewood fauna, indicates that this horizon is Alexandrian (late Edgewood) in age. It is also significant that the strata con- taining Virgiana mayvillensis in Wisconsin seems to occupy about the same position in the Silurian column as do the strata which contain Virgiana barrandet in the Becsie River formation of Anticosti Island. tT. E. Savage and H. F. Crooks, “‘Early Silurian Rocks of the Northern Peninsula of Michigan,” Am. Jour. of Science, XLIV (January, 1918), 59-64. In the lists of fossils given in this paper the name of the species given as Airypa putilla should have been written aff. Atrypa putilla. Lf. E. SAVAGE 338 Sieuspoomabps o/euonusyod eyynd edhity weunl g oauoyouhyy RAIDS B4944 oy ruhagLy sreusjjiahioud pluolbsif\ | I ‘On YIU) D444 Of 492 GF WNIS /DIZ4IA sysus//Ah pus puvibsi, < plessncap oupibsi, Dy~OSSNI2EpP Olivibsi, 4 os ae | — +——} "BDuUOZ OUun/b//A f? Joy I t euyngryy 40Ar s2buisiy pitipseda7 OfOUIEsOLLI4O/ OW/NG Rf E 4RA sipupsb DUI/1IGIOS/ DuUul/qo fs 4OA HIbLIS/YY DY /psDI27 Fs SS : ypA uabuisiy Dnipsade7 i : Sear SISUAYSIMIM ¢ DIF DAOLOSDLLIDD OL wulbsouiy Of Doubs, wy YOA sipues6 pwip/tj20s/ YOA sipuo416 PUI//4206/ SYO4USPIIIO OILMAS SIEUBYSILIAN § 014239040/0LUOD S10162p1930° 08 5 | ESSA _ LOL L NO/(D Ie NOID TY MSNOISIM NYILSY FT NY&OIHDIN NEFHLEYON SNO9FH LNFISTISNT NI ENEIDAA ININIEXLNOD LLEALS NE&ANSTHUS ATALT SO SNULY TILA ODNINOHS LYWHD AXE NOSOTH BAYH INYO EHARLY SILURIAN ROCKS IN HUDSON BAY REGION 330 In Michigan the strata containing Camarotoechia ? wintskensis, Isochilina grandis var. latimarginata, and Leperditia hisingert var. fabulina conformably overlie the Virgzana mayvillensis beds, and thus are thought to correspond in age to about that of the Sexton Creek or Kankakee limestone which overlies the Edgewood in Illinois and Missouri, but they were deposited in a different geologic province. The close correspondence in the fauna of the strata overlying the Virgiana mayvillensis zone in northern Michigan with that of the strata above the horizon of Virgiana decussata in the Hudson Bay and Saskatchewan regions leaves no doubt of the equivalence of the strata containing this fauna in the areas above mentioned. They also prove that the Vzrgiana mayvillensis zone in Wisconsin and Michigan, and the Vzrgzana decussata zone in the Hudson Bay and Saskatchewan localities represent the same stratigraphic horizon. Besides the above-mentioned localities Hume’ has found early Silurian strata containing Camarotoechia? winiskensis, and numer- ous ostracods in the Lake Timiskaming area that he correlates with the Cataract formation, which doubtless corresponds with the Camarotoechia? winiskensis, Isochilina, and Leperditia horizon in the.regions above described. The age assigned to this horizon by Hume agrees with that given by the writer above. Kindle? found Silurian strata several hundred miles north of the Grand Rapids locality, in the vicinity of the Pas, from which he obtained the fossils Camarotoechia? winiskensis, Pterinea ct. occidentalis, and Leperditia cf. hisingert. This fauna also indicates a horizon about equivalent to that of the Silurian in the Lake Timiskaming region and to the strata containing Pterinea occiden- talis, Isochilina, and Leperditia, above the Virgiana decussata horizon in the Grand Rapids section, the latter horizon not being exposed in the more northern locality. From the similarity in the faunas of the Virgiana zone, and of the higher strata containing G.S. Hume, ‘“ Paleozoic Rocks of Lake Timiskaming Area, Geol. Surv. of Canada, Sum. Rept. (1916), pp. 188-92. Fossils reported by Charles Schuchert in a personal letter. 2. M. Kindle, op: cit., p. 12. 340 Lf. E. SAVAGE Camarotoechia? winiskensis, Isochilina grandis var. latimarginata, and Leperditia hisingert var. fabulina in the regions above described it is inferred that during the time these strata were laid down the above-mentioned regions were a part of the same province or basin of deposition, which was rather broadly connected northward with the Arctic Ocean. This extensive northern invasion, together with the nearly synchronous deposits of the Cataract formation in a basin con- nected eastward with the Gulf of St. Lawrence region, and of the Brassfield and Sexton Creek limestones which were deposited in a southern basin, indicates a much more extensive submergence of the continent during early Silurian (late Alexandrian) time than was formerly supposed. The very close correspondence of the middle and late Ordovician and early Silurian rocks and faunas in the Saskatchewan and Hud- son Bay regions is strong evidence that they were deposited in a sea that was continuous between these areas. The presence of middle Ordovician and early Silurian rocks and faunas in the Lake Timiskaming region’ similar to those in the Hudson Bay region, and of late Middle and Upper Devonian strata in the vicinity of James Bay which probably originally extended south to the Timis- kaming region,” indicates that this part of the ancient Laurentian or Canadian shield did not exist continuously as a land surface throughout the Paleozoic era, as has generally been assumed, but that during middle and late Ordovician time, in early Silurian, and probably also during late Middle and Upper Devonian time the northern seas, temporarily at least, covered the greater part of this shield on the south and probably also on the west of Hudson and James bays. Kindle and Burling’ have previously shown that the seas probably also extended widely over the Laurentian upland southeast and east of Hudson Bay during the Paleozoic era. 1M. Y. Williams, ‘‘The Ordovician Rocks of Lake Timiskaming,” Geol. Surv. of Canada, Mus. Bull. No. 17 (Geol. Series No. 27), June 7, 1915. 2G. S. Hume, ‘Paleozoic Rocks of Lake Timiskaming Area,” Geol. Surv. of Canada, Sum. Rept. (1916), p. 190. 3. M. Kindle and L. D. Burling, “‘Structural Relations of the Pre-Cambrian and Paleozoic Rocks North of the Ottawa and St. Lawrence Valleys,’ Geol. Surv.. of Canada, Mus. Bull. No. 18 (Geol. Series, No. 28), July 23, 191s. NOTES ON SEDIMENTATION IN THE MACKENZIE RIVER BASIN’ E. M. KINDLE Geological Survey of Canada, Ottawa INTRODUCTION In the course of field work in the Mackenzie River region during the summer of 1917 the writer had an opportunity to observe various features in connection with the constructive and destruc- tive work of the rivers and lakes traversed. The following notes relate to the lower Peace, the Slave, Athabasca, and upper Mac- kenzie rivers, and Great Slave and Athabasca lakes (Fig. 1). The making of these observations was incidental to other work, and they are assembled here as a fragmentary contribution to a knowl- edge of present-day work in continental sedimentation in the Northwest. The parallel streams of clear and muddy water in the channel of the Mackenzie, its sloping bowlder-paved banks in marked con- trast with the cut banks of the Peace and Slave rivers, and its relatively straight course are among the noteworthy features of this great river. The marked inequality in the rate of sedimen- tation on opposite sides of Great Slave Lake is one of the signifi- cant features in connection with the lacustrine sedimentation of - the region. These and other factors relating to sedimentation in the Mackenzie basin will be discussed in the following notes. MATERIALS OF THE VALLEY FLOORS The valleys of the Peace and Athabasca rivers are throughout the major part of their courses cut deeply into the shales and sandstones of the Cretaceous formations. At Peace River crossing the Peace River flows in a steep-sided valley cut about nine hundred t Published with the permission of the Director of the Geological Survey of Canada. 341 342 E. M. KINDLE feet below the surface of the Cretaceous plateau. The Athabasca, where it joins the Clearwater, has cut its valley into the Cretaceous rocks to a depth of about five hundred feet. The lower portions of both streams, however, flow for considerable distances across a very broad, low, flat plain of recent origin, which is composed Frc. 1.—Sketch map of the Mackenzie River basin chiefly of fine silts and sands. These beds are of both lacustrine and fluvial origin and evidently of postglacial age. Near the river the fluvial beds are more in evidence than the lacustrine, but the latter probably have a far greater extent than the former. In the Peace River region the cut banks of sand rising in places 80 feet high or more above the water represent the bottom deposits of a SEDIMENTATION IN MACKENZIE RIVER BASIN 343 greater Lake Athabasca which was a contemporary of Lake Agassiz. Similar beds occur along the lower Slave River (Fig. 2), which belong to the period when a much greater Slave Lake was developing the Fic. 2.—Lacustrine sediments on the Slave River near Salt River, representing deposits of a former greatly extended stage of Great Slave Lake. Photograph by E. J. Whittaker. wave-cut cliffs and elevated beaches now found far back from its shores. The following section, taken above Little Rapids on the Peace, indicates the general character of these lacustrine beds: MRTOTES De Cuarl dia SOnl ny aah p pend Wien tisees alan t foot ZR BTOWNNSATIG ie Chey iset cn fn Me eu. 3 feet Ge OLeSt DEGaye LC. ne ait) cotaas Mesmvaten el gia end 6 inches to o 4. Gray sand with pebble band in middle....... 1 foot Peelardssanduwithvlimy bands a peer = eT TOOk OB Nihite mar ey Aes oie ysl en aat atm m Ore 4" 340.2 8 inches to o jeg Giay, sang with calcined moots. yan)... 25 feet See Narhyssain dhyeci icy isin ela ene nod ees rusk 2 inches g. Gray even-bedded sand partly covered....... 30 feet The broad low plain of lacustrine sands which have been extensively re-worked by the river, extends from the pre-Cambrian hills east of Slave River and the lower Athabasca many miles to the westward. It meets the foot of a plateau called the Caribou Mountains west of the Slave River, while a similar upland known 344 E. M. KINDLE as the Birch Mountain projects shieldlike from the south into the low land between the Peace and the Athabasca rivers. The banks of the Mackenzie for the first 200 miles below Great Slave Lake are in some places very low, particularly near the head of the river, though rising 50 to 150 feet above the stream in many places. But the materials in which the channel is cut are nearly everywhere either glacial drift or Devonian shales. It will be pointed out later that the till banks of the Mackenzie give rise to certain peculiarities of the river which are not found in the Slave, lower Peace, and Athabasca rivers, whose channels are cut largely in lacustrine beds. RELATIVE TURBIDITY The major part of the great volume of water comprising the upper Mackenzie River is gathered by two large rivers, the Peace and the Athabasca. The latter stream is filtered through Atha- basca Lake before joining the Peace. The Peace, united with the Athabasca under the name of the Slave River, pours its flood of sediment-laden water into Great Slave Lake, which discharges it, freed of its burden of floating trees and suspended mud, into the head of the Mackenzie River. The filtered waters of the Slave and half a dozen other considerable streams which flow into Great Slave Lake unite in the upper Mackenzie to form a stream which, in both volume and clarity, is comparable with the upper St. Lawrence. This clear river issuing from the lower end of Great . Slave Lake receives no stream of notable size for 150 miles, to the point where it is joined by the Liard from the west. The Liard drains a great area including all the easterly valleys of the Rocky Mountains between the head waters of the Peace and the Yukon. The large volume of the Liard is characterized by a high degree of turbidity. Its muddy waters join the clear waters of the Mackenzie at Fort Simpson, but for more than 160 miles below Simpson they fail to mix except in a limited zone near the middle of the river. During the canoe trip down the Mackenzie observations regarding the relative clarity of the water on the two sides of the river were made repeatedly. The following excerpts from my notes indicate the contrasts observed. Twenty-five miles below the mouth of the Liard River, a crossing from the east to the west bank was made to SEDIMENTATION IN MACKENZIE RIVER BASIN 345 observe the relative amounts of sediment on the two sides. On the east side the water was quite clear, the tip of an oar being visible as far down as its length. One third of a mile from the east bank the water showed a trace of sediment, the visibility in looking down into it along the side of an oar being noticeably less. ‘The suspended matter gradually increased toward the west bank for a quarter of a mile. The last third of a mile on the west side held so much sedi- ment that only indistinctly could the bottom of a cup 4 inches deep be seen through it. An oar could not be seen more than 3 or 4 inches below the surface of the water. ‘At old Fort Wrigley, 140 miles below the Liard, the water on the east side of the river is quite clear, no visible sediment is present, and visibility extends down 3 feet or more. On the west side the bottom of a cup 4 inches deep can be discerned through the water only very indistinctly. The contrast in driftwood also continues striking, being very abundant on the west and scarce on the east side.”’ This contrast, though in a somewhat less marked degree, extended as far as New Fort Wrigley, 160 miles below the Liard, the northern limit of my journey. At Fort Norman a large river brings the clear waters of the Great Bear Lake drainage basin into the east side of the Mackenzie. This large accession of clear water to the east side doubtless results in keeping the eastern half of the Mackenzie comparatively clear to the head of the delta. Thus we have the curious phenomenon of two rivers, one a clear and one a highly turbid stream, flowing side by side in the same channel without mixing except ina comparatively narrow zone. As a result the deposition of sediment differs markedly in amount and kind on the two sides of the Mackenzie. The islands of alluvial material which occur at various points in the river below the Liard are all confined to the western half of the stream. The abundant supply of drift logs which comes down the Liard furnishes a large amount of drift timber to the west bank of the Mackenzie. Com- paratively little of it lands on the east bank. DESTRUCTIVE AND CONSTRUCTIVE WORK OF THE RIVERS The Slave River carries vast quantities of sediment into Great Slave Lake. It enters the lake through several narrow channels, 346 E. M. KINDLE the mouths of the easternmost and westernmost of them being — separated by more than ten miles. The lake water in front of the delta is, for some miles out, quite shallow. The submerged margin of the Slave River delta over several square miles is less than a half-foot in depth. This zone is strewn in most places by number- less logs and trees. Inside these comes a broad, irregular fringe of grass-covered land with no willows or other trees, too low for the bank to show. Farther up the delta some of the shores have been built up one foot or a little more above the late summer stage of the water and are covered with willows. Between these islands of the delta, with willow and other low trees, stretches a vast net- work of shoal-water channels and marsh land. Nearly all these ‘hannels are bordered with large quantities of driftwood. Still higher up the delta small poplars are scattered in patches among and behind, or inland from, the willows. Some miles up from the lake, where the banks rise about 3 feet above the water, spruce comes in with poplar. There is, however, no-relief beyond the increase of a few inches or eet in the elevation of the shore above water. With this slight elevation cut banks appear, together with the slumping of trees and sections of the bank into the river. Along the Sawmill branch of the delta willows and alders make a solid wall of low, overhanging brush. The banks under these show great numbers of .ogs projecting from the silts, as noted by McConnell.’ Above the delta the banks rise to an average height of 7 or 8 feet above the ordinary midsummer stage of the river, and to or 15 feet is not unusual. The banks are nearly everywhere of fine silt. The following is a representative section taken several miles above the delta: Soilvandbmeanvamatertalls 255 ay) ween Leer eee te t foot Silt and dark bands of organic matter in alternating [ern Gl See reese ces ae csia et kl ne a A a ea 4 feet LS UNVS SARE iMG 2g te ce oes PRM SI Misia dlole gio obeys 7 feet Cut banks are found everywhere on one side or on both sides. These retreat rapidly during the warm season. A trapper’s cabin was observed at one point partially undermined by the sapping of t Canada Geol. Surv., Ann. Rept., 1V, 1888-89 (1890), p. 66 D. SEDIMENTATION IN MACKENZIE RIVER BASIN 347 the river. The spruce timber is found caving into the river along considerable stretches. Where a strong current sets against the bank caving proceeds as rapidly as the face of the cut bank thaws. The heavy mat of moss and vegetable matter prevents thawing downward beyond a couple of feet, or less in most sections. The lateral thawing on the face of cut banks results in overhanging masses of silt covered with forest trees (Fig. 3). These finally break off from the bank as thawing and undercutting proceeds, and the slumping frequently splits the trunks of trees, leaving half of Fic. 3.—Destructive river work on the outside of a curve, lower Slave River stump on shore. Islands are formed quickly, and many of them disappear quickly. A sandbar first appears; then a multitude of willows spring up. If ice and floods are not too devastating during the next seasons, the small willows persist and materially aid in adding more sediment to the bar. The building of silt islands below places of maximum cutting, or opposite cut banks on the inside of the curve which results from their development, is seen throughout the course of the Slave River below Fort Smith. These islands, if near one bank, are likely to have successive zones of sediment added to them on the sid2 next the channel with the lesser current, until it is closed and they become 348 E, M. KINDLE apart of the mainland. All stages of these islands may be observed, from the sandbar just emerging from the water, with no trace of vegetation, to the island with a mature forest of large spruce. As soon as a bar island is built sufficiently above low-water stages for any vegetation to survive on it, a dense growth of willows covers it. These for some years practically exclude other kinds of trees. Their enormously long roots form a network which protects the loose silty material of the young island from destruction by high water, while their twigs and stems greatly accelerate the accumula- tion of sediment by checking the velocity of the current around Fic. 4.—Constructive river work on the inside of a curve. Note three successive’ growths of willows in front of the tall spruce timber representing periodic increments of silt bands to the river bank. them. The growth of the silt island is therefore rapid after the first growth of willows has become well established. When the island has been built sufficiently high by the annual accretion of sediment, poplars and later spruce begin to displace the willows. Frequently three or more zones may be distinguished around these islands, each a year or more younger than the one inside it, by the height of the willows on them (Fig. 4). Above the mouth of Slave River for about 125 miles cut banks of yellow sand 15 to 18 feet high are common for long stretches. They terminate abruptly against the ordinary alluvial banks and evidently represent a different and earlier set of deposits which probably are of lacustrine origin. Immediately below the Grand SEDIMENTATION IN MACKENZIE RIVER BASIN | 349 Detour cut banks of this yellowish sand and gravel 40 feet high are exposed. The outermost island at the west end of the Grand Detour illustrates the lateral migration of islands which is sometimes observed. The western border is a recent sandbar formed during the present year. Inside this border is a crescent of willows one year old, while the third zone is a very narrow belt of willows nearly mature. The eastern and oldest part of the island is covered with poplars. This east side, however, is a cut bank and is being removed apparently at the same rate at which the western shore is being built up. On the north half of the Grand Detour the cutting is all on the outside bank, but this is not true of the south half, where islands in the channel deflect the current to the inside of the bow, where it now appears to be doing the maximum amount of cutting in beds of buff sand toward the upper end of the bow. Near Fort Smith the sandbanks of the river reach their maxi- mum height. At the steamer landing at Fort Smith the top of the bluff is 125 feet above the level of the late-summer stage of the water. The base of the section is a bed of thinly laminated gray clay 6 feet thick with numerous concretions which have the appear- ance of being built up of a series of disks each smaller than the pre- ceding. Above this laminated clay the beds appear to be composed entirely of sand. For 16 miles above Fort Smith the Slave River flows over a series of granite ledges and between numerous low, rounded, granite islands which interrupt navigation. Above this series of rapids most of the islands are composed either partially or exclusively of granite or limestone instead of silts as in the lower Slave. Above the Stony Islands, where the granite islands rise to a maximum of - more than too feet above the river, numerous low, granite bosses also rise at intervals a few feet above the water surface, in many cases only a foot or two above low-water stage. Some of them are only to to 20 feet in length. Groups of half a dozen or more of these small granite islands are seen in a distance of 200 yards up and down stream. ‘These granite knobs are the nuclei of the many long alluvial islands seen in this part of the river. Frequently a pile of driftwood caps the top of one of the low granite knobs 350 E. M. KINDLE (Fig. 5). Such a drift pile checks the current and furnishes the beginning of the conditions essential to island formation. The river makes a deposit of silt below it, and this, in time, may connect with another similarly formed ‘sland by downstream growth. These low knobs prevail up to the mouth of the main outlet of the Peace. Above the mouth of the Peace the knobs of granite increase in elevation till 25 feet is an average height. Still higher up the Slave River the granite knobs increase in height toward Little Lake and Lake Athabasca, just north, until an elevation of 150 feet or more is reached. In Peace River the same type of island construc- tion and destruction which characterizes the lower Slave is seen. Fic. 5.—Island composed of drift timber, Slave River Great numbers of spruce trees are being undermined constantly and thrown into the river. This timber from the cut banks of the Peace and Slave is the source of the enormous quantities of drift logs which line the shores of Great Slave Lake (Fig. 3). Some of the logs accumulate on the low, granite-island knobs and form islands, some of which appear, at high water, to consist exclusively of logs (Fig. 5). Trees whose roots are heavily Joaded with earth and stones often strand in shoal water and form the nuclei of new islands. The delta of Athabasca River near the western end of Athabasca Lake is very similar to that of Slave River in Great Slave Lake. The lake water in front of it is only 2 to 4 feet deep for a considerable distance except along a narrow, crooked channel. The banks rise above the river (late-summer stage) less than 1 foot for 3 or 4 miles. Higher up from the lake they rise gradually to about 3 SEDIMENTATION IN MACKENZIE RIVER BASIN 351 feet, when willows become common. Ten miles above the river mouth the banks rise to 5 or 6 feet, and large willows and alders are common, but no spruce trees or other evergreens are seen. The banks of alluvium increase gradually in height till at the old Fort Forks at the head of the delta they rise 10 or 12 feet above low water. Numerous good examples occur in the lower part of the Atha- basca, in the cut banks, of tree stumps which have been buried where they grew under from 2 to 6 feet of alluvium by the shifting of the course of the river. They are illustrated by one of E. T. Seaton’s figures.t. O’Neill? has described similar examples of allu- vium-buried forest beds in the delta of the Mackenzie where the buried stumps are much larger in girth than any other trees now growing in the delta. The testimony of the land surveyors who have run their lines across an extensive region between the lower Peace and Athabasca rivers indicates that with the exception of two or three localities bedrock outcrops are wanting over a vast area between these two rivers, an area which is doubtless underlain throughout by fluvial and lacustrine deposits. About 15 miles above the Old Fort forks the river cuts into a sand bluff 75 or 80 feet high. The heavy load of sand acquired by the river at this and other points higher up results in extensive sandbars which are spread over the middle of the river and interfere with steamer navigation at low water. About 3 miles below Point Brule the extensive alluvial and lacustrine deposits are terminated on the east side of the river by land rising 200 feet or more above it. CONTRASTING FEATURES OF THE MACKENZIE RIVER The bowlder pavements are among the most striking features of the Mackenzie River. These marvelous pavements, resembling cobblestone roadways, often stretch along both banks of the great river without interruption for miles (Fig. 6). They frequently extend up the concave banks from below low water to a height of 25 feet or more above it. The shores of many of the islands as well as the banks of the river are paved with bowlders. On the Slave, lower Peace, and lower Athabasca rivers the pavements are entirely ' The Arctic Prairies, p. 197. Charles Scribner’s Sons (1911). 2 Canada Geol. Surv., Sum. Rept. (1915), p- 230. 352 E, M. KINDLE absent. The channels of these streams are cut in lake and river silts which contain no bowlders, while the channel of the upper Mackenzie is cut for the most part in glacial till containing an abundance of bowlders, which through the grinding and sliding of the ice during the spring break-up are pressed deeply into the clay and built into pavements. The contrast between the bowlder- caved banks of the Mackenzie and the bowlder-free banks of the rivers just mentioned is related to another feature in which the Mackenzie contrasts sharply with these rivers in the bowlder-free silts. The latter meander widely, while the former pursues a fairly direct course, its bends showing none of the characteristics of typical meandering streams. Fic. 6.—Bowlder pavement on island opposite Old Fort Wrigley, Mackensie River. The Grand Detour on the lower Slave is an example of the meanders of this stream and the lower Peace. At the Grand Detour the Slave swings abruptly to the westward in a great loop of about 20 miles. The distance across the base of this meander is only one mile. The relatively direct course of the Mackenzie as compared with the meandering lower Slave and Peace rivers can be explained, in part at least, by the protection which the bowlder pavements afford against lateral cutting. These pavements furnish protection against erosion of the banks as effective as artificial riprap, and thus prevent the excessive cutting at the bends which in many streams leads to the formation of loops and oxbows. SEDIMENTATION IN MACKENZIE RIVER BASIN 353 The plowing and gouging action of ice is nearly everywhere in evidence along the Mackenzie. At the head of the river, in the shallow eastern channel, one can see through the clear water numer- ous deep grooves made by ice cakes or bowlders pushed by ice in the bowlder clay of the bottom. In the gravel or silts of low islands the broad grooves made by ice-shoved bowlders or ice blocks can often be traced for a considerable distance (Fig. 7). In some local- ities the plowing and scooping action of the ice carries large quan- tities of mud from the bottom to the banks of the river (Fig. 8). Fic. 7.—Trail left by ice-shoved bowlder or ice cake. Note the cratic course unlike that left by drifted tree roots, Mackenzie River. The upstream ends of some of the low islands are built up in this way several feet higher than the rest of the island. In such cases the ice is likely to build a clay dike across the head of the island at right angles to the course of the river terminating at the top in a sharp ridge. The front of such a dike is frequently bowlder-paved and thus becomes almost as resistant to river erosion as a hard rock cliff (Fig. 9). A feature of the ice work along the banks of the Mackenzie is the distribution of great numbers of a small bivalve, Sphaerium vermontanum, over the higher levels of the bank, much higher than the ordinary stages of the river in summer could carry them. I 354 E. M. KINDLE have never seen this shell in the shallow water of the river, although other shells are common there. It appears, therefore, that it lives Fic. 8.—River bottom clay and silt shoved on bank of an island by ice scour, Mackenzie River. © abundantly in the deeper parts of the river and reaches the local- ities where found on the banks as a result of ice excavation and the vagaries of strong current action and transportation during the Fic. 9.—A clay island protected from erosion by a bowlder pavement spring break-up. When the ice breaks on the river in the spring, ice jams occur which raise the river to abnormal heights at various localities. McConnell' has given the following striking descrip- 1 Geol. Surv. of Canada, Ann. Rept., IV, 1888-89 (1890), p. 87 D. SEDIMENTATION IN MACKENZIE RIVER BASIN 355 tion of this phenomenon as observed by him at the mouth of the Liard: Huge cakes of ice under the enormous pressure were constantly raising themselves on end and falling, and the whole mass, urged forward by the terrible energy of the piled-up waters behind, was battering a way across the Mackenzie. The ice of the latter, fully five feet thick and firm and solid as in midwinter, was cut through like cardboard, and in a few moments two lanes were formed across its entire width, while a third was open for some distance below, before the force of the rush was exhausted and the movement ceased. In the afternoon the crashing of trees in a channel behind the island, con- cealed from view by the intervening forest, was distinctly heard and showed that a temporary vent had been found there, and in front of the fort intermittent fountains played at intervals from holes and crevices in the ice. At midnight Fic. 1o.—A remnant of a mass of river-shoved ice and the bowlder pavement which such ice levels and smoothes, Mackenzie River 15 miles above Fort Wrigley. the dam at the mouth of the Liard gave way and the massive crystal structure was hurled by the liquid energy behind it against the firm ice in front with such force that the whole sheet, for some miles below the fort, was crushed into fragments by the impetuosity of the assault. At the Ramparts ice jams are reported to have sometimes raised the river nearly too feet. . About 15 miles above Fort Wrigley, I observed the remnants of one of these ice jams—an accumulation of great ice blocks which had remained unmelted as late as July 22 (Fig. 10). At this date the principal mass of ice blocks had a thickness of 25 feet. They were covered with a thin veneer of dark mud, and hundreds of specimens of Sphaerium vermontanum were scattered over their surface. 350 E, M. KINDLE Good examples of rock basins which represent apparently one phase of the work done by river ice occur on the limestone island opposite old Fort Wrigley. A considerable area of limestone beds lying approximately horizontal on the northeastern side of the island is covered by the river during the spring break-up, but exposed through the summer. On this area a group of four rock basins has been developed in the limestone. The rim of the largest of these rises from 5 to 10 feet above the bottom. This basin has a maximum length of 65 feet and a width of 30 feet (Fig. 11). Another of these basins has a diameter of 10 feet and a depth of 5 Fic. 11.—A rock basin in limestone between low and high water, Old Fort Wrigley, Mackenzie River. feet. The basins appear to be the product of the plucking action of the river ice which covers them during the late winter stages of the river. LAKE FILLING Two large lakes, Athabasca and Great Slave, lie in the path of the Mackenzie-Athabasca drainage system. Great Slave Lake, much the larger of the two, has a length of about 290 miles from east to west. A recent survey of part of the north shore by A. E. Cameron, of the Canadian Geological Survey, changes the rank in size of Great Slave Lake from fifth to fourth among the great lakes of the continent. As pointed out by McConnell,* it seems originally to have had ‘‘the form of a great cross with one arm 1 Geol. Surv. of Canada, Ann. Rept., TV, 1888-89 (1890), p. 65 D. SEDIMENTATION IN MACKENZIE RIVER BASIN 357 penetrating the crystalline schists while two others stretched north and south along the junction of these with the newer sedimentaries, and the fourth extended itself over the flat-lying Devonian to the west.”’ Lake Athabasca lies almost entirely within the limits of the pre-Cambrian rocks. Inspection of the map (Fig. 1) will show that the drainage of an enormous area in Northwestern Canada, extending from the interior of the Rocky Mountains region far into the pre-Cambrian area west of Hudson Bay, passes through Lake Athabasca and Great Slave Lake. Practically all of the vast quantity of sediment which is annually stripped from this extensive area is left in these great settling basins. A noteworthy feature of this lacustrine sedimenta- tion is the extreme inequality of its distribution. Probably 95 per cent of the immense volume of sediment which enters Great Slave Lake is poured into the south side of the lake. The streams entering the north side of the lake are nearly all small and com- paratively insignificant. In the course of a survey of the north shore of the west arm of Great Slave Lake, A. E. Cameron found that “‘throughout the entire 136 miles of shore line between the Mackenzie River and the north arm only one stream, and it a very minor one, was found entering the lake” (manuscript). The streams which do enter the north shore of the lake lose most of their sedi- ment in passing through small lakes before reaching Great Slave Lake. On the south shore, besides the Slave, which is one of the great sediment-bearing streams of the continent, three other rivers enter, each of which has a considerable volume. These are the Taltson, the Buffalo, and the Hay. It is the zone of lake bottom bordering the 150 miles of the south shore receiving these streams which takes the great bulk of the river-borne sediment. Great Slave Lake opposite the mouth of Slave River, which carries the great bulk of the silts entering the lake, has a width of more than 60 miles. Little or none of the sediment brought in by the Slave has any chance of being deposited in the northern half of the lake. Coastwise currents, however, distribute the silts from the Slave and other south-shore streams widely along the south-shore bottom zone. It is probable that along the shore line between the Hay and the Buffalo rivers the prevailing direction of the coastwise 358 E. M. KINDLE currents is easterly. This was evidently the case off the mouth of the Buffalo when I passed it early in July. The canoe entered muddy water some distance east of the mouth of the stream, but entered clear water immediately west of the west bank of the river, although a breeze from the northeast was blowing at the time. At a later date Mr. Cameron made similar observations at the mouth of the Buffalo. The muddy water of the Slave ceases to be noticeable in the lake ro or 12 miles west of the western side of the delta. There is probably no lake in North America which receives any- thing like the amount of driftwood which is poured into Great Slave Lake, chiefly through the Slave River (Fig. 12). The shores Fic. 12.—Drift timber on the shore of Great Slave Lake are nearly everywhere lined with enormous quantities of logs, many of which came from a thousand miles or more up the Peace. Large quantities of this driftwood must eventually become water-logged and sink. Practically none of it leaves the lake by the Mackenzie. At some localities the drift timber is intimately mixed with the shingle of the beaches (Fig. 13). In Lake Athabasca sedimentation appears to be even more localized than in Great Slave Lake. The maximum length of Lake Athabasca, which nearly equals that of Lake Ontario, lies in an east and west direction. ‘The great bulk of the sediment which the lake receives is poured into the western end by the Athabasca and Peace rivers. It is discharged from this end by way of Little Lake and Slave River. The relationship of the Peace River dis- charge to Lake Athabasca is peculiar and variable. Three or four SEDIMENTATION IN MACKENZIE RIVER BASIN 359 outlet channels of the Peace empty directly into Slave River, but the Quatre Fourches channel, which branches off above the Slave River outlets, empties into Lake Athabasca. Ordinarily by far the greater part of the Peace River water flows directly to Slave Lake via Slave River; but a conjunction of low water in Lake Athabasca and high water in the Peace River may for a time reverse the direction of flow in the upper Slave River and turn all of the Peace River outflow into Lake Athabasca. Such a reversal of drainage involves the temporary obliteration and reversal of Little Rapid, located about 25 miles north of Lake Athabasca in Slave River. Father Lafebre, of the Catholic Mission, informs me Pomel Hs Fic. 13.—A double beach of a and drift wood, shore of Great Slave Lake that he has observed this reversal in May. It thus appears that at times the whole of the combined volume of sediment carried by the Peace and the Athabasca is dumped into Lake Athabasca. When this reversal of current near the head of the Slave River occurs, it is evident that the deposition of sediments must proceed in the western part of Lake Athabasca at an enormously increased rate. Under these special conditions the whole of the Peace River sedi- ment, nearly all of which ordinarily reaches Great Slave Lake, stops in Lake Athabasca. During the unusual seasons when the Peace River contributes largely to the Lake Athabasca sediments, the annual layers of silt would not only be thicker than those ordinarily laid down but would also probably have a much greater easterly extension. 360 E, M. KINDLE Lakes Mamawa and Clair, lying west of Athabasca Lake, appear to represent parts of a former greater Lake Athabasca which have been segregated into separate lakes by lake filling. Lake Clair has a shore line approximating 200 miles in length. Its depth is reported seldom to exceed 8 feet. Little Lake, through which the outflow of Lake Athabasca passes, is extremely shallow except for a channel through which most of the outflow passes to Slave River. East of this channel a large area on the north side of the lake is completely silted up. On the west side large areas of aquatic plants and stranded logs indicate the approach of the final stage and extinc- tion of the lake. Observations made on the relative clarity of the Slave River water above and below the mouths of the Peace show but a slight difference in the clarity of the water. This would indicate that the river water passes through the narrow western part of Lake Atha- basca too quickly to lose nearly all of its sediment. | NOTES ON TEE MiISSISSIPPIAN CHERD OF THE Sf. LOUIS AREA DONALD C. BARTON Cambridge, Massachusetts The origin of chert is a question which is still open to discussion. Although several of the chert series of Europe have been studied somewhat in detail by a number of geologists, among the more important Hull and Hardman, Hinde, Sollas, Renard, and Cayeux, the conclusions reached with regard to the origin of chert have been considerably at variance. In this country Lawson and Palache seem to have demonstrated the organic (radiolarian) origin of certain Californian cherts. ‘The Missouri cherts, chiefly Missis- sippian, and some of the closely associated cherts of neighboring states have been studied, although not in detail, by Shepard, Ball and Smith, and Van Tuyl, and, in thin section only, by Hovey. The conclusions as to the origin of the chert have not been in agree- ment. ‘The present paper presents the results of a detailed study of the Mississippian cherts of the St. Louis area both in the field and in thin sections. OCCURRENCE OF THE MISSISSIPPIAN CHERT OF THE ST. LOUIS AREA The Mississippian chert of the St. Louis area is found in the St. Louis and in the Burlington-Keokuk limestones, to a slight extent in the Warsaw shales, and in very rare, small patches in the 1 E. Hulland E. T. Hardman, Trans. Royal Soc., Dublin, I (1878),71. G.J. Hinde, Geol. Mag. (III), IV (1887), 435-46. W.J.Sollas, Am. Mag. Nat. Hist. (5), VI (1880); VII (1881); Proc. Roy. Soc., Dublin, VI (1887), Part 11. A. F. Renard, Bull. Acad. Roy. Belgique (2), XLVI (1878), 471. L. Cayeux, Ass. frang. pour Pavanc. de Sci. (Carthage, 1906), pp. 220-93. A.C. Lawson and C. Palache, Bull. Geol. Dept. U. of Cal., IL (1902), 354-65. E.M. Shepard, U.S. Geol. Surv., W.S. Paper 195 (1907), p.19. T.M. Van Tuyl, Proc. Iowa Acad. of Sci., XIX (1912),173-74. E. O. Hovey, Mo. Geol. Surv., VII, Part II (1894), pp. 727-39. 361 362 DONALD C. BARTON Salem limestone. The stratigraphic position of the chert beds is shown in the accompanying generalized section of the St. Louis area (Fig. 1). The Burlington-Keokuk limestone wherever exposed Pennsylvanian shales Clastic Mississippian chert St. Louis limestone Massive and argillaceous limestone Bed of cherty limestone Massive limestone, much lithographic and some argillaceous limestone Massive granular limestone with a few thin, shaly layers, chert common at many horizons Salem limestone Thick-bedded granular limestone Chert free Warsaw shale Shales, cherty limestone at the base Burlington-Keokuk limestone Shales intermixed with limestone and chert Thin-bedded limestone and chert Fern Glen limestone Fic. 1.—Generalized section of the St. Louis region in this area is consistently very cherty. The amount of chert in the St. Louis limestone varies considerably from place to place. The presence of more or less chert at the base is characteristic, but it MISSESSTPPIAN: CHERT OF ST. LOUIS AREA 263 was impossible to correlate with certainty throughout the area the higher cherty horizons. MORPHOLOGY OF THE CHERT BEDS The general form of the chert varies considerably from bed to bed, although it seems to be more or less constant for any individual horizon. It is possible to distinguish several distinct types of occurrence of the chert. Most of the chert in the Burlington- Keokuk limestone and some of the chert in the St. Louis limestone are in nodular bands or bands of flattened nodules. The nodules are irregularly elliptical with horizontal diameters subequal and one and a half to twice the vertical diameter, and are seldom less than three centimeters or more than ten centimeters in diameter ver- tically. The nodules are rounded and usually sharply delimited, at least to megascopic examination, from the inclosing limestone. In any given chert bed these nodules show distinct distribution parallel to the stratification and by coalescence form the nodular bands intercalated between the thin limestone beds. In a few chert beds in the Burlington-Keokuk the chert is irregularly rami- fying, with angular outlines, and in general pattern resembles some of the mottled Ordovician limestones. Although the chert of this type usually crosses many distinct layers of the limestone, the greatest development shows distinct distribution parallel to the stratification. A form of the chert very characteristic of the St. Louis limestone and found but very rarely in the Burlington- Keokuk are the bands of nodules, spherical or ovoid in shape and six to sixty centimeters in diameter. The contact of the nodule and the limestone is apparently sharp, but there is usually present a thin chalky-looking transition zone. The nodules in a given band characteristically are well aligned to some horizon, in many cases the middle of a thick limestone bed. A form of the chert that is found both in the St. Louis limestone and in the Burlington- Keokuk limestone is a thin band, pancake-like in the St. Louis and platelike in the Burlington-Keokuk. These bands most commonly are six to ten centimeters thick, but in some cases much more, and fifteen to fifty meters long. The contact of the chert of these bands with the limestone seems to be sharp. 364 DONALD C. BARTON LITHOLOGICAL CHARACTER OF THE CHERT To the naked eye the chert characteristically appears stony, but in some cases granular or chalky, although in the latter cases the chert is not appreciably less tough or hard. In color the chert varies from dirty white to dark gray and, except for one thin band of chalcedonic chert, is mottled and banded. The mottling is simi- lar in effect to that given by the grain to the limestone and is appar- ently a pseudomorphic character of the chert reflecting the granular character of the limestone. The banding is concentric with the form of the nodule, or horizontal. In the latter case it is a retention of the stratification markings. The chert with rare exceptions is fossiliferous wherever the limestone of that horizon is fossiliferous. Crinoid stems are by far the most common fossil. Bryozoa, Spiri- feri, Producti, and other brachiopods, Lithostrotion in the St. Louis limestone and Fusulina are also common. A notable feature about the fossils is that in a very great number of cases they are still cal- careous and do not show the effects of the siliceous replacements of the rest of the rock. In thin section under the microscope the chert is seen to be com- posed chiefly of quartz with more or less calcite, and in some cases with chalcedony, opal, dolomite, pyrite, and iron staining. The quartz making up the mass of the chert is in excessively fine grains, less than o.o1 mm. in diameter, which are not clearly distinguished even under the high power, and it is to the compensation due to the superposition of these small grains that many of the areas dark under crossed nicols are to be attributed. Locally, in many cases within a shell, there are patches of allotromorphic grains of larger size (0.1 mm.). Ina few thin sections there were seen larger, sub- angular, clastic grains. The calcite, abundant through much of the chert, is in small rounded grains that cloud certain areas or form patchy aggregates through the chert, or is in large grains forming the unreplaced shells. The calcedony, when present, is in thin, fibrous, wavy bands that permeate part of the chert, lining shells, and microscopic cavities. ‘The presence of the opal is inferred from the presence in part of the chert of much isotropic material with a moderately low index of refraction. ‘The dolomite, where observed, was in small rhombs scattered through the chert and was distinctly MISSISSIPPIAN CHERT OF ST. LOUIS AREA 265 more abundant in the chert than in the surrounding limestone. Pyrite was present in only a few cases and was in small cubes in the center of quartz-filled cavities. Im much of the chert there is a slight amount of iron staining present. The structure of the chert as revealed under the microscope varies considerably. Part of the chert is massively composed of very fine grains, whose boundaries cannot be made out. Much of the chert is similarly composed in large part, with ramifying, banded microscopic masses of calcedonic material or interlocking quartz grains, which seemed to have filled pre-existing cavities. The interior cavities of shells are in most cases filled with inter- locking quartz grains. The shells, even very minute ones, for the most part are composed of calcite in medium-sized grains, but in some cases show partial or complete replacement. Where replace- ment of a shell has taken place it is mostly by fine allotriomorphic quartz granules. Concentric banding is shown by much of the chert and is caused in some cases by very slight differences in the amount of staining, probably ferruginous, and in other cases by a variation in the amount of admixed calcite grains, the calcite rich areas tending to a chalky white color. In a chert much resembling in marking the mottled Ordovician limestones the mottling is likewise due to a rapid variation of the proportions of calcite to quartz grains. Banding due to stratification is present in some of the chert and is shown by the orientation of minute shells, by variation in the amounts of a cloud of fine black specks, and by variation in the staining. RELATION OF CHERT TO SURROUNDING LIMESTONE The character of the contact of the chert and the limestone apparently varies with the different chert beds. In many cases there is a visible transition extending over a zone of one to two centimeters. This is particularly the case with the spherical nodules of the St. Louis (Fig. 2). But the contact more commonly is sharp, at least under megascopic examination. In the Osage chert of Iowa such contacts are reported by Van Tuyl microscopi- cally to show gradual transition. In the thin sections of St. Louis 366 DONALD C. BARTON chert examined the transition was distinctly sharp, being confined to a zone of 0.2 to o.4 millemeters in width. The lateral contacts of the larger, semispherical nodules can be seen in many cases to be zones of slight displacement. The lines of stratification running into or across the chert are broken and slightly displaced upward, the more so toward the top of the nodule, and not at all or only faintly in reverse at the base. The ends of the line of stratification in the adjoining limestone in many cases tl HANG /, ra ees Fic. 2.—Chert from the St. Louis limestone, St. Louis, Missouri, under the high power: a, chalcedony; 6, dolomite rhombs; c, limonite stain; d, granular quartz; é, silicified shell; f, microgranular groundmass. are bent up near the nodule with the upper ones arching over, but in other cases run to the contact without deviation. Slickensides were found in a very few cases on the lateral contacts, showing relative movement upward of the chert of the nodule. CHARACTER OF THE ROCK INCLOSING THE CHERT Although chert is confined to calcareous rocks, the exact char- acter of the rock in which chert appears varies widely. The grain seems to have no effect on the presence of chert, which is found MISSTISSIPPIAN CHERT OF ST. LOUIS AREA 367 indifferently in fine-grained, even lithographic limestones, and in coarse-grained ones. It is found in massively bedded as well as in thin-bedded limestones. It is not found, to the writer’s knowledge, in pure or only slightly calcareous sandstones or shales, but is found in arenaceous and argillaceous limestones. It is found in highly magnesium limestones and in very pure limestones. A series of analyses of the chert-bearing St. Louis and Burlington-Keokuk limestones at St. Louis show a variation in composition as given in Table I. TABLE I* Limestone SHE Recdde PnOnaE CaCOs MgCO; SP yIEOUISS ae ee oss 1.48-9.56 0.35-1.82 61.88-94.97 | 0.94-24.53 Burlington-Keokuk....| 1.10-4.35 ©.40-1.82 77.95-94.50 | 3.18-14.84 *Anaiyses by A. E. Atwood, Geol. Surv. of Missouri, Bull. No. 3 (1890), p. 77. CONTEMPORANEOUS CHERT OF OTHER AREAS In areal distribution these Mississippian cherts are not restricted to the St. Louis area, but are widespread and are characteristic of the St. Louis limestone and equivalent formations and the Burlington-Keokuk limestone and equivalent formations practically wherever they are found. The Salem limestone of the St. Louis area is free from chert, as is also the Bedford odlite, its equivalent to the east. The exact extent of the distribution of the chert in the St. Louis limestone and the Burlington-Keokuk limestone is best shown graphically in Figs. 3 and 4, on which are plotted the outcrops of these formations and the areas in which they are chert-bearing. The correlation of formations on these maps is taken largely from B. Willis’ “Index to the Stratigraphy of North America,” U.S. Geol. Surv. Prof. Paper 71. While the morphology of these equivalent chert beds varies somewhat from locality to locality, yet there seems to be a greater or smaller constancy of habits of the chert of each formation. The St. Louis is characterized by even, ball-like chert even to Alabama, while the Lauderdale is spoken of as platelike, and the Boone chert and Grand Falls chert of western Missouri are said to be lenslike or sheetlike. It is perhaps worthy 368 DONALD C. BARTON Fic. 3.—Distribution of the chert in the Burlington-Keokuk limestone and equivalent formations. ; Fic. 4.—Distribution of the chert in the St. Louis limestone and equivalent formations. MISSTSSIPPIAN CHERT OF ST. LOUIS AREA 369 of note, although very possibly nothing more than a coincidence, that much chert is found in the lower Carboniferous limestones of England, Ireland, and Belgium. AGE OF CHERT In age the chert of the Mississippian of the St. Louis seems without doubt to lie between the early Pennsylvanian and the time of formation of the chert-bearing limestones. The evidence of the pre-Pennsylvanian age of the chert lies in the presence of the Mis- sissippian chert as clastic fragments in the base of the Pennsylva- nian. In a Missouri-Pacific railroad cut between Kirkwood and Barrett’s Station, St. Louis County, Missouri, and also in a Frisco railroad cut a half-mile east of Meramec Highlands, clastic chert carrying Lithostrotion proliferens and Spirifer Keokuk and showing predepositional weathering is to be found in abundance at the base of the Pennsylvanian shales. In Miller County, Missouri, clastic chert from the Burlington limestone is reported as being present in the Graydon sandstones and the Coal Measures shales. In the Joplin district likewise clastic Burlington chert is reported as lying at the base of the Pennsylvanian. The evidences of the formation of the chert later than the forma- tion of the containing limestone are several. The replacement of fossils has been cited and probably in some cases correctly so. Although the fossils in the great number of cases have not been replaced, the replacement of the fossils, especially the larger ones, when studied in their sections under a microscope seems, partly at least, surely to have taken place after the formation of the chert. The retention in the chert of the markings of the limestone, includ- ing the grain, stratification, styolithes, and fossiliferous character, is valid evidence of the secondary character of the chert. The mottling of the chert in very many cases accurately reproduces the appearance of the granular character of the limestone. The strati- fication markings and the variation in grain and fossil content in the different layers can in many cases be traced into or across the chert. The arching of the stratification and the faulting and slickensiding at the contacts of the nodules are also evidences of the secondary character of the chert. The stratification of the limestone in which 370 DONALD C. BARTON the pancake-like chert bands appear, for example, is more or less contorted, although in the closely associated, chert-free beds it is even and regular. The slight vertical displacement in the slicken- sides that are found at the lateral contacts of some of the larger nodules has already been mentioned. Yet while the chert seems definitely to be secondary, there is an aspect on its part of contemporaneity with the limestone. This is evidenced by the widespread development of the chert in the St. Louis limestone and in the Burlington-Keokuk limestone, although the intervening Salem limestone—Bedford odlite lime- stone—is practically chert-free. The constancy of habit of the chert in the St. Louis limestone and in the Burlington-Keokuk, - respectively, over a wide area has already been noted. The devel- opment of the chert also is parallel to the stratification. In a band of isolated nodules there is characteristically a striking alignment of the nodules at some level in a bed, often a massive one. There are in some cases several bands, and in a few cases there is no align- ment of the nodules, but where the bands are present they are parallel to the stratification. The nodule-containing bed is in some cases three or four feet thick, with no shale partings, is uniform in grain and character throughout, and shows no apparent cause for percolating waters, whether silica-bearing or not, to flow at certain definite and localized levels. The pancake-like chert masses and the chert lenses likewise show conformity to stratification, although in this latter case there is in many places coalescence of several lenses by lateral thickening. As far as could be seen there was no possibility of localization of the chert at definite levels by control of percolating waters by shale partings or such. Ii the chert were purely epigenetic, it would seem probable that the chert bands would show some tendency to cut across the stratification. The various theories that have been proposed to explain the origin of chert may be said in essence to be six. I. The silica is of organic origin, derived chiefly from the spicules of siliceous sponges. The silica may be derived from other of the siliceous organisms, as, for instance, in the case of the radiolarian cherts of California. MISSISSIPPIAN CHERT OF ST. LOUIS AREA Biya: t. The chert is supposed to be derived from colloidal silica which formed from the decomposition of siliceous sponges and collected in the depressions of the sea floor. The chert bands are supposed to represent former sponge beds, where the sponge remained accu- mulated in place over a considerable area. 2. The chert is supposed to form before consolidation of the limestone through the solution of scattered siliceous spicules and the almost immediate replacement of parts of the limestone. 3. The chert is supposed to form after the consolidation of the limestone through the solution by percolating waters of the siliceous spicules and the replacement of part of the limestone by this dis- solved silica. II. The silica is supposed to be of inorganic origin. 4. The chert is supposed to form by the precipitation of silica and the replacement of the limestone in the presence of circulating waters which have passed through sandstones, arenaceous rocks, or rocks containing silicates. 5. The chert is supposed to result from the reaction of the dis- solved silica of sea-water with the limestones, with the consequent precipitation of the silica and with possibly a later concentration. 6. The chert results from the diffusion of silica in solution through a limestone. The concentration will vary in the direction of the diffusion, and the deposition resulting when the concentration is sufficient will be in zones perpendicular to the direction of diffusion. As the conditions for diffusion are more favorable in the early days of the consolidation and as the most likely direction for the diffusion is upward toward the surface or downward from it, the deposition will be parallel to the stratification, although inde- pendent of it. _The development of the chert in successive zones is due to the lowering of the concentration immediately around the first started zone or zones of crystalizing material. The silica may be derived from organic or inorganic sources. ORIGIN OF THE CHERT OF THE ST. LOUIS AREA The source of the silica of the chert of the St. Louis area is not as clear as in the cases of the English cherts and flints, the radio- larian cherts of California, and the cretaceous cherts of Texas, and 372 DONALD C. BARTON apparently must remain a matter of conjecture. Evidence of an organic origin is wanting. In slides of the St. Louis chert spicules were present only in one case. In his study of slides of the Mis- souri cherts Hovey reports that he found only one carrying sponge spicules. Likewise Van Tuy] in his study of the cherts of the Osage series found sponge spicules present in only one sample. Of the presence in numbers in the Mississippian seas of other silica- secreting organisms nothing is known. In the case of the cherts of the St. Louis area the theory that the chert is formed from the collection of colloidal silica on the sea floor is not applicable, since the chert is plainly secondary. Hinde’s application of this theory to explain the presence of unsilicified shells in the chert is not necéssary, as the differential replacement is readily accounted for by the lower solubility of.the shell mate- rial. In this connection a brief series of tests was run on the relative solubility of recent and fossil pelecypod shells, fossil bra- chiopod shells, on crinoid stems, and on chert-bearing and chert-free limestone. The material was powdered to pass through a two- hundred-mesh sieve and was digested in 25 cubic centimeters one- half normal HCl, plus 350 cubic centimeters distilled water, and the time required for neutralization, as shown by methol orange, was noted. A marked tendency was shown toward less solubility on the part of the shells, crinoidal limestone, and chert-free lime- stones, such as the Salem, and greater solubility on the part of the limestones associated with chert. The experiments were not extended enough to be conclusive. That the chert was formed before the consolidation of the lime- stone from silica derived from the solution of siliceous spicules or tests is possible in the case of some of the chert. But it is definitely not possible in the case of most of the chert, as the chert did not form until after the limestone had acquired its granular character. The formation of the chert in a similar manner during, or later than, the consolidation is more possible. The chert is secondary and is pre-Pennsylvanian, and therefore must have formed dur- ing, or not long after, the consolidation of the limestone. There is, however, no positive evidence of the organic origin of the silica. The suggestion that the silica of the chert is exotic and that it has MISSISSIPPIAN CHERT OF ST. LOUIS AREA 373 been introduced from other siliceous formations by the underground circulation, possibly that of the geologic present, does not seem valid. The formation of the chert cannot have taken place through the agency of the present-day circulation, since the presence of the chert at the base of the Pennsylvanian shows the period of forma- tion to have been late Mississippian or early Pennsylvanian. The jasperoid of the Joplin district, furthermore, is a chertlike siliceous deposit that is said to have been deposited by the same under- ground circulation that is responsible for the mineralization of the region. But the jasperoid is later than the chert and distinctly different. The more serious objection, however, is the conformity of the chert with the stratification. Vertical zones following the joints are not found. The chert is found widespread, but is not found in an adjacent formation and is more or less similar over wide areas, but different in aspect in the different formations. The derivation of the silica of the chert through precipitation from sea-water is a possibility. Silica is precipitated from solution by calcite and replaces it when H,CO, is present, and, as the accumulating sediments of the ocean bottom usually contain decaying organic matter, H,CO, should be present. A tenth of the yearly increment of saline material in the ocean is silica, but the silica content of sea-water is practically nil, 1 part in 220 to 460 thousand. The very considerable annual increment of silica must therefore be removed quickly, either by direct chemical precipita- tion or through the action of organic agents. In the case of the Mississippian beds of the St. Louis area it is not known that sili- ceous organisms were present to any important extent at the place and time of the deposition of the beds. The degree of concentra- ° tion of the silica in the ocean-water, in connection with the slow rate of diffusion and proximity or distance from the mouth of a river, may be an important factor, contributing to the lack of chert in some limestones, as, for instance, the Salem and Kemms- wick limestones. Such chert-free limestones may have formed at a distance from the mouths of rivers, and the silica may have been completely precipitated before currents brought these waters to the place of deposition of these beds. 374 DONALD C. BARTON The principles of diffusion as given by Liebesang and Cole in connection with the origin of flint partially explain some of the features of the chert of the St. Louis area. These principles seem- ingly explain the formation of the chert after, but not long after, the formation of the limestone—the position of the chert parallel to the stratification but independent of it, the rhythmic deposition of the chert, and the excessive development of the chert within a few hundred feet of a great uncomformity. They are equally applicable whether the silica is derived from organic or inorganic sources, and would seem to necessitate a rather general distribution at the start of siliceous material through the mass. The particular localization at a given horizon of a chert bed might be affected, however, through the influence of a local excess of siliceous material on the concentration at that horizon. The localization might also be affected by the solubility of the limestone of the various horizons. if SSS epalce ren formation; FE Red and winle ) «y., PEIEe| SEDIMENTA. l6nfous Rocks Jookcarak basan eal rerrerae kaseealick Lake EASTWARD To THE SEA Abal imestone STRUCTURE SECTION ACROSS A PORTION OF THE BELCHER /SLANDS FROM Y Scale: (erizentally: /-inch=1-Mile ROCKS hoe I Silicified and shal ET Ee Graywacke and slate {1 Te stone While To gre fi Peat alee 47 2.—Geological structure section across a portion of the Belcher Islands Fic. 418 E. S. MOORE more rarely chalcopyrite is found. In one large sill a mass of impure dolomite has been inclosed, partly altered to tale and serpentine and impregnated with chalcopyrite. There is not sufficient copper, however, to be of economic interest. The occurrence of cobalt in a calcite vein in one of these sills is interesting because of its frequent association with the Keweena-_ wan diabases in other parts of Canada. A narrow vein of calcite’ was found cutting the basalt and carrying smaltite, chalcopyrite, magnetite, actinolite, and specular hematite. The relation between the igneous rocks and the sediments is not always clear. In most cases it is quite evident that the former are distinctly later than the latter, but in the case of the large mass of diabase and basalt forming the backbone of Tookcarak Island the evidence is inconclusive. It would appear that the sediments above it have been metamorphosed to some extent, but this cannot be proved to be the result of contact action. There are some amygdules of chlorite in its upper portion, but no sign was seen of ellipsoids, so common in some of the distinct flows in the area. Although amygdules are usually considered evidence of extrusion, they cannot be taken as definite evidence of such origin because vesicular dikes and sills are known which must have solidified thousands of feet below the surface. The question of whether an intrusive rock will be amygdaloidal or not depends chiefly upon the porosity of the rock adjoining it, and on the amount and pressure of the gases which it contains. Leith regards some of the sheets in the Richmond Gulf region as flows, but so far the writer has not found any of these interbedded igneous masses which he is sure are flows. There is one great extrusion, however, which is the youngest consolidated rock seen on the Belcher Islands, and which apparently at one time spread over the whole of the islands. It seems probable that it also extended to the Manitounick Islands near the coast of the bay. It immediately overlies the main band of iron-formation at almost every point where it outcrops. Its thickness is uncertain, since the flow is largely under water, but it is doubtless several hundred feet. Near Kasegalick Lake it is about 30 feet, but this does not represent its maximum thickness, since it has suffered much from erosion (Figs. 3, 4,5). This extru- THE ITRON-FORMATION ON BELCHER ISLANDS 419 sive mass consists of more than one flow, because in some places the surface of contact between two flows may be traced for a long distance (Fig. 6). Ellipsoidal and amygdaloidal structures are Fic. 3.—View of Innetallung Island showing in the foreground the limestones and quartzites and in the distance the white quartzite overlain by the iron-formation and it in turn by the great basalt flow. 5 aS Fic. 4.—The great basalt flow overlying cherty shales and jaspilite on the shore of Kasegalick Lake. 420 E. S. MOORE well developed at almost every place where the surface of the basalt is exposed (Fig. 7). Outside of these two bodies all the igneous rock seen showed definite evidence of intrusive origin. There does not appear to be any definite relation between the origin of the iron-formation and these igneous rocks, because it does not seem to matter whether they intrude it or are flows over- lying it, whether they are close to it or are far removed from it. There is in many places a little micaceous and specular hematite near the contact between the diabase and the adjacent sediments, but this seems to be independent of the original iron-formation. Fic. 5.—Large diabase dike cutting the jaspilite near Kasegalick Lake. It apparently served as a feeder for the large flow shown in Fig. 4, since the flow lies over the jaspilite just beyond the upper left-hand corner of the picture. THE ALGAL CONCRETIONARY LIMESTONES In recent years much attention has been paid to the study of the minute organisms which play an important réle as precipitating agents in calcareous and iron-bearing solutions. It has been proved that low forms of piants, chiefly the algae, are at the present day causing to be precipitated great quantities of calcium carbonate in streams, lakes, and seas, and that the iron bacteria are respon- sible for the deposition of much iron in bogs and other bodies of water. Dr. Walcott set a new record when he described the numerous algal structures from the pre-Cambrian rocks of Mon- THE I[RON-FORMATION ON BELCHER ISLANDS 421 tana,’ and it now seems certain that these low forms of life flour- ished well back into pre-Cambrian time, and that they were Fic. 6.—The great basalt flows forming Keepalloo Peninsula. The contact between two thick flows may be seen near the foot of the hump in the center of the photograph. Fic. 7.—Ellipsoidal structure which is very common in the basalt flows responsible for the precipitation of large bodies of calcareous matter whose origin was formerly a matter of doubt. tC. D. Walcott, ‘‘Pre-Cambrian Algonkian Algal Flora,” Smithsonian Miscel- laneous Collections, Publication No. 2271 (July 22, 1914), pp. 77-150, Pls. 4-23. 4227) E. S. MOORE On visiting the Belcher Islands the writer was impressed by the extraordinary development of structures which seemed to resemble so strongly the cryptozoons of the Upper Cambrian that they were at first regarded as species of that fossil. However, there seemed to be some marked differences between the two types, and the Belcher Islands specimens should be regarded as deposits made by a new group of algae. These fossils have an important bearing on the age of the associated rocks, because if they be regarded as cryptozoons we must either change the generally accepted Fic. 8.—‘‘Ribboned”’ ferruginous shale more highly silicified in alternate layers, Tookcarak Island. conclusion that these rocks are pre-Cambrian or we must push the cryptozoons back into the pre-Cambrian. There is no definite evidence that the sediments on the east coast of Hudson Bay are pre-Cambrian, but there is as much evidence as there is for the age determination of most of our pre-Cambrian rocks of Northern Canada. Leith™ points out the remarkable resemblance between them and the Animikie of the Lake Superior region, and Low,’ after considering them as Cambrian and then as Laurentian, finally concluded that they were more like the later pre-Cambrian. LC. Ke Leith ops cite 1p 233. 2A. P. Low, op. cit. THE IRON-FORMATION ON BELCHER ISLANDS 423 Low had observed the concretionary structures mentioned above, and he states regarding the rocks associated with them: ‘“‘No fossils have as yet been discovered in any of the beds of this formation, but the presence of certain concretionary forms in its limestones and the amount of carbon in many of the shales lead to the belief that at least low forms of life existed at the time these rocks were deposited.’* Leith also mentions them, in the Rich- mond Gulf region, as follows: ‘The limestone floor has great con- cretionary structures up to two feet in diameter, the sides of the concretions locally open on one side and locally opening out into waved and crenulated bedding lines, interpretation of which the writer does not attempt, but which probably tell of conditions which if known would indicate the depth of water and other significant conditions of the r formation.’” On the Belcher Islands these bodies form whole reefs in the more or less silicified limestone of the Belcher series, making up a thick- ness of over 400 feet (Figs. 9, 10, 11, 12). There appear to be two types, one smaller than the other, but as to whether these should be classed as distinct forms or simply regarded as concretions in different stages of maturity is not yet settled. The fact that the smaller ones form a large, reeflike mass near the top of the algal limestone and near the line where the rocks change in character suggests that they are incompletely developed specimens of the larger type. Both types are spherical to subspherical bodies con- sisting of concentric layers, and they vary in size from an inch to over fifteen inches in diameter. The larger type is seldom less than four inches, and the smaller seldom more than four inches, in diam- eter. The larger ones are more regular in form, being much more nearly spherical than the smaller ones, and they also show the con- centric lines of growth more distinctly than the smaller ones, which show a tendency to be rather disk-shaped. The distinctly crenu- lated character of Cryptozoon proliferum is not found in many of these specimens. They are probably more like Cryptozoon steeli. The concretions consist chiefly of calcium carbonate and can be almost entirely dissolved in cold hydrochloric acid. There is, 1A. P. Low, Geol. Sur. of Canada Ann. Rept., XIII, Part D, p. 46D. 2C. K. Leith, op. cit., p. 240. 424 E. S. MOORE however, some silica in small grains of various sizes, some chal- cedony infiltrated along the lines between the bodies, and in some Fic. 9.—The algal concretionary limestones on Tookcarak Island. The largest concretion is about 15 inches in diameter. Fic. ro.—Algal concretionary limestone on Tookcarak Island. The largest con- cretion is 14 inches in diameter. places a little carbon may be found under the microscope, especially along the surface separating the concentric rings of growth. This THE IRON-FORMATION ON BELCHER ISLANDS 425 carbon is left on dissolving the rock in cold or hot acid. No definite cell structure can be recognized in this carbon, but in Fic. 12.—The smaller type of algal concretions in the limestone on Tookcarak Island. some of the granules in the calcareous beds of the iron-formation, a little higher up in the series, grains of iron oxide, now replac- ing the calcareous matter, show such a distribution, size, and 426 E. S. MOORE arrangement that they seem to indicate the replacement of organic cells. That these larger concretions are the result of the action of algae, which cause precipitation of calcium carbonate owing to the chemical changes produced in the water by them, there seems to be little doubt. They are very similar to the concretions described from the Algonkian rocks of Montana by Walcott, and the cal- careous concretions described by Roddy? as now forming in the streams of Lancaster County, Pennsylvania (Fig. 13). Numerous Fic. 13.—Recent algal concretion collected by J. Roddy in Conestoga Creek. Lancaster County, Pennsylvania (2.5 inches in diameter). other occurrences of similar deposits have been described by other writers. The two types, large and small, bear certain resemblances to Walcott’s Newlandia concentrica and Collenia? frequens, but they do not seem to be identical with them, and new generic and specific names should be applied. As stated above, no definite organic cell structure has so far been recognized in these large concretions, but the replacement ™C. D. Walcott, op. cit. 2H. J. Roddy, ‘‘Concretions in Streams Formed by the Agency of Blue-Green Algae and Related Plants,” Proc. Amer. Phil. Soc., LIV, No. 218 (August, 1915), pp. 246-58. THE IRON-FORMATION ON BELCHER ISLANDS 427 of calcium carbonate by iron oxide in some small concretions in this same series of rocks was so suggestive that the attention of Dr. J. Ben Hill, of the department of botany, Pennsylvania State College, was called to them. Dr. Hill very kindly examined them, prepared the accompanying camera lucida sketches (Fig. 14), and the follow- ing statement regarding these bodies: The specimens in question are generally smaller than the living Cyano- phyceae, but are not smaller than the smallest of the living species. In fact, the most striking specimens are well above the lower range of size in the living species. The measurements of the objects in the rocks range from o.5 to C20 ae 5 ore 2.5 microns for the smaller forms to : B oe 5 and ro microns for the larger. This ~<— oe oe is exclusive of the very minute and . S&B & ae the very large ones. The shapes of : the objects would suggest species in (She, WicCannaeilinstidin Glenines ber the two classes (Coccogoneae and J. B. Hill of grains of iron oxide resem- Hormogoneae). The isolated spher- pjing replacements of algal cell struc- ical forms resemble some genera of turés. The accompanying scale indicates the Coccogoneae as Chrococcus or the diameter of these bodies. Gleocopsa, and the specimens showing cell-like structures connected to form a filament, some genera of the Hormog- oneae as Nostoc or Anoheana. Although Dr. Hill and the writer both recognized the great difficulty in distinguishing small mineral grains, which often form strings and bunches, from organic structures we feel satisfied that the sizes, shapes, and arrangement are too regular to be the result of simple replacement without some original organic control. A further discussion of these smaller concretions will be found in the section on the iron-formation. The photographs, thin sections, and two specimens of the con- cretions were later sent to Dr. M. A. Howe, of the New York Botanical Gardens, and to him the writer is indebted for his kind- ness in making an examination of these materials. Regarding them Dr. Howe makes the following statement: This Hudson Bay limestone is of obviously organic origin, and the organ- isms contributing to its upbuilding are, it seems to me, in all probability of a vegetal and algal rather than animal nature, though the microscopic structure 428 E. S. MOORE shows little or nothing that would justify a student of the recent algae in refer- ring them to a modern genus, family, or class. I infer that the main organism deserves comparison with Cryptozoon proliferum Hall from Saratoga, New York, and one photograph is a bit suggestive of Walcott’s Collenia? frequens as shown in his Plate 1o, Fig. 3. Unless more definite structure is revealed by future sectioning it seems to me that about as far as we can safely go with these Hudson Bay fossils is to say that they are probable algae. Another structural feature in the limestones of the Belcher series, and one which may also depend upon organic agencies for its origin, is a remarkably regular and uniform banding, which Fo ig Fic. 15.—Very distinctly banded limestone on Innetallung Island. It is believed that the regularity of this banding may be due in some way to the action of algae. is due to alternating very fine-grained and coarser-grained layers of limestone, more or less silicified, especially along the bedding planes (Fig. 15). As seen under the microscope some very small fragments of twinned feldspar scattered through the fine-grained layers and grains of quartz are fairly common. In the hand speci- men the rock looks more like a cherty quartzite than a limestone, it is so dense and fine-grained. ‘The alternate bands weather more rapidly, producing a ribbed effect. The bands are often extremely regular for considerable distances and usually run from one-half inch to three-quarters of an inch in width. It is suggested that this banding may be due to the seasonal work of low forms of life THE IRON-FORMATION ON BELCHER ISLANDS 429 causing variations in the nature of the deposits made during the different seasons of the year. THE KEEPALLOO IRON-FORMATION The term iron-formation is used here as in other writings on pre-Cambrian geology for a group of rocks which vary considerably in composition, but which together contain conspicuously more iron than their associated rocks, and which by natural concentration processes are capable of giving rise to iron-ore deposits. There is a large body of this formation on the Belcher Islands, and the term Keepalloo has been applied as a local name for it, since it is so well exposed on the peninsula along Keepalloo Sound. It consists of jaspilite, iron carbonate, calcite, probably iron- magnesium carbonate, hematite, magnetite, chert, and greenalite. A section on Keepalloo Peninsula from the quartzite up to the basalt is as follows: 2 2 OG FEET a) A mixture of cherty, sandy, jaspery, calcareous granular rock with bandssombprownish=weatnering shale yy eienane 4 Seen ai rs ley Ese Dwikeddishvands brownish nissilersialenie: er vias ee le ays ce ne lee tare 17 GMa ASP ile gna Neem Rn Nees ne MNS Hy aU IRC LRT yas Nd A ME eg 30 @) aspilite withibands of hematite Orel so: uu se eeiie eis Selec a 39 ON, VERSE OMIT ESS Sai sei)s EAL I ea a AC EAN 2c Maa 46 MEN aspcrancdibandsjomleanwhemativterores (ui). 1 ov eyae Wat selec 4): 10 g) Dull, shaly jasper with bands of bright-red jasper, resembling a felsite Ande concalnlMenculbes Om py TILE. oer cuca clea see sigur ete ae see 43 PAR CoybeN| Ves Ce LN oo at Sar RES CO iS a ee cen RUE MR tS 239 In many of these rocks small granules may be recognized with the naked eye, and they occur mostly in those rocks poorer in iron, lying near the top and bottom of the iron-formation. Their greater abundance in these rocks is no doubt due to the fact that the iron oxides lend themselves less favorably to the preservation’ of such structures than silica or carbonates, where there is considerable concentration of iron, and also to the fact that when a great deal of replacement occurs original structures are likely to be lost. A number of thin sections from these rocks were studied, and it was found that in nearly all cases the rocks are made up of granules of various types. A thin section of the red jasper over- lying the quartzite and lying near the base of the iron-formation 430 E. S. MOORE showed dirty-gray granules of a variety of shapes having a maximum size of 1.90 by 1.0 millimeters. They consist of calcite, chert, and iron oxide, the latter as a rule distributed in fine specks through the granules in an unusual manner indicating a probable replace- ment of organic tissues. It was in this section that the minute algae-like cells, previously described, were found (Fig. 16). Numer- ous veinlets of quartz following lines of fracture-cleavage indicate the extensive transfer of silica since these rocks were consolidated. These granules indicate a replacement of calcareous granules, in most places at least. Fic. 16.—Photomicrograph of granules from the iron formation. These consist of hematite, calcite, and silica, and the groundmass is mostly silica. From these granules the camera-lucida drawings were prepared (X 20). Fic. 17.—Photomicrograph of granules consisting chiefly of silica in which some hematite and magnetite occur, the lighter areas being silica (X20). From the cherty iron formation. Similar granules in a cherty matrix and carrying considerable magnetite were found in the rock described under g) in the section and spoken of as felsitic in appearance. Another specimen of red jasper with specks of gray opal-like silica was taken from one of the bright-red bands in the iron- formation. It was found to consist of granules of fine-grained chert, opal, and iron oxide. They show a great variety of shapes varying from ovoid, balloon-shaped, ham-shaped, and roughly THE IRON-FORMATION ON BELCHER ISLANDS 431 triangular to nearly spherical. A few are long and narrow and some are curved (Fig. 17). One measured 1.70 millimeters in diameter. Some granules consist of opal, in others the opal is changing to chert by loss of water, and the chert is in turn changing to granules of quartz by crystallization. Some granules consist chiefly of magnetite, some of hematite, and in others a smattering of both occurs throughout the silica. In a number of granules consisting chiefly of iron oxide the grains of silica are arranged in groups, so that they produce under the microscope a cell-like structure strongly Fic. 18.—A. Photomicrograph showing the hematite and greenalite granules with grains of silica distributed in a cell-like arrangement. The dark concretions are apple green to red and brown. The lighter areas are silica (X20). From the cherty iron-formation. B. From the same specimen as A. resembling what would result if a fragment of a bryozoan were replaced by iron and silica. Its regularity in so many granules is suggestive. The most interesting specimen was taken from the siliceous shales in the upper part of the iron-formation, where they underlie the basalt on Kasegalick Lake. In the hand specimen it is a dense, grayish-black to light-gray, cherty rock containing dark, cherty grains. It weathers to a dark, brownish mass. Under the micro- scope it is seen to be made up almost entirely of granules of various shapes, colors, and sizes (Fig. 18, A and B). In shape they are 432 E. S. MOORE similar to those described in the last section, ovoid, ham-shaped, irregular, and rarely spherical. In size they vary from o.14 to o.gt millimeter in diameter. They are colorless, reddish brown to deep brown, and apple green. ‘The colorless ones consist of very finely granular silica with a matrix of chert and chalcedony. A great deal of calcite is distributed through the section, and one small granule consists entirely of calcite, thus suggesting that all the granules may have been calcite originally. The green granules vary from apple green through brownish green to dark brown, depending upon the amount of iron oxide which has developed by alteration. They show very little double refraction and no pleochroism except in some places where they are altered to little rosettes of extremely small radiating needles of what is apparently an amphibole, with higher bi-refringence than the chlorites, but lower than actinolite or griinerite. Some granules are largely altered to magnetite, and in others rhombs of limonite. indicate the change of siderite to limonite. In many of them the same cell-like arrangement of the quartz grains mentioned in the last section may be seen. : It seems evident that these green granules consist of iron silicate, and of the silicates, thuringite, chamosite, and greenalite commonly found in iron-ore deposits, the characters correspond most nearly to Van Hise and Leith’s description of the greenalite granules of the Mesabi Range." ORIGIN OF THE IRON-FORMATION From the descriptions given above it is evident that we have on the Belcher Islands an unusual development of large concre- tions whose origin can only be attributed to organic processes. We have further hundreds of feet of rocks consisting chiefly of minute granules, some of which at least show good evidence of being of organic origin from the widespread occurrence of apparent traces of plant remains in them. If these granules be compared with certain siliceous granules in the Upper Cambrian limestones of central Pennsylvania, which grade into typical odlites, it will ‘Van Hise and Leith, ‘“‘Geology of the Lake Superior Region,” U.S. Geol. Sur. Mon., LIL (1911), 165. THE ITRON-FORMATION ON BELCHER ISLANDS 433 be seen that they are very similar and that these Upper Cambrian odlites are associated with abundant cryptozoon fossils. Further, if the various types of odlitic iron ores from the Clinton and Cam- brian of America, from the Silurian of Europe, and from the Jurassic of France and England be examined, similar bodies will be found. In fact, so far as the writer’s experience goes with slides from the iron-bearing rocks of the formations later than the pre-Cambrian, such concretionary bodies are only found in those rocks which con- tain other evidence of organic action. Although the writer recognizes that low forms of life, such as algae and iron bacteria, are not essential to the formation of odlites* and related concretions in all cases, it seems probable that they generally serve as the agents which produce the chemical changes causing precipitation of the calcium carbonate and the iron. This action seems to be due chiefly to the removal of the carbon dioxide from the acid carbonates by the algae, and the oxidation of the iron by the iron bacteria, thus in both instances producing insoluble compounds. The occurrence of vast deposits of odlites and related concretions in certain geological formations, in some instances on different continents during the same period, is also suggestive of the influence of certain organisms which reached a high stage of develop- ment at that particular time and caused the precipitation of the iron or other salts. It may also be due to the particular conditions of erosion, which permitted a large amount of any particular kind of salt to be carried to the sea during a certain geological period. In his monumental works on the origin of the pre-Cambrian iron ores in various parts of the continent Dr. Leith has held firmly to the belief that the bulk of the iron has been supplied directly to the sea as magmatic waters accompanying the great eruptions of basic igneous rock. He advocates this theory very strongly for the iron-formations on the east coast of Hudson Bay.? Although the writer readily recognizes the possibility of supplies of iron salts from this source, he cannot see that they have played a rdle at all comparable to the supplies carried into the sea, or into inland « An additional note on ‘‘The Oolitic and Pisolitic Barite from the Saratoga Oil Field, Texas,” Science, N.S., XLVI (October 5, 1917), 342. 2121s Leithops cit pa ean. 434 E. S. MOORE bodies of water, in which it seems probable that many of our pre- Cambrian deposits may have been laid down by processes of weathering. The largest known deposits of high-grade iron ore in the world, the Brazilian deposits, do not show direct relation to igneous rocks.t The great deposits of Lorraine, the Jurassic deposits of England, and our own Clinton ores show no direct relation to igneous eruptions. Going farther back into the pre- Cambrian rocks, it will be found that the greatest deposits of all, those of the Upper Huronian, show far less direct association with the basic igneous rocks than the smaller deposits of the Keewatin. This may be due to a large extent to the conditions of drainage, which must have been much better developed in the Huronian than in the Keewatin, if we can judge from the topographic features which are likely to have been produced during such a volcanic period as the Keewatin, and from the rocks which we now find making up the Keewatin series.?, There would be a tendency to deposit small and isolated bodies of iron-formation in the Keewatin, which later, on erosion, might add materially to the Huronian deposits. The problem of transportation of the silica and iron has always been a big one unless we invoke the aid of hot water and magmatic solutions. However, the work done on colloids in recent years has aided us materially toward a solution of this problem. It has been recognized by Lacroix that colloids are an important product in the weathering actions which produce laterites, and the authors of a recent paper on the origin of the Missouri cherts state that, so far as they know, silica is transported only in the colloidal form and not as a sodium silicate, since such a form dissociates to form colloidal silica. The silica set free from the decomposition of basic rocks would be almost entirely derived from the silicates, and it might be retained readily and carried in the colloidal form. The iron would « —. C. Harder and R. T. Chamberlin, “‘ Geology of Central Minas, Geras, Brazil,”’ Jour. of Geol., XXIII, 358-62, 385-404. 2 In the recent volcanics on the island of Hawaii may be seen almost a complete imitation of the topographic features of certain uncovered Keewatin igneous areas. 3G. H. Cox, R. S. Dean, and V. H. Gottschalk, Studies on the Origin of Missouri Cherts and Zinc Ores. Bull. 2, Vol. III, School of Mines and Metallurgy, University of Missouri. THE IRON-FORMATION ON BELCHER ISLANDS 435 also be in the ferrous form and easily transported. It has generally been considered that practically all the silica in these pre-Cambrian iron-fermations has been carried in solution, and that very little of it has been clastic sediment. Experience with many different areas of these rocks shows that there is almost invariably a great deal of quartzite, arkose, or graywacke, distinct products of weather- ing, associated with the iron-formation, and that these often grade into the jaspers. The distinctly clastic sediments cease, and the cryptocrystalline forms of silica take their places. It scarcely seems reasonable that the deposition of clastic siliceous sediment should be so suddenly cut off in all cases and its place taken by chemical precipitates without a great deal of silt being deposited with the chemical precipitates. While this clastic material cannot now be identified in the jasper and chert, it seems probable that it is there, but indistinguishable because of metamorphism from the finely crystallized silica which makes up the bulk of all these jaspilite formations. Regarding the adequacy of the weathering processes to produce these deposits one has but to observe the great deposits of lateritic iron which have formed, and are continuing to form, in Cuba, India, and other warm countries to be convinced of the efficiency of. the weathering process. It is evident that the weathering of ~ iron-bearing rocks is almost constantly in operation, but it is owing to certain chemical and drainage conditions that the iron remains on the land as laterite and is not carried off to the sea or to other bodies of water. The chemical conditions depend upon two important factors, -one being the presence or absence of suitable solvents for the iron and the other the presence or absence of suitable precipitating agents which may throw the iron out of solution before it reaches the sea. That considerable iron which is left on the surface as a lateritic deposit and later washed to lower levels as a detrital deposit is carried in solution is evident from the fact that some of it takes on the concretionary form after being transported from its original location. From a consideration of the laterites the writer believes that the weathering of the basic igneous rocks would furnish plenty of iron to form the pre-Cambrian iron-formations, and that whether 436 E. S. MOORE the iron will be transported or left as a residual deposit will depend, not only upon the presence of solvents for the iron, but also upon the presence or absence of precipitating agencies. He agrees with Dr. Leith, however, in not regarding the pre-Cambrian iron- formations as laterites in sitw, although certain portions of them, especially those portions consisting largely of limonite or hematite, argillaceous materials, and silica, may very reasonably be regarded as lateritic, mechanical sediments more or less assorted. The constituents of the granular portions of the iron-formation and the iron carbonate must certainly have been carried in solution, probably as colloids, and through the aid of carbon dioxide and other agents. In the case of the ferric and siliceous granules in the Keepalloo iron-formation the presence of calcareous granules suggests that they were the primary granules, and that the iron and silica replaced them on the floor of the body of water in which these sediments were laid down. ‘This is the principle of deposition advocated by Cayeux" for some of the odlitic iron ores of France. There may also have been some primary iron-oxide and iron-silicate granules, as advocated by Hayes? for the Wabana ores of Newfoundland, and it seems probable that the concretionary character of the ore may be due to the action of low forms of life. The work of Harder’ and previous writers has shown that the iron bacteria are the important agents in precipitating iron compounds. These bacteria were found by Harder to be present in almost all iron-bearing waters, Spirophyllum and Gallionella, the latter, often mentioned among the algae by previous writers, being found even in underground workings of mines to a depth of several hundred feet. Harder found further that some solutions were kept under anaérobic condi- tions by passing carbon dioxide through them. In some solutions ferric hydroxide was precipitated, while in others there was no precipitate. The precipitation took place from either ferrous or ferric salts by oxidation. *L. Cayeux, Les minerais de fer odlithique de France. Ministere des Trav. Pub., Paris. 2 A. O. Hayes, ‘“‘Wabana Iron Ore of Newfoundland,” Can. Geol. Surv. Memoir 78, Ottawa, IgIs. 3 E. C. Harder, ‘‘Iron Bacteria,” Science, N.S., XLII, No. 1079, pp. 310-11. THE IRON-FORMATION ON BELCHER ISLANDS 437 In the article previously cited on the origin of the Missouri cherts* it has been demonstrated that the presence of carbon dioxide has a very important effect in precipitating colloidal silica in the presence of calcium carbonate, causing it to be thrown down very quickly. It would therefore appear that in the action of carbon dioxide on colloidal silica and on the processes of the iron bacteria we may have a clue to the cause of the distinct banding in some of our pre-Cambrian iron-formations, provided further studies of these rocks tend to show evidence of the wide distribution of plant life during pre-Cambrian time. The fact that living algae will furnish oxygen to the waters around them and when they decay give off a certain amount of carbon dioxide may cause some seasonal varia- tion in the precipitation of the iron and silica, and thus give rise to the banding in some of these rocks. The distinctness of the banding may later be increased by metamorphism with recrystalliza- tion of the minerals and a certain amount of transfer of materials among the bands under the influence of chemical affinity. SUMMARY The Belcher Islands, which lie about seventy miles from the southeast coast of Hudson Bay, have recently been brought to the attention of geologists through the discovery on them of large areas of iron-formation. The iron-formation forms part of a thick series of sediments consisting of limestones, shales, quartzites, and graywackes, and this series is intruded by sills and overlain by flows of diabase and basalt, making up a group of rocks which in many respects strongly resemble part of the Animikie and Keweenawan formations of the Lake Superior region. The lime- stone of this group is, however, very unusual, since it consists of concretions varying from one inch to over fifteen inches in diameter and so strongly resembling some of the modern concretions formed by blue-green algae that there seems to be little doubt that they are of algal origin. They bear some resemblance to Cryptozoon proliferum, but differ from that fossil too much to be placed in the same genus. Their abundance indicates the presence of vast numbers of low plants in the Hudson Bay basin in pre-Cambrian 1G. H. Cox, R. S. Dean, and V. H. Gottschalk, op. cit., pp. 9-10. 438 E. S. MOORE time, since it has been generally agreed among geologists who have seen the same series of rocks on the east coast of the bay that they are pre-Cambrian in age. The iron-formation consists of jasper, chert, hematite, mag- netite, siderite, and green granules regarded as the iron silicate, greenalite. The chert and hematite are also in concretionary form, and it is suggested that the algae and iron bacteria have been responsible for the precipitation of colloidal silica, hematite, and iron silicate in this granular form, in some places as a direct pre- cipitate on the floor of the basin and in others as a replacement of the calcite granules by the iron compounds. INTERNAL STRUCTURES OF IGNEOUS ROCKS; THEIR SIGNIFICANCE AND ORIGIN; WITH SPECIAL REFERENCE TO THE DULUTH GABBRO: FRANK F. GROUT University of Minnesota INTRODUCTION - It is commonly said that igneous rocks are structureless, or of massive structure, as distinct from stratified or banded rocks of other origin. When considered in detail, however, they are known to show a number of characteristic structures. Under special conditions igneous rocks develop lithophysae, orbicules, bunchy segregations, spherulites, etc. But besides these there are a num- ber of rock masses which show a banded structure. It is this band- ing which is the main subject of this paper, first as to its relation to the form of the rock mass, and later as to its origin. Three somewhat distinguishable features give a plane structure to.an igneous rock unaffected by metamorphism; they will be dis- cussed here as banding, sheeting, and fluxion structure. Various geologists have noted these structures and combinations of them under the terms bedded, stratiform, gneissic, laminated, foliated, trachytoid, schistose, linear, streaked, platy, schlieren, layers, benches, etc. Banding.—The banding noted in many igneous rocks is an alternation of mineralogically unlike layers or flat lenses (Figs. 1, 2,3, and 4). The dip and strike of the bands can be estimated in many cases, but may show minor undulations and bunches. In some cases the layers are all thin, but in others they range more widely, up to a hundred feet. The line of division between bands may be sharp or gradual. The texture of one band is in most cases very little different from the textures of adjacent bands, and the t Published by permission of the directors of the United States Geological Survey and the Minnesota Geological Survey. Appeared first as part of a thesis presented at Yale University. 439 440 FRANK F. GROUT minerals interlock across the contact. In most cases there is no great difference in the mineral constituents of the bands, but only Fic. 1.—The banded gabbro of Duluth, Minnesota. The banding in this outcrop is about as conspicuous as in the average. Fic. 2.—The bands in this gabbro outcrop are irregular and the color contrast is very slight, but the lighter bands are polished by glaciation. in the relative abundance of the minerals. The colors of adjacent bands may be only slightly different, or in some cases may show a strong contrast. In a rock mass containing a variety of minerals INTERNAL STRUCTURES OF IGNEOUS ROCKS 441 any one mineral may be quite completely segregated in certain bands. Where the minerals of any band weather more rapidly Fic. 3.—Faint banding in the Duluth gabbro. These bands curve slightly, and a white band near the hammer divides to the left. Fic. 4.—Conspicuous bands of peridotite and gabbro near the base of the Duluth gabbro. than those of adjacent bands, such bands appear as grooves in the surface. The composition of the bands is independent of the com- position of the wall rocks. There are no transverse dikes or con- nections between bands. 442 FRANK F. GROUT A classic example of banding is that in the gabbro mass of the Isle of Skye (1, 2).! There are many parallel layers of lighter and darker material, and some of the bands curve conspicuously. Another prominent case is that on Orn6é (3) just south of Stock- holm, where the alternate bands are black and white, and the banded rock is said to constitute the periphery of an intrusion. An equally notable color banding appears in the large igneous Ilimausak rock in Greenland (4). The bands are from one to three meters thick, and three main rock types alternate with remarkable regu- larity. ‘The bands are saucer shaped in a large way and there are no apophyses between bands. ‘Transition zones are narrow and the texture is unchanged at the contacts. The Laurentian gneisses have a banding that is in some places clearly an original igneous structure (5). Some bands pinch out, and all are notably different from the roof in composition. There are no sharp contacts and no transverse dikes, though some related pegmatites cut across the bands. Many papers on Canadian igneous tocks mention structures of this sort (6, 7,8). The banded rocks studied under the microscope show in the most positive manner that the structure developed while the rock was still molten, or at most only partly crystalline. There are in many specimens no traces of mineral deformation; nor is there any reason to suppose that recrystallization has obscured the signs of some previous deformation. Mount Johnson, near Montreal (9), shows bands rich in feldspathic material alternating with others richer in iron and magnesian constituents. The dip and strike can be measured. The alkali syenites of eastern Ontario (10) show bands. The Sudbury norite is reported by Mr.. Hugh Roberts, of Minneapolis, on the basis of recent exploration, to show an alternation of mineralogically differing bands. The Cortlandt series in New York has an “original gneissoid”’ structure in which the bands differ in mineral composition. While there are sharp contacts, it is characteristic that the grains in all cases interlock across the contact. None of the series exhibits any great amount of shearing (11). In the Adirondacks, bands one to one hundred feet thick show alternating gray and pink colors (12). 1 Numbers refer to entries in the bibliography at the end of the paper. INTERNAL STRUCTURES OF IGNEOUS ROCKS 443 In Maine some gabbro masses show alternating bands about two inches thick, in some of which segregation of feldspar produces light colors (13). The rocks of Lizard show a linear structure and an occasional distinct banding which is said to have nothing to do with dynamo- metamorphism (14). Such banded rocks are reported from the Himalayas (15), the Kola Peninsula (16), and the British Isles (17, 18). In the “‘Cottian sequence” (19) the banding has been supposed to be metamorphic, but there are some dikes with a folia- Fic. 5.—Apophyses of feldspathic Duluth gabbro into its traprock roof, east of Duluth Heights. tion parallel to their walls and at a high angle to the structure of the schist. The banding of the Duluth gabbro was long ago mentioned (20, 21), but new work has recently been done on the area by the geologists of the Minnesota Geological Survey. The structure is exposed in typical, as well as in some exceptional, conditions at the city of Duluth. The gabbro intrusion (Fig. 5) occurred after the accumulation of a great thickness of diabase and other. flows of the Keweenawan. It spread at or near the base of the flows, and along the unconformity at the base of the Keweenawan sediments a little below the flows. While the roof and floor are not an exactly continuous horizon, the transgression of a few hundred feet in a mass 444 FRANK F. GROUT a hundred miles long is insignificant. The relations are well exposed in the western part of Duluth. By detailed study it is found that the intrusion of gabbro occurred at two or more times, for at Lincoln Park and elsewhere the chilled contact and apophyses of one show that an older gabbro had already cooled. Banding (Fig. 1) is shown chiefly by the later mass, which is much the larger of the two. In some places two rock types alternate, but in most there are several minor rock varieties in irregular alternation. The bands vary in thickness from a fraction of an inch to many feet. It is likely that in the average the gabbro does not show such minute or intimate lamination as some associated sediments,’ but while there may be a general difference, each varies to resemble the other. Some contacts between adjacent bands are abrupt, but more com- monly there is a complete gradation between them. Some neigh- boring bands contrast strongly in color, while others are visible only on careful scrutiny; some are intensified by weathering, pro- ducing black, brown, gray, and white colors; some are conspicuous only from a difference in the degree of glacial polish (Fig. 2). Some large outcrops at Duluth show faint bands as much as fifty feet wide. Itis therefore evident that smaller outcrops a few feet wide may not reveal a banded structure even if it really exists. The whole area has been mapped as banded, because the outcrops which did not show the structure were smal! and not numerous; they may represent other variations of the mass, but are here considered as probably thick bands. .Most of the bands are regular, parallel, and fairly continuous along the strike and dip. However, there are locally lenticular bands, and spots or bunches along the bands, as shown in Fig. 2. Rarely the bands curve and finger out into each other (Fig. 3) and are as complex in structure as the ancient metamorphic gneisses. This irregularity is not as prominent as in the gabbro of the Isle of Skye (1); but locally the average dip of about 25° to the east increases to 80° with some variation also in strike. Although these outcrops may resemble metamorphic gneisses enough to be deceptive, a thorough study of ™U. S. Grant, ‘Contact Metamorphism of a Basic Igneous Rock,” Bull. Geol. Soc. Amer., XI (1900), 508. INTERNAL STRUCTURES OF IGNEOUS ROCKS A445 the rocks shows little trace of any crushing or recrystallization. Poikilitic and ophitic structures remain unaffected, and the min- erals are fresh. The associated earlier flows and sediments show no such structure as would have developed if the gabbro had been metamorphosed. ‘The structure is therefore a primary one. In general, the minerals of one band are the same as those of adjacent bands, and the banding is a consequence of difference in proportions of minerals. Textural changes are slight. Rock types at Duluth range from peridotite to anorthosite as extremes, with magnetite gabbro and troctolite as other variations from normal gabbro. A few measurements were made on thin sections of bands of gabbro, and some have been selected and presented in Table I to show how the bands vary. TABLE I PERCENTAGES BY WEIGHT y eae . Miscella- Plagioclase | Pyroxene Olivine Magnetite aaae 75 ite) IO 4 I Commonibandsiy sss) scien 465 19 IO 5 I 70 18 fe) 12 Two adjacent bands such as 84 12 3 | ° alternate many times 62 15 12 II | fo) 96 3 ° Tiel ° 20 44 ° BO || ° Bands of extreme composition 440 48 fo) 3 fe) 2 15 70 13 Oo 75 2 22 I fo) Detailed observations of the dip and strike of gabbro struc- ture at Duluth show only minor irregularities. Fig. 6 shows the general structure of the gabbro in those townships where observations have been made. It gives the impression of concord- ance with neighboring contacts. Magnetic mapping of the por- tion of the gabbro far to the northeast shows the general parallelism of bands and contacts, as well as the probably lenticular nature of the bands, which in Fig. 7 represent titaniferous magnetite ore. It may be added that some large sills, more or less related to the gabbro (22) and typically exposed at Beaver Bay, on Lake 446 FRANK F. GROUT Superior, show an exactly similar banding. The same gabbro in Wisconsin shows what has been described as a “‘bedded”’ struc- ture (23). There are a number of smaller masses which also show a banding, possibly of similar origin. The Purcell sills gabbro has certain streaks of lighter color (24) in addition to the separation into differentiated zones. A gabbro dike near Boulder, Colorado, has bands of iron ore parallel to the walls (25). A dike in the Isle of Man is similarly banded (26). The Mt. Holmes bysmalith has a color banding parallel to the walls (27). Other examples will no LAG 73 2 iS = 7DuluthGabbro aa Zo8 O} hp Bins eee c Co aes dea aese (Ses) Cy cee a Gearkewecemeer) Animikie eS ? Pre Animihie [SzeS] ag 5 os 26 miles SN ay (—— —— —_ ____ _} Fic. 6.—Sketch of the area of the Duluth gabbro showing the dip and strike of its internal structure. doubt be recalled by those who have worked in igneous rocks. A color banding is visible in many flows, but the difference in mineral content of the bands is not always clear. Fluxion structure—Certain igneous rocks have an abundance of platy or needle-like minerals, notably the feldspars and horn- blende. Many gabbros and syenites show a certain amount of parallelism of such grains (Fig. 8). Most of these rocks show banding of the sort just discussed. The rocks of Lizard (14) and the Adirondacks (12) and Laurentia (5) are ‘‘foliated.”’ The Ilimausak (4) rock has “primary schistose structure.” The Mt. Johnson rocks (9) have a “‘fluidal arrangement of grain.”” The INTERNAL STRUCTURES OF IGNEOUS ROCKS 447 Ontario syenites have an “original foliated or schistose structure” (10). All of these are noted by the authors as a feature in addition to banding. The occurrence at Duluth is a particularly good example of this structural feature as well as of the banding. Both the early, relatively thin feldspathic gabbro and the later banded gabbro show a parallelism of plagioclase grains in many outcrops. The smaller sills referred to also show the fluxion structure. Sheet structure-—When independent of surface changes of temperature, this is probably related to some such feature as the banding and fluxion structure just described, even when they of 0 la Imile ———— Areas ot high magnetic attraction [ __|eabbre % x x* x MLater and outcrops of magnetite. x x *** *IGranite Fic. 7.—Map of three square miles in Cook County, Minnesota, showing in black the lenticular form of the outcrops of bands in the banded Duluth gabbro. In this case the bands carefully mapped are those rich in titaniferous magnetite. themselves may be inconspicuous. Platy parting is recorded in the Ilimausak rocks (4) and the laccoliths of Highwood Mountains (28) and at Tripyramid Mountain (29) and elsewhere. The Duluth gabbro shows such joints in many outcrops (Fig. 9). Combinations. —It is evident from the foregoing notes that several masses show two or three structural features at the same time. This is true for a single outcrop as well as for the mass as a whole.* 1 The term “‘gneiss’’ may be extended to cover such rocks as these showing banding and fluxion structure, but when this is done the name should be qualified as “‘ primary gneiss.”’ The usage is discussed by Barlow, ‘‘ Nipissing and Temiskaming Region,”’ Geol. Survey of Canada, Ann. Rept., X (1897), Part I, p. 49; and Miller, Bull. Geol. Soc. FRANK F. GROUT 448 ‘azIs [eInyeu j[ey-auo ynogy “einjonzqs uorxngy Sutmoys ‘o1rqqes ay} Ul UOTVeSaI9aS a}1]}9USeUT eB pUe OIGGLS Jo SMaIA apis pure doy —'e ‘org INTERNAL STRUCTURES OF IGNEOUS ROCKS 449 THE RELATION OF IGNEOUS STRUCTURES TO THE FORMS OF IGNEOUS MASSES Pirsson finds the parallel arrangement of crystals and the platy parting of the laccoliths of the Highwood Mountains parallel to the roof (28), and has some evidence of a similar relation at Tri- pyramid Mountain (29). Iddings reports the parting and color banding of the Mt. Holmes ‘‘bysmalith”’ (27) parallel to the walls. _F1c. 9.—Sheeted structure in the Duluth gabbro evidently independent of the surface. Spheroidal weathering also appears. Rogers finds that the bands in the Cortlandt gneiss bear no definite relations to the borders of the magma (11). However, banding in the Adirondacks is of several kinds, and Miller records ‘‘a folia- tion that boxes the compass around the borders of the stocks”’ (30). The banding in lava flows and their trachytic structures is often recorded as parallel to the general plane of the flow. Examples Amer., XXVIII, 455. ‘‘Fluidal gneiss” and “‘injection gneiss,’’ as terms recently developed in structural geology, are probably best restricted to another type of struc- ture. It is detected in tracing igneous injections in bands between masses of a schist of other cleaved rock, or even curving in and out among rock fragments. This results in an alternation of the original cleaved rock (of whatever origin) and the igneous rock. Solution of the original rock and its metamorphism by the magma may pro- duce such an intimate intergrowth as to make distinctions between intrusive and intimate intergrowth as to make distinctions between intrusive and intruded rocks difficult. See Leith, Structural Geology, p. 85; and Cross, Science, XXIX, 946. 450 FRANK F. GROUT are seen in the Yellowstone banded obsidians (27) and the flows of the eastern (31) and southwestern states (32). The bands in the Purcell sills ““approximate a position parallel to the upper and lower contacts of the sill’? (24). Barlow records that the strike of the banding is uniform over large areas in the Nipissing and Temis- kaming regions, and shows a “‘marked correspondence in direction with the line of outcrop of the neighboring stratified Huronian rocks”’ (8). Adams finds the banding of Mount Johnson vertical and clearly parallel to the walls of a volcanic plug curving around the mountain (g). Adams and Barlow say that the strike of the banding and foliation of the alkali syenites of eastern Ontario con- forms to that of the adjacent country rock (10). Ussing considers the strata of the Ilimausak mass the upper layers of a batholith, but records that near the walls of the chamber the bands, which are nearly horizontal most of the way, turn up and become parallel to the walls (4). Gregory mentions some dikes which are foliated parallel to the walls, but not parallel to the foliation of the neighbor- ing schists (19). Banding also appears in a dike in the Isle of Man, parallel to its walls (26). Harker says that the banding of the gabbro at Carrock Fell is parallel to “the lie of the intrusion as a whole” with only minor undulations (17). At the Isle of Skye the banding is undoubtedly related to the boundaries. Iddings says that it is ‘‘not locally referable to the form or boundary of the body of a particular igneous rock,’ but he cannot have seen as much of the structure as Geikie and Teall, who say that each sheet of gabbro “consists of many parallel layers . . . . which correspond in direction with the trend of the sheet itself’’(1); or as Harker, who says that the bands dip with the mass as a whole and are in general parallel with the upper and lower surfaces of the sheets (2). A similar disagreement may be recorded in the case of the Duluth gabbro, where Elftman says that the banding is irregular (21), and more recent data show only minor variations from the direction of the contacts. The agreement of strike with the boundaries of the mass is shown in Fig. 6. The agreement in vertical section is more difficult to prove, on account of the scarcity of exposures showing such vertical sections. A single outcrop at «J. P. Iddings, Igneous Rocks, I, 252. INTERNAL STRUCTURES OF IGNEOUS ROCKS 451 Duluth (in the NW. Cor. Sec. 22, T. 50. N., R. 14 W.) reveals the upper contact of the gabbro with a clear exposure of dip. The roof here dips east a trifle irregularly at an angle of about 15°. At Lincoln Park it may be seen further that the later banded gabbro dips east under the earlier feldspathic gabbro. At the base of the gabbro, where one might search for the exposures of the floor, the relations are confused by pegmatitic and aplitic emanations and differentiates of great variety. At the Paulson mine in Cook County the floor apparently consists of eroded iron formation. Though the dip of the contact is not well exposed, drilling was con- ducted on the assumption that the bedding of the sediment and the banding of the gabbro indicated the direction of the contact. As far as exploration went, this proved to be true.’ A floor under the gabbro, conforming to the position of the banding, is also indi- cated by the constancy of the horizon of the gabbro intrusion, and by the arrangement of differentiates; some heavy segrega- tions are on the northwest, as if a floor dipped under them on that side. A review of literature and suggestions to be presented later with regard to the origin of these structures has no reference to any process which would tend to develop a banding independent of the boundaries. The favored theories involve movement during crystallization, and it would be expected that such movement would be more or less controlled by the boundaries of the magma chamber. These results are sufficiently uniform—only one apparent exception—to warrant the assumption that in a large way the fluxion and banded structures, as well as sheet jointing (when not referable to surface weathering), may be a guide to the position of the boundaries of igneous masses, and therefore of great value in mapping igneous forms and interpreting their position. The occurrences include flows, dikes, sills, a plug, a bysmalith, and laccoliths. Exceptions may be found, but even if the idea proves untrustworthy it is worth stating for the sake of stimulating accurate observations and records, which are at present not very numerous. «E. C. Harder, personal communication. 452 FRANK F. GROUT THE ORIGIN OF IGNEOUS BANDING Former suggestions.—The field study of the structure of the Duluth gabbro led the writer to assume a process of convection during crystallization as its cause. On reference to the literature, it was found that Bowen recently eliminated convection from the list of magmatic phenomena which he considered important (33). No clear statement of the relation between convection and structure could be found, and a review of the various explanations of the banded structure was thought desirable. In addition to suggestions made in connection with specific areas already mentioned, there are some discussions of the phenomena in general papers G. 3) and textbooks (34). Banding is so characteristic of metamorphic gneisses that the structure is not rarely referred to secondary processes, but the papers cited above show very conclusively that much of it is primary. Furthermore, as igneous rocks they cannot have been fused in place and retained traces of earlier structure, for the gabbro at Duluth and banded rocks in a number of other places are known to be intrusive into both their roof and floor (Fig. 5), neither of which is much metamorphosed. Of the other possible causes of banding, the following tabulation includes the chief suggestions found: . Partial assimilation of inclusions, forming schlieren . Lit par lit, or fluidal gneiss . Deformation during solidification . Deformation just after solidification . Streaked differentiation, with reference to rhythmic cooling or intrusive action 6. Successive intrusions: a) Cooling separately and successively b) Cooling later, all together 7. Heterogeneous intrusion mM BW ND H The writer would add: 8. Convection during crystallization differentiation Discussion.—The idea of partial assimilation of xenoliths, or lit par lit injection of wall rock as an explanation of banding, loses INTERNAL STRUCTURES OF IGNEOUS ROCKS 453 its main support when it develops (as in Duluth) that the floor and roof are rocks of about the same composition as the average gabbro, and that the bands range from anorthosite to peridotite—with compositions that could hardly be synthesized from any rocks in the region. It is admitted that schlieren, developed from xenoliths, occur in some local spots, but they have no relation to the banding and show no extreme in composition. Deformation during crystallization might explain the orienta- tion of grains, but cannot clearly explain the banding. The process of differentiation has not been described so as to explain the banding. It is, of course, probable that a rhythmic variation in the process of crystallization would give a rhythmic alternation of rock deposited, but that furnishes no explanation of the fluxion structure. None of the theories of differentiation out- line a process that will result in a combination of gravitative arrange- ment, parallel banding, and parallelism of grain. It is well said that the necessary conditions for igneous banding are heterogeneous composition and differential movement (34). Of the suggestions listed above, those which fulfil these two conditions are successive intrusion, heterogeneous intrusion, and deformation during crystallization. It is to be noted that each of these involves movement. The orientation of the platelike grains can hardly be accomplished except by some sort of movement of the magma while the plates are suspended in it. Such orientation is seen in surface flows where it is parallel to the direction of flow, and is often visible in thin sections of trachytes. To be sure, the settling of crystals, which idea is in special favor recently, might be thought of as analogous to the settling of mica plates in a sedi- ment. Those falling on a flat bottom might adjust themselves in horizontal and parallel positions. On the contrary, it does not seem probable that such an orientation would occur in a crystalliz- ingmagma. ‘The settling is slow, so that other crystals might lodge close to the plate and prevent its rotation. Furthermore, the difference in specific gravities, tending to orient the plates, is less than the difference for mica in water, while the viscosity opposing the rotation is much greater. As a final argument against orienta- tion by settling, the relation of structures at Mt. Johnson (g) should 454 FRANK F. GROUT be considered. There orientation is vertical and parallel to the sides of a volcanic plug as if dragged upward by eruptions through the channels, while in other respects the structures seem to be identical with those described elsewhere. It therefore seems necessary to adopt the customary view that the orientation of grains here associated with banding is a result of magma move- ment during crystallization in the general direction of the grains and of the bands, 1.e., parallel to the walls of the chamber. This view is so prevalent that the structure is often called “fluxion structure,’ even when its movement cannot otherwise be deter- mined. The question remains as to the nature of the movement. The common suggestions are movements of intrusion or of deformation. The writer is in favor of a third suggestion, viz., a circulatory movement. The data on which the argument is based are simple. It will be recalled that the banding of many rocks involves, not only a parallelism of grain, but an alternation—many times repeated— of mineralogically unlike bands. It is also known at Duluth that the extreme differentiates have in a large way become distributed in crudely gravitative positions, i.e., heavy near the bottom and light near the top. With these points in mind the several sug- gestions may be considered in detail. Successive intrusions of slightly varying magma are undoubtedly able to produce banded rocks and may even give a crudely gravita- tive arrangement; but the intrusion of successive layers of alter- nating composition, a few inches to a few feet thick, until the whole had a thickness of thousands of feet is inconceivable. The process would have to be extremely minute and often repeated in order to explain the detail of some outcrops. But such minute intrusion can hardly account for an intrusion several miles thick, where the intrusive action must have been on a grand scale. Even a process of crystal settling of each intrusive, combined with a sequence of intrusions, does not explain the alternations that are visible in some outcrops, where dozens of alternating bands appear in as many inches. Turning to heterogeneous intrusion, we find that the idea is accepted without any feeling of shock or surprise when attention is INTERNAL STRUCTURES OF IGNEOUS ROCKS 455 called to the variety sometimes shown in a series of extrusive lava flows, apparently derived from a single large chamber. The most recent statement of the case is incidental to a discussion of differentiation by crystallization and settling (33). It is necessary to introduce some modification to explain the development of the banded structures often seen. If differentiation took place by settling of crystals, and before the mass was all solid some dynamic process squeezed the liquid out from between the settled crystals, this liquid would not be the same in composition as the supernatant magma. ‘These two liquids might be involved in an intrusive layer and produce bands if not thoroughly mixed before crystallization. There are several difficult points in the application of this idea to “such banded rocks as the Duluth gabbro, though it seems clear from the variety of the Keweewanan lava flows that differentiation was well advanced before intrusion. First, the mechanics of the filter-pressing process in a deep reservoir like a batholith is not stated and is a little hard to con- ceive. Pressure on a magma is largely hydrostatic and not differential. Secondly, if heterogeneous liquids were intruded into so large a chamber there would be a great stirring and mixing effect and plenty of time to make the mixture more homogeneous before it crystallized. In general, the larger the mass the more time avail- able for diffusion and mixing. If banding was a result of hetero- geneous intrusion, the larger masses would be least banded. Asa matter of fact, the Duluth gabbro, one of the largest known intru- sions, is most strikingly banded. Thirdly, there is no reason to assume that the differentiation which caused the variation in the magma in the deep reservoir should suddenly cease upon intrusion into an upper horizon. In fact, from the gravitative arrangement it seems almost certain that some differentiation did take place. To be sure, if the hetero- geneous magma varied in specific gravity, the several parts might have been intruded in roughly gravitative position; but even if they were, there was nothing to stop the differentiation until the magma cooled. In so large a mass as the Duluth gabbro there would be plenty of time for further differentiation by settling. The difficulty 456 FRANK F. GROUT arises from the fact that if a crop of crystals settled across the intrusive bands they would destroy the banding and orientation. Fourthly, the alternation of bands found is so varied and extreme in composition—from anorthosite to peridotite—that the process of filter pressing can hardly yield the liquids which would be needed. Finally, the alternating bands, if they represent two liquids imperfectly mixed, should consist of a large volume of the upper liquid phase and a smaller amount of the phase strained off or filter-pressed from below. At Duluth there are found small volumes of granophyr and peridotite, with large volumes of anor- thosite and immense volumes of olivine gabbro. These would hardly result from filter pressing. It thus appears that crystal settling and filter pressing and heterogeneous intrusion will not explain the structures at Duluth. However, other modifications of the idea of heterogeneous intrusion may be suggested. The objections mentioned are enough to make them all unsatisfactory. The magmas may come from two reservoirs or become heterogeneous by any other process, but if they did they would have time to mix, and the mixing and crystal settling would destroy the banding. The magmas might be in- truded, when partly crystallized, as a great mass with “mushy” consistency. Banding and orientation would be satisfactorily explained by this idea, but in such a banded, mushy mass there would be no opportunity for gravitative differentiation. It seems necessary to believe that both differentiation and some sort of motion were involved in the production of the bands, and that these occurred after the magma reached its chamber. It may be best to leave the matter open as to the kind of motion that occurred, but the idea of convection is an attractive one. CONVECTION It is suggested that many cases of igneous banding are related to convection currents during crystallization differentiation. It is not necessary, in conceiving of this action, to regard it as a very thorough stirring, but rather as some degree of circulation following the intrusion of either homogeneous or heterogeneous magma. Neither is the process exclusive. Successive intrusions of hetero- INTERNAL STRUCTURES OF IGNEOUS ROCKS 457 geneous material probably occur, and crystal settling, differentia- tion, and deformation all leave their mark; but these things are apparently not sufficient to produce the structures seen. Convec- tion currents during crystallization result in bands and aid in the differentiation. Such a circulation would drag into parallel posi- tion any crystal formed near the wall of the chamber just as it became lodged in the viscous matrix and was removed from circula- tion. Rhythmic effects in the way of cooling, intrusive action, or gas emanation (all of which are known to be rhythmic) might rhythmically change the mineral composition of the crystals growing along the walls, and thus result in banding. Other features also are favorable and the writer does not find the mechanics of the process at all difficult. SUMMARY A review of the descriptions of banding in igneous rocks and a detailed study of the Duluth gabbro show that the alternation of mineralogically unlike bands is commonly accompanied by a fluxion structure and in some places by a sheet jointing. These structures are found to be parallel to the bounding surfaces of the igneous masses in nearly every case. Exceptions should be carefully studied and the facts in all cases noted, because such a relation of form and structure would be of great value in mapping and economic work. The banding and related structures probably develop during crystallization, while the magma is in convection circulation. SELECTED BIBLIOGRAPHY t. Geikie, A., and Teall, J. J. H., ‘‘On the Banded Structure of the Gabbro in the Isle of Skye,” Quart. Jour. Geol. Soc., L, 648. 2. Harker, A., ‘‘Tertiary Igneous Rocks of Skye,’”’ Mem. Geol. Survey of the United Kingdom, 1904. 3. Hoégbom, A. G., “Zur Petrographie von Orné, Hufvud,” Bull. Geol. Inst. Upsala, X, 150. 4. Ussing, N. V., Geology of Julianehaab, Greenland (Copenhagen, 1911), p. 318. 5. Wilson, M. E., ‘‘ Banded Gneisses of the Laurentian Highlands,” Am. Jour. SY eiig POOsQVA Ly sere} 6. Lawson, A. C., Geol. Survey of Canada, III (1887-88), Part I, pp. 130 f. 7. Adams, F. D., Problems of American Geology (1915), p. 80. 458 FRANK F. GROUT 8. Io. ils WAG Barlow, A. E., “Geology of Nipissing and Temiskaming,” Geol. Survey of Canada, Ann. Rept., X (1897), Part I, p. 61. . Adams, F. D., “‘The Monteregian Hills,” Jour. Geol., X1, 2709. Adams, F. D., and Barlow, A. E., “Alkali Syenites of Eastern Ontario,” Trans. Roy. Soc. Canada, IV (1908), 11, 73. Rogers, G. S., “Original Gneissoid Structure in the Cortlandt Series,” Amer. Jour. Sci., XXX, 125. Miller, W. J., ‘‘Magmatic Differentiation and Assimilation in the Adiron- dack Region,” Geol. Soc. Amer. Bull., XXV, 263. . Dale, T. N., “Granites of Maine,” U.S. Geol. Survey Bull. 313, p. 60. . Bonney, T. G., “Rocks of the Lizard District,” Quart. Jour. Geol. Soc., LIT, 40. . McMahon, C. A., “Gneissose Granite of the Himalayas,” Geol. Mag., Decade 4, IV (1807), 345. . Fennia, XI, No. 2 (1804), p. 97. » Harker; Aj) Carrock Helli?” Overt Sour. Geolt Soc, W310. . “Isle of Rum,” Mem. Geol. Survey of the United Kingdom (1908), p. 60. . Gregory, J. W., ““Gneisses in the ‘Cottian Sequence,” Quart. Jour. Geol. Soc., Iu) 265: . Grant, U. S., Minn. Geol. Survey, Final Rept., 1V, 477, and Bull. Geol. Soc. Amer., XI, 508. . Elftman, A. H., Amer. Geol., XXII, 135, and XXIII, 225. . Van Hise, C. R., and Leith, C. K., “Geology of the Lake Superior Region,” U.S. Geol. Surv. Mon. 52, p. 373- . Geology of Wisconsin, III, 337. 24. Schofield, S. J., “Origin of Granite in the Purcell Sills,’ Canada Dept. Mines, Museum Bull. 2 (1914), pp. 9, 11. . Jennings, E. P., ‘“‘A Titaniferous Iron Ore Deposit,’ Trans. Amer. Inst. of Min. Eng., XLIV (10912), 14. 26. Hobson, B., ““Igneous Rocks of the Isle of Man,” Quart. Jour. Geol. Soc., XLVII (1891), 430. . Iddings, J. P., U.S. Geol. Survey Mon. 32, I1, 66-68. . Pirsson, L. V., “‘The Igneous Rocks of the Highwood Mountains,” U.S. Geol. Survey Bull. 237, pp. 24, 45, 48, 40, 51, 52. . Pirsson, L. V., and Rice, W. N., ‘““The Geology of Tripyramid Mountain,” Am. Jour. Sci., XXXI (1911), 283. . Miller, W. J., “Adirondack Gneisses,”’ Jour. Geol., XXIV, 617. . Williams, G. H., “Ancient Volcanic Rocks,” Jour. Geol., II, 12. . Robinson, H. H., “‘San Franciscan Volcanic Field,’ U.S. Geol. Surv. Prof. Paper 76, p. 206. . Bowen, N. L., “Later Stages in the Evolution of Igneous Rocks,” Jour. Geol., December Supplement, 1915. . Harker, A., Natural History of Igneous Rocks. THE HABITAT OF THE SAUROPOD DINOSAURS CHARLES C. MOOK American Museum of Natural History INTRODUCTORY In the study of the sauropod dinosaurs which has been carried on by the writer for a number of years under the direction of Pro- fessor H. F. Osborn in connection with the preparation of the latter’s monograph on these reptiles, some problems have presented themselves upon which a study of the habitat, or immediate environment, has a bearing. The course, or trend, of evolution in a group of organisms is limited, or controlled, by two things: (1) the heritage or assemblage of characters inherited from the ancestors; and (2) the environment. The environment offers the organism opportunities for developing along a limited number of lines. What these lines will be depends upon the general character of the environment. For instance, upon inland plains advanced aquatic adaptations, such as are characteristic of marine organisms, will be barred out, and under strictly-marine conditions the development of cursorial locomotor apparatus is impossible. This is true no matter what may be the heritage of the organism under discussion. Within certain limits, however, the environment offers the possibilities or opportunities for evolution along a number of lines. The heritage furnishes the material or instruments by which, or by a modification of which, the organism may evolve along one or more of these lines. In working out adaptations and habits in a group of animals such as the Sauropoda, morphology, together with comparison with living forms, will be the most important guide. Morphological - structures have meanings, and if these meanings can be interpreted the habits of the animals possessing the given structures can be determined to a certain extent. A study of the environment of 459 460 CHARLES C. MOOK the group in question may aid in interpreting the structures and may guide us in determining the habits of fossil animals, because certain types of environment definitely exclude certain modes of life, as has been noted above. The habits of the animals must have conformed to the environment which actually surrounded them when they lived. The present discussion is concerned with the environment of the Sauropoda. The environment of a group of organisms is divisible into two components: first, the physical, and second, the biotic. These are related to each other in a complex manner, but with regard to their relation to a given group they may be considered separately. The first is concerned with such things as climate, with mountain, plain, delta, lake, or marine conditions, with the geographic extent of a given type of physical condition, with means of communication and barriers preventing communication or intermingling, with the rate of sedimentation and erosion, with the presence or absence of volcanic phenomena, etc. The second is concerned with food supply, with competition, and with enemies. If in a given case the various factors of each of these compo- nents may be determined, a comprehensive idea may be had of the habitat, or immediate environment, of the group of ani- mals involved. PHYSICAL ENVIRONMENT The physical environment of the Sauropoda is concerned with the geology of the Morrison, Arundel, Wealden, and corresponding Indian, African, Patagonian, and Malagasy formations. The American Morrison may be considered in this connection as an example of sauropod-bearing deposits. The physical characters of the Morrison, its relations to other formations, considered in connection with the general Mesozoic history of Western North America, indicate certain definite things regarding the geography, topography, climate, and dominant physical processes of the time and region in which the Western American Sauropoda lived. An extension of this study to include . world-wide conditions would give a fair idea of the physical environment in general. THE HABITAT OF THE SAUROPOD DINOSAURS 461 The characters of the Morrison may be discussed under the following heads: (1) distribution, present and probable past; (2) lithology; (3) internal structures; (4) stratigraphic relations; (5) conclusions regarding conditions of deposition, physical processes dominant during the period of deposition, and middle Mesozoic history. 1. The present distribution of the Morrison is indicated on a map compiled by the writer.‘ The formation outcrops along the eastern and western borders of the Rocky Mountains in Wyoming, Colorado, and New Mexico; in the rim of the Black Hills; in canyons in southeastern Colorado and northeastern New Mexico; around the borders of the Bighorn and Owl Creek mountains in Montana and Wyoming; in isolated uplifts in Wyoming; in canyons and mesa scarps in northwestern New Mexico, western Colorado, and eastern Utah; and in various other occurrences in the states mentioned. The outcrops are usually not extensive, the formation never being the country rock over a wide area. The total area in which Morrison outcrops occur is, however, very large. There are vast areas where the Morrison must unquestionably underlie younger formations. The areas in which the Morrison was formerly present, but from which it has been removed by erosion, are also very large, their exact size not being known at the present time.- The total area which was formerly covered by Morrison sediments must have been extremely large, very likely exceeding a million square miles in extent. As remains of sauropods are now found at practically every region of Morrison outcrops, it follows that the distribution of the Sauropoda in North America was also very wide. 2. Lithologically the Morrison is composed of a variety of rock types. The formation is frequently described as a series of ‘‘ joint clays,’ fine-grained sediments which appear to be fairly well consolidated when dry, but which crumble or break readily when wet. These are red, brown, gray, or maroon in color. Petro- graphically they are fine grits composed mostly of quartz, with some argillaceous interstitial material which may or may not be t Charles C. Mook, ‘“‘A Study of the Morrison Formation,” Annals of the New York Academy of Sciences, XXVII (1916), 39-101, Pl. VI. 462 CHARLES C. MOOK stained red by hematite, as the case may be. These grits, espe- cially the red ones, are frequently most abundant in the upper levels of the formation; it is in these upper levels that calcareous material is more scarce. Beds of sandstone, sometimes of considerable thickness, occur at various levels; these sandstones are made up principally of quartz, which is often well rounded; feldspar grains, either fresh or more or less altered, occur along with the quartz; in some beds grains of volcanic ash, both fresh and altered, are found; and in a few instances beds of coarse sand several feet thick are made up of volcanic ash. Limestone beds, usually not over a foot or two in thickness, are frequently found in a section; these are sometimes composed largely of the shells of small gastropods. They are more common near the base of the formation than in the upper members. The lower beds are often arkosic, considerable quantities of feldspar being present, often cemented to the accom- panying quartz and to each other by a calcite matrix. Thin beds of agate are found in some sections. Very coarse material is not found in the Morrison. Sandstones of a moderate degree of coarseness are common throughout the entire area of Morrison outcrops. Such sandstones are, however, on the whole thicker and more common in the western exposures than in the eastern. 3. The Morrison contains internal structures of considerable interest. In the first place the various strata often appear to the eye to extend over considerable distances, but when detailed sections are made, even a few miles apart, and compared with each other, it is noticeable that the details of the sections vary con- siderably. A gradual thinning out of beds of one kind of material, and their replacement by another kind, is the rule in this formation. For all of this variation the general aspect of the formation in one locality is very much like that in another. This type of thing has been aptly described by Dr. Lee as “uniformly variable.’”’ In some cases the thinning out of beds is sudden, as in the case of the old stream channel exposed at the site of the old Marsh-Hatcher dinosaur quarry near Cafion City. The variation and at the same time the uniformity of the thick- ness of the formation are of special interest. The greatest recorded THE HABITAT OF THE SAUROPOD DINOSAURS 463 thickness is about goo feet, and it is possible, if not probable, that beds are included in this measurement which are older than the Morrison. The range in thickness is usually from 700 feet in the western sections to less than too feet in the Black Hills region. This thickness is exceedingly slight for a formation of such vast geographic extent as the Morrison. In general, the western sections are thicker than the eastern, but this will not hold as an invariable rule. Sections of 400 feet or less sometimes occur in the western areas, and sections fully 400 feet thick exist in eastern New Mexico. No section of more than 500 feet is known from the eastern areas, however, and the western sections frequently reach that or a greater thickness. The Morrison sediments might perhaps be described from the point of view of thickness as a thin mantle of sandstones and clays extending over a vast area, thickest in the west and thinning out definitely, but irregularly, to the east. Cross-bedding is abundant in the beds of the Morrison, espe- cially in the sandstones. It is represented by the type described by Walther and others as typical of desert deposits, and also by the type usually assigned to stream deposition. 4. The stratigraphic relations of the formation are in a broad way disconformable with regard to.the underlying terranes. The deposits rest upon older formations of various ages, from the Unkpapa sandstone of uppermost Jurassic or earliest Comanchean age in the Black Hills region to Archean crystallines in the Rocky Mountain region. The relation of the Morrison to the overlying sediments appears to be a conformable one. The fact that the Morrison appears to be closely related to the succeeding formations has been pointed out by Lee.’ For further description of the Morrison formation the reader is referred to the above-mentioned article on the Morrison by the writer and to the bibliography contained therein. Taken together, the physical characters of the Morrison indicate a history something like the following: After an extensive period of erosion, during Jurassic and perhaps late Triassic time, Western North America was invaded from the north by the sea, and the tW. T. Lee, ‘‘Reasons for Regarding the Morrison an Introductory Cretaceous Formation,” Bull. Geol. Soc. Amer., XXVI (1915), 303-14. 464 CHARLES C. MOOK sediments of the Sundance formation were laid down; following this the sea retreated, the retreat taking place along with final Jurassic folding in the Sierra Nevada region. Over the plain exposed by the retreat of the Sundance sea the Morrison sediments were spread out in the form of a very broad, very flat alluvial fan from west to east. This fan must have been crossed by many large streams, dotted with lakes, large and small, and characterized by an interlacing type of drainage, much after the manner of the great alluvial plains of Eastern China at the present time. The plain must have been low, and the streams crossing it must have been characterized, for the most part, by a low gradient. Locally, especially in the western areas, there may, in fact must, have been some deposition in relatively swift currents, as indicated by the cross-bedded sandstones. These sandstones being rather fine- grained for the most part and never conglomeratic, true torrential conditions were probably not present in any part of the Morrison area so far studied. The round sand grains, associated with aeolian type of cross-bedding and sudden variations in thickness, indicate that wind deposition was also a factor in the gradual building up of the Morrison sediments. The presence of unaltered or little altered volcanic ash indicates that volcanic activity must have been going on somewhere in the region. Over a plain as broad and flat as this one must have been material could not have been transported rapidly from the original source to the outer limits of the area of sedimentation. The sluggish streams of the plain must have deposited material and later picked it up again and carried it farther very many times before a selected lot of sediment could have reached the outer margins. This will account for the greater relative abundance of finer clays in the eastern areas. ‘The inter- lacing stream, lake, and swamp conditions on such a plain would readily admit of rapid shifting of the courses of streams, areas which were at one time stream beds constantly changing to inter- stream areas, and the reverse. This would result in the slow, gradual shifting of material outward. The end result would be the product of alternate deposition and erosion, erosion and deposition, for a long period of time, the material being very slowly worked eastward. In some such manner as this the relatively thin sheet of THE HABITAT OF THE SAUROPOD DINOSAURS 465 Morrison sediments may have been spread out over a vast area. In this area of deposition, and living while it was in progress, were the Sauropoda and their contemporaries. BIOTIC ENVIRONMENT I. VERTEBRATES The known vertebrate fauna of the Morrison is large and varied. The unknown fauna must have also been larger; perhaps, in fact quite likely, larger still. Of the mountain fauna of Morrison time nothing can be said, but the mountain fauna was not part of the sauropod habitat. 1. Mammalia.—Between twenty and thirty species of mammals have been reported from Morrison beds. These were small tricono- donts, trituberculates, and multituberculates. They are known only from teeth and fragmentary jaws, so that their structure and adaptations cannot be made out. It has been suggested that they were arboreal. They might serve very well for arboreal members of the Morrison fauna. These small mammals could scarcely have competed with the Sauropoda for food; they certainly could not have constituted food for the Sauropoda in themselves; nor could they have been directly formidable as enemies. It has been suggested, however, that they may have fed, in part at least, upon reptilian eggs. If they did, and if they existed in large numbers, they may have been very troublesome companions for the sau- ropods. 2. Aves.—Only one species of bird is known from the Morrison. Undoubtedly more were present in Morrison time, but there is no direct evidence of their existence. It is not likely that the birds had any important effect upon the lives of the sauropods, although they may have had something to do with the distribution of species of plants which perhaps composed part of the sauropod diet. 3. Reptilia.—The reptilia of the Morrison were many and varied. They all represented degrees of organization and stages of evolution which were comparable to the degree of organization and stage of evolution of the member of the fauna under discussion— the Sauropoda. Undoubtedly there was competition of several sorts between the Sauropoda and their reptilian contemporaries. 466 CHARLES C. MOOK The reptilian fauna of the Morrison suggests the following analysis: a) Rhynchocephalia. The only modern representative of this group is non-marine, and there is nothing in the structure of the single Morrison representative of the order to suggest that it was anything different. It may perhaps have been amphibious, or fluviatile, or terrestrial. It is not probable that there was any direct competition between the members of this group and the sauropods. b) Crocodilia. Several species of mesosuchian crocodiles are known to have existed in Morrison time along with the Sauropoda. The modern crocodiles are amphibious creatures, either fluviatile or lacustrine, not marine, and the Morrison forms probably lived in a similar manner. They were good-sized, active, carnivorous, relatively intelligent animals, which may easily have preyed upon the young of the Sauropoda. c) Pterosauria. The pterodactyls were aérial forms. One species is known from the American Morrison. It is hardly to be expected that sauropod dinosaurs and pterosaurs would enter into any direct conflict or competition with each other. d) Squamata. Lizards, snakes, and mosasaurs are entirely unknown from the Morrison. Some of them must have been living somewhere at the time, for lizards are known to have existed since the Triassic, and mosasaurs appear well developed early in the Cretaceous. Consider in this connection the fact that the mosa- saurs were marine animals. e) Chelonia. Turtles at the present time are marine, fluviatile, lacustrine, or terrestrial. Only one species in this group is known from the Morrison, and it is probable that the chelonian ele- ment in the Morrison fauna was not very important. The turtles could hardly have come into severe competition with the Sauropoda. jf) Sauropterygia and Ichthyopterygia. The plesiosaurs and ichthyosaurs were entirely marine. No trace of them has been found in Morrison rocks, though they must have been living in the sea during Morrison time. They are found in marine rocks both younger and older than the Morrison. THE HABITAT OF THE SAUROPOD DINOSAURS 467. g) “Dinosauria.”’ The two great orders of reptiles collectively known as “Dinosauria”’ were represented in the Morrison by a variety of forms. These were all terrestrial or amphibious, perhaps some forms being largely fluviatile or lacustrine. Of the sauris- chians there were present, besides the sauropods, several types of carnivorous forms. Some of these were small, active, and not very formidable. Ornitholestes may be considered as a typical example of this group. Probably these small dinosaurs had little impor- tance, so far as the Sauropoda were concerned, unless, perhaps, they ate sauropod eggs. They were too small and weak to attack the sauropods, and by their activity and carnivorous structure could obtain food that would not be available for gigantic, largely herbivorous swamp dwellers. The larger carnivorous dinosaurs, such as Allosaurus, Creosaurus, and Ceratosaurus undoubtedly played a very important part in sauropod economy. ‘The asso- ciation of carnivore teeth and grooved sauropod bones is in perfect accordance with the idea that the sauropods were prey for the large carnivores. ‘The large carnivores were unquestionably land-living forms. It is not at all likely that they entered the water except under unusual circumstances or along the margins. The gigantic sauropods, therefore, were less likely to be attacked in the water than on land. This may have tended to keep the sauropods in the water, and may have had considerable control over the evolution of the group. The ornithischian dinosaurs must also have had an effect upon the Sauropoda, but as competitors rather than as direct enemies. The stegosaurs and the iguanodonts, both large and small, existed along with the sauropods. These forms, especially the stegosaurs, were probably land animals. They were herbivo- rous beyond all doubt. If the sauropods spent a considerable part of their time on land they must have come into competition with the predentates in the matter of getting food. Perhaps such a competi- tion took place early in the history of the Sauropoda, and may have been instrumental in forcing the latter to take to the streams, and finally to develop some aquatic adaptations and spend the greater part of their time in the water. The predentates may have aided the sauropods in their struggle for existence by furnishing a con- siderable amount of food for the carnivores. If the predentates 468 CHARLES C. MOOK and carnivores were almost exclusively terrestrial and the sauropods largely aquatic in habit, the latter might escape the carnivores much more frequently than if the predentates were not present. 4. Amphibia.—The known amphibian fauna of the Morrison consists of one frog or toad. A fauna of this nature could scarcely have had any effect upon the lives or development of the Sauropoda. 5. Pisces.—The only fishes known to have existed along with the American Sauropoda were a few species of Ceratodus. These could hardly have furnished food for the sauropods, or have had any direct effect upon the security of the latter. II. INVERTEBRATES The known invertebrate fauna of the Morrison is neither very large nor varied. It consists of a number of fresh-water pelecypods and gastropods, together with a few ostracods. None of these were large enough nor abundant enough to have served as an essential part of the diet of the sauropods, but they may have served as accessories, to a very small extent, to the normal sauropod diet. II. FLORA The known flora consisted almost entirely of cycads. These might also have comprised an accessory portion of the sauropod diet, but probably not much more. The remainder of the Morrison flora is very little known. Silicified wood is found occasionally, and rarely some imperfect reeds. The Kootenie formation contains leaves of deciduous trees, but few if any sauropods. The Arundel formation of Maryland contains both sauropod bones and deciduous leaves. The latter also might have formed an accessory part of the food of the sauropods, but could scarcely have been abundant enough to have sustained their huge bulk. In a region such as the one described above there might have been a considerable amount of soft vegetation which would not be easily preserved. In the interstream areas, especially where considerable amounts of windblown sands were being deposited, little vegetation may have been present, certain areas being semiarid. Other areas were probably well covered with vegetation. The character of the known flora suggests a rather warm climate, and the physical THE HABITAT OF THE SAUROPOD DINOSAURS 469 conditions point more to an abundance of rainfall than to a wide- spread lack of it. We may therefore conclude, provisionally at least, that the climate was warm and moist. INTERPRETATION OF HABITAT From the evidence of the physical characters of the Morrison formation and from the nature of its flora and fauna it may be possible to work out, to a certain degree at least, the environment which surrounded our American Sauropoda. In the first place, there was a vast plain in the Western United States, possibly extending northward into Canada; this plain was low throughout its entire extent, but slightly higher in the west than in the east; it was bordered on the west by mountainous country. From this mountainous area streams issued, bringing sediment and depositing it along the western border of the plain. The streams, upon leaving the mountains, became sluggish and split up into a large number of distributaries. As in the great plains of China, streams would be split up into distributaries and united again. Between the streams, and more or less connected with them, were lakes. A considerable amount of vegetation was present, especially along the stream and lake banks. In this respect the plain would resemble the interior of Florida, which has often been suggested as the type of habitat possible for sauropod dinosaurs. Our Morrison plain would differ from Florida, however, in its extent, the central Florida swamps being relatively small and the Morrison plain being a million square miles or more in area. In some of the interstream areas, especially in the west, vegetation may have been more scarce. Active volcanoes were present somewhere, either in the mountains or on the plain. Somewhere to the southeast was the sea, but its exact border is not known, especially with regard to the earlier part of Morrison time. On this plain lived an extensive terrestrial, amphibious, fluviatile, or lacustrine fauna. Little, primitive mammals climbed the trees or scurried over the ground; here and there some birds flew through the air; along the shores of the lakes and rivers lived some Sphenodon-like rhynchocephalians; ptero- saurs flew through the air; some turtles swam about in the water; crocodiles inhabited the stream and lake banks, and infested the 470 CHARLES C. MOOK waters themselves. On land, large and small carnivorous dinosaurs roamed about, seeking what they might devour; stegosaurs and camptosaurs endeavored to escape their voracious contemporaries; some frogs inhabited the swamps. In the rivers and lakes lung fish swam about. Fresh-water mollusks and crustaceans lined the river and lake bottoms in places or swam about freely in the water. Cycads grew in abundance, and soft swamp vegetation probably furnished the food supply for hungry reptiles. In some such environment as this, or at any rate in one very much like it, lived the American sauropod dinosaurs. So far as is indicated by evidence now directly available, conditions of practically the same sort prevailed in other parts of the world inhabited by Sauropoda. PETROLOGICAL ABSTRACTS AND REVIEWS ALBERT JOHANNSEN WASHINGTON, HENRY S. ‘‘Some Lavas of Monte Arci, Sardinia,” Amer. Jour. Sci., XXXVI (1913), 577-90, analyses. The extinct volcano Monte Arci has not been described since 1857. It consists of a core of rhyolite with later dacites, andesites, trachytes, and basalts. The various lavas are described petrographically and chemically. Seven analyses are given and they are recomputed in the C.I.P.W. system. WASHINGTON, Henry S. “The Composition of Rockallite,”’ Quart. Jour. Geol. Soc., LXX (1914), 294-302. This paper describes a peculiar aegirite-granite from Rockall. A chemical analysis gives SiO. 69.80, while a previous and probably accurate analysis of the other half of the same specimen gave 73.60. This shows the great variation possible in a single small specimen. WASHINGTON, Henry S. ‘An Occurrence of Pyroxenite and Hornblendite in Bahia, Brazil,’ Amer. Jour. Sci., XX XVIII (1914), 79-90, analyses. Describes a hornblendite with a border of pyroxenite intruded in gneiss. The rocks contain considerable manganese. They are interesting since they possess absolutely no feldspar and yet show from 18 to 30 per cent in the norms. WasHINGTON, Henry S. “TI Basalti Analcitici della Sardegna,” Boll. Soc. Geol. Ital., XX XIII (1914), 147-67. A translation of ‘The analcite basalts of Sardinia”? which appeared in this Journal, Vol. XXII (1914), 742-53. 471 472 PETROLOGICAL ABSTRACTS AND REVIEWS WASHINGTON, HENRY S. “Contributions to Sardinian Petrog- raphy. I. The Rocks of Monte Ferru,” Amer. Jour. Sci., XXXIX (1915), 513-20, figs. 2, analyses. In this paper are described trachyte, trachytic phonolite, basalt, and analcite-basalt from Monte Ferru, Sardinia. WASHINGTON, HENRY S. ‘The Calculation of Calcium Orthosili- cate in the Norm of Igneous Rocks,” Jour. Wash. Acad. Sci., V (1915), 345-50. Calcium silicate, which was originally calculated in the C.I.P.W. system as 4CaO.3Si02, was later (1912) revised as 3CaO.2Si02. This is again revised and now the orthosilicate 2CaO.. SiO; is used in calculating the norm. A table is given which may be cut out and pasted over the akermanite table given in the C.I.P.W. book. WASHINGTON, HENRY S. ‘The Correlation of Potassium and Magnesium, Sodium and Iron in Igneous Rocks,” Proc. Nat. Acad. Sci., I (1915), 574-78. The writer shows that in igneous magmas potassium and magnesium on the one hand and sodium and iron on the other probably vary together. WASHINGTON, H. S., and Day, ARTHUR L. ‘Present Condition of the Volcanoes of Southern Italy,” Bull. Geol. Soc. Amer., XXVI (1015), 375-88, pls. 9. A record of the state of activity and other conditions at Vesuvius, AXtna, and the Aolian Islands during the summer of 1914. The general results of the observations and studies of gases, salts, and rocks collected will be given in a subsequent paper. Watson, THomMAS L., and TABER, STEPHEN. ‘‘Magmatic Names Proposed in the Quantitative System of Classification for Some New Rock Types in Virginia,” Bull. Phil. Soc. Univ. of Virginia Scientif. Ser., I (1913), 331-33- Proposes four new names in the C.I.P.W. system, for rocks found in Virginia. PETROLOGICAL ABSTRACTS AND REVIEWS 473 Watson, THomAs L., and CLinE, Justus H. “Petrology of a Series of Igneous Dikes in Central Western Virginia,” Bull. Geol. Soc. Amer., XXIV (1913), 301-34, pls. 3, figs. 5, analyses. Describes a series of dikes occurring in Rockbridge, Augusta, Rock- ingham, and Highland counties, Virginia. The rocks are diabase, granite-felsophyre, quartz-gabbro, nephelite-syenite, teschenite, and camptonite. All the rocks are analyzed and their positions in the C.I.P.W. system are determined. Watson, THomas L., and TABER, STEPHEN. “Geology of the Titanium and Apatite Deposits of Virginia.” Bull. III-A, Vargo Gah. (SOs, TOUR, IAD. BelieOlSs By, Wess QA, 7 oon bibliography on titanium. After a general discussion of the titanium minerals and a brief description (28 pp.) of the rutile deposits of the world, the geology of the ore deposits in the Amherst-Nelson region is given. The rocks are described in detail, and chemical analyses of some of the associated rocks—biotite-quartz-monzonite-gneiss, syenite, gabbro, nelsonite, and diabase—are given. In the first the plagioclase is oligoclase, Ab2An;; the second rock is really an andesinite, or andesine-anorthosite as used -by the authors; the rock described as gabbro consists chiefly of andesine and hypersthene with ilmenite, apatite, orthoclase, quartz, biotite, etc.; the nelsonite is a dike-rock consisting essentially of ilmenite and apatite. Wuerry, Epcar T. “A Peculiar Oolite from Bethlehem, Pennsyl- vania, Proc. U.S. National Mus., XLIX (10915), 153-56, pls. 2. Describes an oolite in which the ooids show a ‘‘half-moon”’ effect, the upper portions being light and the lower dark. An explanation for this peculiar character is given. WHERRY, EpGAR T., and GorDON, SAMUEL G. “An Arrangement of Minerals According to Their Occurrence,’ Proc. Acad. Nat. Sct., Philadelphia, 1915, 426-57. Minerals are classified according to their occurrence in igneous rocks, pegmatites, hydrothermal deposits, fumerolic deposits, or sediments. Each of these main divisions is subdivided into various groups, and under each is given a list of the minerals which occur. It is a very useful list. 474 PETROLOGICAL ABSTRACTS AND REVIEWS WHeErRRY, Epcar T. “The Microspectroscope in Mineralogy.” Smithsonian Miss. Col., LXV (1915), No. 5. Pp. 16. Descriptions of the spectra observed in the examination of about 200 minerals with the microspectroscope. Witxman, W. W. “Kaleviska bottenbildningar vid Mélénjarvi.” Bull. com. géol. Finlande, No. 43, 1915. Pp. 36, figs. 11. A geological description of the basal formations near Lake Mdélén- jarvi, in east Finland. Granite, conglomerate, quartzites, various’ schists and phyllites, and basic dikes are described. Wricut, FRED. EuGENE. ‘An Electrical Goniometer Furnace for the Measurement of Crystal Angles and of Refractive Indices at High Temperatures.” Jour. Wash. Acad. Sci., III (1913), 396-401. Describes a furnace for measuring the interfacial angles and refrac- tive indices of crystals at temperatures up to 1225°. WRIGHT, FRED. EUGENE. ‘‘Graphical Methods in Microscopical Petrography,”’ Amer. Jour. Sci., XXXVI (1913), 509-39, pls. 8, figs. 9. ; The writer gives eight charts for the solution of the following equa- tions: sinz=# sin r, sin? 7=n? sin?r, cot A= B.cot C, sin A=sin B. sin C,. ca Ege ee 2, 2 2 a2 Z - sin Grasin Gai tan2 Vie tan Va= neneg : = Te ior I oe \enP WriGuHT, FRED. EUGENE. ‘‘The Change in the Crystal Angles of Quartz with Rise in Temperature,” Jour. Wash. Acad. Sct., III (1913), 485-94, figs. 2. The polar angle p of the unit rhombohedron was found to decrease at an increasing rate until the temperature of 575° was reached. Here, with the inversion of a-quartz to B-quartz there is an abrupt decrease of 2’ in the angle, and the value remains constant thereafter to 1250°. PETROLOGICAL ABSTRACTS AND REVIEWS 475 WRIGHT, FRED. EUGENE. ‘‘The Measurement of the Refractive Index of a Drop of Liquid,” Jour. Wash. Acad. Sct., IV (1914), 269-79, figs. 14. Describes various methods for measuring the refractive index of liquids, five of them new methods for the petrographic microscope. WRIGHT, FRED. EUGENE. “The Determination of the Relative Refringence of Mineral Grains under the Petrographic Micro- scope,” Jour. Wash. Acad. Sci., IV (1914), 389-92. Describes a new diaphragm for observing relative refractive indices. WRIGHT, FRED. EUGENE. ‘“‘The Optical Character of the Faint Interference Figure Observed in High Power Objectives be- tween Crossed Nicols,”’ Jour. Wash. Acad. Sci., IV (1914), 301-09. Explains the positive character of the interference figure produced by the rotation of the vibration plane of the transmitted light. Wricut, FreD. Eucene. “A New Half-Shade Apparatus with ‘Variable Sensibility,’ Wash. Acad. Sci., IV (1914), 309-13. Describes a device for varying the sensibility of a half-shade appa- ratus. WRIGHT, FRED. EuGENE. “A Simple Method for the Accurate Measurement of Relative Strain in Glass,” Jour. Wash. Acad. Sci., IV (1914), 594-08. Describes a wedge with which it is possible to measure path- differences of less than typ. WRIGHT, FRED. EUGENE. ‘Measurements of Refractive Indices on the Principal Optical Sections of Birefracting Minerals in Convergent Polarized Light,” Jour. Wash. Acad. Sci., IV (1914), 534-42. Two refractive indices are measured by the immersion method, the third by measurements made on the rings in the interference figure. 476 PETROLOGICAL ABSTRACTS AND REVIEWS WRIGHT, FRED. EUGENE. ‘The Accurate Measurement of the Refractive Indices of Minute Crystal Grains under the Petro- graphic Microscope,” Jour. Wash. Acad. Sci., V (1915), 101-7. Considers the conditions necessary for the exact measurement of refractive indices of minute crystal particles by the immersion method. WRIGHT, FRED. EUGENE. ‘‘A New Crystal-Grinding Goniometer,” Jour. Wash. Acad. Sci., V (1915), 35-41. Describes the precision goniometer used in the Geophysical Labora- tory, which is capable of grinding a plate to within 1’ of the required position. Wricxt, FreD. EUGENE. ‘Obsidian from Hrafntinnuhrygghur, Iceland. Its lithophysae and Surface Markings,” Bull. Geol. Soc. Amer., XXVI (1915), 255-86, figs. 12. The writer cee the formation of spherulitic, lithophysal, and pumiceous structures in the obsidians of Iceland, as well as certain mark- ings produced by the etching effect of hot volcanic gases. WRIGHT, FRED. EUGENE. “The Position of the Vibration Plane of the Polarizer in the Petrographic Microscope,” Jour. Wash. Acad. Sci., V (1915), 641-44. The writer concludes that there is a slight practical advantage in having the plane of vibration parallel to the vertical cross-hair. WRIGHT, FRED. EUGENE. ‘“‘A Simple Device for the Graphical Solution of the Equation A=B.C,” Jour. Wash. Acad. Scz., VI (1916), 1-5. Describes a device by means of which certain formulae may be solved by the use of a series of scales. WRIGHT, FRED. EuGENE. ‘‘A Geological Protractor,”’ Jour. Wash. Acad. Sci., V1 (1916), 5-7. A device for solving various problems, such as the slope of any bed, in the field. Witrine, E. A., and Becut, K. ‘‘ Uber neue Turmalinanalysen,” Sitsb. Akad. Wiss. Heidelberg, 1913. A-—20. Pp. io, analyses. PETROLOGICAL ABSTRACTS AND REVIEWS 477 Wittrine, E. A. “Uber Kristallwinkel bei verschiedenen Tempera- turen,’’ Sitsb. Akad. Wiss. Heidelberg, 1913. A-—21. Pp. 5, figs. 2. In the determination of the changes produced in crystal angles it is sometimes necessary, on account of the irregular natural faces, to grind artificial faces. The present paper deals with the question as to the most advantageous angle at which they should be cut, and the con- clusion is reached that this is 45° with the base. If possible it is better to measure across the base from one pyramidal face to another, thus (101) to (101). WULFING, E. A., and OPPENHEIMER, L. “Neue Untersuchungen an Cordierit,” Sztzb. Akad. Wiss. Heidelberg, 1914, A-—tio. Ppa ken Wes. 1. New determinations on cordierite from eight localities. Witrine, E. A., and Horner, F. “Die kristallographischen Konstanten des Stauroliths vom St. Gotthard,” Sztzb. Akad. Wiss. Heidelberg, 1915, A-10. Pp. 11. Gives new determinations of interfacial angles and axial ratios. Wttrinc, E. A. “Lassen sich die kristallographischen Funda- mentalwinkel der Plagioklase mit der Zusammensetzung in gesetzmassige Beziehung bringen ?”’? Sztzb. Akad. Wiss. Heidel- DEV Ee OLS: A132.) Jep. 245 tgs. 0: New determinations of crystallographic properties make it possible to say with certainty that these properties follow a definite law. ZIEGLER, VicToR. “The Minerals of the Black Hills.” Bull. zo, South Dakota School Mines, 1914. Pp. 255, map 1, pls. 30, figs. 26, bibliography. In this bulletin the writer describes briefly, for the benefit of the miner, prospector, and student, all the minerals known to occur in the Black Hills. There are included three very good tables for the mega- scopic determination of the minerals, as well as many full-page illustra- tions. The book should be of great value to everyone working in the Black Hills region. REVIEWS Geology and Paleontology of the Raton Mesa and Other Regions in Colorado and New Mexico. By Wits T. LEE and F. H. Know ton. U.S. Geological Survey, Professional Paper IN; WO; MON7e 1e\Os AUK, JOS, EAA, WIGS, a0), This interesting paper presents many details concerning the geology of the areas indicated by the title. To the general reader the large interest of the paper centers in the correlation of disputed formations in the several regions discussed, and in the attempt to establish their time-relations to formations elsewhere. Some of the significant con- clusions are the following: “The coal-bearing rocks of the Raton Mesa region which have formerly been referred to the Laramie constitute two distinct formations separated in time by a period of erosion.” The lower of these two formations (Vermejo) contains a Montana flora, and ‘‘is more closely related to the Mesa Verde of western New Mexico than to any other formation we have examined.” The coal-bearing rocks of the Canyon City field are correlated with the Vermejo, which is thought to be approximately equivalent in age to the Fox Hills formation of the Denver basin. “The upper coal-bearing formation of the Raton Mesa region, to which the name Raton is here applied, is Eocene in age and contains a flora distinct from that of the Laramie of the Denver basin, but similar to that of the post-Laramie formations of that basin and to that of the Eocene Wilcox group of the Gulf coast.” “The unconformity between the Vermejo and Raton formations represents a time interval comparable to that ... . separating the Laramie from the Arapahoe of the Denver basin.” “The coal-bearing rocks of the Cerrillos, Hagan, Tijeras, and Rio Puerco fields are essentially equivalent in age to the Mesa Verde of the San Juan basin.” The plants of the formations described in this paper are discussed by Mr. F. H. Knowlton, who indicates that the flora of the Raton formation indicate “‘a relatively moist, warm situation, whose temperature did not fall much if any below 42°F,” and that the floral hiatus between the 478 REVIEWS 479 Vermejon and Raton floras can be explained only by the lapse of a very long period of time. Specifically Mr. Knowlton, on the basis of plant fossils, correlates the Raton formation with the Wilcox formation, and “probably with the Midway formation of the Gulf region.” The paper tends to confirm the conclusion which has been growing for many years that the Raton and equivalent formations are of Eocene— not of Cretaceous—age. The general conclusion is reached that the Raton formation is essen- tially equivalent to the Arapahoe, the Denver, the Dawson, the Fort Union, the Wilcox, and perhaps the Midway formations. Re DiS: My Reminiscences. By RaPHAEL PuMPELLY. New York: Henry Holt & Company, 1918. 2 vols. Pp. 844, maps, ills. In these two volumes the reader will find a most fascinating story of the very remarkable adventures and varied experiences which were crowded into the long life of this eminent American geologist. As the central figure was an inveterate traveler, roaming over a large portion of the globe when traveling was vastly different from what it now is, the work is first and foremost a book of travel. After studying at Freiberg and taking long vacation rambles through the mountains of Corsica and various other parts of Europe as the fancy struck him, Pumpelly returned to the United States and in 1860 began his professional work at the Santa Rita mines in Arizona. The Apache terror was then at its height, and Pumpelly alone of five successive superintendents of the mine was not murdered. Chance then took him to Japan, where in the employ of the Japanese government he conducted geological investigations and introduced certain improvements into Japanese methods of mining. It was but one step more to China, where a year and a half were devoted to private travels and geological explora- tion for the imperial government—experiences and researches which are vividly sketched. For the return journey to America several alternative routes were open, but with the true instinct of adventurous travel the author chose a winter journey across Mongolia and Siberia to Europe by saddle-horse and sleigh. In our own country the author’s most significant explorations were those in the Lake Superior region between 1867 and 1871, when the Lake Superior iron ores were beginning to attract attention. These reminis- cences are of special interest for the light they throw on the discovery and beginnings of the Menominee and Gogebic iron ranges. It was here in 480 REVIEWS 1871 that Pumpelly, owing to a combination of circumstances, declined the chance to take up all the even-numbered sections along twenty miles of the Gogebic iron range and thus, as he himself expressed it, ‘‘missed the opportunity of a lifetime.”’ In 1903-4 occurred what the author regards as the most interesting part of his hfe—namely, the Carnegie expedition to Turkestan under his leadership. The purpose was to test the hypothesis that Central Asia was the primitive home of the Aryan race, by studying the traces of prehistoric civilizations and seeking evidences of geological and climatic changes during and since the glacial period. In a way it was the climax of the author’s travels and professional work. Excavation at Anau in southern Turkestan revealed, among other things, the fact that the Stone Age inhabitants of this region used stone sickles with sharp cutting edges, while arrowheads and weapons of the chase were unknown to them. From this and other lines of evidence it seemed safe to conclude that the Neolithic peoples of Turkestan were not hunters but agriculturalists. Though sketching a scientific career, the topics treated are so spiced and stocked with amusing anecdotes that they must appeal strongly to the general reader as well as to the geologist. The narrative is lively and always interesting. R. tee THOMAS c CHAMBERLIN AND ROLLIN: D. SALISBURY . With ‘the Active Collaboration. of. 2 f Siang ss my ae Ww. Seitienow Vertebrate Paleontology ; ALBERT JOHANNSEN, Petrology Pn ie T WELLER, Invertebrate Paleontology - ROLLIN Te CHAMBERLIN, DuneaDe Geology oe ALBERT D. Ep Economic Geology a, ASSOCIATE EDITORS HIBALD GEIKIE, ha Britain Roane ; JOSEPH P. IDDIN GS, Ween D. c. id S BARROIS, ‘France 9 : JOHN C. BRANNER, Leland Stanford Junior Gaiecivy. al PENCK, Germany wed Sten) i RICHARD A. F. PENROSE, Jr, a Philadelphia, Pa. WILLIAM H. HOBBS, University of Michigan FRANK D. ADAMS, McGill University ; CHARLES K. LEITH, University of Wisconsin © GROVE we GILBERT, Washington, D.C. _ WALLACE W. ATWOOD, Harvard University CHARLES D. WALCOTT, Smithsonian Institution : WILLIAM H. EMMONS, University of Minnesota NEY. WILLIAMS, Cornell University ARTHUR L. DAY, Carnegie Institution ane -- SEPTEMBER-OCTOBER 1918 TWO- PHASE CONVECTION IN IGNEOUS MAGMAS - - - - FRANK F. Grout 481 ee CARBONIFEROUS CONDITIONS VERSUS PERMO- CARBONIFEROUS TIME i He CACASE 500 NOTES ON THE ‘GEOLOGY OF EASTERN GUATEMALA AND NORTHWESTERN % SPANISH HONDURAS: © 2° Sonije" 2 a as Ph SIMNEY) POWERS” 507 : NOTES ON A COLLECTION OF ROCKS FROM HONDURAS, CENTRAL AMERICA Witpur G. FovE 524 _ LOESS-DEPOSITING WINDS PIN LOUISEMNA Kj or 6 OS ee F. V. EMERSON 532 * VOLUME CHANGES IN METAMORPHISM =~ (= = = > Warpemar LinpcRen © §42 s DESCRIPTIONS OF SOME NEW SPECIES OF DEVONIAN FOSSILS w CiInton R. STAUFFER 555 ON THE NATURE AND ORIGIN OF THE STYLOLITIC STRUCTURE IN TENNESSEE BU RSENS Hs ae in ee ee A pt open a Se SS C. H. Gorpon 561 Pte, VALLEY CIFM GRABEN, UTAH = 0-00 5) 20h Se CL. Dake 4569 Bee SB EMER ee Ne ie caine: aimee Ne aban wie he ae ioe al seca RY tel, | ie as Gthe : RECENT: PUBLICATIONS © = te). (55 04)- + peas = a eB Sia dss = 25 570 TIE UMIVERSITY OF CHICAGO PRESS | CHICAGO, ILLINOIS, U.S.A. THE CAMBRIDGE UNIVERSITY PRESS, Lonpon anp Epinsurca THE MARUZEN-KABUSHIKI-KAISHA, Toxyo, Osaka, Kyoto, FukuoKA, SENDAI ; THE MISSION BOOK COMPANY, SHancHAI price is ant 00 pee year; he price of oak copies ds 5 ee ene (Orie for | service 0 less ee will be ee at the single-copy rate. { Postage is prepaid by: the epubisle ( 2 for all other countries in the Postal Union, 53 cents on annual subscriptions (total $4. ; 53 on s copies" rr cents (total 76 cents.) {[ Patrons are requested to make all remittances payepl to Th a versity of Chicago Press in postal or express money orders or bank drafts. : The following are authorized to quote the prices indicated: see : For the British Empire: The Cambridge University Press, Fetter Lane, Toten E. subscriptions, including postage, 18s. 8d. each; single copies, including postage, 35. ad. each. For Japan and Korea: The Maruzen-Kabushiki-Kaisha, 11 to 16 Nihonbashi Tori Tokyo, Japan. Yearly subscriptions, including esiaee Yen 9.00 each; single con a h - postage, Yen 1.50 each. For China: The Mission Book Company, 18 Peking Road, Shanehat. Vearly sulin @ - $4.00; single copies 65 cents, or their equivalents in Chinese money. Postage extra, if mailed direct ; Suicide of Shanghai, on yearly subscriptions 53 cents, on single copies 11 cents. : : Claims for missing numbers should be made within the month following the regular aah of publi-- oi cation. The publishers expect to supply missing numbers free only when losses have been sustained in _ transit, and when the reserve stock will permit. Deliveries are not guaranteed within the war zones of 5 Europe, nor in other countries where transportation is disturbed by conditions incident to war. Business correspondence should be addressed to The University of Chicago Press, Chicago, Ill. Communications for the editors and manuscripts should be addressed to the Editors” of Tae JouRNAL or GEotoGcy, the University of Chicago, Chicago, Ill, : Contributors are requested to write scientific and proper names with particular care and i in citations : + to follow the form shown in the pages of the Journal. ~ jae dE ors: Separates, if desired, must be ordered in advance of publication. Separates of leading articles will be printed as ordered, of which 50 (without covers) will be furnished gratis, the actual cost of the remainder iz (and covers, if desired) to be paid for by the author. The actual cost may vary, and will depend upon " the amount of work in remaking the pages into forms, presswork, paper, binding, etc. Separates con- taining half-tones may be expected to cost somewhat more, the increase depending upon the number 7 of cuts and the amount of work required upon them. ; a rat Entered as second-class matter, March 20, 1893, at the Post-office at Chicago, Ill., under the Act of March 3h 1879. Acceptance. for mailing at special rate of postage provided for in Section 1103, Act of October 3, ror7, authorized on in July 15, 1918. ‘