Ali—F l . A. i. I"! jun-m U“. ... ‘ ‘31--L'fif'.“lAhp-.I ‘5', o ‘ Fin. 1‘. up A “A a ‘f . A ' t . , _. . _ mlv'a-xrfir-m'zwvr r ..r Ni':'lnp_¢r'—¢..-1— -_"?"" ‘1‘... 1:3... "I '. wru "’11-‘23; “ " ."c : . . -. ; . 9'5"“ a-3_ 1 ‘wmrn- ' g. '.. - A PETROGRAPHIC INVESTIGATION OF THE RELATIONSHIP OF DEPOSITION OF SEDIMENTS IN A GROUP OF ESKERS RELATED TO THE CHARLOTTE TILL PLAIN by George Theodore Schmitt - A Thesis Submitted to the School of Graduate Studies of Michigan State College of Agriculture and Applied Science in partial fullfillment of the requirements for the degree of MASTER OF SCIENCE Department of Geology and Geography 1949 *‘HESIS ACKNOWLEDGEMENTS The writer wishes to eXpress his sincere appreci- ation to Dr. B. T. Sandefur for his helpful suggestions and criticisms in the direction of the problem. The writer also wishes to express his deep thanks to Dr. S. G. BerEQuist and Dr. J. W. Trow for proof read- ing the manuscript and to Dr. F. W. Foster for his aid in the preparation of the maps and graphs used in the thesis. ii LIST OF TABLES . . . . . . . LIST OF ILLUSTRATIONS. . . . INTRODUCTION . . . . . . . . LOCATION OF AREA . . . . . . General . . . . . . . Charlotte Morainic System CONTENTS Distribution of Eskers Charlotte Esker. . mason Esker. . . . Williamston.Esker. Webberville Esker Howell Esker . . . PURPOSE OF STUDY . . . . . . METHOD OF INVESTIGATION . . General. . . . . . . Field Sampling . . . Laboratory Sampling Leaching . . . . . . Sieving. . . . . . . Separation . . . . . mounting for Microscopic Study Identification of Heavy Minerals STATISTICAL METHODS OF CORRELATION . . Quartile Measures_of the Cumulative Weight PercentagGS. . . . . iii Page vi 0\ O\ oooooo-q 10 10 11 ll 12 12 12 14 15 15 18 20 33 33 Page Comparison of Heavy Mineral Suites . . . . . 40 Sphericity and Roundness Measurements . . . 45 CONCLUSIONS . . . . . . . . . . . . . . . . . . . . 49 BIBLIOGRAPHY . . . . . . . . . . . . . . . . . . . . 51 iv Table 1. LIST OF TABLES Frequency Percentage of Heavy Minerals, 65-100 Sieve Range 0 o o o o o o o o 0 Frequency Percentage of Heavy Minerals, 100—150 Sieve Range . . . . . . . . . Frequency Percentage of Heavy Minerals, 150-200 Sieve Range . . . . . . . . . Sorting, Skewness and Kurtosis . . . . Comparison of the Heavy Mineral Suites of the Williamston Esker with the mason Esker by the Use of the "Coefficient of Determination'. . Page 27 28 29 4O #4 Figure l. 10. ll. 12. 13. LIST OF ILLUSTRATIONS Some Eskers on the Charlotte Till Plain, LanairLg-Howell “ea 0 O O O O O O O O O O O O 0 Systematic Procedure for Simultaneous Analysis of Sediments for Shape Analysis and Heavy Mineral Identification . . . . . . . . . . . . Weight Percent Analysis of Sand . . . . . . . . Apparatus for Heavy-liquid Separation . . . . . Frequency Percentage of Heavy Ennerals, 65—100 Sieve Range 0 C O O O O O O O O O C C O O O C 0 Frequency Percentage of Heavy Minerals, 100-150 Sieve Range . . . . . . . . . . . . . . Frequency Percentage of Heavy Minerals, 150-200 Sieve Range 0 o o o o o o o o o o o o 0 Cumulative Curve of the Weight Percentage for the Williamston Esker . . . . . . . . . . . . . Cumulative Curve of the‘Weight Percentage for the Webberville Esker . . . . . . . . . . . . . Cumulative Curve of the Weight Percentage for the Howell Baker 0 O O O O O O O O O O O O O 0 Cumulative Curve of the Weight Percentage for the Charlotte Esker . . . . . Comparison of the Heavy Mineral Suites of the Williamston, Webberville, Howell and Charlotte Eskers with the mason Esker by Means of the "Coefficient of Determination . . . . . . . . . Sphericity and Roundness Measurements for the Sieve Ranges, 48-65 and 65-100 . . . vi Page 13 17. 19 30 31 32 34 35 36 37 INTRODUCTION The esker is an interesting glacial deposit in the form of a long sinuous ridge and composed chiefly of strat- ified drift. Flint“ *Flint, R. F., Eskers and Crevasse Fillings: American Lournal 9;; Science, 5th Series, Vol. 15, 1928, pp. 411- 412. has written a summary of the external features, composition, and structure of eskers: Position 1. Commonly bear a definite relationship to terminal or recessional moraines, lying within and tributary to them. External Features 2. May be very long, length commonly several miles. More than one individual in Maine has a length of 100 to 150 miles. Eskers in eastern Connecticut vary from 12 to 45 feet in height . . . Usually . . . . . trend in a direction parallel to the direction of ice move— ment. ' Commonly broken and discontinuous, so that the entire length of any one esker must be measured by restoring the portions between the isolated remnants . . Commonly . . . . . sinuous in plan, suggesting the winding of streams. Arranged, not infrequently, in tributary and distributary systems, like those characteristic of normal streams. Trend uphill and cross divides, in many cases with no change in volume or texture of deposit. Crests in many instances knobby and hummocky or gently undulatory; rarely level for any long distance. Elevations of crests bear no definite relation to the tOps of surrounding forms. Composition and Structure 10. Component material predominantly coarse. . - . Fine sand and clay are rare. 11. Bedding exceedingly variable and irreg- ular, with common deveIOpment of semi- stratified to non-stratified lenticular masses. . . l2. Transverse sections of every esker suitably exposed in eastern Connecticut, as well as sections of scores of eskers described from other localities, invar- iably exhibit irregular bedding parallel- ing the side slopes of the esker . . . . Crosby* ”Crosby, w. 0., The Origin of Eskers: 1§ostog Society of Natural History; Proceedings, VOl. 30, 1893, pp. 375- 411.'1'L' ' ’ “ Observed that in areas of the earth's surface now covered by glaciers, no conditions exist that are similar to those of the Pleistocene Ice Age. The glaciers of today are primarily of the valley and piedmont types with subglacial drainage. Ridges definitely identified as eskers are not found which have formed recently in the areas of these glaciers. The eskers identified with Pleistocene time were formed on country which had been nearly peneplained by glaciation. .Crosbyfn *Croaby, W. O.._op. cit., pp. 375-411. has proposed a superglacial theory for the origin of eskers. A.large amount of englacial material was eXposed on th e surface of the glacier by melting of ti e tOp of 3 the ice sheet. The base level Of the superglacial streams was determined by the water table within the ice, produced in many cases by a lake barrier formed at the edge of the glacier. This may have caused the base level of the super- glacial streams to be above the base of the ice. Crosby states:* “Crosby, W. 0., 0p. cit., p. 393. "In view of these considerations, there :seems to be no escape from the conclusion that the ice floor of the superglacial stream will be lowered by the super- glacial as well as the basal melting of the ice; that the superglacial melting will be more active and efficient in pro- portion to the extent of aggrading of the channel and the volume of stagnant water saturating the gravel; and that the stream will be lowered at least as rapidly as the interstream surfaces . . . " This theory may explain the manner in which the eskers con- form to the topography on which they lie. Upham” *Upham, w., Evidences of the Derivation of the Kames, Eskers, and Moraines of the NOrth American Ice Sheet Chiefly from its Englacial Drift: gulletin Q: the Geologgpal Sogigty of America, Vol. 5, 1894, pp. 71-86. suggested that eskers were derived chiefly from englacial and finally superglacial drift. This material was gathered by the surface melt water and deposited in ice walled channels as eskers, or at the mouths of glacial rivers as 4 kames. In the vicinity of Winnipeg, hanitoba, which is the center of a vast flat region Upham observed that englacial drift was borne to heights of 500 feet above the surround- ing plain. The subglacial theory as postulated by Davis presents a the origin of eskers as follows: *Davis, W. M., The Subglacial Origin of Certain.Eskers: ggstgn Society of Natural Histor 5 Proceedi s, V01. 25, 1893, PP o _477"499 o A sand plateau, which formed in a glacial lake or above the water table at the edge of the ice sheet, served as an outlet for the esker stream. The building of the sand plateau caused the esker stream to become choked, and aggrad- ing took place in the tunnel or crevasse. Davis does not adequately explain the origin of the tunnel or crevasse in which the esker stream flowed. In a study of certain eskers in Denmark, Anderson? *Anderson, S.A., The waning of the Last Continental Glacier in Denmark as Illustrated by verved Clay and Eskers: The Journal of Geolorr , vel. 39, No. 7, 1931, pp. 609-624. " i i I poStulated that anycrevasses formed during an interval of ice stagnation would not be further compressed for the reason that the flow had ceased. These crevasses acted as tunnels 5 or channels in which the melt water escaped to the ice border. During the stagnant stage a water table which sloped in the direction of flow, deveIOped within the ice, and lowered each time the englacial melt water found a lower escape. This water table represented the ultimate upper limit of de- position of the glacio-fluvial material. When the surface of the ice melted until coincident with this table there was a tendency for the whole border zone of the glacier to be buoyed up. With the lifting of the ice there would be no limit to the width of the tunnels developed beneath it. Trowbridge* *Trowbridge, A. D., The Formation of Eskers: Science, New Series, Vol. 145, 1914, p. 145, abstract. believes that eskers were formed bythe drawing out of kames into long sinuous ridges during the recession of the glacier. A.ridge Of this type was built by the deposition of succes- sive segments, each segment marked by a delta where the esker stream entered a glacial lake. There is a disagreement among geologists relative to the origin of the drift contained in eskers. Trefethenf *Trefethen, J. M., and H. 3.. Lithology of the Kennebec valley Esker: .American Journal o§_Sqience, VOl. 242, 1944, _pp. 5212527. - , 3' ' shOwed that the bulk Of both the fine and coarse material comprising the Kennebec esker of Maine is dominantly of local 6 -origin. This is evidenced by the fact that at various points along the ridge the material consists largely of small frag- ments derived from the bed rock over which the ice moved. LOCATION 93 £1.14 General The Labradorean ice sheet was responsible for the development of two large lobes to the west of the Appalachian mountains, namely: the Lake Michigan and Huroanrie lobes.* ”Leverett, F., The Pleistocene of Indiana and Nuchigan: United States_§eological Survex, Monograph 53, 1915, p. 24. Between these 10063 was developed the smaller Saginaw tongue which occupied a large area in central Enchigan and extended south into Indiana and Ohio. The features produced by these two lobes are considerably different. The Lake Michigan lobe, whose axis was deeply keeled in a preglacial valley, deveIOped a series of similar bulky, but rather smooth morainic ridges as it retreatedfrom Illinois. The moraines of the Huroanrie lobe lack the strong bold ridges of the Michigan lobe. I The Lake Michigan lobe deployed southward through the Lake Michigan basin into central Indiana and east-central Illinois. The Huron-Erie lobe extended across northeastern Ohio into Indiana. »The Saginaw tongue advanced southwestward from Saginaw Bay into northeastern Indiana. The three lobes were coalesced untilthe Saginaw tongue receded from the district south of Kankakee, Illinois. Its recession was more rapid than that of the adjacent lobes because the original 7 advancement was over more elevated country producing a thinner ice sheet. The Saginaw tongue developed the following glacial deposits in order of decreasing age: maxinkuckee, Bremen, New Paris, Middlebury, Lagrange, Sturgis, and Tekonsha moraines and the Kalamazoo and Charlotte morainic systems. Charlotte,Moraig;g_Systeg . I The ice was nearly stagnant during the formation of the Charlotte morainic system, indicated by the large number of eskers on the till plain. Towards the east, in the south- eastern part of Livingston and the southwestern part of Oakland counties this system connects with the Fort wayne morainic system of the HuroneErie lobe. In the west, through western Kent county, it connects with the Valparaiso morainic system of the Lake Michigan lobe. The eskers on the till plain ex- tend southward into the Charlotte morainic system. Leverett says:* *Leverett, F., op. cit., p. 206. "This morainic system, includes perhaps, .more small eskers than any other in Michigan, and it receives the southern terminal of some of the most conspicu- ous eskers of the state. Some of the esker ridges have a nucleus of gravel and sand with a thin capping of till, a feature which suggests that they were formed near the base of the ice sheet at a horizon low enough to permit the deposition of the englacial material on them. Distribution 9; Eskegs 4 Ingham and Livingston counties are covered principally by the Charlotte till plain. This area is traversed by several eskers, which for the most part lie in shallow swampy depressions or esker troughs. The larger eskers extend southward over it, into the outer border of the Charlotte morainic system. Figure 1 shows the distribution of the eskers in the Lansing-Howell area. Charlotte Esker ' - The Charlotte esker heads in Section three, Benton Township, on the north side of Thornapple River and extends south-southwesterly for a distance of nine miles. It termine ates in a fan shaped delta directly east of Charlotte. In some localities there are wide gaps in the esker, but in other places it is continuous for over one mile. The maximum relief of the esker is 20 feet above the till plain; the width varying from 270 to 300 feet. Except for two or three miles at the northern end it crosses the Charlotte morainic system. The esker is composed of medium-coarse gravel. ' lasssfislssr The mason esker is the longest esker in Michigan. The entire esker lies in a‘well defined trough 20 miles long, which heads in southeastern Lansing and terminates in a delta plain southeast of mason. The course of the stream, as shown by the attitude of the bedding, was from north to south, which is reversed to the direction of the present drainage. The ridge varies from 10 to 50 feet in relief, and from 150 in o .3 i'llsfo '26:}: I\\:1.u 3283 (Us ju’ozngazj 22.: 44.... NPPOJcsso “ab .8 §u _ _ £1.80m—u 0:004 u _ _ _ _ _ _ n . _ in _ i... “ ......... o _ I ‘ I... _ \ .38.. .\ :nln \ r _ a x . z \ A _ \ .J i... in i .x ,. u . o a . . J _ .. .K 5.23:; r _ 2.08‘4 L _ zo»mez.>_n_ a2.87 S.G.<2.87 S.G.>2.87 S.G.<2.87 S.G.>2.87 S.Gv<2.87 S.G.>2.87 Slides for Slides for Slides for Slides for Slides for shape shape heavy mineral heavy mineral heavy mineral analysis analysis identification identification identification Fig. 2 14 Origin. Unpublished Master's thesis. Department of Geology and Geography, Michigan State College, 1948, pp. 25- 27. showed that the heavy mineral suites and shape measurements remain nearly the same throughout the length of the mason esker, one of the eskers on the Charlotte till plain. A vertical channel sample, of approximately five gallons, was taken from each exposure, and before collecting the material all slumpwas carefully removed. This material was thoroughly mixed, and with the aid of a Jones splitter was reduced until one quart of the original sample remained. Labggatory Sampligg ' Since the-composition of the eskers is predominatly sand and gravel it was not necessary to disaggregate the material. One half of the field sample was taken for labor- atory analysis and the remainder retained for reference. The very fine muds were removed by washing and decanting. All material smaller than 200 mesh was discarded. The mix- ture was decanted carefully to insure no loss of the parti- cles to be used in the analysis. Krumbein* *Krumbein, W. 0., Manual of Sedimentary Petrography, DwAppleton-Century Company, New York, 1939, p. 110. has shown that, according to Stoke's law, particles one sixteenth millimeter in diameter have a settling velocity of .305 centimeters per second. One minute was ample time for all the particles used in the analysis to settle to the 15 bottom before decanting. Leaching The unweighed sample was dried in a Sargent's electric drying oven, divided into eight parts, to each of which dilute HCl (10% solution) was added. They were then allowed to stand in the acid for several days, with frequent stirrings, until the carbonates had been completely dis- solved, after which the sample was thoroughly washed and dried. Erickson” *Erickson, R. L., op. cit., p. 20. showed there was no correlation in the percentage of carbon- ates in each sample. Sieving . The pebbles were removed with the aid of a 10 mesh screen. The sample was reduced with a Jones splitter until 100 grams of the sample remained. The 100 gram sample was sieved in a Ro-Tap automatic shaking machine for a period of eight minutes. The shaker was equipped with six sieves having 35, 48, 65, 100, 150, and 200 Openings per inch. Krumbein states* *Krumbein,‘W..C., op. cit., pp. 118-119. that there were disadvantages in the use of sieves for the mechanical analysis of sediments. Sieves not only sort grains according to size, but also according to shape. A 16 lath-shaped grain, with the same cross section as a spheri- cal grain, will pass through a given sieve. Despite this disadvantage, sieving is used widely in mechanical analysis, and the data obtained may be employed for a number of pur- poses. After sieving, each size was weighed, placed in a bottle and labeled. Figure 3 shows the weight percentage of the sands from the various eskers. Separation Before proceeding with the separation of the light from the heavy minerals, the sample for each size was quartered. This was done by placing four pieces of rectang- ular paper in such a way that each overlapped one half of the other forming a square. The sample was carefully poured on the center of the square, and alternate quarters were pulled out and combined. This process was continued until the desired volume of the sample was obtained. One gram samples from the 48-65 and 65-100 sizes, and one half gram samples from the smaller sizes were used. These weights for each sieve size used in the analysis were suf- ficient to give a representative fraction for the heavy and light materials. The heavy liquid, bromoform (tribro-methane) with a specific gravity of 2.87 at room temperature, was used for the separation of the minerals. This liquid separated the minerals having a specific gravity less than 2.87 from those having a specific gravity greater than 2.87. The apparatus a; 18 used for all separations is shown in Figure 4. The sand was poured into the upper funnel, which was half filled with bromoform, and thoroughly stirred. The heavy minerals sank and were concentrated above the pinch cock. After complete separation the heavy minerals were drawn off by Opening the pinch cock, allowing them to collect on the filter paper of the second funnel. After all the liquid had passed through the filter paper was removed. The bromoform in which the lighter minerals were floating was run off and the lights were allowed to collect on a second filter paper. The bromo- form was returned to the original bottle after it had been collected in the lower bottle. Both the heavies and lights were rinsed with alcohol and these washings were poured into a bottle marked “bromoform washings", and the bromoform was recovered by washing. Twenty-four separations were carried out simultan- eously in a funnel-battery apparatus. Care was taken, how- ever, to mark the individual filter papers in order to avoid confusion. The fractions were then dried, weighed, put in separate vials and labeled. MOunting £2; Microscopic §§ggy . I'The light fractions were used for Sphericity and roundness measurements. The refractive index of quartz is 1.544-1.553, whereas for Canada balsam it is 1.537. Because the quartz shows little relief in Canada balsam, the grains were mounted in a synthetic resin (n—l.66) in order to show greater relief and detail. To reduce the tendency for the 19 l L W 7 7 \_ \_ 0 man. mm 'I‘fll nun-m om“ 1'0 scum M III" I.“ IL s" i 2O formation of bubbles, the resin was melted in an evaporating dish. The slide on which the quartz grains were mounted was also heated and a small amount of the resin was poured on to the slide. Any bubbles produced in the resin remained in the evaporating dish when the resin was poured on to the slide. - The heavy minerals were mounted in Canada balsam which is common practice.” *Pettijohn, F. J., manual of Sedimentary Petrography, DaAppleton~Century Cogpany, New York, 1938, p. 360._ Identification 93 Heavy Minerals The petrographic microscOpe is.an instrument which aids in the identification of heavy minerals present in sands. .A full description of the microscope can be found in one of several works on microscopy.“ *wahlstrom, E. R., Optical Crystallography, John Wiley and Sons Inc., New YOrk, 1943. *Chamot, E. M. and mason, C. W}, Iandbook of Chemical MicrosOOpy, Vol. 1, Principles of Microscopes and Acces- sories, 1st ed., John'Wiley and Sons, Inc., New York, 1931. *Johannsen, A., Manual of Petrographic methods, 2nd ed., Qnivegsitv qnghica Q Presg, Chicago, 1918. -‘.A mechanical stage, fastened to the revolving stage, ~greatly facilitated the counting of the heavy minerals. With the mechanical stage the object slide may be moved in two directions at right angles, the component of each 21 movement indicated on the scales. Certain minerals, because of slight alteration, had prOperties which made them diffi- cult to identify in the mounted slide. When this was the case, it was necessary to remove the grains of this unknown mineral from the unmounted sand, immerse in index oils, and make accurate indentification. Three sizes, namely: 65-100, 100-150, and 150-200, were used in the identification of the heavy minerals. Sizes larger than 65-100 contained an appreciable amount of rock fragments which could not be used satisfactorily in the statistical analysis. Twenty three minerals were identified from samples taken from the eskers studied. Following is a list of the most important minerals with descriptions as compiled from several sources:* *Dana, E. 8., Descriptive Mineralogy, John'Wiley and Sons, Inc., New Kerk, 191 . “Johannsen, A“, Essentials for the Microscopical Determina- tion of Rock-forming Minerals and Rocks, University of Chicago Press, Chicago, 1914. *Krumbein,'w. 0., Op. cit., pp. 414 ff. *Milner, H. 3., Sedimentary Petrography, ghomas Murby and .93., New York, 1929, pp. 97 ff. Hornblende:l COmplex silicate of Fe, mg, Ca, Al, and Na. Crystal system: monOclinic - Color var. common hornblende, green to brown Index a, 1.658-l.698; b, l.670-l.719; c, 1.679-l.722 22 Birefringence : .026-.027 Optic figure : biaxial negative Elongation : positive Pleochroism : marked Distinctive features: grains elongate; prismatic; Inclined extinction; marked pleochroism; common. Clincpyroxqgg: (Augite, diallage, diopside) ""' Ca (Fe, mg, A1) (3103)} Crystal system: monoclinic Color : brownish-gray to gray-green Index : a, l.696-l.700; b, l.702-l.718; c, l.7l4-l.742 Birefringence : .018-.043 Optic figure : biaxial positive Extinction : 44-49 degrees Distinctive features: grains usually elongate; worn cleavage fragments; poorly rounded; high index; high birefringence; large extinction angle. Garnet: R"R"' (510) where R" is Mg, Fe", Ca, Mn . . and R"' is Al, Fe" , Cr Crystal system: isometric Color : pink and colorless Index : 1.70-1.90 Distinctive features: high relief; isotropism; conchoidal fracture. Chloritic matteg: Essentially silicates of Al, Fe, mg, and hydroxyl Crystal system: monoclinic Color : dirty green Birefringence : .003-.009 Distinctive features: "ultra-blue" abnormal interfer- ence color; compound polarization; pale green color; low birefringence. Zircon: ZrSiO4 _Crystal system: tetragonal Color : colorless ' Index : e 1.985-1.99l; o, 1.926-1.936 Birefringence : .6443 23 Optic figure : uniaxial positive Elongation : positive Extinction : parallel Distinctive features: euhedra common; pyramidal term- inations; basal grains rare; rod-shaped inclusions; high index. Hypersthene: (L , Fe) 3103 Crystal system: orthorhombic Color : pale pink and green Index : a, l.665-1.715; b, l.669-1.728; c, 1.674-l.731 Birefringence : .009-.016 Optic figure : biaxial negatiy Elongation : positive ‘ Pleochroism : marked, X-pink, Y-yellow, Z-green Distinctive features: worn elongate cleavage fragments; highly colored; thin, brown, plate like inclusions (Schiller structure); low birefringence; parallel ex- tinction; striking pleochroism. OrthOpvroxene: (enstatite, bronzite) (Mg, Fe)SiO3 Description same as hypersthene except that enstatite is nearly colorless and is biaxial positive. Enstatite, bronzite, and hypersthene are members of an isomorphous series in the orthorhombic pyroxene group. Staurolite: 2 FeO.5A1203.4SiOQ.H20 Crystal system: orthorhombic Color : yellow, brown Index : a, 1.736; b, 1.741; c, 1.746 Birefringence : .010 Optic figure : biaxial negative Pleochroism : marked, X-colorless, Y-pale yellow, Z-golden yellow Distinctive features: marked by hackly to sub-conchoidal fracture; well formed crystals rare; inclusions numerous; exhibits bright interference colors. Epidote: Ca2(Al, Fe)3813012(0H) Crystal system: monoclinic Color bottle green Index a, l.722-1.729; b, l.742-1.753; c, 1.750-l.780 24 Birefringence: .O28-.051 Optic figure biaxial negative Pleochroism distinct, X-colorless, Y-bottle-green Z-colorless Extinction : 2-5 degrees Distinctive features: grains equidimensional; sub- rounded; distinct pleochroism; bottle green color; high index. Titanite: CflO.T1023102 Crystal system: monoclinic ' Color : pale'yellow, light brown Index : a, 1.900; b, 1.907; c, 2.034 Birefringence : .134 , Optic figure : biaxial positive Elongation : negative Pleochroism : weak Extinction : 51 degrees Distinctive features: conchoidal fracture; diamond shaped euhedral grains; exhibit same color under crossed nicols as in ordinary light owing to high birefringence; many grains fail to show complete extinction in white light due to high dispersion; the grain changes to a blue color as the extinction position is reached. Sillimanite: A1203.5102 Crystal system: orthorhombic Color : colorless ' ' Index : a, 1.659; b, 1.660; c, 1.680 Birefringence : . 1 Optic figure : biaxial positive Extinction : parallel Elongation : positive Distinctive features: grains irregular to short pris- matic; marked by longitudinal splitting and striae parallel to length. Tepaz: 2(A1F)O.5102 Crystal system: orthorhombic Color : colorless Index : a, 1.619; b, 1.620; c, 1.627 Birefringence : . 08 Optic figure biaxial positive 25 Distinctive features: irregular fractured grains; basal grains give well-centered interference figure; interference colors bright. Tourmaline: (Na,Ca)R?(Al,Fe)633816027(O,OH,F)4 with R Mg, Fe ', Fe"',Al,Li,Mn and Cr Crystal system: hexagonal-rhombohedral Color yellow-brown, dark brown Index : e, l.62l-l.658; O, 1.636-1.698 Birefringence : .Ol9-.032 Optic figure : uniaxial negative Pleochroism : strong, dark brown to honey yellow Extinction : parallel Elongation : negative Distinctive features: characterized by color and strong pleochroism; negative uniaxial figure. Monazite: (Ce,La,Nd,Pr)2O3.P205 Crystal system: monoclinic Color yellow ' ' Index a, l.786-l;800; b, 1.788-l.801; c, 1.837-1.849 Birefringence : .049-.051 Optic figure : biaxial positive Pleochroism : faint, X-light yellow, Y-dark yellow, Z-greenish Extinction : 2-10 degrees Cleavage : perfect basal Distinctive features: grains rounded; equidimensional often lying on 001; euhedra rare; exhibit same color between crossed nicols as in ordinary light owing to high birefringence; high relief; light yellow color. Leucoxene: Decomposition product of ilmenite, as yet ill- _ defined. Crystal system: non-crystalline Color : translucent to opaque; white to light yellow under reflected light. Distinctive features: commonly occurs as rounded grains; Opaque in transmitted light, white or yellowish- white in reflected light. 26 Less than one percent Of the grains counted were olivine, biotite, apatite, anatase, and rutile. The pre- dominant grain counted was a composite aggregate which was difficult to determine because of its lack Of Optical pro- perties. Milner describes a composite aggregate as:* *Milner, H. B.,Op. cit., p. 100. “The attachment Of iron-ore to quartz, mica to quartz, rutile to ilmenite, pyrite to chart, and such compound minerals as leucoxene, perthitic inter- growths Of feldspar, mica- chlorite-serpentine aggregates, shimmer aggregates, etc., are common occurrences in sediments, whose diagnosis "may occasionally be troublesome. The results Of the heavy mineral counts for the three sieve sizes are presented in tables 1, 2, and 3. Histograms showing the frequency percentage Of the heavy minerals are shown in figures 5, 6, and 7. TABLE 1 Frequency Percentage of Heavy Minerals 65-100 Range 27 Esker Mineral mason‘William- Webber- Howell Char- . ston ville lptte Composite aggregate 27.4 26.8 20.0 18.0 23.8 Chloritic matter 23.2 19.6 22.1 24.0 22.8 HOrnblende (green) 11.2 14.0 16.9 12.3 21.0 ClinOpyroxene 8.0 6.0 8.1 8.0 3.8 Epidote 6:5 7:1 7.4 5.7 6.8 Garnet (colorless) 4.0 8.5 5.7 7.0 2.8 Garnet (pink) 6.5 3.2 4.3 9.3 3.8 Orthopyroxene' 4.0 5.6 4.3 4.3 2.0 Hornblende (brown) 1.4 1.8 4.8 4.3 4.3 Hypersthene ‘ 2:9 1.1 3.3 3.3 1.5 Sillimanite 0.0 0.0 0.8 2.4 0.8 Mbnazite 1.0 1.3 0.8 0.6 1.3 Staurolite ..O.8. 0.7 0.0 0.6 2:3 Leucoxene 1.9 1.1 0.5 0.0 0.8 All others 1:0 3.3 1.0 0.2 2.2 TABLE 2 Frequency Percentage Of Heavy Minerals 100-150 Range 28 Esker Mineral mason ‘William- Webber- Howell Char- ston ville lotte Composite aggregate 23.0 29.3 26.1 23.2 ~ 33.6 Hornblende (green) 14.4 10.2 24.4 18.6 21.0 Chloritic matter 8:8 9.1 6.4 8.5 11.2 ClinOpyroxene 12.5 7.9 7.2 6.4 5.1 Epidote 5.5 6.6 4.6 7.7 8.1 Garnet (colorless) 6:6 7.7 6.3 8.3 1.1 Orthopyroxene 5.9 4.3 7.8 3.8 3.8 Garnet (pink) 4.4 4.1 3.6 7.6 1.3 Hornblende (brown) 4.1 6.8 4.4 2.2 3.3 Monazite‘ 2.6 3:2 4.1 4.9 1.8 Hypersthene 5.5 2.3 2.5 3.4 1.8 Staurolite 2.2 0.7 0.0 0.0 1.5 Zircon 2.2 0.7 0.0 0.0 0.6 Topaz 0:0 1.1 1.2 2.7 2.- Tourmaline 0.0 1.3 1.0 0.9 1.5 ,All othere' 2:2 4.7 0.4 1.5 2.1 TABLE 3 Frequency Percentage of Heavy Minerals 150-200 Range 29 Esker Mineral Mason ‘William- Iebber- Howell Char- , ston ville lotte Hornblende (green) 19.5 19.7 20.0 14.6 22.0 Composite aggregate 19.2 18.6 17.4 16.4 19.0 Garnet (colorless) 5.9 9.0 11.1 14.7 4.3 Hornblende (brown) 6.1 2.3 11.1 9.7 9.5 ClinOpyroxene 7.5 6.4 8.1 4.5 10.2 Orthopyroxene 6.1 5.1 5.1 5.3 6.7 Epidote 9.3 7.2 4.0 8.2 7.8 Chloritic matter 5:3 7.5 576 4.7 4.8 Menazite 5:9 5:5 4.5 2.7 1.9 Zircon 4:3 3.6 2.8 6.8 0.6 Garnet (pink) 2.4 3.0 1.9 6.0 2.2 Hypersthene 2.1 1.8 0.5 _1.7 1.9 Leucoxene 1.6 3.5 1:1 1.6 2.8 Topaz _ 0:0 -3.5 2.3 1.0 0.0 Titanite _ 3:2 0.0 0.0 0.0 0.9 Staurolite 1.1 0.0 0.7 0.0 2.2 ,All others 0.5 2.3 3.7 2.1 3.2 k“ _._.-.. rumour? M-_.__- *fl v.—- . . - .. ..—- .—» - . - . . o \ _._.____ “nay a, _. _. _ . - _ *- _._._ ~- __._ _ -1 . _ - - . flaw- - . ~--_._._.H_.._._._. , __ - _-7 _ _ _~ ,,_.la__--a- .— y - - , _. _.v - ._.- 1- *_._.._—._.—- g 7- ,- , g _ --‘.__-__.-_.__.—.f - co——o—————._.__... - - . _.._*_.__._ u ..-._..__—._—.74- - m24_-_._.i . .- o . +4._._.-_.—.. -7 - , -,_.__.7*_.-_4_--._-- . *- _-._*_.._._- -- «-4 -._- - W__.._. -_.- ._ . wHw— 4 - an... ---_...--_._._ . _. _, _. _._.- - c u . H.— o , --.. H4_.._. .__._ . ‘ . _.__- -_ H-_.—.__ o 4—. -_.*_._oa._-e - M- .——--._ -7 - ' K -._. ka-.. .._-. - .__.1 ¢ . ---—.__—o-——--—-. *-_._,._..-. . .._--_._..__.-- - _.._.._. - o . 11,-._. - ._,- _- . a . ._- -. 4 u , , . . . - W} O—vo- ,. 7 i f D * .. .-_ ,_.___.-_. 4-- 1-_..___.__.__._ 1_._, »« n » . . . v a —._. . l -_+..._._ . __._-, -_. .._*_. -_*v-!.—-—o—.._.—-.. . -.._. _._a.._.w o 4- . H.“ -1- , . A _ , , ,_ k,._-‘ H- _._--- .._q_.._.._.__.__. .o '_‘ . , -!+-~— . . . . . _._._ .7-__.._ -.. -o— -.-.——. - ~Q—5 >- . . _ -7- _.._.1-_- 1 _.__._-. .__,.__. -1- . ,, . .—7-—.—. o—.- . , | --» . . -—.—. .4__.. 1‘ . a - *._.—. - . -7. - . ”1-1 .__*-- .v- ‘ . -_._+41.--_._._-_._ - +_._._.1-..__._- .- .1 . - _. --- - . .v. ._. ._._~_.--, .1_._.~_‘.-._._._ . -_-, -__._. ._. H - 1H- -_._.._- -__._._._.._._._. I . . .7- - 1 - . . . —~--. ._.-.- - . - R.a.._-_....7 . STATISTICAL METHODS Q: CORRELATION Quartile Eggggggg g: the Cumulative Weight Percentages Quartile measurements may be used to describe sedi- ments statistically. These measurements are confined to the central 80 percent of the frequency distribution curve and therefore the values are not influenced by extreme particle sizes. Three expressions Of quartile measures can be used to describe a sediment. They are quartile devia- tion, quartile skewness, and quartile kurtosis. Each de- scribes a characteristic Of the sediment which may be used as a comparison between samples. _ The median, first and second quartile, and the tenth and ninetieth percentiles are used in the computations Of these measurements. Each is read directly from the cumula- tive curves Of the weight percentages as shown in figures 8, 9, 10, and 11. The median, M, is the point on the curve where 50 percent Of the particles have larger and 50 percent of the particles have smaller diameters. The first quartile, 01, is the diameter which has 25 percent of the distribution smaller and 75 Percent larger than itself. The third quar- tile, Q3’ has 25 percent of the distribution larger and 75 percent smaller than itself. The tenth and ninetieth per- centiles correspond to the ten and ninety percent frequency lines. Quartile deviation? *Krumbein,'w. 0., Op. cit., p. 230. 39: access-am 22.3 3.3.250 In mm. Diameter at the Weight Percentage tor the Curve Fig.8. Cumulative Esker Williameton 35 35:69.... 22.3 $222.50 Diameter in mm. Fig.9.0umulative Curve of the Weight Percentage for the Esker Webberviile Frequency Weight Cumulated ' Diameter in mm. FigJO. Cumulative Curve of the Weight Percentage tor the Howell Esker 36 3:263... 2...; 2.2.50 Diameter in mm. Weight Percentage tor the Esker 'Fqul. Cumulative Curve of the Charlotte 38 is the measure of the average spread which is commonly used with the median. The geometric quartile and log quartile deviations were used in this analysis. The geo- * metric quartile deviation was introduced by Trask *Trask, P. D., Origin and Environment of Source Sediments of Petroleum, Houston, Texas, 1932, pp. 67 ff. and is expressed by the formula: QDg . V 93ml It may also be described as the "sorting coefficient", in! So . The log quartile deviation is shown by the follow- ing formula:* *Krumbein, w. 0., op. cit., p. 231. 10g dog . 10g SO . (10g Q3-log Q1) 2 The geometric quartile deviation eliminates the size factor and the units Of measurement. Because the values of 'So' are geometric rather than arithmetic it can not be said that an 'SO' value Of 3.0 is half as well sorted as another sediment with an ’50' value.0f 1.5. The logs of the 'So' may be compared directly because they form an arithmetic series. Quartile skewness is the departure of the arith- metic mean of the quartile diameter from the median. In addition to an arithmetic eXpression it may be described geometrically and logarithmically. The arithmetic 39 expression 13:“ *Krumbein,'W. 0., op. cit., p. 235. Ska . 3(01+a3-2ml If the value of the skewness is negative the skewness Of the curve is shifted to the left, and conversely if the value is positive the skewness is shifted to the right Of the median. If the curve is perfectly symmetrical the skewness value will be zero. The geometric form is as follows:* *Krumbein, W. 0., Op. cit., p. 235. 8kg . 33%: L When the curve is symmetrical the value of skewness is one. Quartile kurtosis is the degree Of peakedness of the frequency curve, and involves a comparison Of the central portion Of the curve to the spread Of the curve as a whole. It is described by Kelly's equation:* fxeiiy, T. L., Statistical Methods, London, 1924, p. 77. an 2 Ql-Q 2(P9O-P10) As the curve becomes more peaked the value for kurtosis approaches zero. Table 4 is a summary Of the results Of the quartile measures. TABLE 4 Sorting, Skewness and Kurtosis Eskers So log So Ska Skg an Williamston 1.234 .0945 -.007 .953 .279 'Webberville 1.282 .1089 -.006 .948 .300 Howell 1.190 .0766 -.007 .960 .278 Charlotte 1.450 .1622 -.002 .925 '.312 Comparison _o_f Heavy Mineral Suites It is rather difficult to determine the degree Of similarity between two heavy mineral suites in which a cor— relation is desired. A statistical method for the compari- son Of heavy mineral suites was used by Dryden,“ ”Dryden, L., A Statistical Method for the Comparison of Heavy Mineral Suites; American Journal of Science, Vol. 29, Noe 173’ 1935’ ppe 393‘408e For the value Obtained he used the term "coefficient of correlation". The word "correlation" must not be confused with the term as used geologically. Pettijohn states:* *Pettijohn, F. J., op. cit., p. 487. "The term 'correlation' must not be 'confused with correlation in the geological sense. Statistical cor- relation 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." 41 Dryden uses the following formula to determine the "coefficient of correlation". i( xx) -nI-lxMy flame—mag [tune-mg EE-»sign.for the summation of that to which prefixed. n - number of pairs of percentages to be used, in this work, the number of mineral species. M - mean of that to which it is prefixed. X - any percentage from the Mason esker. Y - corresponding percentage, 1.6., of the same mineral Species, from the sample being compared to the Mason esker. I‘ Dryden states:* *Dryden, L., 0p. cit., p. 399. "By using r2 instead of r we can “finally drop the word 'correlation' in a statistical sense, but, more important, we now get a simple per- centage as an expression of the prOportion of elements common to the two samples or suites." The "coefficient of determination" was computed by comparing the heavy mineral suites of the mason esker with those of the Williamston, Webberville, Howell, and Charlotte eskers. The sieve sizes 65-100, 100-150, and 150-200 were used. These values were plotted on a graph, shown in figure 12, with r2 as the y-axis and each esker as the x-axis. The value of the heavy mineral suite for the ‘Milliamston esker as compared to that of the mason esker is .882, showing that thepercentage of elements of the Williamston esker not common to the Mason esker is .118, by 42 simple subtraction. Table 5 shows a sample calculation of the "coefficient of determination". The reliability Of any statistical study increases with an increase in the number of observations. In this work only five samples were used, but the study shows the percentage Of elements from the four eskers common to the Mason esker which is not very easily shown by histo- grams. The final result is a number, but with histograms the comparison of the various samples is shown by a picture which in some cases is misleading. of Determination Coefficient William - Webber- Char- Howell eton ville lotte Fig.l2. Comparison of the Heavy Mineral Suites of the Williameton, Wehherville, Charlotte, and Howell Eekere with the Moeon Esker by meane of the “Coefficient of Determination". 43 TABLE 5 Comparison of the Heavy Mineral Suite of the Williamston esker wi h the Mason esker by use of the "Coefficient of Determination" for 100-150 Sieve Size ibson.Esker Williamston Esker Minerals X X? Y Y2 XX Green hornblende 19.5 380 19.7 388 384 Composite aggregate 19.2 368 18.6 346 357 Garnet (colorless) 5.9 35 9.0 81 53 Brown hornblende 6.1 37 2.3 5 14 Epidote 9.3 86 7.2 52 67 ClinOpyroxene ' 7.5 56 6.4 41 48 OrthOpyroxene 6.1 37 5.1 26 31 Chloritic matter 5.3 28 7.5 56 40 Monazite 5.9 35 5.5 30 32 Zircon ‘ ' 4.3 18 3.6 13 15 Garnet (pink) 2.4 6 3.0 9 7 Leucoxene ‘ 1.6' 3' 3.5 12 6 Hypersthene ' 2.1 4' 1,8 3 4 Topaz ‘ ‘ ‘ 0:0 0 3.5 12 0 Staurolite 1.1 1 0.0 0 0 Titanite 3,2 10 0.0 O 0 Totals 99.5 1104 97.7 1074 1058 ..22.5,- x _ 97.7 - . PIX 16 _ 6.2 11y _ l _ 6.1 mix. 16(6 .2) g 615 IlMy = 16(6 .1) = 596 ma,a, . l6(6.2)(6.1) = 604 r - 1058-604 - .939 r2=(.939)2 = .882 V71104-6lsi-(1074-596) 45 Sphericity gpd Roundness Measurements The shape is one of the fundamental prOperties of sand grains and is the most recent to be studied quantita- tively and statistically. There are many geologic factors involved in the development of the shape of a particle. Pettijohn lists several of these factors:* *Pettijohn, F. J., Op. cit., p. 278. The original shape of the fragment. The structure of the fragment. The durability of the material. The nature and violence of the action to which the fragment is subject. , 5. The time or distance through which the action is extended. PUMP O 0. According to wadell * *wadell, H., Volume, Shape and Roundness of Quartz Particles, Journal 2; Geology, Vol. 43, 1935, pp. 250-280. roundness is themeasure of the angularity of the corners, and sphericity is the ratio between the length and breadth of grains. wadell uses the following formula to compute the sphericity of quartz grains: ¢= d° Dc 0 - degree of sphericity dc - diameter of a circle equal in area to the area of the grain obtained by planimeter measurement Dc - diameter of the smallest circle circum- scribing the particle 46 This method of measuring sphericity is accurate but very time-consuming because each individual grain must be drawn before any measurements can be made, Rittenhousefi *Rittenhouse, c., A.Visual method of Estimating Two Dimensional Sphericity, Journal of Sgdimentarv Petrology, V01. 13, No. 2, 1943, abstract, pp. 79-81. suggested a visual means Of measuring Sphericity which is more rapid but not as accurate as the method proposed by wadell. Riley* *Riiey, N. A., Projection Sphericity, Journal of Sedimentary Petrology, Vol. 11, 1941, pp. 94-97. devised a method which is both accurate and rapid. A con- centric circle protractor is placed in the occular of the microscOpe. Measurements may be read directly, thus elim- inating the drawing of individual grains. He used the formula: ¢ - 1 Dc 1 - diameter of the largest inscribed circle of the sand grain. Dc- diameter of the smallest inscribed circle. This Sphericity approaches the accuracy of the value Obtained from wadell's formula. 47 For the measurement Of roundness Wadell uses the formula: (g) in P:—————— N total degree of roundness. radius of curvature of the corner. - radius of the maximum inscribed circle. - number of corners measured. 2:0’1‘0 The individual grains must be projected and drawn with the aid of a camera lucida before the roundness measurements can be made. Combining Riley's method for measuring sphericity with Wadell's method for measuring roundness an accurate and rapid method was devised by the writer for the measure- ment of these values simultaneously. A camera lucida attached to the occular of the microscOpe projected the grains on Wadell's concentric protractor which had been drawn on a large sheet of white paper rather than plexiglass. The radius of curvature of all the corners and the diameter of the inscribed and circumscribed circles were measured simu- ltaneously. .The results of the shape measurements are shown in figure 13. 49 CONCLUSIONS During the formation of the Charlotte morainic system the Saginaw tongue was stagnant over Livingston, Ingham, Eaton and Barry counties. The eskers associated with the Charlotte till plain were developed at that time. Figure 3, the weight percentage of each sieve size from all the eskers, shows a similar distribution of sands from one esker to another. The sphericity of the particles is nearly constant for all the eskers, but there is a slight variation in the roundness measurements which does not appear to be significant since the material comprising the eskers was not tranSported the same distance in each case. Little variation of the heavy mineral suites is shown in Figures 5, 6, and 7. The comparison of the heavy mineral suites of the various eskers to the Mason esker by the use of Dryden's fcoefficient of determination” depicts a high percentage of elements common to the Mason esker. The final result is a number which gives a clearer picture of the similarity than do the histograms. Table 4 is a summary of the quartile measures. The values of the “sorting coefficient", 'So', are very similar, showing that the relative spread of the curves is very much the same, and the low values indicate the sands are very well sorted. This work showed that the sands of the Howell esker, the_maximum sorted, were apporximately twice as well sorted as those of the Charlotte esker, the minimum sorted. 50 In all cases the arithmetic and geometric skewness values show a slightly higher percentage of coarser grains, and for each esker the values are nearly identical. There is a close relationship of kurtosis indicating a similarity in the peakedness of the frequency curves for each of the eskers. From these facts the following conclusions may be drawn: 1. The sands of the various eskers used in this analysis are very similar. 2. The eskers associated with the Charlotte till plain were formed under similar con- ditions. 3. There is a close correlation between the individual eskers. 4. A petrographic investigation of sands to which statistical methods of correlation are applied, may be used advantageously in the examination Of other glacial deposits. 51 BIBLIOGRAPHY Books Chamot, E. H. and Mason, C.‘W., Handbook of Chemical Micro- SCOpY;,VOl._;, Principles of Microscopes and Accessories; lst ed. New York: John Wiley and Sons, Inc., 1931. Dana, E. 3., Descriptive Mineralogy. New York: John Wiley and Sons, Inc., 1914. Johannsen, at, Essentials for the Microscopical Determination of Rock Forming Mgnerals and Rocks. Chicago: University Of Chicago Press, 1914: .Manual of Petrographic Methods, 2nd ed. Chicago: University of Chicago Press, 1918. Krumbein, W. C., and Pettijohn, F. J., Manual of Sedimentary Petrography. New York: D-Appleton-Century Company, 193 . - » Milner, H. B., Sedimentary PetrOgraphy, New YOrk: Thomas Murby and Company, 1929. Wahlstrom, E. R., Optical Crystallography. New York: John Wiley and Sons, Inc., 1943. articles Anderson, S. .A., "The waning Of the Last Continental Glacier" in Denmark as Illustrated by Varved Clay and Eskers, The Journal of Geolo , Vol.39, No.7, 1931, pp. 009-624. _. Crosby, W. 0., "The Origin of Eskers" Boston Society of Natural History. Proceedi ipgs , V01. 30, 1902, pp. 375- 11. Davis, W. M., "The Subglacial Origin of Certain Eskers", Boston Society of NaturalpHistory.figroceedingg,'V01.25 1893, PP.-477-499. - Dryden, L., "A Statistical Method for the Comparison of Heavy Mineral Suites, "American Journal of Science, Vol. XXIX, 1935. pp. 393-408. Flint, R. F., "Eskers and Crevasse Fillings", American Journal of Science, 5th Series, Vol. 15, 1928, pp. 410-416. 52 Leverett, F., and Taylor, F. B., "The Pleistocene of Indiana and Michigan and the History of the Great Lakes”, United States Geological Survey, Vol. LIII, 1915. Riley, N. A., "Projection Sphericity", Journal of Sedimentary Petrology, Vol. 11, 19El, pp. 94-97. Rittenhouse, G., "A Visual Method of Estimating Two- Dimensional Sphericity”Journal of Sedimentary Petrolo , Vol. 13, 19A}, pp. 79-81, abstract. Trask, P. D., Origin and Environment of Source Sediments of Petroleum, Houston, Texas, 1932, pp. 67 ff. Trefethen, J. M., and Trefethen, H. B., "Lithology of the Kennebec Valley Esker", American Journal of Science, Vol. 292, No. 10, 1945, pp. 521-527. Trowbridge, A. D., "The Formation of Eskers", Science, New Series, Vol. 40, 1914, p. 145, abstract. Upham, W., "Evidence of the Derivation of the Kames, Eskers, and Meraines of the North American Ice Sheet Chiefly from its Englacial Drift '"Geolo ical Society 0: America, Bulletin 5, 189%, pp. 71-85. wadell, H., "Volume, Shape, and Roundness of Quartz Particles," Journal of Geology, Vol. 43, 1935, pp. 250-280. Unpublisheg2Material Erickson, R. L., "A.Petrographica1 Investigation of the Longitudinal Deposition within the Mason Esker Relative to its Origin," Unpublished Master's thesis, Department of Geology and Geography, Michigan State Collzge of Agriculture and Applied Science, 1948, DP. 0 I, .I “ “ ' I ' '- "I‘: 3 . .l.‘ .' '1' C. I," ‘v'! [‘3‘] I '1}. '3 ,1. .' \ . t ‘ ' ‘3 . . ‘\ ‘1‘. » ."'_' '1’!" u I 1 .' - ‘. , -. ‘ ‘ , ,m 1- :w". H: 3‘ “(In D O“ gififfiu .i tt‘n' T? ‘ . .. -. ”2%? ; 3",‘91.." I fi-o; 5,; ”7';va "l . . . - , . I . \f ”Y ‘ . . k, .' ' .* .|.\ J ‘ _: ill! - " o. y ' ‘l‘ I 0;;¢:" '§;Y:idu Bxf‘fii \”r~‘~3}:;‘z;¢‘ b. l a - e3»..- -- «a; ‘1 L ‘ J i '1’! ‘.' .’| ‘ 1 ‘ '_.|~.‘ v . 1‘. . “e - . "Hr ' ‘ ‘ .o 5*... . . .,. ._ . 'i -““ I l I .‘.. r': 9 4.1 ' 5 F; J. - a .h ‘| y: a}. "I v I. "gf. _’.;‘I‘I 4‘ J ‘ t. "a. ';' i. W j: $4310”. 1. ‘aulr, .' "i. ‘.' .. _ . ' V .. ’, a 1.3-_ .‘ .Hd .' :t-(n'? ¥$)¢(\n.*’!‘, 8a Ir I. . . :- 7 J _ , 3.. 3'. {fr / . . _. _ .- '- .' -'. I - .' "1‘ “ . "JV. 5" .1“ r‘ ' , \f‘. ”L M - 1:7"?-;§t:’.‘ ~ ' . _‘...h, L: \ O ' 7. ', “"1." .' 3,1. '5 . . ..: o." ',I".M L . ‘ "s "‘14. _'.fi} ; - . ..!{‘. .» ,' ’. {in {’4‘3? [ - I I' f“; ‘7‘ 4- 8' "* {H t 2* I. m"- " ' 0 T' 3‘ . l .. ' v n .‘p' ' ’l . MICHIGAN STQTE UNIV. LIBRQRIE IIHBIlll1ll|l2lllllHll"llllllllllINNlllllllllllllllllll"WI 293008080149