ll ‘ llll ? i 1 HH \ J 1 1 HI < l p—x l mo \l\l A SPECTRQCHEMBCAL ANALYSis OF 'EHE MSGLEEBLE RESIDUES CD? THE DURDEE LGMESFONE Q? PRESQUE ISLE- CQUNW, MECHSGAN Thesis {0? Hm Degree of M. S. MICHIGAN STATE UNKVERSITY Stanley Harofid Elman 1958 1111111111111111111111 31293 01062 0700 Michigan State University Department of Geology East Lansing, Michigan NOV 6 1958 DATE D TUE DATE DUE DATE DU E LEI L _ 3 fl L J ij 7ifi[;' Affirmative Act Ion/Equal Opportunity institutio n Mans-p 1 A SPECTROCHEMICAL ANALYSIS OF THE INSOLUBLE RESIDUES OF THE DUNDEE LIMESTONE 0F PRESQUE ISLE COUNTY, MICHIGAN by Stanley Harold Elman A.THESIS Submitted to the College of Science and Arts Michigan State University of Agriculture and Applied Science in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE Department of Geology 1958 ABSTRAC A spectrochemical analysis of the insoluble residues of the Dundee Limestone of Presque Isle County, Michigan was done to determine their use in correlation. Fourteen samples were taken; four laterally and ten vertically. A qualitative analysis showed no change in the elements pre- sent. A semi-quantitative analysis showed a change in the blackness or the spectral lines of calcium, cobalt, vana- dium, magnesium, and beryllium at different horizons. The relative blackness of these spectral lines, hence the rela- tive concentrations of the elements, remained the same in the residues of the same horizons. Correlation between.the four horizons was made based on the semi-quantitative data obtained from the relative blackness of the spectral lines of the five afore-mention- ed elements. Stanley Harold Elman ACKNOWLEDGEMENTS The author would like to express his sincere thanks to all those who aided him in this study. Among those who merit acknowledgement are Mr. Paul Thornley and Mr. Duanne Dewey of the Michigan Limestone Quarry who assisted the author in procuring the samples used, Dr. H. B. Stonehouse of the Geology Department of Michigan State University for his encouragement and interest, and Drs. Schwendeman and Sternberg of the Chemistry Department of Michigan State University who placed at the disposal of the author the spectrographic and photographic equipment used in this study. TABLE OF CONTENTS Page Introduction: Background and Purpose ................ 1 Description and Areal Extent of the Dundee Formation .................................... 4 Basic Theory of Spectroscopy ........................ 6 Samples: Location and Field Procedures ..................... 8 Preparation ....................................... 12 Description of Residues ........................... 14 Spectrographic Techniques ......................... 17 Spectrographic Results: Qualitative ....................................... 20 Semi-Quantitative ................................. 28 Interpretation of Results ........................... 32 Conclusions ......................................... 33 Suggestions for Further Study ....................... 34 Biblio(3ra13hy .0OOOOOOIOOOOOOOCOOOOOOOOOOOOOOOOOOOOOO. 36 Figure Figure Figure Figure Table Table Table Table ILLUSTRATIONS 1: Location Map - Southern Peninsula of biichigan OOOOOOOOOOOOOOIOOO 2: Map of the Northwest Section of the Quarry Showing Sample Locations .......... 3: Photograph of the Northwest Section of the Michigan Limestone Quarry ............ 4: Photograph of Sample Location One showing steepness of slope and chert bed near top .00COOOOOOOOOOO0.00.0.0000... 5: Photostat 6: Photostat 7: Photostat 8: Photostat 9: Photostat 10:Photostat 11:Photostat l2:Photostat of of of of of of of of Plate NO. 2 ................. Plate NO. 3 000.000.00.000... Plate N00 OOOOOOOOOOOOOOCOO Plate N00 6 Plate No. 7 ................. 8 9 Plate N0. .OOOOOOOCCOOOOOOO Plate No. 10 ................ Plate NO. 11 QQQQOOOOOOOOOOO. TABLES A: Sample weights and amount of acid used .... B: Spectral lines used in element identifica- tion .0OOOOOOOOOOOOOOOOOOOOOOO0.0.00000COO. C: Qualitative reenlts ooccooooocoooccococcooo 5? Quantitative results — Residue Spectra .... 3b 13 INTRODUCTION: BACKGROUND AND PURPOSE The application of insoluble residue studies to sur- face and subsurface correlation and identification of cal- careous rocks has become widespread in the last two decades. The publication of methods of preparation and terminology by McQueen and Martin (1931) gave the initial impetus to this type of study which has evolved into a major segment of the geologic investigations performed by petroleum com- panies. Up until now, all of the work done with insoluble resi- dues has been with the macroscopic and near-macroscopic fractions. "Clay and fine silt are generally decanted in routine work. Little or no work has been done with the fine residues, and their value for correlation and identification is yet to be determined." (Ireland: Leroy, l950,p. 143) Spectrochemical analysis has been employed for many years; however its use has undergone a rapid advancement in the last fifty years. Up until now, the majority of the publications dealing with the use of spectrography in geo- _ logic research and investigaion have been confined to qualitative and quantitative analysis of ore-bearing rocks and igneous and/or metamorphic complexes. It is the belief of the writer that this method of investigation can be -2- applied to problems of correlation that arise in the inter- pretation of sedimentary formations. Little or no work has been done with the fine fraction of insoluble residues. Insoluble residues reflect the sea bottom conditions, the classic conditions, the action of the currents, and the conditions existing on the adjacent land-masses during deposition. These factors may change independently or together and over long periods of time or rather abruptly. If the conditions mentioned above remain- ed constant over a long period of time, no lithologic change would take place in the sediment; hence, no strati- graphic breaks would occur. If conditions did change, how- ever, but to such a small degree or in some particular way so that no visible lithologic change could be seen in the sediment, formations deposited under different conditions might be considered to be of the same "environment". A study of the residues would show this assumption to be false. In areas where the residues consisted almost entire- ly of fines, the investigation would not be undertaken due to lack of suitable method of analysis. It is the conten- tion of the author that in such cases as described above, a spectrochemical investigation of the fines could yield useful results. The samples used in this study were chosen from the Dundee Formation in the Michigan Basin, since the Dundee -3- has been well investigated and insoluble residue studies have been made (see Bibliography). Also petroleum produc- tion in the basin from this formation is relatively impor- tant. At first, well cuttings were considered as the basis of the investigation. After running some tests and looking at various sets of well samples from different sections of the state, it was decided that the contamination introduced by the drilling tools, the drilling mud, and the cavings would be so great as to prohibit a true investigation. Quarry samples were chosen because of their purity and the Michigan Limestone Quarry at Rogers City, Michigan was cho- sen as the sampling site (see figures 1 and 2). As was la- ter discovered during the digestion of the limestone, the Dundee at this locality is exceptionally pure; yielding less than 1 per-cent (by volume) of residue. In most of the samples, the fines ran as high as 95 Per-cent by vol- ume of the total residue; the remainder being composed of quartz, gypsum and fossil fragments. The large proportion of fines made this location ideal for the study being per- formed. -33- Rogers City /1’U’\\. -\¥ S. ILMAN IOU. Figure 1: Location Mop Southern Peninsula of Michigan mam. :25... .m Mn -3b- 0 of .320 an 20:000.. macaw 00.305 3.000 .26.. .202, 30.. 00.5: 9.0.. 0.694 2.0.356 =4 6. ans. some... $530 Scams... 5°22: 2... no 5768 ammzntoz 9.10 no.2 .n m/ o _ ..o.n_ Ed h mm. A; new» CONN 00: LC) lJ: thl..n. Figure 2 mu .m .9; 383d .. :0; on nN .95... 9.0001 -4- DESCRIPTION AND AREAL EXTENT_OF THE DUNDEE FORMATION The Dundee Formation is quite wideSpread and underlies nearly all of the Southern Peninsula of Michigan. It is Middle Devonian in age and was first defined by Lane (1893, p.25). The Dundee is overlain throughout most of the state by the Rogers City Limestone which is a brownish-buff dolo- mitic limestone or dolomite. In Southeastern Michigan where the Rogers City is absent, the Dundee is overlain un- conformably by the Bell Shale member of the Traverse Group. Throughout the Southern Peninsula of Michigan, the Dundee Formation rests unconformably on the Detroit River Group. The lower extremity of the Dundee is usually placed at the appearance of the first anhydrite bed of the Detroit River Group. The surface extent of the Dundee Formation is very limited with the only outcrops being found "in a narrow belt in the southeastern part of the Peninsula and along the shore of Lake Huron from Mackinaw City, Cheboygen County to Presque Isle, Presque Isle County." (R.A.Smith,1915,p.159) The thickness of the Dundee ranges from 100 feet in the southeastern portion of the Southern Peninsula to over 200 feet in the northern part. Thicknesses of 400 feet or better have been encountered in wells drilled in Tuscola County, Michigan. -5... The Dundee is chiefly a high calcium limestone with some magnesium impurity near the base. It is generally gray to buff to brown in color, crystalline, and locally very fossiliferous. The beds of the Dundee in the Michigan Limestone Quar- ry are dark gray to buff, crystalline and especially high in calcium carbonate. I "The average of 235 analyses of the upper fifty feet of cores (from this quarry gave 97.85 per-cent of calcium carbonate, 1.26 per-cent of magnesium carbonate and 0.34 per-cent of silica". (R.A.Smith, 1915, p.251 -5- BASIC THEORY OF SPECTROSCOPY An atom, when excited, is known to emit radiant energy of definite wavelength. The excitation process causes an electron to be elevated to a higher state than is normal for the particular atom in question. The return of the e— lectron to its normal orbital position is accompanied by a loss of energy which is radiated. The energy lost by the electron during the return can be represented by the energy difference between the two states and this may be equated to hA)where h is the Planck constant (6.62 x 10-27 erg- second) and Q is the frequency of the energy radiated (in vibrations per second). From this, it can be seen that the electron energy loss may be measured in terms of the fre- quency of the radiant energy. The radiant energy, when dispersed, produces a spect- rum that is thus unique for the particular atom or molecule. The means of excitation may be an arc, spark, flame, or discharge tube. A mixture of unknown elements, therefore, can be identified by the interpretation of the spectra emitted during excitation. Since the concentration of the element present is directly proportional to the radiation, and the radiation emitted is proportional to the amount of blackening or the "optical density" of the spectral line produced on the photographic plate, a measure of the ele- -7- ment present in the unknown sample can be determined. Various books, tables, and charts have been published listing the most persistent lines of the elements by wave- length. The tables used for identification in this study can be found in Brode (1943), Ahrens (1950), Harvey (1950), and Harrison (1939). Also used as an aid in identification was the Spectral chart of the most persistant lines of R. U. Powder; this chart also containing the persistant lines of the iron spectrum which was used as an external standard. The complete spectrum or, as Brode states "the spec- tral distribution of energy" (Brode,1943,p.3), runs from the ultra-sonic (large wavelength, low frequency) to cosmic rays (small wavelength, high frequency). The portion of the spectrum used in this study ranges from 2929A to 4383A or from the near ultra-violet into the visible range. This portion was used for the two following reasons: (1) within its boundrics can be found the most persistant lines of most of the common elements, and (2) the instrument used would not photograph the spectrum at higher wavelengths. In some cases, however, photographic plates were exposed at lower wavelengths so as to confirm the identification of certain elements suspected. -8... SAMPLES LOCATION AND FIELD PROCEDURES The samples used in this study were obtained from the Michigan Limestone Quarry at Rogers City located on the shore of Lake Huron in Presque Isle County of the Southern Peninsula of Michigan (see figure 1). The quarry, which is owned and operated bf the Michigan Limestone Division of United States Steel, is the largest limestone quarry in the world. It was selected as the sampling site because it was felt that truly representative samples of the Dundee Forma- tion were to be obtained there and the exposure wa‘ suit- able for the purpose. Figure 2 shows the locations of tLe samples. This section of the quarry was selected for samp- ling as a persistant chert bed occurs, varying in thickness from 2 inches to 8 inches and lying immediately below a do- lomite zone (see figures 3 and 4) Which could be used as a horizon marker; all samples were taken below the chert. Small pockets of soft, gray shale were encountered in sec- tions of the quarry and slumping zones occured locally. These sites were avoided for it was felt that samples from these locals would not be representative of the Dundee Formation. At locations 1, 2, 3, and 4 (see figure 2), a sample was taken three feet below the chert zone. This gives the study an East-West lateral extent of one and one-half miles -9- and a North-South lateral extent of approximately one- fourth mile. At locations 1 and 3, 10 additional samples (5 at each locality) were taken at varying distances below the initial sample. Fig.3: The Northwest Section of the Michigan Limestone Quarry -10- Fig.4: Sample Location One show- ing steepness of slope and chert bed near top Due to the character of the quarry face, which was vertical in many places, sample A (the first.samp1e below one) is 13 feet below the chert. Samples B,C,D, and E, how- ever were taken at two-foot intervals. For the same rea- sons, Sample F (the first sample below 3) is 9 feet below the chert with G,H,I, and J at two-foot intervals. This method of sampling gives 3 additional lateral horizons with ' samples A,B, and C being at the same distance below the chert zone as samples H,I, and J. Other samples were not taken due to the steepness of the quarry wall and/or the thickness of the talus slope. Care was taken to eliminate sample contamination at the sampling site. After getting -11- a sample at the quarry face, it was broken by pounding on another piece of limestone, thus preventing any contamina- tion from the hammer. The samples were broken by this method to a convenient size (approximately 3 by 3 inches). Samples were placed in paper sample bags which were marked as to location and position below the chert. -12- SAMPLES' PREPARATIQE: Because contamination of any sort would cause errors of substantial import, a carefully controlled laboratory technique was developed. A cleaning solution composed of 70 ml. of sodium dichromate and 2000 ml. of concentrated sulfuric acid was prepared and all glassware was then soak- ed in this solution. The glassware was then washed twice in distilled, de-ionized water and left to dry. The acid used to digest the limestone samples was 12 normal hydro- chloric acid diluted one to one with distilled, de-ionized water to 6 normal. The 6 normal acid was then distilled to two-thirds its original volume with the impurities remain- ing in the residue. The samples were then placed in 400 m1. beakers and the distilled hydrochloric acid was added until the reaction had ended. The samples were not broken or crushed prior to digestion. A list of the sample weights and the approximate amount of acid needed for digestion can be found in Table A.. The samples were not heated during solution so that volatile constituents were retained. The residues were filtered and dried in a warm oven (150°F) overnight. They were then placed in small, screw top vials, marked, and stored. Samples of the solution were also pre- served in vials. -13- Some of the residues obtained contained fossil frag- ments. These were removed by picking them off the filter paper with a toothpick. They were placed in their own vials and labeled accordingly. A microscopic inspection of the fossil fragments revealed the original shell material had been completely replaced by silica; much of which was in the form of minute quartz crystals. The shape of the hard parts, however, was preserved. From the size and shape of the fragments, they were identified as brachiopods with some crinoid stems present. Due to the fact that the material contained in these fossil fragments was deemed secondary, they were not analyz- ed with the residue but on a separate photographic plate. IAEL§_A Sample No. Weight (in gms) Amount of Acid (ml) 1 43 200 2 40 200 3 43 200 4 93 300 A 89 ' 300 B 163 600 C 119 450 D 107 400 E 136 500 F 116 450 G 71 275 H 70 275 I 81 300 J 101 400 -14- SAMPLES DESCRIPTION OF RESIDqgfi: During the process of transfer of the residues from the filter paper to the vials, the non-fines were segregat- ed from the fines by use of a toothpick and placed in their own vials. The following is a description of these non- fines. Sample One: Siliceous, fossiliferous material, very frag- mentary; massive gypsum and fragmentary quartz. Sample Two: Siliceous, fragmental fossiliferous material, tentatively identified as brachoipods, iron stain; some fragments of quartz found, no mas- sive gypsum. Sample Three: Consisted wholly of fines. Sample Four: Siliceous brachoipod fragments, some with com- plete shells and hinge line, quartz crystals formed on inside parts of shells causing a geode-like affect; massive gypsum and frag- ments of quartz present, some iron stain. Sample A: Silicified brachoipod fragments, some badly crushed, coated with fines and iron stain. Sample B: Silicified brachiopod fragments coated with fines and iron stain; fragments of quartz with some well defined crystals. Sample C: Sample F: Sample G: Sample H: Sample I: Sample J: -15- Silicified brachiopod fragments coated with fines and iron stain; gray to white massive microcrystalline gypsum. Silicified brachiopod fragments coated with iron stain; gray to white massive, microcry- stalline gypsum; milky to white quartz frag- ments. Silicified brachiopod fragments, some large with complete shells and hinge line visible, iron stain coating; milky to white quartz frag- ments; gypsum of two types: (1) gray to white massive microcrystalline, and (2) colorless, vitreous crystals. Silicified brachiopod fragments with iron stain, some crinoid stem fragments; gray to white mas- sive microcrystalline gypsum; milky to white quart 2 fragment 8. Silicified brachiopod fragments with iron stain; gray white massive microcrystalline gpysum; milky to white quartz fragments. Silicified brachiopod fragments with iron stain; gray to white massive microcrystalline gypsum; milky to white quartz fragments; fossils badly crushed. Silicified brachiopod fragments with iron stain; Sample E -15- gypsum of both types abundant; milky to white quartz fragments; crinoid stem fragments. Silicified brachiopod fragments coated with fines and iron stain; gray to white massive microcrystalline gypsum; milky to white quartz fragments. -17_ SAMPLES SPECTROGRAPHIC TECHNIQUES: The instrument used for these analyses was a Bausch and Lomb Large Littrow Spectrograph which utilizes a quartz Littrow prism for dispersing. The computed dispersion of this instrument for the section of the spectrum used in this study was about 5.8 A/mm. The excitation source used was a Direct Current Arc energized at 70 volts and 4 am- peres; anode excitation was used throughout. Kodak Spec- trum Analysis Number One plates of size 10.2 by 25.4 cms. were used as the recording device. Development of the plates was by use of Kodak Developer Dbl9 for 4 minutes at 70°F. This was followed by a 30 second stop bath, a 15 minute acid fix and a 30 minute washing period. All plates were exposed with a slit width of 10 microns and a slit height of 5 millimeters with an overlap of approximately one millimeter. National spectroscopic carbon electrodes, one-fourth inch by 12 inches, manufactured by the National Carbon Company, were used. The samples were burned according to the following format: Five milligrams of the sample was placed in a shallow cup drilled in the end of the lower electrode. A toothpick was used to transfer the sample from the bottle to the electrode so as to eliminate the contamination that -18.. would have been introduced by a metal strip. The upper electrode was partially beveled producing a smooth burn. The gap length between the electrodes was held at approxi— mately one-fourth inch during all of the arcings in order to produce the best results. The amperage and voltage were controlled through a rheostat on the power supply unit. Arcing of the electrodes was done by means of a striker electrode which was run from the upper to the lower elec- trode causing the arc to jump the gap. The shutter was then opened and the sample burned until the cup had been completely consumed. Complete consumption was considered necessary because "some elements form highly refactory carbides with the electrode and persist in the are long past their expected or noticeable burn- ing time". (Harvey,l950,p.l25) The plate was then racked up 4 millimeters (so as to give an overlap of one millimeter) and the operation repeated for the next sample. This was continued until a convenient number of spectra (usually 4 or 6) were photographed on the same plate. The carbon electrodes were replaced by iron electrodes and the iron spectra were recorded above and be- low the sample series and were used for alinement and ex- ternal standard identification purposes. The fOSSil fragments found in some of the samples were burned in the same manner as the residues. Many methods have been outlined for the preparation -19- and arcing of liquid samples. The author found the follow- ing method to be highly satisfactory. The electrodes were pre-arced (the upper slightly beveled, the lower flat) for approximately 2 minutes. This caused a porosity to develop in the lower electrode. The lower electrode was then dipped approximately one-fourth inch into the liquid allow- ing it to soak up into the electrode. Immediate arcing followed; burning the lower electrode one-fourth inch down from the top (approximately 5 minutes). Although this caused heavy cyanogen bands, the spectral lines present were sharp and clear. The liquids were analyzed in series as were the residues with iron spectra above and below. A few words on the method employed in identification is considered necessary for a complete understanding of the spectrochemical technique and such a discussion follows. For the most part, the identifications were made by comparison with standard charts and tables. Such can be found in the publications already mentioned. When this method was not feasible, the wavelength of the unknown spectrum line was determined by the method outlined in Ahrens (l950,pp.62-63). A positive identification was made by the location and naming of the most persistant lines of the elements and in most cases at least 2 lines were identi- fied before final confirmation was given to the existance of an element in the spectra of the residue or liquid. SEQQTROGRAPHIC RESULTS QUALITATIVE: Figures 5 through 12 are reproductions of the photogra- phic plates utilized in this study. A list of the elements found on these plates is shown in Table E. Figure 7 shows the spectrum of residue number 1 photo- graphed at a lower wavelength of the spectrum than the other plates. This was done to check the existence of sili- con, antimony, end bismuth. Figure 7 also contains the spectra of the liquids of samples 1 through 4. An inspec- tion of Table B will show that only one spectrum line was used for the identification of each of the elements anti- mony and strontium in the liquid spectra. These lines are the most persistant of these elements and, in the case of strontium, the next most persistant line is beyond the scope of the instrument used. An inspection of the plates photographed yielded no qualitative difference between the residues. That is, the same elements persisted throughout all of the residues arc- ed, thus confirming the findings of Sloss and Cooke (1946). The spectra of the liquids, which were photographed and i- dentified so as to complete the research, contained their own suite of elements but these also were monotonously per- sistant throughout the samples. -21- TABLE B SPECTRAL LINES USED IN ELEMENTKIQENTIFICATION 1: Lines used in insoluble residue identifications: Element SpectralLipes - Wavelength g9 A Iron(Fe) 3099.9, 3100.3, 3100.7 Magnesium( M ) 3091.9, 3329.9, 2942.1 Aluminum(Al§ 3944.0, 3092.9, 3082.2 Vanadium(V) 3185.4, 3184.0, 3183.4 Beryllium(Be) 3131.1, 3130.4, 3321.3 Molybdenum(Mo) 3132.6, 3194.0 Calcium(Ca) 3158.9, 3179.3, 3933. 7 Antimony(Sb 3267.5, 2769.9, 2528. 5 Titanium(Ti 3242. 0, 3322.9, 3329. 5 Zinc(Zn) 3345. 5, 3345.0, 2800. 9 Cobalt(Co) 2424. 9, 3453.5, 3502. 3 SiliconESi 3905. 5, 2528.8, 2506.9 Bismuth Bi 3067.7, 2497.7 Sodium(Na) 3302.9, 3302.2 II: Lines used in the liquid identifications: Iron(Fe) 3099.9. 3100.3, 3100.7 Calcium(Cag 3644.4, 3179.3, 3158.9 Vanadium V 3183.4, 3183.9, 3185.4 Titanium Ti) 3192.0, 3371.5, 3981.8 Magnesium.(M ) 4352.1, 3093.1, 3091.1 Antimony(Sb§ 3232.5 Strontium(Sr) 4077.7 III: Lines used in the fossil identifications: Iron(fe) 3020.7. 3471.3, 3608.9 Silicon(Si) 2881.6, 3905.5 Magnesium(m ) 2928.7, 3329.9 Aluminum Al? 3082.2, 3092.7. 3539.4 Vanadium v) 3102.3, 3183.4, 3185.4 Calcium(Ca) 3158099 3179.3: 3844.4 Titanium(Ti) 3239.0, 3989.8 Copper(Cu) 3247.6, 3273.9, 3602.0 Zinc(2n) 3035.8. 3345.0, 3345.5 Gamma 3 $9: a use n 8333 saw as: .8 86on 8 one; MW H. WM V. “N “anvfivpm 7a..“ nub & «n V 3 on. M @vfl V a n f _ :..__:_=:__,__ __ : :i i. i. a: 2.: _ E2... _ 2 i s . . on. .. __._._._._...;..._..;_§._ 2. all... _ rags... m o... m at: . m 7 - -. a m gamma 3 89¢ N was H mongoom was nob” mo 8.30on ”w enemas." . P r fivm a» th um anvavnmn; on. o w v a s M mvtv . __ . : N- :1..r_._:__ ___ 2.. _ :. m_ a c o 5 _ : EL... LI: . 1 :_ :8: an ”9 v _ AdmNmN o» mmmdv slums .N.H aeseauom sank naamuph was noun no caveman «m mnanh "9 53 Fe 8 fl V AI M V to CC V 70 Ca Ca 1'. M k «in 23$ 3 89: a .. H 333g as as: no 282$ Adsumw o» mummv H useauom was nouH mo suaooam «5 agenda r0. H fivfiflvvvvflflmv .TR“ a .u. .N M. i. : ._ . Z. , _ :_ n. 223.. 22:} _ ._ t1; w v a "a , I (c In 0- o. ' Nte‘¢33 mu . i a .1. a : is.” __.h_+._._ at . . w £85 -24. TB mrm. "a 38mm 3 ST: e.H.m.¢.m.m 38:8 use no: mo «Sofia 3H Pawns ”swam anvsvmun. .. MM. 23%. 3 89: 396.33 353% as no: .3 Seesaw um 93»; of | IIIOII'I' S o. T: V St A. HM Van a» “Iv“vm u. «a c h A! Q T" mvnvu V a. £~