A SEDQME NTARY , FE‘FRGGPAP’HK A943 STATISTECAL SYUCSY 0F CERTAM GLACML CLAYS OF MQZNHERN MCHIGAN AS AN MD EN CGRWTiCN Thain For flamers-a :6 M. 3. _ mgcmcm sure mass: @5er M. Tinédopaugh 1955 THFSlS J|\M\\Wfl\WWW\WN 'L 31293 010717118 0 This is to certify that the thesis entitled A Sedimentary, Petrographic and Statistical Study of Certain Glacial Clays of Northern Michigan as an Aid in Correlation presented by $tty M . Tinklepaugh has been accepted towards fulfillment of the requirements for M0— degree in _G9_012£¥_ “731. U Major professor Date April 21.9 1955 0-169 A SEDIMENTARY, PETROGRAPHIC AND STATISTICAL STUDY OF CERTAIN GLACIAL CLAYS OF NORTHERN MICHIGAN AS AN AID IN CORRELATION By BETTY M. TINKLEPAUGH A THESIS Submitted to the School of Graduate Studies of Ndchigan State College of Agriculture and Applied Science in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE Department of Geology and Geography 1955 THESIS ACKNOWLEDGMENTS The author wishes to eXpress her sincere thanks to Dr. B. T. Sandefur for his aid and helpful suggestions throughout this investigation. The writer is also greatly indebted to Dr. S. G. Bergquist for the samples used in this study and for the critical and constructive editing of the manuscript. Grateful acknowledgment is also due to Dr. W. D. Eaten of the Mathematics department for his suggestions and general assistance in the choice of statistical methods employed and statistical interpretations best fitted to this study; to Dr. A. D. Perejda for his assistance in the preparation of the maps and figures; and to Dr. J. W. Trow for a critical reading of the manuscript. The writer is grateful to A. M. McCarthy for assistance in editing the manuscript. ABSTRACT The use of sedimentary, petrographic and statistical methods of analysis appears to be a means of determining the characteristic properties of a sediment. In cor- relating glacial morainic systems in the northwestern portion of the Southern Peninsula of Michigan a distinction has been made between certain glacial clays which are either red or blue in color. In attempting to determine differ- ences or similarities between the "red" and "blue" clays nineteen samples from exposures of these deposits were analyzed with respect to their size distributions and heavy mineral frequencies. Cumulative curves were constructed from.the data ob- tained in the size distribution analyses from.which vari- ous statistical values were derived for interpretation. The coefficient of correlation and the chi-square test, both of which are statistical analyses, were used in comparing the data of the heavy mineral studies. Interpretation of these relationships is attempted in order to determine and to compare the characteristics of the "red" and "blue” clays. iii From the results obtained in this investigation, the author has concluded that there are no significant differ- ences between the "red" and "blue" clays. Therefore, any correlation of morainic systems in the northwestern portion of the lower Peninsula of Michigan based entirely on color alone would not appear to be valid. For a glacial interpretation the laboratory investi- gation would tend to support the opinion of some glacial geologists that the clay core material of the Manistee and Port Huron moraines are similar and Cary in age, and that each core is covered with a younger glacial drift, the mankato, whose terminal position is marked by the Port Huron moraine. iv TABLE OF CONTENTS Page INTRODUCTION ................................... 1 General Information ......................... Purpose of Study ............................ SAMPLE SELECTION ............................... Source of Samples ........................... Collection of Samples ....................... SAMPLE LOCATION AND LITHOLOGIC RELATIONS ....... Location of Area ............................ Distribution of Samples ..................... \OO\O\O\\n\n\thl—' TflflfliStee Ifloraine 0.00....00.000.000.000... Mancelona Spillway and Nanistee outwaSh Plain OOOOOOOOOOOOOOOIO000...... 12 Port Huron Nbraine ....................... 13 Lake Border Moraine ...................... 14 Algonquin Lake Plain ..................... 15 LABORATORY PROCEDURE ........................... 16 General ..................................... 16 Preparation of Sample ....................... l6 Dispersion .................................. l7 Reciprocating Shaker .......... ...... . ...... . l8 Flocculation ................................ l9 Page Soluble Carbonates ........................... 19 Dilution ..................................... 19 The Pipette Method ........................... 20 Stokes' Law ............................... 20 Principles of method ......................... 21 Technique of Analysis ..... .............. ..... 23 Computation of Results ....................... 25 Results of Laboratory Analysis ............... 26 Possibility of Erroneous Data ................ 26 Required Laboratory Equipment for Pipette Analysis ................. ....... ... 3h Sieving ...................................... 35 Separation of Minerals ....................... 36 Mounting of Slides ........................... 39 Heavy Mineral Analysis ....................... 39 Selected Slides for Correlation .............. no Roundness and Sphericity Analyses ............ #3 Laboratory Equipment for the Nuneral Analysis ........................... #3 STATISTICAL ANALYSIS ............................ #5 Introduction ................................. #5 Cumulative Curve Analysis .....................h6 Coefficient of Correlation ................... 7O Chi-Square Test .............................. 71 Chi-Square Versus Coefficient of Correlation . 72 vi Page COMPARISON AND STATISTICAL INTERPRETATION OF DATA . 7h Introduction ................................... Analysis of Size Distribution .................. Summary of Size Analysis ....................... Analysis of Heavy Mineral Distribution ......... Summary of Heavy Mineral Analysis .............. Summary Comparison ............................. GLACIAL mTERPRETATION 0000.0000000000000000000000. CONCLUSION SUGGESTIONS FOR PETE-{TIER STUDY ....OOOOOOOOOOOOOOOOO BIBLIOGRAPHY vii 7A 7A 80 81 85 86 87 89 91 92 TABLE I. II. III. V. VI. LIST OF TABLES Page Time Chart for Pipette Analyses ........ Data from Pipette Analyses ............. Grain Sizes Larger Then 0.5 mm. in Samples Studied OOOOOOOOOOOOOOOOOOOOOOOO Heavy Mineral Frequency Distributions .. Total Grains Counted ................ Total Percentages ................... Percentages and Quartile Calculations from.Pipette Analyses .................. Statistical Correlation Values for Heavy Mineral Analyses ................. viii 22 27-33 37 hl-AZ Al #2 69 82 LIST OF FIGURES FIGURE Pages 1-19 Cumulative Curves of Size Distribution .. 47-65 11 LIST OF MAPS MAP Page I. Areal Extent of the Glacial Moraines Included in the Investigation ............ 7 II. Sample Locations ......................... 8 INTRODUCTION General Information Sedimentary, petrographic and statistical methods of analysis have been used in correlating sandstones, classify- ing types of sediments and determining stratigraphic hori- zons. However, in only a limited amount of the research done by glacial geologists have these methods been employed in correlating, determining mode of deposition, age re- lationships, and sources of glacial materials. With the rapid growth in the field of glacial geology in the past half century many controversial issues have arisen relative to correlation, interpretation, and age re- lationships of the various glacial features remaining in the wake of the relatively recent ice sheets of the Pleis- tocene. D One of the more recent of problems in the glacial geology of Huchigan is that of correlation of morainic sys- tems in the northern portion of the lower Peninsula. The Port Huron and Manistee moraines, the last two sys- tems to be developed in lower Michigan, have become the object of much debate regarding age and correlation with glacial deposits in the state of Wisconsin. It is the opinion of certain glacial geologists that the Port Huron moraine is Cary and that the Nanistee moraine is Mankato in age. (Bretz, 1951) Both the Cary and mankato are substages of the Wisconsin stage of glaciation. Others contend that both the Port Huron and manistee moraines are Mankato in age with a Cary core. It is not the purpose of this paper to debate the pros and cons for either side but to introduce to the reader one of the deciding factors or characteristics of the glacial deposits which has been used almost exclusively in correlat- 1ng and in age determinations. This most important deciding factor has been the over all color of the material deposited in these large morainic systems. The problem, highly sims plified here, becomes one of recognizing "blue" and ”red" glacial tills. W. H. Twenhofel states: "Color of sediments is influenced by all en- vironmental conditions that are concerned with origin, transportation, and deposition. Among these are the character of the parent rocks; methods of rock destruction; climatic conditions; methods, dis- tances and duration of transportation; conditions at places of deposition; and diagenesis subsequent to deposition.” (Twenhofel, 1950) In the light of Twenhofel's statement the color of a sed- imentary deposit may have significance for the solution of a problem concerning it but, on the other hand, it may have no significance at all. Thus the problem of the "red" and "blue" clays was thought by the writer to provide a good study for the possibility of application of sedimentary and petrographic methods to glacial problems. "Red" and "blue" clays, which have been prominent as deciding factors in correlation and in age relationships of glacial deposits of northern Ndchigan, were chosen for the study. It is the contention of Bergquist that the "red" is a weathered portion of the "blue" lying above the water table in the zone of weathering and thus the clays would be contemporary in origin and age-~both being Cary in age and covered later by Mankato drift. (Bergquist, 1952) But if the "red" and "blueR are two entirely different clays, both in origin and in age, they may represent two different ad- vances of the Wisconsin ice sheet. (Bretz, 1951) Purpose of Study The main purpose of this study is to compare the "red" and "blue” clays by sedimentary and petrographic methods to determine if possible whether there are any significant differences or similarities between the two. With this in- formation it may be possible to advance an opinion in re- lationship to age and correlation of the two.morainie sys- tems. A secondary purpose is to attempt a statistical interpretation and correlation of the sedimentary and petro- graphic analyses and to compare them with field and labora- tory findings. Both purposes are an attempt to apply sed- imentary, petrographic and statistical methods to problems in glacial geology. SAMPLE SELECTION Source of Samples The glacial clay samples used in this study were ob- tained from Dr. S. G. Bergquist, Professor of Geology at NHchigan State College, who collected the samples while engaged in field work in northern Michigan. Collection of Samples Due to the unbedded and apparently homogeneous nature of the deposits, channel samples were taken as represent- ative of each exposure of the "blue" and "red" clays. .A channel sample is an elongated sample taken from a rel- atively narrow zone of an exposure and involves a contin- uous strip of material from top to bottom of the channel zone. All of the samples were taken below the zone of leaching. SAMPLE LOCATION AND LITHOLOGIC RELATIONS Location of Area The area included in this investigation is confined to the northwestern portion of the lower Peninsula of Michigan. The sample locations are confined to the following glacial features: 1) the Mhnistee moraine which swings northward through Mhnistee, Benzie, and Leelanau counties and thence eastward across Grand Traverse, Kalkaska and Antrim coun- ties; 2) the Manistee outwash plain and Hancelona epillway bordering the southern flank of the Manistee moraine; 3) the Port Huron moraine running northeastward from.Manistee counp ty to Otsego county; and h) a small segment of the Lake Border moraine in Wexford county. (Map I) Distribution of Samples The nineteen samples selected for this study were col- lected from scattered localities located in the following counties: manistee Wexford Benzie Antrhm Leelanau Otsego Grand Traverse Cheboygan Samples are designated by numbers from.one through nineteen, by a method such that the samples in each glacial 6 MICHIGAN m1 w. .__.._J . AREAL EXTENT or THE ’ i..._.._-!.__-.. GLACIAL MORAINES INCLUDED ., Jul: . ‘r-— IN THE INVESTIGATION 'l' | ! _.l_ __ "r 29 . C‘T‘Lo 4.0 20 o 20 40' t (culled I In" :1“ him D. Pen)“ 1 i M ”I" (allege, lep't. of Tool. 8 Eng. ) u- u- or ..- M (flog Puss (mtlgil "51 MICHIGAN MAP II SAMPLE LOCATIONS SCALE 20 MllES 20 A 40 20 KM. e. . '. ...Il‘ m Ive. om «cl in." l ‘0 u. y- Inn. .’ "0 mm . ‘t . uml ”(a at (k I 433m «vi. '0' LE I-I (epyti'it I”! “Malaysian feature or surface form are completely numbered before pro- gressing generally southward into the next feature of gla- cial deposition. (Map II) Nanistee Moraine The manistee moraine (Mankato in age) appears on the west side of.Michigan, near the city of Manistee, and fol- lows the shore northward through Manistee and Benzie into Leelanau county. (Leverett, 1915) It is the youngest morainic system in Michigan. Sample 1. Manistee County. Locality: Center Sec. 25, T. 22 N., R. 17 W. Base of shore bluff along Lake Michigan- directly north of Orchard Beach State Park, immediately above lake level. Characteristics: Dark, chocolate—brown, highly calcareous clay; compacted, jointed, blocky, resistant, water saturated. Remarks: This clay forms the basal and core material of the Nanistee moraine (Cary) overlain with sandy, gravelly, cobbly drift (mankato) 15-30 feet in thickness. Samples 2, 3, and h. Benzie County. Locality: SW k, Sec. 27, T. 26 N., R. 15 W. One mile due west of Benzonia in a 30 foot read out on Betsie River Road. lO Characteristics: Samples 2 and 3. Upper 12 feet, nearly pebble-free, chocolate-brown, calcareous clay; indurated, jointed, blocky and weathered. Sample A. Lower 18 feet, blue, calcareous clay interfingered at contact; smooth, plastic, pebble-free, unleached, unweathered and water saturated. Remarks: This clay material forms the core of the Manistee moraine (Cary) overlain with sandy, gravelly drift (mankato) 3 to A feet in thickness. Sample 5. Benzie County. Locality: SW k, 860. 9, To 26 Ne, Re lil- we County road cut one quarter mile north of the city of Honor-—15 to 20 foot exposure. Characteristics: Taken at 10 feet below the surface-~dark- brown, calcareous clay, weathered pink in color. Remarks: A.lO-15 foot eXposure of red clay core (Cary) which continues below'road level and overlain by sandy, gravelly drift (mankato) A to 5 feet in thickness. Sample 6. Leelanau County. Locality: SE k, Sec. 32, T. 28 N., R. 12 W. One mile west of Grand Traverse Bay on County Line Road #612 in an open cut exposing 25-30 feet of clay at Radio Tower Hill. ll Characteristics: 25-30 feet of calcareous, chocolate-brown, indurated, jointed clay containing occasional lenses of quartz sand and scattered pebbles; leached to a depth of 30-36 inches; upon weathering forms a mealy structure and becomes very plastic. Remarks: This clay forms the core of the Mhnistee moraine (Cary) and is overlain by sandy, gravelly drift (mankato) 5-40 feet in thickness. Samples 7 and 8. Grand Traverse County. Locality: NE k, Sec. 3, T. 26 N., R. 11 W. Six miles due south of Traverse City in a 50 foot out along railroad on the east side of the Boardman drainage. Characteristics: Sample 7. Chocolate-brown, indurated, Jointed, calcareous clay 20 to 25 feet in thickness. Sample 8. Bluish-gray, hard, blocky, calcareous clay exposed 10 feet above the base of the cut. Remarks: This clay core (Cary) of the moraine is over- lain by sandy, gravelly drift (Nankato) 15- 20 feet in thickness. Samples 9 and 10. Antrim County. Locality: Sec. 11, T. 29 N., R. 7 W. Four and one half miles west of Mancelona in a 20 foot road cut on Highway 88. Characteristics: Sample 9. A.chocolate-brown, Jointed, blocky, calcareous clay; pebble-free. 12 Sample 10. A blue-gray, calcareous clay interfingered at contact with brown clay. Remarks: This clay forms the core of the Manistee moraine (Cary) overlain by gravelly drift (Mankato) 10 feet in thickness. The samples were taken from an isolated morainic segment which is considered a part of the Manistee moraine by some and a portion of the Inner Ridge of the Port Huron moraine by others. mancelona Spillway and Manistee Outwash Plain Sample 11. Antrim County. Locality: SW k, Sec. 20, T. 29 N., R. 6 W. Mancelona spillway--outwash plain; clay pit with a 20 foot exposure on G. Williams' farm. Characteristics: An eight foot exposure of calcareous, indurated, Jointed, blocky, chocolate-brown clay at the bottom of the pit from which the sample was taken. Remarks: A clay core, possibly Cary in age, covered with sandy, gravelly drift (Manhito). Sample 12. Manistee County. Locality: Sec. 28, T. 22 N., R. It W. Manistee outwash plain between the manistee and Port Huron moraines. Characteristics: Below the surface of the eXposure, lh-BO feet, a reddish-brown calcareous clay, 10-15 feet in thickness. 13 Remarks: The clay core, possibly Cary in age, is over- lain with a mantle of outwash gravel and sand 6-20 feet in thickness (mankato). Port Huron Moraine The age of the Port Huron moraine, second youngest morainic system in Michigan, is the object of much debate. Bretz (1951) states that the Port Huron is latest Cary in age and was deposited during final retreat of the Cary ice sheet. Bergquist (1952), however, considers the Port Huron as Nankato in age marking the termdnal position of the Mankato advance with a veneer of drift over a Cary core. (Houeh. 1953) Sample 13. Otsego County. Locality: Corner of Secs. 15, 16, 21 and 22, T. 31 N., R. 3 W. Characteristics: A.chocolate-brown, calcareous clay taken 48 inches below the surface; weathers to a deep red color. Remarks: .A clay core below a shallow mantle of sandy, gravelly till (Mankato) 2h-h0 inches thick. Samples 1b and 15. Otsego County. Locality: SE i, Sec. 21, T. 32 N., R. 2 W. East of vanderbilt in a road cut on the east bank of the Sturgeon River. 1h Characteristics: Sample 1h. A seven foot exposure of chocolate- brown calcareous clay streaked with blue clay. Sample 15. A calcareous gray-blue clay exposed down to the road level; interfingers at contact with brown clay. Remarks: This clay core forms the base of the Port Huron moraine and is overlain by a sandy, gravelly drift (Mankato) 6-10 feet in thickness. Lake Border Moraine The Lake Border moraine lies toward the interior of the state and is the third youngest morainic system.in Michigan. It is generally considered by all glacial geologists as Cary in age. Samples 16, 17 and 18. Wexford County. Locality: T. 22 N., R. 12 W., near the village of Harriette along a tributary of the Manistee River. Characteristics: Sample 16. At a depth of four feet, chocolate- brown, jointed, blocky clay leached from.32- 36 inches; six feet in thickness; silty in text- ure. Sample 17. At a depth of eight feet, greenish- gray clay, blocky and heavily jointed; 2 feet in thickness. Sample 18. At a depth of 12 feet, extending 20 feet to bed of Slagle Creek, an eXposure of cal- careous blue-gray clay. 15 Remarks: The chocolate-brown clayey drift extends to the foot of the Lake Border Moraine and is composed of sandy, gravelly, bouldery drift which is pink to light red in color. Algonquin Lake Plain Sample 19. Cheboygan County. Locality: Sec. 12, T. 36 N., R. 3 W. Along pipe line trench, six feet deep. Characteristics: A mixed red and blue calcareous, blocky clay. Remarks: This sample, because of its mixed color character- istics, was used largely for comparative purposes but it is not significant in relation to the problem as a whole. LABORATORY PROCEDURE General This investigation involves both sedimentary and petrographic studies, employing methods of analysis re- lating weight percentages and quartile measurements as determined by the pipette method, roundness and sphericity measurements of quartz grains, and heavy mineral analyses. Preparation of Sample Field samples for each location were air dried and a preliminary disaggregation performed to avoid having ex- cessively large lumps of material in the test sample. Because each sample hardened during drying it was necessary to crush the material gently with a rubber pestle. (Krumbein, 1932) Care must be taken in this treatment not to destroy individual particles with too vigorous a pestling. The purpose of an initial disaggregation is to prepare the material in such form that it can be quartered into smaller samples and thus avoid selective errors. As each sample included dust-size particles it was necessary to hand quarter the field samples to avoid loss of these finer fractions. (Krumbein, 1938) 16 17 Each sample was then carefully weighed on a chemical balance to 30.00 grams. Krumbein (1932) found it desirable to have a final concentration between two and three per cent for pipetting, so that the proportion of fine material in the quartered sample should lie between 20 and 30 grams for a liter of suspension. (Krumbein, 1932) DiSpersion Since the pipette method involves a continuous set- tling of the particles acting as individuals, it is clear that complete dispersion of the sample is an integral and time consuming aspect of the method. Each 30 gram sample was first placed in a Ball quart Jar and soaked for approximately one hour in 250 cc. of distilled water as the next step in dispersing the par- ticles. In the case of silts and clays chemical as well as a physical dispersion is necessary to completely separate the individual particles. Colloidal particles when dispersed in a fluid medium are electrically charged, the charge in the case of clay being a negative one. By acquiring and holding this common charge the colloidal particles repel each other, thus breaking down aggregates and allowing the particles to act as individuals. The degree to which a given sample is dispersed depends upon the ability of the 18 electrolyte added to the liquid medium to supply the common charges needed. (Galliher, 1933) In order to increase the electrical charge, a peptizer must be added. Krumbein and Pettijohn (1938) found that a 0.01 N sodium oxalate solu- tion was the most effective for this purpose. This solution was prepared by adding 0.67 gram of sodium oxalate to the distilled water in which each sample was soaking in the quart Jars. Reciprocating Shaker The last step in the diSpersion procedure was performed by placing the Ball Jar with a tight fitting lid in a hori- zontal position in the box of a reciprocating shaker. Each sample was then shaken in a to-and-fro motion for a period of approximately 12 hours to insure a completely dispersed suspension. A.12 hour period was chosen by the writer in reference to the findings of Olmstead, Alexander, and Middleton, who state that the breaking of particles was negligible even after 16 hours of shaking. The boiling of samples has many advocates, but as there is little question that prolonged boiling causes some of the colloids to coagulate, the writer refrained from boiling any of the samples during the dispersing procedure. (Krumbein, 1932) l9 Flocculation To determine whether the 12 hour period of shaking was sufficient to produce a completely dispersed suspension each sample was allowed to stand_over night. If visible floccu- lation did not occur within 16 hours the suspension was con- sidered completely dispersed. (Krumbein, 1932) When floccu- lation occurs during the test period, a "fluffy" sediment is observed at the base of the suspension, lighter in color -than that of the layer of coarse silt that also accumulates. Soluble Carbonates No provision was made for the removal of carbonates in the samples during the preliminary treatment. It was noted by Krumbein (1932) that when primary carbonate particles are present, they should not be removed before the mechanical analysis is completed, because they are an integral part of the frequency distribution, and it would materially affect the data secured if they were first dissolved. Because a high percentage of carbonates made up the samples no acid was used before pipetting. Dilution After the 12 hour period of shaking the suspension was poured into a liter graduate with great care so as not to lose any of the dispersed solution. A wash bottle of 20 distilled water was used to completely recover the total sample from the Ball Jar. The sample, plus the washings, were now diluted to 1000 cc. with distilled water. It is well to point out here that all of the samples were dispersed in distilled water. Thus distilled water was used as a constant throughout the pipetting procedure. R. C. Hills (1934) reported that the character of the water affects the tests. He found that tap water invariably gave different results than those obtained with distilled water. The Pipette method Stokes' Law One of the fundamental principles on which mechanical analysis is based is that small particles will settle out with a constant velocity in water or other fluids. The physical laws governing the settling of Spheres are well known, and the theory may be extended to include the ir- regular particles which comprise a fine grained sediment. The classic formula for settling velocities, and the best known, is that of Stokes which confines itself to spheres. If the sphere is very small (of the order of less than 0.1 mm. diameter in water), the resistance offered by the fluid is proportional to the product of the diameter and velocity of the particle, times the viscosity of the fluid. By equating these relations for small particles it 21 can be shown that the settling velocity (v) is proportional to the square of the particle diameter: v = Cld2 where the various constants (particle density, fluid density, acceleration of gravity and fluid viscosity) are included in 01. This expression is the simplest form of Stokes' Law. (Krumbein and Sloss, 1951) For practical purposes, the most convenient manner of using Stokes' equation is from tabulated values of the settling velocities as listed in Table I. Principles of Method The pipette method was established independently in 1922 by Robinson in England, Krauss in Germany, and Jennings, Thomas and Gardner in America. (Krumbein, 1932) The prin- ciples of this method rely upon the assumption that in a dilute suspension the particles settle as individuals, an assumption which is inherent in any method of analysis. If a suspension is thoroughly shaken so that the particles are uniformly distributed, and then set at rest, all particles having a settling velocity greater than h/t will have settled below a plane of depth h DBIOW'the surface, at the end of an interval of time t. All particles having a velocity less than h/t, however, will remain in their original concen- tration at depth h, because they will have settled only a 22 Awmmd Nsaonsdhmv .mo.m op assoc pqmsflvmm map no mpfi>mpm camaommm ammuopm as one «.000« He onspmnomsop a so momma one came» wasp ca mmsdwp one ms Nm mN mo m ”Noooo. msooo.o HN ca NJ NM m msoooo. moooo.o. as m HN as m mmoooo. wmooo.o Ha m HN OH OH meaooo. wmaoo.o m s 0H m OH smooo. mmaoo.o N N m a 2 $08. RN88 N H m N S 028. $8.0 on o a H OH NNNoo. mmoo.o ma 0 an 0 OH msmoo. mNoo.o 0H N 0 ma 0 OH moao. oaao.o Nm m o a: N o as NHNo. omao.o em a o Nm m 0 OH mmao. HNNo.o mm o 0 cm H 0 OH memo. NHmo.o mm o 0 cm a 0 ON :NH. Naso.o mm o 0 mm o 0 ON Nam.o mNeo.o .oom .csz .mpm .omm .asz .msm asap wcwwacA .so .omm \ .so ngoposfiaafis mafiappmm maoscfiscoo as“: was» Nasappmm as pnmsmm as mpfiooams as cosmemfiq mmmand24 maemmHm mom em omm~.~ oo.ooa mm.4 omom.a mafia gnaw m4.cm mo.mm ommm.m am.~m Hm.:~ ompz.~ m:.mm m~.mm ooma.~ maooo.o ”p.50 m:.~a ommp.m m:.~o -.c omao.~ oo.a~ mm.w oo~:.~. moooo.o mm.mm om.m oowH.H o~.oo m~.m omwm.m m:.mo oo.m ooo~.m mmooowo ~4.Hm H~.a omam.m Ho.am mo.m ommm.w «4.4m mmrm coward mmaoo.o H~.H: 00.0 ecum.a mm.~: mm.p omma.~ om.om m:.~H omMP.m mmfloo.o me.mm mm.o omwo.~ HH.mm mm.~ comm.“ mo.mm oo.HH oom:.m cpwoo.o o~.mm Hm.m om-.a mm.- «0.0 ommm.a m4.om 044m oowmwm mmoo.o mm.mw Hm.m omaa.a om.o~ ~:.m om~o.a mo.mH o~.m ooap.a mmoo.o mm.ma oa.o comm.a :m.ma om.; oo~a.a mm.ma mm.m ommmwo mpoo.o ~m.~H am.a omaa.a Jo.oa om.m ooaa.a om.m mo.a omm4.o oaao.o Ho.m do.“ omoo.o 4m.o oo.m ooom.o mm.~ o~.a oomm.o omao.o 00.0 om.H oozm.o am.m om.a oagm.o mo.o ~m.o ooaa.o ammo.o ow.4 Ha.o omma.o ao.~ $4.0 ooqa.o N~.m mo.H ommm4o mamo.o mp.m wa.m ommm.o hm.a 5H.o oomo.o 40.: on.: ooam.a Nazo.o Ho.o Ho.o omma.o o;.H o:.a oom:.o mw.o mm.o omwo.o mmoo.o . ma mangam 4a magaam ma magawm unmpmafiaafia “.35 as: 3 ...E “.55 as: 3 ...E «.25 mi: 3 $3 5 3358 Acoaufipnoov HH gaudy 32 8.88 8.88 88.8 8.88 8.888 888.8 88.88 8.88 888 83.8.8 88.888 88.8 8888.8 88.888 88.8 8858.8 88.888 88.8 8884.8 8888 8888 48.58 mm.08 0m00.0 08.58 40.88 00mm.m 08.88 04.48 0mmm.5 84000.0 88.55 84.8 8888.8 88.58 88. 8884.8 48.58 55.4 8884.8 88888.8 08.85 08.8 00mm.0 85.80 ~4.m 0m80.8 58.80 00.4 0m04.8 88000.0 45.88 85.8 8888.8 88.88 84.8 8888.8 88.58 84.5 8848.8 88888.8 88.48 88.8 8888.8 58.88 84.8 8888.8 84.88 88.8 8858.8 88888.8 5m.mm 80.8 0085.8 54.0: 80.5 0088.8 , 08.84 «0.0 0mom.8 05800.0 48.88 88.5 8888.8 48.88 58.5 8888.8 88.88 88.8 8888.8 8888.8 48.84 88.8 8858.8 54.88 88.8 8888.8 88.88 88.8 8858.8 8888.8 84.88 84.5. 8888.8 58.88 88.5 8888.8 84.88 88.8 8888.8 8588.8 88.88 88.8 8888.8 88.58 84.8 888.8 84.88 58.4 8884.8 888.8 58.88 88.8 8888.8 88.88 88.8 8888.8 84.48 88.4 8884.8 8888.8 88.48 88.8 8888.8 88.5 88.8 8888.8 88.8 58.8 8888.8 8888.8 85.8 88.4 8888.8 58.8 88.8 8888.8 88.8 88.8 8888.8 8888.8 84.4 88.8 8888.8 88.8 88.8 8888.8 88.4 88.8 8858.8 8448.8 05.8 05.8 008m.0 58.8 58.8 0084.0 05.8 05.8 0mmm.0 m800.0 88 888588 58 888388 88 888888 88888888888 8.8.8 8.8: 880 .83 8.58 8.83 880 .83 8.58 8.8: 880 .83 :8 838588 88.88888 88 388.8. 33 88.888 88.888 8888.88 88888 88.888 88.88 8888.4 8888 8:88 88.88 88.88 8888.8 84888.8 00.00 0m.m 00m0.8 80000.0 m8.mm 08.8 0058.8 08000.0 88.88 85.5 8888.8 88888.8 3.84 m5.m 0mN8.8 $8000 05. mm 05. m 305.8 05000.0 00.40 05.m 0m85 .8 8800.0 88.88 88.4 8884.8 8888.8 88.88 88.8 8848.8 8588.8 88.88 88.8 8888.8 8888.8 88.88 88.4 8888.8 8888.8 88.8 85.8 8888.8 8888.8 48.8 88.8 8888.8 8888.8 40.8 08.0 0858.0 8440.0 88.8 88.8 8888.8 8888.8 08 080500 08090588883 «.50 8.8: 83 .83 3.. 833380 Avodgnoov HH H549 3h error nearly constant. Possible errors due to theory rather than technique may be accredited to the settling velocities used in the pipette analysis. These velocities were calculated by Stokes' equation which is concerned only with spherical particles and not irregular particles of sediments. In Table I the values were calculated for a temper- ature of 20°C. and a specific gravity of sediment equal to 2.65. Therefore, temperature variation in the laboratory may have resulted in some small error which could have been equally balanced by both higher and lower variations in the temperature. (Hellman, l9hl) No great temperature changes in the laboratory were noted, however. If any flocculation had taken place by the end of a three day period of continuous settling in the graduate during the pipetting procedure there was no apparent in- dication of it. On the whole, the pipette method is based on sound principles, is simple in operation, and produces reliable results. Required Laboratory Equipment for Pipette Analysis The following laboratory equipment is essential for a quantitative analysis of the samples by the pipette method: 35 Quart Ball-Mason jars with tight fitting lids Reciprocating shaker Distilled water 1 liter graduated cylinders Pipettes--20 cc. Rubber tubing Beakers--50 ml. Electric hot plate and sand bath Analytical balance and weights Forceps Detergent Kodak timer Mortar and rubber pestle Wash bottle Sodium oxalate--peptizing agent Sieving All but three of the samples used in this study con- tained material larger than 1/16 mm. in size. At the con- clusion of the final pipetting, the suspension in the liter graduate was washed on a Tyler sieve, size 230, which is designed to separate the smallest sand-size grains from the largest silt—sized particles. That portion retained on the sieve, larger than 1/16 mm., was dried and analyzed on a set of Tyler sieves, sizes 60, 80, 100, 150 and 200. Wet sieving of the dispersed sample after, rather than before, pipetting was employed in accordance with the pro- cedure adopted by Bouyoucos (1938, 1936). This modification in procedure was prompted by the following: 1) the pre- sence of the material larger than 1/16 mm. did not affect the pipette analysis of the finer material, inasmuch as it settled from.the suspension immediately, 2) in washing the silt and clay particles through the sieve, tap water could 36 be used in an unlimited supply, whereas, if the sieving had been completed before pipetting no more than 1000 cc. of distilled water could be used, and 3) no care need be taken to recover the silt and clay fraction because the analysis of this portion of the sample had been completed. That small portion of the sample, larger than 0.5 mm. in diameter, and consisting of highly irregular shapes and sizes was then analyzed only for the largest particle pre- sent in each of the samples. The results for 16 samples are recorded in Table III. 0f the material analyzed on the set of sieves numbering 60 through 200 that portion between sizes 80-100 was re- served for the shape and roundness analyses to be discussed later in this chapter and that portion between sizes 150-200 was selected for the heavy mineral study for each sample. The remainder of the sand fraction was discarded. Separation of Minerals It will be noted from Table V that the sand fractions in most of the samples are small and with sieving that portion to be used for a heavy mineral analysis was reduced in some instances to less than one gram. Due to the limited quantity of the samples to be used for the heavy mineral studies no acid was used on any of the grains to remove carbonates, thus eliminating the possibility of losing more or all of TABLE III 37 GRAIN SIZES LARGER THAN 0.5 MM. IN saaans STUDIED Wentworth Size Classificationl Grade Limits Samples Classification Diameter in mm. Greater than # S, 11, and Pebble h mm. 19 h-2 mm. # b*, lb, and Granule 16 2-1 mm. # 1, 10*, 12, Very Coarse 13, and 15* Sand 1'0-5 mm. # 2: 3: 6a 9 Coarse Sand and 17* IWentworth Classification taken frolerumbein, 1938. * All "blue" clay samples Note: Samples 7, 8* and 18* had no fraction larger than.0.5 mm. 38 the remaining sample. The separation of the heavy minerals from the light minerals was accomplished with pure bromoform.(CHBr3), a heavy liquid having a specific gravity of 2.87 at 20°C. Quartz grains (sp. gr. 2.66), feldSpar (sp. gr. approx. 2.70) and other light minerals with specific gravities less than 2.87 will float on the surface of the bromoform. A large funnel, with a piece of rubber tubing attached and closed with a wire clamp, was half filled with the pure bromoform. The sample was sprinkled upon the bromoform, stirred and more bromoform added. Time was allowed for the heavy minerals to settle to the bottom of the funnel. The clamp was then released and quickly replaced after allowing the heavy minerals and some liquid to flow into a lower funnel containing a filter paper. The heavies were then washed with alcohol to remove the bromoform and allowed to dry. After removal of the heavy minerals, the bromoform in which the light minerals were floating was run off onto a clean filter paper in the lower funnel. This was also washed ‘with alcohol and allowed to dry. (Krumbein, 1938) All the bromoform used, except that lost by evaporation, can be recovered by washing with water. (Krumbein, 1938) This method was used for separations of both the 1/8 and l/h mm. grade sizes. The heavies, 1/8 mm., were saved and mounted for identification. The lights, 1/4 mm., were saved and mounted for roundness and shape studies. 39 Mounting of Slides All the mineral suites, both those for heavy mineral identification, as well as those for shape and roundness studies, were mounted separately in a medium of piperine (n = 1.68). The entire mineral residue to be mounted was mixed with piperine and placed on a glass slide which was then placed on a hot plate. After the piperine had melted (m.p. - 129.500.) a cover glass was added and the mount completed. Excess piperine around the edges of the cover glass was removed with alcohol or xyol. (Krumbein, 1938) Heavy Mineral Analysis By means of a polarizing microsc0pe each of the mounted heavy mineral suites was carefully studied. It was thought by the writer that identification and frequencies of the minerals for each sample might aid in correlating the clays. Thus each of the minerals present was identified according to characteristic Optical properties. Some of the funda- mental optical properties used in identification were: in- dices of refraction in comparison to the mounting medium by the Becke method; crystallographic orientation, optic sign, and optic axial angle with the aid of the gypsum.and mica plates and the quartz wedge; birefringence; color and pleochroism; and by close comparison with known mineral #0 mounts. (Walstrum, 1947) Each individual mineral grain was identified and record- ed. By employing a mechanical stage, several traverses were made over the whole width of a slide and each grain in turn was counted. Between 300 and 400 grains were counted for each slide except in those cases where the total mineral suite representing the 30 gram sample was less than 300 grains. Those samples with less than that number of grains were completely counted by close traversing of the slides. Selected Slides for Correlation Thirteen of the nineteen samples were identified and counted. (Table IV) Samples 7, 8, and 18 could not be studied because the sand fraction was too small to be sieved or for the heavy minerals to be separated. Samples 6 and 11 were omitted as they were representatives of only one sample taken from a single locality and thus could not be used for a local correlation of the red and blue clays. Sample 19 was also omitted because it consisted of a mixture of "red" and "blue" clays and would not show a true representation of either of the two. The remaining 13 samples were used both for the study of "red" and "blue" clays in general and for local correlation studies. The heavy mineral frequency distribution, both in #1 TABLE IV HEAVY MINERAL FREQUENCY DISTRIBUTIONS FUR RED AND BLUE CLAYS Total.Grains Counted Minerals Samples 1 2 3 hr 5 9 10* 12 13 1h 15* 16 17% Dolomite 75 150 66 hl 206 52 S 133 62 107 h? 153 17 Diopside 16 85 38 9 5h 22 l 31 16 63 19 76 13 Apatite 17 3h 13 12 an 55 S 7 h6 31 7 21 17 Siderite 21 o o o o 82 3o 18 lo 20 22 6b 1b Metallic 66 6 13 2h 5 9 3 Lb 32 19 26 1h 18 Glauconite 7 o o o o 2 35 h 70 65 S h 0 Clino-pyroxene 35 9 8 S 7 1 O 17 12 32 l7 19 10 Epidote 38 6 16 b 6 1h 2 1h 13 13 2h 8 8 Zircon O 22 35 7 10 26 1h 2 15 h 9 7 12 Tourmaline l3 3 S 1 6 1 o 7 5 ll 15 5 3 Cassiterite 3 h 13 0 6 O O 0 ll 13 O O 3 Monazite 8 2 h 2 2 2 o h 9 1h 7 3 2 Hypersthene 13 2 3 2 3 O O 8 ll 6 10 O l Calcite 11 3 o o 3 1 o 13 3 o 6 18 2 Rutile 2 S 13 6 3 2 o o 12 7 h o 3 Garnet. 13 S b 1 o 3 o 7 5 1 1 2 l Olivine 3 o o o o 1 o 9 o S o 12 1 Beryl o o o o o 15 o o 6 lo 6 o o Tremolite o o o o o o o o 12 15 S l o Staurolite o o 0 0 o o o o 11 16 3 0 2 Hornblende h S 3 2 h 1 O O S O h S O TOpaz 7 h o 2 12 o o 2 o o o o o Sillimanite o o o o o S o o h 6 h 2 o Erucite o h 1 o 11 o o o o o o o o Chlorite l 2 o 3 o o o o 7 o o o o Corundum o b o o 2 o o o o o o o o Enstatite o o o o o o o o o 3 o 1 o Titanite o o o o 1 o o o o o o l o GlauCOphane l o 2 o o l o o o o h o o Sphalerite o o o o o o o o o o 2 o o Leucoxene O O O O O O .O O O O 3 1 2 Total Grains 35h 355 237 121 385 29S 95 320 377 h6l 250 kl? 129 * Educ Clay samples 4&2 TABLE IV (Continued) Total Percentages Minerals ‘ Samples 1 2 3 h* S 9 10* 12 13 1b 15* 16 17* Dolomite 21% tzz 28% 3h% 5&3 18% 5% h2% 16% 23% 19% 37% 13% Diopside 5 2h 16 7 1h 7 1 10 h 1h 8 18 10 Apatite 5 lo 5 10 11 19 S 2 12 7 3 S 13 Siderite 6 28 32 6 2 h 9 15 11 Metallic l9 2 S 20 1 3 3 1h 8 h 10 3 1h Glauconite 2 - 37 1 19 1h 2 - Clino-pyroxene lo 3 3 h 2 - 5 3 7 7 h 8 Epidote 11 2 7 3 2 5 2 h 3 3 10 2 6 Zircon 6 15 6 3 9 15 - h - h 1 9 Tourmaline h l 2 - 2 - 2 l 2 6 l 2 Cassiterite l l 5 2 3 3 2 Monazite 2 ~ 2 1 - - 1 2 3 3 - 1 Hypersthene A ~ 1 1 — 3 3 l h - Calcite 3 1 — - h - 2 h l Rutile - l 5 5 — - 3 l l 2 Garnet h 1 2 - l 2 l - - - - Olivine — - 3 l 3 - Beryl 5 l 2 2 Tremolite 3 3 2 _ Staurolite 3 3 1 2 Hornblende 1 l l 2 1 ~ 1 l 1 Topaz 2 l 2 3 - Sillimanite l 1 1 2 - Brucite 1 - 3 Chlorite — — 2 2 Corundum 1 - Enstatite - - Titanite — - GlauCOphane — 1 - 1 Sphalerite _ Leucoxene 1 _ 1 Total 100 100 100 100 100 100 100 100 100 100 100 100 100 - equals less than 1 per cent #3 total numbers and percentages, of each of the 13 samples is given in Table IV. The statistical correlation of the results is presented in the following chapters. Roundness and Sphericity Analyses The light minerals taken from the sand fraction between sizes 80-100 and mounted in piperine were first subjected to microscopic study preliminary to the analyses for roundness and sphericity. Upon close inspection it was found that each slide re- presented such a heterogeneous mixture of shape and round- ness variations, typical of a glacial deposit, that it was decided that any further study would be highly inconclusive and of little value in correlation of the clay samples. Laboratory Equipment for the Nuneral Analysis Bromoform (sp. gr. 2.87) Glass funnels Rubber tubing Wire clamps Beakers--250 ml. Filter paper-medium size Glass stirring rods Small glass vials with caps Methyl alcohol Special funnel for heavy-liquid separation wash bottle for bromoform Glass slides and cover slips Forceps Piperine (mounting media) Kyol Bunsen burner and hot plate Cheese cloth Polarizing microscope with mechanical stage Set of slides--known heavy minerals Mt STATISTICAL ANALYSIS Introduction A thorough laboratory study of sediments includes quantitative data on the sizes, shapes, mineral composition, surface textures, and perhaps the orientation of the grains. These fundamental data are related to the physical and chemi- cal factors in the environment of deposition. To relate characteristics with environment one may investigate the areal variation of the sediment, which implies the compari- son of one sample with anOther. The comparison in this study was between samples of "red" clays with samples of ”blue" clays to determine whether any similarity exists between them. This comparison is most conveniently accomplished by means of statistical analysis. The word "statistics" is defined as: "The science of the collection and classification of facts on the basis of relative number or occurrence as a grounds for induction; systematic compilation of instances for the inference of general truth". (Webster, 1926) This definition reveals that the study of sediments is largely statistical in nature in that the various analyses give the facts and occurrences for the inference of general truths about the nature of the #5 L6 sediment. Statistical technique may be divided into several oper- ations. First, the collection and classification of data. In sedimentary terms this refers to mechanical analysis, mineral analysis, and so On. Second, the presentation of data in the form of tables and graphs. Finally, the data themselves may be analyzed statistically and from the values obtained inferences may be gathered concerning the sediments. As most sediments fall within the class of continuous distributions with a range of sizes from largest to smallest, the data in this study were converted into terms of a con- tinuous frequency distribution by means of the cumulative curve. Cumulative curves were drawn using the weight per- centages of size distribution as determined by pipette analyses. Cumulative Curve Analysis To obtain the geometrical ratios directly, cumulative curves representing the data determined by the pipette analysis were drawn on semilogarithmic paper. The dia- meters of the particles in millimeters were plotted loga- rithmically along the horizontal axis and the cumulated weight percentages were plotted arithmetically along the vertical axis. (Figures 1-19) The curve was drawn by 1+7 >h2300 mmhm_z<2 .mhrofik m_m>4130 m>C<43230 #0.... .12 z. (uhw!<_o Go... SAMPLE NUMBER 9 8 8 5 6 0 0 O O 0 0 O O O 6 9 3 9 7 9 5 8 0 6 I 2 2 5 O O I I 2 3 5 7 I 5 2 I 4 2 o O. O O. 0 O 0. O O. O O O. 0. O u «d 0 1 0 > 0 w u 0 0 t O Y.. — _ ‘0‘ _ . \c\ . _ \ o. _ _ \\\\o m . 2 \ 1sinI.—v:|:-l|.|.f-l11..lrlli\lllTllunll lllll ||I||.Ill+aw-uu F 3:530 ewe: _ _ \d on _ _ a - \Cm 09. b O _ 238.2 m ow Oh lllllll T Ill... IIJIII c Ill! Ilelllllnlu .llulllll Ill.)iuullnvlllllifmulfilfl m...C.mhzaoo .00l38 .00l95 m.szm .00276 .00390 .mhIoG; w_m>n_mDo m>_»<43230 N.@_..._ 0|560 O22 IO .03l20 04420 06250 mikmdno hme ZQEMZ 3:330; 0.5:» OBIV‘IHNHO 39'1N3083d A9 >hzaoo m_Nzwm .mhxgmk m_m>4¢30 m>_hkzaoo .00l38 .00I95 m_szm 00276 .00390 $5.105; m_m>4zno u>CkJD§DO 0I560 022l0 03l20 ¢.O.u 04420 6250 mq;mFZDOO w_Nme .95..on m_w>.._<2< mhhwntn. no m>m30 wits-5230 m.o_u .22 z. Kuhn—34.0 604 OOI38 .00I95 00276 .00390 00550 00780 0|I00 0I560 022l0 03l20 04420 06250 IIIIIII- )II ul I I I II I): fill IIIIIII I I I I I k mnfimqao 2.me 245m: wmflmquao-- owl-.1» '3ev1uaouad OBIVTDNOO 52 >h2300 DdZdJNNJ .mhrofii m_m>._mno wink-5.230 @ .@.u .82 z. Kuhw!<.0 004 SAMPLE NUMBER 9 9 8 8 5 6 O O O O O 0 O O O 4 6 9 3 9 7 9 5 8 O 6 I 2 2 5 0 O 0 I I 2 3 5 7 I 5 2 I 4 2 O O O O O O O O O I I 2 3 4 6 0 0 O O. O o O. O 0 O. 0 O. 0 O O ) ) I ) I y ) 2 r . - \o P O— O _ \ _ \OI ow ILIIlIkr.LfiLII|-IIIII III rllirIInII IIIII\ IIIIIIIIIIII IIIIILYIJlifiIIl-ITI “Li-5.430 emmi \\ on o\_ \ \ O? \ on \0 245m: \ . or \e \ S. v.4 n I.) L '1' ll. lllll III IIIJIIJTII IILI IllwIIIIIIJII IIIIIII III) «III VIIIIIIIII IIr-III \e\ immimqno emf-r \ II: I 00 \ om 4H 00. OBIV'IRNOO 39V1N3083d 53 7 SAMPLE NUMBER .00049 00069 >FZDOO 00098 mmmw>4IDo m>:<43530 h.o_u 0I560 .022IO 03I20 04420 06250 w4:m<30 kwmi 245w: '0' wmm mlqnoo I 0.5.: oawinuno '39V1N3083d 5h SAMPLE NUMBER .00049 00069 00098 >h2300 .00I38 um ¢m><¢ ... .00I95 .00276 ozn_130 w>_._.<43§:0 Q.0_u 0 I 560 022l0 .03l20 04420 06250 Mikadao Emmi z<_om2 wmmmquao -. 0.5:». OBLV'IDNOO 39V1N3083d 55 9 SAMPLE NUMBER .00049 .00069 .00098 .00I38 >k2300 .00I95 23.5.24 .00276 .00390 .mFIoGE w_m>1_mbo m>:.<.5230 0.6.... 0I560 022l0 .03I20 04420 06250 mgr—5.430 meC 2.30mi wmmmqnao- 0.55. '39v1uaou3d OBLVWI’INRO 56 I O SAMPLE NUMBER >FZDOO 2.xh2< .mFIom; m_m>.._IDO m>C.FZDOO .00l38 .00I95 !_mh2< 00276 .00390 .thoG? m_m>4zno m>E<43§30 :.@_u 0|560 022l0 03|20 04420 06250 NJCKQDO Emmi zaami wmmwwquao- 0.5:» O31V10N00 39V1N3983d 58 SAMPLE NUMBER .00049 00069 00098 >HZDOO .00 l 38 .00I95 mmhmid! .00276 .00 390 .mhIoG; m_m>4mjo m>_.r<.:.:230 0I560 022l0 .03I20 N707... 04420 6250 Nita/«DO Emmi zSomZ 372230. 0.3:» OELV'IONI'IO 39V1N3083d 59 I3 SAMPLE NUMBER .00049 00069 00098 >.._.z:oo .00l38 .00I95 Coumho .00276 .00390 .mhIoG; m_m>._m30 m>_h<43230 0|560 .022I0 .03I20 n_.o.u 04420 06 250 mikmdno hme ZSQMZ I'll wmmmquao- or?» 0. ON On 0? On 00 Oh 0m 0m 0. OBLV'II'INI‘IO 39V1N3083d >hzaoo oowmho 60 00049 00069 00098 .00 I 38 00I95 .00276 l4 SAMPLE NUMBER . E, .00390 .mhIoa? m_m>.._130 w>Chzaoo oommbo .mhrofik m_m>J130 m>C<43230 9.0.... .1: z. xuhuzfo co; 9 9 8 8 5 6 0 O O O 0 O O O O 4 6 9 3 9 a: 9 5 8 Au 6 I 04 2 5 O 0 O I I 2 3 5 7 I 5 2 I 4 2 0 O O O O O O O O I I 2 3 4 6 O. O. O O. O. O. O. O. O. O. 0 O. 0. O O I I I I I I I F 5 ’ .5 . l— — LO‘I‘O \1 O \1‘ p U\0 _ x . oA \ 2 a ILfi-IAIIAIrIIIITIIIIIIAIIIL TIILfiII IIIIII II IIIIIII III+ILI+IIIIVI wihmqno Emmi 24.0w: III 1 ll IITI v IL lll-l IC|I1 rIII m4; «Ian: 0 off. YI III I 0. ON on O? Om 00 ON GO om 00. OBLV'IONI'IO 39V1N3083d 62 SAMPLE NUMBER 000459 .00069 .00098 >hzaoo .00I38 .00I95 omouxw? .00276 .00390 .m._.10_w§ m_m>n_<2< .00550 cwhu2mDo m>C<43230 0I560 .022l0 .03I20 2.0:. O 5 2 6 I II I mikmdao meE 04420 zSowZ wmmmquao- or?» OSlV'IflNflO 39V1N3083d 63 I7 SAMPLE NUMBER .00049 00069 00098 .x Its-try. ".13.. III I >h2300 .00I38 I I II IIIIH: omouow .00I95 00276 .00390 .mhIoG; m_m>.._<2< cuhmida 004 .00550 00780 DIIOO whhma_a no m>m30 m>CdJDSDo hfo_u 0I560 022l0 03I20 04420 06250 mq;m<30 hwmi 245w: wmmmqnao- or? OBLVWDNRO 39V1N3983d 61; SAMPLE NUMBER .00049 00069 00098 rhzaoo .00I3B omOuow .00I95 .00276 .00390 .mhIomk w.m>._<2< cubm2<5 90... .00550 .00780 .0lI00 whhmnta no m>mno m>_h<43230 9.67... 0I560 022l0 03I20 04420 06250 mfhmdac hmmE 245m: wmmmquao- 0.3+: 031V1 nuno 39V1N3083d 65 SAMPLE NUMBER .00049 00069 00098 >h2300 .00I38 zomm10 .00I95 00276 .00390 .mhIon m_m>._m30 0I560 .022l0 03|20 m>;.N 40000. 0m400. 44H00. ©0.m~ m>.Nm mm.m *ma mw4. 00.N «@000. H4400. 44H00. m4.H~ om.0m Ho.> 4H mam. Hm.m moooo. oaeoo. omaoo. o4.- mo.ma mm.4 ma mag. mH.m mmooo. mmmoo. moaoo. 40.50 oo.o~ em.o NH «.3. tum 380. «38. #80. 8.2. 8.8 ohm fl Hem. mm.~ mmuoo. omNHo. assoc. mfl.mm om.mm 04.“ *OH «mm. oa.m omwoo. ommao. assoc. 04.4m om.~o om.m m 040. wm.4 05000. 04mH0. momoo. mw.4m om.m4 Hm.m *0 004. 4m.w 40000. mmmoo. omaoo. 4m.m> mm.4m mw.m w 0mm. mm.m «@000. m4m00. 0Hm00. m0.mm mm.~m wm.m 0 cap. ma.m mmooo. omeao. mamoo. mm.ms mm.~m mm.ma m mom. so.m «coco. cameo. ommoo. m~.mo 4m.4m mH.m is mam. m~.m aoooo. ammoo. oumoo. om.mo ~4.sm mm.“ m mmm. mo.m eoooo. assoc. cameo. so.me mm.mn ms.m N Ham. Hm.m Hmooo. omwoo. 0MH00. m>.u4 «H.0m maomm H om mag om \ao \ma mes a mean u aafim u eeum uoaaaam. > 338 mmquéd Hannah 20mm mZOHBEBAa MAHBQB 9:. gdgmmm 7O Coefficient of Correlation To determine the possibility of relationships between the heavy mineral suites of each sample a statistical corre- lation was first applied to the data from each heavy mineral count. L. Dryden (1935) first made use of the "coefficient of correlation" in an attempt to state the degree of simi- larity involved between heavy mineral suites of various samples. "The term "correlation" must not be confused with correlation in the geological sense. Statistical correlation implies no time relationship nor any other causal relation. It only states objectively similarity or mathematical dependence of one set of data upon another set." (Krumbein, 1938) The coefficient of correlation, r, for which the formula is given below, gives a means of determining simi- larity between two sets of data numerically: I‘ = Sum (XV) " n MX MY y(Sum X2 - nMx2)( Sum Y2 - nMyz) Sum - summation n - number of mineral species MI- the mean of a set of data X - frequency data for first sample Y - frequency data for second sample On the recommendation of Eisenhart (1935), the use of true frequencies rather than percentages was made in order to reduce the error in correlation. If the correlation coefficient is equal to positive one, there is a direct relation between the variables; if its value is negative one, there is an opposite relation 71 between the variables; and if the coefficient is zero, there is in all likelihood no fixed relation between them. For values other than zero, but neither positive one nor nega- tive one, the significance of the correlation coefficient depends partly on the nature of the data being examined. Using Dryden's "coefficient of determination", which is equal to the square of the "coefficient of correlation", we obtain a simple percentage as an eXpression of the pro- portion of elements common to the two samples or suites. (Dryden, 1935) Table V. Chi-Square Test The use of the coefficient of correlation for the above purposes, however, has been objected to by Eisenhart (1935) and the chi-square test has been proposed in its stead by many statisticians. (Baten, 1955) As a means of comparing the similarity of the mineral suites of the various samples by the coefficient of correlation with the proportions of each mineral represented in the various suites the chi- square test was applied to the same mineral suites. In applying the chi-square test a table is set up showing the two samples in vertical columns, and the attri- butes to be tested in horizontal rows. Each observed frequency is subtracted from an expected frequency, and the difference is squared and divided by the expected frequency. 72 A series of values are obtained, the sum of which equals the chi-square value. The formula for chi-square is: Chi-Square 3 Sum (Observed value - EXpected value)2 Expected value The observed value of chi-square is located in a table of values opposite an appropriate number for the "degrees of freedom” in the problem, and the probability desired is found. Details of the test and an introduction to the theory are given in Eisenhart's paper, 1935. A chi-square table can be found in C. E. Clark's book: "An Introduction to Statistics", 1953. Chi-Square Versus Coefficient of Correlation Caution must be exercised against indiscriminate appli- cation of any statistical measure of correlation or home- geneity to data collected without a full understanding of the underlying requirements which must be met before the par- ticular statistical device under consideration can be applied. At the present time the use of the coefficient of corre- lation in geologic methods is a controversial one, with the substitution of the chi-square test advocated by many statis- ticians. The writer has merely attempted in this study to apply both methods in reference to heavy mineral analyses and to compare both results with those obtained from the 73 quartile deviation studies from the pipette analyses for each of the samples. The results of these methods, their comparison, and interpretation are taken up in the following chapter. COMPARISON AND STATISTICAL INTERPRETATION OF DATA Introduction Having defined the various quartile terms applied to sedimentation in the preceding chapter and presented the statistical methods and formulae used in interpretating the data, the results obtained in the laboratory investigation can now be compared and interpretations made. General and specific relationships existing between the "blue” and "red" clays were determined from results obtained through methods involving size distribution by the pipette analysis and heavy mineral studies. The total results of the analyses of the 12 samples of red clay were compared with the total results of the six samples of blue clay for general correlation purposes. The results obtained from the study of samples taken at the same locality but at different depths were compared for local relationships. Analysis of Size Distribution The data summarized in Table II were employed in pre- paring the cumulative curves (Figs. 1-19) and the statisti- cal values presented in Table V. From.Table V the values of 74 75 the median size, Md¢; sorting, Sb; and Log of So were comp pared for general and local similarities that might exist between the red and blue clays studied. Considering the medians for all the samples, it is found that red clays range in size from 0.00105 to 0.00770 mm. while the blue clays range from 0.00144 to 0.00670 mm. in size. Although the range of medians was slightly greater in the red samples the difference was not of significance to in- dicate any great dissimilarity between the red and blue clays in general. Essentially, this slight difference may be the result of the variation in the number of red samples as compared with blue samples studied. Thus a wider range of sample medians is more highly probable within the red clays. According to the wentworth classification this range in median size would fall between very fine silt and mediumr size clay fractions. For local comparison four localities were considered where red and blue samples occurred in the same exposure but at different depths. The results are as follows: Samples Medians 3 and h, Benzie County 0.00220 mm. red 0.00230 mm. blue 9 and 10, Antrim County 0.00770 mm. red 0.00670 mm. blue 1L and 15, Otsego County 0.0014h mm. red 0.0014h mm. blue 16 and 17, Wexford County 0.00220 mm. red 0.00223 mm. blue 76 Comparison of size medians showed no significant differences between the red and blue clays in any of the above localities. It was noted, however, that in one locality there was a nota- ble difference between medians of the red and blue clays as indicated below. Samples Medians 7 and 8, Grand Traverse 0.00180 mm. red County 0.00365 mm. blue This difference between the red and blue clays can not be explained from.the laboratory evidence alone. The values for sorting, So, of the red and blue clays were compared as to general relationships in a manner similar to the one used in comparing medians. Sorting varied from 2.10 to 5.13 in the red samples and from 2.35 to 4.37 in the blue samples. Again, the variation was greater in the red clay probably for the same reason as previously stated for variation in the medians. 0n the basis of Trask's classifi- cation (1930) the sorting values obtained vary from that of a well sorted sediment (less than 2.5) to one of a poorly sorted sediment (greater than h.5). No significant differ- ence in sorting between the red and blue clays was evidenced. Local comparisons in sorting, So, for the four locali- ties considered previously are as follows: Samples Sorting, So 3 and h, Benzie County 3.79 red 3.67 blue 9 and 10, Antrim.County 2.10 red 2.35 blue 77 lb and 15, Otsego County 2.66 red 2.76 blue 16 and 17, Wexford County 3.73 red 3.54 blue Comparison of sorting in the red and blue clays showed no significant differences in any of the above samples from the same locality. It was noted, however, that there was a wide difference in sorting between the red and blue samples of the Grand Traverse county locality as shown below. Samples Sorting, SO 7 and 8, Grand Traverse 2.94 red County b.37 blue Comparing the logs of sorting, log So, for samples 7 and 8 (Table V; red-O.h68, blue-0.640), by calculation it was found that the red sample was 0.640/0.A68 = 1.3 times better sorted than the underlying blue clay. This may be an in- dication of later modification subsequent to deposition resulting in a re-sorting of the upper clay deposits. In like manner, the log of sorting can be used to determine the relative degree of sorting between any of the samples listed in Table V. The red samples range from 0.322 to 0.710 in sorting, or 0.710/0.322 = 2.20 times the vari- ation in sorting. Therefore, the best sorted sample is 2.2 times better sorted than the poorest sorted sample of the red clays. The blue samples range from 0.371 to 0.6h0 in sorting, or 0.6h0/0.371 = 1.72 times the variation in sort- ing. 78 Again the results show a greater variation in the red samples, but no significant difference between the red and blue clays studied. Percentages of sand, silt and clay for the red and blue samples were compared in similar manner. The red clays, in general, revealed a higher percentage of sand and a lower percentage of silt than the blue clays, but the percentage of clay was nearly equal in the two. (Table V) Clays Average Percentage Sand Silt Clay Red clays 7.65% 31.74% 60.61% Blue clays 2.73 38.84 58.43 For local comparisons the and clays were as follows: percentages of sand, silt Samples Percentage Sand Silt Clay 3, red 2.53% 34.17% 63.30% 4, blue 2.18 34.54 63.28 9, red 3.30 62.30 34.40 10, blue 2.46 58.36 39.18 14, red 7.61 20.96 34.40 15, blue 3.55 22.79 73.66 16, red 8.20 35033 56eh7 17, blue 2.90 31.47 65.63 In general, the percentages compared were nearly similar far each locality, but two major variations were noted. The percentage of sand was higher in the red samples and the percentage of clay was larger in the blue samples. This may 79 be an indication of later re-sorting of the red deposits, with a removal of the finer clay particles, thus increasing the concentration of sand grains present. The possibility that a higher percentage of sand may originally have been deposited with the red clays must not be overlooked, however. 0f the nineteen samples analyzed for size distribution all but two were found to contain a higher percentage of clay than silt. The exceptions were samples 9 and 10 from the same exposure in Antrim county, but at different depths. In comparing these two with each other there was no signifi- cant difference in the size analyses. However, when com- parisons were made between these two samples and all the other samples studied the results varied greatly from the general trend. This can readily be seen by comparing the cumulative curves (Figs. 9 and 10) with the cumulative curves of the other samples studied. The divergence in these two samples is highly indica- tive of somewhat different conditions of deposition or later modification in the area which the other sixteen samples, or localities, did not experience. This may be suggestive for further field investigation in the area. Grain sizes larger than 0.5 mm. in diameter are listed in Table III for 16 of the 19 samples studied. Three of the samples had no material larger than 0.5 mm. In many of the samples only one or two grains represented the largest size found in the sample. The possibility of any 30 gram sample, 80 split from the original field sample, containing the largest material in the entire sample is very small. 0n the other hand, the possibility of the sample never containing any of the larger pebbles is very high. Thus very little value was accredited to these data. Therefore, a broad and very general conclusion was drawn from Table III. 0f the red and blue clays studied the red samples seem to have the larger particles. The larger number of red samples studied as come pared to the number of blue samples must be taken into con- sideration, however. Summary of Size Analysis In summary, the general and specific comparisons of medians, sorting and the logs of sorting from.the data of size analyses between the red and blue clays indicate no significant difference. There were a few exceptions to this general trend. In light of the small number of samples studied and the nature of the scattered sample-locations the proba- bility of more exceptions occurring than was found is much greater. A comparison of percentages of sand, silt, and clay contained in the red and blue samples would seem to indicate a greater amount of coarser material in the red clays. This difference in the clays studied might be determined by future field investigations or a more thorough knowledge of the glacial history of the areas studied. 81 Analysis of Heavy Mineral Distribution The material summarized in Table IV was used in the cal- culations for the statistical interpretation of the heavy mineral analyses. The values for the ”coefficient of cor- relation" and the "chi-square" test are listed in Table VI. Both general and local correlations again were made between the red and blue clays to determine possible relationships. For a general correlation of the red and blue clays the total results of the heavy mineral analyses for each were compared. In other words, total counts were made for each of the minerals present in all nine slides studied for the red clay. These, in turn, were compared with the total counts made for each of the minerals represented in the four slides studied for the blue clays. The value obtained for the coefficient of correlation (r), using the formula on page 70, was 0.87. Using the value of 1.0 as reflecting a per- fect correlation, a value of 0.87 shows a direct relation between the heavy mineral suites of the red and blue clays studied. Due to the larger number of red samples used in com- parison to the number of blue samples in the above correla- tion, another coefficient of correlation value was calculated by using an equal number of heavy mineral suites from the two sets of clays. Samples from the four localities which were selected for comparing size analyses were used here to represent red and blue samples from the same exposures. The TABLE VI STATISTICAL CORRLLATION VALUmS FOR hnAVY'MINbRAL.ANALXSbS Samples Samples r r2 Chi-square X Y ' test All "red" All "blue" 0.8? 76% negative clays clays Red Blue # 3 # 4 13 15 0.91 83% negative l6 l7 9 10 Local Correlations Red Blue # 3 # 4 0.79 62% positive #13 #14 . 0.70 49% negative #16 #17 0.17 30% negative Areal Correlations #1 #3 0.57 33% negative #3 #14 0.71 50% negative #4 #15 0.57 33% negative #1 #12 0.84 71% negative #1 #16 0.5 7 33% negative # 14 #9 0.44 19% negative #3 #9 0.42 17% negative r equals the Coefficient of Correlation r2 equals the Coefficient of Determination 83 heavy mineral suites were totalled for both the red and the blue clays, respectively, and the coefficient of correlation thus calculated. The value determined in this case was 0.91 which indicates a greater relationship between the heavy mineral suites of the red and blue clays than the preceding value of 0.87. By squaring these two values the "coefficient of deter- mination" was resolved. (Table VI) In the case of the nine red samples and four blue samples correlated, 76 per cent of the minerals in the red clays were common to those found in the blue clays. Whereas, in the case of the four localities compared, 83 per cent of the minerals in the red clays were common to those in the blue clays. The results from the correlation of these two situations statistically are highly indicative of a direct relationship of mineral suites between the red and blue clays studied, with a very high percentage of the minerals common to each of the clays. Local correlations between red and blue clays from.the same exposures are as follows: Samples Ocefficient of Correlation 3 and 4 0.79 13 and 15 0.70 16 and 17 0.17 Samples 3 and 4, and 13 and 15, show a close relationship between the mineral suites taken from the same eXposure but 84 at different depths. Samples 16 and 17, with a value of nearly zero, show no fixed relationship between the heavy mineral suites from the same exposure. This poor correlation may be due to location. Both samples were taken from an older morainic system, the Lake Border moraine, consequently different conditions of sedimentation, alteration, and weathering were experienced as compared to conditions with- in the younger morainic systems. This value of 0.17 was the lowest obtained and no negative values were calculated for any of the relationships between the heavy mineral suites far the red and blue clays studied. Various localities were compared with each other for possible areal relationships (Table VI). Samples 1 and 12, both red clays from manistee county have a high correlation of 0.84 indicating a direct relationship of heavy mineral suites. Samples 1 and 3, both red clays from the core of the manistee moraine, with a correlation value of 0.57 ShOW'a possible relationship of heavy mineral suites of the two samples. Samples 3 and 14, both red clays from.the manistee and Port Huron moraines, respectively, have a high cor- relation of 0.71, indicating a direct relationship of heavy mineral suites. The remainder of the correlations between samples of the various localities in the different morainic systems indicate a highly possible relationship of the heavy mineral suites between: 1) the red and blue clays, 2) the red clays 85 in different localities and 3) the blue clays in different localities. The chi-square method of statistical analysis was ap- plied in each of the preceding correlations to determine the proportions of each of the minerals represented in the differ- ent samples. With the exception of one locality all the results show no relations in the proportions of heavy miner- als found in each of the samples. Several personal observations by the writer during the Study of the heavy mineral suites which were related to the results of the chi-square test are mentioned here. It was found in nearly all cases that the blue clays had a higher percentage of metallic, ferro-magnesium, and ferrous iron minerals present than did the red clays. If, as reasoned by some, the red clays are a weathered portion of the blue this may explain the increase with depth of these more easily weathered minerals. Summary of Heavy.Minera1 Analyses The comparisons of heavy mineral suites of the samples studied by the coefficient of correlation can be summarized as follows: 1. There appears to be a direct relation, or similarity, between the heavy mineral suites of the red and blue clays studied. 2. There may be a possibility of some relation between the heavy mineral suites of the red and blue clays over the entire area included in this study. 86 The chi-square test shows no significant relations in the proportions of heavy minerals represented in each of the samples studied. It is questionable whether the chi-square test is useful in a study of this nature. There is a higher percentage of metallic, ferro-mag- nesium, and ferrous iron minerals present in the blue than in the red samples, although the number of slides were some- what more limited for the study of the blue clays. Summary Comparison The laboratory results obtained and statistical cor- relations presented in this chapter are considered by the writer as general evidence of no significant differences between the red and blue clays used in the study. It is believed that the relationships existing between the size analyses of the samples are more valuable than those of the heavy mineral analyses. The results of the size analyses would tend to indicate similar conditions of deposition of the sediments studied. Because of the characteristic heter- ogeneity of glacial deposits and the large area over which the samples were collected, the variations in the heavy mineral suites and the relationships presented were con- sidered of less value in the correlation of the two clays than the size analyses. GLAC IAL INTERPRETATION From the laboratory investigation it may be possible to explain the different relationships between the red and blue clays. One may infer from a study of the size distribution in the two sets of clay that both were deposited under the same or similar conditions. If the sediments were laid down under conditions of continuous deposition, disregarding color, a single nearly homogeneous deposit would result. If, then, this area were subjected either to submergence or a high water table after glacial retreat the younger sediments, or those in the upper-most portions of the deposit, would ex- perience re-sorting and possible removal of some of the finer materials. Subsequent emergence and lowering of the water table would result in the upper near-surface portions being exposed and subsequent weathering taking place. The smaller percentage of ferro-magnesium and metallic minerals present in the red clays seems indicative of a higher degree of weathering which has resulted in the removal or alteration of these minerals. Thus the red, near-surface clays would re- present the weathered portion of the deposit lying above the water table in the zone of weathering. The lower limits or contact of the red with the blue clays would mark the lowest 87 88 position of the water table in the area at that time. If these clays were deposited at the time of the retreat of the ice during the last of the Cary substage a readvancement of the ice during the following, or Mhnkato substage, would cover these deposits with a younger mantle, resulting in a readjustment of the water table. Thus the buried red and blue clays would become the core material of the morainic system. If this interpretation of the laboratory findings is valid then the core material of both the Manistee and Port Huron moraines may be similar in composition and of the same age. If so, the clay core material would be interpreted as Cary and the mantle of younger drift as Mankato in age. CONCLUSION From a study of the size distribution and heavy miner- al analyses of the red and blue clays and the interpretations of the data, it may be concluded that the characteristics of a sediment as disclosed by laboratory investigation can be determined and compared by the use of sedimentary, petro- graphic and statistical methods of analyses. It should be noted, however, that the interpretation of data and the re- lationship existing between the various statistical methods of analysis are the most important factors in determining the characteristics of and correlations between the two clays. Although the methods used in obtaining statistical results of size distributions are well known, the choice of statistical methods for heavy mineral correlation is highly controversial. Information regarding application of sta- tistical methods, interpretation of the results of any of the methods used, and the relationships existing between the results is highly lacking in the research in various fields of geology. From.the laboratory investigation and interpretations of the data, it was concluded that no significant differ- ences exist between the red and blue clays studied. 89 90 This relationship existing between the red and blue clays may aid in supporting the opinion of some glacial geologists that: 1) the clay cores of both the Manistee and Port Huron moraines are similar and Cary in age, and 2) that each core is covered with a younger glacial drift, the mankato, whose terminal position is marked by the Port Huron moraine. The writer believes that with a systematic plan of sampling and a greater number of samples analyzed by sedi- mentary, petrographic, and statistical methods a very thorough and conclusive study of the blue and red clays of the northwestern portion of the Southern Peninsula of Michigan could be made which would bring the present contro- versial issue of correlation of these clays to a satisfactory conclusion. SUGGESTIONS FOR FURTHER STUDY For a complete analysis of the relationships existing between the red and blue clays of the northwestern portion of the Southern Peninsula of Ndchigan the following suggestions are made: 1. 2. Possible 1. 2. A systematic plan of sampling the drift from each morainic system and intermorainic area. Nbre samples collected from each glacial feature and an equal representation of samples for both the red and blue clays to be studied. field investigations for the solutions to: Why the two samples from Antrim county, red and blue clays at different depths, vary so highly in degree of sorting and size distri- bution from.the other samples studied. Why the two samples in Grand Traverse county from the same eXposure in the manistee moraine, are notably different from the other red and blue samples in the same moraine both in sorting and size distribution. 91 BIBLIOGRAPHY Allen, P., and P. S. welder (1945). Accuracy of Mineral Frequency Analysis of Sediments, Nature, vol. 155, F85 e 1945 e Baten, W., (1955), Oral communication. Bergquist, S. G. (1952), Pleistocene Complex in the Northern Rim of the Southern Peninsula of Michi an, IAbstr.), GedI. Soc. Am. Bull., vol. 63, pp 12 - 1334. Bouyoucos, G. J. (1934). A Comparison Between the Pipette Method and the Hydrometer Method for making MeEh- anical Analyses of SOiIs, Soil Science, vol. 38, I10. 5’ 19310-0 Bouyoucos, G. J. (1936), Directions for makinngechanical Analyses of Soils by the Hydrometer Method,‘SoII SCience, vOI._42, no. 3, Sept. 1936. Bradfield, R. (1931), The Colloidal Chemistry of Soils, Colloidal ChemistryL Theoretical’and Applied, vol. 3, The Chemical CataloglCompany, 193I. Bretz, J. Harlen (1951), The Stages ofLake Chicago: Their Causes and Correlations, Am. Jour. of ci., vol. 249, June, 1951, pp. 401-429. Dryden, Lincoln (1935), A Statistical Method for the Comparison of Heavy Mineral Suites, Am. Jour. of Sci., vol} 29, may, 1935. Eisenhart, Churchill (1935), A Note on WA Statistical Method for the Com arison of Heavy.Mineral SuItes", Am. Jour. of ScI., vol. 30, Dec., 1935. Eisenhart, Churchill (1935), A Test for the Significance of Lithological variatIons, Jour. of Sed. Petrology, V0. , 3,PP¢ - O 92 93 Flint, Richard Foster (1953). Probable Wisconsin Substages and Late-Wisconsin Events in Northeastern United States and Southeastern Canada, Bull. Geol. Soc. of Am., vol. 64, 1953. Galliher, E. W. (1933), Factors in Sedimentation Analysis, Am. Jour. of Sci., vol. 26, Dec., 1933, pp. 564-568. Hellman, N. N., and V. E. NbKelvey (1941), A.Hydrometer- Plpette Method for Mechanical Analysis, Jour. of Sed. Petrology, vel. 11, no. I: 1941. Hills, R. c. (1934). A Comparison of Some Methods Used for - the Fineness Test of Sands and'Clayp, Am. Foundry- men's Ass., Reprint no. 34-23,“I934. Hough, J. L. (1953), Pleistocene Chronology of the Great Lakes Re ion, Office of Naval Research, Univ. 6?- IIIInoIs, Jan.. 1953. 108 pages. Kropschot, R. E. (1953). A Quantitative Sedimentary Anal sis 2£_the Mississippian Deposits inlthe Michigan has n, Unpublished master'sThesIs, Dept. of_§eology and Geography, Mich. State College. Kruger, Fredrick C. (1937). A Sedimentary and Petrographic Study of Certain Glacial Drifts of Minnesota, Am. Jour. of Sci., vel. 34.5N0v., 1937. Krumbein, W. C. (1932), The mechanical_Analysis of Fine- Grained Sediments, Sedimentary Petrology, Dec., 1932. Krumbein, W. C. (1933). Textural and Litholo ical Variations in Glacial Till, Jour. of Geol., vol. XII, no. 4, May. i933- Krumbein, W. C. (1935), A.Time Chart for Mechanical Analyses by the Pipette Method, Jour. Sed. Petrology, vol. 5. n0. 2, Ange, 1935. Krumbein, W. C., and F. J. Pettijohn (1938), Manual of Sedimentary Petrography, Appleton Century Crofts, N0 YO, 938. . Krumbein, W. C., and W. C. Rasmussen (1941), The Probable Error of Sampling Beach Sand for_g§gyy:EI§§EEI—_" ‘Analysis, Sedimentary Petrology, April, 1941. 94 Krumbein, W. C., and L. L. Sloss (1951), Stratigraphy and ‘Sedimentation, W. H. Freeman and Co., 1951. Leverett, F., and F. Taylor (1915), The Pleistocene of Indiana and NHchigan, U.S.G.S..Monograph, vol. 53, 1915. . murray, R. C. (1953). The Petrology of the Cary and Valders Tills of Northeastern Wisconsin, Am. Jour. of Sci., v61. 251,Feb.1953o O'Hara, N. W. (1954), A Statistical and Mechanical Analysis of the Marshall Sandstone in Western Michigan to Determine the Environmental Pattern of the Deposit, Unpublished Master's’Thesis, Dept. of Geology and’ Geography, Mich. State College. Olmstead, L. B., Alexander, and Middleton (1930), A Pipette method of Mechanical Analysis of Soils based on 'Im roved"DIs erSIon Procedure, U’.S. Dept‘. Agric. Tech. BfiIl. no. I70, I930. Otto, George H. (1939),A AMbdifiedL Logarithmic Probability Graph for the Interpretation of Mechanical Analyses of SedIMents, Sedimentary Petrology, Aug., 1939. Rittenhouse, G. (1939), The Pipette Method Modified for Mass Production Nat. Research Council, 'Div. of Geol. and Geog., wash., D'.C., 1939. Shepps, V. C. (1953), Correlation of the Tills of North- eastern Ohio by Size AnalysIs, Sedimentary Petrology, Mar., 1953. Sommers, R. E. (1919), MHcroscopic Examination of Cla s, Jour. of the wash. Ace. of Sci., vol. 9, no. 5,1919. Trask, P.D (1930), mechanical Analysis of Sediments by Centrifuge, Eco. Geol., vol. 25, no. 6, Sept. 1930. Twenhofel, W. H. (1950), Principles of Sedimentation, mcGraweHill, N.Y., 1950. Walstrum. E. E. (1947) I neous Minerals and Rocks John Wiley and Sons,’N.Y., 1947. , \ Websterigzgew International Dictionary, Springfield, mass., I. " ‘ Iv . . 1 | 2 95‘ 1.2.1- a- l...- “'J be 1. A ‘-«.- "IIIIIIIIIIIIIIIIIIIIII