"^ i k mil tin » » » ^ » » » »^'«» •9JBD mjM auinjoA Sim 3[puEi{ 3SB3[J ■♦» » »"» » » » » » fmrnm Digitized by tine Internet Archive in 2009 with funding from Boston Library Consortium IVIember Libraries http://www.archive.org/details/lakebonneville01gilb DEPARTMENT OF THE INTERIOR MONOGRAPHS United States Geological Survey VOLUME I WASHINGTON GOVERNMENT PRINTING OFFICE 1890 U. 8. GFOLOGICAL SURVEY SHORE-LINES ON THE NORTH i' °^ ^^^ OQUIRRH RANGE, UTAH. . H. Holmes. VoV w UNITED STATES GEOLOGICAL SURVEY J. W. POWELL, DIRECTOR LAKE BONNEVILLE BT GROVE K^RL GILBERT WASHINGTON GOVERNMENT PRINTING OFFICE 1890 CONTENTS Fage. Letter op Transmittal , xv Preface xvii Abstract of Volume xix C'liAi'TER I. — Introduction 1 Interior Basins 2 The Great Basin 5 History of Investigation.. 12 Tlie Bonneville Basin 2U Cbronologic Nomenchitnro 22 Chapter II.— The ToPOGRArnic Features of Lake Shores 23 Wave Work 2'J Littoral Krosion 29 The Sea Cliff 34 The Wave-cut Terrace 35 Littoral Transportat ion 37 The Beach , 39 The Barrier 40 The Subaqueous Kidgo 43 Littoral Deposition 4t) Eiu bankments 46 The Spit 47 The Bar 48 The Hook 52 The Loop 55 The Wave-built Terrace 55 The V-Torracoand V-Bar .. 57 Drifting Sand ; Dunes 59 The Distribution of Wave-wrought Shore Features 60 Stream Work; the Delta 65 Ice Work ; the Rampart 71 Submergence and Emergence 72 The Discrimination of Sliore Features 74 Cliffs 75 The Cliff of Differential Degradation 75 The Stream Cliff 75 The Coulde Edge 76 The P^ault Scarp 76 The Land-slip Cliff 77 Comparison 77 Terraces 78 The Terrace by Differential Degradation 78 The Stream Terrace 79 The Moraine Terrace 81 VI CONTENTS. Page. Chapter II— The Topographic Features op Lake Shores— Continued. The Fault Terrace 83 The Laud-slip Terrace 83 Comparison S4 Ridges 86 The Moraine 86 The Osar or Kame 87 Comparison 87 The Recognition of Ancient Shores 88 Chapter III. — Shores of Lake Bonneville 90 The Bonneville Shore-line 93 The Question of a Higher Shore- line 94 More Ancient Lakes 98 Outline of the Lake 101 Extent of the Lake 105 Shore Details 106 Embankment Series Ill Determination of Still- water Level 122 Depth 125 The Map 125 TheProvo Shore-line 126 Outline and Extent 127 Shore Characters 128 Deltas 129 The Underscore 130 Embankment Series 131 The Map 134 The Stansbury Shore-line 134 The Intermediate Shore-lines 135 Description of Embankments 135 Grantsville 135 Preuas Valley 136 The Snow-plow 137 Stockton and Wellsville 137 Dove Creek 137 Comparison of Embankments - 137 Hypothesis of Differential Displacement 140 Hypothesis of Oscillating Water Surface 141 Superposition of Embankments 147 The Snow-plow 147 Reservoir Butte 148 Stockton , 149 Blacksmith's Fork 151 Dove Creek 151 Double Series in Preuss Valley 152 Deltas 153 American Fork Delta 155 Logan Delta 159 Summary - ICG Tufa 167 R4sum6 161) Chapter IV.— The Outlet 171 Red Rock Pass 173 Mar.sh Valley 176 The River 176 CONTENTS. VII Page. Chaptkr IV.— The Outlet — Continued. The Gate of Bear Kiver 178 The Question of an Earlier Discharge 180 The Old River Bed 181 Other Ancient Rivers 184 Outlets and Shore-lines 186 Chapter V. — The Bonneville Beds 188 Lower River Bed Section 189 Lemington Section - 192 Upper River Bed Section 194 Yellow Clay 194 First Gravel '- 194 White Marl 195 Lower Sand 195 Second Gravel 195 Upper Sand 196 Upper Gravel 196 Oscillations of Water Level — 196 Height of the First Maximum 199 The Whiteness of the White Marl 200 Source of Material 203 Composition of Lake Water 204 Experiments 205 Deposition by Desiccation 208 Organic Remains 209 Joint Structure 211 Chapter VI.— The History of the Bonneville Basin 214 The Pre-Bonneville History 214 Alluvial Cones and Aridity 220 The Post-Bonne ville History 222 Subdivision of the Basin 222 Snake Valley Salt Marsh , 223 Sevier Lake 224 Salt Bed 225 Rush Lake 228 Great Salt Lake 230 Surveys 230 Depth 230 Gauging 230 Oscillations since 1875 233 Oscillations prior to 1875 239 Changes in area 243 Causes of Change 244 Future Changes 250 Saline Contents 251 Sources of Saline Matter 254 Rate and Period of Salt Accumulation 255 Fauna 258 The General History of the Bonneville Oscillations 259 The Topographic Interpretation of Lake Oscillations 262 Hydrographic Hypothesis 263 Orogenic Hypothesis 263 Epeirogenic Hypothesis 264 The Climatic Interpretation of Lake Oscillations . 265 Opinions on Correlation with Glaciation 265 VIII CONTENTS. Page. Chapter VI. — The History of the Bonneville Basin— Coiitintiod. The Argument from Aualogy 2C9 Recency i!69 Episodal Character 269 Bi partition 270 Genetic Correlation , 275 The EtFect of a Change in Solar Energy 283 The Evidence from Molluscan Life 297 Depauperation and Cold :{00 Depauperation and Salinity 'Ml The Evidence from Vertebrate Life :50:{ The Evidence from Encroaching Moraines ;;()5 Wasatch-Bouueville Moraines 30G Siena-Mono Moraines 311 Summary of Chapter 310 Chapter VII. — Lake Bonneville and Volcanic Eruption 319 Ice Spring Craters and Lava Field 320 Pavaut Butte 325 Tabernacle Crater and Lava Field 329 Pleistocene Winds • 332 Funiarolo Butte and Lava Field 332 Other Localities of Basalt 335 Pleistocene Eruptions Elsewhere 3.36 Rhyolite 337 Summary and Conclusions 33s Chapter VIII. — Lake Bonneville and Diastrophism 340 Evidence from Faulting; Fault Scarps 340 General Features of Fault Scarps 354 Local Displacements versus Local Loading and Unloading 357 Monutaiu Growth 359 Earthquakes 360 Evidence from Shore-lines 362 Measurements 362 Deformation of the Bonneville Shore-liue 365 Deformation of the Provo Shore-line 371 Deformation during the Provo Epoch 372 Postulate as to the Cause of Deformation - 373 Hypothesis of Gcoidal Dcformatiou 376 Hypothesis of Expansion froui Warming 377 Hypothesis of Terrestrial Deformation by Loading and Unloading 379 Evidence from the Position of Great Salt Lake 384 The Strength of the Earth 387 Chapter IX.— The Age of the Equus Fauna 393 The Fauna and its Physical Relations 393 The Paleontologic Evidence 397 Appendix A.— Altitudes and their Determination. By Albert L. Webster 405 Scheme of Tables 405 Trigonometric Data 406 Barometric Data 406 Lake Records 409 Railroad Records 411 Special Spirit-level Determinations 411 Combination of Data 413 Altitudes of Shorelines and their Differences 416 CONTENTS. IX Page. Appendix B. — On the Deformation of the Geoid by the Removal, through Evapo- ration, OF THE Water of Lake Honnevili.e. By R. S. Woodward 4^)1 Appendix C— On the Elevation of the Surface of the Bonnkvillb Basin by Ex- pansion DUB TO change OF CLIMATE. By R. S. Woodward 42.') Index 407 TABLES. Table I. Dimensions of Lakes 106 II. Enibaukinent Series of the Bonueviilo Shore- line 119 III. Analyses of Bouiievillo Sediments 201 IV. Con that lake has repeatedly risen .and fallen throngh a range of 10 feet.— The history of Lake Bonneville is par- alleled by that of Lake Lahontan, and each is connected willi a history of glaeiation in adjacent iniinntains. This connection, the deiianperatiou of the fossil shells, and an analysis of the climatic conditions of glaeiation, lead to the conclusion that the lacustrine ejiochs were epochs of relative cold. Chaptkr VII: Lakr Boxnkvillkand Volcanic Eruption. —The group of small craters and basaltic lava fields near Fillmore, Utah, are closely related to the lake history. Somi; eruptions took place beneath the water of the lake, others since its disappearance, and others again during the inter- lacustrine epoch. — Numerous basaltic ernptions occurred in the lake area before the lake period, and at still earlier dates rhyolite was extra va.sated. Cii.M'Tku VIII: Lakk Bonnevii.lk and Diastropiiism. — Orogenic change during a period subseijueut to the lake is shown by fault scarps. The formation of fault scarps is accompanicil by earlh- i|uakes. — Epeirogeuic change during a period snbseifi)i(')/f(il, or i)!f('rio>; or eJoscfl, or shut, or drainless basin. If an interior basin exists in a climate so arid that the superficial ffow of water, which constitutes drainage, is only potential and not actual, or else is occasional only and not continuous, it contains no })erennial lak-'^ and is called a dry basin. INTERIOR BASINS. 3 The boundaries separating basins are water-partings or divides, and these are of all characters, from the acute crests of mountain ranges to low rolls of the plain scarcely discernible by the eye. Interior basins are com- ])letelY encncled by lines of water-parting. The existence of interior Ijasins depends on two conditions: a suitable topographic configuration and a suitable climate. The ordinar}^ process of land sculpture by running water does not produce cup-like basins, but tends on the contrary to abolish them. Wherever a topographic cup exists the streams flowing toward it deposit within it their loads of detritus, and if they are antagonized by no other agent eventually till it. If the cup con- tains a lake with outlet the outflowing stream erodes the rim of the basin, and eventually the lake is completely drained. The work of streams occasionally produces topographic cups by the rapid formation of alluvial deposits where two streams meet. If the power of one stream to deposit is greatly increased, or if the power of the other stream to erode is greatly diminished, the one may build a dam athwart the course of the other and thus produce a lake basin. The great agent in the production of lake basins, or the agent which has produced most of the large basins, is diastrophism,^ and in a majority of the cases in which basins are partitioned off by the alluvial process just described, the change in the relative power of the streams is brought about by diastrophism. Other basin-forming agencies are volcanic eruption, limestone sinks, wind waves, dunes, land slides and glaciers. By far the greatest number of topographic cups are due to glaciers; but with these we are not now concerned. The basins of ordinary lakes are distinguished from interior basins by overflow, and that depends on climate. The rainfall of each basin is or may be disposed of by three processes: first, evaporation from tlie soil and 'I finil it advantageous to follow J. W. Powell in the use of diastrophism as a general term for the process or processes of deformation of the earth's crust. The products of diastrophism are conti- nents, plateaus and mountains, ocean beds and valleys, faults and folds. Diastrophism is coordinate with voleanism, and is the synonym of displacement and dislocation in the more general of the two geologic meanings accjuired by each of those words. Its adjective is diastropldc. It is convenient also to divide diastrophism into orogeny (monntaiu-making) .and epeirogeny (continent -making). The words epeirogeny and epeirogenic are defined iu the opening paragraph of chapter VIII. 4 LAKE BONNEVILLE. from the vegetation supported by it; second, evaporation from a lake sur- face; third, outflow. If the rainfall is sufficiently small, it is all retiu'ned to the air l)y evaporation from the s. 4ril-J,'i2. 2U. S. Geol. Surv. of Wyomiuj,'. . . 1870, by F. V. Hayilen. Wa.sliiut;ti)ii, 187'2, pp. 161), 170, 172, 175. ^ Report- of Frauk H. Bradley, in U. S. Geol. Surv. of the Territories, Ropt. for 1S72. Washing- ton, 187:!, pp. 192, 11)6. ■•The Great American Desert, by Henry S. Poolo ; Proc. Nova Scotia Inst. Nat. Sci., vol. 3, pp. 208-220. WHITNEY, KING, HAYDEN. 17 history of the subject, but, as ah-ead}- mentioned, they were partially an- ticipated by those of Simpson and Engelmann, and wholly anticipated by those of King, Hague and Emmons, the geologists of the Fortieth Parallel Exploration. The work of tliis corps covered a belt one hundred miles broad, spanning the Great Basin in its broadest part, and within this belt the Pleistocene lakes were studied and for the first time approximately mapped. It was shown that the corrugated surface of the Great Basin in this latitude is higher in the middle than at the east and west margins, war- ranting a general subdivision into the Utah Basin, the Nevada Plateau and the Nevada Basin; that the Utah Basin formerly contained a large lake, Bonneville, extending b(jtli north and south beyond the belt of survey; that the Nevada Basin contained a similar lake, Lahontan, likewise exceed- ing the limits of the belt; and that the valleys of the central plateau held within the belt no less than eight small Pleistocene lakes. The mechanical sediments and chemical deposits of the lakes were studied, and were ascer- tained to overlie subaerial gravels, thus proving that a dry climate had pre- ceded the humid climate of the lake epoch; and it was inferred from the chemical deposits of Lake Lahontan that the lake had been twice formed and twice dried away.^ The field work that afforded this important body of information was performed chiefly in the years 1867-70, but publication was delayed till 1877-78. In 1872 Howell and the writer, traveling Avith topographic par- ties of the Wheeler Survey, traversed the Utah Basin on many lines, and our reports, printed in 1874 and 1875, contained an account of Lake Bon- neville, the extent of Avhicli we were able to indicate with inconsiderable error, and to which the writer gave a name.^ Thus, by an accident of pub- lication. King and his colleagues lost that literary priority in regard to Lake Bonneville to which they were fairly entitled by priority of investigation. 'Geol. Expl. of the 40th Parallel. Vol. 1, Systematic Geology, by Clarence King. Washington, 1878; vol. 2, Descriptive Geology, by Arnold Hague and S. F. Emmons. Wasliington, 1877. * Prelim. Geol. Rept. by G. K. Gilbert; Appendix D to Progress Rept. Expl. and Sur. W. of the 100th Mer. in 1872. Washington, 1874, pp. 49-50. Explorations* and Surveys west of the 100th Meridian, vol. 3, Geology. Washington, 1H75. Part 1, by G. K. Gilbert, treats of Lake Bonneville on pp. 88-104. Part 3, by Edwin E. Howell, treats of Lake Bonneville on pp. 249-251. MON I 2 18 LAKE BONNEVILLE. In 1877 Pealc observed shore terraces in various parts of (Jaolie valley.' From 1875 to 1878 I spent each summer in Utah as a mem1)er of the Powell survey, and found many ()i)it(trtunities in connection with other work to continue the study of Lake Bonue\'ille. This \vas especiall}' the case in 1877, when the duty of gathering inforaialion as to the irrigable land of the basin of Great Salt Lake led me all about the margin of the Salt Lake desert. When the corps for western surveys were reorganized in 1879, 1 was placed in charge of the Division of the Great Basin, with tlie understanding that the Pleistocene lakes, previously investigated only in an incidental way, should fonn a principal subject of study. Late in the season some months were spent in the field, with Mr. W. D. Johnson as assistant; and a corps was oi'ganized the following year. Of this coi'ps, Mr. Israel C. Russell was \)vin- cipal assistant, and he remained with the work from first to last, l^eing assigned independent investigations after the first season, ilessrs. H. A. Wheeler, W. J. McGee, and Geo. M. Wright took pai-t in the geologic work for shorter periods. Messrs. Gilbert Thompson, Alljcrt L. Webster, Willard D. Johnson, and Eugene Ricksecker, associated with the work at various times as topographers, and Messrs. Fred. D. Owen, J. B. Bernadou, and E. R. Trowbridg-e, temporarily attached to field jjarties as general assistants, all contril^uted to the mapping and illustration of the lake jjhenomena. The field work of the year 1880 was in the Bonneville Basin, and little was afterAvard done in that area. In 1881 Mr. Russell made a ])relinnnary e.\annnation of the vestiges of Lake Lahoiitau in the Nevada IJasin and of the Mono Basin, aiid in the following spring extended his reconnaissance to the lake basins of southeastern Oregon. I was called to Washington in the spring of 1881 on duty supposed to be temporai-y, but remained there until the following year, w hen the work of the Surve}, pre\iously restricted to the western l\'rritories, was extended l)y C<->ngress to the eastern States also. As the enlargement of field and function was not accompanied l)yan equivalent increase of funds, it became necessary to curtail the western work of the Survey, and it was decided to stop the investigation of the Pleistocene lakes as soon as this could lie done witliout •'•reat sacrifice of ' Keport, of A. C. Peale, in U. S. GeqJ. Surv. of tho Territories for 1877, Washiugtou, 1879, pp. 603-606. U S. GEOLOGICAL SURVKY LAKE BONNE^/ILLE PL m 113° 112° 111° 42«^ 41' 39" 38' MAP OF 'L/\KE BONNEMLLE showing ROUTES OF THA\^EL Routes by G. K-Gilbert ^^ "* JVdditional routes by Assistants Julius Bien A Iji.lith Drown by C TfaompBon RUSSELL, DANA, GALL. 19 material already acquired. Mr. Russell completed the study of the Lalion- tau and Mono Basins by the close of the season of 1883 and then returned east. I made a single excursion in the summer of 1883, devoting a few weeks to supplementary observations in the Bonneville, Lahontan, and Mono Basins, and visiting Owens Valley to examine the geologic features of the Inyo earthquake. The examination of the more southerly valleys of the Great Basin, the study of the brines and saline deposits, and the elaborate measurement of post-Pleistocene displacements, are indefinitely deferred. The results of the investigation have been communicated in a series of reports, essays, and memoirs. An outline of the Bonneville history was published by me in 1882,' and an essay on shore topography in 1885.^ Russell's results have appeared in a preliminary report on Lake Lahontan,^ reports on the Oregon basins'* and the Mono Basin, ^ and a monograph on Lake Lahontan.^ An essay on the Pleistocene fresh-water shells ^vas pre- pared and published by Call,' and one on the pseudomorph thinolite by Dana.* The present publication completes the series. 'Coutributions to the history of Lake Bonneville: Second Ann. Kept. U. S. Geol. Survey. Wash- ington, 1882, pp. 169-200. = The topographic features of lake shores: Fifth Ann. Kept. U. S. Geol. Survey. Washington, 1885, pp. 7.5-123. Adiscnssion of post-Bonneville displacement appeared in an address "The luculcation of Scien- tific Method by Example," read to the American Society of Naturalists Dec. 27, 1885, and printed in the Am. Jour. Sci., vol. 31, pp. 284-291). A description of the jointed structure of the Bonneville beds was printed in the Am. Jour. Sci., 3d Series, Vol. 23, 1882, pp. 25-27. 'Sketch of the Geological History of Lake Lahontan : Third Ann. Kept. U. S. Geol. Survey. Wash- ington, 1883, pp. 189-235. ^A geological reconnaissance in Southern Oregon: Fourth Ann. Rept. U. S. Geol. Survey. Wash- ington, 1885, pp. 431-164. •'* Quaternary history of Mono Valley, California : Eighth Ann. Rept. U. S. Geol. Survey. Washing- ton, 1880, pp. 261-394. 15 Geological History of Lake Lahontan : Mon. U. S. Geol. Survey, No 11, Washington, 1885, pp. 302. Other publications by Mr. Kussell containing portions of the same material are — Lakes of the Great Basin: Science, vol. 3, 1884, pp 322-323. Dejiosits of Volcanic Dust in the Great Basin : Bull. Phil. Soc, Washington, vol. 7, 1885, pp. 18-20. Notes on the Faults of the Great Basin, . . . : Bull. Phil. Soc. Washington, vol. 9, 1887, pp. 5-8. The Great Basin. In Overland Monthly, 2d Series, vol. 11, 1888, pp. 420-426. 'On the Quaternary and Recent MoUnscaof the Great Basin, with descriptions of new forms. By E. Ellsworth Call. Introduced by a Sketch of the Quaternary Lakes of the Great Basin, by G. K. Gilbert. Bull. U. S. Geol. Survey No. 11, 1884, 56 pp. "A Crystallographic Study of the Thinolite of Lake Lahontan. By Edward S. Dana. Bull. U. S. Geol. Survey No. 12, 1884. 29 pp. 20 LAKE BONNEVILLE. THE BONNEVILLE BASIN. The Great Basin comprises a large number of sul^sidiary closed basins, each draining to a lake or ])laya. Aliout sixty ot" these could be enuiiicnited from present knoAvledge, and the fvdl nuiiil)er may be as high as one luni- dred. In the last geologic epoch a more humid climate (■(inverted iii;iiiy, or perhaps all, of these playas into lakes, and enlarged all the lakes. Some lakes overflowed the rims of their basins, becoming tributary to others; and the lakes of adjacent basins in many instances expanded until they l)ecaine continent. A few of the overflowing lakes discharged across the rim of the Great Basin, thus becoming tributary to the ocean, and subtracting their catchment basins from the district of interior drainage. In the remaining portion of the district the nmnber of independent drainage areas A\'as reduced by coalescence. The laro-est of the confluent lakes Avere formed at the eastern and western raai-gins of the Great Basin, being separated by the plateau of eastern Nevada. Lake Lahontan at the west was fed chiellA- In' the snows of the Sierra Nevada, Lake Bonneville at the east by those of the Wasatch and Uinta mountains. The catchment basin of Lake Bonneville comprises that part of the Great Basin lying east of the Gosiute, Snake, and Piiion mountains of east- ern Nevada — an oblong ai-ea embracing about five degrees of latitude and three of longitude, and containing about 54,000 squai-e miles, or the fourth part of the area of the Great Basin. Its western two-thirds may be described as a plain ranging in altitude from 4,200 to 5,500 feet above tide, and more or less interrupted by short moiuitain ranges trending north and south. At the north, where the mountains are comparatively few and small, the barren plain is called the Great Salt Lake Desert, and similar open stretches at the south are named the Sevier Desert and the Escalante Desert. The eastern third is much higher, including the lofty Wasatch Range and its dependen- cies, the western end of the still loftier Uinta Range, and the western jjart of the district of the High Plateaus. Several peaks of the Wasatch and Uinta Mountains rise above the level of lL*,000 feet, and the High Plateaus culminate near Beaver in the Tusliar ridge with peaks of similar altitude. THE BONNEVILLE BASIN. 2l The eastern uplands are the only important condensers of moisture, and from them flow a sj'stem of rivers whose Avaters are eva])orated in the salt lakes of the lowlands. The Bear, the Weber, and the Provo-Jordan have their principal sources in the Uinta Mountains, and break through the Wasatch Range on their way to Great Salt Lake. One of the upper val- leys traversed by the Bear River contains Bear Lake, a body of fresh water; and Utah Lake, likewise fresh, receives the Provo and discharges the Jor- dan. The Sevier River, after flowing 1,50 miles nortlnvard among the plateaus, receives the San Pete from the nortli and then turns westward to Sevier Lake, the saline of the Sevier Desert. The eastern uplands are better timbered than any other part of the Great Basin. The upland valleys are fertile, but having a climate too cool for agriculture are devoted to grazing and maintain only a scant population. The western plain is infertile by reason of aridity, and is almost without inhabitants. The lower valleys of the rivers, where they issue from the uplands upon the plain, have a climate suited for agriculture, are rendered fertile by irrigation, and constitute a habitable zone, over which the Mor- mon community has spread. To understand fully the topographic relations described above, the reader should examine the large map of Lake Bonneville (in a pocket attached to the cover of this volume), where the reliefs are expressed by contour lines at each 1,000 feet; and also Plate XII, whereon are marked the boundary of the Bonneville Basin and tlie boundaries of the equivalent group of smaller basins as they exist at the present time. He will find also that the plate supplements the expression of the distribution of the uplands, by contrasting the area above 7,000 feet with the area below; and he can learn from it more readily than through words the relation of the basin to the political diA-isions of the country. By turning again to Plate II he will see that the Bonneville basin adjoins interior drainage only on the west; its northern rim parts it from the basin of Snake River, a branch of the Columbia, its eastern and southern from the basin of the Colorado of the West. The more important streams heading near the northern rim and flowing to the Snake are the Salt, Blackfoot, Portneuf, Bannack, and Raft. In the eastern rim rise Black's Fork, the Uinta, and the Price, all tributary 22 LAKE BONNEVILLE. to the Green before it joins the Colorado, tind the San Rafael, Fremont, and Escalante, immediate tributaries of the Colorado. The Paria, Kanab, and Virffen flow to the Colorado from the southern rim. CHRONOLOGIC NOMENCLATURE. The geologic period to which the Bonneville history has been referred has three names in good standing, Quaternary, Pleistocene, and Glacial. Each name varies more or less in scope as used by different authors, but as ordinarily understood the three are strictly synonymous. In earlier Avrit- ings I have preferred Quaternary, in the present I prefer Pleistocene. No vital principle is involved in either preference, and indeed I am not of those who clamor for the rights of words. In my judgment words have no rights which the users of words are bound to respect. The claim of a word for preference rests only on its utility — its convenience for the com- munication of thought. Glacial connotes glaciers, and was a convenient name while it was sup- posed that a cold climate marked the whole period. But now that interrup- tions of that climate are recognized, it is more convenient to speak of glacial epochs and interglacial epochs of the Quaternary or Pleistocene })eriod. Quaternary connotes a fouribld classification, and is coordinate \vith Tertiary. Pleistocene suggests by its termination coordination with the subdivisions of the Tertiary. Using the scale of time-nouns adopted by the International Congress of Geologists, the Quaternary is an era, having the classificatory rank of the Tertiary era, and the Pleistocene is a period, rank- ing with the Eocene period. It is generally believed that the Pleistocene is comparable in ])oint of diiration with one of the periods of the Tertiar}* era, being less rather than greater, and those who advocate the emplopnent of the name Quaternary recognize the Quaternary era as one containing but a single period. The time division with which we have to deal is, then, from eveiy point of view, a "period," and it is believed that the use of the name Pleistocene Period involves a minimum amount of implication as to higher classification, a subject whose discussion is not here contemplated. CHAPTER II. THE TOPOGRAPHIC FEATURES OF LAKE SHORES. It has been assumed in the i)veceding- pag-es that valleys trom which lakes have recently disappeared are characterized l)y certain features wherebv that lact can he recognized. Perhaps no one observant of natural phenomena will disjjute this. But there is, nevertheless, some diversity of opinion as to what are the peculiar characters to which lakes give rise; and especially has the true interpretation of certain local topographic features been mooted, some geologists ascribing them to waves, and others to dif- ferent agencies. In the investigation of our ancient lake, it has been found necessary not only to discriminate from all other topographic elements the features created by its waves, but also to ascertain the manner in which each was produced, so as to be able to give it the proper interpretation in the recon- struction of the history of the lake. It is proposed in this chapter to pre- sent the more general results of this study, describing in detail the various elements which constitute shore topography, explaining their origin, so far as possible, and finally contrasting them with topographic features of other origin which so far simulate them as to occasion confusion. The play of meteoric ag-ents on the surface of the land is unremitting, so that there is a constant tendency to the production of the forms charac- teristic of their action. All other forms are of the nature of exceptions, and attract the attention of the observer as requiring explanation. The shapes wrought by atmospheric erosion are simple and symmetric and need but to be enumerated to be recognized as normal elements of the sculpture 23 24 LAKE BONNEVILLE. of the land. Along each di-aiuage line there is a gradual and gradually increasing ascent from mouth to source; and this law of increasing acclivity applies to all branches as well as to the mnin stem. Between each ))!iir of adjacent drainage lines is a ridge or hill, standing midway and roundcMl at the top. Wherever two ridges join there is a sunnnit higher than the adja- cent portion of either ridge; and the highest summits of all arc those which, measuring along lines of drainage, are most remote from the ocean. The crests of the ridges are not horizontal but undulate from summit to summit. There are no sharp contrasts of slope; the concave profiles of the drainage lines change their inclination little by little and merge by a gradual transi- tion in the convex profiles of the crests and sununits. The factor Avhich most frequently, and in fact almost universally, inter- rupts these simple curves is heterogeneity of terrane. Under the infiuence of this factor, just as in the case of a homogeneous terrane, the declivities adjust themselves in such way as to oppose a maxinmm resistance to erosion; and with diversit}' of rock texture this adjustment involves diversity of form. Hard rocks survive, while the soft ai"e eaten away. Peaks and clifts are })ro1, pp. 49-117. 'Toronto Harbor — its fonnation and preservation. By Sandford Fleming, C. E. : Canadian Journal, vol. 2, ls54, pp. 103-107, 223-230. Reprinted witb additions as Report on Preservation and Improvement of Toronto Harbor. In Supplen^ent to Canadian Journal, 1854, pp. 15-29. ■"Tbe North American Lakes considered as cbronometers of post-Glacial time. By Dr. Edmund Audrows. Trans Chicago Acad. Sci., Vol. 2, pp. 1-23. i^On the reclamation of tide-lands and its relation to navigation. By Henry Mitchell. Appen- dix No. 5, to Rept. U. S. Coast .Survey for l-^GO. Washington, 1872, pp. 75-104. «Snl moto ondoso del mare e sn le correnti di esso specialmente sii (juello littorali. Alessandio Cialdi, Rom.n, 18G6. 'Studien nberdie Gestaltung dcr Sandku.sten, etc., 11. Keller, Berlin, 18S1. 'Fiihrer fiir Forscbuugsrciscnde, vou Ferdinand Freiherr von Richthofen. Berlin, 1880, pp. 336-365. 9 Report of the Committee on Waves, by Sir John Robinson, and John Scott Russell, Reporter: Rept. British Ass. Adv. Sci., 7th meeting, 1837, pp. 417-496. '"G. B. Airy, Vol. V, Ency. Metrop. "W. J. McQ. Rankine, Philos. Trans. Royal Soc. London, vol. 153, 1863, pp. 127-13-'. LAND SHAPING AND SHORE MAKING. 27 In the following' treatment, of the subject the descri])tion and analysis of the elements of shore topography will be followed by a comparison of certain of these elements with sinuilating features of ditferent origin. First, however, a few words will be devoted to the consideration of shore shaping as a di\'ision of the more general process of earth shaping. The earth owes its spheroidal form to gravity and rotation. It owes its o-reat features of continent and ocean bed to the mieciual distribution of the heterogeneous material of which it is composed. Many of its minor inequalities can be referred to the same cause, but its details of siu'face are chiefly molded by the circulation of the fluids which envelope it. This shaping or molding of the surface may be divided into three parts — sul)- aerial shaping (land sculpture), subaqueous shaping, and littoral shaping. In each case the process is threefold, comprising erosion, transportation, and deposition. In subaerial or land shaping the agents of erosion are meteoric — rain, acting both mechanically and chemically, streams, and frost. The agent of transportation is running water. The condition of deposition is diminishing velocity. In subaqueous shaping, or the molding of surface which takes })lace beneath lakes and oceans, currents constitute the agent of erosion. They constitute also the agent of transportation; and the condition of deposition is, as before, dimiiaishing velocity. In littoral shapings or the modeling- of shore features, waves constitute the agent of erosion. Transportation is performed by waves and cuiTents acting conjointly, and the condition of deposition is increasing depth. On the land the amount of erosion vastly exceeds the amount of dep- osition. Under standing water erosion is either nil or incomj)arably inferior in amoiint to deposition. And these two facts are correlatives, since the product of land erosion is chiefly deposited in lakes and oceans, and the sediments of lakes and oceans are derived chiefly from land erosion. The products of littoral erosion undergo division, going partly to littoral dep- osition and partly to subaqueous deposition. The material for littoral deposition is derived partly from littoral erosion and partly from land erosion. 28 LAKE BONNEVILLE. That is to say, the detritus worn from the land by meteoric agents is transported outward by streams. Normally it is all carried to the coast, but owing to the almost universal complication of erosion with local uplift, there is a certain share of detritus dep the left, and vice versa. In every such case the direction of transportation is the direction of the littoral current. The waves and undertow accomplisli a sorting of the detritus. The finer portion, being lifted up by the agitation of the waves, is held in sus- pension until carried outward to deep water by the undertow. The coarser portion, sinking to the bottom more rapidly, can not be earned beyond the zone of agitation, and remains as a ])art (tf the shore. Only the latter is the subject of littoral transportation. It is called tihore drift. With the shifting of the wind the direction of the littoral cun'ent on any lake shore is occasionall}-, or it niay be frequently, reversed, and the shore drift under its influence travels sometimes in one direction and some- times in the other. In most localities it has a ])revailing direction, not nec- essarily determined l)y the prevailing direction of the shore current, but THE HIGHWAY OF THE SHORE DRIFT. 39 rather by the direction of that shore current which accompanies the greatest waves. This is frequently but not always the direction also of the shore current accomjianing the most violent storms. The source of shore drift is two-fold. A large part is derived from the excavation of sea-cliffs, and is thus the product of littoral erosion. From every sea-cliff a stream of shore drift may be seen to follow the coast in one direction or the other. Another part is contributed by streams depositing at their mouths the heavy part of their detritus, and is more remotely derived from the erosion of the land. The smallest streams merely reinforce the trains of shore drift flowing from sea-cliffs, and their tribute usually cannot be discriminated. Larger streams furnish bodies of shore drift easily referred to their sources. Streams of the first magnitude, as will be explained farther on, overwhelm the shore drift and produce structures of an entirely different nature, known as deltas. The Beach—The zoue occupied by the shore di-ift in transit is called the heach. Its lower margin is beneath the water, a little beyond the line where the great storm waves break. Its upper margin is usually a few feet above the level of still water. Its profile is steeper upon some shores than others, but has a general facies consonant with its wave-wrought origin. At each point in the i^rofile the sloije represents an equilib- ^ ^ ^ 1 Fig. 4.— Section of a Beach. rium in transjjorting power between the inrushing breaker and the outflowing undertow. Where the undertow is relatively potent its efficiency is diminished by a low declivity. • Where the inward dash is relatively potent the undertow is favored by a high de- clivity. The result is a sigmoid profile of gentle flexure, upwardly convex for a short space near its landward end, and concave beyond. In horizontal contour the beach follows the original boundary between land and lake, but does not conform to its irregularities. Small indentations are filled with shore di'ift, small projections are cut away, and smooth, sweep- ing curves are given to the water margin and to the submerged contours within reach of the breakers. 40 LAKE BONNEVILLE. The lieach graduates insensibly into the wave-cut terrace. A cut-terrace lying in the route of shore drift is alternately Imried ^)y drift and swept bare, as the conditions ot wind and breaker vary. Tlie cut-and-built ter- race (Figure .'')), which owes its detrital extension to the agencies detennin- ing the beach profile, may be regarded as a forai intermediate between the beach and the ciit terrace. The Barrier. -Where the sublittoral bottom of the lake has an exceedingly gentle inclination the waves break at a considerable distance from the water Fig. 5.— Section of a Cntand Built Terrace. margin. The most violent agitation of Fig. 6. — Section of a Barrier. the water is along the line of breakers; and the shore di-ift, depending upon agitation for its transportation, follows the line of the breakers instead of the water margin. It is thus built into a continuous outlying ridge at some distance from the water's edge. It will be convenient to speak of this ridge as a harrier The barrier is the functional equiva- lent of the beach. It is the road along which shore drift travels, and it is itself composed of shore drift. Its lakeward face has the typical beach profile, and its crest lies a few feet above the normal level of the Avater. Between the barrier and the land a strip of water is inclosed, consti- tuting a lagoon. This is frequently converted into a marsh by the accumu- lation of silt and vegetable matter, and eventually becomes completely filled, so as to bridge over the interval between land and barrier and convert the latter into a normal beach. The beach and the barrier are absolutely dependent on shore drift for their existence. If the essential contiinious supply of moving detritus is cut off, not only is the structure demolished l)y the waves which formed it, but the work of excavation is carried landward, creating a wave-cut teiTace and a cliff. The principal elements of the theory of shore-drift deposits here set GEOMETRIC RATIO OP EFFECT TO CAUSE, 41 forth are tacitly postulated by many writers on the construction of harbor and coast defenses. According to Cialdi' the potency of currents in con- nection with waves was first announced by Montanari; it has been concisely and, so fixr as appears, independently elucidated by Andrews.^ Still water level is the datum with which all vertical elements of the profile of the beach and barrier are necessarily compared; and, referred to this standard, not only does the maximum height of the beach or barrier vary in thflferent parts of the same shore, liut the profile as a whole stands at different heights. The explanation of these inequalities dej)ends in part on a principle of wide application, which is on the one hand so important and on the other so frecpiently ignored that a paragraph may properly be devoted to it, by way of digression. There are numerous geologic processes in which quantitative variations of a causative factor work immensely greater quantitative varia- tions of the effect. It is somewhat as though the effect was proportioned to an algebraic power of the cause, but the relation is never so simple. Take, for example, the transportation of detritus by a stream. The variable cause is the volume of water; the variable effect is the amount of geologic work done — the quantity of detritus transported. The effect is related to the cause in three different ways: First, increase of water volume augments the velocity of flow, and with increase of velocity the size of the maximmn parti- cle which can be moved increases rapidly. According to Hopkins, the size of the maximum fragment which can be moved varies as the sixth power of the velocity, or (roughly) as the f power of the volume of water. Second, the increase of velocity enlarges the capacity of the water to transport detritus of a given character; that is, the per cent of load to the unit of water is in- creased. Third, increase in the niunber of unit volumes of water increases the load pro rata. The suimnation of these three tendencies gives to the flooded stream a transporting power scarcely to be compared with that of the same stream at its low stage, and it gives to the exceptional flood a ' Loc. cit., p. 394, et seq. Cialdi himself maintains at great length that the work is performed by waves, and that the so-called shore current, a feeble peripheral circulation observed in the Mediter- ranean, is (jualitatively and quantitatively incompetent to jiroduce the observed results. Whether he would deny the efflciency of currents excited by the same winds which produce the waves is not clearly apparent *■ Trans. Chicago. Acad. Sci., vol. a, p. 9. 42 LAKE BONNEVILLE. power greatly in excess of the nomaal or annual flood. Not only is it time that the work accomplished in a few days during' the height of the chief flood of the year is greater than all that is accomplished during the remainder of the year, but it may even be true that the eflect of the maximum flood of the decade ftr generation or century surpasses the combined efl"ects of all mmor floods. It follows that the dimensions of the channel are established by the great flood and adjusted to its needs. In littoral transportation the great storm bears the same relation to the minor storm and to the fair-weather breeze. The waves created by the great storm not only lift more detritus from each unit of the littoral zone, but they act upon a broader zone, and they are competent to move larger masses. The currents which accompany them are correspondingly rapid, and carrv forward the augmented shore di'ift at an accelerated rate. It fol- lows that the habit of the shore, including not only the maximum height of the beach line and the height of its profile, but the dimensions of the wave- cut terrace and of various other wave products presently to be described, is determined by and adjusted to the great storm. It should be said by way of qualification that the low-tide stream and the breeze-lifted wave have a definite though subordinate influence on the topographic configuration. After the great flood has passed by, the shrunken stream works over the finer debris in the bed of the great channel, and by removing at one place and adding at another shapes a small channel adjusted to its volume. After the great storm has passed from the lake and the storm s^vell has subsided, the smaller waves of fair weather construct a miniature beach profile adapted to their size, superposing it on the greater profile. This is done by excavating shore drift along a narrow zone under water and throwing it up in a narrow ridge above the still water level. Thus, as early perceived by De la Beche^ and, Beaumont,^ it is only for a short time innne- diately after the passage of the great storm that the beach profile is a simple curve; it comes afterward to be inteiTupted by a series of superposed ridges produced by storms of difl"erent magnitude. Reverting now to the special conditions controlling tlu; profiles of beach or barrier at an individual locality, it is evident that the chief of these is the ' Mauu.al of Geology, Pbiladelpliia, 1832, p. 72. « Lefoiis, p. 22(5 ami plato IV. THE FETCH OF WAVES, 43 magnitude of the largest waves breaking there. The size of the waves at each locahty depends on the force of the wind and on its direction. A wind bkiwing from the shore lakeward produces no waves on that shore. One from the opposite shore produces waves whose height is approximately pro- portional to the square root of the distance through which they are propa- gated, provided there are no shoals to check their ^augmentation. For a given force of wind, the greatest waves are produced when the direction is such as to command the broadest sweep of water before their incidence at the particular spot, or in the technical phrase, when the fetch is greatest. A second factor is found in the configuration of the bottom. Where the off-shore depth is great the imdertow rapidly returns the water driven forward by the wind, and there is little accumulation against the shore; but where the off-shore depth is small the wind piles the water against the shore, and produces all shore features at a relatively high level. The Subaqueous Ridge.-Various writers liave mentioned low ridges of sand or gravel running parallel to the shore and entirely submerged. As the origin of such ridges is not understood, they have no fixed position in the pres- ent classification, and they are placed next to the barrier only because of similarity of form. The following description was published by Desor in 1 851 : An example of tbis character occurs on the northern shore of Lake Michigan, not far from the fish station of Bark Point (Pointe aux J5corces), under the lee of a prom- ontory, designated on the map as Point Patterson. Here, the shore, after running due east and west for some distance, bends abruptly to the northeast. The voyageur com- ing from the west, after having passed Point Patterson, is struck by the appearance of several bands of shallow water, indicated by a yellowish tint. These bands, which appear to start from the extremity of the point, are caused by subaqueous ridges, which spread, fan-like, to the distance of nearly half a mile to the east, being from three to ten yards wide, and from five to ten feet above the general bed of the lake, at this point. They are not composed, like the flats, of fine sand, but of white limestone pebbles, derived from the adjacent ledges, with an admixture of granitic pebbles, some of which are a foot in diameter. It is difficult to conceive of currents suificiently Ijowerful to transport and arrange such heavy materials, and yet we know of no other means by which this aggregation could have been accomplished. These subaqueous ridges afford, on a small scale, an interesting illustration of the formation of similar ridges now above water. If the north coast of Lake Michigan were to be raised only twenty feet, such a rise would lay dry a wide belt of almost level ground, on which these ridges would appear conspicuously, not unlike those which occur on the south shores of lakes Erie and Ontario, and thus confirm the views of Mr, Whittlesey, that most of these ridges are not ancient beaches, but have been formed under water, by the action of currents,' ' Foster aud Whitney's " Geology of f he Lake Superior Land District." Part 2, p. 258. 44 LAKE BONNEVILLE. Wliittlesey describes no examples on existing coasts, but refers to them as familiar features and relegates to their category numerous inland ridges associated with earlier water surfaces in the basins of Lakes Erie, (Jntario, and Michigan. He says that "their composition is luiiversally coarse water- washed sand and fine gravel", while beaches consist of "clean beach sand and shingle"; and alsoJ:hat beaches are distinguished from subaqueous ridges by the fiict "that the foiTner are narrow and are steepest on the lake side, resembling miniature terraces."' Having personally observed many of the inland ridges described by Wliittlesey and recognized them as barriers, having failed or neglected to observe ridges of this subaqueous type in the Bonneville Basin, and having independent reason to believe that the waters of Lakes Michigan, Erie, and Ontario have recently advanced on their coasts, I leaped to the conclusion that the ridges seen by Desor beneath the water of Lake Michigan, as well as the subaqueous ridges mentioned without enumeration by Whittlesey, were formed as barriers or spits at the water surface and were subsequently submerged by a rise of the water.^ In so doing- I ignored an im[)ortant observation by Andrews, who, writing of the beach at the head of Lake Michigan, describes "a peculiarity in the contour of the deposit, which is uniform in all the sand shores of this part of the coast. As you go out into the lake, the bottom gradually descends from the water line to the depth of about five feet, when it rises again as you recede from the shore, and then descends toward deep water, forming a siibaqueous ridge or 'bar' jiarallel to the beach and some ten or twenty rods from the shore." ^ It is impossible to regard this sand ridge as a beach or barrier sulimerged by the rise of the lake, for it stands within the zone of action of storm waves, and no mole of loose debris can be assumed to successfully oppose their attack. It is to be viewed rather as a product of wave action, or of wave and cuiTent action, under existing relations of land and lake. The subject is advanced by Russell, who visited the eastern shore of Lake Michigan in 1884. He says: Bars of anotber character are also formed along lake margins, at some distance from the land, which agree in many ways with true barrier bars, but differ in being ' Fresh-water Glaci.il Drift of the Northwestern States. By Charles Whittlesey. Sniithsoniao ContribntioM No. 197. W-aHhington, 1S66, pp. 17, lit. 'Fifth Ann. Rcpt. U. S. Gool. Survey, p. 111. 'Traus. Chicago Acad. Sci., vol. ", p. 14. SUBAQUEOUS RIDGES. 45 composed of homogeneous, fine material, usually saud, and in not reaching the lake surface. The character of structures of this uature may be studied about the shores of Lake Michigan, where they can be traced continuously for hundreds of miles. There arc usually two, but occasionally three, distinct sand ridges; the first being about 200 feet from the land, the second 75 or 100 feet beyond the first, and the third, when present, about as far from the second as the second is from the first. Soundings on these ridges show that the first has about 8 feet of water over it, and the second usually about I'i; between, the depth is from 10 to 14 feet. From many commanding points, as the summit of Sleeping Bear Bluff, for example, these submerged ridges may be traced distinctly for many miles. They follow all the main curves of the shore, with- out changing their character or having their continuity broken. They occur in bays as well as about the bases of promontories, and are always composed of clean, homo- geneous sand, although the adjacent beach may be composed of gravel and boulders. They are not shore ridges submerged by a rise of the lake, for the reason that they are in harmony with existing conditions, and are not being eroded or becoming cov- ered with lacustral sediments. In bars of this character the fine debris arising from the comminution of shore drift appears to be accumulated in ridges along the line where the undertow loses its force; the distance of these lines from the land being determined by the force of the storms that carried the waters shoreward. This is only a suggested explanation, however, as the complete history of these structures has not been determined.' In the survey of these lakes by the U. S. Engineers, numerous inshore soundings were made, and while these do not fall near enough together to determine the configuration of subaqueous ridges, they serve to show whetlier the profile of the bottom descends continuously from the beach lakeward. A study of the original manuscrij)t sheets, which give fuller data than the published charts, discovers that bars sunilar to those described by Russell occur along the eastern coast of Lake Michigan wherever the bottom is sandy, being most frequently detectible at a depth of 13 feet, but ranging upward to 3 feet and downward to 18 feet. At the south end of the lake they are not restricted to the 5-foot zone indicated by Andi-ews, but range to 13 feet. A single locality of occurrence was found on the shore of Lake Erie, but none on Lake Ontario. These ridges constitute an exception to the. beach profile, and show that the theory of that profile given above is incomplete. Under conditions not yet apparent, and in a manner equally obscure, there is a rhytlunic action along a certain zone of the bottom. That zone lies lower than the troudi between the greatest storm waves, but the water upon it is violently oscil- 'Geol. Hist, of Lake Lahontan. pp. 92-93. 46 LAKE BONNEVILLE. lated by the jjassing waves. The same water is translated hikeward })y tlie undertow, and the surface water above it is transhxted kindward by the wind, while both move with the shore current parallel to the beach. The rhythm may be assumed to arise from the interaction of the oscillation, the land- ward current, and the undertow. LITTORAL DEPOSITION. The material deposited by shore processes is, first, shore tb-ift; second, stream drift, or the detritus delivered at the shore by tributary streams Increasing depth of water is in each case the condition of littoral deposi- tion. The structures produced by the deposit of shore drift, although some- what varied, have certain conmion features. They will be treated under the generic title of embankments. The sti-uctures produced by the dejjosit of stream drift are deltas. EMBANKMENTS. The current occupying the zone of the shore drift and acting as the coagent of littoral transportation has been described as slow, but it is insepa- rably connected with a movement that is relatively rapid. This latter, which may be called the off-shore current, occupies deeper water and is less impeded by friction. It may in some sense be said to drag the littoral current along with it. The momentum of the off-shore current does not permit it to fol- low the sinuosities of the water margin, and it sweeps from point to point, carrying the littoral current with it. There is even a tendency to generate eddies or return currents in embayments of the coast. The off-shore cur- rent is moreover controlled in part by the configuration of the bottom and by the necessity of a return current. The littoral current, being controlled in large part by the movements of the off-shore current, separates from the water margin in three ways: first, it continues its direction unchanged at points where the shore-line turns landward, as at the entrances of Ijays; sec- ond, it sometimes turns from the land as a surface current; third, it some- times descends and leaves the water margin as a bottom cun-ent. In each of these three cases deposition of shore di-ift takes place by reason of the divorce of shore cuirents and wave action. The depth to THE GENESIS OF SPITS. 47 which wave agitation sufficient for the transportation of shore di-ift extends is small, and when the littoral current by leaving the shore passes into deeper waters the shore di-ift, unable to follow, is thrown down. When the current holds its direction and the shore-line diverges, the embankment takes the form of a S2nt, a Jiook, a hat; or a loop. When the shore-line holds its course and the current diverges, whether superficially or by descent, the embankment usually takes the form of a terrace. The spit.-When a coast line followed by a littoral current turns abruptly landward, as at the entrance of a bay, the current does not turn with it, but holds its course and passes from shallow to deeper water. The water be- tween the diverging current and coast is relatively still, although there is communicated to the portion adjacent to the current a slow motion in the same direction. The waves are propagated indifferently through the flow- ing and the standing water, and reach the coast at all points. The shore drift can not follow the deflected coast line, because the waves that beat against it av^ unaccompanied by a littoral current. It can not follow the littoral current into dee^) water, because at the bottom of the deep water there is not sufficient agitation to luove it. It therefore stops. But the supply of shore di-ift brought to tliis point by the littoral current does not cease, and the necessary result is accumulation. The particles are carried forward to the edge of the deep water and there let fall. In this way an embankment is constructed, and so far as it is built it serves as a road for the transportation of more shore di'ift. The direction in which it is built is that of the littoral current. It takes the form of a ridge following the boundary between the current and the still water. Its initial height brings it just near enough to the surface of the water to enable the wave agitation to move the particles of which it is constructed; and it is naiTOw. But these characters are not long maintained. The causes which lead to the consti'uction of the beach and the barrier are here equally efficient, and cause the embankment to grow in breadth and in height until the cross-profile of its upper surface is identical with that of the beach. The history of its growth is readily deduced from the configuration of its terminus, for the process of growth is there in progress. If the material is coarse the distal portion is very slightly submerged, and is terminated in 48 LAKE BONNEVILLE. the direction of o^rowtli by a steep slope, the suliaqueous "earth-sh)pe" of the particular material. If the material is fine the distal ])ortioii is more deeply submerg-ed, and is not so abru])tly tenuinated. The portion above water is usually narrow throughout, and terminates without reaching the extrem- ity of the embankment. It is flanked on the lakeward side by a submerged plateau, at the outer edge of which the descent is somewhat steep. The profile of the plateau is that normal to the beach, and its contours are con- fluent with those of the beach or barrier on the main shore. Toward the end of the embankment its width diminishes, its outer and limiting contour turning toward the crest line of the spit and finally joining it at the sub- merged extremity. The process of construction is similar to that of a railroad embankment the material for which is derived from an adjacent cutting, carted forward along the crest of the embanlcment and dumped off at the end; and the sym- metry of form is often more perfect than the railway engineer ever accom- plishes. The resemblance to railway structures is very striking in the case of the shores of extinct lakes. As the embankment is carried forward and completed, contact between the current and the inshore water is at first obstructed and finally cut off", so that there is practically no communication of movement from one to the other at the extremity of the spit. At the point of construction the mo^^ng and the standing water are sharply differentiated, and there is hence no uncertainty as to the direction of construction. The spit not only follows the line between the current and still water, but aids in giving definition to that line, and eventually walls in the current by contours adjusted to its natui'al flow. The Bar._If flic curreut determining the foraiation of a spit again touches the shore, the construction of the embankment is continued imtil it spans the entire interval. So long as one end remains free the vernacular of the coast calls it a spit; but when it is completed it becomes a lar. Figure 7 gives an ideal cross-section of a completed embankment. The bar has all the characters of the spit except those of the tenninal end. Its cross-profile shows a })lateau bounded on either hand by a steep slope. The surface of the plateau is not level, but has the beach profile, is BAKS AT THE MOUTHS OF KIVEES. 49 slightly submerged on the windward side and rises somewhat above the ordinary water level at the leeward margin. At each end it is continuous with a beach or bairier. It receives shore drift at one end and delivers it at the other. The bar may connect an island with the shore or with another island, or it Fig. 7.-Section of a Lluear Kmbanknaut. may connect two portions of the same shore. In the last case it crosses the mouth either of a bay or of a river. If maintained entire across the entrance to a bay it converts the water be- tween it and the shore into a lagoon. At the mouth of a river its mainte- nance is antagonized by the outflowing current, and if its integrity is estab- lished at all it is only on rare occasions and for a short time. That is to say, its full height is not maintained; there is no continuous exposed ridge. The shore di'ift is, however, thrown into the river cuiTent, and unless that current is sufficient to sweep it into deep water a submerged bar is tlu'own across it, and maintains itself as a partial obstruction to the flow. The site of this submerged bar is usually also the point at which the current of the stream, meeting the standing water of the lake, loses its velocity and depos- its the coarser paii; of its load of detritus. If the contribution of river drift greatly exceeds that of shore di'ift, a delta is fomied at the river mouth, and this, by changing the configuration of the coast, modifies the littoral current and usually detennines the shore drift to some other course. If the contri- bution of river drift is comparatively small it becomes a simple addition to the shore drift, and does not interrupt the continuity of its transportation. The bars at the mouths of small streams are constituted chiefly of shore drift, and all their characters are determined by their origin. The bars at the mouths of large streams are constituted chiefly of stream di-ift, and belong to the phenomena of deltas. On a preceding page the fact was noted that the horizontal contoiu's of a beach are more regular than those of the original surface against which it rests, small depressions being filled. It is now evident that the process of filling these is identical with that of bar construction. There is no trenchant line of demarkation between the beach and the bar. Each is a carrier of MON I 4 50 LAKE BONNEVILLE. shore drift, and each employs its first load in the construction of a suitable road. Plate IV represents a part of the east shore of Lake ]\Iichigan seen from the hill back of Empire liluflfs. In the extreme distance at the left stand the Sleeping Bear Blutl's, and somewhat nearer on the shore is a tim- bered hill, the lakeward face of which is likewise a sea-cliff. A bar coimects the latter with the land in the foreground and divides the lagoon at the right from the lake at the left. The symmetry of the bar is marred l)y tlie for- mation of dunes, the li<^ter portion of the shore-drift being taken ui) by the wind and carried toward the right so as to initiate the filling of the lagoon. Figure 8 is copied from the U . S. Engineer map of a portion of the south shore of Lake Ontario west of the mouth of the Genesee River. The orig- FiG. 8. — Map of Braddook'B Bay and vicinity, N. T., showing headlands conneoted by Bars. inal contour of the shore was there irregular, consisting of a series of salient and reentrant angles. The waves have truncated some of the salients and have united them all by a continuous bar, behind which several bays or / BARS ACROSS BATS. 51 ponds are inclosed. The movement of the shore drift is in this case from northwest to southeast, and the principal source of the material is a point of land at the extreme west, where a low cliff shows that the land is being eaten by the Avaves. The map in Figure d is also copied from one of the sheets published by the U. 8. Engineers, and represents the bars at the head of 'Lake Supe- rior. These illustrate several elements of the preceding dis- cussion. In the first place they are not formed by the predomi- nant winds, bufby those which brinff the greatest waves. The predominant winds are west- erly, and produce no waves on tin scoast. The shore cWft is de- rived from the south coast, and its motion is first westerly and then northerly. Two bars are exhibited, the western of which is now protected from the lake waves, and must have been com- pleted before the eastern was begun. The place of deposition of shore drift was probably shifted from the western to the eastern by reason of the shoaling of the head of the lake. The converging shores should theoretically produce during easterly storms a powerful undertow, by which a large share of the shore drift A\'ould be carried lakeward and distributed over the bottom. The manner in which the bars terminate against the northern shore without inflection is explica- ble lilvewise by the theory of a strong undertow. If the return current * were superficial the bars would be curved at then- junctions with both shores. An instructive view of an ancient bar will be found in PL IX, repre- senting a portion of the Bonneville shore line. The town of Stockton, Utah, appears at the right. The plain at the left was the bed of the lake- The -■--■ -r- Fig. 9.— Map of the head of Lake Superior, eLowins Baj Bars. 52 LAKE BONNEVILLE. storm waves, moving from left to riglit, carved the sea-cliflP which appears at the base of the mountain at the k^ft, and di'ifting the material toward the right built it into a great spit and a greater bar. The end of the s\nt is close to the town. The bar, which lies slightly lower, having been fonned by the lake at a lower stage of its water, sweeps in a broad curve across the valley to the rocky hill on the opposite side, where the artist stood in making the sketch. The Hook.-Tlie line of direction followed by the spit is usually straight, or has a slight concavity toward the lake. This form is a function of the lit- toral current, to which it owes origin. But that current is not perpetual; it exists only during the continuance of certain determining winds. Other winds, though feebler or accompanied by smaller waves, nevertheless have systems of currents, and these latter currents sometimes modify the form of the spit. Winds which .simply reverse the du-ection of the littoral current retard the construction of the embankment without otherwise affecting it; but a cuirent is sometimes made to flow past the end of the spit in a direction making a high angle with its axis, and such a current modifies its foim. It cuts away a portion of the extremity and rebuilds the material in a smaller spit joining the main one at an angle. If this smaller spit extends lake ward it is demolished by the next stonn; but if it extends landward its position is sheltered, and it remains a permanent feature. It not infi-equently happens that such accessory si)its are formed at intervals during the construction of a long embankment, and are preserved as a series of short branches on the lee side. It may occur also that a spit at a certain stage of its growth becomes especially subject to some conflicting current, so that its noimal gi-owth ceases, and all the shore drift transported along it goes to the construction of the branch. The bent embankment thus produced is called a hook. The currents efficient in the formation of a hook do not cooi)erate simultaneously, but exercise their functions in alternation. The one, during the prevalence of certain winds, brings the shore drift to the angle and accumulates it there; the other, during the prevalence of other winds, de- molishes the new structure and redeposits the material upon the other limb of the hook. HOOKS. 53 In case the land on which it is based is a slender peninsula or a small island, past which the currents incited by various winds sweep with little modification of direction by the local configuration, the hook no longer has the sharp angle due to the action of two currents only, but receives a curved form. Hooks are of comparatively rare occurrence on lake shores, but abound at the mouths of marine estuaries, where littoral and tidal currents conflict. Plate V represents a recurved spit on the shore of Lake Michigan, seen from a neighboring bluff. The general direction of its construction is from left to right, but storms from the right have from time to time tiu-ned its end toward the land and the successive recurvements are clearly discernible near the apex. The mole enclosing Toronto harbor on the shore of Lake Ontario is a hook of unusual complexity, and the fact that its growth threatens to close the entrance to the harbor has led to its thorough study by engineers. Especially has its history been developed by Fleming in a classic essay to Avhich reference has already been made. A hill of drift projects as a cape from the north shore of the lake. The greatest waves reaching it, those having the greatest fetch, are from the east (see Fig. 10), and the cooper- ating current flows from east to west. As the hill gradually yields to the waves, its coarser material trails westward, building a spit. The waves and currents set in mo- ti( >n by southwesterly winds carry the spit end northward, producing a hook. In the ric lo. -Diagram of Lake Ontario, to sbow tho 1 -1 J 1 1 ,1 Futch of Waves reaching Torouto fiom (liH'erent past the westward movement has been the directions. more powerful and the spit has continued to grow in that direction, its north- ern edge being fringed Avith the sand ridges due to successive recurvements, but the shape of the bottom has introduced a change of conditions. The water at the west end of the spit is now deep, and the extension of the embank- ment is correspondingly slow. The northward drift, being no longer sub- ject to frequent shifting of position, has cumulative effect on the terminal hook and gives it a greater length than the others. In the chart of the har- bor (Fig. 11) the composite character of the mole is readily traced. It may 54 LAKE BO]S NEVILLE. also be seen that the ends of the successive hooks are connected by a beach, the work of waves generated within the harbor by northerly winds.^ It will be observed furthermore that while the west end of the spit is continuously fringed by recurved ridges its eastern part is (juite free from them. This does not indicate that the spit was simple and unhooked in the early stages of growth, but that its initial ridge has disappeared. As the cliflf is eroded. Fig. 11. — Map of the harbor and peninsula (Ilook) at Toronto. From charts published by U. T. Hind, in 1854.* its position constantly shifts landward, the shore current follows, and the lakeward face of the spit is carried away so that the waves break over it, and then a new crest is built by the waves just back of the line of the old one.^ By this process of partial destruction and renewal the spit retreats, keeping pace with the retreating clilf. At an earlier stage of the process the spit may have had the position and form indicated by the dotted out- line, but whatever hooks ft-iuged its inner margin have disappeared in the process of retreat. 'The marsh occnpying part of the space between the spit and the inaiuland (Fig. It) is only incidentally connected with the feature under discussion. A small stream, the Don, reaches the shore of the lake within the tract protected from waves by the hook and is thus enabled to construct a delta with its sediment. -Report on the preservation and improvement of Toronto Harbor. In Supplement to Canadian Journal, 1854. 'At the present time the spit is divided near the niiudle, a natural breach having been artificially prevented from healing. The portion of the peninsula fringed by successive hooks stands as au island. LOOPED BARS. 55 The landward shifting illustrated by the Toronto hook affects many embaidvments, but- not all. It ordinarily occurs when the embankment is built in deep water and the source of its material is close at hand. Wherever it is known that an embankment has at some time been breached by the waves, it may be assumed with confidence that retreat is in progress. As retreat progresses the layers constituting- the embankment are trun- cated at top, and new layers are added on the landward side. In the result- ing structure the prevailing di}) is landward (Fig. 12), and it is thereby distinguished from all other forms of lacustrine deposition. This structure was first described and explained by Fleming, who observed it in a railway cutting through an ancient spit.^ The Loop.- Just as the spit, by advancing until it rejoins the shore, becomes a, bar, so the completed hook may with propriety be called a looj) or a looped bar. There is, however, a somewhat different feature to which the name is more strikingly applicable. A small island standing near the main-land is usually furnished on each side with a spit streaming toward the land. These spits are composed of detritus eroded from the lakeward face of the island, against which beat the waves generated through the l)road expanse. The currents accompanj-ing the waves are not unifoi-m in direction, but vary witli the wind tlu'ough a wide angle; and the spits, in sympathy with the varying direction of currents, are curved inward toward the island. If their extremities coalesce, they constitute together a perfect loop, resembling, when mapped, a festoon pendent from the sides of the island. Such a loop in the fossil condition, that is, when preserved as a vestige of the shore of an extinct lake, has the form of a crater rim, the basin of the original lagoon remaining as an undfained hollow. The accompanying illustration (PI. VI) represents an island of Lake Bonneville standing on the -desert near what is known as the "Old River Bed." The nucleus of solid rock was in this instance nearly demolished before the work of the waves was arrested by the lowering of the water. The Wave-built Terrace.-It has already bccu pointed out that when a separa- tion of the littoral current from the coast line is lirought about bv a diverg- ence of the current rather than of the coast line, there are two cases, in the 'Notes on the Daveuport gravel diift. Canaili.ui Joarnal, New Series, vol. 6, 1861, pp. 247-253. 56 LAKE BONNEVILLE. first of which the current continues at the surface, while in the second it dives beneath the surface. It is now necessary to make a further distinc- tion. The cun-ent departing from the sliore, but remaining at tlie surface, may continue with its original velocity or it may assume a greater cross- section and a diminished velocity. In the first case the shore drift is built into a spit or other linear embankment. In the second case it is built into a terrace. The quantity of shore diift moved depends on the magnitude of the waves; but the speed of transit depends on the velocity of the current, and wherever that velocity diminishes, the accession of shore di-ift must exceed the transmission, causing accumulation to take place. This accumu- lation occurs, not at the end of the beach, but on its face, carrpng its entire profile lakeward and producing by the expansion of its crest a tract of new- made land. If afterward the water disappears, as in the case of an extinct lake, the new-made land has the character of a terrace. A cun-ent which leaves the shore by descending, practically produces at the shore a diminu- tion of flow, and the resulting embankment is nearly identical with that of a slackening superficial current. The wave-built terrace is distinct from the wave-cut terrace in that it is a work of construction, being composed entirely of shore drift, while the wave-cut terrace is the result of excavation, and consists of the pre-existent terrane of the locality. The wave-built terrace is an advancing embank- ment, and its internal structure is characterized by a lakeward dip (Fig. 13). It is thus contrasted with the retreating embaidiment (Fig. 12). Fig. 12.— Section of a Linear Embaukmcnt retreating landward. Tliedolti-d line .sliiiws llie oiiyiu^il posili(in of tbe crest i^^^^^i^^iiiiil^^^sl^Mlii^MSsJMi^^Si^^;. Fig. 13.— Section of a Wave-built Terrace. The surface of the wave-built terrace, considered as a whole, is level, but in detail it is uneven, consisting of parallel ridges, usually curved. Each WAVE-BUILT TERRACES. 57 of these is referable to some exceptional siorm, the waves of which threw the shore ch-ift to an unusual height. Wliere the shore drift consists wholly or in large part of sand, and the prevailing winds are toward the shore, the wave-built terrace gives origin to dunes, which are apt to mask its normal ribbed structure. The locality most favorable for the formation of a wave-built terrace is the head of a triangular bay, up which the waves from a large body of water are rolled without obstruction. The wind sweeping up such a bay carries the surface of the water before it, and the only return current is an undertow originating near the head of the bay. The superficial advance of the water constitutes on each shore a littoral current conveying shore drift toward the head of the bay, and as these littoral currents are diminished and finally entirely dissipated by absorption in the undertow, the shore di'ift taken up along the sides of the bay is deposited. If the head of the bay is acute, the first embankment built is a curved bar tangent to the sides and con- cave toward the open water. To the face of this successive additions are made, and a terrace is gradually produced, the component ridges of which are approximately parallel. The sharpest curvature is usually at the ex- treme head of the bay. The converging currents of such a bay give rise to an undertow which is of exceptional velocity, so that it transports with it not only the finest detritus but also coarser mattei', such as elsewhere is usvially retained in the zone of wave action. In effect there is a resorting of the material. The shore drift that has traveled along the sides of the bay toward its head, is divided into two portions, the finer of which passes out with the reinforced undertow, while the coarser only is built into the terrace. The v-Terrace and v-Bar.-It rcmalus to dcscribc a type of terrace for which no satisfactory explanation has been reached. The shores of the ancient Pleis- tocene lakes afford numerous examples, Ijut those of recent lakes are nearly devoid of them, and the writer has never had opportunity to examine one in process of formation. They are triangular in ground plan, and would claim the title of delta were it not appropriated, for they simulate the Greek letter more strikingly than do the river-mouth structures. They are built against coasts of even outline, and usually, but not always, upon slight 58 LAKE BONNEVILLE. salients, and they occur most freqiientl}' in the long, narrow arms of old lakes. \ One side of the triangle rests against the land and the opposite angle points toward the open water. The free sides meet the land with short curves of adjustment, and appear otherwise to he normally straight, although they exhibit convex, concave, and sigmoid flexures. The growth is by ad- ditions to one or both of the free sides; and the nucleus appears always to have been a miniature triangular terrace, closely resembling the final struct- ure in shape. In the Bonneville examples the lake ward slope of the teiTace is usually very steej) down to the line where it joins the preexistent slope of the bottom. There seems no reason to doubt that these embankments, like the others, were built by currents and waves, and such being the case the for- mative currents must have divercced from the shore at one or both the land- ward angles of the terrace, but the condition detennining this divergence does not appear. In some cases the two margins appear to have been deteimined by cur- rents ajiproaching the terrace (doubtless at different times) from oj^posite directions; and then the terrace margins are concave outward, and their confluence is prolonged in a more or less irregular point. In most cases, however, the shore drift appears to have been carried by one cm'rent from the mainland along one margin of the teiTace to the apex, and by another current along- the remaining side of the terrace back to the mainland. The contours are then either straight or convex. In Lake Bonnevnlle it happened that after the best defined of these ter- races had attained nearly their final width the lake increased in size, so that- they Avere immersed beneath a few feet f)f water. Wliile the lake stood at the higher level, additions were made to the terraces by the building of lin- ear embankments at their outer margins. These were carried to the water surface, and a triangular lagoon was imprisoned at eacli l(>ialit\-. The sites of these lagoons are now represented by flat triangular basins, i-ach walled in by a bar bent in the fonn of a V. These Ijars were at first observed without a clear conception of the terrace on which they were founded, and the name W-bar was applied. The V-bar, while a conspicuous feature of a S. GEOLOGICAL SUPVEY LAKE BONNEVILLE PLVH I'LATS Ol' LOOl'Kl) AM) V-SllArKT) KM15AN K'MKNTS, OBSERVED ON TBK SHORES OF LAKE I{0.\NEVlEEi:. o I 8 3 ■tOOO_ _ SCALt: t arrows sfitiw on 171 w/t/r/f x liri/'f/d 1, Siill Mnixli , .■>■„, lit' I'.illii •-', II. A'„.v/ B.i.-.; ,ir IJra\,r Crrrl. Uaii,/, h'<.s;-ni'ir lUillr ,01,1 Hivrr Ui-il . 1.. SlIiWIllKW ■'>. Fri'tt o/' the Mouriliiui 6. Haxl Biixf , Drrp I'r JtU: 7 , U sill,- nl did Hirer Bed \\,S\X\ U'f.y/ NtiAf o/' Fn.-irn Mnnritfiin,. 10,12 Prmxs I'liUev. rinir Wa-iia .s/nuu/ \'i , .Vf rt r StttrkLcn .luliu.t Klcn \ L'o.liUi DiMwn t>v C; Tlioiuluth at the extremity nf the new platform backward to some steeper part of its channel — a continuous grade sufficient to give it a velocity adequate to its load. Tlie postulate is, of course, ideal. The river does not in fact build a level bed and afterward change it to a slope, but carries forward the whole work at once, maintaining continuously an adjustment between its grade and its work. Moreover, since the deposition begins at some distance from the mouth, the lessening load does not require a uniform grade and does not produce it. The grade diminishes gradually lakeward to the foot of the deposit slope, so that the longitudinal profile is slightly concave upward. At the head of the deposit slope there is often an abrupt change of grade. At its foot, where the maximum deposit is made, there is an abrupt change of a double character; the incline of the river surface is exchanged for the horizontal plane of the lake surface; the incline of the river bottom is ex- changed for the steeper incline of the delta front. The river current is swifter in the middle than at the sides, and on a deposit slope, where velocity is nicely adjusted to load, the slight retarda- tion at the sides leads to deposition of suspended matter. A bank is thus produced at either hand, so that the water flows down an elevated sluice of its own construction. The sides are built up pari passu with the bottom, but inasmuch as they can be increased only by overflow, they never quite reach the flood level of the water surface. A river thus contained, and a river channel thus constructed, constitute an unstable combination. So loner as the bank approximates closely to the level of the surface at flood stage, the current across the bank is slovyer than the current of the stream, and deposits silt instead of excavating; but whenever an accidental cause so far lowers the bank at some point that the current across it during flood no longer makes a deposit, there begins an erosion of the bank which increases rapidly as the volume of escaping water is augmented. A side channel is thus produced, which eventually becomes deeper than the 68 LAKE BONNEVlLLli;. main or oriq'inal cliannel and draws in the greater part or perlia]is all of the water. The ability of" the new channel to drain the old one depends on two things: first, the outer slope of the bank, from the circumstances of its con- struction, is steeper than the descent of the bottom of the channel; second, the first-made channel, although originally following the shortest route to the lake, has so far increased its length by the extension of its mouth that the water escaping over its bank may find a shorter route. The river channel is thus shifted, and its mouth is transferred to a new point on the lake shore. Repetition of this process transfers the work of alluAnal deposition from place to place, and causes the river to build a sloping plain instead of a simple dike. The lower edge of the plain is everywhere equidistant from the head of the deposit slope, and has therefore the fonu of a circular arc. The inclination is in all directions the same, varying only with the dimin- ishing grade of the deposit slope, and the fomi of the plain is thus approxi- mately conic. It is, in fact, identical with the product of land-shaping known as the alluvial cone or alluvial fan. The symmetry of the ideal form is never attained in fact, because the process of shifting implies inequality of surface, but the approximation is close in cases where the grade of the deposit slope is high, or where the area of the delta is large as compared with the size of the channel. m ''&^////«ilyX^^Mi, . Fin. 14.— Section of a Delta. At the lake shore the manner of deposition is ditVcrcnt. The heavier and coarser part of the river's detrital load, that which it j)ushes and rolls along the bottom instead of earring by suspension, is emptied into the lake and slides down the face of the delta with no impulse but that given by its own weight. The slojje of the delta face is the angle of repose of this coarse material, subject to such modification as may result from agitation by waves. DELTA STRUCTURE. 69 The finer part of the detritus, that which is transported by sns]:)ension, is carried beyond the delta face, and sinks more or less slowly to the bottom. Its disti'ibution depends on its relative fineness, the extremely fine material being widely diffused, and the coarser falling near the foot of the delta face. The depth of the deposit formed from suspended material is greatest near the delta and diminishes gradually outward, so that tlie sloj)e of the delta face merges by a curve with the slope of the bottom beyond. As the delta is built lakeward, the steeply inclined layers of the delta face are superjiosed over the more level strata of the lake bottom, and in turn come to support the gently inclined layers of the delta plain, so that any vertical section of a normal delta exhibits at the top a zone of coarse material, bedded with a gentle lakeward inclination, then a 7.one of similar coarse material, the laminations of which incline at a high angle, and at bottom a zone of fine material, the laminations of Avhich are gently inclined and unite by curves with those of the middle zone. The characters of the fossil delta, or the delta as it exists after the des- iccation of the lake concerned in its formation, are as follows: The upper surface is a terrace with the form of an alluvial fan. The lower slo})e or face is steep, ranging from 10° to 25°; it joins the upper slope by an angle and the plain below by a curve. The line separating the upper surface from the outer slope or face is horizontal, and, in common with all other horizon- tal contours of the structure, is approximately a circular arc. The upper or landward limit of the upper surface is a line horizontally uneven, depend- ing on the contours of the antecedent topography. The lower limit of the face is a vertically uneven line, depending on the antecedent topography as modified by lake sediments. The material is detrital and well rounded; it exhibits well-marked lines of deposition, rarely taking the character of bed- ding. The structure as seen in section is tripartite (Fig. 15). In the upper division the lines of deposition are parallel to the upper surface of the delta; in the middle division they are parallel to the steep outer face, and in the lower division they are gently inclined. The separation of the middle divis- ion from the lower is obscure. Its separation from the upper is definite and constitutes a horizontal plane. The fossil delta is invariably divided into two parts by a channel running from its apex to some part of its periphery 70 LAKE BONNEVILLE. and occupied l)y a stream, the agent of its construction Ijccoming, under changed conditions of base level, the agent of demolition. The ftm-like outline of the normal delta is iiioditicd wlierever wave :ic- tion lias an importance comparable with that of stream fiction. Among tlie great variety of fonns resulting from the combination of tlic two agencies, there is one wliidi repeats itself with suf- ficient frequency to deserve special men- tion. It occin-s where the force of the )unt tlie delta is inconsiderable. In such case the shore current from either direction is deflected by the mass of the delta, and wave action adjusts the contour of the delta to conformitv with the deflected shore current. If the ANave influences from oj)posite directions are equal, the delta takes the form of a symmetric tri- aiiffle similar to that of tlu' V-terrace. Numerous illustrations are to be seen on the shores of Seneca and CaAiiga Lakes, where the conditions are peculiarly favorable. The lake is long and narrow, so that nil the efficient wave action is associated with strong shore currents, and these alternate in dii'ection. The predominant rock of the sides is a soft slude, so easilv triturated l)v tlie waves tliat the entire product of its erosion escapes with the undertow, and no shore drift remains. The sides are straight, and each tributary stream l)uilds out ii little proinon- torv ;it its month, to wliicli the waves ffive form. Some of these triauii'ular deltas (miiIxmIv perfectlv tlu' Greek letter, Init tliev turn tlu* aj)ex toward tla* wati'i- instead of towanl tlw* laud. Fig. 15. — Verlitul section in :i Delta, .sliowiuj; the i\,p\ cal siiccessiou of strata. THE WALLING IN OF "WALLED" LAKES. 71 ICE WORK ; THE RAMPART. This feature does not belong to lakes in general, but is of locjil and exceptional occurrence. It was named the "Lake Rampart" by Hitchcock, who gave the first satisfactory accoinit of its origin.' Earlier observations, containing the germ of the exjjlanation of the phenomenon, wci-e made by Lee^ and Adams.' A later and indepen- dent explanation was given by White.* Tn ignorance of ITitchcock's description, I yave credit in the Fifth Amuial Keijort of ^ ' Fig. 16.— Section ol a r.nTiipni I. the U. S. Geological Survey to White, and myself proposed the name "Shore Wall." I now substitute Hitclicock's name, "Rampart", being moved thereto not only by the priority and the eminent fitness of the name, bnt by the consideration that "Shore Wall" is liable to be confounded with "Sea Wall", a term applied on some marine coasts to steep-faced endiankments of shingle. The ice on the surface of a lake expands while forming, so as to crowd its edge against the shore. A further lowering of tem})erature produces contraction, and this ordinarily results in the opening of vertical fissures. These admit the water from below, and by the freezing of that water they are filled, so that when expansion follows a subsequent rise of temperature the ice cannot assume its original position. It consequently increases its t( >t;d area and exerts a second thrust upon the shore. Where the shorts is iil)rnpt, the ice itself yields, either bv crushing at the margin or by tlm formation of anticlinals elsewhere; but if the shore is generally shelving, the margin of the ice is forced up the acclivity, and carries with it any 1)owlders or other loose material about which it may have frozen. A second lowering of tem- perature does not withdraw the protruded ice margin, but initiates other cracks and leads to a repetition of the shoreward thrust. The process is repeated from time to time during the winter, but ceases with the melting of 'Lake Ramparts in Vermont. By Clias. H. Hitchcock. In Proe. Am. Ass. Adv. Sci., vol. 1.3, 1860, p. 335. ^C. A. Lee. Am. Jour. Sci., vol. 5, 1822, pp. 34-37, and vol. 9, 1825, pp. 2.39-241. 'J. Adams. Am. Jour. Sci., vol. 9, 1825, p)(. 13(>-144. *C. A. White. Am. Naturalist, vol. 2, IHi/.i, pp. 14G-149. 72 LAKE BONNEVILLE. the ice in the spring. The ice formed the ensuing winter extends only to the water margin, and hy the winter's oscilhitions of temperature can be thrust Landward only to a certain distance, determined by the size of the lake and the local climate. There is thus for each locality a definite limit, beyond whicli the ])rojection of bowlders cannot l)e earned, so tliat all are deposited along a common line, where they constitute a wall oi' ramjinrt. The base of a, rampart stands somewhat above and beyond the ordinary mai'gin of tlie water. It is parallel to the water margin, following its inflec- tions. Its size is ])robably determined in fact by the supply of matenal, but there must also be a limit dependent on the strength of the ice formed in the given locality. Its material is usually coarse, containing bowlders such as the waves generated on the same lake would be unable to move. These iw.vy be either smooth or angular, heavy or light, the process of accunuda- tion involvino- no discrimination. Ramparts are not found on the margins of large lakes, for whatever record the ice of winter may make is obliterated by the storm waves of sum- mer. Neither do they occur on the shores of very deep lakes, for such do not admit of a heavy coating of ice; and for the same reason they are not found in wann climates. So far as the Avrit. is aware, they have never been found in the fossil condition, except that in a single instance a series of them serves to record very recent changes of level. SUBMERGENCK AND EMERGENCE. In tlie preceding discussion the general relation of the water surface to the land has been assumed to be constant. In ])oint of fact it is subject to almost continuous change, and its mutations motlify the products of littoral shaping. Lakes with outlet lower their water surfaces by con-ading the channel of outflow. Lakes without outlet continually oscillate up and down with changes of climate; Jind finally, all large lakes, as well as the ocejin, are aftected by differential movements of the land. The series of displacements which in the geologic past has so many times revolutionized the distribution of laud and water, has not ceased; and earth movements are so nearly uni- versal at the present time that there are few coasts which betray no sjTnntoms THE COASTS OF RISING AND SINKING LANDS. 73 of recent elevation or subsidence. In this place it is unnecessary t(j consider whetlier the relation of water snrftxce to land is affected by mutations of the one or of tlie otlu-i'; and the terms emergence and submergence will be used with the understanding tliat they apply to clianges in the relation without reference to causes of change. Tlie general effect of submergence or emergence is to change the horizon at which shore ])rocesses ai'e carried on; and if a considerable change of level is effected abruptly, the nature of the ])rocesses and the character of their ])roducts are not materially modified. A submerged shore- line retains its configuration until it is gradually buried by sediments. An emerged shore-line is subjected to slow destruction by atmospheric agen- cies. Only the delta is rapidly attacked, and that is merely divided into two parts l)y the stream which formed it. In the case of submergence the new shore constructed at a hi^'her horizon is essentially similar to the one submerged. In the case of emergence the new shore constructed at a lower horizon rests upon the smooth contours wrought by lacustrine sedimenta- tion, and, finding in the configuration little that is incongruous witli its shore currents, carves few cliffs and builds few embankments. The barrier is usually one of its characteristic elements. A slow and gradual submergence modifies the products of littoral action. The erosion of sea-cliffs is exceptionally rapid, because the gradually deep- ening water upon the wave-cut terraces relieves the waves from the task of carving the terraces and enables them to spend their full force against the cliffs. The cliffs are thus beaten back before the advancing tide, and their precipitous character is maintained with constant change of position. A rhythm is introduced in the construction of embankments. For each level of the water surface there is a set of positions appropriate to the initia- tion of embankments, and Avith an advancing tide these positions are suc- cessively nearer and nearer the land ; but with the gradual advance of water the position of embankments is not correspondingly shifted. The embank- ment constructed at a low stage controls the local direction of the shore current, even when its crest is somewhat submerged, and by this control it determines the shore di-ift to follow its original course. It is only when the submergence is sufficiently rapid to produce a considerable depth of water 74 LAKE BONNEVILLE. over the crest of the embankment that a new embankment is initiated behind it. The new embankment in turn controls tlie shore current, and by a rep- etition of the process a series of embankments is produced whose crests differ in height by considerable intervals. A slow and gradual emergence causes the waves, at points of excava- tion, to expend their energies upttn the terraces rather than the cliffs. No great cliffs ui-e produced, but a wave-cut terrace is carried downward with the receding tide. Then; is now no rliytlmi in the construction of embank- ments. At each successive lower level the shore drift takes a course a little farther lakeward, and is built into a lower embankment resting against the outer face of the one just formed. The delta is very sensitive to emergence. As soon as the lake water falls from its edge, the formative stream, having now a lower point of dis- charge, ceases to throw down detritus and begins the corrasion of its chan- nel. It ceases at the same time to shift its course over the surface of the original delta, l)ut retains whatever position it happened to hold when tlie emergence was initiated. Coincidently it begins the constraction of a new or secondary delta, the ajiex of wliicli is at the o^^ter margin of the original structure. With continuous emei'gence a series of new deltas are initiated at points successively farther lakeward, and there is pi-oduced a continuous descending ridge divided by the chaimel of the stream. THE DISCRIMINATION OF SHORE FEATURES, A .shore is the common margin of dry land and a body of water. The elements of its peculiar topography are little liable to confusion so long as they are actually associated with land on one side and water on the other; but after the water has been withdrawn, their recognition is less easy. They consist merely of certain cliffs, terraces, and ridges; and cliffs, terraces, and ridges abound in the topography of land surfaces. In the following pages the topographic features characteristic of ancient shores will be com- pared and contrasted with other topographic elements likely to create con- fusion. Such a discrimination as this lias not before been attempted, although the principal distinctions upon wliich it is based have been the common DISCRIMINATION OF SHORE FEATURKS. 75 property of geologists for many years. The contrast of stream terraces with shore terraces was clearly set forth by Dana in the American Journal of Science in 1849, and has been restated by Geikie in his Text-Book of Ge- ology. It was less clearly enunciated by the elder Hitchcock in his Illus- trations of Surface Geology. CLIFFS. A clitf is a tojwgraphic facet, in itself steep, and at the same time sur- rounded by facets of less inclination. The only variety belonging to the phenomena of shores is that to which the name "sea-cliflf " has been ap})lied. It will be compared with the cliff of differential degradation, the stream cliff, the coulee edge, the fault scarp, and the land-slip cliff. The Cliff of Differential Degradation.-It is a familiar fjxct that cortaiu rocks, maiuly soft, yield more rajjidly to the agents of ei-osion than certain other rocks, mainly hard. It results from this, that in the progressive degradation of a country by subaerial erosion the minor reliefs are generally occupied by hard rocks while the minor depressions mark the positions of soft rocks. Where a hard rock overlies one much softer, the erosion of the latter pro- ceeds so rapidly that the former is sapjjed, and being deprived of its support falls away in Ijlocks, and is thus wrought at its margin into a cliff. In re- gions undergoing rajiid degradation such cliffs are exceedingly abundant. It is the invariable mark of a cliff" oi differential degradation that the rock of the lower part of its face is so constituted as to yield more rapidly to erosion than the rock of the upper part of its face. It is strictly dependent on th(! constitution and structure of the terrane. It may have any form, l)ut since the majority of rocks are stratified in Ijroad, even sheets, and since tlie most abrupt alternations of texture occur in connection with such stratifica- tion, a majority of cliffs of differential degradation exhibit a certain uniformity and parallelisin of parts. The crest of such a cliff is a line parallel to the base, and other associated cliffs run in lines approximately parallel. The most conspicuous of the cliffs of stratified rocks occur where the strata are approximately horizontal; and these more often than any others have been mistaken for sea-cliffs. The Stream Cliff. -The most powerful agent of land erosion is the running stream, and, in regions undergoing rapid degradation, corrasion by streams 76 LAKE BONNEVILLE. so far exceeds the general waste of the surface that their channels are cut down vertically, forming cliffs on either hand. Tliese cliffs are afterward maintained hy lateral corrasion, which opens out the valley of the stream after the establishment of a base level has checked the vertical corrasion. Such cliffs are in a measure independent of the nature of the rock, and are closely associated with the stream. They stand as a rule in pairs facing each other and separated only by the stream and its flood plain. The base of each is a line inclined in the direction of the stream channel and in the same degree. The crest is not parallel thereto, l)ut is an uneven line conforming to no simple law. The Coulee Edge.-The viscosity of a lava stream is so great, and this i-is- cosity is so augmented as its motion is checked l)y gradual cooling, that its margin after congelation is usually marked by a cliff of some height. The distinguishing characters of such a cliff are that the rock is volcanic, ^\•ith the supei-ficial features of a subaerial flow. It has probably never been mis- taken for a sea-cliff, and receives mention here only for the sake of giving generality to the classification of cliffs. The Fault Scarii.-Tlie faultiug of rocks consists in the relative displacement of two masses separated by a fissure. The plane of the fissure is usually more or less vertical, and by virtue of the displacement one mass is made to project someA^'hat above the other. The portion of the fissure wall thus brought to view constitutes a variety of cliff or escarpment, and has been called a fault scarp. In the Great Basin such scai-jis are associated with a great number of mountain ranges, appearing generally at their bases, just where the solid rock of the mountain mass is adjoined by the detrital foot slojie. They occasionally encroach upon the latter, and it is in siu-h case that they are most conspicuous as well as most likel}- to ])e mistaken for sea-cliffs. Although in following the mountain bases tliev do not varv greatly in altitude, yet they never describe exact contours, Ijut ascend and descend the slopes of the foot hills. Tlie crest of such a cliff is usually closely parallel to the base for long distances, but this jtarallclism is not aljsolute. The two lines gradually converge at either end of the displace- ment. In exceptional instances they converge rapidly, giving the cliff a somewhat abrupt termination, and in such case anew clifi'ai)pears en dchelon, COMPARISON OF CLIFFS. 77 continuing the displacement with a slight ofiFset. In Chapter VIII these cliffs are described at length and illustrated by views and diagrams. The Land-Slip Cliff. -Tlic laud-slip dltfcrs from the fault chiefly in the fact that it is a purely superficial phenomenon, having its Avhole history upon a visible external slope. It occurs usually in unconsolidated material, masses of which break loose and move downward short distances. The cliffs pro- duced by their separation from the general or parent mass, are never of great horizontal extent, and have no common element of form except that they are concave outward. They frequently occur in groups, and are apt to con- tain at their bases little basins due to the backward canting which forms part of the motion of the sliding mass. Comparison. -The sca-clift" differs from all others, first, in that its base is horizontal, and, second, in that there is associated with it at one end or other a beach, a barrier, or an embankment. A third valuable -diagnostic feature is its uniform association with the terrace at its base; but in this respect it is not unique, for the cliff of differential degradation often springs from a terrace. Often, too, the latter is nearly horizontal at base, and in such case the readiest comparative test is found in the fact that the sea-cliff is inde- pendent of the texture and structure of the rocks from Avhich it is carved, while the other is closely dependent thereon. The sea-cliff is distinguished from the stream-cliff by the fact that it faces an open valley broad enough and deep enough to permit the genera- tion of efficient waves if occupied b}- a lake. It is distinguished from the coulee edge by its independence of rock structure and by its associated ter- race. It differs from the fault scai-p in all those peculiarities which result from the attitude of its antecedent; the water surface concerned in the for- mation of the sea-cliff is a horizontal plane; the fissure concerned in the formation of the favilt scarp is a less regular but essentially vertical ])lane. The former crosses the inequalities of the preexistent topography as a con- tour, the latter as a traverse line. The land-slip cliff is distinguished by the marked concavity of its face in horizontal contour. The sea-cliff is usually couA-ex, or, if concave, its contours are long and sweeping. The former is distinguished also by its discontinuity. 78 LAKE BONNEVILLE. TERRACES. A terrace is a horizontal or nearly horizontal topographic facet inter- rupting a steeper slope. It is a limited plain, from one edge of which the ground rises more or less steeply, while from tlie opjjosite edge it descends more or less stee})ly. It is the "tread"' of a topographic step. Among the features peculiar to shores are three terraces : the wave-cut, the Avave-built, and the delta. These will be compai-ed with tlie terrace by differential degradation, the stream terrace, the moraine terrace, the fault terrace, and the land-slip terrace. The Terrace by Differential Degradation.-The SaUlO gCUCral cirCUmStaUCCS of Hlck texture- which nnder erosion give rise to cliffs j^roduce also terraces, but the terraces are of less frequent occurrence. The only case in which they are at all abundant, and the only case in which they need be discriminated from littoral terraces, is that in which a system of strata, heterogeneous in texture and lying nearly horizontal, is truncated, either by a fault or by some erosive action, and is afterwards subjected on the truncated section to atraosphei'ic waste. The alternation of hard and soft strata gives rise under such cir- cumstances to a series of alternating cliffs and terraces, the outcrop of each hard stratum appearing in a more or less vertical cliff", and the outcrop of each soft stratum being represented l)y a gently sloping terrace, united to the cliff above by a curve, and, in typical exam})les, separated from the cliff below by an angle. The length of such terraces in the direction of the strike is nsuallv great as compared with their w^idth from cliff' to cliff. They are never level in cross profile, but (1) rise with gradually increasing slo])efrom the crest of one cliff to the base of the next, or (2) descend from the crest of one cliff' to a medial depression, and thence rise with gradually increasing slope to the base of the next. The first case arises where the terrace is narrow or the dip of the strata is toward the lower cliff, the second case where the teiTace is broad and the (lip of the rocks is toward the up})er cliff. In the first case the drainage is outward to the edge of the lower cliff; in the second it is toward the medial depression, whence it escapes by Jie narrow channels carved througli tlic rock of the lower cliff. DEGRADATION TERRACES. 79 The Stream Terrace.-Tlic' coiiditioii of rapid crosiou ill ally region is u})lift. In a tract which has recently been elevated, the rate of" degradation is iine({ual, the waste of the water channels being more rapid than that of the surface in general, so that they are deeply incised. Eventually, however, the corrasion of the water channels so reduces their declivities that the velocities of current suffice merely for the transportation outward of the detritus disengaged by the general waste of surface. In other words, a base level is reached. Then the process of lateral corrasion, always carried on to a certain extent, as- sumes jirominence, and its results are rendered conspicuous. Each stream wears its banks, swinging from side to side in its valley, always cutting at one side, and at the other building a shallow deposit of alluvium, which con- stitutes its flood plain. The valley, having before consisted of the river channel margined on either side by a clift, now consists' of a plain bounded at the sides by cliffs and traversed by the river channel. If now the corrasion of the stream bed is accelerated by a new uplift or other cause, a smaller valley is excavated within the first and at a lower level. So much of the original flood plain as remains constitutes terraces flanking the sides of the new valley. Outwardly one of these terraces is bounded by the base of the old line of cliffs, which may by decay have lost their vertical habit. Inwardly it is bounded by the crest of the new line of cliHs i>roduced by lateral corrasion. Acceleration of downward corrasion is Ijrought about in many ways. As already mentioned, it may be produced by a new uplift, and this stimu- lus is perhaps the most potent of all. It is sometimes produced by the downtlirow of the tract to which the streams discharge, or, \\\\;\t is nearly the same thing, by the degradation of stream channels in that tract. It is also brought about, within a certain range of conditions, by increase of rainfall; and finally, it always ensues sooner or later from the defect of transported material. The general waste of the originally uplifted tract undergoes, after a long period, a diminution in rapidity. The streams have therefore less detritus to transport. Their channels are less clogged, and they are enabled to lower them by corrasion. Perhaps it would be better to say that after the immediate consequences of uplift have so far passed away that an equilibrium of erosive action is established, the degradation of 80 LAKE BONNEVILLE. the entire tract proceeds at a slow continuous rate, the sliglit variations of whic'h are in a sense accidental. Lateral corrasion under such circumstances coexists in all stream channels with downward corrasion, and is the more important process; but the horizon of its action is continuously lowered by the downward corrasion. The terraces which result represent onl}- the stages of a continuous process. In a great number of stream valleys, not one but many ancient flood plains find record in terraces, so that the stream terrace is a familiar topo- graphic feature. When a stream meandering in a flood plain encroaches on a wall of the valley and corrades laterally, it carries its work of excavation down to the level of the bottom of its channel; and afterward, when its course is shifted to some other part of the valley, it leaves a deposit of allu\'ium, the upper surface of which is barely submerged at the flood stage of the stream. The depth of alluvium on the flood plain is therefore measured by the ex- treme depth of the current at high water It constitutes a practically even sheet, resting on the undisturbed terrane beneath. When the stream finally abandons it, and by carving a deeper channel, converts it into a terrace, the terrace is necessarily bipartite. Above, it consists of an even layer of allu- vial material, fine at top and coarse at bottom; below, it consists of the preexistent formation, whatever that may be. Where the lower portion is so constituted as to resist erosion, it loses after a long period its alluvial blanket, and then the terrace consists simply of the floor of hard rock as pared away by the meandering stream. The coarse basal portion of the alluvium is tlie last to disappear; and if it contains Inird bowlders some of these will sui'vive as long as the form of the terrace is recognizable. The elder Hitchcock enumerated and described four types of stream terrace: the lateral terrace, the delta teiTace (groui)ed by the writer with shoi-e terraces), the gorge terrace, and the glacis terrace;' and Miller, whose clear anal}'sis of stream terracing is tlie most recent contribution to the sub- ject,^ adds the ampliitheater terrace, the junction terrace, and the fan ter- race. Such detail is not required in this connection, but it is proper to dis- ' Illustrations of Surface Geology. By Edward Hitchcock, p. .5. 'River-Terracing: its Methods and their results. Bj Hugh Miller. In Proc. Re. Asdescrilx'd in tlie introductory chapter, that district is studded with a -great EARLIER SCULPTURE SUBAEKIAL. 91 number of small mountain ranges, standing in irregular order, but with a nearly constant north-south trend. BetAveen them are narrow valleys Hoored by detritus worn from their summits during the uncounted ages of their existence. At the foot of each range, and piled high against its sides, are great conical heaps of alluvium, each with its apex at a mountain gorge. At top these alluvial cones are separate, but lower down they adjoin, and their bases coalesce into a continuous scolloped slope, the visible footstool of the mountain. The cones, like the valley floors, are composed of detritus eroded from the mountains, but their material is coarser. At the margins of the undrained valleys the cones merge by gentle curvature with the valley floors. In the higher valleys, which drain to the closed basins, cones from the two sides meet along the medial line, giving to the cross profile the form of an obtuse \y- Above the alluvial cones all is of solid rock, and the topographic forms are hard and angular. Every water-pai-ting is a sharp ridge, and every water-way is an acutely V-shaped gorge. The ridge and the gorge are characteristic features of land sculpture, being carved only where rain and running water serve for erosive tools. The alluvial cone is an equally characteristic land feature, being formed only where running water throws down detritus, without itself stopping. They are all the distinctive and exclusive products of land shaping, and could never originate beneath a lake or ocean. These are the features exhibited by the Bonneville Basin above the highest shore-line; and the same features can be traced continuously down- ward past the shore-line and to the bottoms of the once submerged valleys. If one stands at a distance and views the side of a valley, lie will see that each of the great alluvial cones is traceable within the zone of submergence almost as distinctly and quite as surely as above it. Its curving contour formed a part of every individual shore of the series. So, too, of the moimt- ain gorges and ridges; wherever they extend below the ancient water limit the shore-line can be seen to follow their contours in a manner demonstrat- ing that they were already in existence when the lines were drawn. The preexistent topography of the Bonneville Basin was therefore of terrestrial type and of subaerial origin. The sea-cliff's and embankments and sediments of the lake were carved from and built on and spread o^•er a 92 LAKE BONNEVILLE. system of reliefs which originated at a time anterior to the hike, when the (h-ainao-e of the mountains descended without obstruction to the bottoms of the valleys. In this respect, and in other respects to be developed further on, the pi-e-Bonneville conditions were identical with the post-Bonneville. Illustrations of this general fact could be adduced almost witliout limit, for they are aifoi'ded by all the slopes of the basin, but a few will suffice. In Plate VIII there appears at the right a jiortion of the western front of the Frisco Range. The crowded and uneven contour lines mark the posi- tion of steep-faced rock undergoing erosion. At the foot of the range is a system of alluvial cones, represented by contours with smooth curves and regular spaces. Still lower are the contours produced by wave action, and lowest of all is the outline of a playa. A moment's attention will show that the great alluvial cone at a, which, like a trunk glacier, is compounded at its head of a number of single cones, is represented at the base of the slope by the convexity at c. The cone h appears, though less plainly, at f; and the cone d appears at /(. The cone c is greatly disguised at g, being loaded with a group of embankments; Ijut it is jjrobable that it has liad something to do with the deflection of shore-currents whereby those em- bankments were originated. Conversely, the indentation at j represents tlie uiibroken rock-face at i, where for a space of half a mile no debris-convey- ing srorffe issues from the mountain; and the dearth of detritus in the remon A- is represented by the indentation at /. The maj) also suggests, what a study of the ground demonstrates, that the material built into embankments was derived by the paring away of the coast to the north f)f each locality of deposit. Considered by themselves, the monuments of the waves' activity are by no means inconsiderable; each grou]) of embankments contains some hundi'eds of millions of cubic yards of gravel; but they sink into insig- nificance when compared with the stupendous monuments of alluvial activ- ity on which they rest. They are mere appendages, and the erosion of their material from the adjacent slopes has by no means oljliterated, though it lias somewhat defaced, the alluvial forms. Granite Rock, an isolated mountain of the Salt Lake Desert, has at its north end a gorge dividing the extremity into two narrow spurs. About these spurs the Bonneville waters rose to a height several hundred feet U.S. GEOLOGICAL SUPVET LAKE BONNEVILLE PL Vm .luhus BuMi AC..,hil, Drown bv (• Tb'impttnD LAND SHAPING BEFORE SHOKE SHAPING. 93 above any alluvial accumulation. All about the .spurs there is a distinct terrace cut in the granite at the highest water level, and the same can be traced, less continuously but still unmistakably, along the sides of the gorge to its head. This relation could not subsist had not the gorge and the sj)urs been carved out in substantially their pi-esent form before the Avaves attacked them. Bradley, speaking of the canyon of Ogden River, says : It is evident that, when this canyou was originally excavated, the Great Salt Lake was not far, if at all, above its present level ; so that the rushing torrent which wore out this old rounded bottom met no check until it had passed entirely beyond the mouth of thecanyon. There followed a time when the lake tilled nearly or quite to its highest ter- race; and, meanwhile, the Ogtleii River continued to bring down the sand and pebbles which it had before been accustomed to sweep out upon the lower terrace, but now, checke The AnciuDt Outlet of Great Salt Lake. By A. C. Peale, Am. Jour. Sci., 3d series, vol. 15, 1878, pp. 439-444. A DISCREPANCY OF OBSERVATION. 95 both sides of the Portneuf where it comes into Marsh Creek Valley an upper terrace is seen, and in 1872 Prof. ¥. 11. Bradley also readily identi- fied an upper terrace in the Marsh Creek Valley at the level of about 1,000 feet above the stream. In Gentile Valley and in Cache Valley also, traces of this upper terrace exist." In the })assage referred to,^ Bradley mentions this terrace in connection with stream terraces, but does not speak definitely of its origin. Its interpretation as a shore feature therefore rests with Peale, who regards it as identical with the one observed by him "on both sides of the Pt»rtneuf " It has not been seen by me, but I am by no means sure that in seeking it I succeeded in following Bradley's route. With more con- fidence it may be asserted that Marsh Valley is not contoured by any well- marked shore-line. I was careful to study its slopes from stations at various levels and under favorable atmospheric conditions, and I failed to discover even the faintest trace of wave work. The same careful search was made for high-level shore traces in Cache Valley and Gentile Valley, but none were found. There are indeed terraces in Gentile Valley, and these are elsewhere mentioned by Peale, who found their altitudes 5,526, 5,242 and 5,186 feet;^ but they are stream terraces, not shore terraces. It is with reluctance that I record not only my inability to rediscover phenomena which another has reported, but also my opinion that his reported discovery was based on an error of observation ; but the question here in- volved is of such importance in its relation to the Bonneville history that it can not well be ignored. As set forth in the second chapter, there are various other types of ter- races liable to be mistaken for shore terraces ; and the ranging of shore ter- races and other wave-wrought features in the same horizontal line, or })lane, is a characteristic of great imjjortance in their discrimination. To the ob- server who places himself in that plane and views the distant hillside at his own level, certain elements of the various shore features appear united in a horizontal line. If he selects for his observation an hour when the distribu- tion of lights and shadows gives strong expression to the details of the con- figuration, he is able to detect a shore record so fjiint that he might cross and recross it repeatedly without suspecting its existence. Having searched ' Rept. U. S. Geol. Survey Terr, for l>-82, pp. 20a-'203. ^Eept. U. S. Geol. Survey Terr, for 1877, Washington, 1879, p. 001. 96 LAKE BONNEVILLE. with distant view and selected light for the reported high-level shore traces in Marsh and ( "ache Valleys, and having failed to discover them, I am satis- fied that Peale misinterpreted terraces formed in some other Avay. The matter is not fnlly set forth l)y the recital of the conflicting obser- vations. T\w Valley of Marsh Creek falls outside not only the Bi)uneville Hasin l)ut tlic (Jreat Basin. It is di'ained to the great plain of the Snake River by a deep and rather broad canyon which bears the marks of antitj- uity. The sides of this canyon, though of crystalline and schistose rf)cks, are not steep, and at the most constricted point there is a flood-plain a tliou- sand feet broad. If there was, as Peale supposes, a barrier at this point containing the ancient lake, then its cutting must have consumed a long period; and it is incredible that shore terraces have survived the contem- ])oraneous general waste of the surface. If there was no barrier at this point, then the supposed lake was a great inland sea, flooding the plain of the Snake River, and its shore tracings on the margins of that plain should have been much more conspicuous (by reason of the greater magnitude of its waves) than any drawn in Marsh Valley, — but they have not been discovered. Moreover, a body of Avater capable of fomaing the supposed shore ter- races in Marsh Valley would have extended not only to Cache and Gentile Valleys 1)ut to the Great Salt Lake Desert, and the work of its waves should be visible, if anywhere, (m the face of the Wasatch Range. In tliat region, the conditions for the generation of large waves are far more favorable than in the relatively narroAv valley of Marsh Creek. Nevertheless, a higher line has not been observed on the margin of the greater basin. Not only has Peale failed to record it there, but Bradley, Howell, Emmons, Hague, and King have expressly noted the BonncAnlle as the highest shore-line.^ It may be objected that the failure of these numerous observers to de- tect an upper shore-line is negative evidence merely, and should be given little weight in comparison Avith a single positi\'e obser\-ation. But the fail- ure to detect is in this case something more than a negation. Subaerial land sculpture is as positive a fact as AvaA-e-Avrought shore sculpture; and the as- ' F. H. Bradle.y, U. S. Geol. Surv. of Terr. Ann. Rept. 1872, p. 192 E. E. Howell, U. S. Geol. Snrv. West of the 100th Meridian, vol. 3, Geology, p. 2r)0. S. F. Emmons, U. S. Geol. Esplor. -lOth Parallel, vol. 2, Descriptive Geology, p. 441. Arnold Hague, Idem. pp. 421, 428. Clarence King, U. S. Geol. Explor. 40th Parallel, vol. 1. Systematic Geology, p. 491. U. S. GEOLOGICAL SURVEY LAKE BOXNEJVILLE PLATE IX. THE GREAT BAR AT STOCKTON. UTAH. NEGATIVE EVIDENCE. 97 sertion that the Bonneville is the highest shore-line implies the assertion that above it the topography is of the ordinary dry land type. Every recogni- tion of an ancient shore is based, consciously or unconsciously, on an ac- quaintance with the ordinary cliaracteristics of the features of the land as well as with the peculiarities of shores; and ability to discriminate the pres- ence of wave sculpture implies in the same degree ability to note its absence and its limits. The supremacy of the Bonneville shore has been recognized not only by many observers Init in a great number of localities, and an induc- tion resting on so broad a basis may justly demand of a conflicting obser- vation the most rigorous verification. K the reader will turn to Plate IX he will be able to realize the weight of this evidence. The view presents the Bonneville shore at the pass be- tween Tooele and Rush Valleys. The observer stands on the west side of the pass and looks eastward toward the Oquirrh Mountains. At the left lies Tooele Valley, open to the main body of the old lake. At the right is Rush Valley, which held a sheltered bay. The greatest waves came from the north, and, beating on the southeast shore of Tooele Bay, carved out a long line of sea-clififs. The debris was at the same time drifted southward part of it being built into a free spit 7,000 feet long and 150 feet high at the extremity, and another part being accumulated during lower stages of the lake in an immense bay-bar, obstructing the pass. The spit appears in the picture at the right, following the base of the mountain. The bay-bar extends from the center of the view to the foreground. It will be observed that the line of sea-cliff at its most distant point impinges on a spur of the mountain; and at its southern end, near the middle of the picture, it touches another spur, while in the interval it crosses only the alluvial slope. There could scarcely be a greater contrast than between the sculi)turing of the mountain-spurs above the line of sea-cliffs and the smooth contours of the slopes below that level. The cliffs are here of rather unusual height, and the shore emliankments are of exceptional magnitude, so that the separation between subaqueous and subaerial topography is more than ordinarily dis- tinct. This fact does not weaken the evidence that the Bonneville shore-line is the highest, but gives it greater strength. For, if the water had occupied a higher level in Pleistocene time, the waves would have been able to record MON I 7 98 LAKE BONNEVILLE. it at tliis jxtiut l)y a shore-line of unmistakable definition. If shore traces of a greater lake are anywhere preserved they should lie found at such a point as this, where the conditions for wave beating are exceptionally favorable. The same lesson maybe learned fnmi Fiiiun' 21, mikI fnuii tlu* views on I'lu. "Jl— BuiiiiL'Ville and Intermediate ( nibunknieuts near W, Ilsville, Utah. siinwiUji contrast betueeu Lilturul .iiid Subaerial Topo*:rapliy. Plates XXI and XXII, representing the shore topography and mountain topography at Wellsville and Dove Creek. MORE ANCIENT LAKES. Although Peale's supposed discovery is unverified, and though it is believed that an exhaustive investigation would prove it to be illusory, it is nevertheless true that some or all of the mountains of the Bonne^^lle Basin were girt by shore-lines long before the Bonneville epoch, and that if these shore-lines wei'e extant they would, in some places at least, lie higher than the Bonneville. The mountains against which Lake BouncA'ille washed are relatively very old, so old that they were greatly eroded before Tertiary tiiiie. Ever since their first u})lifting they have been wasted by erosion, and during at least a portion of the tune the detritus worn from them has TERTIAKY LAKES OF TDE BONNEVILLE BASIN. 99 been received by the interjacent valleys. The degradation of tlieir crests and the burial of their bases would long ago have obliterated them had they not been preserved by a series of supplementary upliftings, which, like the original, were ditferential, not being shared by the intervening valleys. In the region of the Great Salt Lake Desert, where a plain has been formed by the coalescence of many valleys and the local burial of the ranges, the depth of detritus must be several miles. Of the constitution of this depos- ited mass nothing is known by direct observation. It is smoothly covered by the sediments of Lake Bomieville, and no section is exposed. But indi- rectly we are shown that some ])art of the debris was spread under water, for the uprising mountain ranges have carried with them here and there, clinging to their flanks, small patches of lacustrine strata. It is believed that four separate groups of lake beds have been thus distinguished. The first of these occurs in the southeastern part of the basin, and probably touches the shore of the ancient lake only in the estuary of the Sevier River. No fossils have been found at that point, but there is little reason to doubt that the strata were once continuous with the Pink Cliff formation, which covers large areas farther east, and has been classed as early Eocene. The principal locality of the second is the eastern base of the Ombe Range, where an isolated outcrop of barren strata resting against the mountain dips abruptly beneath the later sediments of the desert. These strata have been correlated on lithologic grounds Avith fossiliferous beds farther west, and are regarded by the geologists of the Fortieth Parallel Survey as of Middle Eocene age. The third group, though yielding no fossils, is believed to be Neocene. It was first noted by Emmons in Rush Valley south of the Great Salt Lake Desert, and has since been found at the narrows of the Jordan River, at Salt Lake City, at the north edge of the desert near Matlin, and at the extreme northwest cornei' of the basin in Cache Valley, whence it extends across the rim of the basin into Marsh Creek Valley. The strata of the fourth group, known chiefly from the investigations of King and Hay- den and their assistants, occur at a number of points along the northern margin of the plain, and are believed to appear also north of -the divide in districts now draining to the Snake River. From Morgan Valley to Cache Valley they occupy a trough between two divisions of the Wasatch Range. 100 LAKE BONNEVILLE. On the low northward continuation of the main Wasatch ridge, where it separates Cache and Malade Valleys, they are seen to be wrapped around a series of low crags of Paleozoic rocks ; and it is evident tliat they liave been raised to their present prominent position by tlie reliftiug of an ancient crest. On the east side of it they have been upturned by the displacement so as to dip at a high angle beneath the Bonneville lacustrine beds of Cache Valley. On the west they are separated from theii- original continuation beneath Malade Valley by a fault, the tlirow of which is probably more than 1,000 feet. Their relation to tlie third group has not been established, and it is possible that they constitute a part of the same series. The local- ity of the fifth group is just north of Salt Lake City, wliere an epaulette of Tertiary gravel and sand rests on a jutting shoulder of the Wasatch Range. This fragment is completely surrounded by faults, its eastern continuation having been lifted high in air and obliterated by erosion, and its prolonga- tion in every other direction having been dropped so lo\\- that it is at once preserved and concealed by the deposits of the plain. This, too, is unfos- siliferous ; and it is here assigned to the upper Neocene merely on the strength of its structural relations. It is needless to enter upon these at this place ; l)ut it should be remarked that the same relations, considered from another point of view, led King to sunnise its Eocene age. Each of these lakes made its contribution to the filling of the basin, receiving, sorting, and spreading the debris from the wasting mountains; but neither can in strictness be called the predecessor of Lake Bonneville, for neither was confined to the area of the Pleistocene basin. So far as in- dicated by observed outcrops, the oldest Eocene lake lay almost entirely outside the Bonneville area; and it may have existed at a time when the greater part of that area was dry land. The second stretched westward far beyond the present drainage of the Salt Lake Desert, and may have over- lapped the Bonneville Basin but slightly. The third and fourth encroached northward on the drainage of the Columliia River. Too little is kuowu of the fifth to indicate its relation to the Bonneville Basin. Their record is exceedingly fragmentary, but if it were full it would still give an hnperfect history of the basin in Tertiary time, for there is no reason to believe that they represent more than a small imrt of that time NO TERTIARY SHORE-LINES. 101 They tell iis, however, that the physical mutations of the period included numerous local elevations and depressions, whereby the di'ainage of the country was repeatedly revolutionized; it was dry land at one time and and lake basin at another. It is quite possible that the lakes were excep- tional phenomena, and that the prevailing condition was one in which the whole area drained to the ocean. It is equally possible that the Bonneville Basin continuously held a lake which, as the land rose and fell unequally, was expanded and contracted, now in one direction, now in another. The character of the lake beds and their relations to the mountains, show in numerous localities that the ranges were not submerged. Waves must therefore have beaten on their flanks, and tlie cliffs, terraces, and em- bankments peculiar to shores must have been wrouglit, but of these there is no known vestige. When the structure of the mountains has been elabo- rately studied, so that those elements of their configuration which depend on the distribution of strata and on faults can be definitely indicated, it may be possible to point out dissected terraces and ruined sea-clifts as remnants of Neocene shores; but for the present such vestiges are beyond recognition. A shore is of the most perishable of geologic phenomena. It is little more than a congeries of forms ; and whether worn away by atmospheric agencies or buried by sedimentation, it ceases to lie available as evidence of a water margin. OUTLINE OF THE LAKE. The outline of Lake Bomieville at its highest stage was intrioate. Its shores presented a succession of promontories and deep bays, and it was beset with islands. Its longer diameter lay north and south, parallel to the trend of the mountain ranges of the district and to nearly all the lines of geologic structure. Its general outline was rudely pear-shaped, with the stem ])ointing southward. A straggling series of promontories and islands crossed it near the middle, dividing it into two principal bodies, of which the northern and Inrger covered the Great Salt Lake Desert, and the south- ern the Sevier Desert. The long southward bay representing the stem of the pear, occupied the Escalante Desert. The main body was joined to the Sevier body by three straits, of which the deepest and broadest lay be- 102 LAKE BONNEVILLE. tween Simpson Mountain at the east and Mt-Dowell Mountain at the west, in the r('(i;i()n now known as the Old River Bed. The EscaUinte Bay was connected with the Sevier body ))y a long strait, most constricted at Tlier- mos S])ving. The following details are of local rather than general interest, l)ut are essential to a full description of the lake. They will be more readily fol- lowed by the aid of the large map accompanying the volume. The trend of the ranges gave character to all the major details of the coast, and the axes of the larger islands, ])eninsulas, and bays lay approxi- mately north and south. Beginning at the north to describe them, we have first Cache Valley bay, an oblong sheet of Avater, tangent at one side to the main l)ody and there joined to it by a broad strait interrupted by several islands. Inside the bay were three islands, whose positions are now marked by Franklin, Cache, and Battle Creek buttes. The butte near Smithfield was likewise an island at first, though finally connected with the land by a bar. The canyons of Bear, Cub, Logan, and Blacksmith rivers were occupied by inlets, and the Bear River inlet may have reached at first to Gentile vallcA'. These were all gradually diminished by the deposits from the sti-eams, and eventually the Bear River inlet was approximately, and the Logan com- pletely, filled. Malade Valley held a long liay running northward from the main bod}', and having an expansion where the towns of Malade and Samaria now stand. Parallel but smaller bays occupied the Pocatello and Blue Spring valleys and the valleys containing Hanzel Spring and the town of Snows- ville. Park Valley was filled by a bay, exceptional in its east and west trend, and separated from the main body by a group of islands. The Prom- ontory range was divided by a strait at the point where it is crossed by the Central Pacific Railroad, the north j)art being a peninsula and the south a narrow, rocky island. Little Mountain, near the town of Corinne, Avas a small island, and the mountain from which Hanzel Spring issues made a group of islands. There were three small islands near the site of Kelton, and one just south of Ter- race. The Ombe range, including Pilot Peak, was an island, sheltering behind it a bay or sound from wliich a narrow arm ran northward to DETAILS OF ANCIENT GEOGRAPHY. 103 Thousand Spring Valley, the extreme limit of the water in a northwest direction. Of the existing islands of Great Salt Lake, Stansbury and Antelope were islands then, and Fremont barely showed its apex above water. Of the "lost mountains" of Great Salt Lake Desert, nearly all overtopped the flood. Silver Islet, Newfoundland, Terrace Mountain, Lakeside Mountain, Granite Rock, and a half-dozen nameless buttes, were circled by rocky and inhospitable coasts, but the Cedar Range west of Skull Valley made a broad and low island, which, bleak and barren as it now is, we may picture as then mantled with verdure. The eastern shore of the main body followed the steep base of the Wasatch Mountains, and had a simple outline except at three points, Avhere it was diversified by the estuaries of Box Elder Creek, Ogden River, and Weber River. The Box Elder estuary extended nearly or quite to the little mountain valley where the Danish settlement of Mantua lies. Ogden Canyon was occupied by a long and narrow strait, conuuunicating with a bay several miles broad, hemmed in by mountains. Through the canyon of the Weber a similar strait connected the main Ijody of the lake with a small bay in Morgan valley, — a bay on which the Weber delta gradually encroached, but wliich was not completely obliterated before the final subsidence of the water. The western shore of the main body followed the eastern base of the Gosiute range, and was characterized by an abundance of small islands. Its only estuary ran southward a short distance into Deep Creek Valley, stop- ping several miles north of the settlement. Southward from the main body ran four long bays, two associated with the east shore and two with the west. The first of these, counting from the east, was divided by a close stricture into an outer bay and an inner, the outer covering the valley of the Jordan River and the inner spreading over Cedar, Utah, and Goshen valleys and a part of Juab Valley. In the inner bay the Goshen Hills made two islands, and the Pelican Hills constituted one large and several small islands. Small estuaries occiipied Emigration and Little Cottonwood canyons, connecting with the outer bay, and the inner bay sent an estuary into Provo Canyon. The shalloAV arm in Juab 104 LAKE BONNEVILLE. Valley was nearl}' closed by one of the Goshen islands. It connected by the canyon of Salt Creek with the division of the bay in Cjoslien \'al]('y, and by the j)ass followed by the Utah Southern Eailroad with the l)ay in Utah Valley. The second of the southward stretchinm a peninsular tract bearing the Frisco and Picacho Mountains. There were two low islands a few miles broad close to the western shore, near Antelope Spring. The apex of Fumarole Butte was slightly emergent, and so was tlie highest point of the contiguous lava naesa. Small islands marked the sites of Pavant and Kanosh buttes, and there were four rocky islands near the mouth of Escalante Bay, one of which is now represented by the more northerly of the Twin Buttes. In Escalante Bay there were five or six islands. EXTENT OF THE LAKE, The area of the Bonneville water surface was 19,750 square miles, a magnitude ranking it with the Laurentian lakes. A fifth part of this belonged to the Sevier body with its dependencies, and the remainder to the main body. Its length, measured in a direct line from Cache Bay to the south end of Escalante Bay, was 346 miles, and its extreme width, from the mouth of Spanish Fork Canyon to a point on the Shoshone Range near Dondon Pass, was 145 miles. If its water surface were given a circular shape, its circum- ference would be 500 miles, but the actual length of coast, exclusive of isl- ands, was 2,550 miles. Its maximum depth was about 1,050 feet. The fol- lowing table will enable the reader to compare these dimensions with the corresponding dimensions of Great Salt Lake and the Laurentian lakes.' ' The area of Lake Bonneville wa.s me.asured by I. C. Russell ; the areas, lengths, and widths of the Laurentian lakes, by A. C. Lane. The length of a lake wa.s, for this purpose, defined to be the length of the longest straight line terminated by two points of the lake shore; its width, the greatest distance between shores iu a direction at right auglo to the hue of the leugth. 106 LAKE BONNEVILLE. Tablk I. Dimensions of Lakes. Bonneville. Great Salt. Superior. Haron. 23, 800 247 215 702 Michigan. Erie. Ontario. Area in aqnare milea 19, 750 346 145 1,050 •2. 170 83 51 f49 31.. 500 377 170 1,008 22, 300 330 106 870 9,900 246 58 210 7,250 197 67 738 Width in miles Extreme depth in feet * In 1869 ; near high stage. t At high stage. The greater part of the desiccated bed is an irreclaimaljle desert, liut its eastern edge is the granary of the Great Basin. The Bear, the Weber, the Jordan, the Sevier, and other tributaries, fed by the snow-banks of a score of mountain ranges and plateaus at the east, carry their hfe-giving moisture to the genial climate of the lowlands, and a belt of oases is the result. If the water were to rise again to its old mark, more than one hun- dred towns and villages would be submerged and 120,000 persons would be di'iven from their homes. The Mormon temple at Salt Lake City would stand in 850 feet of water, and the temple at Logan, the metropolis of Cache Valley, would stand in 500 feet of water. Fort Douglas would be covered to a depth of 150 feet, Ogden 850, Provo 650, Kelton 1,000. Seven hundi-ed miles of railroad would be immersed, and trans-conti- nental passengers would be transferred by boat either from Morgan City or from Spanish Fork to some point near Toano, Nevada, — a voyage of 145 miles for the northern route or 185 miles for the southern. The town of Fillmore would be half covered, the State House barely remaining on di-y land, and Mantua, Paradise, Morgan, and Minersville would be lake ports. Heramon, Bingham, Opliir, Vernon, and Frisco would be peninsular towns ; and the mining settlements of Drum and Buell would be stranded on islands. SHORE DETAILS. The sinuosity of the coast and its diversity of slope and material give to the shore phenomena the utmost variety. Every typical feature of non- tidal shores is well illustrated, and some of the combinations are perhaps unique. The abundance of salients and reentrants, of promontories and inlets, has occasioned a large number of spits and bay bars, while long beaches and liarricrs are rare. SEA-CLIFFS OF BONNEVILLE SHORE. 107 At an early stage of the investigation, the writer thought that the coasts facing in certain directions gave evidence of exceptional amounts of wave work, and imagined that he had discovered therein the record of prev- alent westerly winds or westerly storms in ancient times. This belief was dissipated by further study ; and he discovered, as students of modern shores long ago discovered, that there is a close sympathy between the magnitude of the shore features and the "fetch" of the efficient waves. The greater the distance through which waves travel to reach a given coast, the greater the work accomplished by them. The highest cliffs, the broad- est terraces, and the largest embankments are those wrought by the unob- structed waves of the main body ; and opposite coasts appear to have been equally affected. The most interesting details of the upper shore-line are found at locali- ties where similar details affect the lower shore-lines, and it will be conven- ient to describe them in discussing the order of succession of the shores ; but certain features should be mentioned here. The greatest sea-cliffs are as a rule carved from headlands and from the islands of the main body, but the highest of all occurs in the Jordan Bay at a locality known as the Point of the Mountain. For a distance of half a mile the cliff there has an aver- age height of one thousand feet, the eroded material having been swept to the southwestward and built into a magnificent spit, around the extrem- ity of which the Utah Southern Railroad winds in passing from Draper to Lehi. Another notable cliff occurs on the south face of a butte east of Dove Creek, and is visible from the Central Pacific Railroad between Ombe and Matlin. The eroded material was in this case swept eastward and north- ward, being carried about the angle of the butte, then an island, and dis- tributed in embankments on its eastern face. The cut-terraces of the Bonneville shore are narrow as compared with those of one of the lower shore-lines. They rarely exceed a few rods in width. A good example can be found on the flank of the Wasatch Range just north of Big Cottonwood Canyon and others on the north end of the Oquirrh Range near Black Rock. These are mentioned as being easy of access, but they are less striking than some that are carved on islands at various points nenr the margin of the Great Salt Lake Desert. 108 LAKE BONXEVJLLE. Spits are exceedingly nnmermis, being attached to nearly ;ill (if the ancient islands and to many of tln^ salients of the main coast. < )f tliose having some magnitude, the most accessible are at Stockton (1*1. IX), near Grantsville, Tooele Valley (PI. XV), at the Point of the Mountain between Draper and Lehi, on Kelton Butte near Ombe station, and on the extremi- ties of the Terrace Mountains. Fig. 22.— Butt* near Kelton, Utali. Embankments connecting islands with each other (u- with the main- land are to be seen at the west end of Park Valley, at Smithfield in Cache Valley, on Antelope Island in Great Salt Lake, a few miles east of George's Ranch south oi Deseret, and at the eastern base of the Gosiute Range. V-shaped embankments are most numerous in Snake Vallo}-, where no less than ten occur. Four are attached to tlie Simpson Mountains opposite to the Old River Bed and others were seen in Preuss Valley and in Beaver Creek Valley. THE CUP UF CUP BUTTE. 109 Typical deltas are rare. Certain parts of the valleys of all the }}riiici- pal streams were occupied by inlets or estuaries, and the heads of these inlets received alluvial deposits of the nature of deltas; but the process of accumulation appears usually to have been arrested before the deposit had extended to the open lake ; and afterward, when the lake receded and the streams resumed their work of excavation, all but scattered patches of the alluvium was removed. American Fork, Spanish Fork, and Rock Creek built free deltas in the Utah Bay, and Spring Creek furnished one to the shore of Cedar Bay, but these v^ere exceptional and small. At lower levels great deltas were constructed by many streams, and the deltas of the Bonne- ville shore ai'e described in connection with these in one of the later sections of this chapter Plate VI exhibits a peculiar circular bar observed in a single locality only. The sketch is in part ideal, for there was no commanding point from which to obtain the bird's-eye view necessary for the best presentation of the subject. Near the Old River Bed there is a group of quartzite buttes which were surrounded by deep water and formed a cluster of rocky islands. To the north and northwest the deep lake stretched unbroken for more than one hundred miles, but in all other directions land was near at hand. Each island butte shows a weather side facing the open water and a lee side fac- ing land. Each weather side is marked by a sea-cliff, which looks down on a broad terrace carved from the solid rock. The lee sides have no cliffs, but are embellished by embankments of various forms, built of the debris from the weather sides. In the case of the butte figured, the excavation of the platform was carried so far that only a small remnant of the original island survived, and a comparatively small additional amount of wave work would have sufficed to reduce it to a reef. From each margin of the sur- viving crest, an embankment streams to the leeward, and the two embank- ments, curving toward each other, unite so as to form a complete oval. At their point of junction they are a few feet lower than where they leave the butte. Their material is coarse, ranging up to a diameter of two feet, and is conspicuously angular, exhibiting none of the rounding characteristic of detritus that has been rolled long distances upon a beach. Within the oval rim is a cup 38 feet deep, its sides and lip consisting, on the north, of the no LAKE BONNEVILLE. rocky slope of the butte, and elsewhere of the wall of loosely heaped blocks of quartzite. If the material were volcanic, instead of sedimentary, it would be easy to imagine the cavity an extinct crater. Reservoir Butte, another island of the cluster, is figured in PI. XXIV, and further represented in PI. XXV and in Fig. 3 of PI. VII. It derives its name from a series of natural cups analogous to the one just described. These are attached to its steep slopes at various levels, the process of con- struction having been repeated at as many epochs in the history of the oscil- lating lake. In this connection, only the cups associated with the highest shore-line will be described. The longer diameter of the butte trends north and south. At its northern extremity and along its northwestern face it displays a bold sea-cliff, from 50 to 100 feet high, springing from a terrace at the Bonneville level several hundred feet broad. On the eastern side the cliff and terrace give place near the north end to a massive embankment, which first swings free from the side of the butte and then curves inward toward it, meeting it somewhat south of the middle. From the middle of the western side there starts a similar embankment, which, curving through an oval arc of 150°, joins the butte at its southern extremity. The interval between the tei'mini of the two embankments, a space of 1,000 feet along the southeastern face of the butte, was almost unaffected by the waves, being neither abraded nor covered by debris. The material contained in the embankments was derived exclusively from the weather side of the butte, and though each looped embankment joined the shore at two points, the conveyance of shore-drift along its crest appears to have been in one direction only. It is difficult clearly to realize the process of this con- veyance, but there is no question as to the fact. In one case it left the shore at a small salient, its course being there tangent to the contour, and, curving through an arc of 90°, finally assumed a course directly toward the coast, there almost precipitous. In the other case it left the shore at an obtuse salient, and before returning swung through so great an arc as nearly to reverse its direction. The cups witliin these loops have been somewhat silted u]i in modern times, but still, except for their diyness, they deserve the name of reser- voirs. The eastern was found to be 38 feet deep. The embankments were THE GUPS OF RESEKVOIR BUTTE. HI built in deep water and upon a foundation inclining steeply from the shore. Their forms are independent of the configuration of their foundation. They were not accumulated from the bottom upward, but were constructed by successive additions at the end, the boulders being rolled along the crest of the embankment by the breakers and then dropped in deep water at its extremity. The outer face of the eastern bar has a height above its base of four hundred feet EMBANKMENT SERIES. It might be inferred from the preceding description that the Bonneville shore-line was the product and is the index of a single uniform and continu- ous water stage. Indeed, it has been so regarded by every observer who has published an account of it, and the impression is readily and properly derived from its ordinary phase. There are, however, a few localities where the shore mark is distinctly resolvable, and shown to be compounded of several similar elements at slightly different heights superposed on one another. One of the most striking localities, and at the same time the one which first demonstrated the compound nature of the phenomenon, is repre- resented in PI. X. A rocky cape projecting from the east shore of Snake Valley sheltered on one side a small bay opening to the south. Across this bay the waves built a series of bars, as represented in the map. The outermost of the series, that is, the one farthest from the land, is connected at its eastern end with a shore cliff labeled on the map " Bonneville Sea- cliff"; and this cliff runs for some miles southward along the slope of the valley. A study of the locality demonstrated beyond question that the excava- tion occasioning the cliff and its terrace, furnished the material for the bar, and furthermore, that the same cliff line had previously been connected with each bar of the series. It will be readily understood that the inner bar was the first one to be built, and that the order of position is also the order of age. They stand so nearly at the same level that no one of them could have been formed in the rear of another. Their differences of level therefore record changes in the relation of the water to the land during the period of their formation. If 112 LAKE nONNEVILLE. we call the inner bar No. 1 and its altitude 11 feet, the series will be repre- sented by the following list : Feet. No. 1 11 No. 2 12 No. 3 13 No. 4 4^ No. 5 4i Feet. No. 6 ^ No. 7 8 No. 8 0 No. •) 0 No. 10 18 No importance is to be attached to the individuality of the bars. There is a rhythm of action in the process of their formation which would prevent the construction of a continuous and even-topped terrace under the most uniform conditions. If the bay had been so shallow that the same accu- mulation of shore drift would have abridged it twice as much, there might have been twenty bars instead of ten. The first tlu-ee bars signify but a single epoch, during which the water stood at one level, or perhaps rose slowly. The next thi-ee, which in point of fact are but obscurely indi\id- ualized, represent a succeeding water stage eight feet lower and ])ossibl\' of somewhat greater dm-ation. The seventh bar shows that the next move- ment consisted of a deepening of the water and was not long sustained. The eighth and ninth record the lowest stage of all, and the tenth the highest. The tenth contains so much more material than either of the others, being founded in deeper water and carried higher, that it must be considered as representing a longer time, and may be coordinated with either of the ante- cedent groups. Outside the tenth bar the plain slopes gently lakeward, being inter- rupted within the area of the map oidy by a low bar, indicated in the pro- file. This bar lies so far below the others that, if older, it might not have interfered with the wave action necessary to their formation. Its relati\e age therefore does not appear. The process of construction is clearly demonstrated by the local details. The sea-cliff was excavated from the alluvial foot slope of a mountain range. The derived material consisted primarily of boulders, large and small, sand, and a certain portion of clay. The finer part was immediately washed lakeward by the undertow. That of middle grade was carried along the shore to the bay, and the larger boulders remained in situ until sufficiently U S. GEOLOGICAL SURVEY LAKE BOHNE'vTLLE PL.X MAP OF BAY BARS OF THE BON.XEVILLE SHORE Near \lw Salt Marsh, in Snake Vallev, Ttah Bv ^^' D Johnson JO-ff'€t Contours Y\% Profile. Vfjiical Sralf t/uef times the Hofizontal . Julius Bipti 4 Co.Iith DraifH bu'G Thoinpst SNAKE VALLEY BAY BARS. 113 reduced by attrition to be transported. In the bay tlie surface currents ^^ere concentrated by converging shores, and a powerful undertow was pro- duced, whereby a further separation was effected, the shore drift being de- prived of a coarser grade of debris than that previously eliminated, so that tlie matter actxially deposited consisted of particles ranging from a lialf inch to turn- inches in diameter, — a clean shingle without admixture of sand. The sand and fine gravel thus eliminated by the undertow were deposited in h)rge part near the liead of the bay, causing the water to shoal rapidly, and uhimately determining the breaker line to a new position outside the first, and tlius initiating the construction of a new bar. In this way the depth and length of the bay were at the same time progressively diminished. For purposes of comparison the profile of the Snake Valley bars has been repeated in PI. XI, where a series of .similar phenomena are also drawn to the same scale. A brief description will be given of each locality. At the head of Skull Vallev, a few miles north of Government Creek, there is a low albnial A\'ater-})arting se})arating the valley from the open desert at the west. At the time of the Bonneville water stage this pass was reduced to an istlunus only a few rods in width, and the Avater was shallow on each side. On the Skull Valley side there were formed a series of bay bars, represented in profile in the plate. The winds under the influence of which they were formed, covdd have blown only from the northward. The third profile represents in similar manner a group of bay bars ob- served a few miles east of Sevier Lake. The general trend of the old shore- line is there north and south, but at this particular spot there was a small cove lying on the north side of a rocky promontory. The bars were formed by northwesterly winds. The fourth locality is a few miles east of the third, l)eing on the oppo- site side of the Beaver Creek mountain range near George's Ranch. A small rocky hill was insulated at high-water stage by a narrow and shallow strait, and across this strait embankments were eventually built l)y the north- easterly winds. The first of the embankments, however, did not completely close the passage, and remains as a spit, while the others are completed bars. The topographic relations are shown by Fig. 23. MON I 8 114 LAKE BONNEVILLE. The locality of tlie fiftli profile is the southwestern an\' a discordanci' among I :, s, (;i':iii.oi:li..\l. suiA'KY REPOHT Oyr LAKE BONNEVlLIjE. J 'J, ATE XI HYPOTHESES AND TESTS. 1 1 7 the systems of bars. The theory of lake movement would be sustained by an accordance. An imperfect accordance miglit indicate a combination of the land and lake changes. The facts are assembled in PI. XI, to which the reader is again referred. Each of the profiles represents a section at right angles to the system of bars it illustrates, and all are dra\\u to the same scale, the vertical element being exaggerated three-fold. They are grouped on the page in such man- ner that the outer embankments of the several series appear at the right and fall in the same vertical column. The first consideration affecting the comparison is that each series pre- sumably represents the same period of time, so that, if a correlation is pos- sible, the embankment drawn at the right in one series should correspond to that at the right in the others. That at the extreme left in one should correspond to that at the extreme left in tne others, and the intermediate portions should be comparable. The only exception to that rule is in the case of the Dove Creek series, which, as already explained, may represent only the later portion of the time consumed in the formation of the others. Restricting attention to the first five groups of bars, we note first that the right-hand member of each is higher than any other. The second con- spicuous fact is that the member second in size stands at the extreme left. To this there is a single unimportant exception, which vanishes if we con- sider the three bars at the left of the upper profile to constitute a single member comparable with the individual bars of the other series. It is by no means improbable that a more careful stud}' of the Skull Valley locality would resolve the left-hand maximum into such a series as was found in Snake Valley. The most extended series exhibits a third maximum, lower than either of the others, but intermediate in position and standing somewhat to the right of the middle of the profile. No other profile shows a third maximum, but three of them exhibit bars of approximately the same height, which may be conceived to represent it, if the bars of the second minimum are assumed to have been covered and concealed by the great outer bar. It is easy to understand that a condensed or foreshortened series would exhibit super- ficially only the maxima of a fully extended series. It therefore seems 118 LAKE BONNEVILLE. proper to correlate the intermediate maximum of the upper profile with the bar appearing at the inner base of the outer niaxiiiuiin in tlie second, tliinl, and foui'th profiles. In the fiftli profile, bars representing tlie fii-st mi id third maxima stand in juxtaposition; and it is necessary to assume tluit tlic inter- vening maximum, as well as the two minima, is covered and concealed. It thus appears that, in their most general features, the groups of bars are in accordance, with no greater variation than might readily be ascriljed to local disparity of condition. The difference between the altitude of the outer bar and that of the intermediate maximum was measured in tour localities. In Snake Valley it is 10 feet, in Skull Valley 12 feet, in Sevier Lake Valley 15.3 feet, and at George's Ranch 15 feet. The range of these measurements is 5 feet, and this must be regarded as a real discrej^ancy, though not a gi-eat one. The altitude of the outer bar above the inner maximum was measured at five points and found to be 5 feet, 10 feet, 10 feet, 7 feet, 8 feet, — the enumeration following the order of the diagrams. Here again the range is 5 feet. If the inner bar be compared with the intermediate maximum instead of with the outer bar, the diff"erences are found to be 5 ft, 2.7 ft., 5.3 ft., and 8 ft., showing again a range of 5 feet. Finally, the low bars observed between the inner and intermediate maxima have approximately the same relation in tlie tln-ee localities where they were observed. Compared witli the intermediate maximum, their measured difi"erences ai'e 3.5 ft, 2.3 ft. and 4.7 ft., tlie nuige being 2^ ft. These com])arisons exhaust the data, and they appear to establish the systematic hannony of the phenomena. It is inconceivable that such ac- cord shoidd be fortuitous. The most complete record (that in which the bar system was spread out most broadly, so as to resolve it most completely into its elements) exhibits three maxima with intermediiitc miiiim;i. Tlie record second in extent shows the three maxima and one miniiiimii, — the other minimum being overplaced and concciilcd. Tlie Sevier Lake ri'cord shows the same four elements, but more compactly arranged. At George's Ranch the three maxima are so closely crowded that l)oth minima are con- cealed. At the head of Tooele Vallev, the outer and inner maxima are in ADJUSTMENT BY LEAST SQUARES. 119 juxtaposition and all the intermediate elements ot" the series are buried. The ordinary bay bar, in which all the elements are welded together and covered by the last and highest deposit, is logically the final term of the series of facts. The hypothesis of water movement is therefore sustained. The chang- ing relations of land and water during the formation of that complex record to which we have applied the title of the Bonneville shore-line, were brought about by the alternate rising and falling of the water surface. While the higher bars were being formed, there was more water in the basin; while the lower, less. Having thus established the correlation of the series of profiles by a comparison of the unmodified facts of observation, it is now proper to adjust them to one another for the purpose of ascertaining the mean (juantitative value of changes of water level. Applying the method of least squares, we obtain for the most probable values of the water stages, referred to the low- est of the series as zero and arranged in the order of time:* feet. First maximum 12.3 ± .2 First minimuin 3.0 ± .2 Second m.axiinum 7,3 i .2 Second minininm 0.0 Third maximum 20.1 ± .2 Adjusted to the same zero, the observations at the several localities ex- hibit the following relations: Table II. Embankment Series of the Boiinei^Ule Shore-line. Locality. Allilnile in feet. Variation fioDi ailjusfoil mean. 1st Max. 1st Miu. 2.1 Max. 2a Min. 3ll Mas. lat Max Iflt Min 2il Max 2.1 Mm ad Max Snake V.illey Skull Valley Sevier Lake Valley. George's Ranch Tooele Valley in.o 10. .■! 12 2 13.0 12.2 4.5 5.3 2. 2 8 0 7.6 C.9 5.fi 0.0 18.0 20.3 22. 2 20.6 20.2 + .7 —2.0 — .1 + 1.3 — .1 + .6 + 1.4 -1.7 + .7 + .3 — .4 -1.7 —2.1 + .2 + 2.1 + .5 + •1 1 ' The computation incliuled data from the Dove Creek protile aud from tho PruiiiiB Valley bars. It was performed by Mr. A. L. Webster. 120 LAKE BONNEVILLE. The residual discordance, as shown by the cohimns at the nght, is not large, thougli it is somewhat greater than the range ot" variation found in the longitudinal profile of the crest of a single har. A part of it is i)n)b;il)lv due to inaccuracies of measurement; no instrinnents of j)recision were emj)loyed, and the methods at more than one locality were improvised and crude. There will be no impro})riety in referring the remainhig part to exceptional storms combined with local conditions. Reverting now to the Dove Creek series, wliicli the field observations gave reason to suspect of incomj)leteness, we find by inspection that its two levels can readily be correlated with the second and third maxima of the gen- eralized profile. It is highly probable, therefore, that the earlier water stages, including the first maximum and the first minimum, failed to make an independent record at that ])lace. To convert the data fully into terms of lake history it is necessary to comjiare the epochs of formation of the several l:)ars in the matter of duration as well as in that of water stage. The amount of shore di'ift accumulated in the several bars has to be considered, and likewise the manner in which the varying water stage affected the rate of accunudation. A determination of absolute duration is manifestly out of the (piestion, and any estimate of relative duration is largely a matter of indi\ddual judg- ment. An attempt has been made in Fig. G of PI. XI to represent the oscilla- tions and their periods in a quantitative way, so far as they are dediu-ible from the plienomena. If the facts permitted xis to draw the full curve of oscillation with all its details it would unquestionably be far less simple. The number of minima concealed by the bars of even the most extended series may be very great; and it is even possible that these bars do not re})- resent a continuous history. If, after the series had been ])artly formed, the lake shrank to nuich smaller dimensions, returning to the region of the Bon- neville shore only after a long interval, there seems no wav to determine this fact by the phenomena of tlie shore. Probablv tlie only conclusions deducible from the profiles are; first, that, when the lake basin was full, the position of the water level was unstable; and, second, that of a series of high-water stages, the latest was the highest of all. INTERPRETATION OP V-HARS. 121 It will perhaps occur to the reader that the enumeration and discussion of these facts have been needlessly prolix; and this I am not prepared to deny. But it may be said in extenuation that the phenomena belong;- to a novel tvpe, and that the method of investigation Avas so far new that the simple conclusions finally reached required for their establishment a full presentation of the alternative hyj^otheses eliminated by the investigation. In the sequel it will appear that even these simple conclusions afford a key to the understanding of some of the most important elements in the history of the lake, and through that history are brought into relation to the prob- lem oi' the physical condition of the earth's interior. One result of tlie discovery and interpretation of the groups of bay bars of the Bonneville shore-line was the explanation of certain features of the V-embankments which had previously been problematic. V-embank- ments have already been described as triangular terraces built against mount- ain slopes at the shore level, and margined toward the lake by even-topped parapets. In the light of the conclusions thus detailed it becomes evident that this conformation was occasioned by oscillations of the lake during the period of the formation of the terrace. The space within the parapet is usually occupied by a playa, the surface of which is from five to eight feet below the enclosing rim. This represents a certain amount of silting up of the basin. If there were no filling, it cannot be doubted that the interior of each enclosure would exhibit a series of bars parallel to one or both arms of the jiarapet, and corresponding in height and arrangement to the bay bars. In fact, this very phenomenon was finally observed at several localities. The most interesting are in Preuss Valley along the western base of the Frisco Mountains. In that valle}' the shore features of many different hor- izons afforded an instructive study, and were carefully mapped. PI. VIII gives a general view of the phenomena on the east side of the valley, and it will be noted that the Bonneville shore-line includes three of these tri- angular terraces. The same appear on a somewhat larger scale in Pis. XVI, XVII, and XVIII. The parapet associated with the middle group of embankments (PI. XVII) offers an exception to the general rule, in that it is broken through by the drainage, so that the interior contains no playa. It contains instead the eroded remnants of a system of bars parallel to the 122 LAKE BONNEVILLE. southern pai-apet. In this system it is easy to recognize the equivalents of first and second maxima of the Snake Valley bars, holding their j^roper re- lation to the para])et, which corresponds to the third or outer maximum. The V-enibankment of the south group, PI. XVIII, is undi-ained, but its filling has not progressed so far as to obliterate the inner maximum. Two elements of the bay-bar series are therefore represented ; and the same were found in the north group of embankments. In the case of the middle and southern of these Preuss Valley embank- ments, and in two f)r three other instances, the interior embankments are parallel to one parapet only, so as to constitute with that a series of parallel ridges connecting the remaining parapet with the shore. It seems evident that in these cases the growth of the triangular terrace was chiefly or en- tirely by additions to a single face; and it may not be improper to define the aggregate structure as a spit gradually projected into the lake by recur- rent storms from a certain direction and buttressed by successively formed bay bars connecting its extremity at various stages with other points of the shore, the bay bars being the work of a series of storms from a diff'erenr direction. The variety of contour assumed by the parapets of the V-erabank- ments, and by the crests of the hooks and loops with which they are more or less affiliated, is illustrated by PI. VII. DETERMINATION OF STILL WATER LEVEL. One of the collateral results of the composite nature of the Bonne- ville shore-line is a discrepancy in the evidence aft'orded by different parts of the shore phenomena as to tlie altitude of the ancient water level. Tliose parts of the coast which were given their character by excavation indicate the water level by a line forming the angle between a cliff above and a ter- race below, and this line often represents the lowest of the series of water levels recorded by the bay bars. Tlie im])ression made by the waves at the last and highest level is usually, thougli not always, so faint that it has been obliterated by the falling down of the cliff. On the other hand, those parts of the shore formed by the accunuilation of detritus appear as a rule at the highest water stage only. The localities in which embankments represent- U S.GEOLOGIC^AL S'JRlrEy JjAI-'vE BOiJN'B'.lr.bE PL. xn 42°'- 41' 40' 39° 38° US' !uliu9 Bien * Co, lilK MAP OF E BONNEVILLE I'lIKSKNT UYliRdllRAFMIC DIVISIONS OF THE BONXE\TLLE BASIN n red ) and The areas with alUUide ijreater than 7000 feet. cin "blice ; \'n[c The dotted lines indicate dottbi 42 41° SCA LE : t Dt;i«Ti tfv C Thuiopsi FINDING TUE STILL WATER LEVEL. 123 ing progressive action are differentiated, are exceptional ; and in ordinary- cases the latest additional material covers all the preceding. For an accu- rate determination of the height of the niaxiinnni water level, it is therefore necessary to consider the character of the record to which measurement is aj)plied. The base of a sea-cliff is apt to give too low an indication, while the crest line of an embankment is not. If this element were the only one to 'be taken into accoimt, it would be a simple matter to ascertain in every region, by using emljankments only, the precise height of the old water level; but there is unfortunately a com- plication. The crest of a completed embankment always stands somewhat higher than the still water level of the lake to which it pertains ; and the amount of the difference depends on conditions which are not entirely sim- ple. They include some elements of the configuration of the bottom, and especially the magnitude of the largest incident waves. The same elements of configuration affect also the record embodied in the base line of a cliff, but the magnitude of the waves does not. On a coast foeing deep water the base of the sea-cliffs coincides very closely with the still water level. If, therefore, the surface of Lake Bonneville had not fluctuated while near its highest stage, the sea-clifis would aft'ord a more intelligible record of its precise horizon than the embankments. As the case stands, the best indications are sometimes afforded by one class of facts and sometimes by the other. Wherever it is evident that the sea-cliffs associated \\ ith the maxinuun water stage survive, their base is assumed to give the most authentic record. Where these cannot be dis- criminated, embankments have been em])loyed, an allowance being made for their height above the water line. This allowance is a matter of judg- ment in each individual case. It will be instructive to illustrate the difficulties of the subject by a few examples. If the reader will refer to the general map of the lake, he will see that the Jordan valley was occupied ]>y a large bay receiving waves from the open lake, while the Utah Lake valley was occupied by a land-locked bay affected by no waves but those generated within its own borders. These two bays were joined by a narrow strait at the locality now known as the 124 LAKE BONNEVILLE. Point of the Mountain, and from the coast east of tliis strait tliere was con- structed an iniiiiense triaiij^'uhir terrace, receiving upon one side the (h'tritus rolled by the great waves of the Jordan Bay, and on the other the slioic drift moved by the snitdler waves of the inner bay. The parapets on the two margins of the V-shaped embankment give clear expression to this disparity of conditions. Tliat facing Jordan Bay is the more massive and tlie longer, and the other is Iniilt against it as a sort of appendage. The general altitude of the larger bar is six feet greater than that of the less; and since the latter has all the features of a completed endjank- raent rising above the water level, it follows that the northern or higher eni- bankiuent was built more tlian six feet above the still water level of the lake. Kelton Butte (Fig. 22) projected its apex as a small island above the water level and was surrounded by deep water. From one direction it re- ceived waves propagated through a distance of thirty miles, and by these a cliff and terrace were carved out and an embankment was constructed. The terrace is itself tei-raced in such way as to encoiu-age the belief that the base of the cliff corresponds Avith the highest water stage; but this base is 7i feet lower than the contiguous embankment. At a locality in Preuss Valley, where the conditions did not admit of the generation of waves of great size, an embankment has lieen connected by leveling with a sea-cliff and terrace, and found to be 5 feet higher than the terrace. In this case part of the discrepancy is doubtless referable to the failure of the waves at the highest stage to score a durable record on the face of the sea-cliff" carved at a lower level. A similar measurement was made at Wellsville in Cache Valley, where also the waves were not of the greatest magnitude, and gave a difference of 19 feet. At the opposite end of Cache Valley, near the town of Franklin, tliere is a small indentation in the shore in which an isolated embankment has been preserved with a crest 12 feet above the base of the adjacent sea- cliff; and in a sheltered spot north of the town of Tecoma, in the northwest- ern portion of the Ijasin, the measurement of similar details showed a differ- ence of 20 feet. The state of preservation of the embankments is all that could be de- sired for purposes of measurement. The innjority of them are composed of DEPTH OF THE OLD LAKE. 125 gravel, and are exempted by their ridge-like form from the destructive action of cross-flowing drainage. A few inches at most would express the loss their crests have sustained from the wash of the rain. With tlie sea-cliffs and wave-cut terraces it is different. The decay of a cliff' throws (lo\\n a con- stantly increasing amount of del)ris, which falls to the base ;ui(l foi-ins a talus; and every little drainage channel by which a cliff' is divided spreads-a heap of alluvium upon the terrace below. The base of the cliff, therefore — the element of the jirofile which for purposes of measurement it is most desiral:)le to recognize — has been almost universally covered by the rising alluvium, so that its precise position is a matter of estimation or indirect observation. The discovery that the old Avater line is no longer of uniform height, and the tact that its variations of altitude afford a means of measuring the recent differential movements of the earth's crust within the basin, give occa- sion for great regret that the exact identification of the highest water stage is so difficult a matter. In a majority of instances the range of uncertainty, after all allowances have been made, amounts to five or six feet. DEPTH. The greatest depth of the lake was about 1,050 feet; and this depth obtained over all the western ])art of the present site of Great Salt Lake. The })oint west of Antelope Island, where the deepest water in Great Salt Lake is now found, did not sustain the same relation to Lake Bonneville, ))ut was rivaled and perhaps sur})assed l)y jioints between Promontory and the Terrace mountains. The Great Salt Lake Desert has now a re- markably flat floor, and the ancient de])th of water above it did not vary greatly in diflerent parts. The mean de])th of the main body of L'lke Bon- neville was in the neighborhood of 800 feet. The Sevier body had a max- imum depth of G50 feet, and Esciilante bay of about UO feet. THE MAP. The mapping of the Bonneville shore received careful attention; audit is pro])alde that the extent and fonn of no modern lake in an unsettled country is more accurately known. The determination of certain questions 126 LAKE BONNEVILLK. with reference to overflow necessituted tlie inspection (if a lariic part of the periphery; and the knowledge thus obtained ol' tlic position of tlic coast was afterwards systematically supplemented until a complete ma]) Wecame possi- ble. T\u' insulai- mountains standing- on Oreat Salt Lake Desert were not visited, and the coast lines about their sides were for the most part deduced from the contours of the published maps of the Survey of the Fortieth Parallel; but with this exception all of the coast was seen by some member of the corps and sketched from actual observation. A large pai-t of it was examined by more than one individual. The map is indebted to Mr. C4il- bert Thompson for tlu^ details of the west coast between Deep ('reek and Montello, and for the bays at the north ends of Pocatello and Malade \'al- leys. He delineated also the details west of Sevier Lake and in the southern extension of White Valley. The map is indebted to j\Ir. Thompson and Mr. Albert L. Webster for the outlines of the Escalante Bay. Mr. Willard I). Johnson delineated the shores of the White Valley Bay and the coasts on the Dugway, MacDowell, and Simpson Mountains. The outline in Tintic Valley was furnished by Mr. H. A. Wheeler. Mr. Israel C. Russell map])ed the bay east of the Canyon Range, and is responsible for most of the coast between Fillmore and George's Ranch. He contributed also numerous de- tails in all parts of the basin. The remaining portions of the shore were mapped by me. Some idea of the distribution of responsibility for the maj), as well as of the thoroughness of the exploration, may be derived from an examination of PI. Ill, where the routes of travel are exhibited. THE PROA^O SIIORE-I^INE. Below the Bonneville shore-line are numerous other shore-lines, amting which one is cons})icuous. The name Provo was given to it on account of a great delta, which is at once a notable feature of the shore-line and a prom- inent element of the topograj)hy of Utah N'alley in the vicinity of the town of Provo. The shore mark so far surpasses in strength all others of the series that this character serves for its identification; and it has been recog- nized in all parts of the basin without the necessity either of tracing its meander or of measuring its altitude. It has indeed been recognized with confidence des})ite conflicting determinations of altitude, for it is neither THE PROVO SHORE-LINE. 127 uniform in height nor uniform in its vertical relation to the Bonneville shore- line. In a general way it is 375 feet lower than the Bonneville shore and 625 feet higher tlian the water of Great Salt Lake. The Provo record is more recent than the Bonneville. "^I'liis a])pears, first, from its state of preservation; tlie Provo cliffs are the steeper and sharper and the smaller talus lies at their base. It appears, second, from the absence of lake sediments on the surfaces of the Provo terraces. Dur- ing the formation of the Bonneville shore, the horizon of tlie Provo was sufficiently submerged to receive a layer of fine sediment; and a lake de- posit commensurate in amount with the shore drift accumulated in the Bon- neville embankments would not escape detection if it had rested on the terraces of the Provo shore. The relative age is shown also Ijy the relation of the shores to the outlet of the lake, as will be explained in another chapter. The duration of the water stage recorded by the Provo shore was greater than that of the Bonneville water stage. Although tlie Bonneville is the most conspicuous of all the shore-lines, it does not exhibit the greatest monuments of wave work, but owes its prominence largely to its position at the top of the series, where it is contrasted with topographic features of another type. There are several other shore-lines which rival it, and, al- though it probably outranks in magnitude all except the Provo, its discrim- ination would be a difficult matter were it an intermediate member of the series. The Provo, on the contrary, is rendered conspicuous chiefly by the magnitude of its phenomena. Its embankments are the most massive, and its wave-cut terraces are the broadest. Moreover, the Provo Lake was in every way inferior to the Bonneville as a field for the generation of jiowerful waves. It was narrower and shallower and obstructed by larger islands. To have constructed shores eciual to those of the Bomieville, it must needs have existed a longer time; and still longer to have built its greater struct- urcis. OUTLINE AND EXTENT. The outline of the lower shore was the less tortuous. The sinuosity of the Bonneville shore is due to the fact tliat the water flooded a large num- ber of the narrow trouglis of the Great Basin and was partially divided by 128 LAKE BONNEVILLE. the mountain ridges. When the Avater retreated to tlic I'rovo level, it ;il);iii- doned a considerable number of the valleys and retired on in;iii\ jiarts of the coast from the uneven mountain faces to the smooth contours of the alluvial slopes. Two of the largest bays, the Escalante and the Snake Val- ley, were completely desiccated, and so was a third part of the Sevier Des- ert. The water was withdrawn from Thousand Spring and Buell Valleys, from Gi'ouse Valley and Park Valley, from Ogden Valley and Morgan Valley, from Cedar Valley, Rush Valley, and Tintic Valley, and from both ends of Juab Valley. Of the three straits joining the Sevier Ixtdy with tlie main body of the lake, only the eastern remained. The closhig of the cen- tral and western straits joined to the western peninsula the islands Avhicli hann U S. GEOLOGICAL SURVEY LAKE BONNEVILLE PZyj^ MAP OF THE PASS bPiwHon RUSH AND TOOELE VALLEYS. ULUl. ShoM-ind the NViVVE HUILT B.UJRIER Rv H A VVhoelei- aOOO Q 100 0 2O00 SCALE III I I CCCT l/6> -fegl. Contott.rs . \>Ttiral Spctio n from O" to H u s li L n Ke VeftfitiL Scale tionbJe (fie //orixontul * ISicn S.Co,lia, Qrowu by GTIiompHOn S.GEOLGOICAL oLIRVr-'i LAI-LE BoMNE-.OLLE PL 773 VIEW FROM THE EAST ^^^^te.^ -lulius Ripn A Cu.i.tJ, wn by C.Tli..inpson and W H.Holtn U S-GEOLOOrcAL SUPirEY LAKE BONNEVILLE PLXSH MAP OF SHORE TERRACES NeaT Dove Ci'cpk, Ulnh Bv Gilbert Thoinpsim SCALE ° 1000 I — 2U00 '•^'••■■^'^\ /O/eel t'imlniirs Boitri^, , II f '■ N\\\SvVVAvy- VIEW FFfOM Th'e" s'ouf H EAS'f'" " ■ - , , ^ ^> \\^\>.v\%<- liliurt tlii-n A Co. Ml, Drawn bv li Tli.>iri ATTEMPTS AT CORRELATION. 139 Having thus assembled the data, let us now endeavor to obtain a clear conception of the questions to be answered by their comparison. At the Grantsville locality the shore di-ift is built into a small number of large, definite, individual embankments, differing in height. The analogy of the Bonneville and Provo shores suggests the hypothesis that each of these em- bankments was produced by, and therefore represents, a prolonged mainte- nance of the water surface at a corresponding height. Under this hypothesis there should have been accmuulated at each of the other localities during this time a corresponding embankment; and if all the embankments remain undisturbed in their original position, a complete correlation should readily be made out. For each of the principal embankments at Grantsville there should be found a representative at the same height in each of the other localities. If such correspondence is not found, it is necessary either to abandon the hypothesis, or else to supplement it by the assumption that the relations of the embankments were deranged by differential movements of the earth's crust occurring during the general period of their formation. Examining now another locality, as, for example, the Wellsville, Fig. 8, we find that, although it exhibits a small number of large individual em- bankments, the altitudes of these do not correspond each to each with the altitudes of the Grantsville embankments. However the comparison is made this disparity appears. In the plate the Bonneville horizon is assumed as the common zero for the vertical elements of the profiles. This assumption is purely arbitrary, and Avas not adhered to in making the comparisons. In order to test the matter fully, each group of embankments was represented on a sheet of transparent paper by a system of parallel lines whose intervals were drawn to a scale, so as to agree with the vertical intervals of the em- bankments. These transparent sheets were then superposed in pairs and other combinations, and were tentatively adjusted in numerous ways, in the hope of discovering occult correspondences. Only one element of order was discovered. A horizon from 15 to 25 feet below the Bonneville (marked a on the plate) is discernible in eight of the ten localities. With this single exception, there are no correspondences which can not be referred to fortuitous coincidence. Not only is the series of altitudes different at each locality, but the number of embankments varies 140 LAKE BONNEVILLE. from place to place. It is evident, therefore, that the hypothesis of persistent water stages is tenable only with the addition of a h}"]iothesis of contempf)- raneoxis dis})lacement; and the question arises whether we have any means of subjecting- this phase of it to test. HYPOTHESIS OF DIFFERENTIAL DISPLACEMENT. The supplementary hypothesis is not a priori a violent one. As will be set forth in a following chapter, our investigation has fully demonstrated tlmt the Bonneville shore-line is no longer of equal altitude at all points, but varies within the region comprising these localities through a range of more than 100 feet. The same has been shown with reference to the Provo shore-line; and it has also been shown that a part of the Bonneville derangement oc- curred before the Provo epoch. In the series of localities represented by the profiles, the interval between the Bonneville and Provo shore-lines ranges from 345 feet to 400 feet, exhibiting a difference of 55 feet. It is therefore easy to believe that the localities may have undergone relative displacement after the construction of certain of the Intennediate embankments and prior to the construction of others, or even that local changes of water level may have been thus occasioned at one locality while the process of shore forma- tion was continuous at another. The possibility of confusion thus intro- duced seems at first unlimited, and a rigorous test of the hjqiothesis would be difficult were it not for a fortunate circumstance. The .surveyed locali- ties include several pairs, the members of which are so closely associated geographically that there is a strong presumption against their ha\'ing been affected discordantly by contemporaneous earth movements. The middle and southern localities of Preuss Valley, Figs. 1 and 2, are but two miles apart, and bear the same relation to the adjacent mountain rnnge. Tlie localities of the Old River Bed, Figs. 4 and 5, are five miles apart, and those of Tooele Valley, Figs. 6 and 7, about ten miles apart. The principal recent displacements of the basin have been of the nature of broad, gentle undidations, not aft'ecting the horizontality of the shore-lines, so far as that is distinguishable by the eye. The region including each gi'ou]) of localities may properly be assumed to have risen or fallen in con- sequence of such earth movements without important internal change; ATTEMPTS TO EXPLAIN DISCOKDANCB. 141 and this circumstance leads us to anticipate that the members of each of these groups of embankment localities will be found to correspond with each other better than with the members of other groups or with isolated locali- ties. This expectation is realized in the relation of the Bonneville and Provo sliores. In each of the two Preuss Valley localities tlie Bonneville-Provo interval is 345 feet. At the two localities of the Ohl River Bed it is 400 feet and 398 feet. At the two localities of Tooele Valley it is 375 feet and 378 feet. At the Point of the mountain, 20 miles east of Tooele Valley, it is 375 feet. When, however, the Intermediate shores are considered, no cor- relation is found. The harmonious relations exhibited by the Bonneville and Provo shore- lines at contiguous localities confirm the postulate that a general correlation should be possible in these localities, desjiite the influence of contempora- neous displacement, and compels us to reject displacement as a sufficient explanation of the discordance of the Intermediate shore-lines. By these considerations, and by others which it is unnecessary to de- tail, the writer was led to abandon the hypothesis of persistent water stages, even though a better was not immediately suggested. Eventually another was found, and this is believed to give a satisfactory explanation of the phe- nomena. It may be called the hypothesis of an oscillating water surface. HYPOTHESIS OF OSCILLATING WATER SURFACE. In order to set foi-th this hypothesis, it will be necessary to recur to the general theory of the construction of shore emljankments, page 46, and imagine how the process would be modified by the contemporane- ous oscillation of the water surface. Let us select some point of the coast where the local conditions determine the deposition of shore drift, and assume that a spit has been formed, its crest being slightly higher than the surface of the water when still. Suppose now that the height of the water surface is gradually increased. A portion or the whole of the shore drift contriljuted by the next stonn is deposited upon the top of the embankment, tending to restore the profile to its normal relation with the still-water level. During this restoration the growth of the end of the spit is retarded, or per- 142 LAKE BONNEVILLE. lia])s altogether checked. If the general rise of the water is very slow, the construction of the embankment keeps pace with it, and the crest maintains its nonnal height, but if the rise of water is more rapid, the spit is sooner or later submerged, so that the stonn waves sweep over it. Witli a slight sub- mergence, the course of the shore cuiTent is unchanged, and the waves still break as they reach the line of the spit, so that the conditions of littoral transportation are not there abrogated. A portion of the force of the waves is expended on the land inside the spit, but the shore di-ift is not diverted or divided so long as the position of the shore current remains luichanged. The growth of the spit therefore continues in its submerged condition, and if the water level ceases to rise, the crest of the spit eventually emerges and acquires its normal height. Assume now that the rise of the lake surface, being more rapid than the growth of the spit, does not cease, but continues indefinitely. A time must sooner or later be reached when the depth of water on the submerged spit permits the waves to pass over it ahnost unimpeded, and at the same time penults the shore cm-rent to be deflected inward. The formation of a new spit then begins in a position higher on the sloping side of the basin. Now let the tendency of the water level be reversed, so that it gradu- ally falls. Additions will continue to be made to the new spit by the ac- cumulation of shore di-ift on its weather face and at its end ; but sooner or later the water will reach a stage at which the shore current will be de- flected by the lower-lying spit, and at Avhicli the waves in sweeping over that spit will be broken and diminished in force. Additions to the upper spit will then cease, and the growth of the lower spit will be renewed. If this theory is well founded, there should be produced at the margin of an oscillating lake a series of embankments separated by vertical inter- vals bearing some relation to the magnitude of the waves, and each of these should grow in height every time the oscillating water surface passes its horizon, either in ascending or in descending. The rate of growth would naturally be diff'erent at different points on the margin of the lake ; and the interval between embankments, being a function of wave magnitude, should vary in different regions, being greatest where cu'cumstances are most favor- able for the development of waves. THEORY OF OSCILLATING WATER SURFACE. 143 This relation between the embankment interval and the local conditions affecting wave" magnitude is so evident a consequence of the theory that it may be used to test its applicability to the problem in question, and this may be further tested by considering the phenomena of littoral excavation in connection with those of littoral construction. The conditions which theoretically produce a rhythm in the process of littoral deposition have no similar effect upon the concomitant erosion. In the regions of littoral erosion, the shore currents are not deflected by circumstances associated with the rise and fall of the water level, and the zone subjected to the beating of the waves bears always the same relation to the still water level. An equable rise of the water should therefore pare away the coast in an equable manner; and upon the theory of rhythmic deposition, the Inter- mediate embankments should not be associated with sea-cliff's and cut- ten-aces of comparable magnitude. Proceeding now to the application of the hypothesis to the problem in question, we may premise that the water level has twice risen above the Provo horizon and afterward descended, one rise extending to the Bonne- ville shore-line and the other being nearly as great. The space occupied by the Intermediate embankments has thus been subjected to wave action at least fom- times. These oscillations have been demonstrated by independ- ent evidence; and it is pi'obable that there were also numerous minor oscil- lations. The conditions were therefore favorable for the production of the rhj'tluuic result. The vertical interspaces between the Intermediate embankments yield evidence confirmatory of the hypothesis. Six of the localities represented in the profiles and maps are suitable for comparison. Among these the local conditions indicate the greatest waves at Grantsville and Dove Creek, and at these points the average interspaces between the principal embankments are 72 feet and 75 feet. The conditions are less favorable at Wellsville and the Snowplow, but it is doubtful which of these two localities should rank next At Wells\alle the average interspace is 60 feet. At the Snowplow it is either 71 feet or 61 feet, according as an embankment of doubtful rank is included or excluded. In Preuss Valley, where there was comparatively small scope for the formation of waves, the average interspace is 53 feet. 144 LAKE BONNEVILLE. Eqiuvlly liariiionious i« tlio evidence from the iilieiioniena of littoral excavation. Take, for example, the Siiowplow. The material there aggre- gated was derived from a broad alluvial slope, partly represented in the northern portion of tlie map (PI. XIX). In this region there is a nearly continuous slope from the Provo terrace to the Bonneville terrace; and above the Boinieville cliff there is a continuous slope of undisturbed allu- vium. This latter originally extended over the entire slope, including and beyond the Provo horizon, and it can be restored in imagination so as to realize the magnitude of the excavation. From ten to thirty feet ai)}Xiar to have been removed from the general surface, and this so evenly that there are only one or two points where the presence of sea-cliffs can be indicated; and even these -can not readily be traced to corresponding embankments. The same is true in i\ general way of all localities. Not oidy are the In- termediate embankments nowhere connected \vith a s)'stem c)f differentiated cliffs and ten-aces, but it lias been found impossible, (wherever the attemjjt has lieen made,) to trace their horizons fairly into the region of excavation. At the Snowplow locality, the excavated alluvium is of such nature as to be easily modified by the rain and it does not preserve the minor details of the configixration im])ressed on it by the waves; but elsewhere, on alluvial slopes of coarser material, the inters])ace between the Bonneville and Provo cut-terraces has been observed to be occupied by a continuous s}'stem of naiTow terraces and cliffs, constituting a sort of horizontal striation of the surface. At one point, near Pilot Peak, thirty-three separate teiTaces were counted, the average interspace being less than ten feet. The liy})othesis receives additional support from the structui-e of the individual embankments. The spit built by the waves of a lake with a con- stant level should normally have a certain simplicity of structm'e, the prin- cipal additions to its mass being made at the distal end, and the deposits near the crest having no irregularity, except that referable to the disjjarity, in force and dij'ection, of the constructive storms. A spit consti'ucted by the waves of an oscillating water surface should theoretically be begun at a relatively low level and receive additions in the form of superposed spits of various altitudes and lengths, some extending to the end of the mole and others sto2)ping short. The compound structure is characteristic of the ACCESSORY EVIDENCE. 145 Intermediate embankments. Sectional exposures are indeed rarely to be seen; but from many of the embankments there project, either at the distal extremity or on the shoreward side, shelves or spurs indicating the horizons of the lower wave work and testif}'ing to the composite structure of the mass. Fig. 25 gives an illustration of this, observed near Willow S2)ring, west of the Great Salt Lake Desert. A broad spit is characterized by a hook at its extremity. A study of its details shows that the shore di-ift, under the \WM Fig. 25. — CompouDd Hook of an lutermediate Shore-line near Willow Spring, Great Salt Lake Desert. influence of the dominant waves, here from the north and northeast, traveled from a to h. By less powerful waves from the east and south it was then carried about the end of the embankment to the recurved point c, a point with a peculiar and notable outline. On the lee side of the spit, at a point where the Avaves could have no force after its construction, there are tliree projecting tongues d, e, f, built of beach-rolled gravel and closely resembling the extremity of the point c. The highest is twenty feet below the spit; the others thirty and forty feet. They are evidently more ancient hooks, the MON I 10 146 LAKE BONNEVILLE. appendages of similar but shorter and lower spits, which may fitly be re- garded as progi'essive stages of the huge table ultimately constructed. Finally, the single element of order detected in the accumulated pro- files is by this hypothesis shown to be consistent with the general want of order. The terrace (a, PI. XXIII) lying from 15 to 25 feet below the high- est Bomieville embankment, was preserved because it was the penultimate deposit of the ascending series, and because the ultimate deposit was too meager to mask it. The differentiated series of Bonneville bars described in a preceding section shows that the penultimate water stage was about 20 feet below the ultimate. Wlierever the penultimate contribution to an em- bankment w^as made upon its lakeward face, it escaped concealment by the final contribution, which was small in amount and was perched u2)on the top of the same embankment. The second hypothesis is thus sustained at all points. The Intermedi- ate embankments record the wave action of an oscillating water surface. Within this zone the water level did not long linger at any one horizon, or if it did, the record of that lingering was effaced by later action. It follows as a corollary from this discussion that cut-ten-aces with their associated sea-cliffs afford a more trustworthy record of persistent water stages than do embankments. It is an additional mark of persistent stages that they afford coordinated terraces and embankments. It is impoi'tant to note, however, that neither the sea-cliff nor the cut terrace, if observed alone, affords satisfactory evidence of persistent wave action at one horizon. They must be found together. A slowly rising tide continually abandons the freshly cut teiTace and attacks with its waves the freshly cut cliff above it. In this way a cliff is carried before the ad- vancing water of an oscillating lake ; and when the ma.ximum is reached and recession follows, the cliff is stranded, so to speak, at the upper limit, even though the water margin was retained there a short time only. Sim- ilarly, it is conceivable that a falling lake sm-face may carry before it a cut terrace without leaving at any horizon a sea-cliff of comparable magnitude. The first of these conclusions has an application in the case of the Bonne- ville shore-line, which, as already remarked, is characterized by the great height of its sea-cliffs, but is inferior to the Provo shore-line in the widtli CIIARACTBRS GIVEN BY STABLE WATER LEVEL. 147 of its cut terraces. The considerations here adduced serve to complement the ])artial explanation of this contrast advanced on page 129. As already intimated, the compilation of the Intermediate embankments was the result of a series of oscillations of the ancient lake, whereby a zone of wave action was carried alternately upward and downward over the slojie. The basis for this statement does not lie in the embankments them- selves so much as in the associated lacustrine and alluvial deposits. It is imquestionably true that the entire history of oscillation is embodied in the internal structures of the embankments, but these are not exjjosed for exam- ination, and the external forms afford information for the most part only of the Litest additions. It is a curious fact that these forms of embankments appear to have been moulded by a gradually rising rather than by a falling tide. The last general movement of the water was of course a recession, for the slopes are now dry, but that recession has left so little trace above the Provo horizon that we are led to believe it was far more rapid than the preceding advance. This conclusion is as interesting as it was unexpected ; and it is proper that the evidence on which it rests be presented somewhat fully, especially as it has been assumed by several investigators, including myself, that the several shore marks of the series represent lingerings of the ancient lake during a gradual recession. SUPERPOSITION OF EMBANKMENTS. The snowpiow.- Ill tlio first placc, there are many superficial indications of the overlapping of low embankments by high ones. If the reader will turn to the map of the Snowplow (PI. XIX), he will see that the table lettered a is not entirely supported by the table b, but projects a little on the south side so as to rest partly upon the general slope which is the common founda- tion of both. (It is necessary to restore in imagination the contours inter- rupted by the (h-ainage line southeast of the letter a and dividing the embankment it indicates.) As has already been explained, the material of the Snowplow Avas derived from the region fff, and was di-ifted along the shore from southeast to northwest. That which composes the upper surface of each embankment must have been carried along the southern edge of the 148 LAKE BONNEVILLE. Snowplow by beach action, so that each embankment was, at the time of its completion, connected by a continuous beach with the source of supply. The embajikment h is not so connected, for the evident reason that its southern edge has been overlapped b}' the latest addition to embankment a. If the waves during the recession of the water had made a contribution to the lower embankment, they must either have excavated the side of the upper embank- ment or else have built a platform around it, and in either case the slope from the crest of the upper to the foundation plain would not have the observed uniformity and steepness. A similar relation of parts shows that the em- bankment 1) was completed after the embankment c, so that at least three of the members of the series received their final moulding in ascending order. Reservoir Butte.-At Rcscrvoir Butto Substantially the same story is told, but in different language. The face of the butte turned toward the open lake was rugged in the extreme, and the configuration of the neighboring bot- tom was irregular, so that, as the depth of the water changed, the conditions determining the transfer of shore drift and the construction of em];)ankments were continually modified. The resulting embankments were not built into a synnnetric system but Avere thrown together in an irregular and unique group. By referring to Pis. XXIV and XXV, where they are represented by vertical and horizontal sketches,^ it will ,be seen that, of those above the Provo, the highest is tlic last formed, overlapping all the others. Number 2 (tli^y are numbered in the order of height) has no visible connection by beach with the north or weather face of the butte, whence its material was derived; and its form and relations show that it could not have been con- structed after the completion of Number 1. The third and part of the fourth are in a similar manner overplaced by the second, and were evidently earlier fomied. The fourth is however separable into two parts, which may have been formed at different times; and the outer, marked 4« in the diagram, is not so related to No. 2 as to demonstrate the order of sequence. It is hoAvever overplaced by No. 1. The relative age of the third and fourth is not appar- ent; but tho fifth, which lies; in a bay completely sheltered by the fourth, is evidently of greater age. The sixth and eighth have no detenniued relation ' Tho plat of tlicso eraljankiucnts Riven in PI. XXV cinnot claim the accuracy of other maps of embankments. It was sketched in the field without the aid of instruments, and may be very inaccu- rate in matters unessential to the discussion above. SUPERPOSITION OF EMBANKMENTS. 149 to any other except the first, which they underlie; and the seventh, which projects from beneath the fourth, shows no direct relation with any other. The ninth is the Provo, and this proclaims its recency l)y its relation to the first. Its table extends to the north face of the butte, and not merely passes the face of the first or Bonneville emljankment but is in part carved from it. The Provo waves encroached also upon the eighth embankment. These relations may be tabulated in the following form, in which the word "ante" should be construed to mean completed at an earlier date than. ^ \ ante 2 5 ante 4 \ - > ante 4a • ante 1 ante 9 stockton.-Another unique aggregate of embankments is equally instruct- ive. Previous to the rise of the lake, the drainage of Rush Valley was tribu- tary to that of Tooele Valley, the connecting parts having a continuous descent from south to north, and an ample channel, of which a portion is yet clearly to be seen. At the point of greatest constriction between the two valleys, where the Bonneville strait had a width of only 8,000 feet, the bot- tom of the channel ran about 350 feet below the level afterward marked by the Bonneville shore. At all high stages of the lake the strait received a large quantity of shore diift from the northeast, and a series of curved bars were thrown across it. These bars have a total width of 5,000 feet, and partially overlap each other, so as to constitute a single earthwork of colossal propor- tions. Whenever the water surface fell below the highest completed bar, the Rush Valley bay was completely severed from the main body, and became a lake by itself This lake was so small that its waves were comparatively powerless; and, although traces of their work can be discovered, they did not materially influence the configuration of the earthwork. The locality is exhibited in the foreground of the view in PI. IX and in the map and 2;)ro- file of PI. XX. If the reader will refer to the latter plate and give attention to the profile in connection with the map, he will see that the bars rise in consecutive order from a to g, and that each has a curved axis with concavity toward the north. This curvature, which is characteristic of bay bars in gen- 150 LAKE BONNEVILLE. eral, shows that the waves concerned in their production came from the north. It is evident that after the bar b was constructed, the bar a was protected from all further wave action, a was therefore completed before b was built; and in general the order of construction could not have been other than the order of the letters, — the lowest bar a being the first, and the highest bar g, the last. The order of construction was therefore from low to high. It is to be noted that this order is demonstrated only for the visible or superficial portions of the earthwork. There may be beneath the bar ff, for example, a deeply buried series of bars lower than a, and either younger or older; and so of any other of the higher bars. We have no reason to believe that the whole history is embodied in the visible phenomena. The bar g diff"ers from the others in that it is not unifoiTQ in height thi'oughout its length. The lowest point of its crest is approximately in the position occupied by the letter; and from this there is an ascent of about 30 feet toward either shore. At the Bonneville stage the strait was not closed by a bar, but the shore drift was built into spits. That at the west is short and has the fonn of a hook. It is crested from end to end by a slender ridge, built at the cuhninating water stage. The eastern is straight and broad and 6,000 feet in length. Its proximal end bears two small spits, referable to the cuhninating stage of the water; and its distal end evidently overlajis the lower members of the compound earthwork. So far as outward api)earance goes, this is purely the product of shore action at the Bonneville stage ; but it is possible that similar spits were formed at lower stages, so as to consti- tute a foundation for the Bonneville spit. One of the most striking features of the series of bars is the paucity of wave marks upon the northern face. There is a diminutive bar, character- ized by an abundance of tufa, imposed on the face of the gi-eat bar g four feet below its crest ; and twelve feet lower a wave-cut terrace is barely per- ceptible. These may record an oscillation of the water after the comple- tion of the great bar and before it rose to the Bonneville shore ; or they may have been produced by the receding water after the highest level had been touched. In any event, the final recession must have brought every foot of the northern slope of the earthwork within reach of the waves, and SEQUENCE OF BARS AT STOCKTON. 151 the surviving continuity of the slojie testifies to the rapidity of the reces- sion. The conditions for wave work were unchanged. The alluvial slopes which had furnished the gravel for the several embankments, still offered an inexhaustible supply, and the same currents and waves must have been set in motion by the storm winds ; but the lake seems not to have tarried long enough at any one level to add a terrace to the structure. Another evidence of the rapidity of the final descent of the waters is found in the fixilure of the waves at any of the Intermediate horizons to un- dercut the embankments constructed at the higher stages. If the water tan-ies long at one level, the changes it effects in the form of the shore finally modify the currents so as to shift slowly the districts of erosion and of construction. Spots that were at first excavated are afterward made to receive deposits, and portions of the original deposit are afterward removed. Instances are known in which the Provo waves have pushed their excava- tion to the heart of the Intermediate embankments, so as to undercut even the highest members ; and there are few localities of great wave action which do not exhibit more or less encroachment ; but there is no evidence that the waves of any Intermediate stage have seriously impaired any higher embanlanent. There is a narrow wave-cut terrace on the north face of the Stockton earthwork ; two lines are engraved on the points of Intermediate terraces in the Snowplow; and there is possibly a similar occurrence in Preuss Valley ; but no locality gives evidence of long-continued action. Blacksmith Fork.-Thc uudercuttiug of the Provo shore has in two places exposed instructive sections of the Intermediate embankments. At the south end of Cache Valley, close to the point where Blacksmith Fork issues from the mountain, there is a section, nearly 300 feet in height, show- ing a face of clean gravel, which has slidden down so as to cover the entire surface — if, indeed, it does not constitute the entire mass. At four horizons this is barred across by level lines of cemented gravel marking successive positions of the upper surface of the mass as it was piled. Dove Creek.- A siuiilar cscarpment of gravel is exposed on the soiith face of the Dove Creek group of embankments (see profile diagram on PI. XXII.), and a similar series of parallel lines can be traced across it. They are best seen from a distance, and on close examination prove to consist 152 LAKE BONNEVILLE. merely of a scattering growth of bushes. There is no visible variation in the character of the gravel, but the position of tlie bushes is doubtless di.'- termined by the existence beneath the surface of relatively impervious strata. Whatever the nature of these strata, they are elements of structnr(i iind demonstrate the growth of the series of embankments from the base iipward. The featm-e especially interesting is the relation of the section to the unim- paired eastern face of the embankment group. Each line ttf division is the continuation of the iipper plane of a terrace, so that the terraces are shown to be units of stratiiication. The evidence from external foiTu is thus con- nected with that from internal structure ; and the general conclusion in regard to the succession of the Intermediate terraces is strengthened. Here, as in the other localities mentioned, it is necessary to guard against the impression that the entire history of the lake during the forma- tion of the Intermediate shore-lines is revealed l)y what can be seen of the embankments. These structure lines do not extend through the entire mass, and no other lines replace them. Those portions of the general mass of detritus which lie next to the original hill slope may have been accumulated by rising or falling waters, or, for aught we know, by a surface subjected to many oscillations. In the case of the Snowplow, all that we can predi- cate is that the latest additions to the mass were made in ascending order. With reference to the Stockton eartliAvork, we know only that, of a certain series of visible bar crests, the order of height is also the order of date. It is not only possible but even probable that the series is discontinuous, hav- ing been interrupted by epochs when the water was too low to add to the accumulations at this point. Double Series in Preuss Valley.-But, Avllilc it WOuld liave bcCU iuipOSsiblc tO gaiu a knowledge of the repetitive movements of the lake surface from shore phenomena alone, they nevertheless serve to supplement the information afforded by the lake sediments. Having learned from the sediments, as will be explained in anotlier place, that the wati'r rose at least Uvice from the lower to the higher parts of the basin, besides undergoing many, minor oscillations, it was not difficult to see that certain of the shore embankments were referable to an earlier flood than certain others. The most important locality is illustrated by the map and sketch of PI. XVI, and shows a series DOUBLE RECORD IN PREUSS VALLEY. 153 of curved bars (h hh hj, overlapped by a series of spit-like embankments massed together into a few sloping terraces (t t t). The source of the shore drift was at the north, and the beaches which conveyed it to tlio curved bars are hidden b)' the later erabaidcments. It would be impossible for the bars to originate under the lee of the spits. Moreover, the spits everywhere exhibit their gravelly constitution, but the curved bars are half buried by lake deposits. DELTAS. The earliest allusion to the deltas of the ancient lake is by Bradley, who remarks tliat the lake terraces " are much more numerous near tlie mouths of the streams, where the stream-currents have distril)uted their sediment, when the lake waters were at these higher levels";^ but the first clear discrimination of the deltas from other terraces was by Howell, whose observations were made only a few months later. Speaking of the horizon of the Provo shoi'e-line, he says : — "When the old lake stood at this level, the detritus brought down liy the Provo River formed a delta, cov- ering at least twenty thousand acres. Another delta was formed at this time at the mouth of Spanish Fork Canyon, in the same valley, which covered an area of eight thousand or ten thousand acres."" It was the magnitude of the former of these deltas that led Plowell to suggest the application of the local name Provo to the shore-line at that level. It is now known not only that all of the more notable deltas of the basin appertain to the same shore-line, but that the delta built by each stream at that level equals or exceeds in mass the aggregate of its deltas at all other levels. At higher levels such accumulations are exceedingly rare ; and at lower they appear to have derived their material largely from the partial destruction of the Provo deltas. In attempting to translate these facts into terms of geologic history, the first impression is that the lake surface was held at the Provo level during more than half the period of its existence, but a fuller consideration shows that this conclusion is not warranted. The degradation of the uplands and the offscouring of the rivers are doubtless sufficiently uniform in rate to 'Frank H. Bradley: Geol. Surv. of Terr., Ann. Rept. 1872, p. 192. 2 Edwin E. Howell : Geol. Surv. West of 100th Meridian, vol. 3, p. 250. 154 LAKE BONNEVILLE. afford the basis for a time scale, but there are important modifying condi- tions given by tlie relations of the oscillating lake surface to the configura- tion of the stream valleys. In the discussion of shore processes, it was pointed out that the detritus brought to a lake by a small stream is absorbed by the shore drift, while that brought by a large one overwhehns the shore di-ift and records its acces- sion by a delta. The codeterminauts are, on the one hand, the magnitude of the lake and the consequent force of the waves, and on the other, the volume of the stream's load of detritus. In the case of Lake Bonneville, the number of streams competent to project deltas from the shores of the open lake or of the larger bays, was small; and it is believed that all of their ancient mouths have been examined. With very few exceptions, they enter the lake basin through mountain gorges so deeply eroded before the lake epoch that the rising water set back into them, forming naiTOw estua- ries. Knowing as we do from the study of the Intermediate shore embank- ments that the water rose slowly as it apjijroached the highest level, we can not doubt tliat the stream di-ift was contemporaneously accumulated into a series of deltas within the mountain gorges. Afterward, when the water fell rapidly to the Provo level and there rested, the streams attacked the deltas in the defiles and carried their substance farther lakeward to form new structiires. These new structures began for the most part within the walls of the defiles, and were progressively built outward until they pro- truded into the open lake, where space permitted them to develop into typ- ical fan-shaped deltas. 'The material furnished by the older deltas in the defiles was close at hand, and in a condition peculiarly favorable for removal. Not only was it uncemented, but it was confined to the very courses of the streams, so that it could not escape their action. It must have been rolled to its new position hi an exceedingly short time ; and we need not be surprised that the traces of its original forms are nearl}- ol)literated. The rapidity with wliich delta alluvia are torn up and carried away bv run- ning water finds al)undant illustration at the present time in the irrigation districts of Utah. Wherever the water of a canal breaks through its bank, or is neglected and suifered to discharge unguided down a delta slope, it quickly erodes a canada of formidable proportions. DELTAS OF LAKE BONNEVILLE. 155 Deltas associated with the Provo shore are thus composed not merely of tlie contemporaneous outscour of the catcliment basins of their several streams, but of the detritus antecedently accumulated in the estuaries dur- ing the higher water stages ; and, so far as they afford a time ratio, they represent the entire period during which the water stood at and above the Provo horizon. There are, however, a few exceptional localities where the Bonneville estuaries were so small and shallow that the stream drift not merely filled them but threw out semicircular capes into the Bonneville lake ; and in such cases it is possible to make a comparison between the magnitude of the structures pertaining severally to the Bonneville and Provo epochs. American Fork Deita.-Tlie best locaHty for such obscrvation is on a tributary of Utah Lake known as the American Fork, and this was carefully exam- ined for me by Mr. Russell. The Bonneville delta there displayed has a radius of nearly 5,000 feet, and a height at its outer margin of 120 feet. It is bisected by the creek, and is thus cut nearly or quite to its base. The walls of the channel exhibit a section of the deposit, showing it to consist chiefly of rounded gravel, with some intermingled sand. The gravel, being uncemented, will not hold an escarpment, but flows down in the fonn of a talus wherever it is excavated by the stream, thus masking the greater part of the stiixcture. There is, however, some indication of horizontal bedding. The outer margin of the terrace is fortunately more communicative. Around three-fourths of its periphery there runs a narrow shelf half-way down the steep face ; and the details of this shelf show that it is the protruding edge of an older and lower delta terrace, furnishing the foundation for the upper. At some points lake beds were found intercalated with the alluvial gravels, but they appear to be local deposits and not continuous sheets traversing the whole body. The most complete local section has been intro- duced into the accompanying diagram, and presents the following sequence : 6. Well rounded gravel, forming the top of the upper terrace; 20 feet. 5. Lake beds; laminated clays with .dmnicoJa; 30 feet. 4. Well rounded gravel ; 15 feet. 3. Well rounded gravel cemented at the top by calcareous tufa ; constituting a bench on the face of the terrace ; 20 feet. 2. Well rounded gravel ; constituting locally a distinct bench ; 25 feet. 1. Lake beds, to foot of slope, 10 feet. 156 LAKE BONNEVILLE. The continuity of the gravels 3 and G throughout the whole mass is shown by their relations to tlie topography. Each marks a water stage during which a broad delta was built in the lake. The beds numbered 2 and 4 are identical in charactcu-, and may be salients of similar deltas, here locally brought to light and elsewliere comjdetely buried; or they may be merely local masses of alluvium, marking the i)Ositions held ]jy the creek during temporary fluctuations of the lake level. At another point of the profile, a less complete section was observed, exhibiting a rapid alternation of gravels and clays in the lower part of the mass, and at a few other points short tongues of gravel were seen to project from the table at various levels. Flu. :i*i.— Ucuoralizud soctiou or Deltas at the mouth of American Fork Canyon, Utah. By I. C. Russell. Uonzontal scale, 4,500 feet = 1 inch. Vertical scale, 300 feet = 1 inch. These indications of complexity of structure accord well with such con- ceptions of the oscillation of the lake at these stages as we have derived from the phenomena of the Intermediate embankments. If its surface was incon- stant, rising and falling, like the surface of Great Salt Lake, with an irregu- lar rhythm, all processes of deposition at the mouth of a stream would be successively interrapted, and any detailed section should show e^adence of alternation. A rising tide would induce the formation of a delta far up the slope and give opportunity for the accumulation of lake beds farther do\\n. A falling tide would cause the stream to deepen its channel by the partial erosion of the incipient delta, and perhaps of lake beds also,' and would cause a local deposit of gravel at some lower level. A reascent would re- OLD DELTAS OF AMERICAN FORK. 157 pair the breach in tlie delta, and a redescent might conduct the stream drift in some new direction. The same oscillations would carry the waves to all parts of the surface and enable them to work over the detritus, adding their tribute to the general confusion. Assuming that the water did actually oscillate to and fro during the compilation of the delta, it is manifestly impossible to trace in detail and in true sequence the processes which make up its history. The most that can be affirmed is that a definite stage is marked at the liorizon of Bed No. 3, where the water stood long enough to complete a well developed delta ter- race, and that a similar definite stage is marked by Bed No. 6, which is a continuous delta sheet almost coincident in area with the one below. The lake level rejiresented by this higl ;st delta falls within the range to which the Bonneville shore-line pertains, but was not the absolute maximum. It is probable that the latter is represented by a shoal-water bar which crosses the south part of the delta with a crest about 20 feet higher than the delta margin. The locality thus exhibits at least three ancient deltas, of which the order of position is : — Bonneville delta; capped by Bed No. 6. Intel-mediate delta; capped by Bed No. 3. Provo delta. In the order of time the Intermediate comes first and the Provo last. The Intei-mediate was built; the Bonneville was spread over its back, but failed to cover it coniijletely; the lake fell, and the two were eroded by the creek, the Provo being formed at the same time. Finally tlie Provo shore also Avas abandoned by the lake water, which receded to its i)resent position in Utah Lake. The creek has opened a broad passage through the Provo delta, cutting it at the outer margin to its base, and is engaged in building a modern delta in the modern lake. The apex of this delta lies within the channel through the Provo delta, and is continuous with the flood plain of the upper course of the stx-eam. The modern stream Ijed has a more rapid fall than the ancient, as will be seen by comparing the profiles of the modern flood plain and the Provo delta, as exhibited in the diagram. This is due chiefly to the lowering of 158 LAKE BONNEVILLE. the stream's mouth; but it is also due in part to the elevation of its point of issue from the mountain. A recent fault has lifted the movmtaiii witli refer- ence to the valley through a space of 70 feet. Tliere is perhaps no locality more favorable than this for the estima- tion of the time ratios of the higher lake levels, but even here it is f;n- from satisfactory. The Provo delta of American Fork coalesces with the con- temporaneous delta formed by the next creek to the north in such way that it is impracticable to di-aw a line of separation; and there is no record of the tribute made by American Fork during the rising of the lake until it reached a level barely 100 feet below the Boimeville. Nevertheless, it is instructive to make such comparisons as the circumstances })ermit, and Mr. Russell's field notes have enabled him to compute approximately the vol- umes of alluviiun accumulated at the different levels. MillioDH of cubic yards Voluiuo of Bonneville aud Intermediate deltas before erosion by the creek 330 Volume of alluvium conteniporaueously deposited in mouth of bed-rock canyon 5 Total volume of gravel furnished by American Fork while the lake leA-el was within 100 feet of the highest stage 335 Volume of Provo delta of American Fork (the separation from delta of Dry Creek being arbi- trarily made) - 400 Deduct gravel derived from Bonneville and Intermediate deltas 28 Deduct gravel derived from mouth of bed-rock canyon 5 Total volume of gravel furnished by American Fork while the lake stood at the Provo level »67 If these quantities were well ascertained, instead of being rudely esti- mated, they would show the gravel tribute of the stream to have been sliglith- greater during the Provo epoch than during the last 100 feet of the antece- dent rising, and would warrant the inference that the time during which the lake level lingered within 100 feet of its highest mark was slighth- exceeded l)y the duration of the Provo stage; and, after all allowance has been made for imj)ei'fection of data, there remains a presumption that the Provo epocli is comparable in dui-ation with the epoch or epochs recorded by the upper deltas. Mr. Russell has computed also the volume of gravel furnislied Ijy the creek after the completion of the Intermediate delta, finding it to be 153 ■ TIME RATIOS. 159 million yards. This represents the tribute of the creek for all lake changes within 50 feet of the maximum, and includes the Bonneville tribute. Its ratio to the estimated Provo tribute is as 5 to 12. It is perhaps fair to as- sume that one-half of this mass pertains to the Bonneville shore proper; and on that assumption the indicated ratio of the epochs of the Bonneville and Provo shores is as 1 to 5. Quantitatively, this estimate has not a high value, but qualitatively it serves to confirm the impression derived from the wave work of the Bonneville and Provo shores. It is worthy of note that the only halt of the lake surface which here finds record between the Provo and Bonneville horizons, was a halt of the advance and not of the retreat. The Intermediate delta is unmistakably older than the Bonneville; and there is none younger except the Provo. There was of course no cessation of stream action while the water of the lake was falling from tlie high mark to the low. The creek must have be- gun the erosion of the Bonneville delta as soon as its point of discharge was at all lowered by the recession of the lake; and the product of that erosion must have been deposited at the mouth of the creek in the form of a delta or group of deltas, but the eroded channel was so narrow and the resulting deposits were of so small bulk that later action destroyed them. Wliile the Provo delta was being built the channel through the Bonneville was enlarged nearly to its present dimensions, and no stream terrace sur- vives to mark the earlier stages of its excavation. In the same period the creek tore down and removed whatever deltas it may have built at the shore of the receding lake. If the lake had halted and lingered by the way, the creek would have been able to carve a broad flood plain and spread a broad delta, some vestiges of which would survive ; and we can legitimately infer from their absence that the recession of the lake was rapid and without in- terruption until the Provo level was reached. When the lake afterward shrank away from the Provo delta, its move- ment was less precipitate. The channel then opened by the creek has a maximum depth of only 70 feet, but five separate stream terraces, cut from its right wall, record the hesitation of the water as it fell. Logan Delta.- One of tlio most beautiful and symmetrical of all the deltas is that constructed by Logan River at the Provo stage of the lake. The 160 LAKE BONNEVILLE. river enters Cache Valley from the east, debouching from a bold mountain front through which it has eroded a narrow V-form canyon. At the mouth of the canyon the Bonneville shore-line is engraved on the rock nearly five hundi-ed feet above the river, and the grade of the river bed indicates that when the line was cut the lake water set l)ack into the nan-ow way a dis- tance of about four miles. There are are some slight traces of gravel ac- cumulations within the canyon, but it probably was only partially filled, and certainly no delta was foi-med in the lake at the Bonneville level. If any estuary existed at the Provo stage it was small and quickly filled with alluvium. The apex of the Provo delta is at the mouth of the canyon, and about this point as a center the margin describes an arc of about 130 degi'ees with a radius of 8,000 feet (see map and profile of PI. XXVI). Tlie upper surface is visibly and distinctly conical, having a radial slope in all direc- tions from the apex of 55 feet to the mile, or three-fifths of a degree from the horizontal. At the margin this gentle inclination is abruptly exchanged for a declivity of about 20 degrees. At the north the terrace joins and coa- lesces with a similar and contemporaneous but smaller terrace pertaining to what is now a small creek. The marginal height of tlie terrace is about 125 feet. During its construction the river occupied every part of its surface in turn, and when the construction work was brought to an end l)v the lower- ing of the lake, and the excavation of a channel was begun liy the ri\er, the position of that channel was determined by the chance position (tf the shifting stream. It is not medial, but bears so far to the south tliat the northern remnant of the delta is two or three times greater than the southern. As soon as the erosion of the Provo delta conunenced, the 1)uilding of a new delta Avas begun at a lower level, and the apex of the new delta was at the mouth of the channel through the Provo. With the progressiA'e low- ering of the lake, yet other and lower deltas were built, the construction of each being accompanied by the partial or complete' destruction (»!' tliose above it ; and this continued luitil the desiccation of the valley. For two miles below the Provo delta, each Ijiiiik of tlie modern river is lined by the remnants of these old deposits, four or five lying on each side. One of the most conspicuous has been selected as the site of the Logan Temple, and two lower benches are occupied by the town of Logan. A glance at the I- S. GEOLOGICAL SURVEY LAKE BONNEVILLE FLXXVI MAP OF THE D 1", I.T A S 1.11 mc-a 111 AKE nONNKVIlJ.I': bv llii- LOC.AN KIVEK Rv ^V'^ D Joliiisoii . ''WiJii'flfP"'''"'' Pro 111 e Vertit'ol Sivii- ilouOlc the Borixonin.1 ZS- feet Cyntmu's . Temple Jul,i,« Hicn .vi'o.l.th Dravni by (* Tliompiji OLD DELTAS OF LOGAN RIVER. 161 map will show their arrangement l)etter than any description. The river has developed so broad a flood plain that half their mass has disappearetl, and the dissevered remnants are too fragmentary to be readily correlated across the interval. No attempt has been made to restore their forms and com- pute their volumes, l)ut it is evident by inspection that they included no ri^•al of the great delta above. Their renmants do not exceed in total bulk the mass the river has dug from the upper terrace. They can have no value as a basis for time ratios, because it is impossible to tell how nnich they owe to the reworking of the material of the higher delta and how much to the annual tribute of gravel brought by the river from the mount- ains ; Init they serve to show, first, that the lake lingered by the way as it receded from the Provo shore, and second, that its lingerings were not long. The same lingei'ings have left record within the Provo delta in the form of stream terraces, which abound near the mouth of the canyon. Mr. Russell has recognized ten independent benches on the north side of the stream and three on the south. The view in PI. XX^^1I was sketched from the wall of the Mormon temple standing on one of the lower terraces. It exhibits the Provo delta, divided by the alluvial valley and overlooked by the Bonneville shore mark, which happens to Ik; strengthened immediately above the delta by an accumulation of shore drift. The main delta, and probably all lielow it, rest upon a sloping floor of lacustrine sand and clay. The modern bed of the river runs below the bases of the deltas and within the zone of these sediments, Ijut exposures are rare, l)y reason of tlie tendency of the uncemented delta gravel to slide down and overplace it. The best exhibition at the time of our examination was afforded by a fresh excavation for an irrigation canal along tlie bluff north of the river, and was sketched by Mr. Russell. The strata show many undulations beneath the Provo delta, but are relatively smooth be- yond its margin. Mr. Russell suggests that the disturbance of the strata may have been an incident of the building of the delta. At every stage of the work there was a diftei-ence between the weights borne 1j}' the lake beds beneath the delta and by those beyond it, and the line of sej^aratiou was sharply drawn at the edge of the deposit. The conditions were there- MON I 11 ](J2 LAKE BONNEVILLE. fore favorable for the deformation of the freshly deposited sediments by differential jjressnre, some of the softer layers being made to flow out from beneath the gravel. The difference in weight between the water on one side and the saturated gravel on the other amounted to seventy-five pounds to the square inch. As the delta was progi-essively increased by additions at the outer margin, the zone of unecjual pressure; was correspomlingly ad- vanced, until the whole substructure of the delta hail been subjected to the action and deformed as far as its constitution permitted. fnmo Vrlta, Temple Delta FLond. PtaOv yv E. Fig. 27. — Partial soctinii of Dolta.s at Lo^^aii, Utah, liy I. C. KuasoU. Vi-rtical scalo greatfr than hdrizunlal. Wherever the body of the Provo delta is freshly exposed, it displays an oblique lamination inclining in the direction of the lakeward margin- The dip near the top of the deposit is 15 or 20 degrees, and diminishes downward, the layers being disposed in sweeping, parallel curves. Only a single locality exhibited (1880) the nearly horizontal beds which in a normal delta overlie the inclined — a point half a mile below the canyon's mouth, where the south bluff of the river had freshly fallen down, exposing ninety feet at the top of the face. The series consists of: r>. Fine sand, 5 feet. 4. Gravel, horizontally laminated, 10 feet. 3. Fine sand, 'i.') feet. / 2. A line of small boulders, unconformable to No. 1. 1. Gravel, coarse and fine intermingled; dipping 15° toward the SW. Exposed 50 feet. Other Deitas.-Of tlic otlicr sti'eams of Cache Valley, as many as eight built Provo deltas, and one. Spring Creek, probably formed also a small Bonne- ville delta. The Cub Creek and High Creek deltas are small, and lie within the flaring mouths of the canyons. Smithfield and Bell ville Creeks heaped their tribute just outside the canyons. Blacksmith and Muddy Forks de- bouched close together and built a confluent delta, larger perhaps than that of the Logan, but less symmetric. The original or ante-Bonneville canyon of Blacksmith Fork was so deeply cut that tlie modem .stream lias not yet removed all the debris gathered during the lake period. The mass of allu- L' S. 0EOL(iGlCAI. Srn\ KY 1„\KK BlIN.VEVlLLK PI, XV\ I THE ANCIENT DELTAS OF LOGAN RH'ER, AS SEEN FROM THE TEMPLE. INTERNAL STRUCTURE OF LOGAN DELTAS. 16;3 viiun stored in it at the Provo epoch was great, and contributed to the t'orniation at lower levels ot" a fine series of deltas, on which stands the village of Hyruni. Spring Creek issued from a canyon which was never cut down to the Provo level, and the apex of its Provo delta was (piite outside the can- \(tn. The modern stream is a mere rivulet that one may leap across; but its delta liad a radius of two-thirds of a mile, "^llie history of the Bear River deposits was not a\ ell made out. At the canyon mouth the river now Hows at a level a few feet higher than before the lake })eriod, and tliat level is four Imndred feet below the highest lake shore; but the modern river outside the canyon is walled in by a great deposit, chiefly of sand, through which it has opened a passage. There was clearly no Bonneville delta at this point. The upper surface f)f the sand is a sloping plain, joining the mountain near the canyon only fifty feet below the Bonneville shore. Unfortunately the examination was made while snow lay on the ground, and the structure of the deposit could not be seen. If it is a delta it is probably of the Provo date, and its outer margin must be in the vicinity of Battle Creek Butte, ten miles away. Otherwise it must be regarded as a lake sediment, which owes its exceptionally great volume to the proximity of a silt-bearing river. In either case its source of material is the river drift; and in either case its ac- cumulation was probably contemporaneous with that of the deposits which filled Gentile Valley, a small opening among the mountains at the head of the canyon. Outside of Cache Valley all the notable deltas except that of the Sevier River lie at the western base of the Wasatch Range. The most northerly is near Brigham City, on Box Elder Creek, ^ a stream rising in a small valley just east of the main axis of the range, and cutting across it. In the upper valley there are remains of a detrital filling-, which was probably coe\al with Lake Bonneville, although not in visible contiimity with delta forma- tions. The canyon through the mountain has been swept clean of debris, except at the bottom ; and at its mouth there is a small composite delta, of which the highest element has the Provo height. The history of Ogden River is nearly the same, but its features are on a larger scale. The upper valley contained so large a bay that a discernible shore-line was carved therein ; and it is probable that some of its sloping ter- ' Not to be confounded with the Box Elder Creek of Tooele Valley, mentioned in connection with the Grantsville embankments. 164 LAKE BONNEVILLE. races are remnants of Bonneville deltas. The fall of the lake drained the upper valley and led to the building of a broad delta just outside the mouth of the canyon; but this delta is exce})tional to the general rule in that it is somewhat below the Provo horizon. On the plain beyond it a series of ter- races were afterwards formed similar to those at Logan. The city of Ogden stands at the end of the series, and its suburbs encroach on some of the lower benches. Close to the Ogden deltas lie those of the Weber, less synunetric l)ut far more massive. They extend from four to six miles in nil diroctions from the mouth of the canyon. The channel cut through them by the modern river is several hundred feet deep, and is exceptionally indirect, curving through the fourth part of a circle. The broad flood plain within it supports three agricultural hamlets, and is traversed by the Union Pacific liaihvay. The westward-bound passenger issuing from the rock-bound defile of the Wasatch at Uinta Station finds himself enclosed by walls of delta sand, and does not fully emerge from the lowest terraces until he reaches Ogdt'U (Sta- tion, a ride of eight miles. The greater portion of the stnicture lies on the left or south bank of the river and is locally known as the Sand Ridge. It is the largest of all the deltas of the ancient lake biiilt upon an open ])lain, but, owing to the lightness of its material, the details of its form are imper- fectly preserved. Portions of the interior of the mass ap])oar to be gravelK', but the upper parts are chiefly composed of sand, so fine as to be moved by the wind. The ])rincipal terrace is at the Provo level, and upon this there stands a liill more than 200 feet high, which niav p(issi])lv l)e the remnant of a more ancient and more lofty delta, but is probably a dune accumulated during the Provo epoch. The lower terraces, marking the I'e- cession of the water, were built on the north side. The south fai'e of tlie Provo delta has been supei-ficially modified by subsequent wave action. City Creek, the stream supplying Salt Lake City with water, rises in the Wasatch Range and flows through a long canyon before emerging on the plain. This canyon was capable of storing a large amount *)f alluvium; and it is probably due to this fact that the Provo delta is smaller than tliose at lower levels. The group of deltas constitute "the bench" on both sides of the creek, and are composed of coarse, \\ell rounded gravel. While they OTHER DELTAS. 165 were forming, a large amount of sliore drift soems to liave readied tlie lo- cality from the southeast, and this modified the resulting topograjihic. iorms. The configiu'ation of the bench owes nearly as nuich to the action of waves as to the depositiini of stream drift. The deltas formed by Little Cottonwood and Big Cottonwood Creeks coalesced with each other, and probably with one from the Dry Cotton- wood; but their outlines are greatly obscured by subsequent stream erosion, and they have been further modified by a system of faidts. P^ollowing the l)ase of the Wasatch southward, the next delta i-eached is that of American Fork, already described. Beyond it, is the delta of the Provo River, a broad low terrace of gravel spreading fan- wise from the mouth of the Provo canyon. The radius of the fun is about 4^ miles, and the terrace has a marginal height of 70 feet. It is skii'ted rather th;ni di- vided by the modern river, which turns abruptly southward from the mouth of its canyon. Lower deltas were only obscui'ely differentiated, but the form of the lake shore indicates that the river is now constructing one. The wagon road from Provo to Pleasant Grove crosses the mniii delta; the railroads pass around it. Near Provo City a small stream named Rock Creek issues from a short, steep canyon in tlie mountain. It built a small delta, during the Pxnuieville epoch, and another during the Provo; and these would afford an instructive study in chronology Avere it not for the injury they have suffered from the recent faulting. Hobl)le Creek, which irrigates the farms of Springville, built a, well-marked delta at the Provo level, and proliably a small one at the Bonneville. The subaei-ial alluvium here rests so high against the mountain that it constituted the coast at the Provo stage, and the Provo delta rests against it. Five miles southward Spanish Fork issues from the range, with a northwesterly course. In the Boimeville lake it built a delta with a radius of 4,000 feet, and in the Provo lake a larger delta coalescing with that of Hobble Creek. At Payson a small creek formed a delta at the Provo level. Salt Creek, the next stream issuing from the range, reached the ancient lake only after flowing for some distance across the j^lain. Its highest delta appears to be one at the Provo horizon, and lies at the south end of Goshen Valley. 166 LAKE BONNEVILLE. Apart from the di-ainage system of the Wasatch, only three deltas ^A-ere observed. A small one lies in an open canyon back of the town of Port- age, in Malade Valley. A larger was pro])ably foi-nied by Beaver Creek at the Provo level near George's Ranch; but it is difficult in this case to distinguish stream drift from shore drift. The deltas of the Sevier River are more important. At the Bonneville epoch alluvial terraces were built where the river enters Juab Valle}-, but the topography did not |)ermit the formation of a broad fan. At the ProA^o epoch a broad, low delta fan was built by the river on the i^lain between Lemington and Deseret. summary.-The contributious made by the phenomena of the deltas to the history of the oscillations of the lake may be summarized as follows: First, the Bonneville shore-line antedates the Provo. Second, the Provo epoch was several times longer than the Bonneville. Third, in falling from the Bonneville shore to the Provo the water lin- gered very little, if at all. Fourth, in falling from the Provo level to the bottom of the basin the water occasionally lingered, but its lingerings were brief as compared to the halt at the Provo level. Fifth, the water lingered during its advance antecedent to the Bonne- ville epoch, not standing long at one level, but oscillating up and down. A cei'tain significance attaches likewise to the absence of deltas from the greater portion of the coast of the old lake. All of the olil deltas are associated with modern streams; and all the modern streams of iiuportance built deltas. It would appear, then, that the ancient climate did not create important strc^^ams in regions where the outflow is now small. In the west- ern portion of the basin, there are catchment districts of considerable extent which furnish little or no water to the lowlands by reason of the scantiness of rainfall. If the rainfall in Bonneville times Avas very great, as compared to the modern, these catchment districts should liavc furnislicd tributarv streams; and such streams, flowing over tracts of alluviuni, the accunndation of ages, should have transported large quantities of it ti) the margin of the lake and constructed deltas of it. We seem thus to have an intimation that THE HISTORY TOLD BY DELTAS. 167 the climatic change, whatever its nature, did not affect the rainfall in a de- firree commensurate with the difference in area of lake surface. TUFA. Calcai-eous tufa was deposited by many and perhaps all of the Pleis- tocene lakes. In Lake Lahontan and the other lakes of the western portion of the Great Basin, great masses were accumulated, and their study has resulted in an important contribution to the Pleistocene history. In Lake Bonneville very little tufa was foiTned, and its bearing upon the history of the lake seems to be unimportant. It is associated exclusively with the shores; and its amount upon individual shore-lines is in a general way pro- portional to the magnitude of the other shore features. At least this rule applies to the Bonneville, Intermediate, and Provo shoi'e-lines. The Provo carries most of all; the Bonneville and Intermediate have an equable dis- tribution. Next to tlie Provo the Stansbury is most generously supjdied; Init this shore is not characterized by endiankments and cliffs of great magnitude. The extent of the lake was so greatly reduced at this stage that the i)ower of the Avaves was materially lessened; and it is ])erhaj)s legitimate to infer that the tufa records a })rotracted lingering of the falling water which does not find adequate expression in other shore features. In embankments the ])osition oecnpicd liy the tufa is on the Aveather face a few feet lower than the crest. It lies just beneath the surface, and has the function of a cement, binding the gravel together into a conglomerate. Tlie association is far from being invariable; and indeed the majority of the emT)ankments are uncemented. In regions of excavation the tufa occurs just outside the edge of the cut-terrace, coating the lower slope for a space of 20 or 30 feet. Its zone of maximum deposition was probably from 10 to 20 feet beneath the water surface. Where the deposit is thin, it consists merely of a uniform film, but wherever it acquires a thickness of an inch or more, there is manifested a tendency to assume dendroid forms. These are not uniform in character, but generally consist of branching stems, an eighth or a fourth of an inch 168 LAKE BONNEVILLE. in diameter, frequently dividing and again joining, so as to constitute a reticulated mass in which the interspaces are not large. The composition is shown ])y the following analysis, copied from the report of the Fortieth Parallel Survey, Vol. 1, page 502: Analysis of Tufa\froin Maiv Terrace, liedding SpriiKj, 'Suit Jake Desert, by li. JV. Wooduard. [Specific gravity, 2.4, 2.3, 2.4.] Silicic acid (cbiefly iucluiled sand) Alumina Sfsquiosidu of iron Lime Ma<;ne8ia Soda Potassa Litbia ... Phosphoric acid AV ater Carbonic arid Total Percentages. First, sample. Second sample. 8.40 B.22 1.31 '.20 Tr. Tr. 40.38 i. 50 3.54 3. .52 0.48 0.22 0.54 0.22 Tl. Tr Tr. Tr. 1.71 38.20 1.62 38.33 100.24 100. 14 On p.iges 495 and 496 of the same volume, the microscopic cliaracters of tlio tufa are described by King. The distribution of the tufa along each shore is independent of the nature of the subjacent terrane. The heaviest observed deposits are upon quartzite and granite at a considerable distance from calcareous rocks. The most conspicuous accumulations are upon rock in place, but this difference probably depends u])on the fact that deposits upon unconsolidated material are largely interstitial. A more important peculiarity of the distribution is its relation to wave action. No deposit is found in sheltered bays; and on the open coast those points least protected from the fury of the Avaves seem to have received the most generous coating. These characters indicate, first, that the material did not have a local origin at the shore but was derived from the normal lake-water; second, that the surf afforded a determining condition of deposition. It will appear in a later chajiter that calcareous 'The analysis is headed "Thiuolite (pseudo Gay-Lussite) " — prohably thnmjih inadvertence, for the reference to the analy.sis in the text (p. 4%) iise.s the dpsit;nation tufa only ; and the iheory in res-ird to the origin of the Lahontan tnfa which is cnihodicd in llic term "psendo Gay-Lussite," appears from the context not to have been applied to the Bonneville basin. TUFA. 169 matter constitutes an important part of the fine sediment of the hike bottom, and that this was chiefly or wholly precipitated from solution. It is not easy to see why this deposition should consist of discrete particles in the open lake and be welded into a continuous mass upon the shore; but a par- tial explanation a})pears to be afforded by the hypothesis that the separation was promoted by the aeration of the water. All precipitation being initiated at the surface during storms, coalescence at the shore niay ha\'e resulted from contact at the instant of separation. The suggestion finds a certain amount of support in the part played by nuclei as determinants of precipi- tation. The thickest deposit anywhere observed is on the outer verge of the Provo terrace at the north end of Reservoir Butte, where there is a, maxi- mum of four feet. The tufa, there coats a knob of solid quartzite so situated that while it was fully exposed to the surf, whatever the direction of the wind, it was exempt from attack by shore drift. The locality is exceptional; in most places where the tufa is so abundant as readily to attract attention, its depth is measured by inches. An allied deposit may be mentioned in this connection, namely, oolitic sand. This was fiist observed on the Bonneville shores by Miss Susan Coolidge, of Grantsville, Utah, and was afterward found by Messrs. W. J. McGee and George M. Wright on several shore terraces at the north end of the Oquirrh Range. It is now forming in Great Salt Lake along the coast between the delta of the Jordan and Black Rock, where it constitutes the material of a beach, and is drifted shorcAvard in dunes. Like the tufa, it is exclusively a shore formation, but the circumstances connected with its occurrence on the modern shores of Gi'eat Salt Lake and Pyramid Lake warrant the suspicif)n that it is not e(iually independent of local sources of supply. The locality mentioned on the shore of Great Salt Lake is near the mouth of a stream whose annual tribute of carbonate of lime can not be small, and the only known locality on Pyramid Lake is associated with hot calcareous springs. RE8UMfi. The highest of the shore-lines jireserved on the slopes of the basin, namely, the Bonneville shore-line, has an altitude of 1,000 feet above Great 170 LAKE BONNEVILLE. Salt Lake. By reason of its position at the top of the series, it is the most conspicuous of all; but the one most deeply carved is the Provo, 375 feet lower. Between the Bonneville and Provo are the Intermediate shore- lines, characterized by embankments of great size, but without" correspoiid- ingly great sea-clifFs and terraces. Below the Provo tlie .slopes exliiljjt lake sediments, with occasional shore-lines superposed. Of these latter the Staus- bury is the most prominent. The area of the lake at the Bonneville stage was 19,750 square miles; at the Provo stage, about 13,000 square miles; at the Stansbury stage, about 7,000 square miles. The order of sequence of the shores to which names have been given is: first. Intermediate; second, Bonneville; third, Provo; fourth, Stansbury. During the period of the formation of the Intennediate embankments, tliere were no persistent water stages; but the water surface oscillated uj) and down. The last additions to the embankments were made during a gen- eral advance of the water. The oscillation of the water surface continued through the Bonneville epoch, the Bonneville shore representing the cnm- bined results of wave action at a series of water levels having a vertical range of 20 feet. The last stage of this series was the liigliest, and imme- diately afterward the surface fell ra})i(lly to the Provo horizon, where it remained a long time. The water margin afterward receded from the Provo shore to its present position, halting occasionally by the way, and longest at the Stansbury sliore. CHAPTER IV. THE OUTLET. Tliirteen years ago I had the temerity to predict,^ first, that the position of the Bonneville shore-line would eventually be shown to have been deter- mined by an overflow of the lake, and second, that the Provo shore-line would be found to have been similarly determined. The first of these pre- dictions has been verified in its letter, but not in its spirit; the second has proved to have full warrant. My anticipation was based on the following consideration: A lake without overflow has its extent determined by the ratio of precipitation to evaporation within its basin; and since this ratio is inconstant, fluctuating from year to year and from decade to decade, it is highly improbable that the water level will remain constant long enough to permit its waves to carve a deep record. I failed to take account of the fact that the highest shore-mark of the series is conspicuous by reason of the contrast there exhibited between land sculptni-e and littoral sculpture. We now know that the height of the Bonneville shore-line was determined in a certain sense by overflow, since a discharge limited the rise of the water; but the carving of the shore was essentially completed before the discharge; and as soon as that began, the water level fell. At the Provo horizon, on the contrary, a constant or nearly constant water-level was maintained by discharge for a very long time. The outlet of a lake is necessarily across the lowest point of the rim of its basin; and it is essential that this point be somewhat lower than the water level of the lake. The search for an outlet to Lake Bonneville was therefore a search for a pass in the rim of the basin lower than the neigh- ' Expl. West of the 100th Mer., vol. 3, pp. 90, 91. 171 172 LAKE BONNEVILLE. boring shore-lines. It is e(|ually necessary tluit tli(! liasin on the opposite side of tlie ])ass he connj)etent to receive the discharged water. It nnist either drain to tlic ocean or else be snfficiently large and suflicicntl\' arid to dispose of tlie afflncnt water ])y evaporation. TIh^ conditions of outlet having been satisticd, and a discharge having been produced, it is ('(|ual]y evident that the process of that discharge would modify the topographv in a peculiar manner. A channel would be produced at the pass, and this would descend in one direction only, its sides and bottom merging at the pass into other topographic features. The site of the ancient outlet of Lake Bonneville should therefore exhibit a channel, the bed f)f which is lower than the contiguous shore-line, and the de.scent of which is toward some basin competent to receive and dispose of the water. It is quite C(mceivable that a basin like tlie l)onneville, known to ])e subject to deformation through hypogene agencies, should discharge its surplus water at one time over one pass and afterward over another; and this possibility was one of the considerations leading to an examination i>f its entire coast line. By that examination it was ascertained that all the lower passes of tlie basin's rim are at the north, se})arating the basin from the drainage system of the Columbia River. These passes were systematically visited by competent observers; and it was ascertained that the Bonneville waters discharged at one point only. The trend of the mountain ranges in that region is generally north and south and the passes are siinjdy culminating points in the intervening valleys. As a rule they are not rocky, but con.sist of alluvium, the profiles of which rise gently toward the mountains on either side. South of each sm/h pass the minor drainage lines from each mountain unite and produce a main drainafje channel descendino: toward the basin of Great Salt Lake. At tlie north a similar confluence produces a drainage channel descending toward the tributaries of the Columbia. On the pass the alluvial profiles from the mountains unite with gentle curvature; and there is no channel of drainage. It is a curious fact that in a region characterized by great reliefs of sui-- face, a munber of passes were so nearlv at the same level that a difierence of only a few feet determined the actual point of discharge. The water of the lake rose within 75 feet of the pass north of Kelton, where the Boisd SEARCH FOR THE OUTLET. 173 stage-road crosses from the Salt Lake basin to the head-waters of Raft River; and it rose Avitliin 100 feet and 200 feet, respectively, of the passes north of Snowsville and Curlew. Red Rock pass.-Tlie actual point of discharge was at the north end of Cache Valley, at a point known as Red Rock Pass; the outflowing river entered Marsh Creek valley, and being there joined by the Portneuf, flowed through Portneuf Pass to the valley of the Snake River. The first suggestion of its position was by Bradle}', who crossed the old channel some miles l)elow the ])ass in 1872; and it was independently demonstrated l)y Mr. Gilbert Thompson and by the writer, wlio separately visited the localit-s^ some years later.^ The ascent to Red Rock l*ass from Cache Valley is so gentle as to be scarcely noticeable, and the descent on the opposite side, while ]ijerceptible to the eye, affords an easy grade to the Utah and Northern Railroad. A few miles west of the pass, there rises a lofty mountain ridge separating Cache Valley and Marsh Valley from Malade Valley. On the east are lower mountains, separating Cache Valley and Marsh Valley from Gentile Valley and Basalt Valley. From the base of the range on either side, an alluvial slope descends to the pass, but this is not continuous. Knobs of indurated rock, similar to those constituting the mountain, project through it, testif}'ing to the existence a short distance beneath the alluvium of a rocky sj)ur comiecting the two ranges. At a few points there are exposui-es of less indurated rocks, supposed to be of Tertiary age, but these form no hills by themselves, being buried under the alluvium except where laid bare by recent erosion. The alluvium is further interrupted by the clianncl of the ancient outlet, which is one of the most notable features of the landscape. It has been excavated to a depth of several hundred feet, and has a general I It was iiiiiintained by I'eale tliat tlie orii;inal point of discharge was at Portneuf Pass instead of Red Rock Pass; and the discussion of tliis view gave to the subject of the onth-t and its discovery a, more voluniiuous literature than perhaps it deserved. The writer's diss nt from Pcale's determination has already been recorded in discussing the supremacy of the Bonneville shoreline (p. 94). Readers who care to pursue the subject further will fiud the following references useful :— G. K. Gilber*, in Sur- veys Wist of the 100th Meridian, vol. 3, Geology, p. 91 : E. E. Howell, idem, p. 'Jr.l ; F. H. Bradley, Oeol. Survey of Terr., Ann Rept. for 1872, pp. 'JO'^, 20:i ; Gilbert, Bull. Phil. Soc, Washington, vol. 2, p. 103: A. C. Peale, Geol. Survey of Terrs., Ann. Rept. for 1^~7, pp. 565, 642; Am. Jour. Sci., 3d series, vol. 15, 1H7H, p. C5; Gilbert, idem, M Series, vol. 15, 1878, p. 256; Peale, idem, vol. 15, 1878, p. 439; Gilliert, idem, vol. 19, 1880, p. 342; Lieut. Willard Young, Surveys West 100th Meridian, Ann. Rept. for 1878, n. 121. 174 LAKE BONNEVILLE. width of about one-third of a mile. Five small streams flow from the mountains to the ancient channel, and each of these has carved a deep trench in the alluvium, casting the eroded mateinal into the channel. The gi-eatest of the streams is Marsh Creek, debouching at Hunt's Ranch; and its freshly formed deposit occupies the old channel for a distance of nearly three miles. Three or four miles farther south Five Acre Creek makes a similar tribute, filling the old channel with alluvium for the space of a mile; and the same thing is repeated on a smaller scale by Stockton Creek, two miles farther south. The alluvial fan built by Marsh Creek is a few feet higher than the others, so that the actiial water parting is at Hunt's Ranch. Between the Marsh Creek and Five-Acre Creek alluvia, the old chan- nel is occupied by a marsh three miles in length with an average width of twelve hundred feet; and Avithin this there is a small pond. Between the alluvia of Five-Acre Creek and Stockton Creek there is a larger pond, known as Swan Lake. These marshes and ponds, whenever they accumulate \\ater enough to overflow, drain sovithward to Cache Valley; and all the streams of the pass except Marsh Creek are tributary to them. ]\Iarsh Creek turns abruptly north on entering the channel and flows toward Marsh Valley. Its volume is so small that during the dry season it does not maintain a super- ficial flow through the valley, but repeatedly sinks beneath the smface and reappears below in springs. The knobs of indurated rock, which in the immediate vicinit}' of the pass consist of arenaceous limestone, both adjoin and interrupt the chan- nel. Near Hunt's Ranch there are two buttes, each several hundi-ed feet in height, overlooking the channel from opposite sides, and between them are a nmnber of low reefs projecting throvigh the flood-])lain of ]\Iarsli Creek Constricted by these reefs, the channel has a mininnnn superficial widtli of only GOO feet. The relations of these various features will be better imderstood by reference to the map in PI. XXVKI. The Bonneville shore-line is traceable continuously about Cache Valley to the vicinity of the pass. On the east side its most noi'therly vestige is upon a butte a mile south of Hunt's Ranch. On the west side it is lost on the alluvial slope two miles from Hunt's Ranch. Its height above the marsh U S. GEOLOGICAL SURVEY liAKE BOMNE'/ILLZ, PL Xr^lir MAP OF THE OUTLET OF l^VIvE BONNEVILLE R E D R 0 C Iv P ASS. Oneida Co Idalio Toj)Oqraph\ hv h^ J) . John son . Ofoloqy b\ GK Gilberi Soiitieiille .S'htwehne o Vi 1 SCALE I I =4 WILE 3.^'0 iet't (hntoars ji^^'- ^Oiierrt ^lltwuil Deposits . .luHus Iticn A t'o.lilli Di ovm In- G TliompM RED ItOCK PASS. 175 betAveen Marsh and Five-Acre creeks is 340 feet. The nearest, point at which the Provo shore-line was observed is about eight niih's farther south, in the vicinity of the town of Oxfoi'd. Marsh Creek issues from its canyon in the mountains aljout two and one-half miles east of the old channel. The intervening' sjjfice is occupied by a sloping alluvial plain terminating in a bluff. It is evident that this is an alluvial fan or alluvial cone constructed by the creek before the exca- vation of the Bonneville outlet. It was afterward partially eroded by the outflowing river, and also by Marsh Creek, which has excavated a passage several hundred feet in depth. Where this old alluvial plain approaches nearest to the Bonneville channel, its edge is fifty feet higher than the nearest terrace of the Bonne- ville shore, and a restoration of its profile indicates that it coalesced with slopes from the opposite mountain range at about the level of the Bonne- ville shore. A careful study of the ground has satisfied the writer that the base or outer margin of the alluvial cone was part of the ancient water- parting, and was the point at which the outflow was initiated. The fact that the Bonneville water discharged at first over a barrier of alluvium instead of solid rock had much to do with the subsequent history of the lake. Uncemented alluvium is easily and rapidly torn up and re- moved, and as soon as a current began to flow across the divide, it must have commenced the excavation of a channel. As the channel increased, the vokime of the escaping water became greater, and this increase of vol- ume reacted on the power of erosion. In a short time a mighty river was formed, and the lowering of the lake surface resulted. For a time the out- pouring was a veritable debacle, and it could not have assumed the phase of an ordinary river commensurate with the inflow of the lake until the allu- vial barrier was completely demolished and the resistance of the limestone reef was called into play. When the corrasion of the channel had proceeded so far as to give the river a bed of limestone, the process of excavation was changed from the mere transportation of loose detritus to the corrasion of solid rock, and the rate of excavation was greatly diminished. We have here the ex2)lanation of the rajiidity of the final recession of the lake from the Bonneville level to the Provo. 176 LAKE r.ONNEVlLLE. Marsh Valley—Marsh Vulk'}', like Cache Valle}', is ench).sed Ijetween mount- ain ranges, and has a north and south trend. Its length is aljout thirty-five miles, and its greatest width is eight or ten miles. Twenty miles from Red Rock Pass, the Portneuf River breaks through the eastern mountain chain and enters the valley, turning northward and running parallel ^\itli Mai-sh Creek to the end of the valley. There it receives the creek and then tiii-ns abruptly westward and escapes from the valley through a deep liut ojx-n canyon. The upper canyon of the Portneuf has at some time admitted hna as Avell as water. A succession of basaltic coulees have poured thi-ough it into Marsh Valley and have followed the slope of the valley to the lower canyon. The Portneuf River follows the western mai-gin of the lava beds, and i\Iarsh Creek the eastern, each occu])ying a narrow valley sunk from 'M) to 100 feet below the level of the lava table. A comparison of these val- leys illustrates the disparity between Marsh Creek and its channel. I'urt- neuf River is several times larger than Marsh Creek; l)ut the inuncdiatc valley liy which it is contained is smaller. Indeed, there is every evidence that the valley of Marsh Creek, having been formed by the ancient Bonne- ville river, is now in process of filling. It abounds in meadows and marshes, and at one point contains a lakelet. The River.-It a})pears, however, that the Bonneville river was nut citn- tained during its entire existence in the channel now ((ccuj)ied by Mar.-li Creek. The whole upper surface of the lava tongue, where it has a width of more than a mile, is fluted and polished, and pitted with pdt-holes after the manner nf a river bed; and there seems no escape from the condusidn that it was swept l)y a broad and rapid current. The trenches at the side of the lava may or may not then have existed; l)ut even if they did not, we have to contemplate, as the agent of corrasion, a river comparable with Niagara. Indeed it is even possibles that Niagara might suifer by com- parison. Let us assume that at the time the Bonneville river travtu'sed the la\a- bed the lower channel at the side had not been eroded; and let us furtlu-r assume that its width was somewhat less than that of thi' lava, — sa\' one mile. When the river came into being, the total descent of its bed, from one end of Marsh Valley to the other, was at the rate of 13 feet to the mile. THE DEBACLE. 177 In the last stages of its existence its average grade in the same space was 7 feet to the mile. At all stages the declivity was greater near the pass than in the lower end of Marsh Valley. Let us assume that the slope of the water surface in flowing over the lava was 2^ feet to tlie mile, or one foot in 2,000. If now we assume in addition that the discharge equaled that of the Niagara Rivei', we have all the data necessary for computing the mean depth; and A\e obtain ftir that depth 9 feet. To one who stands upon the lava bed and notes the scale of the carvings which ornament its surface, this determination appears for too small. Twenty feet would better accord whh the phenomena, and twenty feet woidd discharge the flood volume of the Missouri. Another evidence of the magnitude of the outflow is found at the pass. West of the swamp there is an irregular terrace, extending from Swan Lake to Red Rock, the upper surface of which is corrugated with parallel furrows and I'idges trending in the general direction of the current. These consist partly of limestone crags and partly of alluvium. Comparing them with similar flutings in other stream beds, they ap^iear to be explicable only as details of channel-bottom wrought by a torrent of great volume. How long the discharging river maintained its colossal dimensions can not be learned, but the period certainly w^as not great. The entire prism of water between tlie Bonneville and Provo planes would be discharged by the Niagara channel in less than 25 years; and if the Bonneville river reached a greater size, it could liave maintained it only for a shorter time. It is evident that the channel at the pass has been partly filled since the desiccation of its river; but the precise amount of filling is not so evi- dent. A crude estimate was based U2)on the configuration of certain small drainage lines tributary to it. Before the filling began, these drainage lines (as, for example, that of Gooseberry Creek; see PI. XXVIII) found their base of erosion in the main channel, and adjusted their profiles thereto. As the filling of the channel progressed they were likewise partially filled near their mouths; and a study of their configuration yields a crude esti- mate of the amomit of deposition. It is judged to be about thirty feet; and if this estimate is con-ect, the bottom of the channel is 370 feet lower than the Bonneville shore. This is approximately equal to the difference in MON I 12 178 LAKE BONNEVILLE. level of the Bonneville and Provo sliores and it serves to connect the testi- mony of the outlet with that of the shore-lines. It is not easy to estimate the cross section of tlic channel of outflow at any stage of its existence. Undouhtedly it was broader and deeper while its walls and bed consisted of alluvium than afterward when solid rock was reached. The trough now occupied by the marshes and Swan Lake i)roba- bly represents its width after rapid corrasion had ceased ;nid before tJH^ tiiial desiccation of tlic lake was begun; but this is a mere surmise. We ncc); an.l G. K. Gilbert, Amer. Jour. Sci., 3(1 nor., vol. 10, 187C, p. 228. 'J S. GEOLOGICAL SUP,VEY LAKE BOMKE^TLLE. PL.XXXI Juliua BirnAOu.UUi DroHTi bv (• Thompsoi THE OLD RIVER BED. 183 to the River Bed Station its depth increases to 130 feet. In the pass be- tween the mountains its Ijanks coalesce with the steep faces of huttes; and its general depth may he several hundred feet. This description applies merely to its present condition. There is good reason to holieve that, at the time of its desiccation, it was deeper, especially in the southern part. Everywhere it is margined by easily eroded lake sediments; and near the mountains the surfoce of these lies at such an angle that every rain washes down an abundance of mud into th(^ old channel. On the Salt Lake Desert the ])lain is so nearly level that superficial waters have little power of erosion, and the silting of the channel has been less. In the vicinity of the pass the recent deposit lias a probable depth of 100 to 200 feet. The general descent of the channel is from south to north, but this is interrupted at one point in the pass by an alluvial dam, over which the water seems to find its way rarely. The direction of the original descent, or the direction of drainage through the channel, is not demonstrated by the existing levels; but fortunately there is other evidence in the shape of a terrace, marking a flood-plain of the ancient stream when its channel was half excavated. This appears on the banks of the channel north of the River Bed Station, and is capped by a deposit of fine gravel, the pebbles of which are evidently derived from the McDowell and Simpson Mountains. From the head of the chamiel the plain of the Sevier Desert descends southward for many miles; and it is evident that, when the channel was occupied by a river, the desert was covered by a lake. In a word, the channel was opened at a time, during the final desiccation of the lake, when the level of tlie water in the main body fell below the l)ottom of the strait. The inflow of the Sevier body was for a time greater than its re- stricted lake surface could discharge by evaporation, and the surplus flowed over the pass to the main body, opening a channel as it flowed. The upper lake thus preserved on the Sevier Desert Avas both small and shallow, and its shore marks have not been identified. The lower lake was large, and may have left a well marked shore record; but this has not been discrimi- nated from others on the margin of the desert. A rough estimate, based on a general knowledge of the contours of the country, indicates that the up- 184 LAKE BONNEVILLE. per lake had one-eleventli the area of the lower. The lake system had also another member, for the Bonneville shore had then receded from Utah Val- ley, and the outlet of Utah Lake was, as now, an affluent of the Great Salt Lake basin. The continuance of the climatic decadence finally lowered SeAaer Lake below the level of outflow and dried tlie liver bed. It has already been remarked that even in the i)ass between the mount- ains the river bed was carved from the lacustrine strata deposited by Lake Bonneville. The Bonneville strata there rest against steep faces of the rocky buttes; and the relation of these faces to each other, and to the gen- eral course of the channel, indicates that they are the walls of an older channel whose course the post-Bonneville river followed. The history of this older channel is unknown ; and its discovery only tells us that, at some unknoAvn period before the lake, there was free da-ainage from one desert to the other. Tliere seems no waj^ to determine in which direction this drain- age led, nor whether either plain was covered by a lake. OTHER ANCIENT RIVERS. ' • Three other long abandoned stream courses have been observed within the basin. One of these has already been mentioned. The pass between Rush and Tooele valleys is now dammed across by a great system of wave- built bars, which prevent the drainage of Rush Valley from passing through Tooele Valley to Great Salt Lake. Against this dam the water of Rush Valley sometimes accumulates in a lakelet known as Rush Lake, and tliis lakelet occupies a portion of the ancient drainage channel. It has a width of 1,000 feet, and is shallow. Doubtless the depth of the channel has been considerably diminished by recent deposits; and if these were cleared aAvay the width of its bed would be found smaller than the indication given by the lake. This channel is interpreted as showing, not that there was anciently in Rush Valley a water supply com])etent to override and remove such a liar- rier as noAV restrains it, but merely that, before the creation of the Bonne- ville lake, the valley had free drninage northward. A larger channel, whose habit indicates a stream comparable with the smaller rivers of the basin, enters Snake Valley from the south at a point OTHER OLD RIVERS. 185 just east of Wlieeler Peak known as the Snake Valley Settlement. The channel ends at the margin of the old lake, and appears to have contained a stream triljutary to the lake, which disappeared at the same time. It is now occupied near the settlement by a streamlet from the adjacent mount- ain known as Lake Creek, but this enters the channel at its side, and played no important jiart in its formation. Above its confluence the channel has essentially the same dimensions, and these continue as far as it was traced, about twenty miles from its mouth. Circumstances did not permit its far- ther exploration. Near its mouth the ancient stream cut across the base of an immense alluvial fan, poured out from Wheeler Peak, opening a channel 1,000 feet broad, which retains a depth of 50 feet. A secondary alluvial fan, formed by the same mountain stream, and from the material amassed in the first, was afterwards thrown across the channel, damming it and causing a small lake. Still more recently this dam was broken through and a smaller chan- nel Avas opened, whereby the lake was nearly drained, and Lake Creek escaped to Snake Valley. The closing chapter of the history has been con- tributed by man. The denizens of the little hamlet have built another dam within the small channel (a puny and insignificant affair compared with those of Nature's construction), whereby they have created a pond for the storage of water for iri'igation. A third stream course of some magnitude enters the basin in Idaho at the north end of Snowsville Valley, dc'bouching, fi'om a mountain at the west, almost precisely at the divide between the drainage of the Basin and that of the Snake River. It was not traced toward its source, l)ut the grade of its bed indicates that it drains a valley of some size within the mountains. Its flood-plain has a breadth, just before it reaches the Bonneville horizon, of 2,000 feet, and below that liorizon is covered by the lake sediments. Within the lake area it can be traced for several miles, although lined throughout by the lacustrine deposits. Through this channel water rarely finds its way at the present time. The flood-plain is covered by soil and vegetation, which give no evidence of recent disturbance except along a narrow meandering- trench that one may leap across. There is here no delta associated with the Bonneville shore, and the implication seems to be 186 LAKE BONNEVILLE. that the locality was characterized at tsome very ancient date by a climate more liumid than either the lionneville or the present. With tliese exceptions the water courses of the drier coasts are not known to give evidence of modification. All of them are larger than the ordinary streams within them require; but the extraordinary requirements in an arid region are so great that the channels do not seem abiKjnnal. OUTLETS AND SHORE-LINES. The harmony between the conclusions based on the phenomena of the shore-lines and those derived from the features associated with tlie outlet lias a double bearing. On the one hand, it serves to establish the elements of the lake's history thus far set forth; and on the other it defines tlu^ in- fluence of outflow on shore topogra2)hy. Without outflow the level of a lake is inconstant and oscillatory, and unless the water stands long at the same level the waves will not excavate cliffs and ten-aces comparable in magnitude with the embankments constructed. It follows that the Stansbury shore, which gives e\'idence t)f a perma- nent water stage, not merely by its cliffs and terraces but by its accumula- tion of tufa, was determined by an outflow or its equivalent. At one time I supposed that the problem f»f its existence would be solved by the Old River Bed — that its level would be found to have been determined by a discharge from the main Ijody to the Sevier body; but this hypothesis was was overthrown by the study of the river bed, which showed the discharge to have been northward instead of southward. The precise relation of the Stansbury shore to the river bed has not been ascertained, for the shon^ has not been recognized in that -sdcinity, but they do not differ greatly in alti- tude. It is probable that during the Stansbury epoch the main lake did not extend to the Sevier Desert. There is one other valley Avhich might have served as a reservoir for surplus water at the Stansbury stag(>, Init tlie connecting strait has not been critically examined. White Valley contained a large bay during both the Bonneville and Provo epochs, and was deej) enough to have received a considerable discharge at the Stansl>ury stage, if the strait was adjusted to its delivery. Its area is indeed small as compared to the main lake at that level, but it might none the less have served as a THE STANSBUKY PROBLEM. 187 regulator, causing the oscillating lake to linger at a particular level each time it rose. The nature of the problem embodied in the Stansbury shore was not realized until the field examinations were so nearly complete that the op- portunity had passed for visiting the localities important for its discussion. It therefore remains as one of the unanswered questions developed by the investigation. CHAPTER V. THE BONNEVILLE BEDS. A certain series of lacustrine strata have been designated tlie Bonne- ville beds. Tlieir relation to the old shore-lines was first pointed out by Hayden/ and afterward by the geologists of the Fortieth Parallel Siu'vey and the Wheeler Survey. The grounds for the correlation have not been distinctly enunciated, probably because they are so patent to each obsers^er that their statement seems surperfluous. In the present work, however, it is proposed to combine the history derived from the sediments with the history derived from the shore record; and there is a logical necessity for establishing the general synchronism of the two. A brief account has already been given of the Tertiary lacustrine strata observed in the Bonneville basin. While these exhibit considerable variety in texture, they are in general so distinct lithologically from the Bonne^■ille beds that their discrimination has been easy and uuemban-assed by doubt. The Bonneville lieds occupy the lowlands, constituting nearly the entire surface, and retain the attitude of deposition, Ijnng flat on the open plain or gently inclining at the bases of the mountains. Wherever the outcrops of the Tertiaiy beds are associated with these, they exhiliit dips referable to displacement, and they are overlain tuiconformably by the Bonneville. Tlie Bonneville beds are thus seen to be the latest lacustrine deposit of the basin, and this fact indicates tlieir synchronism with the latest littoral evidence of a lacustrine condition. Again, the distribution of the Bonneville beds is strictly limited l)y the Bonneville shore-line; and none of the other groups are so limited. The latter are thus shown to be older than the shore-lines. The Bonne%nlle ' Snn-pictures of Rocky Mountain Scenery, by F. V. Haydeu, New York, 1B70, p. 1S2; Auu. Kept. Geol. Survey Terr, for 1870, p. 170. 188 COEEELATION OP SEDIMENTS AND SUOKE-LINES. 189 beds are not traceable outward from the center of the basin to all parts of the Bonneville shore-lines, or at least they do not to that limit hold their familiar characters; but they bear to the shore-line certain definite relations, \\liich may be stated. Where the margin of the basin is steep and the shore-line is high, the lake beds reach to the foot of the slope; where the basin margin is gently inclined, as in the shallow bays, they extend nearly to the outer limit of wave work. Finally, as has been fully set forth by King,' the Bonneville beds are in places interstratified with alluvial deposits; they rest upon the principal mass of alluvium from the mountains and support alluvium of recent trans- portation. Tins relation is strictly paralleled by the shore-lines, which rest upon the alluvial cones of the mountain bases and are themselves overplaced by recent alluvium. Adding to these facts the a priori consideration that the deltas contain only the coarser material brought by streams, the finer having been car- ried in suspension to the lake, and that the shore embankments represent only the coarser part of the product of littoral erosion, the finer having been carried lakeward by the undertow, so that there must have been fine lake sediments contemporaneous with the deltas and embankments of the shore, the general correspondence of the Bonneville beds with the Bonneville shore-lines is clearly established. It is only in regard to details that the correlation is less clear than could be desired. One result of the deposition of the sediments was the raising of the base level of ex'osion of all streams tributary to the basin, so as to make them agents of deposition along their lower courses in post- Bonneville time. The localities are therefore exceedingly rare where even partial sections of the Bonneville beds can be observed; and it is only at their extreme outer limits, where they rise toward the shore, that their base is ever seen. LOWER RIVER BED SECTION. The deepest section of the lake beds, or more strictly the section repre- senting the largest fraction of the Bonneville Period, is exposed in the walls of the Old River Bed near the point where it is crossed by the Overland ' Geol. 40th Par., vol. 1, p. 493. 190 LAKE BJNNEVILLE. Stafi;'e-r()ad. It lias some title to be reg'arded as the typical section, and exhibits tlio following' iriendjers: 1. (At base.) The Yellow Clay, a fine argillaceous deposit, laminated throughout, olive gray on its fresh exposure, but weathering to a pale yellow. In this are occasional passages of sand, but these are local and discontin- uous. Nodules of selenite, consisting of grouped arrow-head crystals, are abundant; and jointage cracks sometimes contain rosettes of recrystallized gypsum. Bivalve shells of several species are included. The base is not seen; a thickness of 90 feet is exposed. 2. The White Marl, a fine calcareous clay or argillaceous marl, light gray or cream-colored on fresh exposure, nearly white on weathered sur- face. Contains some gypsum, but less than No. 1. Overlies No. 1 with unconformity by erosion, and is at its- base crowded with shells represent- ing nearly the same fauna. Thickness, 10 feet. 3. The marl passes upward into a fine sand, the transition being grad- ual and the continuity perfect. The sand contains also the same species of shells. Thickness, about 10 feet, the upper limit being obsciu'ed by a recent eolian deposit of similar texture. The distribution of the Yellow Clay and White Marl is universal through- out the lower parts of the basin, and they ascend in the shallower bays toward the upper shore-lines. At low levels their physical charactcu-s undergo little change, and they are readily discriminated by their diff"erence in color. At very low levels a yellow clay appears over the White Marl, blending with it as though continuously deposited. This may be the equivalent of the sandy member in the typi(;al section, which is not everywhere foinid. The unconformity between the Clay and the Marl does not include any observed diff"erence in inclination, and is not always detectable, but it was observed at localities so widely distrilmted as to indicate that it is not a mere local phenomenon. Against the steeper coasts the beds appear to terminate somewhat abruptly at low levels; but on gentle slojies they con- tinue with a change of character, acquiring sand both by admixture and by intercalation. By these changes their distinctive chai-acters are lost, and at high levels their separation is for the most part impossible. THE TYPE SECTION. 191 The exposures of the Yellow Clay are so rare and so small that its special mutations can not be characterized, but abundant opportiniity is atlbrded for observation of the White Marl. As the shore is approached, the arenaceous capping increases in relative thickness, encroaching on the marl below. The base is the last to change, holding its white color on many l)arts of the coast to levels above the Provo shore. At numerous points between the Bomieville and Provo liorizons, sedi- mentary deposits are seen to alternate with littoral, the former consisting of marls, clays, and sands, and the latter of shore drift in the form of spits and bars. We have not succeeded in correlating these sublittoral deposits either with each other or with the lacustrine sediments of the center of the basin; and the phenomena, although numerous, are so fragmentary that there seems no advantage in placing their details on record. Their only contribution to the deduced history of the lake is the confirmation they afford of the con- clusion indepeiidently reached that the surface of the lake, when not limited by outflow, was subject to many minor oscillations. At a few localities there was observed an abnormal development of the lacustrine section, a result of what may be called redeposition. A single illustration will suffice. Snowsville Valley contained at the Bonneville stage a bay eight miles broad and rumiing twenty miles inland. At the Provo stage its linear dimensions were reduced one-half, and it became shallow. At a later and lower stage, possibly the Stansbury, the water barely reached to the entrance of the bay; and at this time the freshly deposited muds of the bay appear to have been washed lakeward in great volume, accumulat- ing at the mouth of the bay in a series of sheets inclined at an angle of 3 or 4 degrees toward the lake. This may perhaps be called a delta deposit, but it differs from typical deltas in the fineness of its material and the conse- quent low angle of crosS lamination. The last addition to the deposit con- stitutes the face of a percejjtible terrace, ascended by the road from Curlew to Snowsville. Through this terrace Deep Creek or Deseret Creek, the drain of the valley, has excavated a channel from twenty to thirty feet in depth, exposing the structure of the mass. The deposit has a general resemblance to the normal lake beds, but exhibits four or five alternations of the typical yellow and white colors. 192 LAKE 130NNEV1LLE. LEMINGTON SECTION. The uiicniif'orinity of tlio White Marl upon the Yellow Clay iiidieates (liscoutiimity of lacustrine eouditious; and at two localities this evidence is supplemented by the occurrence of subaifrial d(;posits at the horizon of un- conformity. ( )ue of these hjcalities is at Lemin«.^-ton, where the .Sevier River, issuing from its narrow valley iu the Canyon Range, enters the Sevier Desert. During the highest water stages, no delta was foi-med at this point, because the land-locked bay on the east side of th(^ range received and i-etained all the coarser alluvium; but a great amount of tine matter was washed into the lake, and this was deposited with exceptional rapidity ;d)(Kit the mouth of the estuary. The total local deposit must have amounted to several hundred feet, and recent erosion by the river has exposed 150 feet of this to view. The point of sj^ecial interest is just outside the canyon mouth, where the lacustrine strata are seen to abut against the steep face of Fig. 2h, — Section .showing; snccessiou nf Lacnstriiit* iind Alluvial Ut-posits at Leniini:t«ii. Ut'li. 1. Piilt'ozoic sandstoiKV 2. Tlit! Yellow ('lay (Lower Homieville). 'i. \VtHlj;e of alluvial ;irav«l. 4. The White Marl (Upper IJuuneville). 5. Keceut alluvial j;tavel. G. liuiiiieville shore uutcli, with recent talus. , quartzite constituting the mountain front. The material of the lake beds is here coarser than in the typical section, and the contrast in color between the upper and lower series is barely discernible. The Yellow Clay incdudes through nearly its whole depth a considerable percentage of fine sand, and the White Marl has a fine texture only at its base, consisting above of coarse and fine sands. SECTION ON THE SEVIER RIVER. 193 Associated with the lake beds are two wedges of alluvium, the tliicker ends of which abut ag,ainst the quartzite of the mountain. The upper of these is a modern deposit, receiving- additions at every storm; the loAver, which otherwise is similar in all its characters, is inserted between the White Marl and the Yellow Clay. The Marl and its associated sand have here a joint thickness of 50 feet, and the Yellow Clay a visible thickness of 100 feet, the base being con- cealed. Tlie Bonneville shore-line, here taking tlie form of a terrace and clitf, runs 50 feet above the upper limit of the White Marl and 120 feet above the upper limit of the Yellow Clay. The series of events by which these relations were produced can not be mistaken. While the lake stood at a liigli level the Yellow Clay was de- posited against the base of the mountain; and as the de])osit extends to within 120 feet of the Bonneville .shore, the lake level must have a])proaclied this maxiiuimi very nearly. Then the water receded so for as to l)ring sub- aerial agencies locally into jjlay. The waste from the mountain face was washed by the rain into the margin of the lacustrine deposit, and accumu- lated there in a talus or alluvial slope of low inclination. Afterward the water returned, and remained at a high level during the deposition of the White Marl; and at the sanae time the Bonneville shore terrace was cut by the waves. The locality was carefully studied for the purpose of discovering other intercalary alluvial wedges, but none were found; and the exposures were sufficiently complete to warrant the confident assertion that none exist within the range of the section. Their a])sence indicates that during the deposition of the visible portion of the lower sedimentary formation the water did not fall more than 200 feet below the Bonneville horizon, and that during the period represented by the upper deposit the water did not fall more than 150 feet below the Bonneville horizon; that is to say, the locality records twf) high stages of the lake separated by an epoch of lower water, and [)recludes the hypothesis of a larger number of great oscillations of water surface within t!ie limits indicated by the local deposits. MON I 13 194 LAKE BOXNEVILLE. UPPER RIVER BED SECTION. The second locality ;it which the clay and marl are separated by sub- aerial deposits is at the Old River Bed, about five miles south of the point at which the typical section of the lake deposits was observed. The sedi- ments here lie about seventy feet higher, rising gradually toward the mountains and buttes between which the River Bed passes. The numljer of distinct members in the series is greater than in the northern part of the River Bed, and the relations are complicated by at lea.st one other uncom- formity. They are exhibited in the map on PI. XXXII and in the sectional diagram. Fig. 21). The letters designating formations are made to correspond in the two illustrations. Fio. 29.— Tbo Upper Eivir Bod Section; running from AA to Ulf on Pl.ite XXXH. f7. — Upper Sand. .S"G — .Second Gravel, /y = Lower Sand. If = White Marl. FG ^ First Gravel. 0= Yelliiw Clay. Vertical .scale greater than horizontal. On the left or southwest bank of the River Bed, the paleozoic terrane is largely exposed, consisting of limestones and sandstones or quartzites, be- lieved to be of Silurian age, though not yielding fossils at this precise point. The structure of the mass is not essential to the Pleistocene history. On the opposite side of the River Bed are five small buttes of trachyte and l)itchstone, nearly buried by the later deposits. These are so ancient and worn that their forms convey no information as to the original extent of the masses from wliich they have been carved. Yellow ciay.-The lowest member of the later series of formations is a fine laminated clay, which rests-against the Silurian wall on the side of the River Bed, and presumably surrounds the bases of the buttes, although its contact is not seen. Tliis is olive on fracture and yellow on weathered surfaces, and is visil)ly continuous witli tlio Yellow (*hn' of tlic tv])e .sectiim. First Gravel.- Resting on the clav, with a sliglit uncont'oniiity by erosion, are several masses of gravel. The largest runs southward from the more southerly buttes, and has protected the underlying clay from erosion. It is U S.JEOLCOICAL SUF'/EY hAl<£ B'jHHE\aLLE PL. XX>II I c I )?//,.» riav 0 LI) inVER BE D, U TAH Topo^rn-phy hv W D Johns* !()- t'i't't CoiKoiLlS .Iiil.u.H Hicn ^Vo.Uih Ur.iiwu hv li.TbtimpKf UPPER RIVER BED SECTION. 195 lenticuliir in cross-section, and has a niaxinmni tliickness of fifty feet. Its pjbbles are well rounded, and are relatively small at bottom, but at top include boulders six inches in diameter. Near the surface there is in places a calcareous cement, binding the pebbles together; and there are also rosettes or mushroom-like masses of calcareous tufa. The majority of the pebbles are of pitchstone and trachyte, similar to the material of the adjacent buttes, but there are also examples of other volcanic nicks not known to occur in situ within several miles, and also, limestone and ([uartzite, such as constitute the mountain ranges on both sides and are distributed through all the large alluvial cones of the neighborhood. At the west margin the mass can be seen to terminate in a wedge separating the Yellow Cla}' from the next member of the series, and beyond the limit of the mass there is a ribbon of sand, witli occasional pebbles, marking its horizon. Half a mile farther west this ribbon expands into a l)ed of cciarse sand and gravel, four or five feet in thickness, and half a mile north there is an independent outcrop of similar material at the same horizon. These masses are not of subaqueous deposi- tion. The form of the one first described, the associated tufa, and the pre- ponderance of boulders of local derivation, indicate shore action, but it is possible that an interlacustrine river was the agent of transportation. What- ever their origin, the gravels mark a period when the lake level \\as much lovv-er than during the deposition either of the Yellow Clay or of the suc- ceeding deposit. White Marl—Next lu ordcr is a bed of Avhite marl, eight feet in thickness, deposited uniformly over the undulating surface of the gravel and clay This is in visible continuity with the White Marl of the type section Lower sand-Tlic uiarl graduatcs upward into a bed of sand, fine below and coarse above, with a total de})th of 45 feet. The sand and marl are conformable throughout, but were both eroded before the deposition of the next bed. Second Gravel.- Above tlic saud is a second gravel, which rests unconforma- bly on the marl as well as the sand, and probably on the first gravel, from which it could not be separated at the point of contact Its pebbles are small and are mingled with a coarse sand, the whole having a thickness of about two feet. 196 LAKE BONNEVILLE. Upper sand.-Above the secoud jiravel is an up|)t'r bed of sand, conic )rinal)lo with it so far as conld be ascertained, but exhibitinjj;- little structure. This has an observed thickness of 32 feet, l)ut may have jj^ained or lost by the action of the wind, wliich throws its surface into waves, and has ciiused it to bury at the north the exposure of tlic lower formations. Upper Gravei.-Finally, tlierc appears about the bases of the more northerly buttes a fine gravel of alluvial habit. It rests on the second gravel; l)ut its relation to the upper sand Avas not seen. On the opposite side of the River Bed there are a few remnants of the White Marl capping the Yellow Clay; and at one point a small tract of sand appears, which may belong either to the lower or upper series. In terms of lake oscillation, this section bears the following interpreta- tion; first, an epoch of deep submergence, during which the Yellow Clay was deposited; second, an epoch of emergence, during which the surface of the Yellow Clay was slightly eroded and the first gravel was deposited, either by Avave action or by running water; third, a second epoch of deep submergence, during which the White Marl was thrown down; fourth, a continuance of submergence, but with a less depth, during the deposition of the lower sand; fifth, a second epoch of emei'gence, during AA'hich the lower sand and White Marl were eroded and the second gravel was deposited; sixth, a third submergence, permitting the accumulation of the upper sand as a shallow-water deposit; seventh, the final emergence and tlie erosion of the River Bed. The locality has thus been three times submerged and as many times laid bare and subjected to atmospheric erosion. It will be convenient to refer to this locality as the Upper River Bed. It is coimected by continuous outcrop with th(> Lower River Bed, where the type section of thc^ lake sediments is exhibited; but there is no such connection with Lemington, forty miles away. It is al)out sevent}' feet higher than the Lower River Bed, and about 4r)() feet lower than Lemington. OSCILIjATIONS of AVATEIl LKTEL,. At the Lower River Bed locality two emergences are recorded; at the Upper River Bed, three; at Lemington, two; ;uid it is imjiortant to the determination of the history of the oscillation that the relations of these several emergences be ascertained. COMBINING THE RECORDS. 197 There can l)e no error in referrin<>' tlie latest of tlio indicated emer- gences at each of the three locaHties to tlie final subsidence of tlie lake a,nd desiccation of the basin. There were, of course, intervals between the appearances of the several localities, tlie hig'hest being first exposed by the receding water, but the existence of these intervals does not contravene the general fact. We may therefore restrict our attention to the temporary emergences, of which the Upper River Bed witnessed two and the other localities one each. Continuity of outcrop demonstrates the identity of the first emergence at the Upper River Bed with the emergence recorded at the Lower River Bed; and there is stratigraphic evidence of a cumulative na- ture in favor of correlating the Lemington emergence with these two. 8ince this is not direct and positive, it is necessary to state it somewhat fully, in order to exhibit the weakness of the argument as well as its strength. The temporary emergence is recorded at the Lower River Bed by an unconformity — by the erosion of the surface of the Yellow Clay before the deposition of the White Marl. The section includes in descending order: (1.) White Marl, crowded with shells at the base; (2.) Unconformity; (3.) Yellow Clay. All the elements of this section are traceable continu- ously to the Upper River Bed locality, and they are repeated at several other localities low down in the basin. A few of these are higher on the slopes of the basin than the Upper River Bed, and one attains an altitude of 250 feet above the latter locality, falling only 200 feet short of the Le- mington locality. The unconformity may therefore be said to have been traced by a harmonious series of observations within 200 feet of the level of the Lemington locality. At Lemington the stratigraphic series is com- parable, but not identical. It contains all the enumerated elements except the White Marl, and this is replaced by a white clay. On the other hand, the second emergence recorded at the Upper River Bed has not been recog- nized elsewhere, so that there is some warrant for the belief that the oscil- lation of lake surface causing it had not a great amplitude. Finally, the sediment recording the latest submergence at the Upper River Bed is a sand merely, indicating that the depth of the water was not great; and if this submergence did not include the Lemington locality, the preceding emerg- ence, as recorded at the River Bed, could in no manner be separated, at Lemington, from the final emergence. 198 LAKE BONNEVILLE. The accompanying' diagram, Fig. 30, expresses graphically the con- clusions reached from the joint consideration of tlui threes localities. The vertical scale represents heiglit of water surface, ranging from tlic level of Great Salt Lake to that of the Bonneville shore. The horizontal scale represents (from left to right) the oi-der of sequence, but witlioiit any attempt to exjjress the relative duration of the several elements of the UPP£RRIV£RBU LOWtlililVtnBEl) Fig. 30. — Diagram ut' Lake Of^cillatiuas uil'enud fium Deposits aod Erosion)'. history. The curve exhibits the progressive rise and fall of the lake. Beginning at the left, we have high water represented by the Yellow (Jlay at all three localities, then an ei)oc]i of low water represented by the allu- vium at Lemington, by the first gravel at the Upper River Bed, and l)y unconformity at the Lower River Bed. llow low the water fell, does not appear. 8o far as this evidence goes, it niay have fallen only to the bottom of the Old River Bed, or it niay have descended to the level of Great Salt Lake, or even lower. Then came a second and shorter epoch of deep water, represented at Lemington by white chu- and sand, nt tlie Tpper River Bed locality by the White Marl and the lower sand, and at the Lowei- River Bed by the White Marl. The final emergence is recorded at Lemington by the superficial alluvium and by the erosion of the modem cliannel of the Sevier River. Tt is recorded at the Lower River Bed by the erosion of the River Bed and l)v its ])artial filling with alluvimn. .Vt the Upper River Bed the THE TWO FLOODS COMPARED. 199 second and third gravels, witli the intervening sand, record a general de- scent of the water, interrupted by i'n n})ward movement of small extent. It is not to be understood that this curve exhibits any more of the historj^ of oscillation than is derivable from the deposits and unconformities at these three localities. The additional elements derived from the study of the shore-lines are purposely ignored, and innumerable minor oscillations are perforce omitted. If sections of all the alluvial, littoral, and lacustrine deposits of the basin were accessible; and if these were elaborately studied, it can not ]k' doubted that the simi>le curves here drawn to represent the two great submergences of the basin would have to be replaced by lines with innumerable small inflecti(ms, similar to that deduced from the upper deposits at the Upper River Bed. In the sequel the data embodied in this curve will be combined with other data in our possession, including that from the shore-lines and outlet, and a more accurate curve will be drawn. HEIGHT OF THE FIRST MAXIMUM. If the first submergence had been carried so far as to produce outflow, the corrasion of the channel of outflow would have made it impossible for the second submergence to extend higher than the Provo level. Knowing, as we do from tlie phenomena of the shores and the features of Red Rock Pass, that the second submergence was characterized by outflow, we are warranted in concluding that the first rise was somewhat less tluni tlie sec- oiul. The amount of tlu* difference appears to be indicated by the embank- ments of Preuss \alle}', to which allusion has aln^ady been made. At the north group of embankments, figured in PI. XVI, there is an older series j)artly buried l)v a newer; and the hig'hest mend^er of this lies 90 feet below tlie Boimeville horizon. It is probable that this represents the extreme advance of the earlier flood. At the Leming-ton locality the Bonneville shore-line is the only one represented by a sea-clifl" and ten-ace; but at lower levels there are lines of tufa adhering to tlie (juartzite and apparently marking temporary positions of the water level. Probably the relation of the waves to the contiguous slopes enabled them to employ shore drift in attacking the mountain face at the Bonneville horizon, l)ut did not afford them that aid at lower levels. 200 LAKE BONNEVILLE. Tlie unarmed waves not only were unable to tear down the cliff, l)ut were compelled by their peculiar chemical constitution to add a iiiiiici-al cdatino- to its face. These lines of tufa are all covered l»y tli(! lacustrine deposits except where exposed by recent denudation; and it is assumed that certain of them now buried by the White Marl l)eds were formed durin<>- tin- d('])o- sition of some portion of the Yellow Clay. The ]iiji,liest of tliesc; is se])arated from the Bonneville shore-line by an interspace of 90 feet (aneroid mea- surement). THE WHITENESS OF THE WHITE MARL. As soon as the wide distribution of the White Marl and the Yellow Clay and the constancy of their contrast came to be appreciated, attention was directed to the determination of the cause of their difference. It is easy to luiderstand a gradation in texture and composition of strata as one passes from the margin of a, l)asin toward its center, or from the vicinity of sea- clitfs and river mouths, where the supply of detritus is great, to quieter and remoter places, reached only by sediment long held in suspension; but it is not so easy to understand why there should be an abrupt change in the sedimentary sequence throughout an entire basin. If the true explanation of the difference between these strata can be reached, it should contribute something to the history of the lake. For the purpose of seeking such an explanation, the character of the two deposits has been examined Ixith chem- icallv and microscopically. Two samples each of the White ]\Iarl and Yel- low Clay were analyzed by Prof. 0. D. Allen of New Haven, with the results exhibited in Table III. CHEMICAL COMPOSITION OF THE CLAY AND THE MARL. 201 Table III. — Jnali/ses of Bonneville Setii, inents. I. White Marl from the Ohl River Beii. n. White M.-111 IVo.u ne.ir Willow SpriuR, .it the eastern h.i30 of the Deep Creek Mouutiins. III. Upper part of Yellow Clay, Old Kiver Bed. IV. Lower part of Yellow Clay. Ohl River lied. lusoluble; percentage Soluble; percentage 100 parts of the Insoluble portion coutain- Silica Aluiuiua Ferric o.xide Potaaaa Soda Lime Magnesia Carbon dioxide Water . 100 parts of the determined; Soluble constitueut.s eoutaiu- Sulphiiric oxide Lime ■ Magnesium Potash Soda Sodium oxide** Chlorine Nitric acid** Boric acid- Carbonic acid Lithium 4.5. 03 8.03 2.«5 1.70 .68 19.08 2.71 16.25 2.33 Oxygen equivalent to chlorine. 90.32 23. 539 .916 1 U« .534 47. 039 96.84 3.16 23. 05 3.20 1.10 .70 .54 3 i. 08 2.87 31.49 1.23 33. 857 tr'ace trace trace 20. 204 8.9i;6 .721 1.363 39. 295 Probable couibination of soluble ciuiaJiruents- Calcium snlphato Magnesium sulphate Pota-siuni sulphate Sodium sulphate Calcium chloride Magnesium chloride Potassium chloride Sodium chloride Sodium oxide** 107. 633 7.633 100. 000 2.225 3.444 .987 36. 079 38. 029 trace III. 0.71 43.84 13.85 4.04 2.40 .44 12.43 4.54 11.88 2.84* 4.111 lY. 100.43 8.806 2.341 5.980 1.792 50. 742 95. 57 4.43 41.74 13.00 3.61 1.87 .70 16.01 4.96 15.78 3.78 100.45 39. 169 present 2.045 4.322 1.897 .370 50. 637 52. 594 Iiresent 108.578 1 8. .578 I 108. 836 8.836 100. 000 I 100. 000 56. 789 1.476 100. 000 21. 775 2.163 2.521 8. .51 1 5.685 8.193 62. 659 2.371 111.865 11.865 100. 000 7.729 2.835 52. 830 22. 728 100. UOO 5.727 3.759 .586 76. 336 10.115 100. 000 * Water lost at 100° C. t Water lost liy ignition J The total wei-bt chloride would be th tions of the solution. stitueuts were delerm por- con- ♦.„ "The sodium oxide reported among the constituents" is not ass'imed to be free, but to exist as sodium oitrntp *:i^Ll ",' yi':?."j'_.'T:'» f^^'V" e.'ol' '"Stance; and in the c;,se of the thir.l and fourth' santnles it"anto,,nt' L' c ' »; t^'f: ^^ii^^Ss^^kfsr- '^-'^ --'^^^ '*-«^^- ^^^>^^^^^:^\:'iS'o} ^^:^ 202 LAKE BONNEVILLE. The soluble constituents need not concta-n us at present, for they do not materially affect the color of the beds. Indeed the characteristic colors are everywhere recognized by the weathered- surfaces, from which the solu- ble materials are nearly or completely leached. Tlie carbonic acid in each of the samples is nearly sufficient to sati.sfy the lime and magnesia; and it maybe assumed to have been all combined with tliose bases. The alumina, iron, soda, and the remaining lime and magnesia, undoubtedly exist in the form of silicates, while the unsatisfied silica is free. The microscopic characters indicate that the silicates are chiefly feldspars; and if we assume orthoclase to be predominate, the bases are barely satisfied in the case of one sample and there i^ an excess of silica in each of the others. It is prob- able that the following table represents the constitution of the earths nearly enough for the purposes of the present discussion. Table IV.— Condensed Results of Analyses in Table III. Sample Sample II. Sample III. Sample White Marl: Mean uf 1 and II. Yellow Clay: Mean of III and IV. Carbonates of lime and magnesia- - Silicates Per cent. 36 54 10 Per cent. 70 18 12 Per cent. 26 74 0 Percent. 34 62 * 1 Per cent. 53 36 n Per cent. 30 68 2 Free Silica , Totals 100 100 100 100 100 iOO Under the microscope the White Marl is seen to contain, first, numerous minute crystals exhibiting double refraction; second, minute particles, ap- parently clastic, likewise doubly refracting; third, siliceous organisms. The crystals are too snuill for meiisureinent. They appear in gencrMl to be taj)ering pyramids whose longer diameters are three or four times their shorter. They undoubtedly represent the carbonates. The clastic matter is conceived to represent, in like manner, the silicates, and possil)h- ;i portion of the free silica. The remainder of the silica, or })ossil)ly th(^ whole of it, is contained in tlie microscopic organisms. These are ])artly diatoinaceous, but include also numerous slender tubes witli punctate or jJiipilkite walls which may be spiculae of sponges. Unfortunately, a majority of the samples of Yellow Clay which should have been examined for comparative purposes, were lost in transportation CARBONATES VERSUS SILICATES. 203 before tlio microscope was applied to them The only two })reserved are from a subhttoral deposit at Lemingtou and from the type section in the Old River Bed. These exhibit only rounded grains of crystalline matter, for the most part clear, uncolored, and doubly refracting-. Neither diatoms nor cr}'stals were discovered. In brief, the White Marl and Yellow Clay resemble each other in com- position, but the former is characterized by a relatively great amount of earthy carbonates and by free silica, while in the latter the argillaceous element predominates. In the former the carbonates were largely thrown down as a chemical precipitate, and St least a portion of the silica is an organic precipitate. The whiteness of the marl appears to be largely due to its precipitated elements. These differences in the characters of the two deposits were unques- tionably determined by some event in the history of the lake; during the intervening epoch of low water the conditions of sedimentation underwent some change. A double interest attaches to the determination f)f the nature of this change; on the one hand its discovery would add an element to the history of the lake; and on the other it might lead to the establishment of some law of sedimentation hitherto unrecognized. Much thought has there- fore been given to the subject, hypotheses have been framed and many experiments have been made, but the results of the experiments are unfor- tunately negative, and the j)i'o]>lem can not be regarded as solvcil. It is necessary, however, to give some consideration in this place to certain of the hypotheses for the ])urpose of showing the grounds ujjon which one of them was so seriously entertained as to receive a provisional jiublication. SOURCE OF MATERIAL. The simplest explanation of the change in sedimentation is that the nature of the material supplied to the lake by tributary streams was for some reason different. In the interval of time between the two epochs of deposi- tion, the deformation of the earth's cnist may have wrought changes in the area of the basin, either cutting off some important element of the detritid contribution or making some equally inqiortant addition. The prime diffi- 204 LAKE BONNEVILLE. culty with tliis hypothesis is that the configuration of the region offers no way of rendering it h)cal and concrete. The calcareous tribute of the basin must flow chiefly from the limestones of the Wasatch and associated ranges, and the drainage system by which it is conveyed seems to have been estab- lished before the Pleistocene. The possibility of an ancient modification in the drainage system of the Bear River will be discussed in the next chapter; but such modification, if it occurred, can not have had so late a date as the epoch of the White Marl. COMPOSITION OF LAKE WATER. A second explanation is that the conditions of sedimentation and pre- cipitation in the basin were ixiodified after the epoch of the Yellow Clay by a change in the mineral contents of the water of the lake. It is well known that the precipitation of certain substances from solution is favored by the presence of certain other substances, and by yet others is retarded. It is equally well known that the fall of minute suspended particles is similarly accelerated by the presence of various substances; and their fall is probably retarded by other substances. Is there any ground for postulating a change in the mineral contents of the lake which would account for the observed change in the natui'e of the deposit? There are three different changes of this sort readily conceived. First, the water having been relatively pure during the deposition of the Yellow Clay, it may have acquired, during the interval of recession, a large amount of mineral matter, so as to be a brine at the time of its second flooding. Second, the water of tlie first great lake, having been a feeble brine, may have become so concentrated during the epoch of low water as to precipi- tate its less hygroscopic minerals, with the result that, when the second fiood came, a mother liquor was diluted instead of the normal brine. Third, tlie water of the first great lake, having been a feeble brine, may have been in the interval not merely concentrated but completely evaporated, the desic- cation product being mingled with and buried by mechanical sediments, so as not to be redissolved at the time of the second flood. On the first sup- position, the White Marl epoch was characterized by a stronger brine than the Yellow Clay epoch. On the second, it was characterized by the min- DID THE WATER CnANGE IN COMPOSITION? 205 erals pecular to mother liquors. On the third, it was characterized by purer water. Each of these postulated changes may be supposed to have acted in either of two ways; first, the })eculiar })roperties of the menstruum of the second flood may have caused the precipitation of an exceptionally large proportion of the calcareous matter in the center of the basin, and may have determined the assumption of the crystalline form; second, its prop- erties may have determined the precipitation of argillaceous sediment near the shore, thereby diminishing its importance in the center of the basin and thus increasing the relative percentage of calcareous matter. No at- tempt has been made to test the first of these assumptions experimentally, for the reason tliat the natural reactions could not be fairly represented by the necessarily rapid processes of the laboratory. It may be said, also, that the assumption is less accordant with what is known of the distribution of calcareous matter in the basin. From the second point of view a series of experiments was instituted, the investigation being conducted by my assist- ant, Mr. I. C. Russell. Experiments.-In tlic couduct of thcsc experiments no attempt was made to discuss the general problem of the properties of dissolved substances as the precipitants of sediments, but attention was confined, to the specific problem presented by the lake sediments. With the excej)tion of distilled water, the only materials used were those which occur in the basin and are concerned with the practical problem. The brine of Great Salt Lake in various stages of dilution was assumed to represent the water of Lake Bonneville, the diluent being in each case the approximately fresh water of some stream now tril)utary to Great Salt Lake and anciently tributary to Lake Bonneville. The fine sediment employed was a sample of the Yellow Clay. The water of the selected stream was mixed in various proportions with the Ijrine, and ecpial quantities of the mixtures were arranged in a series of similar vessels, tlie pure stream water and pure brine constituting the first and last terms of the series. Equal portions of the finely divided clay were then added to each vessel and mingled with the water by shaking or stirring, after which the vessels were allowed to stand for several days and notes were made of the relative rates of precipitation. 206 LAKE BONNEVILLE. The first, streuin water einj)l()yetl was that of City Creek, the sample' being' arative experiments witli distilled water. Salt Lake brine and distilled water ai;ree in retaininrj a residuary milkiuess for an indefinite period, but the approximate clearing of the brine is by far the more rapid. 'Second Ann. Kept. U. S. Geol. Survey, pp. 177-180 EXPERIMENTS IN SEDIMENTATION. 207 and while the experiments with Bear River water do not negative the theory broached in the i)reliniinary pubUcation, they serionwly weaken its snpport. It is a cnrions fact that the City Creek and Utah Lake waters, having simihxr jn'operties as precipitants, yet diflfer widely in their mineral constit- uents ; and that the water of Bear River, while behaving very difterently as a precipitant, yet closely resembles in constitution that of City Creek. The accompanying table of analyses (Table V.) shows that the water of Utah Lake is characterized by the sulphate of lime, while the waters of City Creek and Bear River are characterized by the carbonate. Table V. Mineral Contents of Fresh Waters in the Salt Lake Basin. I. Water of City Creek, taken at head of Main .Street, Salt Lake City, December 3(1, 1883. II. Water of Bear Hirer, taken at Evan.ston, Wyoming. III. Water of Utah Lake, taken Deeember, 1883. [t. analyzeil -jy T. M. Chatardi 11 anil III, by F. VV. Clarke.] Grams to tbe litre. Per cent, of total solids. I. n. III. I. II. III. Calcium .0589 .0174 .0091 . 1280 .0070 .0131 .0010 .0090 . 0432 . 0125 . 0082 . 0982* .0105 .0049 . 0.-.58 .OIKC .0178 . 0008* . 1300 .0124 24.19 7.15 3.74 52. 57 2.87 5.38 0.41 3.69 23.41 6.78 4.44 53.24* 5.09 2.65 18.24 6.08 5.81 19. 88* 42.68 4.04 Snlplmric Acid Silica .0070 .0100 3.79 3.27 .2435 .1845 .3060 100. 00 100. 00 100.00 PEOBABLE COMBINATION. Calcium Carbonate Magnesium Carbonate Sodium Carbonate . 1400 .0606 .0014 .0099 .1080 .0438 .0038 .0644 .0204 .1849 .0204 .0100 57.49 24.88 0.57 4.07 8.87 0.42 3.70 59.20 24.01 1.25 21.19 6.71 60.84 6.71 .0135 .0081 .0070 8.48 4.49 Sodium Chloride .0216 .0010 .0090 Silica 3.82 3.30 .2435 . 1824 . 3039 100. 00 100. Oil 100.00 "Estimated by difference. The postulate that the second flood diluted a brine which by fractional preci})itation had accjuired the character of a* mother liquor, was tested in the following manner: Samples of the brine of Great Salt Lake were evap- orated until various portions of the saline contents had been precipitated, 208 LIKE BONNEVILLE. and tlic residuiu-)- licjuurs were tlien diluted with distilled water and coni- pared with similar dilutions of the Salt Lake l)riiie. It was found tluit sedi- ment sejjarated with eijual rapidity from the lirine juid the motlicr liiiuors; and parallel results were obtained tVom their corresponding derivatives. The only one, then, of the alternative hyj)otheses suggested above which tinds any support in the experimental results is the one of wliich pub- lication has l)een already made, and the support accorded it is insutticient to inspire confidence. If the water of Bear River instead of City Creek had l)een first subjected to experiment, the theory would have been at once abandoned. Nevertheless, since it is not controverted by the experiments, and since it has practically no competitor, it is proper that its relation to the general question of lake history be fully set forth. DEPOSITION BY DESICCATION. Fully stated, it takes the following form. During the first rise of the lake, or at least during that part of it represented by the visible portion of the Yellow Clay, the saline matter was held in solution in such })roportioii that tlie precipitation of mechanical sediment was slow. The clay intro- duced by the streams and by the undertow remained in suspension a long time, and was therefore widely distributed, covering the whole Ixtttom of the liasin. At tlie close of the Yellow Clay epoch the liasin was completely desiccated, the saline matter l)eing gathered in the lowest depression and there precipitated. The raiiifiill of the basin, however, 3. POST-BOi^NEVlLLE JOINTS. 213 deposition, they can not have resulted from horizontal pressure and com- pression. I had no explanation to offer, but my inquiry led to the ]jidjlica- tion of one so accordant with the phenomena that it at once takes rank as the working hypothesis for the origin of all ])arallel jointing except slaty cleavage. It was offered independently by Crosby^ and Walling,^ and the force appealed to is the earthquake. During the passage of an earthquake wave the earth material traversed is subjected to momentary strains of com- pression and tension in the direction of wave transmission, and to shearing strains, instantly reversed, in a direction normal to that of wave transmission. At each instant the similar elements of the wave constitute a surface approx- imately spherical or ellipsoidal, with the locus of wave origin at its center, and at any locality remote from the locus of origin such surface is sensibly a vertical plane. Assuming the competence of the strains to create a rock structure, their directions and arrangement show that the structure should ordinarily exhibit vertical parallel planes. Under this theory the two series of joints at the Old River Bed indicate two earthquake directions and at least two efficient earthquakes. As the joints extend as simple regular planes to the very margin of the old channel, and as they determine the directions of arroyos initiated immediately after the excavation of tlie channel, it is probable that they were formed while the lake sediments were yet continuous and unchanneled. We are thus told of eartliquakes occurring just before the retreat of the lake laid bai'e the White Marl. That the Bonneville Basin was subject in Bonneville and post-Bonne- ville time to numerous earthquakes of the type of the great Californiau earthqiiake of 1872, is abundantly shown by the phenomena of fault scarps described in Chapter VIII; and the distribution of the fault scarps, so far a.s it is known, accords well with the strike of the principal system of joints. ' On the classification and origin of joint-structure. By W. O. Crosby. Proc. Boston Soc. Nat. Hist. vol. 22, 1882, pp. 72-85. *0n the origin of joint cracks. By H. F. Walling. Am. Ass. Adv. Sci. vol. 31, Montreal meeting, 1882, p. 417. CHAPTER VI. THE HISTORY OF THE BONNEVILLE BASIN. THE PRE-BONNEVIIiliE HISTORY. The latest Tertiary series outcr()})ping within the Bonneville basin has a distribution quite independent of the basin. Not only do its strata occiir in the mountains above the shore-lines, but they override some of the passes on the rim of the hydi'Ographic basin .and extend continuously to the di'ain- age of the Snake River, and possibly to that of the Humboldt. On the other hand, the Neocene strata have not been found in the southern third of the Bonneville area. It is probable, therefore, that the hydrography of the Neocene and that of the Pleistocene corresponded to configurations of the surface essentially different. The Bonneville Basin was not m existence during the period when the Neocene sediments were deposited; its history began at some later date, after the deformation of the earth's crust which elevated the Neocene strata uj)on the mountain flanks had wrought im- portant changes in the face of the land. The area formerly covered by the main body of Lake Bonneville is now a plain, conspicuous for its flatness. Great Salt Lake, i-esting on its surface, has a mean depth of but fifteen feet; and a rise of a few feet only, as pointed out by Stansbury, would extend it westward over the greater portion of what is known as the Great Salt Lake Desert. The (K^currcnce of such a plain at an elevation of 4000 feet above the sea, and in the midst of a region characterized by mountains, admits of but one explanation, namely, lacustrine sedimentation. Th(^ narrow ridges tliat in places inter- rupt tlie continuity of tlie plain sliow tliat tlie district did not escape the general process of erogenic corrugation to wliich tlie (Jreat Basin was sub- 214 THE FLATNESS OF THE DESERT. 215 jected, and there seems no reason to believe that the disphiceraents were here less profound than elsewhere. Certainly the degradation of the sum- mits has been sufficient to lay bare in places Cambrian and even Archean i-ocks. Moreover, the habit of these ridges is peculiar, and itself indicates burial. The normal mountain ridge of the Great Basin is acutely serrate along its crest, and disjjlays naked rock, deeply carved into gorges and amphitheaters down to a certain line. Below that line the slopes are gentler, the contours are smooth, and the material is alluvial, the waste from the sculpture above. The gorges above and the alluvial cones below are to a certain extent correlative, but the mass of the latter is derived from the general degradation of the mountain summit as well as the excavation of the canyons. The mountains and buttes of the Salt Lake Desert conform to the Great Basin type in the characters of their summits, but are almost devoid of alluvial cones. They spring from the plain so abruptly that the frontiersman as well as the geologist has I'ecognized them as incomplete, or rather, as partially submerged, and has named them accordingly. One of them is known as Newfoundland, another as Silver Islet, a third, which towers 3,000 feet above its base, as Granite Rock; and geuerically they are spoken of as "lost mountains". How deep beneath the lacustrine ))lain their bases lie, it is impossible to say, but 2,000 feet is certainly a moderate estimate. Not all of this lacustrine filling can be ascribed to the Pleistocene, and not all of it belongs to tlie history of the Bonneville Basin as such. The Neocene lake, and possildy earlier lakes, have contributed a sliai-e, and this before the hydrographic basin of Lake Bonneville was established. Since the establishment of the basin, sedimentation has been practically continuous in its lowest depression. If we conceive the local climate to have under- gone a rliythmic series of clianges, the area of lacustrine sedimentation lias alternately expanded and coiitracted, and lias always iiududed tlie lowest depression; and even witli a climate so dry as to maintain no })ereimial lake, the temporary floods occasioned by exceptional storms must still have continued the process of accumulation. The situation of the lowest dei)res- sion may have varied from time to time, as local displacements of the earth's crust modified the configuration, but wherever it was, it was the scene of 216 LAKE BONNEVILLE. sedimentation, and the constant tendency of tlie lacustrine process was to fill the minor depressions and reduce the floor of the basin to a level surface. The evenness of the desert plain testifies to its lacustrine origin. Tlic process of filling' might have been modified, but would not have been interru})ted, by an overflow of the water of the liasin such as occui-red in the Bonneville epoch. As long as the basin was not drained to its lowest depths, those depths would continue to receive detrital deposits, and the out- flowing water would carry with it only the soluble products of the degra- dation of the surface of the basin. Whether such an overflow ever took place is not apparent; but if it did, we may l)e sure that its date was remote as compared to the Bonneville epoch. The lower passes of tlu^ l>asin's rim show no traces of an ancient channel, and the time necessary for the efface- ment of such traces must be reckoned as long in comparison to the antiquity of the Bonneville shore-lines. Upon most of the passes the process would include the growth of great alluvial fans; and at Red Rock Pass, where the Bonneville discharge took place, the record of an earlier discharge could ha^•e been obliterated only by the restoration of the Marsh Creek alluvial fan, and its extension so as to fill the channel of outflow for many miles in Marsh Creek Valley. When we consider that no stream so small as Marsh Creek is known to have built a delta on either the Bonneville or the Provo shore, it becomes evident that such obliteration implies a period vastly longer tlian that consumed by the Bonneville oscillations. As far back, then, as we may hope to obtain a consecutive view of the history of the basin, its waters had no period < )f discharge save that of the Bonneville epoch. It was a closed basin, and the area of its lake surface was determined by the relation be- tween its water supjjly and the rate of evaporation. Tlie lake area was, therefore, a function of climate, provided the extent of the hydrogra])hic basin remained unchanged. To avoid any possible misinterpretation of the climatic historj' it is important that the possibility of variation in the hydrographic basin receive full attention. The general altitude of the country to the east of the basin is several thousand feet greater than that to the west, north and south, and at least 95 per cent, of all the water flowing into the modern lakes is furnished by the eastern highlands. These include the Wasatch Mountains, a portion of the ANTIQUITY OF THE BONNEVILLE BASIN. 217 High Plateaus lying to the south, a portion of the Uinta Mountains lying to the east, and a mountainous tract lying to tlie northeast in western Wyom- ing and southeastern Idaho. Tlie low country to the west of the basin is di\idcd l)y mountain ranges into numerous independent drainage districts, and these have not been so thoroughly studied as to determine what would be their hydrographic combinations in the event of a more generous rainfall. We know, however, that they contribute nothing now to the water sup[)ly of the basin, and that in Bonneville times their tribute was small; and we are thus assured that in pre-Bonneville times the supply from that side was not less than at present. It will be shown hereafter that the possibility of a greater contrilnition from this region does not materially affect the conclu- sions in regard to climate. The same remarks apply to the region south of the Escalante Desert. North of the Bonneville Basin the conligurntion of the country about the water parting does not suggest any possible change in its position during the period under consideration. The water supply from the east reaches the lower portions of the basin by four rivers: the Sevier, the Jordan, the Weber, and the Bear; and its drainage system is correspondingly divided into four j)arts. The Sevier River rises in what Button has called the High Plateaus, and is separated by high divides from the drainage tif the Fremont, tlu- Escalante, the Paria, and the Virgen, branches of the Colorado of the West. The Paunsagunt and tlie ^larkagunt })lateaus, which constitute the most southerly elements of its drainage, are slowly diminishing in area through the sapping and recession of cliffs, and the hydrograjjhic basins of the Paria and Virgen are thus grow- ing at the expense of the Sevier. A less considerable change of the opposite tendency is in ])rogress at the head of Moraine Valley, where a plateau draining to the Fremont River is encroached on by the recession of cliffs draining to the Sevier. The effect of these slow changes upon the water sujiply of the Bonneville Basin can not have been important, and there is no evidence that any considerable tracts have bodily transferred their allegiance. The Jordan includes among its branches the American Fork, the Provo, the Spanish Fork, and Salt Creek. It is quite possible that Salt Creek has changed its course within the basin, and that it was at one time connected with the Sevier and not per- 218 LAKE BONNEVILLE. manentlv with tlie Jordiin, ])ut ^uch ;i cliaiige is of no nioiiR'nt in this con- nection. American Fork and Sjnini.sli Fork head a<^ainst liigh divides, who.se po.sition must liave l)een permanent for a hmg period. Tlie same remark ai)pHes to thc^ I'rovo River, but there is one point in its course where its chainiel is not contained by soHd rock and its water could easily be diverted. Kamas Prairie is a .small valley lying- athwart the western end of the Finta Rauffe. The Provo River crosses the southern end of the vallev, enterinjj by one canyon and leaving by another; and the Weber Riv6r in like manner crosses its northern end. The configuration of the plain .shows that the streams have not always been separate; at one time the Provo turned northward in the valley and was tributary to the Weber. Here, however, as in the case of Halt Creek, the modifications of the drainage do not affect tlio water supply of the Bonneville Basin. The drainage district of the Weber is so nearly embraced at the east by the basins of the Bear and the Jordan, that the only portion of its boundary coincident with that of the Bonneville drainage district is a high crest in the Uinta ^fountains two or three miles in length. Variations in its course and drainage area are therefore unimportant to the present discus- sion; and the same remai-k applies to the American Fork and to the series of creeks issuing from the west face of the Wasatch ]\Ioinitains. The Bear is the most important of all the rivers, and has many tribu- taries. Its main brandi lieads in the Uinta Mountains, and, .so far as may l)e judged from tlu^ maj)s of th(^ Fortieth Parallel Survey, is surrounded by high divides, affonling little opportunity for tran.snmtations of drainage. Smith F(jrk and Tliomas Fork, which join it in midcourse, occupy ])asins contiguous to those of Salt River and John Day River, tributaries to the Snake. These basins have been mapped by the Geological Survey of the Territories, and the testimony of the contours is sustained by that of Mr. Henry Gaimett, who perfonned the topographic work and who states that the conformation indicates permanence of drahiage. In its lower course (in ( "aclie Valley) the river receives a large innnl)cr of tributaries, l)ut mtne of their drainage districts extend to the rim of the Bonneville Basin. The sources of the river appear thus to offer no suggestion of an ancient variation of the drainage area; but there is one point in its course of which the same POSSIBLE CHANGES OF CATCHMENT AREA. 219 can not be said. After receiving the waters of Smith Fork and Thomas Fork, and before entering Cache Valley, the river swings far to the north, apiiroaching very near to the rim of the Bonneville Basin. At Soda Springs it is separated from tlie sonth fork of the lilackfoot River, a l)ninc]i of the Snake, hy a divide rising fonr or five hnndred feet above the Bear, bnt only slightly elevated above the Blackfoot. A few miles lower down it crosses the sonthern end of a broad open valley (IJ5asalt Valley), the northern end of wliieli is traversed by the Portnenf River, likewise a branch of the Snake. The Portnenf is here the lower stream, and the water parting between the two rivers runs close to the course of tlie Bear. It is probably not more than one or two liundred feet above the bed of the Bear. In the Soda Springs pass, the sumniit is ftirmcd liy ])asalt, l}ing in horizontal sheets and associated with cinder cones and other evidence of recent eruption. The princii)al masses are jjrobably more ancient tliiiu tlie Bonneville epoch, l)ut they have not suffered those dislocations which are apt to be observed in this region in the case of rocks dating far back in the Tertiary. It is believed by Mr. Gannett and by ^\r. Gilbert Thomj)son that their eruption has affected the drainage system of the region in ways that are yet discernible, and it is possible that they have wrought a separation of the Blackfoot and the Bear. If the two streams were anciently united, it is most probable that the Blackfoot was tril)utary to the Bear; but the reverse is possible. At the Basalt Valley i)ass the phenomena are essentially the same. The broad valley extending from the chaimcl oi' tlic Bear to that of the Portnenf is covered throughout by basaltic la\a, and portions of this lava are so recent that associated scoriaceous craters are still preserved. Befoi-e the epoch of eruption, the Bear and Portnenf Rivers may have been joined, and their united water may have flowed either to the Snake River or to the Bonneville Basin. If the south fork of the Blackfoot were uo\\ to be diverted to the valley of the Bear River, as, according to Mr. Thompson, it readily might be, the Salt Lake drainage basin would be increased by 350 square miles of upland. If the canyon of the Portnenf below Basalt Valley were dammed, so as to turn its water toward Bear River, 500 square miles would be added to the basin. If another eriq)tion were to dam Bear River aljove Gentile Valley 220 LAKE BONNEVILLE. and divert it to tlie v;vlloy of the Portnenf, the Bonneville Basin would lose about one-fourth of its water supply. All speculation in rejj^ard to the pre- Bonneville climate of the l)asin is therefore subject to the possi])ility that the catchment basin niay on the one hand have been sliyhtly greater or may on the other have been very materially less. ALLUVIAL CONES AND ARIDITY. The principnl evidence bearing on the pre-Bonneville history of the l)asin is embodied in the alluvial cones. These extend nearly to the l^ottom of the basin, and since they could not have been shaped in the pi*esence of a large lake, it is concluded that the epoch of their formation was an epocii of low water. The dependent conclusion that the pre-Bonneville ei)ocli was characterized by aridity is of such importance that a little space Avill be de^•oted to the amplification of these propositions. 'V\n' drainage of a mountain mass, starting in innumerable rills, gathers into a smaller number of rivulets, and is finally aggregated into a verv few main streams before issuing from its self-carved gorges. The outward borne detritus is therefore delivered to the adjacent valley at a limited number of points separated by interspaces. Each point of issue becomes the apex of a sloping mass of alluvium whose surface inclines equably in all directions. A series of such alluvial cones is usually to be found along the base of each mountain range, constituting a foot slope, the contours of which are scalloped. The topographic configuration which thus arises is peculiar and not liable to be confounded with any other. It has already been stated that the alluvial bases of the insular mount- ains of the Bonneville Basin are buried by lacustrine sediments. Those of the peripheral mountains are not so buried, or at least are not so dee})ly buried, and the forms of their cones can at many localities be traced down- ward to the lower levels of the basin. The shore-lines are locall\- mai-kcd upon the cones, cliffs and terraces being excavated from them and ciuhaiik- ments built against them; but where the cones are large, these modifica- tions are relativelv small ;ind do not materialh- impair the general con- figuration. Good illustrations are to be found in i'reuss Valley, in White Valley, at the eastern base o( the Deep Creek and Gosiute Mountains, and DRY CLIMATE BEFORE LAKE EPOCH. 221 on 1 )()th sides of Pilot Creek. There are fine examples also in Tooele Valley, Skull Valley and Blue Creek Valley. The phenomena of the Bonneville shores illustrate the fact that the buildinji- of alluvial cones is arrested by lacustrine conditions. p]ither the stream constructs a delta, which is an alluvial fan above the water but ter- minates in a submerged cliff at the water edye; or else, the stream being small, its load of detritus is absorbed by the shore drift. In the latter case, some point of the alluvial cone is usually trenched on by the waves, a cliff and terrace being jjroduced; and whenever the stream, which had ])reviously shifted its course over the whole surfiice of the cone, assumes a direction leading to this cliff, it is enabled })y the lowering of its l^ase level to exca- vate a more permanent channel, from which it does not quickly escape. It is therefore leg-itimate to reg'ard the formation of alluvial cones as a stiictly subaerial process, and to conclude that the Bonne\ille Basin con- tained no large lake during the pre-Bonneville period when its alluvial cones were formed. I do not overlook the possibility that traces of an epoch Avlien the waves held sway may have been obliterated by the alluviation of a later epoch, but in my judgment such considerations do not impair the general conclusion. Within the masses of the alluvial cones there niay be liuried shore cliffs, shore embankments, and lacustrine sediments, but the time necessary for the oliliteration in this nuiuner of a record similar to that of the Bonneville lake is as long as tlie time necessary for the obliteration of a channel of outflow, and is certainly very long as compared to the dura- tion of the Bonneville epoch. Let us call this relatively long epoch antecedent to the Bonneville, the pre-Bonneville epoch. We have found reason to believe, first, that the basin had then no outlet, and, second, that the basin did not then contain a large lake. The size of an inclosed lake being determined l)y the ratio of water supply to rate of evaporation, it follows that that ratio was small. If the hydrographic area remained unchanged, the water supply as well as tlie rate of evaporation depended upon climate, and tlie climate must have been arid. If the main branch of Bear River was then tributary to the Portneuf Basin instead of the Bonneville, a greater climatic change would have been 222 LAKE BONNEVILLE. necessary to flood the basin, and the hidicated aridity of climate is corre- spondingly less. THE POST-BONNEVILLB HISTORY. l^lie closiu'i' event of the Hoiiiicn illc liist(ii'\' \v;is the (Icsiccatioii uf the basin. \ ft'W stiion the route near Uie :?Stli and li'.Uli parallels, explored by Capt. J. W. Gnnnison : I'acilie lvailroare, 2 inches. 5. Fine sand containing fresh water shells, 6 inches. 6. Gray clay. MON I 15 226 LAKE BONNEVILLE. Section at margin of Sevier Lake salt bed, Anguat 20, 1880. 1. (Top). Sixliiim chloriilo, foriiiiiij; a colicroiit crust : i inoli. '.;. Soiliiiin chloride, with Hodiiim 8uli)hato and magnesium sulpbato; free crystals luiugled with water: li iuclies. :!. Sodium siilph.ate, witli sodium chloride ; a crust of coherent crystals: i inch. 4. Sodium chloride, with magnesium sulphate; incoherent crystals mingled with water: IJ inches. C>. Sodium chloride, with sodium sulphate, cliemii-ally identical with No. 2 hut (ine-Kraim-d and with the consistence of an ooze; color white above with occasional passages of pink, green heueath: i inch. (). Dark gray mud : 2 feet. Tlie sulijoined table of" analyses exlii1)its in detail the constitution of the saline de})osits in each section, and the composition of the original brine is added for comparison. The con- spicuous fact is that the sodium sul- pliate is concentrated in the middle (if the basin, while the sodium chlo- ride is chiefly deposited at the mar- gin. The sulphates of magnesium and potassium likewise occur exclu- sively at the margin. It is note- \\(»rthy also that magnesium is re- ported in larger proportion in the Ijrine of the lake than in any layer of the desiccation products at either point of determination. The mag- nesium chloride reported in the brine implies three per cent, of magnesium. The magnesium sulphate in the richest layer of the desiccation prod- FlG. 31.— Sevi.r L,iko in 1873 (Nell). The white areas uct impHeS Oldy 1.7 per Ceut. of with dotted biniuduriu.s show ealt bods in 1880 (JoIidbou). magnesium. The brine of the lake was analyzed by Dr. Oscar Loew; the desicca- tion products from the center of the area by Prof S. A. Lattimore; those fit mi tlie margin by Prof O. D. Allen. The brine contained 8.G4 per cent, of saline matter; the constituents are here reported in percentages of total solid matter. The constituents of the desiccation products are likewise DRYING OF SEVIEK LAKE. 227 reported in percentages. The figures for tlu^ total deposit are obtained by combining tliose of tlie separate layers, making allowani'e for n^lative thickness. A few weeks after our oljservation of tlie salt bed, Mr. Uusscll and I separately visited the southern portion er. Second layer. Third layer. Fourth layer. Fifth layer. Total. Sodinm Sulphate Sodium Carbonate Sodinm Chloride Calcium Sulphate Ma^nesiun) Sulphate .. Ma^ueBium Clilorido .. 87.65 1.08 2.34 trace trace 71.23 89.10 84.6 .4 7.0 4.78 5.51 83.79 2.71 5.04 14.3 15.5 23.86 trace trace 2.65 91.39 trace 1.83 79.86 7.83 13.84 trace 1.33 88.49 5.29 80.62 .39- 8.32 75.8 5.5 72.1 .5 11.9 Potassium Sulphate . . . trace .34 .26 .11 4.03 trace trace .92 .68 .7 Boric Acid Water 8.90 trace 4.90 trace 8.20 trace 8.0 2.00 6.46 .78 3.40 3.6 .1 Total ... 99.97 99.08 99.95 100.0 100, 00 100.00 100. 00 100.00 100. 00 100. 00 100. 00 The desiccation of this lake is to be ascribed to human agency. The water of its sole tributary flows for nearly 200 miles through valleys con- taining more or less arable land, and has gradually been monopolized by the agriculturist for the purpose of irrigation. The supply is however not completely cut off. It is reported that during the spring freshets, caused by the melting of the snow on the plateaus and mountains, the lake bottom receives considerable inflow, and that the desiccated condition obtains dur- ing only a portion of the year. The principal salt deposit was estimated to extend eight miles north and south and to have an extreme width of about five miles. The accom- 228 LAKE BONNEVILLE. panying sketch shows the form and area of the lake iu 1872 and the approximate extent and position of the salt beds iu 1880. RUSH LAKE. The lowest depression of Rush Valley contains a pond or lakelet which has been observed to undergo considerable fluctuation. It will be recalled that Rush Valley in pre-Bonneville time drained freely to Tooele Valley and that this drainage was cut off by an embankment built by the Bonne- ville waves. The lake occupies a portion of the old drainage channel close to the embankment. It is partially delineated in the map on PI. XX. The earliest record of it appears on Stansbury's map (1850)/ but it is not men- tioned in his text. It is there assigned a length of about 1^ miles, but there is circumstantial evidence that no measurement was made. In 1 855 it was included in a military reservation laid out by Lieut. Col. E. J. Steptoe for the purpose of securing to the military post at Camp Floyd the meadow and pasturage about the lake shore. The map made for the purpose of defining the reservation, assigned to the lake a length of 2| miles, and indicated that the water was shallow and marshy. The land surveys in the valley in 1856 did not include the military reservation, but showed the existence upon it of a lake. According to Gen. P. E. Connor, who succeeded Col. Steptoe in 18G2, there was then only a small pond, the remainder of the lake bed being occupied by meadow land. In 18G5 the water began to increase, the greatest height being attained in 187G or 1877, since which time it has sub- sided. The rise of the water submerged the meadow land and rendered the reservation useless for its original purpose. It was therefore ofticially relinf[uished by the War Department in 1869. In 1872, the water being near its highest stage, the lake was surveyed in connection with the surrounding country by one of the parties of the Wheeler Survey, and the length was determined to be 4^ miles. In 1880, Avhen the lake was visited by the writer, it was said l)y residents to have shrunken to half its ninximuni size. The position of the highest ' Expl. ami Siirv. V;i]lcut a few inches, on the northern slop<^ of a, small limestone knoll just south of the railroad track at Black Rock. Its top is dressed square, al)out 10 by 10 inches, and is marked with a -f . A sketch-map (PI. LI) was made of the locality in 1877, at the time of the establishment of the bench, and it is hoped that this will serve for its identification at any future time. 232 LAKE BONNEVILLE. Observations of lake level were made on tlie IJlack Rock gauge troiu Sep- tember, 1875, to October, 1876, and single observations were made in July and October, 1877. The P'armington gauge was used from November, 1877, t(i November, 1879; the Lake Shore gauge fmiii November, 1879, to Sep- tember, 1881 ; the Garfield Landing gauge from Api-il, iSSl, to June, iSSd. The Garfield Landing gauge was inspected l)y mcml)ers of tlic corps from time to time until 1884, when Salt Lake City ceased to ))e a base for field operations. In 1886 Prof. Marcus E. Jones of that city ascertained and reported that the gauge had suffered accidents wlierel>y its zero liad been raised three tir four inches, but the dates of change were not learned. In June of the same year it was destroyed by a storm. Prof. Jones then began observations of the water height, and eventually prepared and in- stalled a new gauge, placing it near the position of the old one at Garfield Landing, and fixing its zero at the same height. This gauge, which will 1 le called the New Garfield, is still in use. All of the gauges except the New Garfield have by various accidents become displaced, so that the authenticity and coherence of the i-ecords depend wholly on the leveling and other observations conducted tt) deter- mine the relative heights of the gauge zeros. Connection between the Farmington and Lake Shore gauges was established by the writer by spirit- level at the time of the institution of the latter gauge. The Lake Shore and Garfield Landing gauges, which are separated by a space of more than 20 miles, were observed simultaneou.sly for a period of five days in March, 1881, the lake being at the time little disturbed by wind. In 1877 the late Mr. Jesse W. Fox and the writer ran levels from the Black Rock "■iiuffe to tlie Hlack Rock bench; and in 1881 Mr. Russell, by the aid of the .s])irit- level and the level aff"orded by the calm lake surface, connected the Garfield gauge in like manner with the Black Rock bench. These various determinations, together with others, have been compiled and reduced to a system by Mr. Wel)ster, Avhose report on the hypsometric work performed in connection with the Boimeville investigation will be found in Appendix A. He has selected the zero of the Lake Shore gauge as the datum or reference point for all heights within the basin. I insert a table of gauge heights based on liis compilation. GAUGING GREAT SALT LAKE. 233 Tablk VII. Datum PoUtts coiincvttd with iht gaiiyitnj of Great Salt Lake, lilack Rock Bencli Kiirminston Beuch lihick Itock Gau{:;e Zt^ro. . Farniiugton Gauj;o Zero . , Lake Shore Gaiifje Zero . Garticld Gauge Zero New Garfield Gauge Zero Feet. +41.8 + 1G. 7 + 5.3 + 3.K 0.0 - 4.0 - 4.6 Oscillations since i875.-Tlie followiug' ta])le sliows iill tlie trustwoi'tliy obsei'va- tions recorded by the observers at these several stations. It does not cover the entire i)eriod from 1875, but the breaks are unimjxirtant. Table VIII. Record of Oscillations of Great Salt iMke. Referred Gauge. Observer. Year. Day. Keatling. to Lake Shore Zero. Ft. In. Feet. lilack Eock J.T. Mitcliell... 1875 Sept. 14 0 0 5.8 22 0 5i 5.7 25 0 5 5.7 Oct. 0 0 4i 5.0 12 0 4 5.0 18 0 :ii 5.0 20 0 3 5.5 Xov. 9 0 2 5.4 10 0 n 5.4 23 0 4 5.6 29 0 .-.i 5.7 Dec. 7 0 5 5.7 14 0 51 5.7 21 0 6 6.8 1870 .Ian. 5 0 8 5.9 11 0 K* 0.0 29 0 9 6.0 Feb. 1 0 9 0.0 15 0 95 6.1 22 0 9i 6.1 Miir. 15 0 11 0.2 22 1 0 6.3 28 1 04 0.3 Apl. 17 1 2 6.4 25 1 3 6.5 May 2 1 4 6.6 22 1 9 7.0 J"une 2 1 11 7.2 8 2 0 7.3 13 2 2 7.4 - 23 2 4 7.6 234 LAKE BONNEVILLE. Table VIII. Record of Oacillationa of Great Sail Zaic— Continued. Ileforred (r.anKo. Observer. Year. Day. Reading to Lako Shftre Zero. Fl. In. Feet. m.vk Kork J.T. Mitchell . 1876 Juno 30 2 6 7.8 July 18 2 3 7.5 25 2 4 7.0 Auk. 1 2 3 7.5 10 2 2 7.4 22 1 9 7.0 20 1 8 0.9 30 1 8 B. 9 .Sept. 14 1 7 6.9 10 1 6i 0.8 2G 1 6 6.8 Oct. 0 1 51 6.7 G. K. Gilbert .. 1877 July 12 2 0 7.3 Oct. 19 0 10 6.1 Nov. 24 2 1 5.8 Farmington .T Mlllor 1878 Jan. 21 2 li 2 2i 5.9 Mcli. 28 6.0 May 2 5 0.2 June .10 2 C 6.3 July 18 2 3J 6.1 Nov. 1 1 0 4.8 Doc. 11 0 11 4.7 1870 May 2 1 4 5.0 Lake ehoro E. Gam Nov. 19 Doc. 2 2 6 2 C 2.5 2.5 10 2 7J 2.B 31 2 9 2.7 1880 Jan. 14 2 9J 2.8 29 2 7J 2.6 Fob. 23 2 7J 2.6 Mar. 10 2 9i 2.8 30 2 10 2.8 Apr. 15 2 lO.i 2.9 28 2 Il.i 3.0 May 12 3 1 3. 1 2fi 3 3.5 3.3 Juno 10 3 4 3.3 28 3 4i 3.4 July 13 3 3« 3.3 30 3 1 3.1 Aug. U 2 11 2.9 29 2 8 2.7 Sept. 14 2 5 2.4 20 2 2 2. 2 Oct. 15 1 ll.i 2.0 29 1 lOJ 1.9 Nov. 12 1 9 1.7 29 1 8i 1.7 Doc. 11 1 8J 1.7 14 1 9 1.7 RISE AND FALL OF GREAT SALT LAKE. 235 Table VIII. Record of Oscillations of Oreat Salt Lake — Continued. Referred Gauge. Observer. Tear. Day. Reading. to Lake Shore Zero. Ft. In. Feet. Lake ^horo E. Garn 1«80 . 1S81 27 Jan. U 1 10 I 10 1.8 1.8 28 2 2 2.2 Fob. 14 2 0 2.5 28 2 6i 2,5 Mar. 14 2 7J 2.0 Garfield Landing . . . T. Douris Apr. 1 7 3 2.6 16 7 4J 2.7 May 1 7 8 3.0 16 7 11 3.3 Juno 1 8 0 3.4 10 8 OJ 3.4 July 1 7 lOJ 3.2 16 7 10 3.2 23 7 9 3.1 Aug. 2 7 6 2.9 19 7 4 2.7 Sopt. 8 7 0 2.4 10 0 n 2.3 Oit. 2 0 9 2. 1 10 6 9 2.1 Nov. 2 0 8 2.0 10 6 8 2.0 Die. 1 0 8 2.0 15 6 !l 2.1 lKS-2 Jan. 2 6 0 2.1 ' 16 0 10 2.2 Fob. 2 6 inj 2.2 10 6 11 2.3 M;ir 2 0 IIJ 2.3 21 7 OJ 2.4 A pi. 1 7 n 2.0 IB 7 ;i 2.0 May 2 7 5 2.8 16 7 0 2.9 June 2 7 OJ 2.9 10 7 0 2.9 July 2 7 4 2.7 17 7 24 2.0 Aug. 2 7 0 2.4 15 6 10 2 2 Sei>t. 2 0 5 1.8 10 6 3 1.0 Oot,. 2 6 IJ 1.5 l.'i 6 0 1. 4 IVc. l.--. 0 0 1 4 30 0 0 1.4 18K3 Jan. 15 6 0 1.4 1 30 0 0 1.4 236 LAKE BONNEVILLE. Taiilk VIII. Record of Oscillationn of Great Salt Lake — Continued. Koferroil Gauge. Obsi-rvi-r. Tear. Day. Heading. to Lak. Sliun- Zero. Ft. In. Feet. (lai'tii^lil Landing .. T. Oiiuris IKHI Feb. ir, G 1 1.5 .-SO 0 H 1.5 Mar. 15 G 2 1.5 A pr. 2 G 4 1.7 Si-]it, :i 0 0 1.9 10 0 2 1.5 Oi:t. 3 5 8 1.0 IS 5 5 O.R Nov. 1 5 3 O.G 15 5 0 0.4 Doe. 2 5 0 0.4 15 5 0 0.4 I8K4 Jan. 2 5 0 0.4 15 5 Oi 0.4 Feb. 2 8 OJ 0.4 15 5 IJ 0.5 Mar. 1 5 2J 0.6 15 5 6 0.9 Apr. 1 5 8 1.0 15 5 11 1.3 May 2 G 2 1.6 15 0 5 1.9 June 1 7 0 2.4 15 7 3 2.6 July 1 7 5J 2.8 15 7 5J 2.8 Aug. 2 7 2J 2.6 15 7 OJ 2.4 Sept. 1 7 0 2.4 15 7 0 2.4 Oct. 2 7 0 2.4 15 G 11 2.3 Nov. 1 6 11 2.3 15 6 10 2.2 Doc. 2 0 1(1 2.2 15 0 11 2.3 1885 Jan. 2 7 1 2.5 15 7 2* 2.6 Feb. 2 7 3i 2.7 IG 7 5 2.8 Mar. 2 7 6 2.9 16 7 8J 3.1 Apr. 3 7 10 3.2 10 7 U 3.3 May 2 8 1 3.5 15 8 3 3.0 June 1 8 G 3.9 IG 8 9 4.1 July 2 8 10 4.2 15 8 n 4.2 lilSE AND FALL OF GliEAT SALT LAKE. Tablk VIII. Itecord of OscillaHoiis of Great Salt toAe— Continued. 237 Gauge. Observer. Year. Day. Reading Referred to Lake Shore Zero. ^ n ^^ ^^ -JU -^ V A S^^ ■^I^ A /X 4 Ay'^ V Kt V- ^-5 I ^^ A At ^ 7 ^ '-t- K A4^ ^ X3I ^'»».^'^""*^\^ + 6 ft / ^/^N > S + 4 / \ Itt / > + 2 2 ^^. ^^1 ^ ^. 0 / ^\, / / \ / - 2 ^ 3 ^ ■^ S^ ^ y^r-A. _j ^^ "^ -.^ X i^ ^ t z ^==^ T y^' \ V r-" "^^ >, -u-"^ \ ^^^ i + -—^ \,' + 2- \ -t-r ^v S I^ A \^ UA t^ ^V^ ^^ ^ 0 it V V^ • \ M rf \ \ h A i ^ / -1° V^^ S-^ ^ \4 ^ C V ^^ \r ^ -t-tt ^ - Z' T ^ 1 L.l 1 III Julius Bien * Lo.hdi CLIMATE CURVES LAKE CURVES AND CLIMATE CURVES. 247 This curve was found to be almost identical with that derived from the San Francisco observations alone and to be closely simulated by the curves of the other individual stations. It was therefore deemed legitimate to employ the San Francisco curve as representative of the district for the period antecedent to the institution of the Signal Service observations. The San Francisco observations, however, were not employed alone. Mr. Schott has combined with them data from Alcatraz Island, Angel Island, Fort Point and Presidio, all of which stations were in the immediate vicinity. His results are published in the form of mean annual temperatures, and these have been prepared for the present purpose by subtracting from each the mean of the series. The residuals thus obtained and the residuals de- rived from the Signal Service observations are plotted in curve V of PI. XXXIV. This curve may be considered to represent, with a fair degree of approximation, the non-periodic oscillations of temperature within the indicated period in the district of the Great Basin and Pacific Coast. Here, too, it is evident that a direct comparison with the curve of lake oscillation should not be made; whatever the influence of temperature upon the volume of the lake, whether through rainfall or evaporation, it would be semi-cumulative. The temperature determinations have therefore been submitted to the same process of special integration as the precipitation determinations; it was again assumed that the influence of each year's temperatin-e would diminish in arithmetic ratio so as to disappear in ten years. The deduced curve, IV, is far more regular than that derived from precipitation, and presumably represents the slow secular oscillation. In comparing the integrated temperature curve with the curve of lake oscillation the question arises whether the maxima of the former should be compared with the maxima or the minima of the latter. If temperature affects the lake chiefly through rate of evaporation, the maxima of one curve should coincide with the minima of the other. If its chief influence is ex- erted through precipitation, the correspondence should probably be found in the same way; but about this there is diftereuce of opinion. Fortunately, it is unnecessary to discuss the subject in this connection, for whether the comparison be made directly or by inversion, it is equally evident that the curves are inharmonious. 248 LAKE BONNEVILLE. Tlie integrated precipitation, curve IT, resembles the curve of oscilla- tion in several })articulars. Its maxinmm from 1852 to 1855 is comparable with the lake maximmn in 1855 and is^fi. Its minimnm from 1858 to 18G0 is comparable with the lake mininmm in 1860 and 1861; and dui-ing the great maxinunn of the lake from 1867 to 187!) the precipitation curve is for the ]nost part above its mean line. The only great disparity occurs in the years 1S(;8 to 1865, when the precipitation curve shows a minimum unrepresented in the curve of lake oscillation. The precipitation curve is therefore on the whole similar, and indeed its correspondence is quite as close as could be expected by one who realizes how imperfectly the average precipitation of a region is rejjresented by the observed precipitation at a small number of stations. There is, therefore, some su])port for the hypothesis entertained by many ])ersous that the exceptional rise of Great Salt Lake which culminated in 1873 was due to au increase of precip- itation.^ Turning noAv to the consideration of the influences exerted upon the lake by man, we find them separable into two classes; first, those which cause a greater proportion of the precipitation falling on the Innd to be gathered by the streams and carried to the lake; second, those which cause a smaller pro]iortion of the precipitation to reach the lake. The supposed influence of deforesting on the rainfall itself need not be discussed, because in this region no considerable body of forest has been destroyed. The chief influence of man in increasinfj the inflow of the lake is through the grazing industry. In their virgin condition many of the lowland vallej-s and all the upland or mountain valleys were covered by grass and other herbaceous vegetation. These have been eaten off" Ijy the herds of the white man, and in their place has sprung up a sparse gi'owth of low bushes between which the ground is bare. From this bare svirface it is believed that the water falling as rain or freed by the melting of snow, runs off more readily than from the original grassy surface, so that a smaller share of it is evap- orated in situ and a larger share flows through the water courses to the lake. This change has affected a large total area; and if its influence ujjon ' The observational data discussed close with the year 1883. As the niauusoript goes to press they are available to 1889. The later data have not been systematically treated, but their inspection shows that the general conclusion is sustained by them. MAN'S INFLUENCE ON GREAT SALT LAKE. 249 water supply is here coiTOctly interpretecl, it is a factor of importance. Another tactor of tlie same tendency is the draining- of marshes and beaver ponds. IMany of the small streams of the basin were clogged by beaver dams, and the courses of some of these have been opened by the white man for the purpose of increasino- the supply of water for irrigation. The in- creased su])ply has been utilized for irrigation during a portion only of the year, and at other times has joined the streams flowing to the lake. Plowing and irrigation have the contrary effect. Land broken up for cultivation is tliei-eby rendered more porous, so as to retain a larger portion of the rain falling upon it. I'liis i-etained portion is chiefly returned to the atmosphere by evaporation and is tluis lost to the lake. The effect of irriga- tion is precisely similar. The water diverted from the streams and spread out on the land for the })urpose oH nourishing crops is restored to the atmos- l)here by evaporation from the surface of the soil and from the leaves of plants. In 1877 the writer estimated that the inflow of Great Salt Lake was diminished six per cent by this cause. With the exception of iri-igation, it is impossible to give quantitative expression to these factors. Those which tend to increase the lake })robably culminated fifteen or twenty years ago, and have .since remained constant. Those which tend to diminish the lake have increased continuously for the last 35 years. The time is probably past when the net tendency toward lake increment was at a maximum, but it is not entirely clear whether the present sum of human agencies tends toward lake expansion or lake con- traction. Li any case the consideration of the qualitative relation of the several factors suffices to sliow that a curve representative of the influence of hinnan agencies could have but a single maximum, and could not corre- spond in detail with the determined curve of oscillation. Ten years ago I discussed at some length the comparative merits of the climatic theory and the theory of human agencies,' concluding that neither was inconsistent with the facts and that the truth might include both. I pointed out that the former appealed to a cause that may be ade- quate but is not independently known to exist, while the latter appealed to causes known to exist but quantitatively undetei'mined. Since that time ' Lands of the Arid Region, pp. 68-77. 250 LAKE BONNEVILLE. the publication of the second edition of Mr. Schott's discussion of i-ainfall and the progress of the work of the U. S. Signal Corps have rendered it j^ossible to construct the most iinportant fom])arative climatic curves, and the subject is here resumed for the purpose of exhibiting the relation of these curves to the curve of oscillation. The coiTcsjiondence of the inte- grated precipitation curve to the curve of lake oscillation is siifficientlv close to indicate a causal relation, especially in view of the fact that rainfall is the climatic factor to which hypothesis most naturally appeals. In the present as])ect f)f the problem, precipitation seems entitled to rank as the dominant factor, the results of its variation being only slightly modified by the variations of temperature and the changes introduced by gi-azing and agriculture. Future changes.-Thosc humau ageucles which tend to increase the water supplv of the lake, namely, grazing and draining, have acquired a status that is practically permanent, but those which tend to diminish the supplv, namely, plowing and irrigation, have not yet ceased to increase. In 1877, when the consumption of water by irrigation w.as estimated at six per cent, of the inflow of the lake, the intervention of the irrigator was restricted to the minor streams of the basin. The main bodies of the Bear and of the Jordan, the largest of all the streams, flowed unimpeded to the lake. Since that time, the diversion of the water of the Jordan has been undertaken on a large scale ; and the time can not be distant when its entire volume will be utilized. The Bear River presents greater engineering difficulties, and has not yet been brought under control; but sooner or later a large district will be redeemed by means of its water, and the lake will be correspondingly deprived of tribute. Human agency is thus destined to play an iniportant part in the detemaination of the future history of the lake. The next ten years will witness its shrinkage, for lack of affluent water, to a size smaller than has before been observed. It is not to be expected that it will ever share the fate of Sevier Lake, because the conservation of all the stream water for irrigation is not economically practicable, but it will })i"obably be so reduced in voliune as to precipitate a portion of its salt. The final system of irrigation will include the storage in artificial reservoirs of the flood water of all the minor streams, and will cause the lake FUTURE SHRINKING PROPHESIED. 251 to be deprived of all inflow except from saline creeks and from the unused share of Bear River, l)ut this system is not likely to be established by the present generation. The expansion of the methods now in vogue to a limit dependent on the extent of tlie readily available arable land, together with the construction of reservoirs on the most available sites, will employ about two-thirds of the water supplv, and will proportionately reduce the area of the lake. One effect of si;ch a contraction of the lake will be to simplifv its out- line. Antelope, Stansbury, Carrington, Hat, and Dolphin islands will he permanently united to the land. Bear River Bay will be drained nearly to the southern extremity of Promontory, and the bay east of Antelope Island will be drained nearly to the northern end of that island. The Jordan, the Weber, and the Bear will iniite their deltas in the vicinity of Fremont Island, and will eventually fill up all of the sound east of that island, reducing the lake to a linear body lying east of Stansbury Island and the Promontory. With a lowering of the lake siu-face the projection of deltas ^will be a rapid process. During the recent high stage of the lake the chan- nels of the three principal rivers have been converted, in their lower por- tions, into estuaries whose sluggish current has permitted the accumulation of silt. The volume of this silt has been at the same time increased by the culti^'ation of the soil, an industry which always augments the detrital loads of the streams. The lowering of base-level incident to the falling of. the lake surface will cause the streams to erode this detritus and transport it to the shore of the lake. Saline Contents-Auother cffcct will bc the concentration of the brine. The lake is so shallow that its volume is greatly affected by small changes of level, and since the total amount of contained salts undergoes no appre- ciable change, the strength of the solution is affected. Variations of salinity have been observed by persons engaged in the manufacture of salt from the brine, and quantitative expression lias been given to the same facts by the analyses made from samples gathered at different dates. With the lake at its lowest observed stage, 1850, Stansbury collected a sample of the brine containing 22.4 per cent, of solid matter. From a sample gathered in 1873, when the lake was at its highest stage, Bassett obtained 13.7 per cent, of 252 LAKE BONNEVILLE. solid matter. At an intemiediate stage King- collected in 1869 a sample containing 14.8 per cent, and Talmage in 1885 and 1889 obtained samples yielding 1(1.7 and lit.G per cent. It would appear from a comparison of the extreme results that with a rise of the lake surface of 10^ feet the salinity was decreased by 39 per cent, of its amount; and, assuming that the quan- tity of saline matter in solution remained unchanged, the volume of water in the lake was at the same time increased 73 per cent. While these results are approximately true, they should not pass with- out (pialification. Careful comparisons of the several determinations of salinity ^\\t\\ tlie several determinations of density and with the correspond- ing determinations of height of water surface, reveal numerous discrepancies. The comparison of salinities with densities shows that there are errors in determinations of salinities or densities. Discrepancies between determined salinities or densities on the one hand, and heights of water surface on the other, suggest several sources of error. No collector of water samples has placed on record the spot where the collection was made; one may have stopped near the mouth of a stream and obtained too low a salinity; another may have visited a lagoon of the shore with abnormally high salinity. Stansbury and King neglected to record the dates of sampling; and of the five samples analyzed three were collected before the establishment of gauges; there is thus some uncertainty in determinations of the height of the lake when its brine was sampled. The accompanying analyses embody all our knowledge of the nature of the brine and they accord so poorly with one another that they wai-rant our speaking with confidence oidy of the most striking characteristics. The principal base is sodium, and this exists chiefly in the form of chloride, but also as sulphate ; next in rank is potassium, and then follow magnesium and calcium. Despite the fact that calcium carbonate is precipitated on the shore in the form of an oolitic sand, none of the analysts have succeeded in iinding it in the brine; and it is probable that the weighable calcium found in two of the samples exists in the form of sulphate. The theoretic combination of acids and bases given in the lower division of the table is in the main tentative only; but the readiness with which sodium sulphate is ol)tained from the brine Avarrants the belief that it is one of the actual constituents. THE SALT LAKE BRINE. 253 Wlien in winter the temperature of the water falls below 20° F., the precipi- tation of this salt begins, and it sometimes accumulates in such quantity as to be readily gathered from the bottom, or is even thrown upon the shore by the waves. The sodium chloride has become the basis of a large industry, being manufactured for table and dairj- use as well as for metallurgic purposes. This industry has so expanded since the close of my work in Utah that a statement of its condition at that time Avould have historical value only. It is re})orted that the output in 1886 was 23,000 tons; in 1887, 40,000 tons; in 1888, 21,000 tons. For several years sodium sulpliate cast on the shore Ijy the waves in winter has been gathered, and its utilization for the production of various sodium salts of commercial importance is already undertaken.^ The quantity of sodium chloride contained in the lake is about 400 millions tons; of sodium sulphate, 30 millions tons. Table IX. Anahjses of Wate>- of Great Salt Lake. I. Sample taken id 1850; analysis by L. D. Gale. II. Suniple taken in summer of 1869; analysis by O. D. Allen. III. Sample taken in Auj;u8t. 1873; analysis by H. liassett. IV. Sample taken in December, t885; analysis by J. E. Talmage. V. Sample taken in August, 1889 ; analysis by J. E. Talmage. I. II. III. IV. V. Total aolids in 1000 parts of water. . . 224.2 1.170 148.2 [1.111] 136.7 1,102 167.2 1.122 195.5 1.157 First arrangement ofreaults; by acids and bases. Parts in 1000 of water. Per cent, of total sulids. I. 11. III. lY. V. I. II. III. IV. V. 124.5 12.4 85.3 84.0 9.9 49.6 2.4 .2 3.8 Trace Trace 73.6 8.8 38.3 9.9 .6 3.0 90.7 13 1 58.2 1.9 .4 2.9 110.5 11.7 6.5.3 2.1 .8 5.1 55.8 6.0 38.3 .3 50.0 6.6 33.1 1.0 .2 2.5 51.9 6.6 28.6 7.4 .4 2.2 54.3 7.8 34.8 1.1 .3 1.7 56.5 6.0 33.4 1.1 • 4 2.6 Sulphuric acid (SO4) . . . Trace .6 Total 222.8 149.9 134.2 167.2 195.5 100.0 100.0 100.0 100.0 100.0 'The waters of Great Salt Lake. By James E. Talmage. Scieuoe, vol. 14, 1889, pp. 444-446. 254 LAKE BONNEVILLE. Second arrangement of results ; hy theoretic cnmhinatiotis of acids and bases. Parts in 1000 of water. Per cent, of total solids. I. II. III. IV. V. I. II. III. IV. V. Sodium chloride 202.0 118.6 88.5 18.9 11.9 10.9 2.0 2.0 135.9 l.'.7.4 90.7 79.1 65.9 14.1 8.9 8.1 1.5 1.5 B1.3 80.5 Mafjuoaiuin chloride ... Sodium sulphate 2.5 18.3 14.9 9.3 5.3 .9 .9 11..1. 14.2 4.3 1.5 20.1 10.5 4.7 2.8 1.1 8.2 9.9 6.2 3.6 .6 .6 6.7 8.5 2.6 .9 10.3 5.4 2.4 1.4 Chlorine (excea.s) 222.8 149. 9 134.2 167.2 195. S 100.0 100.0 100.0 100.0 lOO.O Xote. The first sample of water was collected by Stansbary, and its analysis is reported on p. 419 of the "Expedition to the Gre.it Salt Lake." The second was collected by the Fortieth PariUel Survey, .nnd is rt-poi ted in Systematic Gtolojjy, vol. I, p. .'">02, and Descriptive Geology, vol 2, p. 433. The third was collected by Dr. W. Marcet in August, l>^73, and is reported in the Chemical News for Nov. 7th, 1873 (vol. 28, p. 2:tG) by n. Cassett. The fourth and fifth were collected by J. E. I'almage in December, 1885, and August, 1889, and :ire rei)ortcd in Science, vol. H, 1889, p. 445. Gale reported the salt.s as hero given in the first column of the second table. Allen's repoit includes two forais, the salts being given in one and the alkalis and acids in the other. Allen's figures, a.s printed, aro not perfectly consistent: the report of the combined salts baa been used in deriving the figures here published. Basaett's report was published in the form here given in the third column of the first taltle. The entire error of analysis is computed iu chlorine in the second table, columns II and III. Talma, e'a reaulta were published in the form given in the first part of the second table. Gale's and Tal- mage's errors of analysis do not appear.. Sources of Saline Matter.-Tlie sources of tliG saliiie material may be considered in two classes; the first including the rivers, the second the littoral springs. The Bear, the Welder, the Jordan and a small number of creeks rise in up- lands above the horizon of the Bonneville shore and bring to the lake water Avhich is sensibly fresh, containing only minute quantities of mineral matter. A cordon of springs about the shore of the lake rise through the Bonneville beds, and are so far charged with salts leached from the sediments as to be perceptibly brackish. With these should be classed also the Malade River, the upjjer course of which is fresh, while the lower is rendered brackish by the accession of saline water from thermal springs rising in the lied of the stream' Avithin the Bonneville area. With only our 2)resent knowledge it is iinjjossilile to say whether the fresh rivers or the In-ackish springs furnish the greater saline tribute to the lake. The rivers only have been subjected to chemical examination. The constitution of the Jordan water was determined from a sample collected in Utah Lake, the source of tlie river, and this determination is taken to represent about one-third of the inflow of the lake. Bear Kiver was sampled at Evanston, where the stream lias proliably two-thirds of its ACCUMULATION PERIOD. 255 maximum volume Since this river furnishes alxnit luilf the water supply of the lake, the sample is taken to represent one-third of that suppl)-. The two analyses exhibit the constitution of two-thirds of the fresh-water tribute of the lake, and it will l)e assumed that their mean shows the character of the entire fresh-water tribute. In the following- table this mean is compared with the analysis of the lake water as reported by Allen: Table X. Accumulation Periods for Suhslnncen contained in the hrine of Great Salt Lake. Sabatance. Parts in 1000. V. Accumu- lation Period. I. Bear River Water. II. Utah Lake Water. III. Meau of I and II. IV. Great Salt Lake Water. Cblorine .0040 . 0105 .0082 . 01'J4 .1306 .0178 Trace .0558 .0186 .0086 . 0703 .0130 Trace .0405 .0155 84. 00 9.87 49.05 2.40 .25 3.77 Trace Trace Tears. 34, 200 490 13, 400 18 850 Su'pbniic acid .. Sodium Calcium Magnesium .0432 .0125 Phoapbonis Rate and Period of Salt Accumulation.-At the tlmC wllCU Allcu's Sample of briuC was collected the lake had a mean depth of about 19 feet. The annual inflow to the lake has been appi'oximately estimated as sufficient to add 5^ feet to its depth.^ The lake volume is therefore equaled by the inflow in three and a half years, and in that period the saline strength of the lake is increased by an amount equal to the saline strength of the inflow. Disregarding for the present the supply from littoral springs, and considering only the supply from rivers, we may, by the aid of these considerations, deduce from the table tlie time necessary to store up in the lake the observed amount of each of its mineral constituents. The results of such comj)utation appear in the right-hand column of the table. One of the most conspicuous features of these results is their variety. The streams carry enough calcium to charge the lake to the observed extent in eighteen years, but 34,000 years are necessary to similarly charge it ' Lands of the Arid Region, p. 72. 256 LAKE BONNEVILLE. witli chlorine. Tlie explanation lies in the relative supply of these sub- stuuccs and tlieir relative solubility. In the mountains from -vvhich the rivers flo\v, calcium is afforded in luilimited quantity, while the su})ply of chlorine is relatively very small. Chlorine, on the other hand, e-xisting as it does in combination with sodium, is highly solulilc; while calcium, exist- ing for the most part hi combination with carbonic acid, is sparingly soluble. Chlorine therefore accumulates in the lake, while calcium is precipitated. It is a matter of observation that calcium carbonate gathers on the shore of the lake as oolitic sand, and it is probable that it also falls to the bottom as a marly constituent of the lacustrine sediment. Calcium has therefore reached its limit and is an unvarying constituent of the brine. The annual accession is balanced by the annual precipitation. The same remark applies to the magnesium. It is presumably precipi- tated with the calcium, just as it was from the waters of Lake Bonneville, and chemical analysis shows that a small portion of it is accumulated in the oolite of the shore. The short period necessary to accumulate the lake's store of sulphuric acid, 490 years, indicates that it, too, has passed the saturation limit and is being precipitated. It appears to exist in the lake in the form of sodium sulphate, and it is probabl}' precipitated in that combination. The fact that sodium sulphate is discharged from the lake by the extreme cold of winter indicates that it must exist at ordinary temperatures in quantitities not far from the saturation limit; and it is found to be the first mineral to separate from the brine when evaporated by insolation. There remain two substances whose long accumulation periods permit us to doubt whether they have reach('(l tlie stage in whicli accession and loss are equal. Sodium and chlorine, in their combination as sodium chloride, constitute the most al)undant mineral, and no analysis has indicated that the brine is fully saturated therewith. If it be true, as surmised, that the annual supply of sul})huric acid is discharged from the lake by the precipi- tation of sodium sulphate, the accumulation period for sodium chloride is not properly represented by the period conq)uted for sodium. It is more likely to be represented by the period estimated for the chlorine, namely, 34,200 years. now OLD IS GREAT SALT LAKE? 257 If iio\y we recall to attention tlie tribute of the littoral springs, tem- porarily ignored, it is at once apparent that onr table nnderestiniates the annual tribute of sodium chloride and corres])ondingly overestimates its accumulation period. We have no present means of determining the extent of this ovei'estimate, but Ave can safely say that the period necessary to charge the lake with common salt by means of the present sources and rate of supply is not more than 25,000 years. Shall we conclude that 25,000 years ago the lake was fresh? or is there reason to believe that sodium chloride, like the other constituents, is being precipitated by the lake as rapidly as received? To this question a satisfactory answer can not be given, but there are several considerations favoring the second alternative. First, the circumstances coimected with the old storm line, to which reference has already been made, indicate that the lake was smaller and therefore more concentrated, for at least a few decades preceding the settlement of the country, than it has been since. It may Avell be that a portion of the salt was thrown down during this preliistoric period, and that it was condjined with mechanical sediment in sucli way as to be preserved from resolu- tion. Second, it is known that under special circumstances salt is now precipitated at some points on the margin of the lake. Where a broad expanse of water near the shore is exceedingly shallow, the local evapora- tion is not compensated by the circulation, and the resulting- high concen- tration leads to a discharge of salt. In passing from Grantsville to Stansbury Island in 1881, Mr. Russell rode for a mile across a deposit of this character an inch in thickness. Such a deposit as this would vmdoubtedly be redis- solved if the lake rose, or if it fell so as to permit the action of rain; but the fact of its formation indicates how triAdal are the conditions which may determine precipitation. On the whole, it is not unreasonable to suppose that each of the minima which occur in the ordinary history of the oscilla- tions of the lake marks an epoch of precipitation, when a portion of the saline matter is discharged .and a smaller portion is so combined with other sediments as to remain a permanent deposit. While it can not be true that the annual precipitation counterbalances the annual supply, it is quite conceivable that a century's precipitation disposes of a century's supply. MON I 17 258 LAKE BONNEVILLE. There seems thus a possibility, if not indeed a j)r()l)ability, that none of the substances which have been ([uantitatively determined in the; Ijrine and in the tributary rivers are undergoing accunudation in th(! lake; ])ut it does not foUow tliat this equation of supply and discharge h;is sid)siste(l for a long period. There are certain soluble l)ut very rare substances, such as the comjiounds of Ijoron, lithium, iodine and bi-omine, which tend to accu- mulate in inland lakes of great antiquity and have come to be regarded as the diagnostic characters of age. Only one of these has been detected in the water of Great Salt Lake, and that one is not found in measurable quantity. The conclusion that the brine is recently accumulated accords with the facts derived from the Bonneville history, for at the time of the outflow the salts stored in the lake must have been discharged beyond the limits of the basin. The age of the Gi'eat Salt Lake brine can not then be greater than the antiquity of the second Bonneville flood. We might conclude that the age of the brine is precisely equal to the antiquity of the Bonneville flood were it not for the possibility that the lake has since then been freshened by desiccation. Russell finds excellent reason to believe that in the Lahontan basin, which is in many respects a duplicate of the Bonneville, the flood epoch has been followed by one of very low ebb, in which the residuaiy lakes have so dried away that all their saline matter lias become entangled with mechanical sediment.^ A more recent accession of water has produced a number of slightly brackish lakes, whose feeble brines contain in their constituents no hint of great age. If the Salt Lake basin has passed through a similar recent epoch of desiccation, it is not easy to see how we should become cognizant of it. Provided the antiquity of the epoch was sufficient to permit the subsequent accunudation of the sodium chloride, the character of the brine would be sub.stantially as we find it. For the present, at least, we must regard it as an open ques- tion whether the existing lake with its characteristic brine dates from the cessation of Bonneville overflow or from a subsequent epoch of extreme aridity. Fauna.-The animal life of the lake has been described by Packard, who finds it to consist ot two species only, a brine shrinq), Artvmia yraciHs- ' Geol, Hist, of Lake Lahoutau, ]ip. 2'H-i'iO. THE BKINE SHKIMP, 259 Verrill/ and the larva of a fly, Epln/did (/racilis Packard. Tlieso are very abundant in certain seasons of the year. They feed upon alga', of which three species have been recognized. The meagerness of tliis fauna is to be ascribed tt) the rarity among animal sj)ecies of the power to li\'e in concen- trated brine. Packard ascribes the phenomenal abundance of the Artemia to the absence of enemies, for the brine sustains no carnivorous species of anv sort. The genus is not known to live in fresh water or water of feeble salinity, but it connnonly makes its appearance when feebly saline waters are concentrated by evaporation. It has been ascertained that a European species takes on the characters of another genus, Branchinecta, when it is bred through a series of generations in brine gradually diluted to freshness, and conversely, that it may be derived from Branclihieda by gradual increase in the salinity of the medium. It is found, moreover, tliat its eggs remain fertile for indefinite periods in the dry condition, so that whatever may have been the history of the climate of the Bonneville Basin, the present occurrence of the Artemia involves no mystery. During the Bonne- ville epoch its ancestors may have lived in the fresh waters of the basin, and during the epoch of extreme desiccation, when the l)ed of Great Salt Lake assumed the playa condition and was diy a portion of the year, the persistent fertility of its eggs may have preserved the race. Or, if the playa condition with its concomitant sedimentation was fatal to the species, it may be that the alternative fresh water form survived in upper lakes and streams of the basin, so as to restock the lower lake whenever it aftbrded favorable conditions. THE GESTKRALi HISTORY OF BONNE VIIiLE OSCIIjIjATIONS. We may now assemble the conclusions derived froni the discussions in preceding chapters and in the j^receding sections of this chapter, and exhibit a complete history of the oscillation of lake surface within tlie Bonneville Basin, so far as it is known. The relation of the alluvial cones to the shore-lines, and the condition of the low passes on the rim of the basin, show that before the Bonneville ' A monograph of the Phyllopod Crustacea of North America. By A. S. Packard, Jr. U. S. Geol. and Geog. Surv. of the Terr. 12th Ann. Kept., Part 1, 1883, pp. 295-592. Artemia graciUa on pp. 330-334. 260 LAKE BONNEVILLE. flooding' the water level was low. This we may call the pre-Bonneville low-water epoch. It was of great duration compared with those enumer- ated below. The first Bonneville epoch of higli water is stratigraphically repre- sented by the Yellow Clay. Peculiarities of the shore-lines, and the })he- nomena at Red Rock and other passes, shoAV that the water did not rise to the rim of the basin and was not discharged. After the deposition of the Yellow Clay the water subsided, and the basin was nearly or perhaps completely desiccated. The stratigraphic evi- dence of this subsidence is found in the unconformity betAveen the Yellow Clay and the White Marl and in the alluvial deposits occurring at that horizon. The possibility of complete desiccation is suggested by the differ- ence in character between the antecedent and subsequent deposits, Avhich difference may have been occasioned by a change in the conditions of sedi- mentary precipitation. This may be called the iuter-Bouneville epoch of low water. The second Bonneville epoch of high water is represented stratigraph- ically by the White Marl. Before the close of the epoch the water over- flowed at Red Rock Pass, forming a channel of outflow which was excavated to a depth of 375 feet. The Bonne\'ille shore-line records the water surface at the date of initial outflow. The Provo shore-line records its position after the channel of outflow had attained its maximum depth. The existing state of affairs was brought about by the recession of the lake surface from the Provo shore, and is stratigraphically re})resented by the formation of local alluvial deposits on the sui-face of the White Marl. The sedimentary deposits and shore embankments marking the high- water stages have been more or less eroded by the modem streams, and the ancient deltas especially have been deeply trenched. The basin has been diA-ided into a number of minor hydrographic units. This modern epoch may be called the post-Bonneville epoch of low water. Nothing is known of the absohite duration of these epochs, and in the study of their relative duration no trustworthy means has been found for comparing a high-water epoch with a low-Avater epoch. The deposit mark- ing the first high-water epoch is thicker than that marking the second, and SUMMARY OP BONNEVILLE HISTORY. 261 we may hence conclude that the first epoch was the hinger, but the amount of this difference is rendered indefinite by the fact that the base of the lower deposit is not exposed. The comparison is further comphcated by the difference in the two deposits, the lo^ver containing in the center of the basin a larger per cent, of clay than the upper. If it be true that the water was so constituted during the second flood as to precipitate a relatively large share of the clay near the shore, and that the difference of constitu- tion did not affect the precipitation of the calcareous matter, a time ratio may be based upon the calcareous factors of the two elements of the exposed section. A computation under this postulate indicates that the first high-water epoch was not less than five times as long as the second. Data do not exist for the quantitative estimation of the relative dura- tion of the low-water epochs, but their order of magnitude is unmistakable. A comparison of the few alluvial wedges referable to the inter-Bonneville epoch with their local representatives formed during the post-Bonneville epoch shows the former to be invariably the larger, and indicates that the time between the two Bonneville floods was longer than post-Bonneville time. The pre-Bonneville low- water epoch represented by the great alluvial cones of the movmtain flanks is still less amenable to numerical statement, in that its beginning is undefined; but it is unquestionable that it far tran- scended in length the inter-Bonneville epoch. It will be observed that in all respects our knowledge of the high-water epochs is relatively definite. Not only are we able a^jproximately to com- jiare tlie two high-water epochs in duration, but we know that on the sec- ond occasion the water rose higher than on the first. But of the decree of desiccation attained in the pre-Bonneville and inter-Bonneville epochs we are practically without information. We have observed and approximately determined two important maxima of an undulating curve, and have dem- onstrated that they are the only great maxima of the curve; but we know practically nothing of the remainder of the curve and are unable to indicate the position of any minima, properly .so called. The knowledge we have gleaned is graphically exhibited in Fig. 34, where the upper and lower horizontal lines represent the horizons of the Bonneville shore and the surface of Great Salt Lake. Horizontal distances 262 LAKE BONNEVILLE. represont time, counted t'roni left to right. 'J'lie curve represents tlie lieig-lit of tlie oscilliiting water surface, and the shaded area indicates ignoi'ance. Fig. 34. — Rise and Fall of water in the Bonuovillo Basin. THE TOPOGRAPHIC INTERPRETATION OF LAKE OSCILLATIONS. (!)ne of the most important siilijects to ■wliicli the discussion of the Bon- nevilk' history sliouhl contribute is tliat of geologic climate. The oscilla- tions of the lake were in all })ro]iability caused ])y oscillations of climate; and if we can satisfy ourselves as to the nature of the |)articular climatic movements associated witia the rise and the fall of the lake, we can imme- diatel}', by changing the notation of oin- curve, convert it into a record of geologic climate. But in order to be fully satisfied that the curve has cli- matic .significance, it is necessary at the outset to give consideration to other possible modes of interpretation. For this purpose we revert once more to the fundamental conditions controlling the size of a closed lake. The size depends on the ratio between the suj)ply of water and the rate of evapora- tion. Rate of evaporation is purely a function of climate; but water supply depends quite as much on topographic configuration as on meteorologic conditions. We are therefore called upon to inrpiire whether the water su})ply of the Bonneville Basin may have been modified by to})Ographic changes in such way as to account for the demonstrated rise and fall of the lake. It is conceivable, first, that local oscillations of land surface, or volcanic eruption, or the Inu'sting of barriers may at one time have increased the Bonneville drainage district at the expense of some other district, and mav afterwards have diminished it. It is conceivable, second, that crust move- ments may have affected the altitude of the nimuit.iins whence the wati'r supply of the basin floAvs, in such way as to cause them to intercept a greater share of atmospheric moisture at some times than at others. It is conceiva- WHAT CONTEOLLED THE WATER SUITLY? 263 ble, third, that still grander crust movements have, by raising and lowering a great area including the basin, produced corresponding modifications of its general climate. Hydrographic Hypothesis.-Tlie posslbility that tlic Bomievillc drainage district has gained or lost Ijy the slow shifting of water partings or the diversion of rivers has already been considered in the first section of this chapter; and it is there shown that the only important changes it is admissible to postu- late are such as affect the supply afforded by Bear River. It is quite pos- sible that the Blackfoot, which now belongs to another drainage district, once contributed its waters to the Bear; and on the other hand, it is quite possible that the main trunk of the Bear was once turned from the Bomie- ville Basin to that of the Columbia; but the first of these possibilities is quantitatively and the second is qualitatively inadequate to explain the Bonneville oscillations. If the Blackfoot were now to be restored to the Bear River, there would result an increase in the area and depth of Great Salt Lake, but such change is not to be compared in magnitude with the changes involved in the Bonneville history ; the depth of the lake would be increased only five or ten feet at most. If the main trunk of Bear River were to be converted into a tributary of the Columbia a more important result would be produced, but the Bonneville status would not be restored; on the con- trary, the area and depth of Great Salt Lake would be diminished. It may be added that the condition of the basaltic sheets occupying tlie passes Ijetween the Bear River and the tributaries of the Columbia does not indicate that they are sufficiently recent to be appealed to in ex])lana- tion of the changes during the Bonneville epoch. There are lavas within tlie lake area which, judged by their condition with respect to weathering, are newer than those on the northern passes, and yet are demonsti'ably older than the epoch of the Yellow Clay. orogenic Hypothesis.-Thc mouutains affordiug the chief Avater supjjly of the basin are the Wasatch and the Uinta. The Wasatch is known to have in- creased in height, by faidting, since the last Bonneville flood, and both ranges are known to have been somewhat u])lifted since the deiDOsition of Neocene strata. It is highly probable that they ex])erienced upward movements during Pleistocene time; and it is indubitable that every such 264 LAKE BONNEVILLE. movement would result in an increase of the local precipitation and of the consequent mag-nitude of the streams. On the other hand, it is hig-jily iin- prohable that either of these mountains has been subject to displacements of such nature as to reduce its height. The conjoint influence of rhythmic upheaval and equable degradation undoubtedly produces alternate gains and losses in altitude, and there must be corresponding gains and losses in the pre- cipitation and outflow ; but however plausible such a hypothesis may appear ni)()ii a merely qualitative statement, it must be regarded as quantitatively inadequate. We have an approximate measure of the extent of the degra- dation in the lacustrine deposits which derive their material chiefly from that source, and we can not suppose, for example, that the removal of the entire mass of the White Marl from the uplands at the east would sufficiently aff'ect their altitude to diminish the water-supidy of the basin as it has been diminished since the White Marl epoch. There is, moreover, a general objection to any explanation appealing to merely local changes, whether of drainage or altitude. The history of Lake Lahontan, as developed by Russell, corresponds in a remarkable way with that of Bonneville. It includes two maxima and two only, the first being the longer and the second the higher.^ It is therefore in a high de- gree probable that the phenomena have a common cause, and such cause must be of a general nature. Epeirogenic Hypothesis.-Tliis difficulty is oscapcd by the third hypothesis, in which a large area, including both lake basins, is conceived to have been siiccessively elevated and depressed to an extent sufficient to reform its climate. Of the adequacy of such a cause there can ])e no question, l)ut we are without evidence of its actuality. There are, indeed, in the basins of the Columbia and Frazer, systems of terraces indicative of recent changes in the relation of the ocean to the continent; but these serve only to indi- cate the fact of wide-spread change and do not demonstrate sucli changes as are necessary to account for the flooding of the Lahontan and Bonneville Basins. If that flooding is the index of a local climate wrought ])y conti- nental movement, the humid condition should theoretically be the result of continental elevation and the last change should have been a subsidence; ' Geol. Hist, of Lake Lahontan, p. 237. OPINIONS ON CORRELATION OF LAKES AND GLACIERS. 265 whereas, in the basins of the Cokimbia and Frazer, tlie hist chanjre appears to have ])een an elevation. Since the suggested continental movements could affect tlie lakes only through the mediation of local climate, the hypotliesis which appeals to them is essentiall)' a climatic hypothesis; and its further consideration may be deferred until its proper place is reached in the discussion of the intluence of changes in terrestial climate. THE ClilMATIC INTERPRETATION OF LAKE OSCILtiATIONS. OPINIONS ON CORRELATION WITH GLACIATION. Turning now to the subject of climatic interpretation, we find an almost universal agreement among geologists in the view that the lake maxima were in some way associated with the history of glaciation. Tlie idea tliat the rise of a lake contained in a closed basin is a phenomenon properly cor- related with the formation or extension of glaciers appears to have been independently suggested by Jamieson, Lartet, and Whitney. Jamieson, speaking in 1863 of the climate of Central Asia,' said: The great basin of the continental streams, larger than the area of Europe, is remarkable for its inland lakes from whence no streams ever reach the ocean, owing to the great heat drying up tlie water. Now this heat and dryness being much lessened during the glacial period, there must have resulted a much smaller evaporation, which would no longer balance the indow. These lakes therefore would swell and rise iu level, . . . Two years later, Lartet wrote: Tlie level of tlie Dead Sea must therefore have been constantly regulated by the conditions of equilibrium between atmospheric preci[iitation and evaiioration. The extension of the waters of this lake, at a certain e[ioch, revealed by the sediments now laid bare, which cover such vast surfaces to the north and to the s.uth of its present limits, bears witness to a great change supervened since then iu the atmos- pheric conditions to which the liydrograpliic regime of the country was subjected. In the absence of fossils in the sediments anciently dei)osited by the lake, it is impossible to assign a i)recise age to the elevation of its waters. However, taking account of the probable duration of the phenomena which must have preceded and followed this important phase of the history of the Dead Sea, one would be led to attribute to it a date close to the end of the Tertiary and the beginning of the Quatei- 'Ou tlie parallel roads of Glen Roy and their place in the history of the glacial period, by Thomas F. Jamieson, Quarterly Journal Geological Soc, Loudon, vol. 19, pp. 235-259. The passage cited occurs on p. 258. 266 LAKE BONNEVILLE. nary poriod. Ono would thou bo alilo to soo in tliis riso of tlic, surface, of tlio lake an ettect of the glainal plieiiomeiia whose, iritliieiice seems to have extiMidecl, at tliese epochs, to ueighboriug resioiis. Tliis, inor(M)ver, would accord ipiite well with the observation of traces of ancient moraines which Dr. D. Hooker tliought he recognized on the slopes of Lebanon.' Only a few months later, Whitney, treating, in the first volume of the Geology of California, of the former extension of ^Mono Lake, said: Whatever cause gave rise to the immense body of ice, in the form of glaciers, which, as we have seen, formerly covered the summit of the Sierra in this region and extended down for 5,000 feet or more from the crest, this would undoubtedly have been snfticient to siipi)ly water enough to raise the lake to the height which the ter- races about it show that it must once Iiave liad.- It is not certain tliat he adheres to this view at present, for in his memoir on the Climatic Changes of Later Geological Times (1882), he characterizes the glaciation of the Sierra as an episode (p. 2), but regards the desiccation of the Great Basin as a continuous process of which the beginning dates far beyond the Pleistocene. On p. 190 he says: Before advancing another stage in our discussion, however, we have to make it clear that the diminution of the rivers, the disappearance of the lakes, and all the other phenomena indicative of a gradual but persistent tendency to aridity over vast areas once fertile and well watered, do not form a transient i)hase of a precedent Claeial epoch, but are the result of some cause which began to act before that period, and is still continuing without any connection with it. In my original description of Lake Bonneville I argued its correlation with the Pleistocene Period in the following language: The Bonneville epoch and the Glacial epoch were alike climatal episodes, and they oci urred in the same general division of geological time, namely, the division of which modern time is the immediate sequel. If it can be .«hown that the climatic changes were of the same kind, there need be no hesitation in assuming the identity of the epochs. The glacial climate we commonly regard as merely cold, and a low temperature was doubtless its chief characteristic; but it admits, ne ertlieless, of another view, The climatic comlition essential to the formation of glaciers is, tint the summer's heat shall be inadequate to dissipate th(> winter's snow, and this may lie brought about, either by a lowering of temperature, or by an increase of winter pre- ci[)itatio;i. The jirofuse |)recipitation of our northwestern coast woultl maintain s'leat glaciers if the climate were cold enough; rivers of ice would follow the higher valleys of the Rocky Mountains if the snow-fall were heavy. 'Louis Lartot, Comptcs Roiuliis fc. U. S. G. & G. Snrv. of Terra, for 1877, p. 641. 2 Dr. F. M. Endlirh ; Ann. Kept. U. S. G. «fc G. Snrv. of Terrs, for 187.-), p. S25. REGENCY OF LAKES AND GLACIERS. 269 than now, their humidity must have been acquired before they reached the district of the hxkes. THE ARGUMENT FROM ANALOGY. Reverting- now to the correlation of lacustrine and glacial phenomena, as suggested and developed by Jamieson, Lartet, Whitney, King, Russell, and myself, the data on which the correlation is based will l)e examined in detail. Up to the present time all reasoning on the subject has been based upon analog v. The identity of the two classes of phenomena in time and cause has been inferred, tirst, from their recency; second, from their exceptional nature; third, from the parallelism of their recurrence; and, fourth, from the belief that it is possible to account for them by the same modifications of climatic conditions. These elements of analogy will be taken up in the indicated order. Recency.-The reccncy of the lacustrine events and the recency of the glacial events are severally inferred from the excellent preservation of their vestiges. The atmospheric agencies which sculpture the land, rapidly oblit- erate all topographic features which do not conform to their types, and they attack with especial vigor masses of unconsolidated material which stand in relief. The embankments of the ancient shore-lines and the moraines of the ancient glaciers agree in their susceptibility to rapid modification by erosion, and they agree in exhibiting a condition of almost perfect preservation. In the case of the moraines, this remark applies onlj- to those which were latest formed; but it is these which can most properly be compared, for the earlier- formed shore embankments are not visible, having been overplaced by those of later date. The recency of phenomena thus demonstrated is qualitative merely: So far as we are able to interpret the evidence from preservation, the embankments may be twice as old as the moraines, or the moraines twice as old as the embankments. Episodai character.-Tlie exccptional iiaturc of the Pleistocene glacial phe- nomena is generally recognized, and is illustrated in a striking manner in the immediate vicinity of the Great Basin. As first pointed out by Whit- ney, the great glaciers of the Sierra Nevada occupied an antecedent system of valleys, shown by their form to be the product of stream erosion. The 270 LAKE BONNEVILLE. period of ice was therefore preceded by h period wlieii tlii're was no ice, or little ice, and this antecedent period was of relatively great duration. The episodal nature of the lacustrine j)lienoraena of tlie Great Basin has been recognized by all ol)servers, with tlie possil)le exception of Whit- ney; and the evidence in relation to the Bonneville Basin has been full}' set forth in the preceding pages. The pre-Bonneville period was characterized by ariditv, and it was long as compared to the Boinieville period. Tlie formation and extension of glaciers and the formation and extension of lakes have thus the common character of episodes, interrupting a course of events \\'hich was resumed after their disappearance. Bipartition.-A tlurd polut of aualogy is parallelism of reciiiTence. The history of Lake Bonneville and the history of Lake Lahontan have been independently shown to be bipartite, and the similarity of the series of oscillations in the two basins gives great contidence to the conclusion that they were synchronous. If it be true, as believed by many geologists, that the history of the glacial period is similarly bi2)artite, the argument in fa\dr of the synchronism and the common origin of the lacustrine and glacial phenomena acquires great strength. It is pertinent, therefore, to inquire what support the belief in a double glacial period finds in the facts of obser- vation ; but since this inquiry Avould involve too great a digression from the subject in hand, attention will be limited to the t^uestion of the support the belief finds in the opinion of those most competent to discuss tlie phenomena. It is to be observed at the outset that a Ijelief in the double nature of the glacial epoch implies a belief in its actuality as a general phenomenon of geologic climate. If the truth lies with those who aflinii that the ancient glacial phenomena depend upon strictly local conditiinis, and are not widely synchronous,' it is evident that the bipartition of the plieiionieiia can not be general, and that the only analogy pertinent to the present incpiirv would arise from the discovery of evidence of recurrent glacial extension in the mountain ranges which border the Great Basin. Reference will be made ' See J. D. Whitney, Climatic changes cit later Geological Time: Mem. >Siiseum of Compariitive Zoology, vol. 7, No. 2, pp. 191, 21)8, :!«7; J. F. Campbell, Glacial periods; Quart. Jonrn. Geol. Soc. Loudon, vol. 35, p. 9K; Rev. James IJrodie, On the action of Ice in what is usually termed the Glacial Period : Brit. Ass. Rep't, 1875, p. 63. (Sections.) EUEOPEAN OPINIONS ON CIPAKTITION. 271 ill the sequel to a I'riigment of local evidence of this nature; but attention •wiW at present be restricted to the testimony in regard to a general duplica- tion of glacial history. The tendency of the testimony will be sufficiently indicated l)y citing those conclusions of held geologists whidi appear to represent the liroadest survey of phenomena and to be least hampered l)y general theories. Penck, who has studied the glacial phenomena of the northern face of the Alps, has supplemented the presentation of his own results l)\' a histori- cal digest of those of his predecessors.' He confirms the recognition by Morlot and others, of two great advances of the glaciers, and announces traces of a third. The greatest advance occurred in the second of the three ice epochs, and the least advance in the first. Briickner, likewise a student of the northern face, agrees with Penck in recognizing three epochs of glaciation, but he considers the first advance slightly greater than the second and the third least of all." French geologists who have examined the western portion of the Alps are practically unanimous in asserting the unity of the phenomena. Falsan admits more or less protracted phases of progression and recession of the old glaciers, but denies the existence of any adequate evidence of an inter- glacial period.^ Those who have given special attention to the southern or Italian slope of the Alps are divided in opinion. Sto})})ani and Gastaldi regard the gla- cial period as a unit,^ while Taramelli distinguishes two phases of glacial expansion, separated by a long interval marked by hydrographic changes and slight oscillations of level.^ James Geikie recognizes no fewer than four glacial epochs, separated by intervening epochs of mild climatic conditions." In the English deposits 'The Glaciatiou of tbc German Alps. . . . By Dr. Albrecht Penck. pp. 220, ^61, 311, :5->2. -Die Eiszeit iu deu Alpeu. vou Dr. Eduard Briickner. Mittheil. Geogr. Gesell. Hamburg, 1887-88, pp. 10-12. ^A. Falsan, Esquisse gdologique du terrain errati(|ue et des anciens glaciers do la region ceutralo dn bassin du Rhone. Lyon, 1883. (Cited at second band.) Also, La piSriode glaciaire. Paris, 1889, pp. 24-2-245. ■■A. Stoppani, Geologia d'ltalia, Part 2, Milan, 1880. Gastaldi, Realo Accademia delle Scienze di Torino. Atti. 1872-73. 8°. Page 410, " Appuuti sulla Memoria del Sig. Geikie F. R. S. E , On changes of cliiiinte during the glacial epoch." sTaramelli, Atti della Reale Accademia dei Liucei, 1881-82, 3d series, vol. 13. Roma, 18S2, p. 508. ^Prehistoric Europe, p. 2G5 272 LAKE BONNEVILLE. the first -8;i3, 896. -On thu complete series of Su])erlici;il Formations in Northeastern Iowa. Hy W. ,J. MeGee. Proc. .\ni. Ass. Adv. Sci. vol. 27, 1879, pp. 198--,':il. •'Tlie Drift Deposits of Indiana, l),v J. S. Newberry; in lllli .\nn. Kep. (ieol. and Nat. Jlist. of Indiana, by .John Collett, 1884, p. 90. ■•Warren Upluuii; Proc. Am. Ass. Adv. Sei. vol. :52. 1884, pp. SJ-J, 223. See also Geol. and Xal. Hist. Snrvey of Minnesota, vol. 1 of Final Kept., 1884, pp. 40(i, 481, .'■)80. *Tlie Driftless Area of the Upper Mississippi. By T. C. Chamberlin and R. D. Salisbury. Sixth Ann. Kept. U. S. Geol. Survey, 188"., p. 212. AMERICAN OPINIONS OF BIPARTITION. 273 Epochs. Sabepochs or Episodes. Attendant or characteristic phenomena. Not yet sati-'^factorily dis- tinguished from the Plio- cene. First aubepoch or epj.'iode. ^ Interglacial subepoch or episode of doglaclation. Sucood subepoch urepisode - Drift sheet with attenuited border; absence or nieagerness of coarse ultra-marKiual drainage drift. Decomposition, oxidation, ferrugination. vegetal accumulation. Drift sheet with attenuated border; loess contem- poraneous with closing stage. Elevation of the upper Mississippi region 1.000± feet. Erosion of old dritt, decomposition, oxida- ti'tn, ferrugination, vegetal accuniulatious. Till sheet bordered by the Kettle or Altamont moraine. Vegetal deposits. Till sheet bordered by the Gary moraine. Till bordered by the Antelope moraine. Marked by teiminal uiuraines ot undetermined importance. Marine deposition in the Champlain and Saiot Lawrence valleys and on Atlantic border; lacus- trine deposits about the Great Lakes. Marked by fluvial excavation, uotahly of the flood plains of second glacial epoch. n. Earlier j;lacial epnch III. Chief interglacial epoch IV. Later glacial epoch First episode or subepoch - - Episode of deslaciatiou .. ■^ .Second 8t;ign or suhepoch . Episode of deglaciation 1 Third episode VI. Terrace epocli According to Newberry " there were two maxima of cold separated by a long interval in wliich the climate was ameliorated"; but this climate was still cool, and the ice probably did not retreat far beyond the Great Lakes/ While the conclusions of McGee, Upham, Chamberlin and Newberry are based jjrimarily on studies in contiguous districts, include to a large extent the same phenomena, and agree in recognizing two maxima of cold, those of Chamberlin and Upham are the only ones in complete accord. Newberry differs from the others in that he regards the inter-maximum ics retreat as relatively small. Chamberlin and McGee, agreeing that glacia- tion was interrupted by a long epoch of warmth, and that it was also varied by episodes of local or temporary retreat of the ice sheet, differ in their ref- erence of an important bed of till, and hence draw differently their lines of primary classification. McGee's interglacial period "of immense duration" is Chamberlin's "interglacial subepoch or episode of deglaciation ", and McGee's " second and last great glacial advance " is Chamberlin's " second subepoch " of the " earlier glacial epoch."^ By later investigation McGee finds evidence as to epochs of cold in the ]ilienomena of the deposits and erosions of the Atlantic border south of the Drift. From this investigation he concludes that the Pleistocene included 'Nortb America in the Ice Period. By John S. Newberry. Pop. Sci. Monthly, vol. 30, 1886, p. 9. ^See McGee in Am. Jour. Sci. :{(1 series, vol. 35, 1888, pp. 458-461. MON I 18 274 LAKE BONNEVILLE. two and only two great epochs of cold ; that tliese epochs were separated by an interval three, five, or ten times as long as the i)ost-glacial interval ; and that the earlier cold endm-ed mnch the longer and was the less intense.' These inferences are harmonions either with Chanil)erlin's conclnsions or with his own results in Iowa, taken separately, and they correspond closely with my reading of Bonneville history; by substituting the terms "wet" and "lacustral" for "cold" and "glacial," the Bonneville story can be summed up in tlie same words as McGee's story of the Atlantic border. Wright early advocated the unity of the period of glaciation in America and still adheres to tliat view. In a recent publication he states that " most of the facts adduced to support the theory of distiiict epochs are cajiable of explanation on the theory of but one epoch with the natural oscillations accompanying the retreat of so vast an ice-front."^ The latest word on the subject is from James Geikie,' whose digest of results obtained by geologists of continental Europe comes to hand \\hile these pages are in proof The plain of northern Germany was twice over- run by the Scandinavian ice sheet, and experienced a temperate climate in the interval. Students of Alpine di-ift recognize more than two epochs of glacier extension, and it is possible that the interglacial deposits of the northern plain do not all belong to the same interglacial epoch. From this summary of opinions it aj)pears that the relatively simj)le conception of Pleistocene history which belonged to the early stages of its investigation has been generally replaced b}- the \"iew that its climate was characterized by great oscillations. This result has been reached separately and through independent methods by European and American students. l)ut while the fact of oscillation is widely accepted for each continent, the progress of investigation seems not yet to have rendered the two histories so definite that the question of their similarity and svnclu'onism can profitably be discussed. Whatever confidence we may have that the Plei-stocene gla- ciation was a recurrent phenomenon, it must be admitted that ])arallelism of recurrence remains to be proven. It follows that, for the present at least, ' Am. Joiir. Sci. 3d series, vol. 35, la-'H, p. 403. ■^Tho Ice Age in North America. By G. Froilerick Wright. New York, l!J89, p. 500. 'Address to the Geological section of the B. A. A. S., September, 1889. GENETIC CORRELATION OF LAKES AND GLACIERS. 275 parallelism of recurrence can not with confidence be ajipealed to in tlie correlation of the lacustral history with the glacial history. Genetic correiation—Tlic fourtli point of analogy is genetic. It is generally believed that any climatic change competent to restore the g'laciers of Cali- fornia and Utah would likewise restore the ancient lakes of the Great Basin. From this belief there has been no dissent, and it is certainly plausible ; but it must nevertheless be admitted that meteorology in its present stage affords it no satisfactory basis. Tlie general subject of climate is highly complex, and Its laws are not so well understood that the results of new combinations of conditions can be foretold. The size of lakes and the size of glaciers are determined by three processes : A. Precipitation of rain and snow. B. Evaporation of water, snow and ice. C. Melting of snow and ice. The essential elements of local climate upon which the local I'ates of these three processes depend are at least four in number, and may conven- iently be indicated under five heads : (rt) The temperature of the air. (b) The vapor tension or vapor content of the air, or the temperature of the dew point.' (c) The general velocity of the wind. (d) The degree of cylonic activity ; and finally, (e) The variation of these, and the distribution of their variations through the year. ' For the iintechnical reader, these terms may stand in need of deliuition. The invisible moist- ure contained in the air is called aqueous vapor, and has the properties of .a gas. By virtue of its elasticity it exerts a certain tension, and this tension is the measure of the amount pretsent at any point. Vapor tension and ra^jor coH(eH( are therefore synonymous. The amount of moisture .air will hold without condensation is limited, and the limiting amount varies with temperature. For each temperature there is a maximum vapor tension known as the tension of saturation ; for each vapor tension there is a minimum temperature known as the drw point. The temperature of the dew point at any place and time is thus an index of the existing vapor tension. Relative humidity is the ratio of the actual vapor tension to the saturation tension corresponding to the actual temperature ; it is the humidity reckoned in terms of saturation as unity. 276 LAKE BONNEVILLE. The move general terrestrial conditions wliich immediately determine these local elements may likewise be enumerated under five heads. They are : (1) The latitude of the locality. (2) Tlie altitude of the locality, and the system of altitudes in its vicinity. (3) The distribution of land and water in a very large district includ- ing the locality. (4) The system of currents in oceans within this district (a function of 1 and 3). (.O) The wind dfrection (a function of 1, 3, and 4). Directly or indirectly, each of these five conditions affects each of thfe five elements of local climate, so that there is a most intricate plexus of cause and effect. In a qiialitative way much is known of the nature of these relations, but quantitatively very little is known. It i.s perhaps fair to say that the relations of temperature and humidity to latitude and altitude are the only ones whose numerical laws have been successfully investigated, either theoretically or empirically. Gradually the various climates of the earth are being explained and referred to their proximate causes; but the time has not come when the meteorologist can trace out the quantitative relations, or even in any fullness the qualitative relations, of a specific hypothetic change in one of the conditions of climate. Such a pr(il)lcni as the distribution of climates if the direction of terrestrial rotation wvre reversed can at present be solved only in a very rude way. In the presence of such complexity, theories are nec^ssaril}- based upon partial views, and the hypothesis or opinion that the magnitudes of enclosed lakes and of glaciers are similarly aftected by climatic changes ajjpi'ars to depend upon such a partial view. This was certainly tlu' case when I advanced the opinion in an earlier paper. Let us assume that in the region of the Great Basin and tlic surround- ing mountains the aqueous vapor, the wind v(■lo(•it^•, the cvclonir activity, and the annual oscillations of these climatic elements remain constant, while tlie tenq)erature alone undergoes variation. The cause of the tenqierature change lies of course in a modification of some climatic condition, and such modification would necessarily have its effect upon vapor, wind velocity, EFFECT OF LOCAL TEMPERATURE CHANGE. 277 etc., but this effect is by the present assumption ignored. Conceive, first, a hiwering of local temperature. The vapor tension remaining the same, the relative humidity of the air would be greater than at present; and cyclonic activity remaining the same, the increase in relative humidity would cause increase in precipitation of rain or snow. The wind velocity remain- ing the same, the lowering of temperature would retard evaporation, a smaller share of the moisture precipitated on the land surfaces of the Great Basin would return to the air, and a larger share would gather in streams and flow to the lakes. Evaporation from the lake surfaces would be sloAver, and the lakes, with increased supply and diminished dissipation, would grow deeper and broader, just as they did of old. In the mountains the lower- ing of temperature would increase the length of the season during which precipitation takes the solid form, and a greater proportion of the total precipitation would be in snow. The increased relative humidity of the atmosphere would occasion a greater total precipitation, and the winter's accumulation of snow would thus be doubly augmented. The same cause would diminish the annual evaporation of snow, and the shorter and cooler summer would have less melting power. In every way the accumulation of snow and ice would be promoted and its dissipation checked. The small glaciers which hang about some of the highest crests would wax in size and others would reoccupy the empty cii'ques, until finally a broad mantle of snow and ice would cover the high district of the Sierra, and ice streams would flow to the valleys on either side, just as of old. Conceive now a rise of local temj^jerature. Tlie relative humidity of the air would be less than at present; the precipitation in rain and snow would be less; the evaporation would be more rapid, antl a smaller share of the diminished precipitation would gather in streams and flow to. the lakes. The lakes, with decreased supply and increased dissipation, would grow shallower and smaller. In the mountains the winter would be shorter, and a smaller share of the diminished precipitation would take the form of snow. The evaporation of snow would be more rapid, and the longer and warmer summer would have greater melting power. The supply of snow wovild be diminished and its dissipation would be promoted. The existing small glaciers would disappear. 278 LAKE BONNEVILLE. Let us now assume that in the same region the temperature, wind velocity, etc., remain constant, while the vapor tension alone undergoes variation. Conceive, first, an increase of local vapor tension. The tem- perature remaining the same, the relative humidity of the air is increased, and this increase in relative humidity causes increase in precipitation of rnin and snow. It induces also a slower evaporation. The supply of water to the lakes is increased, their superfcial waste is diminished, and they grow in size. On the mountains the snowfall is increased, though its period remains the same. The dissipation of snow by evaporation is less, the melting of snow l)y direct insolation is sensibly unchanged, l)ut its melting by sunmier rains is accelerated. In the region of the Sierra glaciers the sunnner pre- cipitation is so small as compared with the winter that this last factor can not be important; and we need not doubt that accumulation of snow would exceed dissipation, causing an extension of the glaciers. Conceive now a diminution of vapor tension. The preceding relations are evidently reversed. The lakes of the Great Basin receive less from the streams and part witli more to the air, and therefore shrink. The glaciers of the Sierra receive less snow, lose more by evaporation and lose slightly less l)y melting, and they will therefore shrink. It thus appears that a local change in temperature alone or a local change in moisture alone would cause the lakes of the Great Basin and the glaciers of the Sierra simultaneously to enlarge or simultaneously to con- tract. But wdien Ave consider their concurrent change, no such definite conclusion is possible. If rise of temperature is accompanied bA- diminu- tion of vapor tension, there will be a common shrinkage of lakes and glaciers, for these climatic changes have the same tendency. Similarl\-, if fall of temperature is accompanied by increase of vapor tension, lakes and glaciers will grow; but a rise of temperatin-e and an increase of vapor, or a fall of temperature and a decrease of vapor, will have antagonistic efiects upon both lakes and glaciers, and the nature of their resultant can not be determined without quantitative data. We need greatly to extend our knowledge, not only of climatic laws, but of the climate and phvsical geogra})hy of the Great Basin, to enable us to determine wliar increase of vapor tension is adequate to neutralize the effect of one degree's rise of ^ EFFECT OP LOCAL HUMIDITY CHANGE. 279 temperature upon the size of the hikes; and we need in addition greatly to extend our knowledge of the climate of the Sierra Nevada to enable us to determine what increase of vapor tension will neutralize the effect of one degree's rise of temperature upon the size of the glaciers. It is only in the case that these two increments of vapor tension are equal, that increase of lakes and increase of glaciers will be invariably coordinate. If they are unequal, then it is possible to assume simultaneous changes of tempera- ture and vapor tension under whose influence the lakes will expand, while the glaciers shrink, and vice versa. But this view of the case is still only partial. Any change in the alti- tude of the district, in the position of the adjacent coast of the Pacific, in the nature of the currents of the North Pacific, or in the direction of the prevailing wind, would not only modify the temperature and humidity of the district under consideration, but would affect the wind velocity, the cyclonic activity, and the cycle of annual climatic change. A variation of wind velocity Avould make itself felt in the rate of dissipation of lakes and glaciers; a variation in cyclonic activity would manifest itself in the supply of Avater and .snow to lakes and glaciers; and a variation in the anmial cycle of climate might affect lakes and glaciers not only miequally but diversely. Too little is known of these last mentioned influences to warrant any attempt to discuss them here. For this reason, and for this ouh', they will be ignored in the following paragraphs; but it is understood that the con- siderations about to be advanced are subject to whatever modification -^ev- tains to the omitted factors. Restricting attention to the two elements of local climate, temperature and vapor tension, we will now endeavor to ascertain how the lakes and glaciers of the district would be affected through them b}' various postulated changes of climatic conditions. Let us inquire, first, what will result from a general change of altitude, or more specifically, from a bodily uplift of the entire district, including the Great Basin and the adjacent mountains. It is Avell known that both temperature and A-apor tension are inverse functions of altitude; the tem- perature of the district will be lowered by the uplift, and the moisture normal to the new altitude will be less. The atniosphere covering- this 280 LAKE BONNEVILLE. distru't is part of a great eastward-tending current which derives its moist- ure from the North Pacific Ocean. The hypothetic change of altitude will not affect its humidity where it enters the district. Its vapor tension can be reduced to the noniial only by precipitation, and if not thus reduced, there will l)e an increase of relative humidity, owing to the lowering of temperature. We shall have, then, for the district, either an increase of precipitation or an increase of relative humidity. The former woidd aug- ment the supply of Avater for the lakes and of snow for the glaciers; the latter woidd retard evaporation and thus diminish the waste of water and ice. The loAvering of temperature likewise will not only retard evaporation, but will retard melting, and will extend the season in which precipitation takes the form of snow. Thus, in every way, the growth of lakes and glaciers will be favored. Conversely, a general depression of the district will diminish lakes and glaciers. Let us inquire, in the second place, how the climate will be affected by changing the distribution of land and water. Evidently, the number of different changes which might be postulated is unlimited, but there is one particular change to which the district is peculiarly sensitive, and which may stand for a large class. This change is an eastward or Avestward move- ment of the coast line of California, so as to diminish or increase the belt of land between the Sierra Nevada and the ocean. Let us postulate a west- ward movement, or an increase of the land. The general movement of the atmosphere in this region is from the ocean to the land, and the moisture gathered from the surface of the ocean is the store whence all the precipita- tion of the land is derived. The addition of a belt of land will inci-ease the area of uncompensated precipitation, and will thus duninish the general vapor tension of the atmosphere of the district. It has been pointed out by Button,^ that the portion of the ocean under consideration has a temjierature lower than the normal for the latitude, so that the air current grows warmer in passing over the land. The intervention of an additional belt of land will add its quota of heat to the air, and thus render the general tempera- ture of the district higher. An addition to the coast will therefore induce 'On the cause of the arid climate of the western portion of the United States, by Capt. C. E. Uutton, Am. Jour. Sci., :!d scries, vol.22, p. 247. See also, Haun's Handbuch der Klimatologie, p. 13(>. EFFECT OF HYPOTHETIC OCEANIC CHANGES. 281 a diminution of vapor and a rise of temperature, and these changes, as we have .seen, are competent to diminish lakes and glaciers. The reverse effects \\ ill of course be wrought by a dimiiuition of the coast area. Tliird, let us endeavor to see how our district would be affected by a moditication of ocean currents, The influence of such currents u^ion cli- mates is exerted through their temperature; and we will postulate a rise in the temperature of the current which follows the coast of California from north to south. A warmer ocean will give a higher temperatvxre to the land- ward-flowing air, and at the same time impart to that air a greater load of aqueous vapor. Since the oceanic district in question is now cooler than the land district whose atmosphere it tempers, a warming of the ocean will tend to diminish the contrast of temperatures. The warming of the air during its landward progress will therefore be less, and there will be a tendency towards a higher relative humidity. Precipitation will thus be promoted. Evaporation will be favored by the higher temperature, but opposed by the higher relative humidity; and it is not easy to see which tendency will prevail. The melting of snow and ice will be promoted both by the higher temperature and by the greater length of the summer, while the winter, or the season in which precipitation takes the form of snow, will be shortened. So long as only a small change is considered, the merely qualitative statement does not clearly show whether the increased rate of snowfall will he more or less than compensated by the increased rate of melting; and the uncertainty in regard to evaporation leaves us in doubt whether the lakes will swell or shrink. If, however, we pass to an extreme case, there is no room for doubt. A great increase of oceanic temperature, say ten or twenty Fahreidieit degrees, would reverse the contrast of temperature between land and shore. The eastward-flowing air, instead of being warmed by the land, would be cooled; and the resulting pi-ecipitation would far surpass any possible in- crease of evaporation. The Great Basin would become a basin of great lakes. The same temperature change would so abridge the winter season in the mountains, and so enhance the melting power of the sunuuer, that no glacier could possibly survive. The converse follows. 282 LAKE BONNEVILLE. Finall|)', let us ask what will result from a change iu the direction of the generixl air current. This direction belongs to the great syst^jin of atmospheric circulation, and a large change is practically out of the ques- tion. We are at liberty, however, to assume small changes, based upon local conditions; and -we A\ill |)ostulate that the wind becomes more south- erly. With such a course, it will derive its temperature and moisture from a portion of the Pacific Ocean warmer than that now traversed by it; and the ])rincipal effects in the mountain district under consideration will be identical with those deduced in the last paragraph, as resulting from a warmer ocean. Minor effects will be conditioned by the configuration of the belt of land traversed by the wind before reaching the interior district, and the distribution of climate within the district will be modified; but the probable importance of these considerations is not sufficient to warrant their discussion. It appears, then, that lakes and glaciers would simultaneously increase if the district as a whole Avere to be uplifted, or if the Pacific Ocean were to encroach upon the California coast; and the conclusion is less confidently reached that the lakes of the Great Basin would increase, and the glaciers of the Sierra Nevada decrease, if the North Pacific Ocean Avere wanner, or if the coastward Avinds traA'ersed a Avarmer tract. But the subject is by no means exhausted. We mig-lit consider the various combinations of these four postulated changes of condition, or, going beyond them, Ave might turn our attention to those more remote causes of change to Avliirh theories liaA-e appealed in explanation of Pleistocene glaciation. Whether Ave attempted to trace out the consequences of far-reaching geographic changes, of varia- tions in the eccentricity of the earth's orbit, or of the terre.striid \\ andcring of the earth's axis of rotation, Ave should equally find ourseh-es iuA-olved in a maze of complexity, and ultimately brought face to face Avith the imjier- fection of the science of meteorology. RevieAving the innnediately preceding discussion, avc see that tlu' partial view Avhich takes account of temperature onh', or of a(pie(ius \apor only, results in a definite conclusion. The Ijmadcr but still partial aIcw wliicli takes account of temperature aud iupicdus Aapor conjointh-, l)ut neglects other climatic elements, leads to no definite conclusion. Certain climatic ARGUMENT FROM CLIMATIC CHRONOLOGY SUMMED. 283 conditions, manifesting themselves through temperatm-e and liumidity, affect hxkes and ghxciers in the same way, while other climatic conditions affect them in opposite ways. Reviewing- the entire discussion of climatic analogies, we are forced to the conclusion that the weight of the analogic argument for the correlation of lakes and glaciers has been overestimated. The fact remains that the lake epoch and the ice epoch belong- to the same short division of geologic time; so does the further fact that each was a peculiar episode, interrupting a distinct and ver}' different course of events. These two facts establish a presimiption in favor of their correlation, but this presumption gains only moderate support from the parallel bipartition of the two sets of phe- nomena, since the duality of the glacial epoch is not generally accepted; and it gains no su]:)port, as we have just seen, from the consideration of the climatic conditions affecting the lakes and glaciers of the Great Basin. The correlation of the phenomena remains as a working li3-pothesis, but before it can regain its position as a fully credited theory, it must be sustained by new arguments. Fortunately, the data for its further discussion have been developed by tlie geologic researches in the Great Basin, and to these data we shall presently proceed. THE EFFECT OF A CHANGE IN SOLAR ENERGY. The jjresent place, however, is more convenient than any other for the discussion of a climatic question Avhose answer is of prime im2)ortance in the interpretation of the geologic dat.t just referred to. The question is that of tlie influence of a general change of temperature upon the growth of glaciers. If the radiant energy of the sun were to becf)me greater or less, how would the glaciers of the earth be affected? Would an increase in the accession of solar heat, or would a decrease in its accession, cause the present glaciers to expand and new areas to be glaciated? It is a familiar fact that the glaciers of the present day are restricted to regions where the temperature is low. They are more immerous and of greater size in polar regions, and there oidy do they reach the ocean; in temperate and tropical climates they occur only^ on high mountains, and their lower limit varies with the altitude, being highest at the equator and lowest at the poles. These facts of distriljution have occasioned the preva- 284 LAKE BONNEVILLE. lent (tpiiiiou that cold is the ])riiuary condition of g'laciation, and that the climate of the glacial epoch or ejjochs was a cold climate. If it were believed by all, as it is by some, that Pleistocene glaciation was produced by a va- riation in solar radiation, the majority would conceive that variation as a diminution. N(;vertheless, there are not wanting iiivestigators who enter- tain the opposite view; and so long as these include men of such weight as Frankland,' Tyudall,- GrolV King,* Whitney,^ and Becker," the majority should at least refrain from dogmatic assertion. I am therefore not content, as one of that majority, to let the sul)ject pass with a mere expression of opinion. Generally speaking, the vapor tension of the atmosphere is greatest at sea level, and it decreases rapidly upward. If the air did not circulate, but remained stationary, the elastic force of the aqueous vapor would cause it to be diffused ujiward, and the product of evaporation from the ocean surface would be continuously added and diffused until there was complete saturation throughout. The theoretic static condition of the atmo.sphere with reference to moisture is one of saturation. The actual condition of imperfect saturation is caused by the vertical movements of the air. These, in accordance with well known laws, produce precipitation, and it results that the vapor tension of the air at every level is, generally speaking, con- siderably below the tension of saturation. Strachey, and afterward Ilann, by studying the records of numerous observations at different altitudes and in diffei'ent rescions, have deduced the g'eneral law of vertical distribution of moisture.^ It is, that the relative humidity of the air is not a function ' Oa the physical cause of the Glacial Epoch, By E. Fraukland. Philosophical Magazine, vol. 27, 1864, p. 321. ^Tbe Foruis of Water, by John Tyiidall, p. 151. Also, Heat considered as a Mode of Motion, Ch.ap. VI. •'Climate and Time in their Geological Relations, By James CroU, New York, 1875, p. 79. ■•The Geological Exploration of the Fortieth Parallel, by Clarence King, vol. 1, p. 52.'i. 'The climatic changes of later geological times, by J. D. Whitney, Mem. Mus. Comp., Zool. vol. 7, No. 2, pp. 20.^-6, ;i21, '.iSS. <■ Temperature and glaciation, by G. F. Becker, in American Journal of Science, 3d series, vol. 2G, pp. 167-175; also vol. 27, pp. 473-476. 'Ou the distribution of aqueous vapor in the upper parts of the atmosphere, by Lieut. Col. Richard Strachey, F. R. S., Proceedings Royal Society of Loudon, vol. II, 1860, p. 182. Ou the diminution of aqueous vapor with increiisiug altitude in the atmosphere, by Dr. Julius Hann, Zoitschrift Oest. Met. Gesell., lrJ74, vol. 11, p. 193. (Cited from translation by Cleveland Abbe in Smithsonian Report for 1877, p. 376.) Strachey notes that the conclusion was originally reached by Ur. Joseph Hooker, but Hooker's inforeuce was based ouly upon observations in the Himalayas. EFFECT OF GENERAL TEMPERATURE CHANGE. 285 of altitude, or, in other words, that for each altitude the vapor tension bears the same relation to the tension of satui-ation. It is not to be supposed that this law is ordinai'ily illustrated by the condition of a local atmospheric column at a given instant; it is exemplified only through the comparison of the means of large bodies of observations. Notwithstanding the empiric nature of this law, it is possible to extend its application somewhat beyond the existing order of things; for it is evi- dent that under the influence of atmospheric circulation the humidity of each isothermal and isoliygral stratum of the atmosphere is determined by the humidify of the stratum beneath it, the humidity of the lowest of all being determined by the rate of evaporation from the surface of the ocean. A universal rise in the temperature of tlie atmosphere, unless it was suffi- cient to materially accelerate the circulation, would have the effect merely of raising all the isothermal strata and inserting a warmer stratum at the base of the series. This, by virtue of its higher temperature, would accel- erate the oceanic evaporation, and thus be enabled to maintain the relative humidity required by Strachey's law. Tliis conclusion implies that rates of oceanic evaporation are proportional to the saturation tensions of the air at the surface of the ocean, so long as the relative humidity is unchanged; a proposition readily deducible from the accepted law of evaporation.^ In stating the above propositions, it has not been possible to incorporate continuously the qualification that they are of the most general character and ignore the extreme variability in time and place wliich characterizes both temperature and humidity. Despite this qualification, they appear to ' In an article "On the depeniloiice of water evaporation ou tlio temperature of the water anil the movement of the air", published in the Repertoriiim fur Meteorologie, St. Petersburg, l-^TT, Article 3, p. 6, Stelliug deduces and applies the follovvinj; formula: » = A(S — 8) -t-B (S — ?)"•> in which v is the rate of evaporation, S is the saturation vapor tension corresponding to the tempera- ture of the evaporating water, .v is tlie vapor tension of the air in contact with the water, ?« is the velocity of the wind, and A anil 15 are constants. Siuce for the present purpose wo may ignore local variations, we are enabled to simplify the formula by regarding the contiguous air and water as of the same temperature, and by regarding the wind as coustaut. With this modification the formula becomes: » = Constant X (S' — s), or r = Constant x S' (1 — ^,)^ in which S' is the saturation tension of the air. The fraction |^ expresses the relative humidity, and since this is by postulate constant, we have t', the rate of evaporation, a simple function of S', the sat- uration tension of the air. 286 LAKE BONNEVILLE. me to warrant the following corollary. If a general rise should take place in terrestrial temi)erature, affecting all local temperatures alike, tlie local moisture condition would be similarly affected. The local capacity for moistun^ hcing everywhere greater, the local vapor tension would likewise be greater, but the relative humidity for each locality would remain the same. The evaporation not only from the ocean, but from lakes and sur- faces of ice and snow, would be increased in the ratio of the increase in the local saturation tension. The increase in capacit}' for moisture for every unit of temi)erature change is not in precisely the same ratio at all temperatures, being somcwluit less for high temperatures. But the difference is so small that no material error is introduced by saying that the evaporation of moisture from the entire earth's surface is proportional to the saturation tension corresponding to the mean temperature of the surface. Since the total evaporation is precisely equal to the total precipitation, it follows that the latter likcAvise is a simi)le function of the saturation tension, and the distribution of temper- ature remaining the same, the local precipitation follows the same hnv of change as the local evaporation. Up to this point it has been assumed that the movements of the atmos- phere in direction and velocity are unaffected by a general change of tem- perature, and it now remains to consider the validity (^f this assumption. The rate of evaporation is known to depend in part on the velocity of the wind, and the rate of precipitation is known to depend in part upon the amount and intensity of cyclonic action. We will give first consideration to wind velocity. The mean temperature of the surface of the earth, reckone(l from the freezing point of water, is about + IG° (J. The absolute zero of tempera- ture is considered to he — 273° C, so that the mean absolute temperature of the earth's surface may be taken as 289°. If the constitution of the atmosphere were fixed, it is prol^able that there would be required, to in- crease of temperature of the earth's surface by 10°, an augmentation of solar heat amounting to ^ or i of the present amount. In fiict, however, the constitution of the atmosphere is variable ; at higher temperatures it con- tains a larger amount of aqueous vapor, and its power to absorb and retain GLACIATION AND SOLAR EADIATION, 287 heat and thus acquire temperature is reciprocally augmeuted l)y aqueous vapor. For this reason, tlie ratio of solar radiation to lie added tor 1(P rise of temperature is something less than ^^. l^eing unable to evaluate this qualification, we shall make use of the fraction unchanged, with the under- standing' that it is too larofe. Owiuff to the difference in attitude of tlie various portions of the earth with refei^enee to the sun, the distribution of solar energy is unequal, and hence arise the ])rincij)al contrasts of tempera- ture on the earth's surface. These contrasts cause the atmospheric circula- tion, by means of which a partial equalization of temperature is eff"ected. The difference between the solar energy received in high latitudes and that received in low, or the diff"erential solar energy, is the force manifested in the winds, and its work is the friction of the circulation. The differential energy is directly proportional to the total solar energy. The law of aerial friction is not known, but it is commonly assumed to be a function of the square of the velocity. If this assumption is coi'rect, then the square of the velocity of circulation varies as the solar energy, and an increment of ^ in solar energy will produce an increment of i in velocity. Considerations connected with the conveyance of heat through the circulation of moisture show that this estimate is somewhat too large, but as we are unable to give them a quantitative expression, we pass them by. The formula for rate of evaporation given by Stelling (see note to page 285) makes that rate a direct function of the velocity of the wind, but in such way that on the average the rate varies only about ^ as rapidly as the wind. The ratio of wind acceleration for 10° rise in the mean temperature of the earth's surface being less than j.'^, the ratio by which evaporation would be accelerated through wind velocity by the same rise of temperature is less than j^^. The smallness of this ratio assures us that the acceleration of the wind may safely be disregarded in a discussion of such general changes of tempera- - ture as may reasonably be postulated to account for Pleistocene glaciation. The conditions under which cyclones are generated are comparatively obscure; but in the ultimate analysis they are necessarily referred to differ- ential temperatures created by the sun. It is probable, therefore, that, like the general winds, they would be affected little by a general rise in the temperature of the atmosphere. 288 LAKE BONNEVILLE. It is to be noted that an increase in wind velocity, by increasing evaporation, would raise the relative humidity, and thereby increase the preci})itation. An increase in cyclonism, on the other hand, by increasing precipitation, would decrease the relative humidity, and thereby increase evaporation. The conjoint effect upon evaporation and precipitation is therefore cumulative, while the effect on relative humidity is, at least par- tially, compensatory. Finding no ground for important ([ualificatron on account of varpng intensity of atmosjjheric circulation, we return to the original deductions as substantially accurate: First, a general rise of terrestrial temperature will increase evaporation, general and local, in the ratio of the saturation ten- sions corresponding to the initial and final temperatures. Second, it will increase precipitation, general and local, in the same ratio. We are now prepared to discuss the immediate conditions of glacier growth, and will first consider a region in which the temperature never rises above the freezing point. In such a region, the only factors affecting the accumulation of snow are precipitation and evaporation. If the former is in excess, there is an accumulation, and its amount is measured by the difference of the two factors. Since each of these factors follows the same law in regard to temperature, that law applies also to their arithmetical dif- ference; and a change in the mean annual temperature will affect the snow accumulation in the same ratio that it affects the saturation vapor tension of the air. If the temperature rises so as to exceed the centigrade zero during a portion of the year, the annual cvcle of climate becomes immedi- ately divided into two portions, which it will l)e convenient to call winter and sunnner. Snow accumulation, then, has a higher rate, by reason of the higher temperature, but tliis higlicr rate is restricted to a sliorter period. With progressive advance of amuinl mean tcmperatiu'i', the rate of snow accumulation is progressively increased, while its period is progressively shortened, until finally, when the annual temperature c3-cle falls entirelv above the freezing ])oint, snow accunudation ceases altogether. As soon as the temperatm-e cycle includes sunnner, a thii-d factor is inti'oduced — melting. Snow is melted in part by contact with warm air, in part by heat radiation from the lower part of the atmosjjhere, in part by GLACIATION AND SOLAR RADIATION. 289 direct insolation, in part by the heat liberated in the formation of dew, and in part by warm rain. Tlie rate of melting is thns a complex function of the temperature of the air, the humidity of the air, the clearness of the sky, and the temperature of the rain. But these four factors are so i-elated among- themselves that a single one, the temperature of the air, may fairly Ije regarded as the measure of the rate of melting. The temperatui-e of the lower air is itself conditioned by the clearness of the sky, the humidity of the air is, broadly speaking, conditioned by its temperature, and the temperature of the rain is conditioned by that of the air. The total annual loss by melting depends likewise on the length of summer, and for present purposes its measure may be assumed to be the product of the length of summer into the mean temperature of summer, exjjressed in centigrade degrees. For the purpose of bringing together the conclusions of the j^i'eceding paragraphs, we shall now resort to a gi-apliic method. By the aid of a few temporary postulates, the law of snowfall and the law of snow-melting may each be given the form of a curve, and the relation of these curves will exhibit the law of n^vd accumulation. In Fig. 35 the line X X' is a scale of temperatures, each point rep- resenting a mean annual temperature of a particular district. The tempera" ^lu-es are reckoned in centigrade de grees, and at every tenth degree a vertical is erected. Vertical distances represent rates of snow accumulation and of snow melting. For the con- struction of the curves, three postu- lates were made. First, that whatever the mean temperature of the locality, its temperature range or the difference between the mean temperatures of its coldest and warmest months is 20° C. Second, that its annual curve of temperature change is of the usual type for cold regions. Third, that the rate of precipitation is uniform throughout the year. The line C D E is the curve of snow accumulation. For all temperatures below —10° its MON I 19 — AO —to O +IO' Fig. 35. — Fir.st Diagram of Glaciation Theory. Hori- zontal distances repl'esent Mean Annual Tenii)erature in Centigrade deffi-ee.s. Tlie ordinate^ oi C D E are rates of Snowfall (leas evaporation). Tbo ordinates of A B are rates of Melting. 290 LAKE BONNEVILLE. ordinates .ire proportioned to the corresponding saturation tensions. For each point between — IC and +10'^, the ordinate represents the product of the corresponding saturation tension ])y tlie k^ngth of winter, expressed as a fraction of the year. The hue A B is the curve of mehing. Each of its ordinates represents, for the corresponding mean annual temperature, the product of the lengtli of sunruner into the mean temperature of summer. To the left of A it coincides with the axis A X. Each of these curves rep- resents a system of ratios, and the unit in each system is arbitrarily assumed. Any other assiimption of relative magnitude might have been made ^^•ith equal propriety, but such assumption would not affect the essential charac- ters of the curves. Since each ordinate of the curve C D E represents a rate of snow accumulation, as affected by precipitation and evaporation, wliile each ordinate of the curve A B represents a rate of melting, the diti'erential ordinate included between corresponding points of the two curves (to the left of their intersection) represents that portion of the winter's snow which survives the summer's melting. It represents the net accumulation. Its maximum value is at A, corresponding to the mean annual temperature of — 10°. With progressive fall of temperature it diminislies, at first rapidly and afterward slowly. With progressive rise of temperature it diminishes, at first slowly and afterward rapidly to the point of intersection, I. We may now, before drawing final conclusions, examine our postulates, and inquire what errors they introduce. In addition to tliose stated above there are several implied postulates Avhich are worth}' of consideration. First, it is assumed that the annual temperature range, or, more pre- cisely, the range of the monthly means of temperature, is 20° C. This is not far from the average temperature range in existhig glacier regions, but there are some localities where the range is somewhat less, and others where it is much greater. The assumption of a different range would produce in the diagram a pair of curves diftering in proportions but identical in type. Secondly, it is ])Ostulated that a clinnge in the general tenq)erature is not accompanied by a change in tlic local annual temperature range, or, in other words, that the temperature i-ange is constant, 'i'he 2«'ecise nature of errors introduced by this postulate is not easily seen, but considerations EXAMINATION OF POSTULATES. 291 analogous to those to which attention was called in discussing' the variations of wind velocity suggest that a rise in general temperature would produce a slight expansion of local temperature range. The corresponding corrective modification of the curves would fall entirely to the right of the ordinate A D, and would be unimportant. Thirdly, it is postulated that the local annual curve of temperature is of the type usually observed in cold regions. If observation afforded us information in regard to the temperature cycles of n('\(' districts, their type would be the one to employ in the construction of our curves; but there is no reason to believe that the error incurred by our ignorance of this point is considerable. Fourthly, it is j)ostulated that the curve derived from the monthly means fully represents the temperature oscillations of the year. This is manifestly untrue, for not only is there a diurnal oscillation, often comparable in range to the annual, but there are also non-periodic oscillations of considerable magnitude. It is a matter of ordinary experience that a melting of snow often takes place during the warm jiortion of a day whose mean tempera- ture is below the freezing point, and that precipitation sometimes takes the fonn of snow during the cold part of a day whose mean temperature is above the freezing point; and that snows may fall in tlie midst of summer and thaws occur in the midst of winter. Thus the actual temperature range in any individual year is greater than the range obtained by the method of monthly means. It is impossible to make satisfactory allowance for this in the construction of f)ur cur^'es, for the reason that the importance of the diurnal and non-})eriodic oscillations varies greatly with latitude and with distance from the ocean. The curves as drawn represent sufficiently well the relations of snow accumulation and melting at maritime stations, but not at interior stations. The general nature of the modifications necessary to adapt them to interior stations is easily indicated. With the mean annual temperature at 0° C, the ratios of precipitation and melting are tmaffected by the neglected oscillations. With the mean annual temperature at or near — 10°, the ratio of precipitation is diminished and that of melting increased. With the mean annual temperature at + 10°, the ratio of pre- cipitation is increased and that of melting diminished. The application of 292 LA.KE BONXEVILLE, these corrections to the diaf^'ram would lower the curve C D E in the im- mediate vicinity of D, smoothing- out the angle at that point, would leave it unchanged where it intersects the ordinate of 0°, and would carry the point E farther to the right. It would raise the curve A B at A, and lower it at B, leaving the central })ortion unchanged. The j)oint A, or the intersection with the horizontal axis, woidd be thrown to the left. Fifthly, in the construction of the cm-ves no allowance was made for evaporation during summer. The curve 1) E includes onh' winter evapora- tion, the curve A B only summer melting. The rate of evaporation for snow and ice has its maximum at 0°, its law changing at that point. In the general law for aqueous evaporation, the rate of evaporation is a func- tion of the difference between the saturation tension corresponding to the temperature of the evaporated substance and the actital vapor tension of the evaporating air. Since snow and ice can not rise in temperature above 0'^, thev can only be evaporated when the aqueous tension of the air in contact with them is less than the saturation tension ior 0°. If it rises above that, moisture is deposited on the ice as dew, instead of being abstracted from it. In all but very exceptional cases the range of summer tempera- tures under which nevd can evaporate is small — from 0° to 5° or 6°. The effect of the evaporation is to retard the wasting of the ice, for the energy consumed by it is deducted from that available for melting, and a unit of solar heat can melt seven times as nnich ice as it can evaporate.^ The cor- rection, if applied to the curve of melting, would slightly increase its upward concavity. Sixthly, the winter evaporation embodied with the winter precipitation in the curve D E is tacitly assumed to have a rate corresponding to the mean annual teini)erature; its rate is reallv less, being a function of the mean winter temperature. Au error is thus manifestlv introduced, and this error is greatest for the amuial tem})eratures corresponding to short winters. A corresponding correction of the diagram would raise the line D E by amounts increasing ])rogTessively from D to E. 'The conditioriH deteriiiinin<; tlie evaporation of ice .iiid the formation of f the Bonneville shore-line. The absence of Intermediate shores tells us that it was completed after their date. A portion of the mole may have been thrown up in the earlier part of the second lake epoch or at any previous time, but if so, it was completely buried by the product of the final eruption at the time of the Bonneville shore-line. This determination of date de])ends on our knowledge of the shore- line history derived from other localities, but the same information may be obtained from data purely local. At numerous points on the north side there is exhibited an unconformity in the bedding of the tufa, and a study of this unconformity shows that after the waves liiul notched the profile on tluit PAVANT BUTTE. 327 side, producing a sea-cliff and a terrace, the renewal of eruption partially filled the notch, the newer layers dipping at a higher angle than the old. We thus learn by consistent and cumu- lative evidence that an eruption took place here while Lake Bonneville was at its liio'h- o est stage, and beneath a body of water 350 feet deep. The resulting cone was built not only to the surface of the water biit 450 feet '^ Fig. 37.— IMa^am to illustrate the Alter- higher. Eruption ceased with the fall of °''«<"' "f voicauic Eruption and Littorai Ero ^ ■*■ aion on Pavaut Butte. the water and has not been resumed. Notwithstanding the recency of the cone, its sides are conspicuously furrowed by erosion, and it is in that respect contrasted with most frag- mental volcanic cones of the vicinity. Where the lapilli are uncemented, all rain is swallowed by the interstices, and escapes gradually and quietly at the base. On Pavant Butte this is prevented by the cement, and the rain flows down the surface, accomplishing its usual work of erosion. The sides of the furrows exhibit to some extent the internal structure of the mass, and show it to be a fine type of its kind. There ai'e no partings between the layers of tuff, but lines of deposition are plainly to be seen, and these exhibit on the inner side a dip toward the crater at 35 degrees, and on the outer face an opposite dijj of from 15 to 25 degrees, the two systems being joined along the crest by anticlinal curves. A figure illus- trating this arrangement is here reproduced from the Wheeler report (Fig. 38). The general distribution of yellow and gi'ay colors indicates that the yellow is original and the gray a result of weathering. The sections exposed by recent erosion show the main mass to be yellow, but there are occasional thin bands of gray, and these are inferred to record the temporary cessation of eruption. The old sea-cliff against which the newer tuff rests unconformably does not show the gray color, a fact consonant with our belief that the latest eruption interrupted rather than followed the destruc- tive work of the Bonneville waves. Fig. 38.— Section of Pavant Butte. O=0at8ide of Crater. /^Inside of Crater. B=Bonneville shore-line. 328 LAKE BONNEVILLE. •v^-- - - ^,- /■ iXwr Fir,. 39.— Section at base of Pavant Bntte, showing Eeninant of earlier Tuff Cone. The dotted lines indicate theoretic sMiictnre of parts concealed or removed. From the northwestern base there jut a number of ragged spurs, con- sisting, like the main mass, of tuff, but exhibiting dips toward tlic liill instead of from it. A study of their dips shows that the spurs are remnants of an older crater rim, on whose ruins the surviving rim was built. The diagram, Fig. 39, shows by full lines the observed relation of dijjs, and Ij}- dotted lines the theoretic structure of the parts concealed or re- moved. The earlier crater was somewhat smaller than the later, and its center was forther noi-tli. Tlie tuff ex- hibits, throughout, the gray color referred to weathering. The date of the structure is uncertain. Its tuffaceous^ character indicates subaqueous eruption. Its color suggests prolonged exposure to the atmosphere after the chief work of demolition was perfonned. It may have been built during the earlier part of the epoch of the "WHiite Marl, while the oscillating lake Avas beginning the for- mation of the Intermediate shore-lines, or still earlier in the epoch of the Yellow Clay. The surface of the plain for a short distance in all directions from the cone is composed of debris derived from it. Beyond this southward outcrops the White Marl, and beneath the White Marl a field of lava. The White Marl seems to be but two or tln-ee feet thick, and as there appears no reason why the open plain at this point should not receive the full deposit, it is inferred that only the upper portion is visible, the lower being beneath the lava. As the Bonneville and Proro shore-lines are contemporaneous with the upper portion of the Marl, the question arises whether the lava bed may not be contemporaneous with the later tuff, and derived from the same vent. The surface of the lava is as perfectly preserved as that of the Ice Spring field, but is of an entirely different type, corresponding to tlie pahoehoe of the Sandwich Islands. It exhibits fine examples of the curved convolutions ' Tufa and tuff, etyinologically the same word, have both been used to designate a calcareous deposit from solntioii ami alHO a cohcront aggregate of lapilli. Following GiMkie, I have in the.se pages allotted the two words in .wveralty to the two functions, applying tnfa and tu/aceous to the deposit from .solntion. ami luff -.mil tuffaceoua to the volcanic product. ERUPTION BENEATH LAKE BONNEVILLE. 329 or wi-inkles that are so suggestive of coils of rope. At the time of my exam- iuatiou I was disposed to refer these to the inter-Bonneville dry epoch, for it appeared to me a priori that a lava stream flowing beneath the water would part with its heat so rapidly that its smooth sui'face would be shat- tered into fragments. But I am informed by Captain Button that where Hawaiian lava streams of the smooth type have entered the sea, their surface characters have not been affected. The evidence comprised in the thinness of the White Marl and the perfect preservation of the lava surface beneath it may therefore be accepted as showing that a lava was here spread under the water during the second lacustrine epoch ; and the close association of the field with the Pavant tuff is probable. Its area is undetermined, for it is overlain not only by the marl, but also in places by a belt of sand dunes. In a southwesterly direction it is visible at intervals for several miles. TABERNACLE CRATER AND LAVA FIELD. The typical phenomena of the Ice Spring and Pavant localities simplify the interpretation of the Tabernacle eruptions. The Tabernacle field lies immediately south of the Ice Spring, and is mapped on PL XXXV. It is approximately circular, with an average diameter of tlu-ee miles and an area of about seven square miles. The point of issue is not central biit lies near the southeast margin. The crater has two rims, an outer and an inner. The outer rim is the older and is composed chiefly of yellow tuff. It contains also some slag- like material colored dark red and grey. Its contours, which are in detail the result of weathering, are smooth, except where broken by slaggy crags. Its surface is largely composed of discrete lapilli, just beneath which the tuff may be found in place. Two-thirds of the original annulus is preserved, the part toward the northwest having been absorbed or buried by later eruptions. The span of the annulus from crest to crest is 2200 feet, and the ridge is highest on the east side, where it rises 120 feet above the lava field. Probably a part of its base is concealed by the lava. Its profile as seen from the Miter crater (PI. XXXIX) resembles the Mormon Tabernacle at Salt Lake City, sug'gesting an appropriate name. The internal structure of the ridge is not well displayed, but an outward dip was observed in the higher part. 330 LAKE BONNEVILLE. The inner rim is characterized by a great abundance of scoriaceous matter that evidently reached its position while still pasty and adhesive. It is not greatly inflated, and its general habit is rather slaggy than scoria- ceous. The rim is exceedingly uneven, and abounds in rough pinnacles. Comparing these features with those of Pavant and the Ice S[)ring craters, we infer with confidence that water was present in the crater during the greater part of the formation of the outer rim and was absent dm'ing the formation of the inner rim. When compact hand specimens of the Tabernacle and Ice Spring lavas are compared, little difference is seen, but their streams differ widely in habit. The Tabernacle field, though by no means smooth, is far less rugged than the Ice Spring. Some of the surface is broken into blocks, which are so far displaced that they are not easily traversed on horseback ; but the greater part is comparatively even, and exhibits the ropy structure charac- teristic of pahoehoe. A conspicuous character of the streams was the con- gelation of their upper portions and the subsequent escape of the liquid matter beneath. This is shown in a few places by the preservation of tubu- lar caves, and more frequently by depressed areas, where the lava crust has manifestly settled down as its support was withdrawn. The constituent streams of the field are partially separable, and the latest may be traced to the inner rim of the crater. At its outer margin the lava field terminates in most directions in a cliff — not such a cliff as results from the undercutting' of a lava bed resting on softer material, but a cliff of original formation contemporaneous with the upper surface. At a point on the eastern side it was measured and found to have a height of 65 feet. On the face of this cliff, near the top, is a ])and of calcareous tufii adhering to the Ijasalt, and above it there was detected at some points a terrace of wave erosion. These are features of the Provo shore-line. The crater rims Ijear no trace of wave work, and this negative evidence is reinforced by the absence of all lacustrine deposits from the crater, from the general surface of the field, and from the sunken areas and caves. The inner rim and tlie field were never sul>merged; the outer may possibly have been covered at the epoch of the Bonneville shore, but not at that of the Intermediate shores. THE TABEENACLE. 331 Lying just above the Provo level, and yet showing no trace of sub- mergence, the lava field must have been formed after the tall of the water from the Bonneville level to the Provo. Bearing the Provo shore mark, it must have been spread before the close of the Provo epoch. It therefore originated during the Provo epoch. The inner rim of the crater has the same date. The outer rim is older than the inner and younger than the Intermediate shores; it belongs to the Bonneville shore epoch or to the earlier part of the Provo epoch. The presence in it of some slaggy matter suggests irregularity in the supply of water and indicates the later date. The most probable history is as follows: When the Pleistocene lake fell to the Provo level, it had a depth of from fifty to seventy-five feet over the present site of these craters and lava fields, and there it remained for many centm'ies. An eruption occurred beneath its surface. At fii'st, or at least during an early stage, the eruption was explosive, its violence, possibly stimulated by the water, being so great that the circle of maximum deposit was more than a thousand feet from the vent. Eventually the growing rampart shut out the water, the explosions becaine less violent, and the ejecta became pasty. Quiet eruption followed, developing a low, black island, which received a wave record before the final desiccation. The closing phase of eruption was explosive. The geologic date of this lava field is so well determined that special interest attaches to the degree of freshness of its surface. Decay has pro- gressed far enough to obliterate the finer convolutions and somewhat obscure the coarser — two to six inches across. Probably salient parts have yielded an inch to atmospheric waste. The minor depressions contain an inch or two of soil, and small cracks are filled. Large cracks remain open. Judged by its color, the soil is less the product of local disintegration than of eolian deposition. The principal vegetation is the common sage of the country. In the caves the eolian deposit, reinforced by the di'oppings of bats and probably other animals, has a depth of one or two feet. The ground just north of the Tabernacle field is traversed by a fault, with a throw of fifteen or twenty feet to the west. It divides the lava, also, and was traced with diminishing throw half way to the crater. In the opposite direction it disappears at the edge of the Ice Spring field, being overplaced by that eru2:)tion. 332 LAKE BONNEVILLE. At the side of the fauh is a k)w hill of scoriae, against and around which the Tabernacle lava flowed. It is a vestige, ill preserved, of some long anterior bnt dateless eruption. Another vestige, equally vague as to time, appears in an inclined fragment of a basalt sheet, brought up l)y a fault at the south margin of the Tabernacle field. This fault is overplaced by the Tabernacle lava. PLEISTOCENE WINDS. The circular wall of a crater often grows more rapidly on one side than another. This must sometimes be occasioned by the obliquity of the flue, but observers have generally refen-ed it to the deflection of flying fragments by the wind. If a group of extinct craters are oriented in the same way, it seems legitimate to infer the prevailing dhection of the wind at the time of their formation. In the Fillmore district there is practical harmony of orientation. The Crescent, the Miter, and the smaller crater between them have their highest walls at the east. That of the Terrace crater is at the northeast. The outer rim of the Tabernacle culminates on the east side, the inner rim on the north. The apex of Pavaiit Butte stands north of the crater. The entire range of the seven is from north to east, and the indication is that winds from the south, southwest, and west prevailed. There are no meteorologic stations competent to tell us whence the winds blow at the present time, but the prevailing air movement is recorded by nature in a satisfactory manner. In the vicinity of George's Ranch, at the south end of the eastern lobe of the Sevier Desert, the Provo shore-line consists of a series of massive bay bars, composed largely of sand. These are the source of a broad train of dunes which traverse the desert, and which demonstrate by their northeasterly course the prevalence of southwesterly winds. The phenomena consist with the theory that the gen- eral air cun-ents of this region during the Pleistocene were similar in du-ec- tion to those of the present time. FUMAROLE BUTTE AND LAVA FIELD. The most important locality remaining to be described is at the north- ern edge of the Sevier Desert, close to the head of the Old River Bed. A basaltic mesa five miles across in either direction is half divided by a valley FUMAROLE BDTTE. 333 opening to the northeast. (See PI. XXXI, near bottom.) At its head this valley is a mile wide, and is floored by red scoriae. In it stands a rough tower about 160 feet high with a truncated and obscurely crateriform sum- mit. The predominant colors of the tower are red and gray, and its material ranges from firm scoriae to compact basalt. These are roughly bedded, and exhibit a centi-ipetal dip at a high angle. The inten-elations of these featm-es are easily understood, at least in a general Avay. The tower, Fu- marole Butte, marks the position of the volcanic vent. About this vent scoriae were piled (as restored in the diagram) in an annular mole, and from it escaped the lava of the surrounding mesa. The last phase of erujition was non-explosive, and compact rock was fonned in the flue. Subsequent erosion carried away much of the scoriaceous rim, but left the resistant core and the equally resistant lava field. c .: e^— Fig. 40.— Theoretic section of Fumarole Butte. The Cinder Cone is restored by dotted lines. Before visiting this butte I had listened with incredulous interest to the statement that smoke or steam Avas sometimes seen to rise from it, but personal observation subsequently removed all doubt. About the outer edge of the summit are thirty or forty crevices from which wann, moist air gently flows. The permanence of the phenomenon is attested by the ver- dure lining the openings — a deep green moss glistening with moisture and vividly contrasting alike with the somber rocks and the sparse, ashen vege- tation without. In diff"erent openings I found the temperatures 62°, 70°, 72°, and 73.5° Fahr., all above the atmospheric mean for the locality, which is approximately 55°. At the time of observation the outer air had a tem- peratui-e of 30°, and was dry. A little mist formed over some of the open- ings, but was reevaporated within a few feet. On days that are moist, cool and still, a conspicuous cloud must arise. It can hardly be doubted that this thermal manifestation testifies to a residuum of volcanic heat in the old flue. A group of hot springs at the southeastern base of the mesa may have the same significance. Their temperatures range from 110° to 178° Fahr. 334 LAKE BONNEVILLE. Just north of the mesa is a basaltic hill whose apex overlooks the mesa and has about the height of the butte. This hill is terraced l)y wave action, exhibiting especially the Bonneville and Provo shores. Tlie Bonneville terrace appears also about thirty feet above the base of the butte, and a single point of the mesa was high enough to receive it. . The I'elation of these shore benches to the valley about the butte shows clearly that the excavation of the valley was antecedent and was subaerial. The littoral excavation was trivial in comparison. The wet-weather di-ainage of the mesa crosses its liounding cliff at numerous points, and at each of these a narrow, notch-like valley has been eroded from the basalt. These notches were cut before the Bonneville epoch, and during that epoch were partly filled by lake deposits. Subse- quent erosion has not wholly removed these deposits, and the remnants show that both Yellow Clay and White Marl were present. These facts demonstrate that not only the volcanic eruption but the principal erosion of the volcanic formations took place in Tertiary time. The surface of the mesa has lost all details of its original configuration. One can not say whether the flowing lava assumed the rough or the smooth type. It is far from smooth, but its unevenness apparently depends on ine- quality of disintegration and erosion. The rock is superficially red from decomposition, and is generally bare of soil, the slopes of surface sufficing for the rapid removal of disintegrated material. The margins of the table on the east and south (where alone they were examined) are cliffs by sap- ping— that is to say, blocks of rock have fallen away in consequence of the yielding of a softer substratum. Probably the lava was spread on the plain before the first establishment of drainage on the line of the Old River Bed. The carving of that channel lowered the base level of erosion for the i-egion and induced the general degradation of the plain, so that the field of obdu- rate basalt became a hill of circumdenudatiou. The greater share of this process also must be referred to the Tertiary. The most impressive phenomenon of the locality is the secular persist- ence of the volcanic heat. At the time of eruption the rocks adjacent to the conduit or conduits became heated, and the lava remaining in dikes and chimneys added to the store of heat. Since that time conduction has steadily U S.OEOLOOrOAL SUFA'EY layj: bci'Inkvjll:-: fl xli Juliu» Itirn ft Co.Iiih Dt-nMn bj- C ThoippBon ANCIENT CRATER STILL WARM. 335 carried this lieat in all directions, and the convection of snbterranean water has helped to discharge it to tlie atmosphere, and yet enough remains to sustain a fumarole ten centigrade degrees warmer than the air. The period of heat dissipation includes the whole of the Pleistocene period and an antecedent period of erosion probably of equal length. OTHER LOCALITIES OF BASALT. The remaining basaltic masses of the lake area, so far as they were inspected, do not declare their age by visible phenomena of superposition, ])ut tlie majority can be referred with probability to the Tertiary from a comparison of their condition of preservation with that of the Tabernacle field oil the one hand and the Fumarole on the other. This statement applies to all localities mapped in PI. XLI north of the fortietli parallel excepting that on Bear River. It applies also to two localities at the west edge of the Sevier body of the lake, to two near Preuss Bay, to two which trench on Escalante Bay, to the buttes near Corn Creek (southwest of Fill- moi'e) and a large table west of them, and to a table lying west of Pavant Butte and south of the town of Deseret. Fig. 41— Duuckrburg Butte. 3:!6 LAKE BONNEVILLE. Between this last-named table and the north end of the Beaver Creek range stands Dunderberg Butte, the remnant of what may have been a large cone of scorioe. Its lapilli are coherent, Init have not the yellow color of the tuff cones. Their mass is traversed by dikes and .sheets of vesicular liasalt. Some of the basalt vesicles contain calcite and zeolitic minerals. The top is flat, except where dikes project, having been trun- cated by the waves at the Provo epoch. The date of eruption can be judged only from the progress of demolition. It was probably Tertiary, but may have been inter-Bonneville. Equally in doubt are a basaltic table north of Pavant Butte and another south of it and extending nearly to the Ice Spring field. PLEISTOCENE ERUPTIONS ELSEWHERE. The same criteria of discrimination may be a})plied with equal pro- priety outside the lake area, so far as the conditions of rock decay are sim- ilar. Carefully applied, they would serve to classify the greater number of basaltic eruptions of the Arid Region as severally Tertiary or Pleistocene. While engaged in general geologic exploration, I have seen in Utah, Idaho, Nevada, California, Aiizona and New Mexico about two hundred fields of lava, judged by their color and habit to be basaltic, and as many as tlu-ee hundred and fifty cones of basaltic scoria?. My attention was usually not called to their state of preservation, but the data contained in note books and memory nevertheless afford a basis for judgment, and I have attempted a classification, with the following result: Of the streams and fields, 15 per cent, are judged to be Pleistocene; of the cones, 60 per cent.; the remainder are regarded as Tertiary. Of the eruptions thus classed as Pleistocene a certain number admit of no question, and these are enumer- ated in the following paragraph. On the Markaguut Plateau in southern Utah, close to its western edge, are three or more lava fields of the rougher type, all fresher in apjDearance than the Tabernacle field, and ^vith them are ten or twelve cinder cones, red and black. It is said that Panguitch Lake, a few miles towai-d tlie northeast, owes its existence to the danaming of its valley by a lava stream nearly as fresh. On the face of the cliff which bounds the Pownsagunt PLEISTOCENE ERUPTIONS. 337 Plateau on the south, a cinder cone marks the position of a vent from which a black stream has flowed down the slope toward the valley of Kanab Creek. This stream has weathered somewhat more than has the Tabernacle lava, but recency is indicated by the small amount of subsequent erosion in a country whose whole configuration indicates rapid degradation. In the heart of the Uinkaret Mountains of northern Arizona, surrounded by scores of basaltic streams and craters, the majority of which are probably Ter- tiary, there is one field of intense blackness rivaling the Ice Spring field in freshness. South of the Grand Canyon of the Colorado there is a similar forest of cratered cones about the base of San Francisco Mountain, and as one surveys them from that peak, his eye is arrested by a lava field at the east on which vegetation has not yet encroached, and by several craters near it of exceptional jJerfection. On the source of the San Jose in New Mexico a stream of lava preserves the wrinkles of viscous flow, and its siu'face has scarcely yielded to the corrasion of a brooklet that crosses it. At the southwestern base of the Zuili Plateau, near El Moro, is " a long-, broad lava stream, comparable in age with the Tabernacle field. In, south- eastern California, on the grand alluvial cones of the eastern front of the High Sierra there are a dozen bright red and black cinder cones marking vents whence basalt has descended toward Owen's River. Farther north, in the same structural meridian, a small basaltic mass overlies one of the glacial moraines of Mono Valley. RHYOLITE. ' Besides basalt, the only important volcanic rock of the Bonneville area is rhyolite. It stands next also in point of recency, Ijut is far older than Lake Bonneville. So far as observation extended, its most recent example is a body lying just east of Coyote Spring, at the south end of the Sevier Desert. This had an original depth of three hundred feet or more, and an extent in each direction of several miles; but it has been so dis- sected by erosion along its lines of drainage that its original configuration is suggested rather than shown. Its system of valleys has a general depth of at least two hundred feet, and these are so related to the Bonneville shore-line as to show their earlier formation. MON I 22 338 LAKE BONNEVILLE. Just east of the Tabernacle lava field is a liill of <,n-ey rhyolite one or two hundred feet high. It is a worn remnant, with nothing in its a.spect to aid conjecture as to its original extent. Its Ijase is concealed by the lake beds, and its sides show terracing by the waves of Provo and Intermediate times. Lying to the leeward of a gy^isum playa, it has acqviired a white mantle of gypseous sand dunes, whence it is called "White Mountain" (see page 223 and PI. XXXY). A portion of the Dug way range, on the south margin of the Great Salt Lake Desert, is of rhyolite and rhyolitic tuff. It is of such antiquity that the original shapes due to eruption have been replaced by those of atmospheric sculpture. From its gorges, as from other mountain gorges, there are spread great fans of alluvium, and across these completed fans are traced the shore-lines of Bonneville. SUMMARY AND CONCLUSIONS. The extravasation of rhyolite in the immediate vicinity of Lake Bonne- ville was long anterior to the epoch of the lake. The same may be said of the' earlier extravasations of basalt, but the period of basaltic erujjtion includes the period of lake extension. In the Fillmore district basalt was extruded at various times during the epoch of the White Marl (later Pleis- tocene), and from one vent there were eruptions after the final desiccation (post-glacial). The states of preservation of lava beds of various determined epochs afford a rude scale for the chronologic classification of lava beds not other- wise correlated, and warrant the conclusion that in Utah, Nevada, New Mexico, Arizona, and California the majority of basalt flows are Tertiary; a small minority are Pleistocene, and of these a few are post-glacial. The post-glacial eruptions are found in each of the indicated States and Terri- tories except Nevada, and belong to eight distinct volcanic districts. Although human history fails to give satisfactory record of the occur- rence of any of these eruptions, their antiquity, as measiu'ed in years, can not be great, and an application of the general law of ])robabilities leads us to look forward to a resumption of volcanic activity. The subteiranean reaction of which basaltic extravasation is the consequence has continued VOLCANIC EPOCH NOT CLOSED. 339 in the broad region not only tlirough the Pleistocene but through a much longer period of preceding time. The intermittence of eruption doe-s not argue discontinuity of the subterranean process, for, whatever that process may be, it involves the production of an unstable equilibrium that is con- verted to stable equilibi-ium only by eruption, and such conversion is always rhythmic. The abrupt cessation of a process so widely spread and so long sustained is highly improbable, and its gradual cessation would naturally include not only growing infrequency of eruption l^ut the suc- cessive extinction of eruption districts. The number of post-glacial erup- tions and the number of districts among which these were distributed alike assure us that the end is not yet. Their distribution in time and space indicates that the volcanoes and the lakes have been genetically independent. The Fumarole volcano broke out during an epoch of aridity, long before the first expansion of the lake; the Pavant and the Tabernacle were built on sublacustrine foundations; the Ice Spring volcanoes continued the series after the water had subsided. Outside the basin there was a parallel volcanic history, and though the volcanic districts are irregularly disposed, one can not say that they are either more or less abundant in the vicinity of the site of the lake. CHAPTER VIII. LAKE BONNEVILLE AND DIASTROPIIISM. The displacements of the earth's crust which produce mountain ridges are called orogenic. For the broader displacements causing continents and plateaus, ocean beds and continental basins, our language affords no term of equal convenience. Having occasion to contrast the phenomena of the narrower geographic waves with those of the broader swells, I shall take the liberty to apply to the broader movements the adjective epeirogenic, founding the term on the Greek word tJTreipo?, a continent. The process of mountain formation is orogeny, the process of continent formation is epeirogeng, and the two collectively are diastrophism.^ It may be that orogenic and epeirogenic forces and processes are one, but so long at least as both are unknown it is convenient to consider them separately. The mountain ranges so thickly set in the Bonneville district, and generally in the Great Basin, are orogenic phenomena; the concavity of the Bonneville Basin, whereby it is constituted an area of interior drainage, is epeirogenic. Neither process of displacement belongs exclusively to the remote past, but both are associated with the lake history. The e\adence of this association is of three kinds, consisting (1) of the phenomena of faults, (2) of departure of shore-lines from horizontality, and (3) of the anomalous ])osition of Great Salt Lake. EVIDENCE FROM FAULTING ; FAULT SCARPS. In the district of the Great Basin the characteristic structure of mount- ain ranges is one in which faults play an important part. Foldings of strata are not wanting, but the greater features of relief appear to have ' Seo iioto on page 3. 340 OROGENY AND EPEIROGENY. 341 been wrought by the displacement of orographic blocks along lines of" fault. Sometimes a mountain range consists of a great block of strata cut off along one side by a profound fault, and inclined in the opposite direction until it descends beneath the plain constituted by the alluvial deposits of the adja- cent valley. More frequently there are other faults within the range, trend- ing parallel to its length, and having throws on the same side with the throw of the greater fault at the base. It was probably these internal faults which originally suggested the structure of the ranges as faulted orographic blocks ; but the structure was soon connected with a certain set of topographic features, and came to be recognized by means of these. A range consisting of a faulted block gen- erally has a bold front on the side of the fault, and is less abrupt on the opposite slope. On the side of the bold front the line separating the rock of the mountain from the alluvium of the valley is simple and direct, while on the opposite side it is tortuous. On the side of the fault the strata usually dip away from the adjacent valley; on the opposite side, toward it. It was not until after the structure had been discovered and described by several geologists that the more decisive evidence afforded by the fault scarp was brought to bear. The writer first became aware in the summer of 1876 that lines of faulting may sometimes be traced upon the ground by means of low cliffs or scarps due to displacement of so recent date that the atmospheric processes of sculpture liave not yet restored the ordinary forms of topographic detail. Since that time he has observed many such scarps in various parts of the Bonneville Basin, and in other portions of the Great Basin, and the observation has been still further extended by others, especially by Russell.^ The observed fault scarps for the most part follow the outcrops of fault planes whose position had previously been inferred from the configuration of the adjacent mountains, but they have served also to betray a number of faults whose existence might otherwise not be suspected. An illustration of this is found on the west side of the A([ui range of mountains, where the strata constituting the range dip down apparently beneath the allu\ium ' Fourth Ann. Kept. U. S. Geol. Survey, pp. 445, 448," 449, 452. Geological history of Lake Lahon- tan, Chap. X. 342 LAKE BONNEVILLE. of Skull viillcv. The typical aspcft of the faulted mountain front is here wanting, and the actual fault, demonstrated by a superficial scarp, naturally escaped the attention of the geologists who have described and ligured the structure of the range. A case of more frecpient occurrence is that in which the fault along the base of the range is compound, one portion following the visible edge of the rock, and another portion lying some furlongs or even some miles val- ley-ward. The orogenic block between the two fault planes lies far lower than the one constituting the mountain range, and may be far higher than the one beneath the valley. Occasionally some portion of it is visible, but it is usually completely buried by the alluvium constituting the foot slope of the mountain, so that the surface affords no intimation of its existence, unless some recent faulting records the position of its margin l)y a scarp. It was at the base of the Wasatch Range that the fault scarj) was iirst discriminated as a distinct toi)ographic feature, and up to the jjresent time that range has afforded the best illustrations. A descrij)tion of the phe- nomena there exhibited will now be given somewhat in detail, following the order from south to north. It should be premised that the fault scarps were at no time a leading subject of investigation; the region was traversed upon other errands, and the faults were obser\ed incidentally. The record therefore, although involving much detail, is far from full or exhaustive. The Wasatch Range, using the term in the most restricted sense, may be said to extend from the town of Nephi, near which it culminates in Mount Nebo, northward to the Gate of the ]leav River, where its axis is ver\' low. The geireral course is a little west of north, and there are two angles just north of ]\rount Nebo, which have the effect of offsetting the axis some miles to the eastward. Near the town of Santa(pnn there is a low spur projecting westward and continued across the valley in a line of hills. Forty miles farther north a higher spur, kno\A-n as the Tra\erse Range, runs westward. A third spur lies just north of Salt Lake City, and a fourth a few miles north of Ogden, near the town of lionneville. These orographic features and tlic i)()sitions of the localities described in the following para- grapli can be lu'st made out b\- the aid of the large map ol Lake Bonne- ville. FAULT SCARPS. 343 From Nephi to the pass near Santaquin the range is lofty, and has a rather high alluvial foot slope toward Juab valley. At a variable distance from the mountain base this foot slope is traversed by a fault scarp from ten to thirty feet in height. It is for the most part single, l)ut in places it is divided into two parts, and it was observed at several points to fade out, being coincidently replaced by a similar scarp a few rods up or down the slope, and lapping past it. Toward the north it swings nearer to the mount- ain base, and it was finally seen to leave the valley altogether and strike across the neck of the Santaquin spur. Juab Valley lies at such an alti- tude that the water of Lake Bonneville covered only its lowest part, and the shore-lines lie far lower on the slope than the fault scarp. There is thus no direct relation establishing the order of sequence of the lake and the displacements, but the relative recency of the last displacement is inferred from the state of preservation of the scarp. Evidence of faulting was next seen in the ancient deltas on the Spanish Fork, deltas lying in the reentrant angle produced by the inflection of the mountain axis north of Mount Nebo. There were distinguished two deltas, synchronous with the Bonne^^lle and Provo shore-lines, the Bonneville delta being widely trenched by erosion and containing the head of the Pi'ovo between its surviving segments. The fault scarps are numerous, producing a confused topography, and their zone is at least a mile broad. The majority traverse the upper delta only, and the abrupt manner in which certain scarps terminate at the edge of this demonstrates that they were produced after the formation of the upjier delta and before the completion of the lower. The greatest throw of a single fault observed on the upper delta is more than 150 feet; the greatest throw on the lower delta is about 40 feet. The throw of all the faults is toward the went, but the strips of delta plain lying between the parallel faults are inclined toward the east. The net displacement was evidently such as to increase the height of the mountain with reference to the \-alley, but its amount was not ascertained. Thence to Hobble Creek, five miles, the zone of displacement follows the margin of the alluvial slope where it adjoins the mountain face, and usually includes from two to half a dozen fault scarj^s. These in the main trend parallel to the base of the mountain range, but a few scarps depart from it at high angles. 344 LAKE BONNEVILLE. At Hobble Creek the fault scarps are numerous, and tlicv are well exhibited on tlie surface of the Bonneville delta. Their total tln-ow was estimated, \\itli the aid of an aneroid barometer, to be 125 feet. Their states of preservation indicate that they are of varifuis dates, and the latest formed is so fresh that vegetation has not yet entirely covered its slope. A little farther north a fault is seen to traverse a beach line of the Inter- mediate series, giving the contiguous portions of the l)eacli a difference in altitude of about thirty feet. Near the city of Provo, a small mountain torrent issues from a gorge called Rock Canyon. At the mouth of the canyon is a delta terrace at the Bonneville level, with a radius of about 1,700 feet, and divided midway by the stream. The stream has opened a passage several hundi'ed feet broad, and is flanked on one side by a stream teirace. The greater portion of the delta terrace on both sides of the stream is corrugated by faulting, being ridged to such an extent that elevated aqueducts have been resorted to in conducting water over it for purposes of irrigation. Figure 42 exhibits two Fig. 42.— Profiles (1,000 feet apart) of the Kock Canton Delta, illustrating its displacement by F.aulting measured profiles traversing the southern half of the delta at right angles to the strike of the fiiult scarps. If the reader will bear in mind that these deltas are normally characterized by simple profiles, sloping with great uni- formity from apex to margin, he may obtain from the diagrams some idea of the nature of the irregidarities introduced by fixulting Tlie rods; of the mountain is indicated at the right, and tlic cliff, a, at the extreme left is that belonging to the margin of the delta The |)ositions of faults are shown liy vertical broken lines, and the letters h c d c mark fault scarps wliich traverse both lines of section. The lines of section are alxiut 1.0(10 feet apart, and FAULTS AT ROCK CANYON. 345 their differences fairly represent the ordinary variahiHty observed in the details of fault belts when followed in the direction of their strike. A little farther north thnn the position of the upper profile the faults h and c approach each other, and the fallen block between them wedges out. Where they join, the trough gives })lace to a ridge about five feet high, and this ridge, after running a short distance on the plain of the delta, reaches the edo-e overlooking- the stream and follows down the stream cliff to the flood plain. A portion of the fault scarp d likewise descends the stream cliff, but all the other scarps of the terrace end at its northern margin. It thus appears that the greater part of the displacement took place before the creek performed its last work of lateral corrasion on the south side of its channel, but that two of the movements are of later date The ])henomena are of special interest because they exhibit the hades of faults, features very diffi- cult of observation where the faulted material is alluvium. The hade is nearly vertical, but inclines slightly toward the valley. These features are shown in Figure 43, in which the stream cliff is represented as seen from Fig. 43. — South half of Rock Canyon Delta, showing Fault Scarps. the north, the artist standing on the northern half of the divided delta and looking across the valley of the stream. The creek itself is hidden liy a stream teiTace which occupies the foreground of the sketch, and it will be observed that this terrace is likewise traversed by two small faiilt scarps, facing each other. Their height is only from two to four feet, and by 346 LAKE BONNEVILLE. contrast with tlie greater scaqis on tlie dc^lta teiTace beyond, tliey serve to show how small a portion of the entire disturbance has occurred since the principal excavation of the stream channel. The next observation was made at the American Fork, ^hicli debouches from the moiintain twelve miles farther north. There, too, a delta of the Bonneville shore-line is centrally divided by stream erosion. Both halves of the delta are traversed close to the mountain base by a fault scarp GO or 70 feet high. The same displacement traverses the flood })lain of the stream, but its throw there is only 15 feet, showing that the entire displacement of the delta was not accomplished in a single movement. The last disturbance of the flood plain was so recent that a rapid still marks the acclivity it pro- duced in the bowlder-paved stream channel. A few miles northward the scarp was seen to traverse the Pleistocene alluvial plain at the mouth of Dry Canyon, and also the moraine with which that plain is associated. This locality is close to the jwint where the Tra- verse Range joins the Wasatch, but the fault was not traced far enough to ascertain its relation to the junction. There can be no question, however, that the great fault passes between the two ranges, and it is jirobable that a recent movement has characterized it here as elsewhere. On tlie north side of the Traverse Range the fault scarp at the base of the Wasatch was traced quite to the junction and seen to rise in the groin between the two masses. In the next ten miles northward, there issue from the Wasatch three creeks, known as Dry Cottonwood, Little Cottonwood, and Big Cotton- wood,' and the fault was continuously traced by its scarps past all these. In the vicinity of the streams and in the intervals between them the surface disturbances are complicated, and for a distance of about 5 miles there run op])osing scarps, between which a block has been depressed. At the mouths of Dry Cottonwood and Little Cottonwood canyons the scarps cross a sys- tem of moraines, described in Chapter VI and represented in PI. XLII, and materially modify their forms. The lateral and terminal moraines of Drv Cottonwood Canyon originally constituted a loop, the extremitA' of \\ liicli was notched by the creek. The depressed block, traversing the latcial moraines, has carried doAvn segments of them, leaving the distal portions as ' In PI. XLII tlio name " Big Cottonwood " is erroneously attached to Dry Cottonwood Creek. U S. GEOLOGICAL SUF.VEY IlAKE BONNEVILLE, PL.XUI MAP OF THE M () r T II s 111-- LITTLK .\N7) DRY COTTONWOOD CANONS, At Ihc Wi-slcni IJasc ul' llir Wasalc li Movuilains, I'Lili , VV shewing GLACIAL MOIJAINKS ANI1 FALL T S Topography I'v lUIberr Tlionipsoii Cc'oIo^Sv bv OK Gilbert Ltitcrnl tiitil Ttrnnn*tl v ICjOO SCALE 't^ ^ ^ r 2000 3000 --^— : FEET S{)-/fe^ Ion. U'li r .IuUu« Ilicn ft I'o.Iith DriiwTi by li TliouipS' w.' FAULTS AT THE COTTONWOODS. 347 Fin. 44.— Profile of the South Moraine at the mouth of Little Cot- tonwood Canyon, ahowing the effect of Faulting. a pair of outlying hills. 'I'lic .southci-u lateral moraiiiu of Ivittle Cottonwood, an acute and originally sj-minetric ridge, has assumed the profile repre- sented in Fig. 4*4. Tlie iiortlieni lateral, being broad and Hat, exhibits a conspicuous trench where crossed by the depressed block (see PI. XLIII). The walls of this trench are among the freshest of the fault scarps, being bare of vegetation along their upper courses, and in places too steep to be climljed. (_)n the side nearest the mountain their height is from 40 to 60 feet. Here again it is evident tliat the total displacement was accomjjlished by a series of efforts, for lietween the two moraines the phenomena of the depressed block appear in the alluvial jilain of Little Cottonwood Creek, and the greatest scarp in the plain has a height of only 20 feet. At Big Cottonwood Creek the total displacement is about 40 feet, and at a ])oint between the two streams a single scarp was observed with a throw of 100 feet. Fig. 45, gi"v"ing a profile of fault scarps near Big Cottonwood Creek, is not based on measurement, but reprodtices a rough field sketch. It is probable that faults traverse the ancient deltas of Little C'otton- wood Creek at a distance of some miles from the moinitain base, but this fact was not fully established. From a point about one mile noi-tli of Big Cottonwood Creek to Salt Lake Cit}^, a distance often miles, the fault records are obscure, and it is prob- able that there have been no very recent movements. No scarps at all were seen close to the rock of the mountain. It was thought, that an old one could be traced a short distance along the middle of the alluvial slope below Fort Douglas, and there is a more decided indication at the foot of the same slope in the eastern suburbs of Salt Lake City. Both of these are ancient as compared with the scarps previously described, and they may even have been washed by the later waters of Lake Bonneville. Fig. 45.— Profile of Fault Scarp.s near Big Cottonwood Canyon. 348 LAKE BONNEVILLE. Salt Lake City is built just soutli of a spur which jji-qjects four or five miles westward from the front of the Wasatch. Tliis spur rcpi-esents an orogenic block distinct from that of the main range. It is separated from the mountain mass by a fault plane along which the Wasatch block has, relatively speaking, risen, and it is separated from the valley on the remain- ing three sides by a curved fault plane along which the block underlying the valley has, relatively speaking, fallen. The first of these faults has been determined from the rock .structure, as I am informed by Mr. J. E. Clayton of Salt Lake City. It is also indicated at its northern end by a fault scarp, which can be traced for a short distance up the groin. The fault on the side of the valley is exhibited at the west and northwest by a series of scarps, which begin in the northern suburbs of Salt Lake City near the Warm Springs. At this point the flat alluvial plain of the Jordan reaches the steep rock face of the spur, the line of separation being marked by an abrupt change of slope. A little north of the springs there can be seen clinging to the rock at a height of 40 feet a line of conglomerate fragments, formed within the plain by the cementation of debris to the limestone, and brought by faulting into the present position. The surface of the plain below is thrown by the same faulting into irregular waves, and at one point it is distinctly terraced. On one of the faulted benches an ore-reducing establishment has been built, utilizing a lower bench as a dmTiping ground for its slag. Between this point and the liot spring an alluvial cone, Ijuilt against the face of the spur, is traversed by a typical scarp, which was sketched by Mr. Holmes. Tlie sketch is reproduced in PI. XLIV, where may be seen not only the scarp but its relation to other elements of the local geologic history. The face of the spur consists of a paleozoic limestone, inclined at various high angles. The horizontal terraces it liears are shore marks of the ancient lake. It is evident that the principal features of its relief had been carved before the production of these terraces, so that the main displacement — that to which the spur owes its origin — nuist have occurred long before the Bonneville epoch. The alhn ial cone may or may not have been constructed before the epoch of the lake, but by the absence of shore-lines and lake beds from its surface we are assured tliat its outer layers at least are of post-Bonueville deposition. The disi)lacements pro- U. S. GEOLOGICAL SURVEY LA^:e eOflNEVILLE PL, XLIV V .> '^" ^0Mi^l^ '%■ '"^•-\\- R^ -t0B-'N}'£!KH.l- S'c'^ts-s FAULT SCARP CROSSING ALLUVIAL CONE, NEAR SALT LAKE CITY. Drawn by W. H. Holmes. FAULTS NEAR SALT LAKE CITY. 349 ducing the fault scarps are therefore subsequent not only to the lake but to a certain amount of post-lacustral alluviation. The portion of the alluvial cone that lies above the fault scarp is chan- neled by the stream, and a study of the system of terraces bordering this channel shows that the total displacement of 30 feet was produced by at least tln-ee independent movements, the measures of the parts being 15 feet, 5 feet, and 10 feet. At this point and elsewhere in the vicinity the scarp is utilized by burners of lime, who construct their kilns against its face and use the terraces above and below for the two approaches needed in the man- agement of the kilns. The proprietor of the kiln represented in the plate enjoys the further convenience of quarrying his limestone from the adja- cent cliff. The hot spring at the apex of the spvir is on the line of the fai;lt, and a scarp can be traced from it in either direction. The powder houses stand- ing a little farther northward are partly above and partly below the fault scarp. Many of the fault features in this vicinity, including those figured in PI. XLIV, may be seen from the car windows of trains passing between Salt Lake City and Ogdeu. From the point where the spur joins the main ridge northward to the ancient delta of the Weber, a continuous scarp follows the mountain base, its throw ranging from 25 to 75 feet. Opposite the village of Farmington its course is less direct than the trend of the mountain front, causing it to ascend and descend the narrow alluvial foot slope in the manner represented in Fig. 46. The broad Weber delta, which belongs chiefly to the Provo epoch, is crossed from side to side by the scarp, the general throw being from 40 to 50 feet. A recent alluvial cone resting upon the southern half of the delta has suffered a displacement only one-third as great as the adja- cent delta. On the northern half of the delta the scarps constitute a sys- tem similar to that in the delta of Rock Canyon, and there are transverse branches running half a mile westward into the plain. At one point the falling of a block has produced on the surface a closed basin, which with a little artificial improvement has been made to serve for the storage of water for irrigation. 350 LAKli BONNEVILLE. Thence to North Ogden Canyon scai7)s were seen at numerous points usually in (>'rou])s of two oi- more. Fig. 47 gives iiii iiunicMsunMl prnliU Fig. 40.— Sh.'u- luir,-, .uj.l I'.uili Sc.up .a iiulns, ,,i ii.r W',.,.,., (■ill F.iriiiiii^tuii, i' Fig. 47.— Profile of Fault Scarps near Ogden Canyon, Utah. across the displacement near Ogden Canyon, and contains an extreme illus- tration of the reversed slope frequently given to blocks of alluvium between parallel faults. A few miles farther north a small closed basin has been foriiu'il in this manner. In the same vicinity one of the fault scarps crosses the line of the Bonneville shore terrace, dis})lacing it about 20 feet. At North Ogden Canyon the axis of the range turns westward for a few miles, and then resumes its northerly course. At the salient angle a low- spur is apijended, similar to that at Salt Lake City, ]:)ut of smaller dimen- sions. The scar]) runs bcliiiid the spur, and none was seen alx.ut its tare; but it can not In- doubted that its boundai-y on the valley side also is deter- mined liv a fauh. .V hot spring rises near its western l»ase. 'riieiice north- FAULTS IN CACHE VALLEY. 351 ward to the town of Willard the fault scarp follows the mountain base with an average throw of 20 feet, and it gradually diminishes and disappears before reaching the next settlement, Brigham City. Beyond Brighani City a single locality only, near the settlement of Honeyville, gave evidence of recent movement on the plane of the great Wasatch fault. The total distance from Nephi to Honeyville is 125 miles, and it is probable that more than 100 miles of that distance is characterized by post- Bonneville ftiult scarps. The average displacement is 30 or 40 feet. North of Honeyville the crest line of the Wasatch falls so low that it was overflowed by the Bonneville waters. The axis rises beyond into a range of importance, Ijut the name Wasatch is not there applied. If the western margin of this range is determined by a continuation of the Wasatch fault, no record of the fai-t was observed in recent scarps. A few scarps were seen on the opposite (eastern) side of the range, especially in the vicinity of Clarkston. Twenty miles farther north, and approximately in the same structural trend, there itre fault scarps at the western margin of Marsh Val- ley, but these are outside the Bonneville Basin. The fault mentioned at Clarkston follows the western margin of Cache Valley. The eastern wall of the valley is an important mountain range, whose bold western front has the topographic configuration of a worn fault cliff. At its base there are obscure indications of late movements, either during or just after the lake epoch, and at one point, near Logan, a post- lacustrine fault scarp crosses a delta of Provo date. The displacement is about six feet. At the north end of the valley a " weathered scarp was observed near the base of the alluvial cone of Marsh Creek, close to the outlet channel of Lake Bonneville. The direction of its throw indicates that it belongs to the eastern side of the valley, but it is several miles from the mountain front proper. The range bordering Cache Valley on the east extends southward parallel to the Wasatch, and exhibits in Morgan Valley, at its intersection by the Weber River, an old fault scarp, judged from its imperfect preserva- tion to be pre-Bonneville. Passing west of the Wasatch meridian, we have at the north a single instance of recent faulting. The small range lying east of the town of Snows- 352 LAKE BONNEVILLE. ville is marked at base by a low scarp, — a scarp more defaced by erosion than are the Bonneville terraces lower down on the same slope. In tlie same meridian and far to the south are the fiiults described in the last chap- ter as associated with the Ice Spring craters. They are probably referable to the volcanic phenomena rather than to those of mountain uplift; and the same remark applies to a scarp observed by Mr. Russell 20 miles farther south. Midway between these are two fault lines, associated with the Oquirrh and Aqui ranges. These ranges are parallel to each other and to the Wa- satch, and agree with that range in having their main lines of displacement on the western side. The scarp at the western base of the Oquin-h runs southward from Lake Point a distance of four miles, exhibiting a throw of 25 feet. Its position is at the base of the pteep mountain face, and the Bonne- ville and Provo terraces are carved in the x'ock above. It was next seen a few miles farther south, where it follows the contour of an embajnuent of the mountain side. It is there partly above and partly below the level of the Bonneville shore-line. Near the town of Tooele it appears to strike across a transverse spur, reappearing southward at the mouth of what is called Dry Canyon, and continuing thence to East Canyon and the canyon which contains the mining hamlet of Lewiston. At the mouth of East Can- yon it intersects alluvial terraces in such way as to show two separate movements with an aggregate tlu'ow of 50 feet. Although the course of the scarp was not traced, it is believed that it could be followed continu- ously for a distance of 25 miles. The southern portion nms above the hor- izon of the lake shores, and is therefore not directly comparable Avith them, . but it is considered probable that post-Bonneville movements have occurred at all points of observation. The scarp on the Aqui Range is low, and there is small basis for judgment as to its date. It was best seen in the vicinity of Knowlton's ranch. Following westward along the system of ranges which separate the main body of Lake Bonneville from the Sevier body observation is purely nega- tive until the House Range is reached. It is proper to say, however, that so much attention was given to mountain foot slopes in connection witli the study of shore-lines that the absence of notable fault scarps may be asserted U S.GBOLOGICAL SURVEY LAKE BONNEVILLE PL. XDV i 113° 112° 111 111° Julius Bicn i t'o, Ulh Drmm bv G.TI...uipi« FAULTS OF WESTERN UTAB. 353 of the southern portion of the Cedar Rang-e, of the eastern face of the Simp- son and the western face of the McDowell, of Granite Rock, and of the northern portion of the Dugway Range. The Plouse Range was long ago recognized as a faulted monocline in which the direction of displacement is reversed midway. The northern third of the range exhibits a westerly dip, and is faulted along the eastern base ; the southern part has an easterly dip and is faulted on the western base.^ This determination was subsequently confirmed .by the discovery of a \vell defined fault scarp in the vicinity of Fish Si)ring, and an obscure and probably very ancient scarp at the western base of the southern division. The next mountain body to the west is the Confusion Range, an assemblage of small ridges, and associated with these a, single scarp was found. This lies near Knoll Springs, on the east side of Snake Valley. It is low and worn, and follows the rock base closely. The Deep Creek Range, which forms part of the western boundary of the Bonneville Basin, is faulted on both sides. In the vicinity of the old overland road crossing the ridge from Willow Spring to Deep Creek settle- ment, to which vicinity observation was restricted, the range is flanked on the east by a broad and high alluvial slope. No fault scarp was seen, but near the lower margin of the slope a partial section of the lake sediments shows that they were disturbed during the period of their deposition. The Yellow Clay at one place suifered uplift and erosion before the deposition of the White Marl, so that there is unconformity of dips, and at another point the Yellow Clay and White Marl together are so greatly disturbed that their inclination is toward the mountain. The superficial topograjihy that must have been created by these disturbances was obliterated by wave work, and at the locality of the section the upper edge of the inclined block was planed away in the formation of a terrace of the Provo shore. On the west side of the range an ancient and nearly obliterated scarp crosses the alluvial slope near its upper edge. On the opposite side of Deep Creek Valley a better preserved fault scarp follows the eastern base of the Gosiute range. It lies far above the Bonneville shore-line, and was not critically examined. ' Surveys West of the 100th Meridian, vol. 3, pp. 27-28. MON I 23 354 LAKE BONNEVILLE. GENERAL FEATURES OF FAULT SCARPS. Except in the yolcanic district of the Sevier Desert, the fault scarps follow the bases of mountain ranges or run })arallel to them. Where there is but a single scarp, it invariably fiices toward the valley and away from the mountain. Where there are several scarps, frequently one or more face toward the mf)untain, but the one nearest the mountain always faces toward the valley, and the net displacement is always of such nature as to inci-ease the height of the mountain with reference to the valley. The mountains are rising ar the valleys sinking. The scarps are rarely found at the contact of the rock of the mountain with the alluvium of the valley; they usually occur in the alluvium several scores or lumdreds of feet from the contact. The segments of alluvial plain included between parallel scarps rarely retain their original slope. In a few instances, and for short distances, their rate of descent toward the valley is increased by the disturbance, but as a general rule the slope valleyward is diminished, or even i-eversed. The tendency of the dissevered blocks to incline away from the side of the downthrow is almost as pronounced as in the case of land slides. The assumption that the attitudes of these alluvial surfaces are representative of the attitudes of large down-reaching masses continuous with them seems untenable, because such masses would mutually interfere. The hade of a fault is usually difficult of determination unless expo.sed by mining operations, and the difficulty is peculiarly great where the walls are of incoherent detritus. The freshest of tlie fault scarps have some talus, and prove only that the hade does not depari widely from verticality. The best observation w^as made in the Rock ('nnvon delta, where, as already described, several scarps descend a stream cliti' standing at tlu- angle of stabilitv. They show a hade toward the valley of less tlian li\ e degrees. That this approximate verticality is more than a superficial feature of the great Wasatch fiinlt, is seri(iusl\- ([uestioiied, for several reasons. In the first place the faults witliin the Hasin Ranges, so far as niv observation shows, hade at consideralih- angles, and it is liiglily probable that this fault belongs to the same system. Second, tlie secular motion of the mountain being upward with reference to the valley, it is prol)able that the roek face THEORY OF FAULT SCARPS. 355 at the contact Avith alluvium lias lieeu little Avasted by erosion, and is essen- tially the protruded foot-wall of the fault, and if so, the visible fault iu alluvium is not in the plane of the great fault, but is a branch with less hade. Finally, the last hypothesis affords an easy explanation of the super- ficial details of the faulting, as will appear by the following explanation. Fig. 48 is constituted of four diagrams illustrating the supposed method of faulting. In the first diagram the line ,r ij represents in section the Wasatch fault, with an assumed hade of 30°. To the right of this line is Fig. 48. — Diagram to illuatrate Theory of Grouped Fault Scarps in Alluvium, the firm rock of the mountain, its surface being somewhat reduced by ei'O- sion above the point a, where the alluvial slope of the valley side adjoins it. To the left of the line x y the material represented is detrital and incoherent, being chiefly alluvial. The alluvial sui'face previous to the last faulting is represented by a c. The direction of motion in faulting is parallel to the plane x y, and the plane of motion is assumed to coincide Avith that plane up to the point c, and then curA-e to &, so that a triangular pi-ism of allu- vium, ft h e, remains attached to the rock, constituting the foot-Avall of the fault. This movement opens a fissure, h e (J. The material traversed by it being incoherent or feebly coherent, the fissure cannot remain open, but is immediately filled by the settling of one or both of the Avails. The remain- ing three diagrams indicate hypothetical methods of closing the fissure. In the second diagram it is supposed that the hanging wall yields Avithout defi- nite fracture, Imt l»y differential moA'einent distributed throughout tlu* mass, so that the triangular prism included between the points // d e is made to assume the form and position // ./' e. There then remains a fai^lt scarp, h f, giving an exaggerated measure of the actual throAv of the fault h d, and 35(j LAKE BONNEVILLE. accompanied at its base hy a reversed inclina,tion of the surface/// In the third diagram it is assumed that the hanging wall is divided by a fract- ure, /; e, and that the prism h d c settles and spreads so as to occupy the space i k e. There result two fault scarps, h k and Ji '/, facing in opposite directions and api)roximately representing l)y their difference the true throw h (J. The fourth diagram supi)oses that the triangular prism h I o in cleaved from the up})er part of the foot-wall and slides down so as to take the posi- tion m n e. This gives two fault scarps, / n and m d, whose sum would ordinarily att'ord an overestimate of the actual movement of the great fault plane. If now we consider that there have been repeated movements along the same general plane of faulting, and that these repetitive displacements have often divided the alluvium in different places, it becomes evident that these hypothetic elementary profiles can be so combined as to jjroduce all the complicated profiles actually observed. While, as just mentioned, a number of successive movements may occa- sion the same number of separate scarps, they may also coincide in locus and produce but one, and it is probable that coincidence is the rule. In general, each scarp represents a series of distinct movements. Indeed, so far as the phenomena of the Bonneville Basin instruct us, the process of faulting might be conceived as one of continuous sl(i\\- motion, and it is only through the phenomena of earth(piakes in other districts that we become acquainted with the rhythmic and paroxysmal nature of dis- placement on surfaces of fracture. The features of the fault scarps accord fully with the general theory that the growth of mountains is a gradual ])rocess, secular in duration, though (•atastrophic in detail. The freshness of some of the scarps points to an anti(puty measured in >ears rather than centuries. A large number have l)een ])roduced since the final retirement of the Bonneville waters. A few were synchronous wilh the I'rovo shore-line. One movement l)elongs to inter-lSoiniexille time. ( )t earlier dates, nothing can be said with precision. Inside the lake area, it is to be sujjposed that scarps older than the Bonneville shore-line were olilit- erated l)^- littoral sculpture and lacustrine sedimeiitatiiin. ( tutside the Bon- neville shore-line the only discovered index of anti(piity is the state of pres- ervation, a criterion affording no precision. Discrimination is further em- OLD FAULTS AND YOUNG. 357 barrassed by the recunviict' of ilisplncemeiit nloiii;- tlic same lines, so tlint tlie qualified indications of date in the preceding pages apph' as a lailc onl\- to the latest of the local movements. LOCAL DISPLACEMENTS VERSUS LOCAL LOADING AND UNLOADING. The phenomena of earthquakes indicate that the orogenic forces, what- ever they may be, slo\vl\- generate and accumulate strains in the crust, until finally the cohesion or static friction is overcome, and a sudden Nield- ing results in a fault and an earthquake. In such a district as the liomie- ville Basin, where the planes of faulting, su})erficially at least, are approxi- mately vertical, it seems probable that the determination of rupture may be hastened or retarded by anything affecting the weight of the orogenic block on either side of the plane of movement. It is coimnonly held by students of physical geology that the degradation of the uplifted block and the accu- mulation of sediment on the downthrown block constitute an unloading and a loading, which consijire with and aid the forces primarily concerned in the displacement, and it is maintained by some that when once the displace- ment along a great fault line has been initiated, the process of loading and unloading is competent to continue the depression of the lower block and the upheaval of the higher without further aid from the forces that initiated the disturbance. Now the filling of the Bonneville Basin with water added a very considerable weight to the vallej's, and therefore to the down-thrown blocks, and made no corresponding addition to the uplifted blocks repre- sented in the mountain ranges. The contemporaneous glaciers were indeed sustained by uplifted blocks, but these were restricted to a short section of the Wasatch, and in that section their weight was much less than that of the water in the adjacent valley.^ It is therefore theoretically conceivalile that during the presence of the lake the pi-ocess of faulting along the.mount- ain bases was stimulated, and that after the evaporation of the water the process was corre.spondingly retarded. That the load of water was quanti- tatively sufficient is readily shown. If the transfer of rocky matter from the mountain block to the valle)' block is the cause ordinarily operative in 'The ,irea of ice on tbo Wasatch Range inay be compared with the contemporaneous area of water in Lake Bonneville by reference to PI. XLIX. The areas of ice there represented on the Wasatch and Uintah Mountains are copied from Kin^^'s map in Volume I of the Fortieth Parallel Report. 358 LAKE BONNEVILLE. generating the stress which renews movement along the fault plane between the blocks, then the dejjth of rock necessary to be removed from one block and added to the other in order to overcome the adhesion on the fault i)lane is measured by one-half the resulting movement. For the Wasatch range this measiu-e is less than five feet. The load of water held by the valley blocks was e(|ui\alent in the vicinity of Great Salt Lake to a layer of rock of the density of the surrounding mountains and with a tliickness of 300 feet, and at the Provo stage the load of water wns equivalent to 200 feet of rock. The stress due to the water was therefore many times greaU'v than that needed to over})ower the adhesion, and the load of A\iiter was com- petent to act, provided the erogenic blocks possessed the tlieoretic susceji- til)ilitv to lonil. If tlie orogenic blocks rest on a ])lastic substratmn, or if thev are oth- erwise conditioned so as to obey the hydrostatic law and yield freely to external stresses, then the valley blocks should have been depressed several hundred feet by the adilidon (if the water, should have partially recovered from this depression during tlie abrupt lowering of the lake from the Bon- neville shore to the Provt), and should have risen still further during the final desiccation of the basin, except in regions where the orogenic forces operated Avitli sufficient rapidity to counteract the tendency. Instead of this, we find that the post-Bonneville movement of the valley blocks, wdierever it has occurred, has been one of depression, and so far as the })henoraena go we find no evidence that the depression of the valleys was more rapid during the epochs of tha Bonneville and Provo shores than it has been in more recent times. We are forced to conclude that the mountain ranges of the Bonneville Basin and the valle}'s between tlicin do not, with reference to each other, obey tlie law of flotation. It follows with eqiud cogency that the faults do not penetrate to a layer characterized by fluidity or semi-fluidity — implying by these terms the ])ower to flow under small shearing strain — but terminate in a region of rigidity — im]dving by that term the ability to withstand relatively large shearing strain. I conceive them to ternnnate at the up])er limit of the i-egiou of plasticity by pressure — implying by that phrase that at and below FAULTING INDEPENDENT OF LAKE HISTORY. 359 » a certain deptli the rocks of tlie crust, liowevor riyid, iire subject to such pressure that their j'ielding under shearing strains exceeding tlie ehistic limit is not by fracture but l)y flow. I conceive the orogenic blocks as confluent with the subjacent layer, excepting such as may wedge out by the convergence of fault ])lanes. MOUNTAIN GROWTH. The height of ;) mountain, considei'ed as a to])ographic feature, is the altitude of its crest, not above sea-level, but above the surroiuiding country. From this ])oint of view it is pertinent to inquire whether the mountains of the Bonneville basin are now growing. The C[uestion is more easily asked than answered, but its consideration may not be unprofitable even thougli the residt is indefinite. In the case of mountains whose uplift takes place along faidt i)lanes, the amount of faulting is a measure of the uplift. If the faulting is at one margin only and the f)ther margin suflers no displacement, then the general uplift above the adjacent valleys is one-half the uplift at the fault lines. The processes of degradation tend constantly to pare away the mountain top and thus reduce its height, and in the district under consideration tlie processes of valley sedimentation likewise reduce the mountain height by building up the valleys and thereby raising the plane of reference. When- ever and wherever diastrophism is the more active, the mountain grows; when degradation and sedimentation are moi'e active, the mountain becomes smaller. The post-Bonneville faulting of the Wasatch Range is restricted, so far as known, to the western base, and there amounts to about 40 feet. Tlie general u])lift of the range may therefore be taken at 20 feet. Tlie product of the simultaneous degradation of tlie niount;iin finds its way to Utah Lake and Great Salt Lake, where its coarser j)art is accumulated in the deltas of the Provo, the Jordan and the Weber, while its finer j^ortion is spread over the lake bottoms. P)ut the deltas and lake beds afford no simple measure of the mountain waste, for the same rivers receive also detritus from other land areas, and in the same lakes are gathered the silts from other streams. The deposits, moreover, are unex])lored, and if they were explored, it would be no easy matter to discriminate the post-Bonne- 360 LAKE BONNEVILLE, ville deposits from the Bonneville beds beneath them. The problem might be attacked by a consideivntion of the annual outwash of the mountain tor- rents, but if this difficult measure were made, we should still need to know the antiquity in years of the last Bonneville flood, a factor for the present entirely unknown. But though a categorical answer is unattainable, a qualified result is not necessarily so. The recent uplift of the Wasatch Range is greater than that of any other range in the basin. That of the Oquirrh may be one half as great, but no other range is at all to be compared in this respect, and many ranges show no fault scarps whatever. It may therefore be said with confidence that if any range of tlie district is actually growing at the present time, the Wasatch is growing, and this l)rings us to a theorem of Powell's which here finds illustration. Powell pointed out' that a high mountain is subject to more rapid degradation than a low one, and that the rate of degradation is a geometric function of the height. It is therefore impossible for a mountain to become tall unless it is uplifted rapidly, and when uplift ceases or becomes slow, only a sliort measure of geologic time is necessary to reduce the height. High mountains are therefore always yoinig mountains. They may be constituted of very ancient rocks, — their initial uplift may have taken place at a remote date, but the great upheaval which produced the present movmtain is geologically recent. The Wasatch, springing ])oldly from a base plain 8,000 feet below its pinnacles, is a young range, and as its recent uplifting has been more rapid than that of any of its neighbors, we may fairly assume that present uplift is in excess of pres- ent waste, and that the mountain is now growing. EARTHQUAKES. The extreme recency of tlie last orogenic movements in the most populous portion of Utah, and the high ])robal)ility of their recuiTence in the future, have a practical bearing as well as a scientific, for it is now generally imderstood that earthquakes are due to paroxysmal yieldings of the earth's crust, and it is e(|uall\- well known that tlie dangers attending earthquakes can be greatly diminislu-d b\- precautionary measures. It is ' Geology of the Eastern Portion of the Uinta Mountains, by J. W. Powell, Washington, 1876, p. 196. FAULT SCARPS AND EARTHQUAKES. 361 indeed true that the fault scarps at the base of the Wasatch Mountams have not been directly connected with earth tremors, but the association of identical phenomena has been elsewhere observed. The earthquake of 1872, one of the most violent ever felt in the United .States, originated in Owen's Valley, California, and its origin was accomjjanied by the sinking of strips of land in such way as to jjroduce fault scarps identical in their general features with those described in the preceding pages. The principal scarp follows the base of the alluvial foot slope of the Sierra Nevada, and has a maximum height of about 20 feet. Where this height is attained, there is a coinpanion fault scai-p, 10 feet high, facing in the opposite direc- tion, so that the net displacement is about 10 feet. At other points the main scarp is associated with others miming nearly ])arallel and facing in the same direction. As I saw them, eleven years after their formation, they appeared little fresher than some of the Wasatch scarps. The earthquake that shook Sonora and southern Arizona on the third of May, 1887, produced a fault scarp which Avas critically examined by Goodfellow and traced for a distance of 35 miles. It intersects the alluvium along the base of a mountain range or ranges, and has an average height of seven feet. Like the Wasatch scarp, it is often divided or furnished with branches, but ludike that of the Wasatch it is exceptionally small where it . intersects the alluvia of streams issuing from the mountains.^ The association of earthquakes with fault scar])s has likewise been determined in New Zealand, where McKay and Hector not merely refer certain scarps to earthquakes of the years 1848 and l^f)b, but recognize them as the indices of modern slip.s on old planes of dislocation, and use them in tracing out important structure features.'' - It is legitimate to infer that the belt of fertile vallevs tliat follows the western base of the great mountain range of Utah is an earthquake district, and this despite the feet that since its first settlement in 18.'')0 no impoitant tremors have been recorded. It is a matter of geologic history that the Wasatch range is gradually rising, and that this rise is not uniform in time ■George E. Goodfellow. The .Sonora Eartli(|u.ake; Scieuce. vol. 11, p. 102. ''Oil the geology of tlie eastern part of Marlborough Provincial district. By Alexander McKay. In Colonial Mns. and Geol. snrvey of New Zealand ; Keports of Geological explorations during lo85. Faults on pp. 129-133. Also James Hector, in same volume, p. xv. 362 LAKE BONNEVILLE. and place, but is accom})lishe(l by small and sudden displacements more or less localizenly their relations to one another can be discussed, and it has been found convenient to refer them all to the water surface of Great Salt Lake ; and since that surface is a fluctuating one, a particular point has been arbitrarily assumed within the range of modern fluctuation. That point is the zero of the "Lake Shore" gauge. As the relation of the altitudes to sea level will not be again refer- red to, it is proper to say here that the zero of the "Lake Shore" gauge is 95 feet lower than the track of the Pacific Railroad at Ogden, and 4,208 feet higher than mean tide. The implied altitude of Ogden, 4,303 feet, is that accepted by Gannett in his dictionary f)f altitudes.^ As the various measurements employing the water of Great Salt Lake as a datum were executed on different days and in diff'erent years, it was necessary to take account of the fluctuations of the lake surface, and this was done by means of the series of gauge observations already described (see page 233). A more important difficulty was encountered in connecting the lines of leveling with the plane of the f»ld water surface, for it was never ])ossible to decide just how the mean level of the old water surface was related to a particular feature of its shore record. At some ])laces the measurement was made to a cut-terrace, and at others to an eml)ankm('nt, niid wherever both these were found in juxtaposition and measured, it was ascertained that the embankment stood higher than any ])art of the cut-terrace. It was found, moreover, that the diflerence between these twt) features was 'A Dictionary of Altitudes in the Unitod States, compiled by Henry Oannett: Bull. T'. S. Geol. Survey No. 5. 1884. ALTITUDE MEASUREMENTS. 365 less in sheltered localities than on coasts facing the open lake, where the fetch of the waves was great. The inference of tW, phuu! of tlie ^\■at(•r surface drawn from tlie local shore record wns thus necessaril\- n inatfer of judgment, and tliis judgment was usually exercised upon llie gniund, wliere the most .satisfactory consideration could be given to the local conditions. Despite all precautions, an uncertainty of several feet attaches to each such determination, and tliis uncertainty is iiu'luded in the estimation of llie probable errors of the measurements of altitude. Most of the barometric observations and all the l^arometric computa- tions were made by Mr. A. L. Webster. He has also combined, unified, and tabulated all the determinations of altitude, and has prepared a report upon them which appears (Appendix A) at the end of this volume. For all matters of detail the critical reader is referred to his report. DEFORMATION OF THE BONNEVILLE SHORE-LINE. A summary of the measurements is contained in tables XIII, XIV, and XV, and their geographical distribution is indicated to the eye in Pis. XLVI, XLVII and XLVIII. Attention wdll first be directed to the table and plate which exhibit the measured altitudes of the highest Avater line. Table XIII. — Height of the BonveviUe Shore-line, at varioiin 2)oi>ils, above Great Salt Lulce {Zero of "Lake Shore" gauge). Locality. 1. Sautaqiiiii, south of Utah Lake 2. Leuiiuj;tou, U. S. K. U 3. Milfonl, U.S. R. R 4. Red Rock Vass, north end of Cache Valley T). Franklin, CJaehe Valley li. Logan, Caeho Valley 7. Point of the Monntain; 22 niile.s south ol' Salt Lake City 8. Ogden 9. Fort Douglas, near Salt Lake City 10. Tecouia, Nevada 11. Willard, east shore of Great Salt Lake 12. Black Rock, north end of Oquirrh Range 13. Stockton, head of Tooele Valley 14. Kelton Butte, near Onibe Station. C. P. R. R Height. Feet. 902 -(- 3 902 -t 5 904 i 10 900 ± 4 940 -I- 3 942 i 4 9.'30 ± 3 980 4-5 980 ±5 981 ±5 98.1 ± 3 1008 ± 3 1014 ±5 1019 ± 3 366 LAKE BONNEVILLE. Table XUl.—Bcight of Ihe Bonnerille Shore-line, at varioux pohilH, above Great Salt Lake (Zero of " Lake Sliore" gauge) — Coutiniied. Locality. iri. Promontory, 10 miles south of Promontory Rtatlon, C. P. R. R If). North end of Aqiii range; 12 miles northwest of Grantsville... 17. Two niile.s east of Thermos Spring, Escalante Desert H. Pavant Hnttc, Sevier Desert 19. Seven miles sontli of Mil ford 20. Four miles south of Thermos Spring, Escalante Desert 21. Seven niiU>s south of Thermos Spring, Escalante Desert 22. Fillmore, east edge of Sevier Desert 23. South Twin Peali, south end of Sevier Desert 24. Kanosh Butte, south end of Sevier Desert 25. North Twin Peak, south end of Sevier Desert 2G. Antelope Spring, Escalante Desert 27. Sulphur Spring, Escalante Desert 28. Pinto Canyon, Piscalaute Desert 29. Shoal Creek Canyon, Escalante Desert 30. Meadow Cieek Canyon, Escalante Desert Height. Feet. 10.^0 -t 3 1070 -{- 3 893 ± 2.5 902 J- 15 921 ± 20 921 ± 25 927 i 25 938-1-8 9.39 ± 20 953 -t 15 971 ±20 1008 ± 30 101.-) -J- 25 1175 ± 35 1227 ± 35 1256 ±35 It appears by inspection that the range of ahitude is about 350 feet, the determination of the amount having an uncertainty of less than 50 feet. The distribution of altitudes does not follow any simple law, l)ut yet exhil)- its certain general features. There appear to be two areas in which the water mark is especially high, the first coinciding approximately with the central meridian of Great Salt Lake, and the second occupying the Esca- lante Desei't, especially its southern portion. Along tlie eastern border of the basin, from the extreme north to the extreme south, there is a general increase of altitude from east to west. At the south this is continued west- ward to the limit of the area covered by the observations, ami is greatly accented. At the north, where the observations have the greatest range in loiiiritude, the westward rise is rcijhu'cd bcNoiid tlie I'l-oiiiontorv Ivauiic bv a westward decline. It ap])ears, moreover, that to ;ill tliese general ruU's there are local exceptions, and that where a rise in a certain direction i.-< continuously indicated by a series of localities, its rate from point to j)oint is not unifonu. DISPLACEMENTS OF BONNEVILLE SHORE-LIN^,. 367 A comparison of the measured heights of shore-hne with the system of faults in the same region indicates in general that they are not closely related, and in particular that the faults cannot be appealed to as a suffi- cient explanation of the displacements of the shoi'e-line. A good illustra- tion of this is found in the latitude of Salt Lake City, where the height of the shore-line has been measured on three adjacent parallel ranges. On the Wasatch it is DSO feet, on the (Jquirrh 1,008 feet, and on the Aqui 1,070 feet. Now each of these ranges has suffered a post-Bonneville faulting at its western margin, as represented in Fig. 49, and the throw of each fault is to the west. The eft'ect of these faults, if there were no other diastrophic V. A T A si h •^ / \ TOOELE VALLEY / ,\ JORP A tJ VA LLEY % =5 5 / / t \ o 5 S / 0: \ / or V__^ 5 v__ i K 1 h. ?lf ?| cj \ Fig. 49.— Generalized cross-profilo of mountains and valleys, illustrating post-Bonneville dia3troi)bic ( hanges. Ver- tical scale greatly exaggerated. Lower liorizontal lino = level of Great Salt Lake. Dotted line = 1,000 feet above Great Salt Lake. Y T & = Bonneville shore-line. changes, would be to lift the Wasatch higher than the Oquirrh, and both higher than the Aqui, but the shore measurements show the reverse of this. If we assume that the portion of the earth's crust included between each pair of the observed faults is rigid, so as to move as a unit without flexure, then the post-Bonneville changes determined l)v the observations on faults and shore-lines are correctly represented (except in exaggeration of vertical scale) in fig. 50, where the base line indicates the level of Great Salt Lake, Fir;. 50. — Diaj;ran» of po-st-Roiineville tii;istropliic changes. Easo line — level or (tftat Suit I,jiK<^. l*(i3 -t 3 672 ± 3 679 ± 3 ESCALANTE BAY. 371 DEFORMATION OF THE PROVO SHORE-LINE. We will now turn to the consideration of Table XIV and PI. XLVII, which record in their several ways the various determinations of the heig'ht of the Provo shore-line. The special criterion by which the identity and synchronism of the Bonneville shore-line were established throughout the greater part of the liasin cannot be applied in the case of the Provo. Where the embankments successively formed during Provo time are sepa- rated from one another so as to be independently measured, they exhibit differences of height, but these diiferences are neither uniform nor constant at the various localities where they were observed. The conclusion has already been reached (page 133) that there were changes of relative height while the wave record was being made. Nevertheless, it was always easy to recognize the Provo shore-line and discriminate it from others by reason of the exceptional magnitude of the wave work accomplished at that level. The cut terraces are broader than any other within the basin, and the embankments are larger. At most points it is impossible to determine from the features of the shore what was the local history of oscillation during the persistence of the outlet, for the later work of the waves has effectually obliterated the earlier. It is highly probable that those of its features to which measurement was extended represent the final portion of the long period during which the water stood at approximately the same height. The number of measurements is smaller than in the case of the Bonne- ville shore-line, only 10 having been secured; but these are so much more harmonious that it was found possible to draw a system of smooth contours representing intervals of 25 feet only. A comparison of Pis. XLVI and XLVII shows at a glance that these correspond in position and arrange- ment with the contours adjusted to the Boinieville data in the same area. The area of maximum elevation indicated by them lies over the western portion of Great Salt Lake. There is a descent thence to the east, and more gently to the southwest and south, while a single station indicates descent to the northwest also. 372 LAKE BONNEVILLE. Table XV. — Difference in altitude of the Bonneville and I'rovo shore-lines at various points. Locality. Prciiss valley, South series of embankments Preuss valley, Middle series of embankments , Kelton Butte, near Ombe Station, C. P. R. R Black Rock, north end of Oquirrh Range Willard, east shore of Great Salt Lake Snowsville, north edge of Great Salt Lake Desert Logan, Cache Valley Point of the Mountain, 22 miles south of Salt Lake City Franklin, Cache Valley Promontory, 10 miles south of Promontory Station, C. P. R. R Tooelo Valley between Tooele and Stockton Tooelo Valley between Tooele and Grantsville Wellsville, Cache Valley Fish Spring, south edge of Great Salt Lake Desert Fillmore, east edge of Sevier Desert North end of Aqui Range; 12 miles northwest of Grantsville Cup Butte, Old River Bed Suowplow, Old River Bed Terrace Mts., 8 miles southeast of Matlin Station, C. P. R. R Dove Creek, between Matlin and Ombe, C. P. R. R Height. Feet. 311 ± 345± 3o6-|- 360 -J- 361 i 360-1- ^r,r, ± 370 -t 371 -t 374 J- 374-1- 380 -t 382 ± 382 ± 385-1- 389 i 392 i 397 -t 411-1- 413 ± DEFORMATION DURING THE PROVO EPOCH. Table XV and PI. XLVIII show the measured differences in aUitude of the Bonneville and Provo shore-lines at various points. The localities at which these differences were measured coincide partly with localities of the two ])receding tables, but are also in part independent; for it was sometimes found possible to make the differential measurement where tlie lack of an available datum point prevented the reference of either to the level of Great Salt Lake. The range of variation is not large, and whatever order mi\\ charac- terize them is so far concealed by irregularities that it was found impossible to classify them by any system of smootli contours. Ikit, as will j)resently appear, when they are classified with reference to the contours of the IJon- neville and Provo .shore-lines, thev betray a certain amount of harmonv. It will be recalled that when the lake attain0 950 to 1000 1000 to 1050 1 Above 1050 Determinations of Bonneville shore-liue above Great Salt Lake ., (- 902 902 904 906 940 942 9.'j0 965 980 981 1008 1014 1019 1050 1070 Mean . ........ .... .. 916 969 1023 1070 Determiuations of Prove shore-line Mean .- 553 569 577 580 624 640 640 663 672 679 566 602 654 679 Determinations of difference between Bonne- ville and Prove shore-lines Mean ... - :J41 345 365 - 371 385 361 365 370 382 382 392 397 356 360 374 374 380 411 413 389 361 378 381 389 By arranging the determinations in this way and then taking means, it was hoped to eliminate so much of the in-egularity due to orogenic dis- US .GEOLOGICAL SURVEY LAKE BONNEVILLE ?L:ZLI2 US' 113° n2° xa." Julius Bien it Co. lith Druwu bv C Thompsc U. S. GEOLOGICAL SURVEY LAKE BONNEVILLE PL. L SCALE ILES THEORETIC CURVES OF POST-BONNEVILLE DEFORMATION. DESICCATION AND UPLIFT. 375 placement and to errors of measurement as to render the data fairly com- parable. The measurements of the Bonneville shore-line having been used as a basis for th-awing the contours, their means, as a matter of course, constitute a series; and it was anticipated that the Provo determinations, having given rise on PI. XLVII to very similar contours, would likewise furnish, as they do, a progressive series of means; but a similar corre- spondence could not have been confidently predicted for the observations of difference in altitude of the two shores, for these are so irregular in detail that representative contours could not be drawn. Nevertheless, their means as thus classified fall into line with remarkable regularity. It appears to be a legitimate inference that the epeirogenic deformation occurring during the Provo epoch was identical in locus and general character with that occurring subsequently, and with the total deformation of which it is a part; and this accords with the postulate, for if the withdrawal of the entire mass of water produced the quaquaversal uparcliing of the basin, then the par- tial emptying of the basin by the draining off through the outlet of a layer of water 375 feet deep should produce a similar uparcliing, differing only in amount. Opposed to the postulate, we have the general fact that the Great Basin appears to have been characterized by epeirogenic movements, varied in character, through Tertiary and Pleistocene time, and that as these move- ments successively created and destroyed lake basins, they must be sup- posed to have generally originated in a different way. It is therefore pos- sible that the coincidence in time and place of the uplift under consideration with the disappearance of Lake Bonneville is a coincidence merely. A second and very serious element of weakness in the postulate inheres in the fact that the observations are mainly confined to the eastern lialf of the basin. Only two points of observation lie west of the max- imum area, and only one measurement was made in the extreme western portion of the basin. The area well covered by points of determination is at most not more than two-thirds of the entire area to which the postulate is applied. These considerations pro and con hardly admit of explicit summation. The predilections of each geological reader will determine the relative 376 LAKE BONNEVILLE. weight he asisigns to them, and his cniisoqneiit confidence or lack of confi- dence in the conclusions which follow. There are at least three ways in wliicli the removal of the water may have given rise to the observed variation of altitudes. First, the geoid may have been locally deformed by a change in the local attraction; second, the surface of the laud may have been deformed by local expansion due to the post-Bonne ville change of climate; third, the earth itself may have been locally deformed in consequence of the removal of the weight of the water. These three hypotheses will be considered in order; and it will be found advantageous to inquire Avith reference to each how the deformation it is competent to produce compares in amount with the observed defoniintion. The maximum measure of the observed deformation is 1070 — !M)2r= l(>s feet; but as this may, and probably does, involve local orogenic displace- ments, it will be better to use for the present purpose a measure obtained by comparing a number of the highest measurements collectively with a number of the lowest. The mean of the five observations of height falling within the 1000-foot contour is 1032 feet. The mean of the four lowest determinations is 903 feet, and their difference, 121) feet, will be compared with the various amounts inferable from the three hypothetic causes. HYPOTHESIS OF GEOIDAL DEFORMATION. The siu'face of a body of standing water is level, but is not plane. Being a part of the surface of the earth, it is ai)proximately ellipsoidal. If there were no inequalities of surface, and the density of the earth were uni- form throughout, or varied only in accoi'dance with certain laws, a level surface carried completely around the globe would be a perfect ellii)soid. The actual inequalities of surface and irregularities of density produce local irregularities of attraction and corresponding irregularities of the level sur- face. To distinguish the deformed level surface from the s])heroid to which it approximates it is called the "gcoiil". Any change in tlie superficial dis- tribution of matter modifies tlie geoid, and tlie i-einoval ot the lake water from the Bimneville Basin was such a change. The effect of refilling the basin would be to increase the local attraction and locally uparch the geoidal DEFOKMATION OF PLANE OP KEFEKENCE. 377 surface; its emptying- unquestionably tended to flatten the geoidal surface. Assuming the configiu-ation of the country unclaanged, the Bonneville sur- fiice was more sharply convex than the Salt Lake surface, and the engineer's level should now find the Bonneville shore-line higher on central islands than on peripheral slopes. The theoretic change corresponds in kind with the observed; does it agree in amount? My mathematical resources not being adequate to this question, it was submitted to my colleague, Mr. R. S. Woodward, who gave it full consideration. It happened that tlie cognate problem of the deformation of the geoid by a continental ice mass was sub- mitted to him at about the same time by Dr. T. C Chamberlin, and he was thus led to a comprehensive discussion of the general subject to which the special problems belong. In the application of his formulfe to the present case no account was taken of topographic details, but the mass of water in the main body of Lake Bonneville was assumed to have the form of a circular lens two degrees (138 miles) in diameter with a maximum depth of 1000 feet. It was found that the maximum de^iression of the geoidal surface referable to the subti'action of such a mass is 2.01 feet, an amount too small to be considered in comparison with the observed deforma- tion of 129 feet. The phenomena are therefore not to be explained as changes in the plane of reference, but must be referred to changes in the relative altitude of portions of the basin. The reader will hud an abstract of Mr. Woodward's treatment of the problem in Appendix B. HYPOTHESIS OF EXPANSION FROM WARMING. The second hypothesis involves considei'ations of temperature. The temperature of the earth's crust at the surface is identical with the mean annual temperature of the contiguous fluid, air or water, and at all subter- ranean points it is warmer, the change of temperature with depth being gradual. Every change of climate prcxluces a corresponding change in the sui-face temperature of the crust, and this change is slowly propagated down- ward. When the Bonneville Basin was full of water, there can be little question that the surface temperature was lower than at present, and it is possible that the corresponding diff'erence between the temperatures of the 378 LAKE BONNEVILLE. adjoiniiig laud, then and now, has not been equal in amount, in which case the i)ost-Bouneville warming of the crust beneath the lake area has been greater than the coincident warming of the crust underlying contiguous areas. Rise of temperature carries with it expansion, and the hypothesis is that such differential expansion pniducedthe observed differential altitudes. ( )ur (piantitative data are here less precise than in the case of the preceding hypothesis, but it is not difficult to assign to them reasonable limiting values, so as to obtain a practical test of the hypothesis. The mean annual tempera- ture at Salt Lake City is 51° F., and this may be assumed for the entire basin. Its ancient climate was somewhat colder, but the moderate development of glaciers permits us to entertain the assumj)tion that the difference was small. The lake, as we know from its wave wox-k, was not frozen, and as it had great depth, we are assured by the analogy of modern examples that its bottom temperature was that of water of maximum density, about 39°. The surface temperature of the crust in the lacustral area was therefore 12 degrees lower than now. If we assume that the surface temperature of the surrounding land was only two degrees lower than now, we are certain to underestimate the climatic change, and thus allow a maximum or limitinsf difference between the crustal changes under the old lake and under the old land. The problem then takes the form: What uplift can be referred to the expansion of the upper portion of the earth's crust consequent on a superficial rise of temperature of 10 degrees occurring at the close of the Bonneville epoch. The remaining constants necessary for its solution are obtained by assuming the coefficient of expansion of the rock involved to ])e 0.000006 for each degree, by adopting Sir William Thomson's coefficient of dift'usion of heat in the earth, and by assigning to post-Bonneville time a duration of 100,000 years, an estimate intentionally large. For the com- putation of the vertical rise of the basin from these numerical data I am indebted once more to Mr. Woodward, who has recently re\'iewed the sub- ject of subterranean temperatures from the mathematical side. His result (see Appendix C) is 1.28 feet, an amount quite too small for om- considera- tion in this connection. DEFORMATION^ OP EARTH'S CRUST. 379 HYPOTHESIS OF TERRESTRIAL DEFORMATION BY LOADING AND UNLOADING. The third hypothesis exphiius the pheiiouieiui Ijy assuming that when the BonneviHe Basin was tiUed with water, the earth yiekled to the weig-ht of the water, permitting a (h^pressiou of the headed area, and tliat when the water was afterward remo\'ed, tliere was a corresponding rise of the unhjaded area. The manner of yiekhng, the amount of vertical change, and tlie figure of deformation all depend on the constitution of the earth, and as that constitution is unknown, it is necessary to make assumptions regarding it iu order to discuss the quantitative sufficiency of the hypothesis. If the earth were perfectly rigid, the removal of the Bonneville load would not affect its form; if the earth were completely liquid, the removal of the load would cause the load to be replaced by the uprising of an equal weight of matter. Neither of these extreme conceptions can be entertained, for the visible portion of the eartli is neither liquid nor perfectly rigid, but between them is room for an infinite variety of special assumptions under each of which some deformation of the basin must be assigned to the unloading. In order to learn the order of magnitude of the greatest possible deformation, let us assume for a moment that the earth is constituted by a thin solid crust resting upon a liquid substratum, and that the rigidity of this crust is very small in comparison with the stresses applied to it by the removal of the water from the Bonneville Basin. The floor of the basin will then rise under the action of these stresses in some sort of arch, whose interior will be filled by liquid matter derived from surrounding- regions. The weight of tlie liquid matter thus introduced will lie approxi- mately equal to the weight of the water removed by evaporation, and the height of the crustal arch will be related to the depth of tlie water in (approximately) the inverse ratio of the densities of the two liquids. Tlie liquid rock may be assumed to agree in density with the average density of visible rocks at the surface, 2.75, and this gives us as the heiglit of the I'esulting arch the quotient of 1000 feet by 2.75, or 364 feet. This is the height attainable by the arch on the supposition that the strength of the crust is a vanishing quantity, and it is the superior limit of all possible values for the height of the arch. With the strength a vanishing quantity, 380 LAKE BONNEVILLE. the vertical stresses due to unloading are equilibrated by vei'tical stresses due to gravitation, and the height of arch is 304 ft ; with the strength finite, the stresses of unloading are equilibrated partly l)y stresses of gravitation and partly by elastic strains, and the height of arch is a function of tlie stresses of gravitation. While the natvire of this function is more conq)lcx than that of simple proportion, it is fair to infer from a comparison of tlic observed height of the arch of deformation, 129 ft., with tlie limiting heiglit, 3G4 ft., that under this hypothesis the stresses from unloading arc; satisiicd chiefly by elastic strains and secondarily by gravitational stresses.' Tluit this implies great strength of crust Ijecomes apparent when the magnitude of the load removed and the width of the affected area are considered. For the sake of illustration, assume that 129 feet of uplift satisfy tlie stres.ses due to 355 feet (129 X 2.75) of the removed water; there remain the stresses due to 645 feet to be satisfied by strains in the crust. Call the basin floor a beam, 120 miles long, supported at the ends, and sustained throughout by flotation so far as its own weight is concerned. Call the modulus of rupture of its material 3,000 pounds to the stpxare inch, and introduce no fiictor of safety. Consider the beam to be suljjected to upward stress 1)}' the removal of 645 feet of water from its entire upper surface, and compute by the engineer's formula the depth of beam necessary to stand the strain. It is about 32 miles.- The illustration is a rude one, because the floor of the basin, being attached all about its periphery, is stronger than a beam supported only at the ends; because a crust graduating into a li(piid beneath is weaker than a homogeneous crust; because the modulus of • If we postulate a tliick crust it is proper to postulate also that the matter flowing In beneath the dome has a greater density than superficial rock. Wilh the density 3.5 — an rxtronie assumption — tlie limiting height of arch is 280 feet. -The engineers' formula is where W is the breaking stress in pounds, the stress being evenly distributed over the upper surface of the beam; R Is the modulus of rupturi' of the material in poumls pir scinare inch ; b is the lireadlh of the beam, (fits depth and I its kiigtli. In the case under cunsideratlon W =D(j/h, in which I) is the depth, in feet, of water removed, and (/ is the weight, in pounds, of a column of water one inch scjuaro and one foot in height. Substituting this value for W in the formula, transposing and reducing, wo obtain 2 = .434 pound. Making 1 = 120 miles, D = G45 feet aud U = 3000 pounds, we find d = M.7 miles. UiSTLOADING AND UPAEOHING. 381 viscous distortion is less — possibly far less — tlian the modulus of rupture; and for other reasons; l)ut it nevertheless assists the iniaginatiou in i-ealiz- ing the relation of bulk to strength. Witli its aid I trust tlie reader \\ ill follow me in tlic conclusion that the hypothesis of local deformation of tlic earth l)y local unloading aifords results of the same order of magnitude as the observed distortion of the plane of the Bonneville shore, and is quanti- tatively adequate. The first and second hypotheses having been foinid quantitatively in- adequate, the third is the only one meriting further discussion. A thoi-ongli treatment is on the one hand highly desirable and on the other beset with difficulties. It is desirable because it pi'oraises to throw some light on the condition of the interior of the earth; a solid earth would not yield the same deformation as an earth partly liquid; a highly rigid earth would be- have differently from one of feebler rigidity. It is difficult because it must deal with magnitudes and pressures far beyond the field of experimenta- tion, and can be accomplished only Ijy the aid of comprehensive mathe- matical analysis. It requires an analytic theory of the strains set up by a stress applied locally to the surface of the earth and of the resulting defor- mation, and this theory must be so general as to include divers assumptions as to the variation of elasticity with depth from the surface, and as to the relation of the strains to the limits of elasticity.' The evolution of such a theory is beyond my power, but in the belief that it is worthy of the attention of the mathematician and physicist, I will endeavor to state the problem. Assume, first, that the rigidity of the earth is uniform throughout, or at least for some hundreds of miles from the surface, its modulus of elas- ticity being that of granite, fi)r example. Then conceive the application to the surface of a lenticulai- Ixxly of water etpiivalent to Lake Boinic\il]e, 'As defined by Sir Williani Thomson, " Elasticity of matter is that property in virtim of wbith a body requires force to chan};e its bulk or shape, and requires a coritiimoiis apiillciitioii of tlio force to maintain the chauj^e, and .spriuss baclc when tlic force is removed, and if left at rest without the force, does not remain at rest except in its ))rovions bulk and shape." Elasticity of bnlk and elaslicity of shape are distinct properties, which coexist iu solids, but not in liqnids. Kiii;iility is synonymous with elasticity of shape. Solids differ iu regard to rigidity in two ways. They have dilfcrcnt moduli of ri^dity and differeut limits of rigidity or elasticity. The niodulns of rigidity depends upon the .stress necessary to jiroduce a unit of deformation, or upon the deformation produced by a unit of stress. The limit of rigidity is reached when the force applied is so great that after its removal the solid does not return to its original shape. 382 LAKE BONNEVILLE. but spnmetric. To imagine the result, it is necessary to divest the mind of the ideas of brittleness and great strength ordinarih- associated with granite and other massive rocks. Brittleness is a, surface phenomenon only; at a depth of a few thousand feet, or at most a few miles, the tend- ency to fracture is effectively opposed by ])ressure. Strength is condi- tioned by magnitude, and in relation to magnitude it is a diininishing func- tion. Structures of the same form and material are not strong in pro])or- tion to their size but are relatively weaker as they are larger until hnally they can not sustain their own weight. In a general way strength increases with the square of the linear dimension; weight and otlicr luads increase M itli tlie cube. Giving due weight to these considerations, it is not improper to compare the earth when loaded by the water of Lake Bonneville with a bowl of jelly upon wliich a coin has been laid. The results in either case are, tirst, the depression of the area beneath the load, second, the formation of an annular ridge about it, and third, the production of strains within the mass. Conversely, the removal of the water of Lake Bonneville would pro- duce an uprising of the central area of the basin and an annular depression all about, and would either relieve the strains previoitsly produced by the addition of the water, or, if these strains had been otherwise relieved, would set up a new system with opposite signs. It is easy to understand from the homologous phenomena of jellies that the precise figure of tlie superficial deformation Avould de])end on the modulus of elasticity of the earth material. With a low elasticity the central arch would be high; with a high elasticity the figure of deformation would be comparatively low. There are two elements of complexity that inhere in the subject. In the first place, the deformation of the earth is resisted not only l)y the elas- ticit}' of the material but l)v gravitation, which always tends to give the siu'face the normal configuration of the gcoid. In the second place, the stresses created l)v the removal of the Homicvillc water wnuld have certain effects through the property of bulk elasticitv as well as that of shape elas- ticity. It is not improbaljle that a suitable discussion of tlie subject would demonstrate that the deformations ascribable to ])ulk elasticity are too small for consideration in connection with those referable to shape elasticity, l)ut to this extent at least thev Avould need to be considered. UNLOADING AND U PARCHING. 383 Add now a third element of complexity, by assuming tliat tlie strains set lip by the removal of the water are not entirely within the limit of elas- ticity of the material. Wherever they exceed the elastic limit, change of another sort occurs, probably not fracture, as in laboratory experiments on the limits of elasticity, but flow — such flow as Tresca's experiments have demonstrated for colloids.^ The plastic yielding of the rock in the region of greatest strains woidd cause a partial redistribution of strains in adja- cent regions, and would correspondingly modity the figure of deformation. The height of the central arch would be increased. Now add yet one other element of complexity, by assuming that the modulus of shape elasticity and the limit of shape elasticity vary (simulta- neously and harmoniously) in accordance with some law involving the dis- tance from the surface. They niay increase from the surface down^^^ard, or they may decrease from the surface downward, and in the latter case liquidity will at some depth be reached. The actual deformation should be comparatively low if the elasticity increases downward, and comparatively high if the elasticity diminishes downward. The application of an analytic theory of these relations could yield the best results only with a better determination than we now have of the elastici- ties of rocks, and with a better determination of the figure of the deforma- tion of the Bonneville Basin; but even with the imperfect data at liand it might establish a presumption for or against the existence of a liquid sub- stratvim beneath the rigid crust, and if the mathematical difficulties were surmounted, there can be little question that the observational data would be supplied, for their procurement is opposed by little beside their expense. Without waiting for the mathematician, we may conclude in a general way that the floor of the Bonneville area arched upward when the load of water was removed, and that this deformation was permitted by the feeble elasticity or the imperfect elasticity, or l)oth, of the portion of the earth affected; the conclusion being qualified l)y whatever weakness inheres in the postulate that the coincidence in time and place of crust unloading and crust deformation is not fortuitous >M6m de I'Inst. Savants strangers, vol. 18, 186S. 384 LAKE BONNEVILLE. EVIDENCE FROM THE POSITION OF GREAT SALT LAKE. In an earlier chapter attention has been called to the fact that in the central portion of the basin of the main body of Lake Bonneville mountain ridges are so nearly buried by lacustrine sediments that only their summits remain visible, jutting forth from the plain after the manner of islands. The amount of sedimentation implied is great, and its magnitude is like- wise indicated by the general evenness of the plain. Wherever the writer has crossed a portion of this plain, he has found himself, after leaving the foot slope of the contiguous mountains, upon a plnya Hoor with no discern- ible inclination, and nearly bare of vegetation. The saltness of tlie soil testifies that water does not flow across it, but rather stands upon it and evaporates. Another evidence of the general evenness of surface is the shallowness of Great Salt Lake, which has a mean depth of less than IT) feet. At the present time the principal contrilnition of debris toward the iill- ino- of the basin comes from the east. On the coast of Great Salt Lake deltas have been observed only at the mouths of the Jordan, the Weber, and the Bear, all rising in the Wasatch and Uinta Mountains and entering the lake on the eastern side. The western coast shows capes only where rocky hills stand near, and bays are found where it receives the intermit- tent drainage of the surrounding valleys. In Bonneville times the same contrast existed. The deltas of the old lake are found almost exclusively where it received streanis from the east, namely, the rivers just mentioned, their principal tributaries, which then entered the lake directly, and the Sevier River. No delta terraces were observed about the nortliern, west- ern, and southern margins, unless possibly in the Escalante Desert. If this deposition, so great in amount, iind dcrivccl so largely from tlie east, were the only factor concerned in the dctcnniHatioii of the configura- tion of tlie desert floor, that floor would be a gently-sloping jjlain, with its higher margin at the east and its lower at the west, and Great Salt Lake would lie at tlu^ base of the Gosiute Mountains instead of the Wasatch. The easterly position of the lake is unquestionnbly due to crustal move- ment, either orogenic or epeirogenic. (See PI. XLVIl.) ECCENTRICITY OF GREAT SALT LAKE. 385 Let us first consider tlie possibility of an erogenic cause. The most conspicuous recent orogenic change in the region is that shown by the fault scarps at the base of the Wasatch Range. These scarps show differential movement, either ascent of the mountain or descent of the valley, or both. The great size of the mountain range, as argued on an earlier page, assures us that a rising of the range is at least a part of the displacement, but is not opposed to the idea that the sinking of the valley is a correlative and perhaps equal part. It is consistent with this idea that the water of Great Salt Lake between the Bear and Weber deltas, and again between the Weber and Jordan deltas, approaches within about a mile of tlie great fault at the mountain base. Epeirogenic causes may be considered from two points of view: first, as belonging to a system of changes correlated with the emptying of the basin by evaporation; second, as belonging to the more general system of changes to which the basin, as such, may be ascribed. Taking the first point of view, we have a post-Bonneville rising of the central area amount- ing to more than 100 feet, and it is conceivable that this has divided the plain into two basins, of which the lake occupies one, while the other con- tains only occasional playa lakes, such as the scant rainfall of the tributary regions is able to produce. Too little is known of the configuration of the desert west of the lake to determine whether it is partitioned off by a bar- rier of such sort, or is in time of great rainfall tributary to Great Salt Lake. But there are other reasons why the hypothesis can not be seriously enter- tained. In the first place, the area of maximum uplift, so far as our meas- urements determine it, coincides with the western portion of the lake instead of with the line of low ridges beyond it. The old shore-line is higher on Promontory Ridge than on the Terrace Mountains to the west- ward. It must also be borne in mind that the present condition of the Ijasin as affected by climate is substantially identical with the [)re-Bonneville condition, and the arid phase was of long continuance before the Bonne- ville flood. Whatever central elevation is recorded by the surviving shore- line is merely the correlative of central depression during the lake period, and to assume the post-Bonneville uplifting of the plain into a barrier ade- MON I 25 386 LAKE BONNEVILLE. (piate to contain the lake is to assume that during the existence of tlie lake the central depression was filled by sediments so as to pi-oduce a lake bot- tom almost absolutely level. From wliat we know by observation of the slopes on which the Bonne\'ille sediments were able to lie, we can not believe that this was accomplished, but rather that throughout the deeper portion of the lake there was an equable deposition over gentle slopes, the depth of deposit increasing rather toward the source of the material at the east than toward the center of the lake. It is pn)l)iible that post-Bouneville changes in the configuration of the plain, so far as they have depended epei- rogenically on the removal of the water, have been the simple converse of changes due to the previous imposition of the water, and have practically restored the preexisting condition. Turning to epeirogenic considerations of a more general nature, we see that the Bonneville Basin is a region of depression, surrounded on the south, west and north by regions of somewhat greater elevation, and on the east by a tract whose mean altitude is several thousand feet higher— an irregular plateau, along the edge of whit-h the Wasatch Range stands as a parapet. The forces which produced this bashi and the plateau to tlie east of it are of necessity independent of the loading and unloading of the basin, and of a more general nature. Whatever they may be, it is not irrational to appeal to them as the cause of the local depression containing Great Salt Lake and to regard that depression as a result of the mere continuance, with possibly greater localization, of the process which created the larger basin. Whether, then, we regard the peculiar position of the lake as a result of orogenic or of epeirogenic dis])lacement, we are comijelled to forego tlie assignment, even tentatively, of a special hypotliesis as to its causr. I*ci'- haps tlie most valuable cont'lusiou to be drawn is tliat, as drposition witliiii tlie liasin, during Imiiiid and arid phases of climate alike, lias (•(Hitiiiiially tended to build the eastern lialf of the plain liiglicr tliaii the western, and. as this tendency has continued to the present time, the sul)sidence opposing and thwarting it has likewise continued to a late epoch and is probably still in progress. COROLLARY. 387 THE STRENGTH OF THE EARTH. The writer has been led by the discussion of these phenomena to a conception of the rigidity or strength of the earth, more definite than he had previously entertained. It would not be proper to call this conception a conclusion from the data here presented, or a result to which they rigor- ously and necessarily lead. It is rather a working hypothesis suggested by the study of Lake Bonneville. If the earth possessed no rigidity, its materials would arrange them- selves in accordance with the laws of hydrostatic equilibrium. The matter specifically heaviest would assume the lowest position, and there would be a gradation upward to the matter specifically lightest, which would consti- tute the entire surface. The surface would be regularly ellipsoidal, and would be completely covered by the ocean. Elevations and depressions, mountains and valleys, continents and ocean basins, are rendered possible by the property of rigidity, but the phenomena of diastrophism, and espe- cially those of plication, show that this rigidity has its limits, and the phenomena of volcanism demonsti-ate that its distribution is not uniform. It has been computed by Darwin' that if the earth were homogeneous throughout, the stress differences occasioned by the weight of continents would be as great as those necessary to crush granite. The stress differ- ence necessary to produce viscous flow in granite and allied rocks is not known, but if different from the crushing stress, it is less; and Darwin's discussion therefore tends to show that the earth, if homogeneous, would require a strength equal to or greater than that of granite. Tliat the earth is not homogeneous as regards density (and does not consist of symmetric homogeneous shells) is shown l)y the massing of land ;ireas in one hemis- phere; and the hypothesis that the crust has low density beneath continents and high density beneath oceans is sustained by observations on the local direction and local force of gravitation at various points.- The general proposition, tacitly postulated by Babbage and Herschel, advocated more ' On the stresses caused in the interior of the earth hy the weight of continents and nioiiutains, by G. H. Darwin. Phil. Trans. Royal Soc, pt. 1, 1882. 2 On the argument from geodetic station errors see John H. Pratt, Figure of the Earth, p. 201. On the argument from pendulum observations see H. Faye iu Revue scientifniue for Feb- 20 and March 27, 188G. 388 LAKE ];ONNEVILLK. recently by Dutton and Fisher, and entertained ])y most modem writers, is that the radial elements of the sphere have the same weight on all sides, the product of the height of each unit colunui into its mean density being everywhere the same. With such a distribution of densities the stresses and strains resulting from the existence of continental elevations do not disappear, but they are less than those derived by Darwin on th^ hyi)oth- sis of homogeneity. How much less has not been shoAvn, but it is fair to say that, so far as the evidence from continents is concerned, the (juestion of the degree of rigidity of the earth's nucleus is still an open one. If a weight l)e added to a limited portion of the surface of the globe, there will result a system of strains beneath and aljout the area, and a defonnation of the surface accordant witli the system of strains. If the weight is small, and if the effect is not complicated by preexistent strains, the resulting strains will at every point fall within the limit of elasticity of the material, and the deformation will be small. If the weight is sufficiently large, the resulting strains will in some places exceed the limit of elasticity, and other consequences will follow. Among these, rupture and faulting may in special cases be included, but the ordinary and predominant res\dt will be viscous flow. The viscous flow will consume time, and when it lias ceased, there Avill remain a system of elastic strains. Beyond the elastic limits, the laws of change for loading the surface of the earth (and similarly for unloading) are quasi-hydi'ostatic. The point on which the Bonneville jdienomena appear to throw light is the magnitude of the load necessary to overpower rigidit}-. The })lie- nomena of faulting at the base of the Wasatch, whether considered liy themselves or in connection with the filling of the adjacent valley ^^ itli water and its subsequent emptying, appear to my mind best accordiint with the idea that the Wasatch Range and the paralh'l ranges lying west of it are not sustained at their existing heights above the adjacent plains and valleys by reason of the inferior specific density of their masses and of the under- lying portions of the crust, but chiefly and perhaps entirely in virtue of the rigidity or strength of the crust. The phenomena of deformntion of the Boimeville shore-line accord best with the idea that the imi)osition of the Bjnneville load of water and its subsequent i-emoval strained the subjacent MEASURE OF RIGIDITY. 389 portions of the crust beyond the elastic limit, the stresses due to tlie load- ing- and unloading- being- partly equilibrated by crustal strains, and partly relieved by crustal flow and a resulting redistribution of the stresses due to gravitation. It is indicated that the limit of terrestrial rigidity falls some- where between that measured by the weight of the Wasatch Range and that measured by the weight of the water of the main body of Lake Bon- neville, or in more general terms, that a mountain of the tirst class is the greatest load that can be held up by the earth, and is therefore an expression of its strength or of the limit of elasticity of the material of its outer layers. Fully to realize the nature of this measure, it is necessary to give it numerical expression, and to this end a few computations have been made. It is evident that the maximum strain produced by a load depends in part on its distribution, and especially that a long ridge taxes rigidity less than a compact mountain mass of the same weight. It appears to me that a very long range causes no greater strains than a shorter one having the same cross section, and I have therefore conceived the Wasatch Range to be fairly represented for this purpose by a division of it including the highest peaks and having a length not quite double its width. This di\ision extends from the Provo River northward to the low pass at the head of Parley's Canyon. Its estimated volume is 200 cubic miles. Similar considerations lead me to base the estimate for Lake Bonne- ville on the main body instead of the entire lake, excluding not only the Sevier body but Snake Valley, Wliite Valley, and Utah Valley bays. Thus defined, the load of water amounted to about 2000 cubic miles, equiv- alent in weight to 730 cubic miles of rock. On the assumption that the strains produced by the lifting of this load were only in minor part relieved by viscous flow, it is inferred that the limit to tlie superficial rigidity of the earth is expressed by a load of 400 to 600 cubic miles of rock (1670 to 2500 cubic kilometers). There are four classes of topographic features with which tliis measure may advantageously be compared, and by which it may perlia])s be tested. The first is mountains of addition, or mountains produced by the mere addi- tion of matter to the surface of the earth. IMost volcanic cones are of this class. The second class consists of mountains by subtraction, or residuary 390 LAKE BONNEVILLE. mountains clue to the removal of surrounding material. The third class is intermediate, including addition and subtraction, as when the extrusion or intrusion of volcanic matter produces a resistant mass cnpable of preserving against erosion a residuary mountain. The fourth consists of valleys by subtraction, or valleys eroded fi'om plateaus. Mountains and valleys due directly to diastropliism are not in point, because, as they an; the super- ficial expression of indviiown subterranean changes, we can nut be sure in individual cases that their existence is independent of the sul)tcrr;inean dis- tribution of densities. For similar reason, a volcanic inoinit.iin whose building has been accompanied by subsidence of tlu^ subjacent tcrrane can not be used for comparison. The contour maps ])repared by the geogi-a])hic branch of the Survey enable me to give the volumes of some of the most imjxtrtant American examples of these various classes with a degree of precision (piite sufficient for the purpose. By their aid each of the following' features was referred, not to sea level, but to the plane of the surrounding country, and its vol- ume was computed. San Francisco Mountain is a volcanic cone standing alone on a high plain, and tlie strata about its base are almost undisturbed ; it is a typical mountain by addition. Its volume is 40 cul^ic miles. Mount Shasta is a volcanic cone standing in a region of disturbed strata, but there is no evidence of subsidence due to its load. Its volume is SO cubic miles. Mount Taylor is a volcanic cone standing on a plain tioored with hard lavas. The degradation of the surrounding country has converted the mA- canic plain' into a great mesa or table mountain. The cone and mesa together, constituting a mountain by combined addition and subtraction, have a vol nine of 190 cubic miles. Tlie Henry Mountains and the Sierra La Sal consist each of a grou]) of laccolites — volcanic additions by intrusion — and of other rocks preserved bv them from the erosive reduction sustained bv the sin-rounding ])lateau. Their vohimes are respectively 230 and 250 cubic miles. The Tavapiits plateaii of the Green River basin, otherwise Ivnown as Roan Mountain, is a great mass of inclined strata carved out by the unequal U S. GEOLOGICAL SURVEY U^J-'Jl B^'I'II'TE-.TLLE FL. LI Juhut:. Bien it <;•:■ Lith .., U I H }:vi.sl..i SKETCH MAP OF BLACK ROCK AND VICINITY, UTAH PREPARED TO SHOW THE POSITION OF THE GRANITE POST KNOWN AS THE BLACK ROCK BENCH. Surveyed in 1877, by G.K.GilbeiL . MOUNTAIN VOLUMES. 391 degradation of a still greater anticlinal. Its determining cause is a thick layer of resistant rock lying between thick layers of yielding rock, and it stands between two nionoclinal valleys due to the excavation of the yielding layers. Its volume standing above the level of the adjacent valleys is about 700 cubic miles. The Grand Canyon of the Colorado is a valley cut from a great })lateau of stratified rock. The plateau has a fault structure of its own, but the canyon and the fault structure have different directions and are manifestly independent. The volume excavated to form the deeper part of tlie canyon, from the mouth of the Little Colorado to the mouth o( Kanab Creek, is 350 cubic miles. The Appalachian Mountains are traversed for nearly a thousand miles l)y a great valley following the outcrop of yielding rocks, and it is probaV)le that we have here a valley by subtraction. For the same reason that determined the selection for measurement of a portion only of the Wasatch Range and of a portion only of Lake Bonneville, measurement was not made of the whole of this valley, but only of a limited part. It was assumed that a section with length fifty per cent, greater than breadth, and selected where tlie valley is broadest, fairly represents the strain-producing power of the whole valley. The portion thus selected lies 600 feet below the mean height of the Cumberland Plateau on the northwest and 1000 feet below the mean height of the mountain district of North Carolina on the southeast, and its volume, computed from the mean of these, is 800 cvubic miles. All of these various features except two fall within the indicated limit of GOO cubic miles, but the limit is exceeded by the Tavaputs Plateau with 700 and the Appalachian valley with 800 cubic miles. There are qualify- ing considerations in each case. The plane above which the volume of the Tavaputs Plateau was computed was that of the low valleys adjoining it; |)erhaps a more suitable plane of reference would have been the general level of the surrounding country. The density of the rock of the plateau is probably less than 2.75, the density assumed in reducing the volume of the abstracted lake water to equivalent rock volume. The Appalachian valley lies in a region of great corrugation, and its trend coincides with the strike of the orogenic structure. That structure unquestionably involves 392 LAKE BONNEVILLE. inequalities in the distribution of subteiTanean densities, and it is possible that the strains due to the valley are lessened by the presence beneath it of exceptionally heavy matter. But after giving due weight to these considera- tions, it must still be admitted that the measure of strength does not stand well the test applied. It is indeed possible that a true measure has Ijeen found, and that it is illustrated by the Bonneville, Tavaputs, and Ai)palachian phenomena, but we can not deny the equal possibility, first, that the strength of the earth varies so widely, in different places that a measure discovered in the Bonneville basin serves merely to indicate the order of magnitude of a measure of the average strength, or second, that the unloading of the Bonneville basin occasioned no greater strains than the crust was able to endure, and that the coincidence of unloading and uparching was a coin- cidence merely. CHAPTER IX. THE AGE OF THE EQUUS FAUNA. THE FAUNA AND ITS PHYSICAL RELATIONS. As the Equus fauna is not known to occur in the Bonneville Basin, the presence of this chapter requires explanation. In considering the relation of the Bonneville history to glacial history, it has been found necessary to consider also the glacial and lacustrine records of the Mono and Lahontan Basins; hence the sixth chapter contains an exceptionally full discussion of the relation of the later lacustrine history of the Great Basin to general geologic chronology. The Equus fauna is so connected with that lacustrine history that the geologist can best discuss its age in that connection. The present chapter is a corollary to Chapter VI. The same explanation serves to account for the discussion of the fauna by the present writer, who has not visited the chief localities of its occur- rence, but derives his knowledge of its geologic relations from the writings' and notes of Russell and McGee. Equus appears to have been first used in the nomenclature of geologic history by Marsh, in an address read to the American Association for the Advancement of Science in 1877." The Equus beds are there made an upper division of the Pliocene, and they are characterized in a table accom- panying the address by the genera Equus, Tapirus, and Elephas. An exam- ination of the text shows that none of these genera are credited to the lower Pliocene, but that all are credited to the post-Tertiary. The characteriza- tion thus fails to separate the Equus fauna from the Pleistocene, and as no 'Fourth Ann. Kept. U. S. Geol. Survey, pp. 4r)8-461. Science, vol. 3, 1884, pp. 322-323. -The Introduction and Succession of Vertebrate Life in America. By O. C. Marsh. Proc. A. A. A. S., vol. 26, 1878, p. 211. 393 394 LAKE BONNEVILLE. locality is mentioned, it leaves the fauna undefined. Two years later the fauna was charactei'ized hy Cope by the following list of mammalian spe- cies.^ Those of the left hand column are extinct, those of the ridit hiiiid column living. Mylodon sodalis. Tliomomys near vlushis. Lutra uenT pincinaria. Thomomys talpoides. ElephiDi primifjeuhis. Caslor filirr. Equus occidentaUs. Canis lalruns. Equus major. Anchcnia henierna. Avchenia magna. Anchenia vitakeriana. Cerims fortis. As the species of this list Avere found together at one horizon and in the same locality, they afford a definite and tangible basis for discussion, and I shall consider them as the Equus fauna, despite the fact that they fail to include the genus Tapirus referred to it by Marsh. The locality was de- scribed by Cope as lying thirty or forty miles east of Silver Lake, Oregon,^ and he styled it "Fossil Lake." Russell, who visited the place in 1882, speaks of it as a few miles eastward of Christmas Lake. The formation in which the bones occur is lacustrine, as shown by its shells. It constitutes the floor of a desert valley, and has suffered scarcely any erosion, though the sand dunes traveling over it suggest that its surface may have been somewhat degraded by wind action. All about the sides of the valley are shore-lines, and above these shore-lines the lake beds are not found. Just as in the Bonneville and Lahontan basins, the physical relations indicate that the shore-lines and lacustrine sediments are coordi- nate products of the same expansion of lake waters. The Christmas Lake basin is part of the Great Basin, and lies L50 miles northwest from the Lahontan shore-lines. Each closed vallev of the intervening region has its ancient shore-line and associated lake beds. Each of the old lakes thus demonstrated stands witness to climatic oscillation, and their geograi)hic relations leave no room for question that they jjertain to the same climatic oscillation and therefore have the same date. 'E. U. Cope: Bull. U. S. Geol. & Geofr. Survey of tlio Territories, vol. 5, 1879, p. 48. 'Americau Naturalist, vol. 16, 18e2, p. 194. CORRELATION OF EQUUS AND LAHONTAN FAUNAS. 395 The mammalian remains obtained from the Lahontan l)c(ls inckide a great proboscidian (^Elephas or Mastodon), a llama, one or more horses, and an ox. No skeletons were found, and the dissociated bones and fragments of bones are not such as to permit the recognition of species; but Prof. Marsh, to whom they were submitted, was able to say with entire confi- dence that the specimens as a whole belong to the Equus fauna. Having myself compared the Lahontan collection with the collection made by Mr. Russell at the Christmas Lake locality, I may be permitted to add that I share Prof. Marsh's confidence in the identity of the faunas. The correlation receives additional support from the lacustrine shells. Russell repoi'ts from the l)one beds near Christmas Lake the following species:' Sphwriiim dcntatum. Limnopln/fta hiilimoides. PifiuliHin nltrnmontanum. Garinifex newhcrryi. Helisoma trirolvis. Valvatn rireiis. Gyraiihis rermicularis. None of these are extinct, and all have been found in Lahontan strata. Nearly all of the bones obtained from the Lahonton strata were found at a horizon somewhat above the middle of the upper division of lake beds. At "Fossil Lake" the bones were found at the top of the formation, but we know nothino- of the thickness of the formation. Unless the Fossil Lake formation is much thinner than the Lahonton, the date of its discov- ered mammalian fjiuna is a trifle later. The physical relations recited above, and the associated paleontologic relations, show that the Equus fauna, as illustrated by its type locality, belongs to the epoch of the Upper Lahontan. It therefore falls, as a mat- ter of general chronology, in the later Pleistpcene. This conclusion ditfers widely from that reached by purely paleonto- logic methods, for these refer the fauna to the later Pliocene. Before they are considered, attention will be called to a possible ambiguity, and one of the lines of physical evidence will be amplified. The term Pleistocene is used by geologists in two senses, one of which may be characterized as chronologic or general and the other as physical ' Fourth Ann. Eept. U. S. Gcol. Snrv., p. 460. 396 LAKE BONNEVILLE. or local. In Europe the later part of Cenozoic time was tUstingtiislied 1)\- a series of physical events including one or more epochs of exce})ti()nal cold and exceptional expansi(»n of glaciers. In European nomenclature Pleistocene is applied to the period of time occupied by these events, and also to the events themselves, and this without confusion. In North Ajner- ica the later Cenozoic history included a series of events of the same gen- eral character, and for these we have borro\\ed the name Pleistocene, or its synonym. Quaternary. The time covered by these events may or may not coincide with the Pleistocene period, and until it is shown so to coin- cide, our imported term is ambiguous. It is primarily in the physical rather than the chronologic sense of the term that the Upper Lahontan and the Fossil Lake beds are found to be late Pleistocene. Properly to characterize them in the chronologic sense — with reference to the period including the glacial and interglacial epochs of Europe — it is necessary to take account of the work of land sculpture and its relative progress in dif- ferent places. When a surface shaped by some agent other than atmospheric — a sea floor, for example, a moraine, a shore terrace, or a terrace modeled by man — is exposed to atmospheric agencies, its sculpture begins. For a long time its original feattires continue to be the characteristic ones, but they eventually become subordinate and finally disappear. The original foriTis at first are new and fresh, then old, worn, and hard to discover; and finally the fact that they once existed can be known only from the internal structure of the deposits to which they belonged. So long as the original form is discernible, it yields to the geologist evidence of relative newness or rela- tive age. Such evidence as this is not readily formulated, but it is con- stantly employed by the field geologist in the study of the surface. Indeed it affords one of the most important liases of tlie wide spread opinion tliat glaciation was simultaneous in Europe and America. The abandoned lake shores of Christmas Valley and of the Lahontan Basin, the lacustrine plains below them, and the correlated glacial moraines, are all of youthful habit, as youthful as the "parallel roads" of Glen Koy and other surface features marking the wane of glaciation in Scotland. The lake shores and sea shores associated with tlie latest Pliocene beds of COMl'AKATIVE SUULPTUKE. 397 Europe are eitlior iinrecog-nized, or else, as in the case of the Enghsh Cray, known only by their internal structure. The plains of their upper surfaces, where not covered by glacial or volcanic deposits, are either obsolete or obsolescent. The topogra})liy created in the presence t)f the Equus fauna is young; that created in the presence of the European Pliocene fauna is old. With the aid of this additional link in the chain of physical evidence, the geologist ties the Equus fauna, not merely to the American glacial or Pleistocene history, but to the Pleistocene time division. The ancient Lake Bonneville, the ancient Lake Lahontan, the ancient lake of the Mono Basin, the ancient lake of the Christmas Lake Basin, and numerous smaller extinct lakes of Oregon and Nevada, are tied together by community of physical characters — freshly bared sediments, conforming to the slopes of surface and surrounded by freshl}- formed shore-lines. Many have yielded shells of recent species. Two, those of the Lahontan and Christmas lake basins, have yielded the same mammalian fauna. The two largest, Lahontan and Bonneville, have yielded detailed and parallel physical histories. The analysis of climatic factors correlates them with ancient glaciation in neighboring mountains, and their shores are carved from and built around late-formed moraines of the Wasatch Rangre and the Sierra Nevada. The detailed history shows two lacustral epochs corre- sponding to two glacial epochs, and correlates the mammalian fauna with the later half of the later glacial e[)i)ch. Presumptively this date falls very late in the Pleistocene period. The phenomena of comparative sculpture show that it is at least later than tlic latest Pliocene of Europe. THE PALEONTOLOGIC EVIDENCE, So far as I am aware. Cope alone has stated the jjaleontologic grounds for referring the Equus fauna to tlie Pliocene. Comparing it with the sub- Appenine fauna of Europe (Pliocene), he says— "The characteristic of this fauna is the fact that the species belong mostU' to existing genera. . . In the Equus beds of Oregon, a few extinct genera in like manner share the field with various recent ones, while not a few of the bones are not distin- guishable from those of recent species." Li a succeeding paragraph he adds: "As a conclusion of the comparison of the American Equus beds in 398 LAKE BONNEVILLE. general with those of Europe it may be stated that the number of identical genera is so large that we may not hesitate to [)arallelize them as strati- grapliioallv the same.'" Three eategories of evidence are here used: (1) the relative abundance of extiuft genera in the two faunas, (2) the relative abundance of extinct species iu the two faunas, (3) the abundance of genera common to Ijoth faunas. The first and second categories embody the nu;thod devised by Lyell for the classifit-ation of Tertiary formations, a inetliod Ijased on the })ercentage in each fauna of living or extinct forms. Faunas with the hjwest per cent of recent forms were grouped together as Eocene, those with a certain higher per cent were called Miocene, and so for the Pliocene. The method rests on a generalization from observation and on a postulate. The generaliza- tion is that from the earliest Eocene time the fades of life has "■raduallv approached the present fiicies. The postulate is that the rate of change has been uniform in all places. If the postulate is true, the method of L\-ell can yield exact time correlation; otherwise it can }'ield only approximate time results. Lyell himself disclaims belief in the postulate and regards his classification as cln-onologically imperfect.^ 'These passages occur ou pages 47 and 48 of a paper on Thu Kelatiousof the Horizous of Extinct ViTtel>rata of Eiirupo ami North America, ]mbli,she(l in volume V of the Bulletins of the IT. S. Survey of the Tirritoiies. Ou paj;e 4'.) the correlation of the Eiiuns beds with the Pliocene is characterized as the "exact idcutilicatiou " of a restricted division. The autlior's conlidence iu the correlation was not materially shaken by a pridiininary statenieut of the physical evidence made by the writer to the National Academy of Science iu 1886. See American Naturalist, vol. \.\I, 18-i7, p. 4.jlt. In the jiassa^e last referred to Cope says: "This gentleman [Gilbert] has expressed the belief that the beds of this age are not older than the glacial ejioch, because they embrace the bases of some of the moraines of some of the ancient glaciers of the Sierra Nevada. It remains to be proven, however, that these moraines are of true glacial age, since they are of entirely local character. The preseuco of so many mammals of the fauna of this valley of Me.xico would not support the belief iu a cold climate." When the moraines referred to were lieing formed, the Sierra Nevada bore eu its back a mer-de- glace as extensivi; as that of the Alps, and a host of glaciers llowed from this to the valleys below, reaching altitudes from (i,(IOII to ",),000 feet lower tiian the littli> glaciers that now cling to a few of its peaks. At the sane^ time there were also great glaiuers lu the Wasatch Mountains. Whatever infer- ences these phcncMuena yield as to the contemporaneous climate of \\w. Great Hasin a])pears to me ipiiti' independent of the question of thinr correlation with a glacial ep leh souu'where else. If the glaciers prove a cold climate iu the Great Basin, then the animals that Icit their bones in the contemporaneous lake sediments of the Basin lived iu a cold climate. If the animals could not live iu a cold climate, then it is shown that the valleys of the Great Basin were warm despite the icoou the high ii.onntains. The question ot' geologic date is not involved. The value of the Ecpius fauna as an index of contemporaneous climate has already been discussed in chapter VI of this volnmi'. ■'Sir Charles Lyell. M.anual of Geology, IJth ed. New York. p. III!. METHODS OF PALEONTOLOGIO CORRELATiOISl. 399 The third category of evidence, the abundance of common elements in two faunas compared, is that ordinarily used in paleontologic correlation, and it aj^plies to the older formations as well as to the Cenozoic. The method of using it is analogous to the assignment of commercial colors to their approximate positions on the prismatic scale, and may be character- ized as a method of matching. Having in one district a number of faunas determined by physical relations to be successive, the paleontologist com- pares a single fauna of another district with each of these severally and " correlates " it with tlie one with whicli it has most in common. The prin- cipal check on this method lies in the consistency or inconsistency of its results with one another. When two faunas of one district are separately compared with the faunal scale of another district, their relative ages as inferred from the results of matching is usually the same as shown by their physical relations, but there are a few exceptions to this. Again wlien l)iotic data of two or more kinds, as for example vertebrate fossils, inverte- brate fossils and fossil plants, are separately employed for correlation by matching, the results are often accordant, but they are also often discord- ant. How far the discrepancies of result are due to imperfection of method and how far to imperfection of data, is not known, but it is generally admitted that there are limits to the applicability of the method. The greatest discrepancies in its resuhs liave been found wlieu the formations compared lie far apart, so as to fall in different faunal j)rovinces; audit may be said in general that its value varies dii'ectly with the degree of resemblance'of the faunas compared. Where the whole number of common forms or of common types is small, cttrrelation is less precise than where the lunnber is large. In order to gauge the Equus fauna by the accepted scale, I iia\e selected a series of European faunas more or less restricted geographically and of well-known age. They are (1) the Lower Pliocene of Montpellier, France, (2) tlie Upper Pliocene of the Arno Valley, Italy, (3) the Pleisto- cene of Great Britain, (4) the living fauna of Europe. The genera and species of the land mammals of these faunas have been compared with those of the Equus fauna and the accompanying ta]:)le constructed. 400 LAKE BONNEVILLE. The table includes only mammalian faunas. Cope has reported from the same Oregon locality ten species of birds' and two of fishes,^ but these are not at present available for purposes of correlation. As it is known that the general rate of evolution differs in different classes of animals, the entire Fossil Lake fauna can not be considered together. The birds can not be separately used because of the scantiness of avian data in the Euro- pean faunal scale. The fishes are themselves too few for profitable com- parison. Table XVII.— Summary of Paleontologic Data for Ike Determination of the Age of the Equua Fauna. TciTustrial mamnjaliau fauDas. Available for com- pariBon. Method of Lyett.— Peicentace of ex- tinct Method by match- ing.— Xutubi-r in cui'ntnnn with the Eiimirt fiiiina. genera, speciea. genera. ypeciea. genera. species. mauy | many 27 48 9 ' 13 18 29 14 15 0 7 *n 0 19 69 4 6 1 2 Pkiatiiceno (Great Britain)' UpptT Pliocene (Val d'Arno)*.. Lower Pliocene (Moutpellier)*... 11 21 100 100 «6 0 2 0 < Britisli Pleistocene Mammalia. By W. B. Dawlsins and \V. A. Sanfonl. Palaeontographical Society, vols. 18 and 32, 1866 and 1878. Plcistocei.o climate, etc. By W. Boyd Dawkin.i, Pop. Sci. Review, vol. 10, 1871, pp. 388-397. 'V. I. Forsyth Major. Atti Soc. Tosc. Sci. Xat., vol 1, pp. 39-40 and "Proc. verb.," vol. 1, p. v. ^Gervais, quoted by Major. Atti Soc. To.hc. Sci. "Nat., vol. 1. pp. 224-225. *ln a publication Hubaequent to the one tin wiiicli Ibis table is b.aaed. Cope cstablislies a now genus, Holomenucug, to which ho transfers the species o( Ajichenia in the Equus fauna. This doubles the number of e.\tinct genera in tlie fauna and rai8e,s its percentage fioni U to 22. ^Tliis number includes the genus iufra, which is not reported from this formation. As it is reported from Ihe preceding and following formations, its existence at that time can not be questioned. The numerical results by the matching method appear in the two col- imms ;it tlie right. The six geneni of tlie Equus fauna foxmd in the upper Pliocene are identical with those of the Pleistocene, and include those of the lower Pliocene and living faunas. The two genera found in the Pleis- tocene l>ut not in the living fauna of Europe are Equus and ElvpJias, which persist in other continents. One species, Castor fiher, is conuuon to the Equus, Pleistocene, and Recent faunas. EJcphds jmmigenius, common to the Eqiuis and Pleistocene, is said to occur in Europe exclusively in the Pleistocene. The evidence from genera is ambiguous. That from s])ecies > Bull. U. S. Survey Terrs., vol. 4, 1878, p. 369. «Americau Naturalist, vol. 12, 1878, p. 125. THE PALEONTOLOGIO EVIDENCE. 401 tends to correlate the Equus fauna with the Pleistocene of Great Britain, but the number of common foiTns is so small that their testimony lias little \\'eight. The numerical results by the Lyellian method a])})ear in tiie middle pair of colunnis. The Equus fauna agrees with the Up})er Pliocene in its ratio of extinct genera; and in its ratio of extinct species it stands rather nearer the Pliocene than the Pleistocene. Tlie evidence from genera is weakened by the fact that the numbers involved are very small; of 9 gen- era from Fossil Lake 1 is extinct, of 1innin<>" of the Pleis- tocene. It certainly does not agree with the physical evidence in indicating late Pleistocene. If all this paleontologic evidence coidd lie pro])erlv coml)ined, giving each element its due weight, the resulting indication of date would he later tlian tlu^ upper Pliocene of the Arno Valley and earlier tlian the middle of the Pleistocene of Great Britain. It might fall in an assumed interval between the two time divisions, or it miglit fall in tlie earlier part of the Pleistocene. At the very l)est, the ilate inferred from the physical tacts and the date inferred from the biotic facts differ liy more than half the extent of the Pleistocene jieriod. Botli can not be triie; which sliould l)e accepted? For my own jiart 1 do not hesitate to prefer the physical I'videiice and llie later date. I hold with Lyell that "we can not presume tliat tlie rate of foi-mer alterations in the animate world, or the continual going out and coming- in of species, has been everywhere exactly ecpial in equal quantities of time;" and the Equus fauna seems to me to illustrate the principle. It may perhaps be found, wdien the fauna is much better known, that its features correspond closely with those of the contemporary fauna in Europe, but for the present it appears that the mammalian fauna of the MON I 26 402 LAKE BONNEVILLE. Groat Basin experienced a greater change at the close of the Pleistocene tliau did that of Europe. In the study of tlie Pleistocene of Europe, geology and paleontology have worked together with adinii-able results. The geologic relations have given to paleontology tlie sequence of its faunas; paleontology has recii> rocated by correlating the deposits of extrn-ghicial regions with elements of the glacial history; and through such cooperation a bewildering multi- plicity of data are being mai'shaled into a consistent though complex sys- tem. In America the same benefit should result from the same coopera- tion. Some Pleistocene deposits can be assigned dates through their rela- tions to glaciation, and when the faunas and floras of these are known, paleontology can contribute much toward the discovery of the Pleistocene history of districts remote from glaciers. For this purpose the Lyellian method of percentages is, in my judgment, far less valual)le than the method by matching; but the standard scale for matching should be an American scale, based on physical studies in the region of Pleistocene glaciation and its immediate vicinity.' ' While these pages are passing through the press, a vohime is published by Messrs. Felix and Loiik, cimtaiiiiug an account of Pleistocene lacustrine formations in the Great Valley of Mexico. In a general way the phenonicuii of the Ijonneville and Lahontau basins are there repealed, but the history ot the climatic oscillation has nut been fully nuidc out. In undisturbed strata, forniiug a continuous series with lake sediments now being deposited, there have been found bones of thirteen maunnaliau species, and two of these species are identical with members of tlu^ Christ ui.-i.s Lake fauna. (Beitrjigo znr Geologio und Paliioutologio der Kepublik Mexico, Vou Dr. J. Felix uud Dr. 11. Leuk. Part 1. Leiiizig, 18U0, pp, (55-06, 7y-88.) APPENDIXES. A.— Altitudes luid tbeir detcrmluatiou. By Albert L. Webster. 15. — On the deforiuatiou of the seoiil '»y the removal, tbrousli evaiioiation, of the water of Lake Bonneville. By K. S. Woodward. C. — On the elevation of the surface of the Bonneville Basin by expansion due to change of climate; By R. S. Woodward. 403 APPENDIX A. ALTITUDES AND TIlEIIi DETERMINATION. By Albert L. Webster. In connection witli the study of the records of the ivnciont Lake Bonneville, it hecame a matter of interest to ascertain the present relative altitudes of points scat- tered alonj;- its fornier perimeter. A comi)lete and thoroughly satisfactory investiga- tion of the subject being' impracticable from economic considerations, it was made ■subsidiary to the more general historic study of the lake, and its results are accord- ingly incomplete or lacking where such study would not permit of a more extended investigation. As far as practicable altitudes were obtained of points representative of the entire shoreline. To accomplish this a large area of country had to be traversed, and it was -ifcessary to employ all available means and methods for the collection of tlie data. All heights are referred for comparison to a common datum l»()iiit, arbitrarily cliosen, the zero mark of the lake gauge at the Lake Shore bathing resort. The measurements and observations here brought togetlier are not my own alone, but were made by many persons and at various times. In the following pages the attempt is made to ari'auge them in such order that the critical reader can readily learn the essential nature of all the data on which each separate determination of altitude is based. SCHEME OP TABLES. Taule XVIII. Differences of aUitiulo determined by trigonometric oljservations. XIX. Dififereuces of altitude determined Ijy Ij.arometric oliservations. XX. Reduction of various lake gange zeros to tlie Lake Shore datum. XXI. Gauge records, showing the height of the water surface of Great Salt Lake at various dates. XXII. Ditferences of altitiule from railroad survey records. XXIII. DitFereiJCes of altitude by special spirit-level determinations. XXIV. Reduction of results to Lake .Shore gauge zero as a common datum. XXV. Comparative schedule of altitudes of points on the Bonneville shoreline XXVI. Comparative scliednle of altitudes of j>ointa on the Provo shore-line. XXVII. Comparative schedule of altitudes of points on the Stansbury shore-line. XXVIII. Ditferences in altitude of the Bonneville aud Provo shore-lines at various localities. XXIX. Differences in altitude of the Provo and Stansbury shore-lines at various localities. By reference to this scheme of tables it will be seen that hypsometric material 403 has been gathered from the five following sources 406 LAKE BONNEVILLE. (1) From deterrainatious based upon trigonometric observations. (2) From determinations based upon concurrent barometric observations. (3) From the records of the fluctuations of the present Great Salt Lake. (4) FroQi the records of various railroad surveys. (o) From especial determinations made with the surveyor's spirit-level. TRIGONOMETRIC DATA. The few results obtained by the first method and jucscnted in Table XVIII were derived by comi)utation from measurements of angles of elevation and depression with accompanyiuf;- short base-lines. Tiie angles were measured with the ordinary surveyor's (ransit, reading to minutes on the vertical limb. The base-lines were measured with a steel tape. The results are recorded in feet and tenths of feet, but it is not intended to assert that they are true to the nearest tenth. They are probably true to the nearest foot. In combining determinatiims of various kinds it has been found convenient to use the same notation for all, and the tenth of a foot has been <-hosen as expressing the pre- cision of the most accurate of all the measurements — thc^ shorter lines of spirit-level- ing. For the purposes of the Bonneville investigation it would be sufiicient to stop at the decimal point, as all the results of measurement are combined with observa- tions involving an uncertainty of several feet; i)nt it is conceived that some of the data may have other uses, and for the sake of these the tenths are retained. Table XVIII. — Differences of Altitude deiermiind hij Triijntinmetrie Olmerrntion-i. Vicinity of- Feet. 415.1 301.1 410. 7 310.0 3G3.0 Dovo Ureek — Kelton Bonneville abore-line above Prove sbore-line ,io Mallin do . . . . Matlin Snowavillo BAROMETRIC DATA. The section of country including the long southern arm of the old bike, now the Escalante Valley, was practically accessible to no better hyiisometiic. method than that of concurrent barometric observation, and that method was accordingly adojited for its investigation. This region in general lies two huntlred miles south of Salt Lake City, and its nearest barometric base was the U. S. Signal Ofilec in that city. It was deemed advisal)le to establish an intermediate sub-base station in the nearer neighborhood of the field of itinerary ob.servation, to which to refer the new stations. The village ol Fillmore, lying one hundred miles sonth of Salt Lake City, ottered especial natural advantages for the location of snch a sub-base. It includes within its limits a jiortion of the Bonneville siiore-line, thus allowing but slight disitaiity in altitude iietwet-n the reference station and now stations. It is moreover situated about midway between the southern field of study and the Salt Lake City i)rimary base, and affords, by the comparison of its series of observations with that of the Signal Oflice. a criterion for judging of the value of results from the observations at the new stations. HEIGHTSBY BAROMETER. 407 Two barometers and psycbrometers were left here in the charge of an observer, Mr. R. II. Smith, from July 29th to October 3ril, 1881. U])on the former of these, hourly observations were made each day from 7 A. M. to 9 V. M. inclusive ; ui)0m the latter, readings were taken daily at 7 A. M., 2 P. M. and 9 P. M. The Survey did not establish a base station at Salt Lake City, but made use of the ordinary observations by the U. S. Signal Service observer. Through the cour- tesy of the Chief Signal Oflicer of the Army we were furnished with copies of sucii ])ortions of the records as were needed for our work, viz., the readings of barometer, thernioineter and psychronieter at 7 A. M., 2 P. M. and 9 P. M., during the period covered by the observations at Fillmore. The altitude of the sub-base above the Signal Ofth^e at Salt Lake City was coni- l)uted from a selected portion of the concurrent oi)servations at the two plaiies. In order to avoid observations affected by abnormal atmosi)heric conditions, the "reduced" barometric readings at the two stations were platted gra|)hically in close proximity, with a common time scale. A marked parallelism of the resulting curves between the dates of July 29tli and August 17th led to the acceptance of the recoids included between those dates as a basis for the computation, and they alone weie employed. Three somewhat independent results were obtained for tlie difference of altitude by considering sei)arately the means of the 7 A. M., 2 P. M., and the 9 P. M. reduced readings at the two stations, Williamson's method and tables being employed.* In each determination the terms t+t' and a + a' are identical, being derived from the means of tlie temperature and humidity terms of the 7 A. M., 2 P. M., and 9 P. M. records for the selected period. 7 a. HI. 2 p. m. 9 p.m. k. Me.in of rrdiicol readings :it Salt Lako City II. Mn:in of reduced rcadinjis at Fillmoro Inr.hrtt. 2.'i. (ifi8 24.947 Fret. 24, 720. 9 23, 973. 8 Tvekea. 2.) 041 21. 91)7 Fret. 24, 693. 2 21,931.7 Inehes. 2.-). 013 24.81)8 Fret. 24.004.6 23, 922. 3 t+t' from means of 7 a. m., 2 p. m., and !) p. ni. tem- perature reading.s at Salt Lake City and Fillmore ~ 1510.92 F. a + a' fi-om moans of 7 a. in., 2 p. m . ami 9 p. m. rela- tive humidity reiluctions for Salt Lake City and Fill- more = 0.60. From Table l>i with argument h First approximate difl'ercDce of altitude From Table Dii 747.1 +66. 80 761. 5 +08. 15 742.3 +06.44 813. 90 -h6.21 829. 05 +0.32 808. 74 -1-6.14 Tiibles Dili to D,ii, inclusive, with general arsumrnts ( + (' = 151°.92 F., a +a'= 0.60, lat.= 40°, and seoond ajiproxi- mate dilTerence of altitude, give additional correction . . . 820. 17 835. 97 814. H8 Accepted result, mean of tlie three determinations, 823.7 feet. 'Professional papers of the Corps of Engiuccrs, U. S. Army. No. 15, Appeudix. 408 LAKE BONNEVILLE. To the Fillmore station alone, as a base, have been referred all the itinerary barometric records taken in tlie district south of it. At the new stations no *'dry bulb" thermometer or psychroraeter readiufj^s were talvon, and wliere such diita were necessary in the comi)utation of their altitu*hoie-line, 1 mile west of Spring, altove Fillmore anb- ba.se harOTueter. Sub-ba.se cistern barometer rtbrice H. S. Signal Office barometer at Salt Lake City, Utah. Bonneville shore-line above Provo shore line Bonneville .shoreline on Kannsh Bnttf hil»w Fillmore anb-base barometer. Bonneville .shoreline, 1 mile east of tiitrance to canyon, above Fillmore sub-base barometer. Bonneville shore-line, 1 mile west of entr.ince to canyon, above Fillmore sub-base barometer. Camp on east bank Beaver River, hehnv Fillmore sub base batom- etcr. Bonneville aliore-line, 1 mile northwest,of Alilfonl, below I-'illmore anb-haae barometer. Bonneville sbore-line, 7 miles south of Milfonl. below Fillmore sub-base barometer. Bimneville ahore-line, east base of Peak, aboiw Fillmore sub base barometer. Bonneville ahorc-lino, ca.st baseof Butte, below Fillmore sub-base barometer. Bonneville sliore-line. west of entrance to canyon, above Fillmore Mub-ba.se harometer. Bonneville shctre-line, north of entrance to canyon, above Fillmore riuh bxso barometer. Bonneville almre-line, west base of Peak, below Fillmore sub base barometer. Bonneville shore-line above Fillmore suli-hase harometer DitlVrence in Altituile. Bonneville shore-line, 2 miles east of Springs, behnv Filhiioie sub- base baronuiter. Bonneville Nhore-liue, 4 miles south of Springs, below Fillmore sub-base barometer. Bonuevilh> shore line. 7 miles south of Si)rings, below Fillmore aub-base liarometer. Camp, belo7t> Fillmore suhbuae barometer Feet. 3fl. 7 ft2:t.7 381.3 17.3 296. G 294.5 21P.6 7fi 0 48. J> L2 67.6 214.1 26.-.. ri .12,5 45.2 76.6 48.7 42.1 474.1 .HEIGHTS B-Y BAKOMETEK. 409 LAKE RECORDS. At various times spirit lovol lines Iiare been run from the surface of Great Salt Lake to points on the ancient beaches in tlie near nei^liborbood of its present shore. The records of altitudes thus obtained are not, however, directly comparable, suice the surface of the lake is in a state of continual fluctuation, the records of which have lieen referred to independent jjauges. It was accordingly nece.'sary to determine pii- marily the relative altitudes of the zeros of the various gauges. Previous to 1875 the record of the rise and fall of the lake is iturely a tradition;il one. Such evidence, however, as is reliable, has been presented by Mr. Gilbert in his chapter on " Water Supply," Powell's " Lands of the Arid Region," in which the rec- ords have been referred to the level of the Antelope Island Bar as a datum.' In 1875 a granite inonuineiit, graduated to feet anns were made at intervals on the newly erected giuige until Octo- ber, 1879, when it was rendered useless by the occurrence of a succession of hea\y winds from the westward, which effectually barred the entrance of the inlet witii sand, thus cutting ofl' its direct cominunicatiou with the lake. In anticipation of such an occurrence, a third gnnge had been established at Lake Shore, live miles soiitii of Farmiugton, and monthly records begun November 19th, 1879. This is known as the Lake Sliore Gauge, and to its zero as a datum have been referred the various deter- minations of which tiiis ai)pendix treats. (See Table XXIV.) A general taliing tendency of tlie Lake for several years portended disqualifica- tion of this gauge, and rendered the erection of a deeper set scale a matter of pieciin tionary advisability. A fourth gauge was accordingly established at Garfield Land- ing, three miles west of Black Kock. It consists of a stout strip of scantling, nine- teen feet long, firmly si>iked to one of the piles of the steamer pier. It is graduated to feet and iiicbes.^ On the 23d of -luly, ISSl, the Black Rock bench was found by spirit level to be 38.7 feet above the surface of the lake; at the same time the water washed the 7 ft. 9 in. mark of the Garfield gauge. Thus the zero of the latter is 4G.4 feet below the Black Rock bench. Table XX indicates the steps by which the various gauges have been reduced to the Lake Shore zero. 'Ttie traditional record is repeated, with an addition, in this volume, pp. 2:W-243. G. K. G. ^Since tlie preparation of this Appendix, the Garfield gauge has been destroyed and renewed Oee p. 232. G. K. G. 410 LAKE BONNEVILLE. Table XX. — Reduction of various Lake Gauge Zeros to the Lake Shore Datum. Point. Intermediate Datum. Uato. Refi^ned to Intermediate Datnm. Referred to Lake Slioro Gauge Zero. Jan. 2:1, 1880 ... + 2.r< + 3.8 A "Temporary Bpnrh'' ,it Farmington. Lako Surface do ... + 1.3 Farmington Gauge Zero A "Temporary Bencli " at Farmington. Nov. .■!, 1879 . . - 0.1 + 3.7 Farmington Gauge Zero do +12. n fie.G + 2.0 M!»r. 21 lo Mot. 25, 1881. Qarfield Landing Gauge Zero Mean Lake Surface .. do - 7.2 - 4.6 Lake Surface G.arfifld Landin" Gange Zero July 23, 1881 . . . + 7.7 + 3.1 Black Hack Bench do +38.7 -34.5 - 2.0 + .8 +41.8 + 7.3 + 5.3 + 0.1 Lake Surface Black Rock Monument Zero.. Lake Surface IJlaclt liock Bencli July 12, 1877 .. . do Lake Surface Black Rock Monument Zero Oct. 10, 1877 . . Antelope Island Bar in tlio Lako Surface ....do - 0.5 - 3.4 "little cbannel." A uote of uncertainty relative to tlie results depenilent on the Black Bock ob- servation of July 12, 1877, must be introduced here. The observer's record of that observation reads as follows: Jut)/ 12, 1877. — Water washed higbest font mark of graduation on Dr. Park's [Black Rock] iiion- iinient ; supposed to bo the two-foot mark. The scale is neither numbered nor lettered, but subsequent conversation witli Dr. Park led to the acceptance of the record in conformity with the su[)position of the observer. Confirmatory evidence is found in the close agreement of this determination of the monument zero with a second determination, which joins t)ie monument zero to the Farmington zero by reference to the lake surface. The difference in the two results is less than two-tenths of a foot. As an interval of fifteen hours elapsed between the readings of the two gauges the second determination was considered only as a general check for large errors, and was not used in the reduction. A table and platted curve showing the rise and fall of the present (Jreat Salt Lake from September, 1875, to June, 1889, will be found on pages 233-243 of the mono- graph of which these pages form an appendix. By means of the data contained in that table the lines of leveling at various times connected with the water surface of the lake were referred to the Lake Shore gauge zero. The specific data thus used are here repeated in Table XXL HEIGHTS OF LAKE SURFACE. 411 Table XXI. — Gauge Records, showing the height of the Water Surface of Great Salt Lal;e at various dates. Gauge. Date. Ri'adinji. Hciulit ..f Trauiio Zoro aliovn Zi'io of Lake Shore Gauge. Height of Water Sui fare ahove Zi'io of Lake Shore Gauge. Rlack Rock Do Fannington LaUe Shore Do July 12,1877.. Oct. 19. 1877 . . May 2, 1879 .. Nov. 9. I87!>... Nov. 12, 1880.. Nov. 29, 1880 . . Dec. 11, 1880... Ft. In. 2 0 0 10 1 4 2 G 1 9 1 8J 1 8J Feet. .'i. n .1.7 IF 0 0 0 Feel. 7.3 0.1 .I.O 2.5 1.7 1.7 1.7 Do. .. , Do RAILROAD RECORDS. A fourth .source from wliicli data liave l>i>cii obtained to assist in the general comi)i!ation, is Found in the records of various railroad surveys. Tiie results appear- inj; in Table XXII in some eases have been derived from Gannett's " Lists of Eleva- tions", 1877, and such are indicated by a star (*); iii other cases they are from tran- scripts of official profiles kindly furnished by the engineers of the dittereut roads. TAni.n XXII. — Differences of Attilude derived from Railroad Surrey Records. Vicinity of Poiuts Dutorinincd anil Poiuts of Kcferonco. Feet. Corinne Station (Central Paciflc R. R.) below Ogilen St.ation Franklin Station ( Utah Nortliern R. R.) alioite 0;;.m. 0 HIack liock Itench above Black Rock Monument zero 34. B Provo shore-line ofiorc lake aurfacc, July 12, 1877 033. U 247.0 .Stauftbury shore-lini' above lake surface, July 12, 1877 Corinne Station (C. P. 11. It.) above lake snrfac o. May, 1873, (Wliechr 22.6 Cup Butte l-'illiiiore IJonnoville ahoro-lino above Provo sliort>-line 397.01' 19.4' 309. 01 Finli Spring Uunne\ille cut- terrace above Prove cut-terrace Bonneville shore-line on Franklin liutte afcorc l''raiiKliii Station (U. N i:. R) KG.O Provo .shoreline on Franklin Bntte above Franklin Stariou (U.N. R.K.) 201. 0 Kelton Bonneville .shore lino above lake .surface, Aug. 11, 1877 (checked by 1017.5 ,52.4' r.akc Shore GauKe zero below Salt Lake City Station, tl. S. R. R Bonneville shore-Iiuo above Lemington Station (tJ. S. R. R. extension) 380. C< Logiin Bonneville shore-line above Lo^an .Station (U. N. R. R.) 632.9 Provo shore-line rttoV(? Lo*;an Station (f.N.R.R.) 270.2 Milford Bonneville shore-line above Millnnl Station (U. S. K. R. extension) - 152. 7 • Camp on e.ast hank Beaver River Womi Milfor.l Station (U.S. R.R. 0>;il<-n 7.0 Bonneville shore-line above Ogden Station U. C. R. 11., (Prof. F. H. Bradley, Ilavden Survey). -. 87G. Of 329. 1* 358.0 Pavant Batte Point of the Mountain Bonneville shore-line above Summit (U. S. R. R.) Picusa Valley Provo shore-line 6eioic Bonneville shore line 375.5 343.2* 'Ncuth Group," Bonneville .shore-line above Pi ovo shoreline " Middle Group," Bonneville shore-lino above Provo shore-line 346.4* Promontory Red Rock Pass Bonneville shore-line a&ore lake auiface, Aug. 23, 1877 1037.7 «B5. 8 303. U Provo shore-line a&oi'c lake surfu'o, Aui^. 23, 1877 .. , Bonneville shoreline above Swan Lake Station (U. N. R. R.) Salt Lake City Bonneville shore. line above Salt Lake City, Meridian M'lnnment ... 845.9* .Meridian Monument below U. S. Signal Si-rvice barometer 12. 6* SaU Lake City St.ation |U. S. R. R.) Mow Meridian Monument .... 72.0* S.alt Lake City Station {U.S. R.R.) above Lake surface, Dec. 11, 1881. 50.7* Santaqain Bonneville shore-lino above Sautaqnin Summit (U. S. \l. R ) 75.0 Bonneville fthoro-line «6f*«f Provo shore-line 401.0* 1011.0 Stockton Bonneville shoreline above lake surface, Mar., 1873 (M. F. Burgess) Prove shore-line (j/'iow Bonneville shore-line 375,0 Bonneville shore-line above Teronia Station (C. P. R. R.) Bonneville shore-line aborc Provo shore-line 367.8 Wellsrille AVhite Mountain (Fill- 383.7 Provo shore-line on White Mountain above While Monnlnin more.) camp 68.9 Provo tufa deposits on Tabernacle Bntte Uva bed above camp 42.9 Willard Bonneville shoreline above lake surface Oct 28 1879 974. Ot 621. Of Provo shore-line above lake surface, Oct. 28. 1879. UEIGUTS BY LEVELING. 413 COMBINATION OF DATA. lu the schedule followiug (Table XXIV) a collection and combiuation is made of results appeariuj? in some of the six tables preceding, so as to reduce tlie stations to which they apply to the arbitrarily assumed Lake Shore zero datum. Tiie table is arranged with reference to the latitudes of the points determined, beginning with the most northerly. Table X'K.IV. —Reditclion of Rumiltn to lite Lake Shore Gauge Zero as a Common Daliim. Point. L.^ke Surfaci-, IK-c. 1 1, 1880 Salt LakeCily Sl.ilioii (U.S. K. It.) Ogdell Statiou Fraiikliu Station ilT. N. K. R.) Swan Laks Station (D.N. Iv. K.) Bonneville aUore-lino, vicinity of Rod Itock Paaa Franklin Station HonneviUe sbore-linf* ini Franklin Butto - Trovo 8bore-line on Franklin Butto OK'len Station Lo2!in StJltion (U. N. E. R.) Bonneville .shore-line, vicitiity oT Logan . Pi-ovo .shoi-e-line, vicinity of Logan Lake aui'facp, Ang. 11, 1877; interpolated Bonneville shore line, vicinity of Kelt." In ascertaining the precision of the bar(unetric work, use was niaile of the long series of simultaneous observations at Fillmore and Salt Lake City. Sixty indei)end- ent comi)ntations were made of the difference in altitude of the two stations, each computation being based on a single set of concurrent observations. A computation based on the discrrepancies of the sixty results showed the luoliable error of a single determination to be IrliS feet. The errors assigned to the barometric determinations were estimated on this basis, allowance being made for distance and other special con- ditions. A part of the leveling work was dui)licated, and an examination of the records of such duplicated work led to the belief that, as executed by u.s, a line of levels not exceeding five miles in length nor 1000 feet in vertical range, need not be assigned a 'A (lisi^Mssiou of this siilijiTt will ho fouiiil in Cliaiilcr III of tliis volume, iinJer the ht'a<]iii<;s " EmbaukiniMit Series" aud " Doteniiiiiatiou of Still-water Level," i)p. Ul-l'i'). G. K. G. HEIGHTS OF SOORE-LIXES. 417 greater probable error than one foot. Locke's liaml level, when supported by a staff and used on a steep hillside, was found to have a probable error of about one foot in SOO feet of ascent. The probable errors recorded in the following tables were obtained by combining the estimated probable errors of measurement with the estimated probable errors of identification of the plane of tlie ancient water surface. It is recognized that any individual determination, not duplicated, msiy involve some gross error for which no allowance is made, but if such errors exist their number is small. Tablk. XW.—Coniptiralive Schedule of Jltit tides of Poiiils on the Boiinerille Shoreline. Locality. Description of Determined Point. Di'terniint'd Alliliulii almve Iho Lake Shore Gau;lo ZiTo. Iiiferreil hiyli-water level, above Lake Shore Gauge Zero. Feet. Feet. Red Kock Pasa Inner edge of a cut-terrnco 897.7 906 ± 4 .... do 933.7 933. 0 940+ 3 942+ 4 .... do Kelton Butte Crest of an enibnukmont 1024. 3 1019+ 3 1014.2 1050+ 3 AVniard. do 976. G 985+ 3 971.1 981 ± 5 970.7 970.9 980 ± 5 979+ 5 Salt Lake Citv Dous^laa. By first determination. S-ilt, LaVe Citv Inner edge of a ctit-tprrace back of Fort Douglas. By .second determination. 970.2 984+ 5 I'Uick Kock ■ TiHier fdge of a cnt-toTrncu 1000. 3 1008+ 3 North ci.d of Aijiii Uauj^e.. Inner edgeof cut terrace. By first determi nation. 1000. 1 1068 i 3 Nintli end of Aqiii UaiiL^o. . Innei" edge of a «ut-ti*rr;n-n. By second ilt^- tonnination. lOOG. 0 1074+ 4 I'oiutof the Mmiiilain . .. 957. 4 1019.0 899.4 894. 0 911. S 893. G 943.9 902. 4 9)0+ 3 1014+ 5 902+ 3 902+ 5 938+ 8 902 i 15 953 + 15 971+20 do do Pavaiit Butte Near outer edge of a cut-terrace Middle of a cut-torrace , Outer edge of a out-terrace B;ise of North Twin Peak. . Base of South Twin Peak.. .. do. 928. 7 939+20 Milford ... End of a V-embanknient. The elevation given in the third columu is the general 900. 3 904 + 10 mean of the three determinations of the point given in Table XXIV, weighting the first at 5, the second at ■'J. and the third at 1. 7 miles aouth of Milford ... Outer edge of a cnt-terraco 912. 3 921 + ''O 2 miles east of Thermo=» Middle of a narrow cut-terrace 884.6 893+25 4 miles aouth of Tliernibs . 7 miles south of Thermoa Middle of a cut-terrace 912.5 919.1 921 ±25 927+25 do MON I- -27 418 LAKE BONNEVILLE. Tablk XXV.— Comparative Schedule of Altitudes of Points on the Bonn.eville Shore-line — Contiuued. Locality. Description of Determined Point. Det<-ruiincd Altitude aliove the Lake Shore ft, , / 0 V^'os 2> — cos a I ^ ' ' wherein a is the angular distancie of anj- ])oint of the disturbed surface frotn the axis of the disturbing mass, a and v are thus polar coordinates of the disturbed sea surface. The effect of the rearranged sea-water, ignored above, is simply to produce an exaggeration of the type of surface defined by (2), ami this exaggeration may be expressed by a series of rapidly converging terms (see §§ 20-24 of paper on Form and 'If Hi be an element of the (U.stiul)iiig mass aiiil r its dislaiice tVoiu the point in question, thi; potential of the mass is the sum of all I lie iinotieuts '", or V ^ 3 '" . The non mathematical reader )• r shonld distinguish carefully between potential and attraction, the latter bcini; a dirivative of the former. DEFORMATION OF GEOID. 423 Position of the Sea Level), but for tbo small masses bere considered tbe snm of these additional terms is iiisigiiiflcant. In all cases, indeed, tbe characteristic effects are expressed by equation (13). For lenticular masses of tbe type assumed iu the text, the thickness is given by the expression nn \ (4) If we represent the values of the definite integral in tiiis equation for points along tbe border and at the center of the mass by S2 and Si respectively and denote tbe corresponding values of v by V2 and »i respectively, we find ih hop { S2 - S, } (5) This expresses tbe difference in altitude of tbe disturbed surface at tbe center and at tbe border of the disturbing mass. When, as in the present case, tbe disturbing mass is water in a lake basin, we must substitute for /j tbe difference in density of water and superficial rocks. That is, p = 1 — 2.8 = — 1.8 approximately. Finally, if we wish to ascenain tbe seiiaration at tbe center of tbe basin, due to a change in the density of its contents, of eipiipotential surfaces which intersect along 424 LAKE BONNEVILLE. the border, we have only to diflerentiate (5) refjarding (i'2— »i) and p as variables and substitute for Zip the change in density of Mie contents of the basin. Thus, the sepa- ration is expressed by Tlie vaUies of S| and S2 in (•^) and (G) may lie found from the following oxpres- n b, = ^^ _^ J n sin ^/i„. S. = ^ sin hft„ \ ., (~_^j + jy (,^ ^^^ (1 + sin^ }Ji,) + \ Rions The march of the above functions Si and Sj and the corresponding values of (v., — ?'i) and J (i'2 — lu) is illus-trated by the numerical results given in the table below. The data for these results are the foUowiu" : 1000 feet, /y„ = arc of 1", r, r. IJ =~ LS', zJp = -I. The results in the fifth column show how niucli nearer to the center of the earth the assumed lake surface is at the middle of the basin than at its border; and the results in the sixth column show how much ashore trace at tlie middle of the basin would be found to be above the contemporaneous trace at the border, by a line of spirit levels run after the removal of the water. TaBlk XXX. — Valncs skoiriiiy relative posUlnna of Lend Surfaces in a lake hasiii 140 miles in diameter and of 1000 feet maximnm (axial) depth. n s, Sj s,-s, 1), — t), ^(i>2 — »,) FMt. Feet. 1 n. 00436 O.OOICI 0. 00^75 2.70 1. .-,0 2 . 0D.)82 .00:.'45 . 00337 3.31 1.84 3 . 00034 . 002DS . OOJ50 3. 50 1.94 4 . OU0U.S . 00333 . 00365 3. 58 1 99 0 . 011727 . 00359 . 00368 3.01 2.01 G . 00748 . 00379 . 0036!) 3.62 2. 01 7 . 00761 . 00395 . 00160 3.62 2.01 8 . 00776 .00407 . 00369 3.02 2 01 9 .007^(0 .00418 . 003(18 3. Gl 2.01 10 .00793 . U0427 . 00366 3. .59 1.69 00 . 00(173 . 0U5S6 . 00317 3.11 1.73 ATPENDIX C. ON THE ELEVATION OF THE SUEFAOE OF THE BONNEVILLE BASIN BY EXl'ANSION DCE TO CHANGE OF CLIMATE. By E. S. Woodward. The folIowiDg problems were submittey interference, tbe total eiil)ic expansion was expressed in vertical dilatation. How many leet was tbe surface of the groniul lifted ? (2) Same as above for period of 100,000 years. (3) Same as above for period of l,00O,00u years. The cooling by coiuliiotion of a large s[)liere like the earth from an initial uiii- (brm temperature, gives rise to cubictil contraction whose amount is assigned approxi- mately by the following formula:' V 71 zlV = Syrrhiea ^1 - in which r = the radius of the sphere, u z= the initial unitbrin e.Kces.s in temperature of the sphere over that of the sur- rounding medium, a^ = the coelidcient of diffusion, assumed constant for the whole sphere, e = tbe coetiticient of cubical contraction, assumed constant, t = the titne after tlie initial epoch, ;r = 3.1415+. This formula will apply to the earth for l,0()(),()O0,()0l» years subsequent to the initial epoch witliout iiitioducing errors greater than those involved in the assump- tion of constancy of a and e. Conversely, the above formula will give the cubical expansion of a sphere, con sequent upon being immer.sed in a medium which maintains a constant surface tem- perature u degrees higher than tlie initial temperature of the sphere. ' Forcorapleto formula see Aunals of Mathematics Vol. Ill, No. 5. 425 426 LAKE BONNEVILLE. If in the latter case wc suppose tbe total volumetric expansion to result in ver- tical uplift, an eflcct wliicli would follow from heating the earth's crust if it behaved under expansion like a liquid, the amount of the uplift will be expressed very closely by the quotient of equation (1) divided by the area of the surface of the sphere. Thus, calling the amount of the uplift z/r, we have STir^uea Ar = n inr^ = 2uea il- (2) tJsing the year and the British foot as units, Sir W. Thomson finds a = 20. With this value and with u = 10° F. and c = 0.000018, (2) becomes Foot. zJr=z 0.00406 V 7. This gives the following values of Ar corresponding to several values of <: 1. ^r. Tears. 10, 000 100, 000 1,000,000 Feet. 0.41 1.28 4.00 INDEX. Page. Aa at Ice Spring 323 Adams, J., lake rampirts 71 Adolescent coast lines. *53 Airy, G. B., tbeory of waves 26,29 Alg» 259 Allen, O. r>., analyses of Bonneville earths 200 analyses of Sevier Lake desiccation products. . . 226 analysis of water of Great Salt Lake 253 Alluvial cone and fault scarps, view 349 Alluvial cones, Bonneville Basm 91 Frisco Kango 92 Marsh Creek ■- HB Lake Creek 185 aridity and 220 Alluvial fans 81 Alluvial terraces and fault scarps, Kock Canyon 344 American Fork 346 near Salt Lake City 349 East Canyon 352 Alluvial cone terrace 81 Altitudes and their determination 405 Altitudes of shore-lines 362,427 American Fork, deltas 155, 346 fault scarps 346 Analyses, tufa 168 White Marl and Yellow Clay 201 waters of City Creek, BearKiver, and L'tah Lake 207 Sevier Lake briue and desiccation products 226 water of Great Salt Lake 252 waters of Bear River and Utah Lake 254 Andrews. Edmund, theory of littoral transportation . 26, 41 subaqueous rid;;e3 44 Antelope I-land bar 240, 243, 410 Appalachian Valley 391 Arpii Range, fault structure 341 fault scarps 352 heights of shoredines 366,370,372 measurement of shore lines 412, 414, 417, 418, 419 Area, Great Basin 5 Bonneville Basin 20 Lake Bonneville at highest stage 105 Lake Bonneville at Prove stage 134 Sevier Lake 2i5 Areas, interior basins of Arizona, New Mexico, and Texas li various lakes 106 Great Salt Lake 243,244 Aridity and alluvial cones 220 Aridity of Great Basin, described 6 cause 10, 280 Page. Arizona, interior basins 11 Pleistocene eruptions 337 earthquake 361 Arno Valley, Pliccene fauna 399,400 Arrow point, fossil 303 Artemia gracilis 258 Barometric measurement of shore lines 363,406 Barometric measurements, probable errors 416 Barrier, described 40 compared with other ridges 87 Bairy,W.C., tbeory of salt harvest 224 Bars. (See alao Bay bars) 48 Basalt Valley 219 Basaltic eruptions, Bonneville Basin 319, 325, 338 map - 334 Basin Ranges, type of at ructuro 5 of the Bonneville Basin 91 Basins, hydrogr-iphic 2 interior, of Aiizona, New Mexico, and Texas... 11 of the Bonneville Basin 122,222 Basaett, H., analysis of water of Great Salt Lake 251, 253 Bay bars, origin and character 48 Snake Valley 111,112 Tooele Valley , 131,132 Beach, origin 39 profile. 39,42,45 Bear River, deposits in Cache Valley. 163 gate of ITS possible changes 218, 263 irrigation 250 Bear River water, precipitation experiments 206 analysis 207,254 Beaumont, Elie de, shore topography 26 limitation of tidal action 29 variation of beach profile 42 Beaver, fossil 303,394,400 Beaver Creek delta 166 Becker, G. F., cited 284 Beckwith, E. G., cited 14 Bellville Creek delta 162 Bench-mark at Black Rock, installation 231, 409 leveling 232 height 233,410 map -. 390 Benchmark at Farmington 409 Biinadou, J. B 18 Big Cottonwood Creek delta 165 Big Willow Creek moraines 309 Bipartition of lacustrine and glacial epochs 270 427 428 INDEX. Pago. Birds, fossil 303,304 nisoii, fossil 211 lilmk Kock, view of laku terraces i lieiRht of shore-lines 365,370,372 niea!*iiieiuenl of shorc-linoa 412,414, 417,418,419 Hlack Kot;k beucb, iustallation 231,409 Unchnj; 232 hoight 233,410 map 3!to Black Rock gauge, iuslallaiion 23 1 , 409 leveling 3:t2 Iifight 233.410 noord 233 lUuckfoot River, possible changes 219,203 lUacksnjitb Fork, superpoailion of embauktueuts ... 151 dt'ltas 162 lUake, William P., cited 15 Illoody Cauyon nioraiiios 313, 315 Bouiioville Basin, description 20 ninp of subdivisions 122 bistory 214,316 subdivisions 222 possible changes 262 Bonnevilb' beds (see, also, White Marl and Tello^v Clitii) 188 llonnox ille, B. L. E., explorations 12 Buuuevillo fossils 209 Bonne vilbi Lake, outline at highest stage 101 area and depth 105 depth 125 authorities for map 125 outline at Provo stage 127,128 composition of water 204 large map (in pocket of cover.) Bonneville sliore-Une, highest 91, 94, 97 general description 93 clilfs and terraces 107 V-embauknionts , 108 spits and loops 108 deltas. 109.153 embankment series Ill uncertainty of still-water level 125 near outlet 174 on Pavant Butte 326,328 in Esc.ilanio Desert 362 deformation 365 height at various points 365 curves of equal height 3G8 synchronism .. 369 (See, also, Intermediate Hhore-lincs, Provo shore- line^ and Stanshury shoreline.) Box Elder Creek deltas 163 Braddock's Bay 50,63 Bradley, Frank H., observations on Lake Bonneville. 10 cited on ancient delta of Ogdcn River 93 cited on terraces in Marsh Valley 95 cited on highest shoreline 96 cited on dt-ltas 153 cited on i)utlet of Lake Bonneville 173 leveling at Ogden 412 Branchinecta 259 Brewer, W. H., cited 206 Brigham City, deltas near 163 Brine of Great Salt Lake 251 Bi ine of Se vicr Lake 226 Page. Brine shrimp 254 Brodie, James, cited 270 Uriickner, Eduard, cit*'d 271 Burgess, M. F., leveling data 412 C.icho Valley, terraces 95,96 Tertiary lake beds 99 Bonneville Bay .' 102,178 del taa 1 59, 1 62 fault scarps 351 Call, R. EUswctrth, recent and fossil shells of Great Basin 19,297 Bonnrville shells 210 Campbell, J. F., cited 270 Cedar Range 103,128 Cbadbtmrne, P. A . , cited 211 CIiamberlin,T. C, cited 272 Cbatard,T.M., analysis 207 Christmas Lake fossils 303, 394 Ciiurcb Lake 300 Cialdi, Alessandio, coast processes 26 tlieory of littoral transportation 41 City Creek deltas 164 City Creek water, precipitation experiments 206 analysis 207 Clarke, F.W., analyses 207 Clarkstor, fnult scarps 351 Clayton, .I.E., cited 348 CliiTs, formation by waves 34 classification 75 comparison 75, 77 Climate and interior basins 3 Climate and moraines 398 Climate curves 246 Climate of (ireat Basin 6 Climate of lake ejioeh, as inferred from fossil shells. 297 as inferred fritm fossil bones 303 as iiifeired from moraines 305 Climatic factors affecting lakes and glaciers 275 Climatic interpretation of lake oscillations. 265 Cloud-burst channels. 9 Coast lines, local phases 60 adolescent and mature. 63 simplification 63 of rising and sinking land 72 Cold, <-orrelation with buMiidity. 265 corrt'lation with biiihlingof houks 52 Curve of precipitation change for Great Basin 245,249 Curve of rise and fall of Great Salt Lake, annual 239 non-pe riodic 243, 240 Curveof secularclimalic change in BonnevilleBa-in. 2G2 Curve of temperature change for Great Basin 240 Curves of etiual height, Bonneville shore-lino 3C8 Provo shoreline 372 Curves of snow-fall and melting 289, 293 Curves, theoretic, of post-Bonneville deformation 374 Cnt-and-built terrace 30,40 Cut-terraces, mode of furmation 35 of Bonneville shore-line 107 of Prove shore-line 127,128 of Intermediate shore-lines 144 Cypris 210 Dana, Edward S., bulletin by 19 Darwin, G. H., cited 387 Datum fur gauges, map 390 Datum jioints connected with gauging of Great Salt Lake : 233,409 Davidson, George, cited 10 D ii V i s, W . M . . e i ted 1 80 Dawk ins, W. B., Pleistocene mammals . 400 Dead Sea history and glacial liisLory 2G5 Death Valley 8 Deep C reisk Kange, faults 353 Deer, fossil 211,303,394 Deformation, crustal, by loading and unloadini: 3.'i7,379 of Bimnoville shore-line 'iG\ 308 of Prnvo slnire-lino 371, 372 duiing Prcivo epoch. 372 ipiestion of cause 373 curves of theoretic 374 nf geoid 421 of Bonneville Basin by expansion 42r> Degradation clift" compared with other cliffs 75, 77 Degradation terrace, compared with other terraces 78, 81 De la Bcche, Henry T., writings on shore topography. 20 variation of beach profile 42 Delta terrace, compared with other terraces 84 Deltas, origin 65 internal structure 69, 70 of emergent coasts 74 of Ogden River 93 of Bonneville shore-lino 109 Pago. Deltas, Provo shore-line 129 of Lake Bonneville 153 history deduced from 166 of Spanish Vork 343 of Weber River 349 Depauperation of lossil sheiks 299 De|>osition, littoral 40 Deposition of salts by desiccation, Bonneville Basin 204.208,258 Rush Lake 2_'9 Depth of Lake Bonneville 125 Di-pths of lakes, table 100 Desiccation, deposition by. See Deposil'ion. Desoi", E., limitation of tidal action 29 cited on subaqueous ridges 43 Diastrophism, defined 3 and interiiir basins 304 and Lake Bonneville 340 of Jordan and Tooele valleys 307 Diatoms 210 Differential degradation cliff, compared with other cliffs 75,77 Differential degradation terrace, compared with other terraces 78, 84 Discrimination of shore features. 74 I>i.-*placement. See I>iastfophism and Deformation. DistiiUulion of basalt, map 334 Distribution of fault scarps, map 352 Distribution of wave-wrought shore features 60 Divides, shifting of 217 Douris, T., gauge readings 235 Dove Creek, sea-cliff near ]07 Bonneville embankment series 112, 114, 117, 120 Provo ombankment series 131 Intermediate embankments 137 map and view ]38 embankment interval 143 superposition of emhanknieitts 151 measurements of heights 372, 406, 419 Drainage system of Bonneville Basin 21 Drainage system of Great Basin 7 Drew, Frederic, alluvial fans HI Dry Canyon, fault >carp 340 Dry Cottonwood Canyon, moraines 309,340 fault scarps 340 Dug way Range, ihyolite 338 Dunderberg Butte 335,330 Dunes ,S9 Dunes of gvpsum 'Z'S.i Dunes on Sevier Desert 332 Dutch Point 53 Dutton,<'. E., eitedoneauseof aridity of Great Basin. 10, 280 cited on isostasy 388 Karth shaping 27 Karlli, strength of the 38? Earthquake waves and joints 213 E irtli quakes ;iGO East Canyon, fault scarps 3.')2 El Moro, Pleistocene eruptions 337 Elephant, fossil 21 1, 303. 304, 394, 400 Elevation of Bonneville Basin by expansion 427 Emergence, effect on shores 72 Embankment, compared with other ridges 87 Embankment series, Bonneville shore-line 111,369 430 INDEX. Embankment serios, Provo sliore-liuo 131, 132 ErabankmentH, littoral 46 rhythmic 73, 137 i»f Bonni'villo Mbore-line 1U8 of I'rnvo Hliort'-lino 127, 131, 132 of InteinuMliato shore-lines 135 conipouml 144 calcareouM cement 107 Emmons, S. F., iuvestigation of Pleistocene lakes ... 17 cited ou hijjbest shore-lino 90 (;ite(l on Tertiary in Rush Valley 99 cited on Litt!« Cottonwood glaciers 305 Empire BlutT.i 50 Eiidlirh.F.M., cited. 2C8 Engelmann, Ilt-nry, iuveHtigalionof Lake Bonneville. 15 lionrieville shells 209 Eocene lake beds 90 Epcirogeny defined 340 Ephedra jrracilis 259 E' 111 i potential surfaces 421 Etiuus fauna, question of age 393 Erowion by waves . . 29 Erosion cliff, compared with other cliffs 75, 77 Erosion terrace, compared with other terraces 78,84 Eruption, recency of latest 324 Escalante Basin, map 122 Escalante Bay, depth 125 question of synchronism 369 Escalante Desert, barometric measurements 362,406 heights of shore-lines 366, 415, 417, 418 Escalante Lake, theory of 363 Escalante, Padre, explorations 12 Sevier Lake 224 Evaporation formula 285 Evaporation rate in Great Basin 7 Expansion as a cause of post Bonneville deformation. 377, 427 Experiments in precipitation of sediments 205 Falsan, A., cited 271 Fans, alluvial 81 Farmiugton, installation of lake gauge 231,232,409 height of gauge 233,410 record of lake level 234 observations of lake changes 240 fault scarp 349 bench-mark 409.410 Fault scarp, compared with other cliffs 76, 77 Fault scarps, of Bonneville Basin 3t0 map showing distribution 352 general features 354 tlates of foi-mation 356 relation to *'arth Isoatasy 387 Jiimiesou, Tbomaa F., cited 265 Johnson, Willard D., field work on Lake Bonneville. 18 map of bay bars iu Snake Valley 112 survey of White Valley Bay 126 map of I^of;au River delta terraces 160 map of KimI Kock Pass. 174 map of Old liiver Bed 182 map of portion of Old River Bed 194 exidoration of Sevier Lake salt beds 225 map of Sevier Lake and salt beds . 227 Joint atrtictnre . 211 Jones, Marcus E-.iiauging Groat Salt Lake 232,237 Jordan Kiver, Tertiary lake beds 99 irrigation 250 anal^-ais of water 254 Jordan Valley, diastrophism 307 Joidati- Utah Bay 103 Juab Valley Bay 103 Juab Valley, fault scarps 343 Kamas Prairie, change of drainage ,... 218 Kame.s, cou)i)ared with other ridgea 87 Kanab Creek, I'leistoceue eruption 337 Kanosh, measurement of height 408,415,417 Keller, H., on littoral processes 26 Kelton Butte, view J08 discrepant .shore records 124 heights of shore-lines 365, 370,372 measurement of f*hore-lines 400,419 King, Clarence, aeknowledgmi iits to xv investigation of Pleistocene lakes 17 cited on highest shore-line yc Eocene near Salt Lake City 100 citetl on correlation of sedimentsand shorelines. 189 fossil mammals 2!1 brine of Croat Salt Lake...- r.52, 254 cited on correlation of lake epochs with glacial epochs 267 cited on ;;i.u'iation and heat 284 tlieory of Esi-alante shore-line 363 Ring Survey. See Fordrth Parallel Exploration. K noli Spi ing, fault sca?p 353 Knowl ton's ranch, fault scarji 352 La Sal. Sierra, volume 390 Lagging of lakes behind glaciers 314 Lahontan llnsin. conipletii dusiucation ....: 258 Lalumian mamtnalian fauna 395 Lake basins 2 Lake beds 188 Lake beds intorslratilied with delta gravels 150 Lake Bonneville. See Bonneville Lake. Lake Creek 185 Lake formerly in Colorado Desert 15 Lake rampart, mode of forma tiou 71 Page. Lake rampart, compared with other ridgoa 87 Lake ridges iu Ohio 43,44 Lake shores, topographic features 23 Lake Point, fault scarp 35*^ Lake Shore gaug ■, installation 231.409 connoction with other gauges 2'Ai litught 233.364,410 record 234 Lakes, Plei.stoceue, of the Great Basin, map- 6 of Great Basin 8 earlier than Bonneville 98 table of dimensions 106 correlation with glaciers , 265 Land sculpture 27 Landslip cliff, compared with other cliffs 77 Landslip teirace. compared with other terraces 83.84 Lartet, Loui.s, cited 205 Lattimore, S. A., analyses of Sevier Lake desiccation y>roduets , 220 Lava field, Ice Spring 320 Pavant Butte 328 Tabernacle 329 Furaarole 332 Lava, liquidity 322 Lee, C. A,, lake rampaits 71 Leevining Creek, glacial moraines 312 Leniington, geologic section 192 record of first water maxiiniim 19!) height ot shoreline 36'. measurement of heights ...411, 412, 414, 417 Lenk, 11., Pleistocene lakes of Mexico 402 Level of still water 122 Leveling, account of work 304,411 probable error 417 Limuiiphysapalustris 300, 301 Little Cottonwood Creek, ancient delta 165 moraines 305,306,340 fault scarps 34© map 340 Lit(le(iull Lake yoo Littoial deposition , 40 Litloral erosion 29 Littoral topoijiapby 23 Litttnal trausjiortation 37 Llam;is, fossil 303,391,400 Loading, unhiadiiig, and tlefnriuation 357,379, 421 Loew, t»scar, analysis of Sevier Lake water 220 Logan, deltas if,ip map of deltas inu view of del [as 102 I'.iitlt seatp 351 heights of shore-lines 365. 370, 372 measurement of heights 411,412,413,417,4:8.41!) Lone Pine, tarthipiake 361.302 Loops, origin ami eharacler 55 outline maps 5g of Bonneville shore-line 109 Lost mountains 215, 320 Lower River Bed section 189 Lyell. Charles, method (.f Tertiary clas8t6cfttion 39H cited on principles of e(u relation 401 Main body 101,122 Major, C. I. Forsyth, Pliocene niamm.ils 4O0 Malade Valley Bay 102 INDEX. 433 Page. Mammalian fossils, from Eonnevillo beds 210 and Bonneville climate 303 from Christinas Lake 394 Map of Lake Bonneville, autborities 125 Marcet, W., brinoof Great Salt Lake 254 Markatjunt Plateau, Pleistocene eruptions 336 Marsh Creek, description 174 alluvial terrace 175 lower course 176 alternation of tribute 178 fault scarps 351 Marsh, O. C, cited on Equus fauna 393, 395 Marsh Valley, terraces i'5 general features 17G Matching 399,402 Matlin, Tertiary lake beds 99 measurement of heights 372, 40C, 419 Mature coast lines 03 McGee.W. J., field work 18 oolitic aand 169 cited on number of glacial epochs 272, 274 cited on deflection of glaciers 315 Equus fauna 393 McKay, Alexander, cited 361 Meadow Creek, measurement of heights 408,415,418 Measurements of shore-line heights. 362, 405 Melting curve 289, 293 Mexico, Great Valley 402 Michigan Lake, evaporation 7 subaqueous ridges 43, 44, 45 view of bay bar 48 bay bar 50 hook at Dutch Point 53 Mil ford, height of Bonneville shore-line 365 raeasurement of heights 408, 411, 412, 415, 417 Mill Creek, moraines 311 Miller, Hugh, classification of stream terraces 80 Miller, Jacob, Farmington gauge 231,234 rise and fall of Great Salt Lake L'40 soundings on Antelope Island bar 211 Mimbrcs Basin 11 Mitchell, Henry, formation of beaches 26 Mitchell, John T., gauge observations 231,233 Miter Crater 321, 322 Mohave River 8 Molluflks, from Bonneville beds 209 and Bonneville climate 297 from Christmas Lake 395 Mono Lake, observations by J. D. AVhitney 16 Mono Valley, shore-lines and moraines 306, 311 Pleistocene eruptions 337 Montanari, theory of littoral transportation 41 Montpellier, Pliocene fauna 399, 400 Moiaine terrace, compared with other terraces 81, 84 Moraines, compared with other ridges 86,87 and ancient shore-lines. 305 and fault scarps 346 and climate 398 Morgan Valley. Tertiary lake beds 99 bay of Lake Bonneville 103 fault scarp 351 Mountains, of Bonneville Basin 91 view of buried 320 growth of 359 Muddy Pork, deltas 162 MON I 28 Page. Murray, John, cited 12 Musk ox, fossil 211, 303 Mylodon sodalis 391 Neocene and Equus faunas 393 Neocene geography of Bonneville Basin 214 Neocene lake beds 99, 173 Nell, Louis, survey of Sevier Lake 225 New Garfield gnage 232,233,237 New Mexico, interior basins 11 Pleistocene eruptions 337 New Zealand, earthquakes 361 Newberry, John S., cited on number of glacinl epochs 272, 273 Nomenclature, geologic 22 Ocean currents in relation to glaciation 281 Ogden, altitude 3tJ4 height of Bonneville shore-lino 365 measurement of shore-lines 411,412, 413,417 Ogden Canyon, fault scarps 350 Ogden Piver, ancient deltas 93, 163 Old River Bed, map of V-embankments 58 description 181 map 182 lower section 189 upper section 194 geologic map of portion 194 Ombe Range, Tertiary lake beds 99 island in Lake Bonneville 102 Ontario Lake, headlands and bay bars 50 fetch of waves reaching Toronto 53 simplification of coast-line 63 distribution of mature and adolescent coasts ... 65 Oolitic sand 169,252 Oquiri h Range, view of lake terraces i fault scarps 352 Oregon, Equus beds 394, 397 Orogeny discriminated Irnm epeirogcny , ... 340 Oaar, compared with otlier ridges 87 Otter.fossil 303,304 Outlet channels, characters 171 relation to shore-lines 186 Outlet of Lake Bonneville, description 171,173 literature 173, 182 map 174 view 176 question of earlier 1 80, 216 Owen, Fred. D., general assi.'itant 18 sketch of head of Tooele ^nlley 96 Owen's Valley, earthquake 361, 362 Ox, fossil 303 Packard, A. S., cited on Old River Bed 182 fauna of Great Salt Lake 258 Pahoehoe, Pavant Butte 328 Tabernacle Butte 330 Paleoutologic evidence on ago of Equus fauna 397 Paleontologic methods of correlation 398 Parallel Roads of Glen Roy 65 Park. John R., gauging Great Salt Lake 231, 409 Park Valley Bay 102 Pass between Tooele and Rash valleys, description 52, 97 map of hook 58 view 96 map 138 434 INDEX. Pass between Tooele and Hash valleys, saperposition of embankments ancient river Pavant Butte, descnption view height of Bonneville Hhore-liue measurement of shore-lines -108, Pavant Kange Pa.vsoii, di'lta near Peale, A . C, observations on Lake Bonneville cited on shoreline higher than the Bonneville .. cited on outlet of Lake Bonneville ^ cited on age of Bonneville beds Penck, Albreeht, cited Physa ampullacea Physa gyrina Physiographic evidence on ago of formations rhywingraphy Pilot Peak, terraces Pink Cliff formation on Sevier liivcr Pinto Cauyou, measurement of heights 408, Plant, fossil Playa de los Pimas Plaj'as of the Bonneville Basin Pleistciceue, shortest of the periods lakes, map name preferred to Quaternary climate volcanic eruptions 323, 326, 330, winds Eqiius fauna two uses of term mammalian fauna, Great Britain lakes, Mexico (See, also, Bonneville beds. White Marl, and TeUo^v Clay.) Pliocene and Equus faunas Pliocene fauna of Aruo Talley Pliocene fauna of Montpellier Point of the Mountain, lu-.tn of V-bar sca-elitf iiue(|ual I'liibankuK-nts piolilu uf omhankuR-nts heights of shore-lines 365, measureuient of shore-linos 411,412,414, 417, I*noU', Henry S., obsei-vatious on Lake Bonneville. . . Portage, didta near Portnenf lliver, terraces lower canyon in Marsh Valley possible chiingos Post -Bonneville history Powell, J. W., acknowledgment a to cited on yonth ()f high mountains Powell Survey, fiild work on Lake Bonneville gauging Great Salt Lake Pratt, John H., cited Pre- Bonneville history Piecipitati(m and interior basins in Great Basin secular curve for Great Basin Precipitation of sediments. expL-riments Prenss Lake Preiias V:illey, V-hars map of east side 149 184 325 328 366 415,417 319 165 18 94,95 173 267 271 300 301 396 27 144 99 415,418 210 11 222 1 6 22 265 336, 338 332 393 395 399, 400 402 393 399, 400 399, 400 58 107 123 138 370, 372 418,419 16 16ii 95 96 176 219 222 XV 350 18 230, 409 387 214 4 (i 245, 246 205 2J4 58, 121 92 Pago. Preuss Valley, discrepant shore records 124 maps of embankments 136 embankments 137 profiles of embankments 138 interval between embankments 141, 143 double series of embankments 152 record of tirst water maximum 109 measurement of heights 372,412,419 Probable errors 4 16 Profiles, Bonneville Bay bars 116 Provo shore-line 132 Intermediate embankments 137, 138 Promontory Mountain, an island in Lake Bonneville. 102 at the Provo stage 128 heights of shore-lines 306, 370, 372 measurement of shore-lines 412,413,417,418, 419 Provo epoch, displacements 372 Provo River, sucieut deltas 153, 165 change of course 218 Provo shore line, north end Oquirrh range, view i origin of name. 120, 153 outline and extent 127 later than Bonneville shore-line 127 cut-terraces 128 deltas 129,153 underscore 130 embankment series 131 area included 134 map 134 tufa 167 on Pavant Butte 326 on Tabernacle lava field 330 altitudes at various points 370,418 deformation 371 curves of equal height 372 Publication of work of Great Basin Division XVII, 19 Quaternary. (See Pleistocene.) Kailroad altitudes 411 Kainfall, interior h.isius aiul 3 of Great Basin 6 secuhir curve f^r Great Basin 245.246 Uanipart, mode of formation 71 comparctl with otlu^r ridges 87 Kamsey, precipitation of sediments 20 Skull Valley, embankment aeries 112,113,122 Shith. fossil 303 STiiitli, R. II., barometer ohse.rver 407 Smitliliold Creek, delta , 162 Snake Valley, map of V-bars 58 bay of Lake Bonueville 104 V-embaukmeiits 108 embankment series 111,112 salt marsh 223 fault scarps 353 Snowfall Curve 289, 293 Snow-plow, map of V-bar 58 embankments. 137 map and view 138 embankment intervals 141,143 supei'ijositiou of emb nkmenta 147 measurement of shore-lines 372, 412, 419 Snows ville Valley, river channel 185 Bonneville beds 191 fault scarp 351 measurement of shore-lines 406,419 Sodium sulphate, precipitation fromGreatSaltLake. 252 Sonora, eaithquake 361 Spanish Fork, deltas 153,165,343 fault scarps 343 Spirit-level measurements 364,411 Spits, mode of formation 47 of Bonneville shoreline 108 near Grants ville, map 134 Spring: Creek, deltas 162,168 Stansbury, Howard, cited on shores of Lake Bonne- ville 13 map of Rush Lake 228 survey of Great Salt Lake 230 map of Great Salt Lake 243,244 brine of Great Salt Lake 251,254 Stansbury Island bar 241,243 Stansbury shore-line, described 134 tufa 167 bypotUetic explanation 186 height 418 Stelling, formula for evaporation 2S5 Sternberg, C. H., collection of fossil bones 303 Stillwater level 122 Stockton, shore-lines near 52 V-bar 58 view of shore-lines 97 IntermedialL'ombankmonta 137, 138, 149 map 138 height of shore-lines 365,370,372 measurement of shore-lines 412, 414, 417, 418, 4! 9 Stoppani, A., cited 271 Strachey, Richard, cited 2S4 Stream Cliff, compared with other cliflfs 75, 77 Page. Stream terraces, compared with other terraces 79.84 claaaiticatinn 8U Sub-Appenine fauna 397,399 Subaciiumus ridge 43 Submergence, effect on shores 72 Sulphur Springs, mea«uroment of 8horc-lino.366, 408, 415, 418 Superior Lake, bay bars 51 Survoy of the Rocky ilounlaiii Region, fiehl work on Lake Ilonuevilh) ]8 gauging Great Salt Lake 230,409 Survey of the Territories, Neocene lake beds 99 Surveys West of the lOOth Miaidiaii, investigation of Lake Bonneville 17 map of Rush Lake 228 measurement (»f Hhoreline 362,414 Synchronism of Bonneville shore-line 309 Tabernacle crater and lava 6eld, map 328 view ^ 328 description 329 Talmage, J. E., analyses of water of Great Salt Lake 252, 253 Taramelli, cited 271 Tavaputs Plateau, volume 390 Taylor, volume of Mount 390 Teconia, height of Bonneville shore-line 365 measurement of height 411,412, 417 Temperature, secular curves for Great Basin 246 and humidity . 265 relation to glaciation 276, 283 relation to growth of mollusks 300 Temperatures of fumaroles 333 Temperatures of hot springs 333 Terrace Crater 322 Terrace Mountain, spits of Bonneville shore-line 108 Provo embankment series 131, 133 shore-lino measureuieuts 372 Terraces, north end of Ofiuirrh Range, view i wave-cut 35 cut and-huilt 36,40 wave-built 55 classification 78 comparison 84 of disputed origin 95 of Provo shoreline 127, 128 Tertiary. See a.\so Keoccne and Pliocene. Tertiary beds. Red Rock Pass 173 Tertiary lakes 98 Texas, interior basin 11 Thermos Springs, measurement of heights 408,415,417 Thompsou, Gilbert, contributions toEonnevillemap. 126 map of embankments near Dove Creek 138 discovery of outlet. 173 cited on Bear River drainage 219 map of mouths of Little and Dry Cottonwood Canyons :!46 Thomson, AViUiam, cited on elasticity 381 coetticient of diffusion 42G Tidal shores 28 Time ratios 159,260 Tintic Valley Bay 104 Tooele Valley, ancient shore-lines 14 Provo embankment series 131, 132 fault scarps 353 INDEX. 437 Page. Tooele Valley, diaatropbism 367 (See alao Grantsville and Pass between Tooele and Rash Valleys. Tooele-Rusb Bay 104 Topograpliic features of lake shores, described 23 distribution 6 discriujinated 74 Titpnnirnpbic interpretation of lake oscillationa 262 Toronto Harbor, structure of peninsula 5-i map 54 Towiia on site of Lake Bonucvillo lOG Trans-Pecips interior basin 11 Transportation, littoral 37 Traverse Range, fault scarp 346 Trees uf Great Basin 9 Trrsca, flow of solids 383 Triaugulation, heights measured by 406 Trowbriujie, E, R., general assistant 18 Tuilla Valley. See Tooele Valley. Tufa, near Pyramid Lake 33 of old shorelines 167 not found in Cache Valley 179 on Tabernacle lava bed 330 Tuff.Pavant Butio 326,329 TaberiLtrlo Butte 329 Twin Peaks, measurement of shore-line 408,415,417 Tyudall, John, cited 284 Uiakaret Mduutains Pleistocene eruptions 337 Uuconfoimityof White Marl on Yellow Clay. 190, 192, 194, 197 Undercutting 15L Underscore 130,132 U:;clertow. origin 30 function 33, 38 pulsation 33 Uphaui, Warren, cited 272 Upper River Bed. section 194 geologic map 191 Utah Lake water, precipitation experiments. 206 analysis 207,254 Utah Valley, fault scarps 343 Valleys of Bonneville Basin 91 Valleys of Great Basin 6 Viisey, George, identification of fossil plant 210 V-b:ira, description 57 outlines 58 of Bonne villu shore-lino 108 interpretation 121 Vegetation of Great Basin , 9 Vertebrate faunas, compared 397 Vertebrate fossils and Bonneville climale ..'... 303 Volcanic district near Fillmore, map 320 Volcanic epoch nut clos-jd 339 Volcanic formation of Bonneville Basin 319 Walled lakes 71 Walling, H. F., theory of joint structure 213 Warm Spring Lake 300 Wasatch glaciers and Lake Bonneville 305,306 Wasatch Range, fault scarps 342 loading and displacement 357 now growing 359 volume 389 Water analy sea Water of Great Salt Lake Water of Lake Bonneville Water of Sevier Lake Wave-built terrace, described compared with other terraces Wave-cut terrace, described compared with other terraces Waves, shore-forming agents theory of refraction function in littoral tiansixutation fetch function in building embankments fetch, on Lake Ontario Webster, Albert L., computation. survey of Escalaute Bay map of shore features at Wellsvillo compilation of gauge data map of Fillmore volcanic district barometric work and compilation of altitudes. .. appendix on altitudes Weber River, deltas change of drainage fault scarps Wells ville. view of terraces /. . discrepant shoie records embankments 137^ map and view of ombaukments measurement of heights 372, Wheeler, George M., position of Sevier Lake Wheeler, H. A., survey of Tintic V;illey Bay map of embankments near Grantsville map of Snowplow map of pass between Tooele and Rush Valleys. . map of Fillmore volcanic district Wheeler survey, investigation of Lake Bonneville.. map of Rush Lake measurement of shoreline White, C. A., explanation of lake ramparts White Marl, character ami distribution upper River Bed section cause of whiteness analyses over lava White Mountain, gypsum map rhyolite height of Prove shore-line measurement of heights 408,412, White Valley and Stausbury shore-line White Valley Bay Whitney, J. D., observations on ancient shore-lines of Mono Basin cited on cause of cstenaion of Mono Lake cited on climatic history of Great Basin cited on synchronism of glaciatiun cited on glaciatum and heat cited on relations of Plei-stocene lakes and gla- ciers '. Owen's Valley earthquake Whittlesey, Charles, cited on subaqueous ridges Willard. heights of shoredines 365, measurement of heights 412, 413,417, Page. 207, 254 252 204 226 55 84 35 84 29 29 30 27 43, 107 46, -17 53 110 126. 370 138 232,409 320 365 405 164, 349 218 349 98 124 139,143 138 412,419 224 126 134 138 138 320 17 228 362, 414 71 190 195 200 201 328, 334 223 320 338 370 415, 418 186 104 16 266 266 270 284 314 3G2 43,44 370. 372 418,419 438 INDEX. Willow Springs, hook near 145 \V iiiii waves, theory 29 Winds, Pleistoccno 332 Woir, fossil 303,394,400 Wocidwanl, R. S., on tho deformation of the gcoid by tlie removal, tliroiigli evaporution, of the water of Lake liouni-ville 377,421 on the elevation of the surface of tho r.onni-villn Basin by expansion due t" change of rli mate. .'I7H, 425 Papo. Wi.oilwanl, R. W.. analysis of tufa 168 Wright, O. Frederiek, cited 274 VVright, George M., field work 18 observation of oolitic sand 169 Yellow Clay, charatter and dialribntion 100 upper Jviver Bed section 194 analyses 2U1 Young, Williaid, riled 173 V