A semi-quantitative x-ray diffraction technique for estimation of smectite, illite, and kaolinite by Roger W E Hopper A thesis submitted in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE in Soils Montana State University © Copyright by Roger W E Hopper (1981) Abstract: Two studies are reported. I. An assessment of the major sources of error in the X-ray diffraction procedure was conducted using a nested design and ANOVA for peak area and clay mineral composition. Clay separation, slide preparation and slide positioning were significant sources of error. II. Modifications of the factor method for semi-quantitative characterization of clay mineral composition by X-ray diffraction analysis were tested. Samples used in the study were from early Tertiary aged sediments of the Fort Union Formation and associated soils in Southeastern Montana. Estimates of the total CEC of the clay-sized fraction were based on X-ray diffraction results. The accuracy of estimation for each modification was tested by linear regression comparing these estimates with measured CEC values. Variation in measured CEC explained 90% of the variation in estimated CEC, 92% of the variation in smectite composition, and 82% of the variation in kaolinite composition. Percent illite was compared with illite content estimated by total K analysis. Variation in measured illite content accounted for 74% of the variation in estimated illite content. A modification of the factor method is presented that provides relatively fast and reasonably accurate estimations of percent smectite, illite, and kaolinite for material that does not contain significant portions of vermiculite or chlorite.  STATEMENT OF PERMISSION TO COPY In presenting this thesis in partial fulfillment of the requirements for an advanced degree at Montana State University, I agree that the Library shall make it freely available for inspection. I further agree that permission for extensive copying of this thesis for schol­ arly purposes may be granted by my major professor, or, in his absence, by the Director of Libraries. It is understood that any copying or publication of this thesis for financial gain shall not be allowed without my written permission. Signature_________________ „ not* / W A SEMI-QUANTITATIVE X-RAY DIFFRACTION TECHNIQUE FOR ESTIMATION OF SMECTITE, !ELITE, AND KAOLINITE by ROGER W E HOPPER A thesis submitted in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE in Soils Approved: Chairperson, Gradpdm Committee Head, Major Department Graduate Dean MONTANA STATE UNIVERSITY Bozeman, Montana November, 1981 iii ACKNOWLEDGMENT The author wishes to express his gratitude to Dr. Murray G. Klages for providing: suggestions, constructive criticism, and unfailing patience and support during the extended time taken to complete this thesis. In addition, appreciation is extended to Dr. Hayden Ferguson, Dr. Gerald Nielsen, and Dr. Theodore Weaver for serving on the author’s graduate committee, for providing needed assistance and for providing clear insights into the complicated interactions of our natural environment. The friendships made while at Montana State University helped create the perfect atmosphere in which to study and work. Finally, I must thank my wife, Glenna. Over the past four years she provided the best combination of patience, support, logic and cajolery without which I might never have completed this thesis. IVITA................................... .......... ............................................... .................................. ii ACKNOWLEDGMENT............................................................................................. i .......... iu TABLE OF CONTENTS......................................... ! ............................................................ iv LIST OF TABLES.............................................. .............................................................. ...... vi LIST OF FIGURES.............................. .......... ....................... ! .................................... .. . . ; ix ABSTRACT.......................................................... x INTRODUCTION...................................................... .........................................................., I LITERATURE REVIEW..................................... ............./. Sample Dispersion and Particle Size Segregation . . . . . Sample Preparation and Presentation........................ Quantitative Estimation of Clay Mineral Components, MATERIALS AND METHODS --MAIN STUDY........... ........................................... 12 Samples............................ 12 Sample Preparation........................; ................... ..................................................... .... 12 Total Potassium Determination.......................... 13 Cation Exchange Capacity Determination................... 14 X-ray Diffraction Analysis. ;■...................................................................................... 14 Cation Exchange Capacity Estimation........... ......................... .................................... 19. Statistical Methods .............................................. ........................................... .. 20 MATERIALS AND METHODS-PRELIMINARY STUDY............. ..................... ............ 21 RESULTS AND DISCUSSION............................................................ : ..............................23 I. Preliminary Study—Sources of Error in Laboratory Technique..................... ...... 23 Significant Main Effects in Determining Peak Area Over All Soils Tested ................................ ........................... .............................. .............. 25- Sources of Error in Determining Clay Mineral Composition Between Soils................................. ..................................................................... 26 . II. Main Study—Quantification ...................... ......................................................28 Cation Exchange Capacity Estimation......................................... ................. 28 Smectite Estimation........................................... ............................................. . . . 3 6 TABLE OF CONTENTS Page m cn vi l> VPage Illite Estimation........................................................ . 4 2 Models derived assuming 8.3% elemental K per unit cell i l l i te ........................43 Models derived assuming 5.1% elemental K per unit cell i l l i te ........................47 Kaolinite Estimation............................................................................................. : 51 SUMMARY AND CONCLUSIONS................................ 57 LITERATURE CITED............................................................................................................. 61 APPENDIX 67 LIST OF TABLES Table Page 1. Sample Identification and Description................. ................................................... 13 2. Percent of Total Variance for Main Effects on Peak Area for All Soils Studied................................................................................................................... 24 3. Percent of Total Variance for Main Effects on the Determination of Clay Mineral Composition Over All Soils Studied................................................37 4. Linear Regression Models of the Measured CEC on Estimated CEC ......................28. 5. Selected Linear Regression Models of Measured CEC on Estimated CEC..................................................................... 29 6. Selected Linear Regression Models of Estimated CEC on Measured CEC........................................................................................................................... 32 7. Selected Linear Regression Models of Estimated Smectite Content on Measured Cation Exchange Capacity......................................................................36 8. Selected Linear Regression Models of the Difference in Estimated and Measured Cation Exchange Capacity on Estimated Smectite Content............................................................................................................................ 40 9. Selected Linear Regression Models of Estimated Illite Content on Measured Illite Content on Measured Illite Content (Assuming 8.3% K per Unit Cell Illite). . . . ; ......................................... .................................... 43 10. Selected Linear Regression Models of Estimated Illite Content on Measured Illite Content (Assuming 5.1% K per Unit Cell Illite).................................47 11. Selected Linear Regression Models of the Difference in Estimated and Measured Illite Contents on Estimated Illite Content (Assuming 8.3% K per Unit Cell Illite) ................................ 48 12. Selected Linear Regression Models of Estimated Kaolinite Content on Measured Cation Exchange Capacity..................................................................... 51 13. Selected Linear Regression Models of the Difference in Estimated and Measured Cation Exchange Capacity on Estimated Kaolinite Content............................................................... 53 vi Table Page 14. ANOVA Tablfe (Over All Thr^e Soils T ested )...........................................................68 15. Peak Area Measurements (in2) for the Preliminary Study....................................... 69 16. Percent Clay Mineral Composition for the Preliminary Study. . . .......................... 75 17. ANOVA for Main Effects on the 17A Peak (MgEG) for All Three Soils Studied—Preliminary Study . . .............................................................................82 18. ANOVA for Main Effects on the IOA Peak (MgEG) for All Three Soils Studied-Preliminary S tu d y ................................................................. 82 19. ANOVA for Main Effects on the 14A Peak (MgEG) for All Three Soils Studied—Preliminary S tu d y ..................... 83 20. ANOVA for Main Effects on the 14A Peak (K350) for All Three Soils Studied—Preliminary S tu d y ............... 83 21. ANOVA for Main Effects on the 7A Peak (MgEG) for All Three Soils Studied—Preliminary S tu d y ............... 84 22. ANOVA for Main Effects on the 7A Peak (K350) for All Three Soils Studied—Preliminary S tu d y ................................................. 84 23. ANOVA for Main Effects on the 3.5A Peak (MgEG) for All Soils Studied—Preliminary Study........................................................................................... 85 24. ANOVA for Main Effects on the 3.3A Peak (MgEG) for All Three Soils Studied—Preliminary S tu d y ..................... 85 25. ANOVA of Percent Smectite Composition for All Three Soils Studied—Preliminary Study...................................................... 86 26. ANOVA of Percent Illite Composition for All Three Soils Studied -Preliminary Study........... ................................................. 86 27. ANOVA of Percent Vermiculite Composition for All Three Soils Studied—Preliminary Study.......................................................... 87 28. ANOVA of Percent Chlorite Composition for All Three Soils Studied-Preliminary Study................... 87 29. ANOVA of Percent Kaolinite Composition for All Three Soils Studied—Preliminary Study............................................................ : ..........................88 vii Table Page 30. ANOVA of Percent Quartz Composition for All Soils Studied —Preliminary Study................................................ 88 31. Complete Area Measurements (in2) for thd Main S tudy ................................... 89 32. Complete Data Table for Modification IV ........................................................... 93 33: . Complete Data Table for Modification V ....................................... ......................... 95 34. Complete Data Table for Modification IX ..................................................; .............97 35. Complete Data Table for Modification X ........... ........................................................ 99 viii LIST OF FIGURES Figure Page 1. Peak Area Determination . .......................................................................................... 18 2. Linear Regression Model of Measured CEC on Estimated CEC for Modification X .................................................................................................... .. 3 1 3. Linear Regression Model of Estimated CEC on Measured CEC for Modification X ......................................... 34 4. Linear Regression Model of Estimated Smectite Content on Measured Cation Exchange Capacity.......................... 37 5. Linear Regression Model of the Difference in Estimated and Measured Cation Exchange Capacity on Estimated Smectite Content................................................................. 41 6. Linear Regression Model of Estimated Illite Content on Measured Illite Content (8.3% K) for Modification V ................................................................. 45 7. Linear Regression Model of Estimated Illite Content on Measured Illite Content (8.3% K) for Modification X ...............................................................46 8. Linear Regression Model of the Difference in Estimated and Measured Illite Content on Estimated Illite Content (assuming 8.3% K per unit cell illite)..............................................................................................50 9. Linear Regression Model of Estimated Kaolinite Content of Measured Cation Exchange Capacity........................................... 54 10. Linear Regression Model of the Difference in Estimated and Measured Cation Exchange Capacity on Estimated Kaolinite Content...................................................................................; .................................... 56 ix ABSTRACT Two studies are reported. I. An assessment of the major sources of error in the X-ray diffraction procedure was conducted using a nested design and ANOVA for peak area and clay mineral compo­ sition. Clay separation, slide preparation and slide positioning were significant sources of error. II. Modifications of the factor method for semi-quantitative characterization of clay mineral composition by X-ray diffraction analysis were tested. Samples used in the study were from early Tertiary aged sediments of the Fort Union Formation and asso­ ciated soils in Southeastern Montana. Estimates of the total CEC of the clay-sized fraction were based on X-ray diffraction results. The accuracy of estimation for each modification was tested by linear regression comparing these estimates with measured CEC values. Vari­ ation in measured CEC explained 90% of the variation in estimated CEC, 92% of the vari­ ation in smectite composition, and 82% of the variation in kaolihite composition. Percent illite was compared with illite content estimated by total K analysis. Variation in measured illite content accounted for 74% of the variation in estimated illite content. A modification of the factor method is presented that provides relatively fast and reasonably accurate estimations of percent smectite, illite, and kaolinite for material that does not contain significant portions of vermiculite or chlorite. INTRODUCTION X-ray diffraction methods are central to studies of the clay fraction of soils with the exception of those soils suspected to contain large amounts of amorphous material. A study of the components of the clay fraction of soil must begin with the proper identifi­ cation of the minerals present. Quantitative estimations of the clay mineral components have applications to many disciplines. Several methods have been employed to make quantitative estimations of clay minerals in soil investigations. The factor method used in making the investigations reported here has the advantages of being relatively rapid and precise. However, estimations obtained are relative values. Consequently, the method has been referred to as semi-quantitative. Few investigations report tests of the accuracy of quantitative estimations. McNeal [31] used a combination of chemical and X-ray diffraction methods to make quantitative esti­ mations of the mineralogy of arid and semi-arid land soils. The objective of this study was to derive a relatively fast and reasonably accurate method of determining clay mineral composition of soils and associated parent material for application to large numbers of samples. Ten modifications of the factor method for semi- quantitative characterization of clay mineral composition by X-ray diffraction analysis were tested for accuracy and precision. Appropriate factors were determined. Estimations of relative mineral composition and cation exchange capacity derived from X-ray diffraction results were tested against cation exchange capacity values obtained by chemical methods. Estimates of the percent illite derived from X-ray diffraction analysis were tested against percent illite values obtained from total potassium analysis. 2In addition an attempt was made to assess the major sources of error in the method of X-ray diffraction analysis used in this study, and to identify the diffraction maxima that may be measured with the greatest precision. LITERATURE REVIEW Quantitative applications are firmly based on sound theoretical considerations. However, almost every procedural step in X-ray diffraction methods may be considered as a potential source of error [16,43,32,39]. In this review of pertinent literature, an attempt is made to briefly survey the major procedural steps in making quantitative clay mineral estimations by X-ray diffraction methods. These steps may be identified as: (I) sample dis­ persion and particle size segregation, (2) sample preparation and presentation, and (3) quantitative estimation of clay mineral components. Sample Disperson and Particle Size Segregation Day [7], Kunze [27], Kittrick and Hope [23], Jackson [18] and Watson [53] provide reviews of procedures applicable to sample dispersion. It is generally accepted that cementing agents and free oxides and salts should be removed to some extent to aid in dis­ persion. Apparent disagreement does exist, however, as to the severity of the pretreatment required. The procedures described by Jackson [18] are generally rigorous. It has been demonstrated that less severe pretreatments are adequate to obtain satisfactory sample dis­ persion and X-ray diffraction results [23]. Several authors have found that sample pretreat­ ment can seriously affect apparent clay mineral composition. Harward, Theisen, and Evans [17] compared the effects of several different dispersion methods. Generally, they found that, although iron removal enhanced dispersion, it also resulted in significant differences in apparent clay mineral composition. More rigorous iron removal and dispersion treatments generally resulted in a greater number of clay minerals identified, however, this was also dependent upon the soil itself. 4The choice of the proper combination of pretreatment and dispersion methods remains to the discretion of the investigator. A combination of methods might be most worthwhile. While it is important to identify the maximum number of mineral components present, it is also advantageous to use those methods which retain the real mineral as­ semblages as they exist in situ [17]. Use of ultrasonic vibrations to obtain sample dispersion may eliminate need for drastic pretreatment. Olmstead [37] found that sonic vibrations could be used in con­ junction with chemical treatments to obtain stable dispersed suspensions of soil colloids. However, his work was largely overlooked for almost thirty years. Recently Edwards and Bremner [9,8] found, using soils having a wide range of characteristics, that sample dis­ persion could be obtained with most soils using only distilled water, thus reducing both the time involved in treating the sample and the possibility of destruction of natural mineral structure. This work has been corroborated by Genrich and Bremner [12]. Vladimirov [52] suggested the value of ultrasonic dispersion methods in studying highly calcareous soils where chemical treatments disallow a particle size investigation of carbonate salts. Emerson [10] found that sodium hexametaphosphate improved dispersion of soils particularly high in organic matter or soluble salts. In most cases it has been reported that abrasion of miner­ als is lower using ultrasonic methods than with either shaker or mixer methods of mechani­ cal dispersion except in the case of biotite [9]. Particle size segregation may be obtained by settling or centrifugation [18]. Tan­ ner and Jackson [48], in considering settling and centrifugation techniques, have pub­ lished nomographs by which the sedimentation of particles having a particular effective radius may be predicted according to time temperature, particle density and centrifuge 5speed. Procedures employing density gradient centrifugation and heavy liquid techniques have not received much attention in clay mineralogy as yet. Towe [51] suggested these latter techniques while critically considering the use of the less-than-2-micron particle size fraction in making typical clay mineral studies. Towe seriously questioned the ability of current sedimentation and centrifuge techniques to yield accurate representative samples of this size fraction based on inherent differences in particle density and settling times. Sample Preparation and Presentation Preferential orientation is rather easily obtained because of the shape of most layer-silicate minerals. Orientation results in the enhancement of basal (OOl) diffraction maxima and thus permits greater sensitivity to small amounts of the mineral components present [27]. The length of the specimen irradiated and the depth to which the X-ray beam pene­ trates are functions of the angle at which the X-ray beam intersects with the sample (0). The length of the irradiated specimen (L) may be calculated by the relationship: L = aR/2sin0 , where a represents the divergence slit width (in radians) and R represents the radius of the goniometer [38]. The irradiated specimen length increases rapidly at smaller angles of 0. This results in a maximum d-spacing that may accurately be measured and a minimum sample length. It is interesting to note that the majority of studies reported in the litera­ ture use CuKa radiation in conjunction with a divergence slit width of 1° to study soil clay minerals. According to the values reported by Parrish [38] the maximum d-spacing accu­ rately measured under these conditions is 5.2A, a value well below even the relatively small c dimension of the kaolinite minerals (approximately I A). 6Cullity [6, pp. 269-272] described the effective depth of X-ray penetration in terms of the fraction (Gx) of the total diffracted intensity contributed by a surface layer of a certain thickness (x) by the relationship: Gx = ( I - B - 2^xZsine) , where ju represents an appropriate mass absorption coefficient. Gibbs [13] used this equation to calculate penetration values. The need for a uniform sample in which no parti­ cle segregation has occurred is imperative. In a comparison of several techniques Gibbs [13] found that particle segregation was best avoided by smear-on-glass techniques as described by Theiseri and Harward [50] and suction-on-ceramic tile techniques described by Kinter and Diamond [22]. Centrifuge methods for deposition on either ceramic tiles or glass were found to cause particle size segregation and thus bias estimates toward the finer grained smectite minerals. An additional consideration in preparing oriented samples is the degree of orien­ tation that is actually achieved. Departure from the preferred orientation can cause a reduction in the peak interisity. Taylor and Norrish [49], using the suction-on-ceramic tile method reported significant variations in the degree of preferred orientation between specific minerals. They also found that variations in the degree of orientation for specific minerals may vary between duplicates. Quakemaat [41] reported relatively low absolute orientation for all minerals studied using a suction-on-plastic membrane technique. Schultz [44], in a study of kaolinite-illite mixtures using a smear-on-glass technique, found that the degree of preferred orientation in pure kaolinite samples, was greater than the orientation of either kaolinite or illite in mixed samples. However, for any one mixture, 7the preferred orientation of the kaolinite and illite was about the same. He concluded that the effect of orientation was eliminated within a single slide. At present no clear advantage is held by either the smear-on-glass or the suction-on- ceramic tile techniques in !comparison with each other [50]. Quantitative Estimation of Clay Mineral Components Quantitative X-ray diffraction methods fall into three basic approaches described here as: (I) the Theoretical Method, (2) the Standard Clay Mixture Methods, and (3) the Factor Method. The Theoretical Method. The work reported by Alexander and Klug and their associates [1,24,25] form the theoretical basis for current quantitative methods. In its simplest form, the method of Alexander and Klug [ I ] reduces to: S TAEMN wP ’ where Ip equals the diffraction intensity of the P component iri a multiphase mixture, OF p equals the diffraction intensity of the P component in pure form (the external standard), and Wp equals the weight fraction of the P component in the mixture. This method assumes that the mass absorption coefficient of the P component (ju*p) is equal to the mass absorp­ tion coefficient of the matrix containing the rest of the components of the mixture (#*%). This assumption is not strictly true and can lead to large errors. Leroux, Lennox, and Kay [28] attempted to correct for this by extending Eq. I to include a ratio of the mass absorption coefficient of the P component to the average mass absorption coefficient of the mixture (ju ^ 1): . V 1OjP = Gx0V ^ - 8Tabulated values of P R for several minerals are available [3]. Assuming an investigator has previously determined OF p, the application of this technique requires only a measurement of Ip and Williams [57] provided an improved method for determining the average mass absorption coefficient. The Standard Clay Mixtures Method. Methods using mixtures of standard clays have been applied through two basic avenues for quantification: (I) the calibration curve approach and (2) the empirical factor approach. The calibration curve approach uses mixtures of known weighed amounts of stand­ ard clay minerals to calibrate the method. Probably the most extensive use of this tech­ nique was that of Willis, Pennington, and Jackson [58]. They used 141 standard clay mix­ tures containing either 2, 3, 4, 5, or 6 components based on their conception oLthe weathering sequence of clay size material. Talvenheimo and White [47] used a diffractome­ ter in developing a standard clay mixture method for multiphase system containing kaoli- nite, illite, and bentonite. With this technique they reported 5 to 10% accuracy. Internal standards have been employed in the calibration curve approach. Com­ pounds such as MgO, LiF, and CaF2 having low absorption coefficients and high sym­ metry are normally used [3] so that small amounts may be incorporated in the sample to be measured without disrupting the desired degree of orientation. The internal stand­ ard method of quantitative analysis is based upon the ratio of the integrated intensity of a component in a clay mixture with the integrated intensity of an internal standard added to the mixture in a constant amount. Calibration curves are normally prepared using syn­ thetic mixtures of standard clay samples together with a constant amount of the internal standard. The use of an internal standard circumvents the need to know mass absorption 9coefficients or crystal lattice parameters. Because of this advantage the method has been applied to the study of soil clays by several investigators. Many of these investigations were done using photographic techniques on random powder mounts [55,19]. In a more recent study, Glenn and Handy [15] applied the internal standard method using a diffractometer. Orientation problems have been approached by Quakernaat [41] by the use of molybdenite as an orientation indicator. Compensating for deviations from preferred orientation, he set up quantity intervals using standard mineral mixtures. In determining quantities of kaolinite, illite, and smectite, he claimed an accuracy of about 7 percent. Estimations of chlorite, vermiculite, and . pyrophyllite were within about 10 percent accuracy. The empirical factor approach is best exemplified by the work of Schultz [44,45]. Basically this method uses standard clay minerals in binary combinations to obtain ratios of integrated diffraction intensities for two minerals. These ratios were than applied to a multiphase mixture, to characterize the peak intensities to obtain relative clay mineral com­ positions. As stated previously, Schultz recognized that such factors not only resulted from characteristics of the composition and lattice structure of the minerals, but from orientation effects as well. Schultz reported that in 50/50 mixtures by weight of several kaOlinites to Fithian Illite, the ratio of the integrated intensities was approximately 1/1. He found no consistent ratio for chlorite minerals. In a similar study Moore [33] reported the accuracy to be within 2 percent of the actual values. The major problem shared by the methods employing mixtures of standard clays is the difficulty faced in obtaining mineral standards that are comparable to the clays naturally occurring in soils. Gibbs [14], however, has reported a ,technique in which he 10 obtained standard minerals directly from the samples to be studied. Coupling the approach of Schultz as described above together with an internal standard method, he avoided the problems of absorption and crystallinity differences between the standards and the unknowns. The Factor Method. The factor method incorporates the use of an empirical multi­ plication factor by which measured peak intensities or integrated intensities are character­ ized. These factors may be derived experimentally as in the case of studies reported by Weaver [54] and Freas [11], or by calculations based on chemical and crystal lattice parameters [4,42]. The method has several advantages in that it is relatively rapid and generally has good precision. Any of the diffraction maxima may be used in the calcu­ lations along with a careful and reasonable choice of multiplication factors. The method as outlined by Johns, Grim, and Bradley [20] is probably the most often cited of all quantitative procedures. Basically it uses illite somewhat like an internal standard. The integrated intensities of the diffraction maxima of the other minerals are then related to the integrated intensity of the illite peak by appropriate multiplication factors. They used two illite peaks for comparison purposes. The IOA illite peak was multi­ plied by 4 to allow direct comparison with the 17A peak of smectite. The 3.3A peak was compared directly with the 3.5A maximum for chlorite and kaolinite. Heat treatments were used to discern minerals which occur concurrently in a peak. An apparently arbi­ trary correction for quartz was applied with the 3.3A peak of illite. Similar applications of this method have been reported by several authors and dif­ fer from the method of Johns et al. by either the method used to determine peak intensity 11 or integrated intensity, in the multiplication factors used, and/or in the diffraction maxima being measured. Weaver [54] used a factor of 2.5 in comparing the 7A peak with the IOA peak for the determination of kaolinite. Freas [11], on the other hand, in comparing all minerals present to the (001) diffraction maximum of kaolinite at 7A, used factors of 3, 3, and I for comparison with the (001) reflections of iltite, chlorite, and smectite, respec­ tively. Biscayne [2] in comparing all minerals to the 17A peak of montmorillonite used a factor of 4 for the 10A peak of illite and a factor of 2 for a comparison with the 7A peak. The relative composition of chlorite and kaolinite was further discerned by using the doublet occurring near 3.5A. Meade [29] assumed that smeptite, Jcaolinite, and illite reflected X-rays at the same intensity. In addition, he used different intensity factors for Type A chlorite (x2) and Type B chlorite (x 1.5) at 7A for comparison with the 10A peak of illite. The method of Keller and Richards [21] is closely similar to that of Johns et al. with the exception that a factor of 3 was used to compare the 17A peak to the 10A peak. Npiheisel and Weaver [34] used a factor of 2 for comparing the 17A and 7A peaks with the 10A peak. MATERIALS AND METHODS MAIN STUDY-QUANTIFICATION Samples The fifty samples used in this study were obtained from the Decker Coal Company, Decker, Montana. The material consists of early Tertiary aged sediments of the Fort Union Formation, together with soil formed on this moderately indurated material. The samples were chosen on the basis of clay mineral composition estimated from preliminary X-ray diffraction analysis in an effort to obtain a wide range of clay mineral composition. The description of those samples used in this study appeared in Table I. Sample Preparation The samples were first ground to pass a 2mm sieve. Sample dispersion was obtained by using a probe-type ultrasound machine (120 volts, 4 amps, 60 cycles) manufactured by Blackstone Ultrasonics, Inc. Ten grams of each sample were placed in 50 ml of 0.01% Na3CO3 and subjected to ultrasound for 2 minutes. Excessive heating of the samples was experienced using longer periods of dispersion. Stock clay-sized (< 2p) particle suspensions were prepared for each sample by five washings using 0.01% Na2CO3, centrifuging at 500 RPM according to the nomographs of Tanner and Jackson [48], and saving the supernatant from each wash. Four 25 ml sub­ samples were removed from these stock suspensions and prepared for X-ray diffraction analysis as described below. The remaining stock suspensions were saturated with calcium by centrifuge washing three times with N CaCl2. Excess salt was removed by simple dialy­ sis until a test for chloride was negative. Upon completion of dialysis the samples were air dried and hand ground with an agate mortar and pestle to pass a 60 mesh sieve. 13 Table I . Sample Identification and Description Sample No. Lab. Ident. No. Description (Depth in Feet) Sample No. Lab. Ident. No. Description (Depth in Feet) I 44518 0 -2 26 45257 54-60 2 44519 2 -5 27 . 45258 60-65 3 44520 5-11 28 . 45261 75-77 4 44521 11-15 29 45262 129-135 5 . ,44522 15 -20 30 45263 135-140 6 44523 20-23 31 45264 140 -145 7 . 44524 23 - 27 32 1235-4 55-60 8 44525 27-29 33 1235-5 60 - 66 9 44528 . 40-45 34 1235-6 73-78 10 44529 45-50 35 1235-8 103-107 11 44530 . 50-55 36 1235-9 107-117 12 44531 55-60 37 1235-10 117-126 13 44532 60-65 38 .. 1237-1 5-10 14 . 45245 0 -2 39 1237-2 10-20 15 45246 2 -5 40 1237-3 20-25 16 45247 5-11 41 1237-4 25-35 17 45248 11 -15 42 ■ 1237-5 35 -45. 18 45249 15 -20 43 1237-6 45-50 19 45250 20-25 44 1237-8 60-70 20 45251 25-28 45 1237-10 98.8 -105.3 21 45252 28-34 46 1237-11 105.3-110.0 22 45253 . 34-40 47 1237-12 111.8-120 23 45254 40-44 48 1237-13 120-130 24 45255 44-49 49 1256-2 42-52 25 45256 49 - 54 50 1256-12 138-142 Total Potassium Determination Duplicate 0.0500 g clay samples were weighed. The HFrHClO4 decomposition method was employed as suggested by Pratt [40]. The extract was diluted to lOO ml. so that the resulting solution contained 0.5% Sr as SrCl2. The concentration of potassium ions in solution was determined by atomic absorption. The results were reported in terms of illite, expressed as a percentage of the total clay fraction as calculated assuming 8.3% 14 [30] and 5.1% [54] elemental potassium per unit cell illite. The resulting estimations of the percent illite were tested against the percent illite estimated by X-ray diffraction analy­ sis using linear regression methods. Cation Exchange Capacity Determination Free carbonates were removed from the dry Ca-saturated clay samples by a modifi­ cation of the method described by Jackson [18]. The modification involved four centri­ fuge washings with normal sodium acetate buffer (pH 5.0) without heating the sample. Following carbonate removal air-dried Ca-saturated clay samples were prepared by the method previously described. The cation exchange capabilities of the clay samples were determined by a Ca//Mg exchange system. Duplicate 0.050 g samples were centrifuge washed four times using 10 ml aliquots of N MgCl2, saving the supernatant following each wash. The resulting extract was diluted to 50 ml so that the resulting solution additionally contained 0.5% Sr as SrCl2. The concentration, of Ca in the extract was determined by atomic absorption and the results reported in terms of meq/100 gm of clay. X-ray Diffraction Analysis One subsample of each clay sample was saturated with Mg by centrifuge washing three times with 25 ini aliquots of N MgCl2. Excess salt was removed by washing twice with distilled water. A second subsample was saturated similarly with potassium using N KC1. Excess salt was removed by washing, once with distilled water followed by a second wash with 50% ethanol. SybsampIes were duplicated for each clay sample. 15 Parallel oriented samples were prepared by the paste method of Theisen and Har- ward [50]. The Mg-saturated samples were ethylene glycol solvated by the condensation method described by Kunze [26]. K-saturated samples were heated to both 350°C and 550° C for three hour periods. X-ray diffraction analysis was carried out on a General Electric XRD-5 diffractome­ ter using Ni filtered CuKa radiation at 45Kv and 18ma with beam and detector slit widths of 1° and 0.2°, respectively. Medium range collimating assemblies were used for both the incident and reflected beams. Scanning speed of the goniometer was 2° 20 per minute and the chart speed was I inch per minute, giving a 2° 26 per inch diffractogram scale for all samples. All Mg-saturated, ethylene glycol solvated samples were scanned through a 20 range of 2°-30°. A 2°-15° 20 range was used for K-saturated samples for both heat treat­ ments. The criteria used to identify the clay minerals present in the samples were taken from [18], [56], and [5] and are as follows: Mineral Group Identification Characteristics Smectite d(001) maximum at approximately 17A under Mg-saturation and glycol solvation. K- saturation together with heat treatments cause progressive collapse of interlayer space resulting in a d(001) maximum at approxi­ mately IOA for the K-saturated, 550°C heat treatment. 16 Vermiculite d(001) maximum at approximately 14A under Mg-saturatioh and ethylene glycol sol­ vation. Total collapse of the interlayer space and a consequent d(001) maximum of ap­ proximately IOA result from K-saturation to­ gether with heat treatments. . Chlorite d (001) maximum at approximately 14A for all treatments, d (002) maximum may or may not be present in the K-saturated, 550°C heat treatment. Illite d(001) maximum at approximately 10A for all treatments. Kaolinite d (001) maximum at approximately 7A and a d(002) maximum at 3.SA for all treatments except K-saturated, SSO0C heat treated sam­ ples. On heating to approximately SSO0C the mineral reported here as kaolinite becomes amorphous to X-rays due to the collapse of crystalline structure. Quartz a diffraction maximum at approximately 3,3A and coincides with an accompanying . id(003) maximum of illite. \17- Peak intensities of the d(001) reflections were measured to a hand-drawn back­ ground line. The areas under the peaks were estimated by multiplying the peak height by the width of the peak at half the peak height [36], as illustrated in Fig. I. Characterization of the minerals followed the factor method of Johns, Grim, and Bradley [20] as modified by Wilding [59]. Modifications in.this method involved both the peaks and factors used to characterize the minerals considered. First order basal reflections were used in the characterization of all clay minerals. The 3.3A reflection was used to characterize quartz. Often the 14A peak of the Mg-saturated, ethylene glycolated slide ap­ pears as a shoulder on the high angle side of the 17A peak. In such cases the low angle side of the 14A peak was estimated, as in Fig. I, and the area calculated. The area of the'TVA peak was then corrected by subtracting the area of the 14A peak from the area of the 17A peak. Teh modifications were tested involving different factors for smectite and kaolinite. Computer programs were used to complete the characterization. The following compu­ tations were used to calculate characterized peak areas: Modification I. 17A Mg-sat. E.G./4 = Smec. Peak Area (14A Mg-sat. E.G. minus 14A K-sat. 350°C)/2 = Verm. Peak Area 14A K-sat. 350"C/2 = Chlor. Peak Area IOA Mg-sat. E.G./1 = 111. Peak Area (7A Mg-sat. E.G. minus 7A K-sat. 550°C)/4 =? Kaol. Peak Area (3.3A Mg-sat. E.G. minus 3/4(10A Mg-sat. E.G.j/4 = Quar. Peak Area 18 Figure I . Peak Area Determination A = Hand drawn background line B = Hand drawn line estimating the extent of the peak C = Area of peak overlap II = Maximum height of peak measured from A W = Peak width measured at 11/2 Peak Area = H xW 19 Modifications II, III, IV, and V were similar to Modification I except that the (7A Mg-sat. E.G. minus IA K-sat. 550°C)peak area was divided by 3, 2.5, 2, and I, respectively, for kaolinite estimates. Modifications VI, VII, VIII, IX, and X were similar to Modifi­ cations I-V except that the 17A Mg-sat. E.G. peak area was divided by 5 for smectite esti­ mates. For each modification tested, the characterized peak areas were totaled and the relative percent composition of each mineral calculated according: Smec. Pk. + Verm. Pk. Area + Chlor. Pk. Area + Kaol. Pk. Area + Quar. Pk. Area = Total Peak Area Percent Smectite = Smec. Pk. Afea/Total Pk. Area X 100 Percent Vermiculite = Verm. Pk. Area/Total Pk. Area X 100 Percent Chlorite = Chlor. Pk. Area/Total Pk. Area X 100 Percent Illite = 111. Pk. Area/Total Pk. Area X 100 Percent Kaolinite.= Kaol. Pk. Area/Total Pk. Area X 100 Percent Quartz = Quar. Pk. Area/Total Pk. Area X 100 Cation Exchange Capacity Estimation Cation exchange capacity values were estimated for the < 2p particle size fraction of each sample by multiplying the percent mineral compositions estimated by each modi­ fication with cation exchange capacity values reported by McNeal [31] for the minerals considered: Smectite 100 meq/100 g Illite 25 meq/100 g Kaolinite . 8 meq/100 g Chlorite 25 meq/100 g Vermiculite 175 meq/100 g Quartz 2 meq/100 g 20 The resulting estimated cation exchange capacity values were tested against the cation exchange capacity values determined by the laboratory procedure previously described, using linear regression methods. Statistical Methods Correlations were computed using, the Bivariate Correlation Analysis routine, sub­ program Scattergram and the Multiple Regression routine, subprogram Regression of the Statistical Package for the Social Sciences [35]. For each modification, linear regression models were developed for the following:. dependent vs. independent . MCEC ECEC ECEC MCEC ESM MCEC ECDIF ESM EIL MILS EIL MILS ILDIFB EIL EKA MCEC : ECDlF EKA where, , , ' . _ MCEG = Cation exchange capacity determined by chemical means: measured CEC. ' ' ECEC = Cation exchange capacity estimated from X-ray diffraction results. MILS = Relative Illite content of the clay fraction as determined by total potas­ sium analysis assuming 8.3% K per unit cell of IIlite; measured illife content. . MILS = Relative Illite content of the clay fraction as determined by total potas­ sium analysis assuming 5.1% K per unit cell of Illite; measured illite content. EIL = Percent Illite content of the clay fraction estimated from X-ray diffrac- u tion results. ’ ESM , . = Percent Smectite content of the clay fraction estimated from X-fay dif- ' fraction results. 21 EKA = Percent Kaolinite content of the clay fraction estimated from X-ray dif­ fraction results. ECDIF = Difference in CEC of ECEC minus MCEC ILDIF8 = Difference in the percent Illite content given by EIL minus MILS. The estimating accuracy and precision of a modification in the factor method was primarily based on comparisons of the slope, the y-intercept, the correlation coefficient (r), the coefficient of determination (r2), and the standard error of the estimate (SEE) for the linear regression models listed above. PRELIMINARY STUDY-SOURCES OF ERROR IN LABORATORY TECHNIQUE Three soils were chosen from samples obtained from the Decker Coal Company, Decker, Montana, and the Coal Mine Reclamation Program, Montana State University, on the basis of their relative smectite content: Ident. No. Description SoilA 44523 20-30 feet (High) Soil B 44531 55-60 feet (None) Soil C 15 Colstrip—C.M. (low to moderate) Watershed #1 From each soil, three 10 g samples were taken and clay size (< 2//) material sepa­ rated by the method previously described. From the resulting stock clay suspensions three Mg-saturated slides and three K-saturated slides were prepared. The Mg-saturated slides were ethylene glycolated and scanned through a range of 2°-30°, 29. K-saturated slides underwent successive heat treatments of 350°C and 550°C and were scanned through a range of 20-15°, 20 following each heat treatment. Each slide was positioned to create three arbitrary fields on the slide corresponding to center, slightly left of center, and 22 slightly right of center. Three readings were made at each position. The design results in 81 samples from each soil and 243 samples for the whole experiment. Characterization and relative clay mineral composition were determined by Modi­ fication No. I described previously in Materials and Methods-Main Study. One-way analysis of variance for the nested design [46] was computed over all soils tested considering peak area (measured in square inches), and relative clay mineral compo­ sition. In the analysis Of peak area the following peaks were considered: Mg-E.G. treatment: 17A, 14A, 10A, 7A, 3.5A, 3.3A K-35d°C treatment: 14A, 7A K-550°C treatment: 7A These peaks have been used in the past by numerous investigators in making quantitative estimations of clay mineral composition. RESULTS AND DISCUSSION The results herein reported and discussed were derived from two separate studies. The two studies are reported separately under the headings: I. Preliminary Study-Sources of Error in Laboratory Technique, and II. Main Study-Quantification. I. PRELIMINARY STUDY-SOURCES OF ERROR IN LABORATORY TECHNIQUE Peak area measurements (in2 ) and percent clay mineral composition data appear in Tables 15 and 16, respectively. Analysis of variance was conducted according to the model of ANOVA table appearing in Table 14 by computer. Computations were con­ ducted by Dr. Erwin Smith. Complete analysis of variance statistics for peak area appear in Tables 17-24. Complete analysis of variance statistics for relative percent composition appear in Tables 25-30. Percent of the total variance for the main effects was calculated from variance components' for all analyses and appear in Table 2. 1 Variance Components were calculated for the Analysis over all soils by: ^BcA ~ M^soils - M.^clay sep/bccln ^CcB ~ ^ c la y se p . ~ ^ s l id e /C(^ n 8DcC = MSslide " MSfield/slide/dn sNcD - Afield/slide - S2 - MSreacJ , Total Variance = S2 + S^jc Q + I FCY sCeB+ sBeA - Table 2. Percent of Total Variance for Main Effects on Peak Area for All Soils Studied Treatment MgEG Mg EG MgEG MgEG Mg EG Mg EG K350 . K350 Peak 17A 14A 10A 7A 3.5A 3.3A 14A . 7 A Soil 90.38* 42.46* 81.79* 65.62* 65.98* 62.26* 11.78* 45.74* Clay Sep./Soil 0.00 7.52* 1.59 . . 2.89 0.00 6.33 ' 0.00 25.06* Slide/Clay Sep. 0.00 0.00 0.00 14.96* 14.20* 9.66* 0.00 2.45 Field/Slide 8.69* 36.49* 11.98* 13.04* 13.36* 18.22* 75.88* 26.30* Read/Field 0.92 13.57 4.64 3.49 6.47 3.53 12.34 0.45 to f t * Significant for a = 0.05 for variance components within columns. 25 The percent of the variance as it appears in the tables is a useful statistic, although the magnitude of the values can be somewhat misleading. For this reason asterisks have been used to indicate main effects significant at the 5% level based on F-test values (Tables 17-30). Significant Main Effects in Determining Peak Area Over All Soils Tested The X-ray diffraction method used was shown to be sensitive to the variation in mineralogy of the clay size fraction of the soils used when considering peak area. As may be seen in Table 2, soil was shown to be the biggest source of variation. This was expected since the soils were chosen on their apparent clay mineral composition (high smectite, low to moderate smectite, arid no smectite). Positioning of the slide (field/slide) was found to have a significant effect in deter­ mining peak area for.all peaks considered in this study. This points to a definite need to be consistent in the positioning of the samples in the diffractometer. It may be further inferred that, in attempting to use different treatments, requiring removal and reposition­ ing of the same or a different slide, a significant sourpe of variation may be incurred in repositioning. This effect is not simply the effect of variations in the thickness of the speci­ men, but also the effect of the length of the specimen exposed to irradiation. Positioning , of the slide significantly off center has the effect of shortening the effective length of.the specimen irradiated by allowing a substantial portion of the incident X-ray beam to miss the sample [39]. For Mg-saturated, ethylene glycol solvated samples, between, slide (slide/clay sepa­ ration) variations were found to have significant effects in determining peak area for the 26 7A, 3.5A and 3.3A peaks. This points to possible clay.mineral segregation during the cen­ trifuge washing technique following Mg-saturation, difference in preferred orientation, variation of clay film thickness, or a combination of these effects. This, seems to be espe­ cially important when considering minerals in the high 26 angle region such as kaolinite and quartz, since the peaks showing significant variation are the first order peaks for these minerals. While these peaks also represent high order reflections for other minerals (chlo­ rite, vermiculite, and illite) similar variation was not recognized in the first order reflec­ tions for these minerals. For these samples, variations due to the clay separation procedure were found to be significant in determining peak area for the 14A and 7A peaks. Again this points to clay mineral segregation, in this instance during the clay separation procedure. The effect seems to be significant in considering chlorite and kaolinite. Vermiculite also exhibits a first order reflection at 14A but was not recognized in many samples. Sources of Error in Determining Clay Mineral Composition Between Soils The X-ray diffraction method used in this study was .found to be sensitive to the variation in mineralogy of the clay size fraction of the soils. As may be seen in Table 3, the soil was found to be a significant source of variation. Soil was the major source of variation for smectite, illite, and kaolinite. This was antici- . . I . • pated since the soils were chosen on their apparent clay mineral.composition. The positioning of. the slide was also found to have a significant effect in deter­ mining the relative clay mineral composition for each of the minerals tested. 27 Table 3. Percent of Total Variance for Main Effects on the Determination of Clay Mineral Composition Over All Soils Studied Smectite Illite Kaolinite Soil 98.56* 94.10* 87.31* Clay Sep./Soil 0.09 1.44* 2.20* Slide/Sep. 0.00 0.00 0.00 Field/Slide 1.12* 2.58* 8.19* Read./Field 0.23 1.89 2.30 Coef. of Var. 6.10 . 8.33 6.80 * - significant at a = 0.05 The clay separation was found to be significant in determining the relative clay mineral composition for all the minerals considered in this study except smectite. The crystallite size of smectite is relatively small. Variation in temperature, time, and centri­ fuge speed not compensated for during the centrifugation procedure would have a greater effect on the amounts of clay minerals having larger particle size than smectite. 28 II. MAIN STUDY-QUANTIFICATION The accuracy and precision of each modification of the factor method tested in this study were determined by linear regression analysis. In this way measured parameters, were compared with related estimated parameters. Complete peak area measurements ap­ pear in Table 31. Cation Exchange Capacity Estimation . Linear regression models were developed to determine the relationship of the cation exchange capacity of the clay-sized fraction as estimated from X-ray diffraction results (ECEC) to the CEC of the same clay fraction as determined by chemical methods (MCEC) for each of the modifications tested. Regression models of MCEC as estiinated from levels of ECEC also permit an assessment of the variability of MCEC for any level of ECEC. Table 4 contains the linear models and associated statistics developed for the ten modifications tested in this study. Table 4. Linear Regression Models of the Measured CEC on Estimated CEC Slope Intercept r F SEE Mod. I 0.683 3.66 0.927* 0.859 6.52 Mod. II 0.684 4.43 0.936* 0.877 6.10 Mod. Ill .0.688 4.97 0.939* ■ 0.881 6.01 Mod. IV 0.697 . 5.50 0.941* 0.886 5.87 Mod. V 0.750 7.11 0.949* 0.900 5.51. Mod. VI 0.733 3.01 0:932* 0.869 6.29 Mod. VII 0.739 3.82 0.935* 0.875 6.15 Mod. VIII 0.741 ‘ . 4.43 0.937* 0.878 6.07 Mod. IX 0.750 5.02 0.940* 0.883 5.94 Mod. X 0.815 6.64 0.946* 0.895 5.63 * - significant at a = 0.05 29 Based on the interdependence between MCEC and ECEC as measured by the cor­ relation coefficient (r), the coefficient of determination (r2), and the dispersion of MCEC about the regression line as measured by the standard error of the estimate (SEE), Modi­ fications IV, V, IX, and X were chosen as the best modifications tested. Complete data tables for these, four modifications appear in Tables 32-35. It may be assumed that if a modification accurately accounted for the CEC as measured by chemical methods, the regression model developed would be of the form MCEC = ECEC, where the regression line passes through the origin and has a slope of 1.0. The modification providing a linear relationship with a slope and y-intercept closest to these values may be assumed to provide the most accurate assessment of the suite of clay minerals present in the clay-sized fraction. Statistical results for these four modifications appear in Table 5. Confidence limits have been applied for consideration of both the inherent accuracy and precision of these modifications. Table 5. Selected Linear Regression Models of Measured CEC on Estimated CEC Mod. No. Slope Intercept r r2 SEE IV 0.697 ±0.072 5.50+1.66 . 0.94* 0.89 . 5.87 V 0.750+0.068 7.11 ±1.56 0.95* 0.90 5.51 IX 0.750+0.073 5.02 ±1.68 . 0.94* 0.85 5.94 X 0.815+0.069 6.64 ±1.59 . 0.95* 0.90 5.63 \ * - significant at a = 0.05 30 A value for the y-intercept greater than 0.0 is evidence of a failure of the modifi­ cation to completely account for the CEC of the clay-sized fraction as measured by chemi­ cal methods and/of the error inherent in the methods employed in obtaining values for both MCEC and ECEC. Since the values of the y-intercepts do not significantly differ from each other for the four models considered here it might be assumed that the error involved in estimating CEC is constant for all of the modifications discussed. The positive intercepts could be caused by one or more of several things. A few erroneously high measured CEC values (MCEC) on samples dominated by low CEC clays would have this effect. Another possible explanation is that the CEC of the clay minerals occurring in samples varied from those assumed in estimating the CEC of the clay fraction. Alternatively, the positive inter­ cepts may indicate the presence, of additional minor, low CEC constituents of the clay­ sized fraction, such as feldspars. Smectite, when present in a suite of clays, significantly affects both the estimated and measured values of CEC. Overestimation of smectite would tend to reduce the slope of these graphs to some value less than 1.0. A comparison of the slopes obtained for the regression models of MCEC on ECEC for the four modifications discussed here reveals that all of the modifications apparently overestimate smectite, since all of the slopes are signifi­ cantly less than 1.0 (a = 0.05). Of these four modifications, Modification X has the liighest value for the slope of the model, coupled with favorable values for the y-intercept, .r, r2., and SEE. Graphical representation of this relationship for Modification X appears in Fig. 2. The regression model given by MCEC = 0.815 (ECEC) + 6.64 approximates the expected relationship given by ECEC = MCEC represented by the dashed line. The graphs are simul- 31 y = 0.815x + 6.64 r = 0.95 r2 = 0.90 SEE = 5.63 10 20 30 40 50 60 70 80 Estimated CEC (meq/100 g clay) Figure 2. Linear Regression Model of Measured CEC on Estimated CEC for Modification X. 32 • I taneous at value of 36 meq/100 g clay. The model underestimates CEC of the clay fraction for values of MCEC less than this value and overestimates for values greater. Regression models of ECEC as estimated from values of MCEC provide a measure of the ability of MCEC to predict estimated values of CEC. This approach also allows an assessment of the variability of the estimated CEC for various levels of the measured CEC as well as an indication of the error involved in applying the regression model at different levels of MCEC. If a modification is making an accurate assessment of the clay minerals in a suite of clays, the regression model would be ECEC = MCEC. Values for the y-intercept that are less than 0.0 would reflect either a failure of the modification to entirely account for the actual CEC, or a use of inappropriate CEC values in estimating the. total CEC of the clay fraction. Overestimation of the CEC could be attributed primarily to overestimation of the smectite minerals and would result in a slope of the regression line that is significantly greater than 1.0. Statistical results for the four modifications discussed appear in Table 6. Table 6. Selected Linear Regression Models of Estimated CEC on Measured CEC Mod. No. Slope Intercept r r2 SEE IV 1.27±0.13 -2.40±2.25 0.94* 0.89 7.98 V 1.29±0.12 -4.98+1.97 0.95* 0.90 6.96 IX 1.18±0.12 -1.46+1.73 0.94* 0.88 7.18 X 1.10±0.11 -3.90±1.83 0.95* 0.90 6.48 * - significant at a = 0.05 33 If, for any one value of CEC for the clay fraction one and only one suite of clay minerals exists, it might be assumed that the SEE for the regression models of MCEC on ECEC would equal the SEE for the regression models of ECEC on MCEC. Values of SEE for the regression models of ECEC on MCEC3 for all the modifications discussed, are higher than those reported for the regression models of MCEC on ECEC, The variability of ECEC for any level of MCEC is greater than the variability of MCEC for any value of ECEC. This is because the clay mineral estimates exhibit greater variability than do values for measured CEC. This indicates that, in practice, due to the variability in X-ray diffraction results, one measured value of CEC does not represent a unique suite of clay minerals. As in the case of the regression models of MCEC on ECEC, the mean value for the slopes and intercepts indicate that these modifications tend to underestimate the CEC at lower levels of MCEC and overestimate the CEC when significant amounts of smectite are present. If any of the values for the CEC of each mineral group varies significantly from the actual Value it would be reflected in a slope different than 1.0. The slope obtained for the regression model testing Modification X does not significantly (a = 0.05) differ from 1.0. The value for the y-intercept does significantly differ from 0.0. Since the y-intercepts of the models discussed do not significantly differ from each other, the difference between the y-intercepts of the expected model and the model actually obtained is probably due either to erroneously high measured CEC values or to the presence of a mineral group that was not considered in estimating the composition of the clay-sized fraction. This parallels the results of the regression models developed for MCEC on ECEC. From Fig. 3 it may be seen that the regression model of Modification X for ECEC on MCEC given by ECEC = 1.10(MCEC) - 3.90, closely follows the expected relationship 34 I.IOx - 3.9 10 20 30 40 50 60 70 80 90 Measured CEC (meq/100 g clay) Figure 3. Linear Regression Model of Estimated CEC on Measured CEC for Modification X. 35 given by ECEC = MCEC represented by the dashed line. The graphs are seen to be simul­ taneous at a value of 39 meq/100 g. The model underestimates the CEC for values of MCEC less than this value and overestimates for values of MCEC greater than 39 meq/ 100 g. The percent error may be calculated by the following: where, y = I .IO(MCEC) - 3.90 = the predicted ECEC and y = MCEC = the expected value of ECEC assuming ECEC = MCEC y - y then %-error = ------ X 100, and the sign indicates underestimation (-) y or overestimation (+). From these calculations it was found that in the portion of the regression model corresponding to values of MCEC less than 20 meq/100 g the error is greater than 11%. It should be remembered, however, that in this portion of the graph the values of both MCEC and ECEC are themselves small and relatively small errors in terms of meq/100 g cor­ respond to large errors when expressed as percent. At these low levels of CEC, illite and kaolinite typically dominate the exchange complex. Consequently, small differences in the ECEC could be the product of significant errors in estimating the amounts of these minerals. The percent error involved in estimating CEC from this regression model in the area of the graph corresponding to values of MCEC greater than 30 meq/100 g is less than 5%. Consequently, Modification X appears to be most sensitive in determining the clay mineral composition for those clay-sized fractions containing significant amounts of smectite. Further, Modification X provides the greatest range in values over which it acceptably esti­ mates clay mineral composition of the clay fraction. 36 Smectite Estimation A more usable relationship is that of the estimated percent smectite composition (ESM) as determined by the measured CEC. These regression models present a direct pre­ dictive tool for the determination of smectite based on the measured CEC of the clay­ sized fraction and permit a measure of the precision of the estimate (SEE). Statistical results and confidence limits (a = 0.05) for the slope and y-intercept appear in Table 7. Table 7. Selected Linear Regression Models of Estimated Smectite Content on Measured Cation Exchange Capacity Mod. No. Slope Intercept r T2 SEE IV 1.49+0.13 *1.22+0.13 -21.7812.27 0.96** 0.91 8.03 V 1.35+0.11 *1.21 ±0.11 -20.7811.94 0.96** 0.92 6.87 IX 1.37+0.12 *1.21+0.12 -20.5912.10 0.96** 0.91 7.42 X 1.23±0.10 *1.19+0.10 -19.41 + 1.78 0.96** 0.92 6.30 * - slopes of the expected regression model derived by substituting 100meq/l OOg at 100% smectite content ** - significant at a = 0.05 Expected models were derived from the regression models developed for each modification assuming a MCEC of 100 meq/100 g at 100% smectite. This expected model appears as a dashed line in Fig. 4 along with the graphical representation of the confidence interval for the regression model obtained (dotted line). ; 37 . •• // 1.23x-19.41 i = 0.96 10 20 30 40 50 60 70 80 90 100 Measured CEC (meq/100 g clay) Figure 4. Linear Regression Model of Estimated Smectite Content on Measured Cation Exchange Capacity. 38 The y-intercepts obtained for these relationships did not significantly differ from each other and are related to the x-intercepts which correspond to the average CEC of the clay-sized fraction when no smectite is present. As may be seen from Table 7 the slopes of the expected regression models do not significantly differ. Consequently, the expected models are similar for all of the modifications tested. Expected models were derived from the regression models developed for each modification assuming a MCEC of 100 meq/100 g at 100% smectite. This expected model appears as a dashed Ifne in Fig. 4 along with the graphical representation of the confidence interval for the regression model obtained (dotted line). The y-intercepts obtained for these relationships did not significantly differ from each other and are related to the x-intercepts which correspond to the average CEC of the clay-sized fraction when no smectite is present. As may be seen from Table 7 the slopes of the expected regression models do not significantly differ. Consequently, the expected models are similar for all of the modifications tested. Overestimatioh of the percent smectite composition would tend to increase the slopes of these relationships. Assuming that the values for the y-intercepts relate to a con­ stant and true average of the CEC of the clay-sized fraction when no smectite is present, the preferable modification would be indicated by a regression model that closely esti­ mated the expected regression model and did not significantly differ from it. It may be seen from Fig. 4 that this is true of the regression model developed for Modification X. The graph of the expected and derived regression models are nearly concurrent and the expected model lies well within the 95% confidence interval applied to the derived regres­ sion model for this modification. The coefficient of determination indicates that approxi­ 39 mately 92% of the variation in the percent smectite composition was explained by vari­ ations in the measured CEC. The error involved between the expected and the derived models was approxi­ mately 3% at MCEC =100 meq/100 g clay. Less than 3% error is incurred in estimating the relative smectite content at lower values of measured CEC. A limiting factor in applying this relationship to the prediction of levels of smec­ tite content is the apparent low level of precision as indicated by an SEE = 6.30. The 95% confidence interval for any value of the percent smectite in a sample predicted from the measured CEC of the clay-sized fraction would be approximately ± 12.6 percent smectite. The difference between the measured CEC (MCEC) subtracted from the estimated CEC (ECEC) was used as a dependent variable (ECDIF) in testing relationships with ESM to assess any effect the apparent relative smectite composition might have on estimating accuracy. In the above discussions it was shown that accuracy is generally lower for esti­ mates of the smectite content in samples containing low amounts of smectite when the error is expressed on a percentage basis. The regression models developed for ECDIF on ESM provide a direct indication of bias and an avenue for determining the significance of the apparent bias. Statistical results for the four best modifications tested appear in Table 8. Confi­ dence limits have been applied to the slope and y-intercept to facilitate the assessment of the significance of the apparent bias implied by the slope and intercept of the model. Modifications providing accurate and unbiased estimates of the suite of clay min­ erals should reveal a relationship that is concurrent with the x-axis and is of the form ECDIF = 0, with r = 0, r2 = 0, and SEE = 0.0. Regression models testing the modifications 40 Table 8. Selected Linear Regression Models of the Difference in Estimated and Measured Cation Exchange Capacity on Estimated Smectite Content Mod. No. Slope Intercept .. r r2 SEE IV 0.248+0.067 -0.32±1.77 0.73* 0.53 6.28 V 0.217 ±0.067 -3.81 ±1.62 0.68* 0.46 5.72 IX 0.198 ±0.073 -0.63 ±1.79 0.62* 0.38 6.33 X 0.151 ±0.076 -3.83 ±1.66 0.50* 0.25 5.86 * - significant a t« = 0.05 that significantly differ from this perfect fit relationship indicate significant bias in making estimates of the clay mineral composition. Ovefestimation of the amount of smectite would tend to increase the slopes of these models and decrease the value of the y-intercept. Of primary importance in analyzing these relationships is the slope and y-intercept of the regression models obtained. From Table 8 it may be seen that the slopes of all the regression models significantly differ from 0.0. Consequently, it may be assumed that bias is incurred in estimating the clay mineral composition by any of the four modifications dis­ cussed. However, the results indicate that Modification X provides the most accurate and relatively unbiased assessment of the clay mineral composition. Modification X has the lowest slope; Modifications IV and IX do include 0.0 in the 95% confidence interval. The y-intercept for Modification X is reasonably close to zero. In addition the r, r2, and SEE values for Modification X are lowest of the modifications discussed. These results tend to support those previously reported for the regression models developed for ECEC on MCEC, MCEC on ECEC and ESM on MCEC. From Fig. 5 it is observed that the graphs of the expected and derived models are simultaneous at a value 41 y = 0.151x-3.83 r = 0.499 I2 = 0.249 SEE = 5.86 10 20 30 40 50 60 70 80 Estimated Smectite Content (%) Figure 5. Linear Regression Model of the Difference in Estimated and Measured Cation Exchange Capacity on Estimated Smectite Content. 42 of approximately 25 percent smectite. The regression model indicates that Modification X tends to overestimate the amount of smectite present when it is greater than this amount and underestimate the amount o f smectite when it is less than this amount. Modification X also tends to favor lower estimated amounts of illite. In the portion of the graph represent­ ing low smectite composition the suite of clay minerals would be dominated by illite and kaolinite. Consequently, Modification X should be expected to yield lower estimates of the CEC in samples dominated by illite. Illite Estimation As a further test of the accuracy of the mineral estimates, regression models were developed to examine the relationship of the relative composition of illite (EIL) in each sample with the percent illite based on total K analysis assuming 8.3% (MILS) and 5.1% (MILS) elemental K per unit cell of illite. Regression models of EIL as estimated from values of MILS and MILS provide a measure of the sensitivity of the chemical methods employed to account for variation in apparent percent composition of illite in the clay­ sized fraction. They also allow an assessment of the variability of the estimated percent illite for various levels of MILS and MILS as well as the error involved in applying the regression model at various levels on MILS and MILS. It may be assumed that if a modification accurately estimated the percent illite and all of the K. present in the clay-sized fraction was a component of the crystalline structure of illite then the expected regression model developed would be PIL = MILS or MILS. This regression line would pass through the origin and have a slope of 1.0. The modification providing a linear relationship with a slope and y-intercept closest to these 43 values may be assumed to provide the most accurate assessment of the relative amount of illite present in the clay-sized fraction. Values for the y-intercept that are less than 0.0 would reflect a failure of the modification to entirely account for the relative amount of illite as estimated by total K analysis. Models derived assuming 8.3% elemental K per unit cell illite. Statistical results for the four best modifications appear in Table 9. Confidence limits (a = 0.05) have been ap­ plied for consideration of both the accuracy and the precision of these modifications. Table 9. Selected Linear Regression Models of Estimated Illite Content on Measured Illite Content (Assuming 8.3% K per Unit Cell Illite) Mod. No. Slope Intercept r r$ SEE IV 1.63 ±0.24 -11.12±1.73 0.87* 0.75 6.10 V 1.03 ±0.15 -4.83±1.10 0.87* 0.75 3.90 IX 1.58±0.23 -9.22±1.67 . 0.87* 0.76 5.92 X 0.99 ±0.15 -3.31 ±1.09 0.86* 0.74 3.87 * - significant at a = 0.05 The y-intercepts of all the regression models (Table 9) are significantly less than 0.0. These negative intercepts support the conclusions made from the regression models developed for ECEC on MCEC, particularly that the inability of the modifications to com­ pletely explain the percent illite derived from total K analysis is due to the presence of a K-bearing mineral in the clay-sized fraction that caused an overestimation of the relative amount of illite. It may be seen, however, that there is a significant difference between the y-intercepts obtained for these models. While the y-intercepts of the regression models for Modifications IV and IX do pot significantly differ from each other, they do significantly 44 differ from the intercepts obtained for the models testing Modifications V and X. The latter modifications do not significantly differ from each other. It should be noted that Modifications IV and IX tend to cause higher estimates of smectite and illite while Modi­ fications V and X tend to favor higher estimates of kaolinite at the expense of the esti­ mated smectite and illite contents. If the modification is sensitive to the changes in illite content as determined by total potassium analysis and the K-content of the illite is 8.3%, the slope of the regres­ sion model testing that modification should not significantly differ from 1.0. In Figs. 6 and 7 it may be seen that the regression models closely parallel the expected models and differ in both cases by a relatively constant amount. The constancy of the difference between the expected and derived regression model might also indicate the use of an inappropriate combination of coefficients in the characterization of the peak areas. As stated earlier it might be reasonable to expect such a relationship if the procedure used to estimate the mineral components of the clay­ sized fraction failed to recognize a K-bearing mineral of minor, relatively constant, proportions. Values for SEE indicate lower variability for EIL at any given level of MILS for Modifications V and X. SEE values were much higher for the regression models testing Modifications IV and IX. Modification X seems to provide the most favorable combination of accuracy and precision for the estimation of illite of the four modifications discussed here. The regres­ sion model testing this modification is given by EIL = 0.99 (MILS) - 3.31. This closely folr lows the expected relationship given by EIL = MILSi The coefficient of determination 45 10 20 30 40 50 60 Measured Fllite Content (%) Figure 6. Linear Regession Model of Estimated Illite Content on Measured Illite Content (8.3% K) for Modification V. 46 0.99x-3.31 Measured Dlite Content (%) Figure 7. Linear Regression Model of Estimated Illite Content on Measured Illite Content (8.3% K) for Modification X. 47 indicates that 75% of the variation in the EIL values was explained by the variation of the MILS values. This is considered further evidence of the presence of additional K-bear- ing minerals in the clay-sized fraction. The percent error was dependent on the relative amount of illite present, with more error, on a percent basis, being incurred on the estimation of low amounts of illite. The percent error was greatly reduced, to about 10%, in considering estimates of larger amounts of illite. This parallels the findings reported by several authors and might be antic­ ipated from the parallel nature of the expected and derived models. Models derived assuming 5.1% elemental K per unit cell illite. The effect of assum­ ing 5.1% K per unit cell of illite is to increase the measured relative composition of illite in all samples compared to values derived assuming 8.3% elemental K. Modifications IV and IX favor higher estimates of smectite and illite. Based on the regression models developed for EIL on MILS, it was anticipated that the estimated illite composition derived by these two modifications might more closely relate to the measured illite composition assuming 5.1% K per unit cell illite rather than 8.3% K. Table 10 contains the statistical results for the four best modifications tested. Table 10. Selected Linear Regression Models of Estimated Illite Content on Measured Illite Content (Assuming 5.1% K per Unit Cell Illite) Mod. No. Slope Intercept r I 2 SEE IV 1.00±0.04 -11.12±0.50 0.87«" . 0.75 1.75 V 0.63 ±0.03 -4.83±0.40 0.87* 0.75 1.40 IX 0.97 ±0.04 -9.22±0.47 0.87* 0.76 1.67 X 0.61 ±0.03 -3.31 ±0.38 0.86* 0.74 1.35 * - significant at a = 0.05 48 The slopes of the regression models testing Modifications IV and IX do not differ from the perfect fit slope of 1.0. However, these modifications more seriously underesti­ mated relative Eite content in comparison with measured values assuming 5.1% K per unit cell illite than did Modifications V and X in similar comparisons assuming 8.3% K. These results indicate that for the samples used in this study the 8.3% value is a more accurate measure for the relative K-content per unit cell E ite than is 5.1%. Consequently, the dis­ cussion and conclusions drawn based on the models of EIL derived from MILS appear to be more appropriate. To assess the bias in estimating the relative Eite composition that is related to the estimated percent composition of Eite, the difference between MILS subtracted from EIL was used as a dependent variable (ILDIF8) in developing regression models with EIL. Sta­ tistical results for the four best modifications tested appear in Table 11. Table 11. Selected Linear Regression Models of the Difference in Estimated and Measured Illite Contents on Estimated Illite Content (Assuming 5;1% K per Unit Cell IEte) Mod. No. Slope Intercept i r2 SEE IV 0.519 ±0.075 -10.30±0.98 0.89* 0.80 3.48 V 0.237 ±0.116 -8.61 ±0.97 0.51* 0.26 3.43 IX 0.501 ±0.077 -9.46 ±0.98 0.88* 0.78 3.45 X 0.219±0.124 -7.83 ±1.00 0.45* 0.45 3.54 * - significant at a = 0.05 Modifications providing accurate and unbiased estimates of the suite of clay min­ erals should reveal a relationsEp that is concurrent with the x-axis and is of the form ILDIF8 = 0, with r = 0, r2 = 0 , and SEE = 0. Underestimation of the relative amount of illite present would tend to decrease the value of the y-intercept. A positive slope for the regression line indicates that the modification provides biased estimates of the percent illite composition such that, above some point at which the derived and expected models are simultaneous, the amount of overestimation is directly proportional to an increase in the apparent relative composition of illite. Conversely, the amount of underestimation is directly prdportipnal to a decrease in the percent illite composition. In addition, statisti­ cally significant values for the r and r2 further indicate that the bias inherent in the modi­ fication is significant. From Table 11 it may be seen that the slope and correlation coefficients of all the regression models discussed significantly differ from 0.0. It follows that bias is incur­ red in estimating the relative composition of illite by any of the modifications discussed here. It may be anticipated, from the results of the regression models developed for EIL on MILS (Table 9) that the y-intercepts will significantly differ from 0.0. This was found to be the case for all four modifications discussed. Modifications X and V combine the lowest slope with a value for the y-intercept that is in line with the results obtained for EIL on MILS. The r and r2 values are the lowest of the modifications discussed. The SEE is slightly higher for Modification X, but it is not felt that this indicates a significant increase in variabUity of the estimates compared to the other modifications. The regression model as shown in Fig. 8 indicates that Modification X underesti­ mates the relative amount of illite present below a value of 35.8% illite (EIL). This cor­ responds to the upper range of the percent illite composition as estimated by Modification X. Consistent apparent underestimation of EIL coupled with a low, yet significant cor- 49 50 0.219x - 7.83 r = 0.454 T2 = 0.206 SEE = 3.54 Estimated Illite Content (%) Figure 8. Linear Regression Model of the Difference in Estimated and Measured Illite Con­ tent on Estimated Illite Content (assuming 8.3% K per unit cell I Hite). 51 relation coefficient, is interpreted to indicate a consistent overestimation of MILS. This could be due to the presence of. an additional, but unmeasured, K-bearing mineral that composes a small, relatively constant portion of the clay-sized fraction. However, a defi­ nite conclusion based on the results of this study concerning the apparent underestimation of illite is not possible. When the results of these regression models are considered in con­ junction with the results previously reported here, it is felt that they indicate Modification X is more responsive to the amount of illite in a sample. Kaolinite Estimation Measurements of the kaolinite content such as those made for illite content were not made during this study. As a result the accuracy and precision of a modification’s ability to estimate the relative composition of kaolinite was inferred from linear regression models of the estimated kaolinite content (EKA) on the measured CEC (MCEC). Statistical results and confidence limits (oz = 0.05) for the slope and intercept appear in Table 12 for the four best modifications tested. Table 12. Selected Linear Regression Models of Estimated Kaolinite Content on Measured Cation Exchange Capacity Mod. No. Slope Intercept r r? SEE IV -1.05 ±0.17 68.45 ±2.87 -0.88* q.77 . 9.98 V -1.22±0.16 87.39±2.77 -0.91* 0.83 9.79 IX -1.02±0.16 . 68.36 ±2.74 -0.88* 0.77 9.70 X -1.17±0.16 86.96±2,68 -0.91* 0.82 9.49 * - significant at a = 0.05 52 In general, samples exhibiting low values for measured CEC will possess high kaoli- nite contents and low smectite content. The reverse is assumed to be true at high values for measured CEC. As a result the slopes of the regression models of EKA oh MCEC should be negative. As may be seen in Table 12 the slopes of the regression models testing the four modifications discussed exhibit this negative relationship. Kaolinite typically exhibits a relatively small CEC. In the presence of measureable amounts of smectite, kaolinite is expected to compose a relatively small portion of the total CEC of the clay-sized fraction. Further, it is assumed that relatively large Variations in the estimated kaolinite content would not affect the CEC of the clay fraction as much as a similar variation in smectite content. Consequently, it could be anticipated that the I accuracy and precision of the modifications as measured by the regression models of EKA on MCEC would be lower than that indicated by similar models developed for smectite composition. The high values for the SEE obtained for all the modifications tested indicate that these modifications estimate kaolinite content with less precision than either smectite or illite. Modification X exhibited the lowest SEE (9.49). The four modifications discussed exhibit significant correlation between the esti­ mated kaolinite content and the measured CEC. The values of the intercepts and their confidence intervals indicate that all four modifications significantly underestimate kaoli­ nite at 100% kaolinite content. Modifications V and X tend to favor higher estimates for kaolinite. These two modifications significantly differed from Modifications IV and IX but not from each other. Since regression models testing the estimation of smectite and illite content indicated that Modification X provided the best estimates of these minerals it was assumed that this modification provided adequate estimation of kaolinite content as 53 well. The regression model for Modification X is given by EKA = -1.17 (MCEC) + 86.96, with r = -0.91, r2 = 0.82, and SEE = 9.49. This relation is graphically presented in Fig. 9. Linear regression models of ECDIF as related to EKA were developed to assess the effect of the relative kaolinite composition has on the estimating accuracy. They appear in Table 13. Table 13. Selected Linear Regression Models of the Difference in Estimated and Measured Cation Exchange Capacity on Estimated Kaolinite Content Mod. No. Slope Intercept r r2 SEE IV -0.318±0.1p6 17.28±1.81 -0.72* 0.52 8.69 V -0.223 ±0.096 12.09±1.63 -0.67* 0.50 7.20 IX -0.238±0.106 . . 12.55±1.82 -0.60* 0.36 7.31 X -0.150±0.096 6.64±1.66 -0.50* 0.25 6.03 * - significant at a = 0.05 The expected regression model is concurrent with the x-axis and is of the form ECDIF = 0, with r = 0, r2 = 0, and SEE = 0. A significant difference between this perfect fit model and the derived model indiates a significant bias in making estimates of the kaoli­ nite content. Underestimation of the kaolinite content would tend to steepen the slopes of the regression models and mcrease the value of the y-intercept. Data in Table 13 show that . the slopes and y-intercepts significantly differ from 0.0. The correlation coefficients indi­ cate relationships significant at a = 0.05. It may be assumed that bias is incurred in estimat­ ing the relatively kaolinite content. The results indicate that Modification X provides the most accurate and relatively unbiased assessment of the kaolinite composition. It combines 54 -1.17x +86.96 -0.91 SEE = 9.49 10 20 30 40 50 60 70 80 90 100 Measured CEC (meq/100 g clay) Figure 9. Linear Regression Model of Estimated Kaolinite Content on Measured Cation Exchange Capacity. 55 the lowest slope and y-intercept with the lowest value of r, r2, and SEE for any of the modifications discussed. These results seem to corroborate those previously reported for estimates of smec­ tite and illite composition. The graphs of the expected and derived models for Modifi­ cation X (Fig. 10) are simultaneous at a value of approximately 44% kaplinite. This indi­ cates that Modification X underestimates CEC of the clay fraction at kaolinite contents higher than 44% and overestimates CEC at kaolinite contents lower than this value. This could result from an underestimation of kaolinite and/or an overestimation of smectite. Previously it was shown that Modification X overestimates CEC and smectite at values of smectite composition greater than 25 percent. The results indicate that Modification X tends to underestimate kaolinite and overestimate smectite at values of relative kaolinite and smectite composition greater than 44% and 25%, respectively. An apparent overesti­ mation of kaolinite and underestimation of smectite occurs at compositions below these values. Since regression models testing the accuracy of the illite estimations indicated a consistent error not directly involved with the X-ray diffraction pattern characterization • .- process, it is inferred that Modification X overestimates smectite and underestimates kaolinite. These results indicate that further accuracy might be obtained by trying other modifications of the factor method that reduce the characterized peak area of the 17A peak while it increases the characterized peak area of the 7A peak. 56 y = -0.150x + 6.64 r = -0.50 T2 = 0.25 SEE = 6.03 — e. 10 20 30 40 50 60 70 80 90 Estimated Kaolinite Content (%) Figure 10. Linear Regression Model of the Difference in Estimated and Measured Cation Exchange Capacity on Estimated Kaolinite Content. SUMMARY AND CONCLUSIONS An assessment of the major sources of error in the X-ray diffraction procedure was conducted in a preliminary study using a one-way analysis of variance for a nested design for peak area and percent mineral composition. When considered over all soils tested, the X-ray diffraction method used was found to be sensitive to the variations in the mineralogy of the clay fraction of the soils used when considering peak area and rela­ tive mineral composition. The positioning of the slide was found to be significant in determining peak area and relative composition for all soils tested. Clay separation and between slide effects were found to be significant sources of variation in specific instances and seem to be particularly important in considering kaolinite content and those peaks associated with kaolinite determination. The results indicate the need for extreme care and consistency during all phases of sample preparation and presentation. In the main study the accuracy of ten modifications of the factor method was investigated. The results indicate that the factor method may be applied in X-ray diffrac­ tion analysis to yield a relatively fast and reasonably accurate assessment of the dominant clay minerals composing the clay fraction of arid and semi-arid land soils and their parent material in southeastern Montana. The method seems particularly useful in making quanti- tative estimations of smectite, illite, and kaolinite. Linear regression models indicated that Modification X is most responsive to the amount of smectite, illite and kaolinite in a sample. This modification yields accurate esti­ mates for these clay minerals in material that does not contain significant amount of ver- miculite and chlorite. The following computations are used to calculate characterized peak areas for Modification X: 58 17A Mg-sat. E.G./5 = Smectite Peak Area (14A Mg-sat. E.G. minus 14A K-sat. 350°C/2 = Vermiculite Peak Area 14A K-sat. 350°C/2 = Chlorite Peak Area IOA Mg-sat. E.G./1 . = Illite Peak Area (7A Mg-sat. E.G. minus 7A K-sat. 550°C)/4 = Kaolinite Peak Area (3.3A Mg-sat. E.G. minus 3/4 ( IOA Mg-sat. E.G.))/4 = Quartz Peak Area Characterized peak areas are totaled and the relative percent composition of each mineral calculated according to: Smec. Pk. Area + Verm. Pk. Area + Chlor. Pk. Area + KaoL Pk. Area + Quar. Pk. Area = Total Peak Area Percent Smectite = Smec. Pk. Area/Total Pk. Area X 100 Percent Vermiculite = Verm. Pk. Area/Total Pk. Area X 100 Percent Chlorite = Chlor. Pk. Area/Total Pk. Area X 100 Percent Illite = 111. Pk. Area/Total Pk. Area X 100 Percent Kaolinite = Kaol. Pk. Area/Total Pk. Area X 100 Percent Quartz = Quar. Pk. Area/Total Pk. Area X 100 Linear regression models involving CEC determined by chemical means (MCEC) . and the CEC estimated from X-ray diffraction analysis (ECEC) showed that 90% of the variability of the predicted parameter could be accounted for by the measured parameter using estimates derived by Modification X. The variability of ECEC for any level of MCEC is greater than the variability of MCEC for any level of ECEC. This indicates that in prac- , tice, due to the variability in the X-ray results, one measured value of CEC does not repre­ sent a unique suite of clay minerals as determined by X-ray diffraction analysis. Measurement of illite content by total K analysis was conducted assuming 8.3% and 5.1% elemental K per unit cell illite. The results of linear regression analysis indicated that measurements made using the 8.3% K value more accurately depicted the estimated illite content. Using estimates derived by Modification X predictive regression models for the esti­ mation of the relative compositions of smectite, illite and kaolinite were developed based on values of measured CEC and the percent composition of illite as determined by total K analysis (MILS): %-smectite (± 12.6%)= 1.23(MCEC) - 19.41, r2> 0.92 %-illite . (+; 7.7%)= 0.99(MILS)- 3.31, r2 =0.74 %-kaolinite(± 19.0%) = -1 .17(MCEC) + 86.96,r2 = 0.82 . The relative amount of illite was frequently underestimated. The apparent under­ estimation could be caused by variability in the K per unit cell of illite resulting in over­ estimation of illite from total K analysis. This is supported by the moderate r2 value. Im­ proper factors employed in characterizing peak areas might have been responsible for the apparent underestimation, but, this is not supported by the slope of the regression. Alter­ natively, the y-intercept of this model could indicate either a threshold amount of illite that must be present to be recognized by the X-ray diffraction machine or could indicate the presence of an additional, K-bearing, nonmicaceous mineral such as feldspar in small constant amounts that was not considered in estimating the mineral composition of the clay-sized fraction of the samples studied. Modification X fended to overestimate smectite and underestimate kaolinite at values of relative smectite and kaolinite composition greater than 25% and 44%, respec­ tively. An apparent overestimation of kaolinite and underestimation of smectite occurs at compositions below these values. Since regression models testing the accuracy of the 59 60 illite estimations indicated a consistent error not directly involved with the X-ray diffrac­ tion pattern characterization process, it was inferred that Modification X overestimates smectite and underestimates kaolinite. Significant bias caused by the apparent amount of the mineral present was found for all modifications tested in estimating smectite, illite and kaolinite. However, Modifi­ cation X was shown to be least affected by the amount of mineral present. Based on a comparison of the SEE and r2 values, the estimates of the relative illite composition ap­ pear to be less accurate than estimates of the relative smectite content. Illite and smectite are estimated with greater accuracy and precision than kaolinite. LITERATURE CITED LITERATURE CITED I . Alexander, L. and H. P. King. 1948. Basic aspects of X-ray absorption in quantitative diffraction analysis of powder mixtures. Anal. Chem. 20:886-889. 2. Biscayne, P. E. 1965. Mineralogy and sedimentation of recent deep-sea clay in the ■ Atlantic Ocean and adjacent seas and oceans. Geol. Soc. Amer. Bull. 76:803-832. 3. Brindley, G. W.. 1961. Quantitative analysis of clay mixtures. /« G. Brown, ed. The X-ray Identification and Crystal Structures of Clay Minerals. Mineralogical Society^ London. 4. Brindley, G. W. and Hsein Ming Wan. 1974. Use of long-spacing alcohols and alkanes for calibration of long spacing from layer silicates particularly clay minerals. Clays and Clay Minerals 22:313-317. 5. Carrol, D. 1970. Clay Minerals: A guide to their X-ray identification. Special Paper 126, The Geological Society of America. Boulder, Colorado. 6. Cullity, B. D. 1956. Elements of X-ray Diffraction. Addison-Wesley Publishing Co., Inc., Reading, Massachusetts. 7. Day, P. R. 1965. Particle fractionation and particle-size analysis, pp. 545-567. In C. A. Black, ed. Methods of Soil Analysis, Part I. Agronomy No. 9. Amer. Soc. of Agron., Madison, Wisconsin. 8. Edwards, A. P. and J. M. Bremrier. 1964. Use of sonic vibration for separation of. soil particles. Can. J. of Soil Sci. 44:366. 9. Edwards, A. P. and J. M. Bremner. 1967. Dispersion of soil particles by sonic vibra­ tion. Jour. Soil Sci. 18:47-63. 10. Emerson, W. W. 1971. Determination of the contents.of clay-sized particles in soils. Jour. Soil Sci. 22:50-59. 11. Freas, D. H. 1962. Occurrence, mineralogy and origin of the, lower Golden Valley kaolinite clay deposits near Dickinson, North Dakota. Bull. Geol. Soc. Amer. 73: 1341-1363. 12. Genrich, D. A. and J. M. Bremner. 1972. A reevaluation of the ultrasonic-vibration . method of dispersing soils. Soil Sci. Soc. Amer. Proc. 36:944-947. 63 13. Gibbs, R. J. 1965. Error due to segregation in quantitative clay mineral X-ray diffrac­ tion mounting techniques. Am. Mineral. 50:741-751. 14. Gibbs, R. I. 1967. Quantitative diffraction analysis using clay mineral standards ex­ tracted from the samples to be analyzed. Clay Minerals 7:79-90. 15. Glenn, G. R. and R. L. Handy. 1961. Quantitative determination of soil mohtmoril- lonite by X-ray diffraction. Am. Soc. Testing and Materials Proc. 61:1277-1289. 16. Harward, M. E. and A. A. Thiesen. 1962. Problems in clay mineral identification by X-ray diffraction. Soil Sci. Soc. Amer. Proc. 26:336-341. 17. Harward, M. E., A. A. Thiesen, and D. D. Evans. 1962. Effect of iron-removal and dispersion methods on clay mineral identification by X-ray diffraction. Soil Sci. Soc. Amer. Proc. 26:535-541. 18. Jackson, M. L. 1956. Soil Chemical Analysis—Advanced Course. Published by author. University of Wisconsin, College of Agriculture, Dept, of Soils, Madison, Wisconsin. 19. Jarvis, N. L., R. D. Dragsdorf, and R. Ellis, Jr. 1957. Quantitative determination of clay mineral mixtures by X-ray diffraction. Soil Sci. Soc. Amer. Proc. 21:257-260. 20. Johns, W. D., R. E. Grim, and W. R. Bradley. 1954. Quantitative estimations of clay minerals by diffraction methods. Jour, of Sed. Petrol. 24:242-251. 21. Keller, G. H. and A. F. Richards. 1967. Sediments of the Malacca Strait, Southeast Asia. Jour, of Sed. Petrol. 37:102-127. 22. Kinter, E. B. and S. Diamond. 1956. A new method for preparation and treatment of oriented aggregate specimens of soil clays for X-ray diffraction analysis. Soil Sci. 81:111-120. 23. Kittrick, J. A. and E. V. Hope. 1963. A procedure for the particle-size separation of soils for X-ray diffraction analysis. Soil Sci. 96:319-325. 24. Klug, H. P. 1953. Quantitative analysis of powder mixtures with the Geiger-counter spectrometer. Anal. Chem. 25:704-708. 25. Klug, H. P., L. Alexander, and E. Kimmer. 1948. Quantitative analysis with the X-ray spectrometer—accuracy and reproducibility. Anal. Chem. 20:607-609. 64 26. Kunze, G. W. 1955. Anomalies in the ethylene glycol solvation techniques used in X-ray diffraction. Proc. of the 3rd Natl. Conf. on Clays and Clay Minerals 3:88-93.. 27. Kunze, G. W. 1965. Pretreatment for mineralogical analysis, pp. 568-577. In C. A. Black, ed. Methods of Soil Analysis, Part I. Agronomy No. 9, Amer. Soc. of Agron., Inc., Madison, Wisconsin. 28. Leroux, J., D. H. Lennox, and K. Kay. 1953. Direct quantitative X-ray analysis by diffraction-absorption technique. Anal. Chem. 25:740-743. 29. Meade, R. H. 1967. Petrology of sediments underlying areas of land subsidences in Central California. U.S. Geological Survey Prof. Paper 497-C. 30. Mehra, 0 . P. and M. L. Jackson. 1959. Constancy of the sum of mica unit cell potas­ sium surface and interlayer sorption surface in vermiculite-illite clays. Soil Sci. Soc. Amer. Proc. 23:101-105. 31. McNeal, B. L. 1966. Clay mineral variability in some Punjab soils. Soil Sci. 102:53-58. 32. McNeal, B. L. 1968. Limitations of quantitative soil clay mineralogy. Soil Sci. Soc. Amer. Proc. 32:119-121. 33. Moore, C. A. 1968. Quantitative analyses of naturally occurring multicomponent min­ eral systems by X-ray diffraction. Clays and Clay. Mineral. 16:325-326. 34. Neiheisel, J. and C. E. Weaver. 1967. Transport and deposition of clay minerals, Southeastern United States. Jour, of Sed. Petrol. 37 :1084-1116. . 35. Nie, N. H., C. H. Hull, J. G. Jenkins, K. Steinbrenner, and D. Brent. 1970. SSPS, Sta­ tistical Package for the Social Sciences, Second Edition. McGraw-Hill Book Company, Inc., New York, NY. b75pp. 36. Norrish, K. and R. M. Taylor. 1962. Quantitative analysis by X-ray diffraction. Clay. Minerals Bull. 5:98-109. 37. Olmstead, L. B. 1931. Dispersion of soils by a supersonic method. Jour. Agr. Res. 42:841-852. 38. Parrish, W. 1960. Advances in X-ray diffractdmetry of clay minerals. Proc. of the 7th Natl. Conf. on Clays and Clay Mineral. 7:230-259. 65 39. Pierce, J. W. and F. R. Siegel. 1964. Quantification in clay mineral studies of sedi- ments and sedimentary rocks. Jour, of Sed. Petrol. 39:187-193. 40. Pratt, P. F. 1965. Potassium. In C. A. Black, ed. Methods of Soil Analysis, Part II. pp. 671-698. Amer. Soc. of Agron., Inc., Madison, Wisconsin. 41. Quakemaat, J. 1970. Direct diffractrometric quantitative analysis of synthetic clay mineral mixtures with MoS2 as orientation indicator. Jour. Sed. Petrol. 40:506-513. 42. Schoen, R. 1962. Semi-quantitative analysis of chlorites by X-ray diffraction. Amer. Mineral. 47:1384-1392. 43. Schoen, R., E. Foord, and D. Wagner. 1972. Quantitative analysis of clays. Problems, achievements, and outlook. Proc. of the Internat. Clay Confi 1972. Madrid, Spain. 787-796. 44. Schultz, L. G. 1955. Quantitative evaluation of the kaolinite and illite in underclays. Proc. of the 3rd Natl. Cohfi on Clays and Clay Mineral. 3:421-429. 45. Schultz, L. G. 1960. Quantitative X-ray determinations of some aluminous clay min­ erals in rocks. Proc. of the 7th Natl. Confi on Clays and Clay Mineral. 7:216-224. 46. Sokal, R. R. and F. J. Rohlf. 1969; Biometry. W. H. Freeman and Company, San Francisco, CA. 47. Talvenheimo, G. and J. L. White. 1952. Quantitative analysis of clay minerals with the X-ray spectrometer. Anal. Chem. 24:1784-1789. 48. Tanner, C. B. and M. L. Jackson. 1947. Nomographs of sedimentation times for soil particles under gravity or centrifugal acceleration. Soil Sci. Soc. Anier. Proc. 12:60-65. 49. Taylor, R. M. and K. Norrish. 1966. The measurement of orientation distribution and its application to quantitative X-ray diffraction analysis. Clay Mineral. 6:127-124. 50. Thiesen, A. A. and M. E. Harward. 1962. A paste method for preparation of slides for clay mineral identification by X-ray diffraction analysis. Soil Sci. Soc. Amer. Proc. 26:90-91. 51. To we, K. M. 1974. Quantitative clay petrology: The trees but not the forest? Clays and Clay Mineral. 22:375-378. 52. Vladimirov, V. Ye. 1968. Study of the effects of acoustic vibrations on the physio- chemical properties of a soil suspension. Soviet Soil Scii 5:654-659. 53. Watson, J. R. 1971. Ultrasonic vibration as a method of soil dispersion. Soils and Pert. 34:127-134. . . . 54. Weaver, C. E. 1958. Geologic interpretation of argillaceous sediments. Part I. Origin and significance of clay minerals in sedimentary rocks. Bull. Amer.. Assoc. Petrol. ■ Geol. 42:254-271. 55. Whiteside, E. P. 1948. Preliminary X-ray studies of loess deposits in Illinois. Soil ScL Soc. Amer. Proc. 12:415-419. 56. Whittig, L. D. 1965, X-ray diffraction techniques for mineral identification and mineralogical composition: pp. 671-698../« C, A. Black, ed. Methods of Soil Analysis, Part I. Amer. Soc. Agron., Inc., Madison, Wisconsin. 57. Williams, P. P. 1959. Direct quantitative diffractometric analysis. Anal. Chem. 31: 1942-1944. 58. Willis, H. L., R. P. Pennington, and M. L. Jackson. 1957. Mineral standards for quanti- . tatiye X-ray diffraction analysis of soil clays: I. Abridgement of.component percent- , ages based dn weathering sequence. Soil Sci. Soc. Amer. Proc. 12:400-406. 59. Wilding, L. D. Procedure as written in mimeographed laboratory handout. APPENDIX Table 14. ANOVA Table (Over AU Three SoUs Tested) Source of Variation DF SS MS F Expected MS - Y Among Soils 2 a b c d n S(SSSSY)2 -CT ssSoils MSSoils ^2+nffDCC + ndffCCB bcdn 2 MSClay Sep. + ndcffBCAx ^ u Sff2+ ndcbPT Yg-Y^ Among Clay Sep. within Soils 6 a b c d n a b c d n SS(SZSY)2 _ S(SSSSY)2 ab cdn bcdn ^sClay Sep. 6 ^ sClay Sep. MSglide ff2 +nofjec + ndofceg + ndcffBCA Yc - Yg Among Slides 18 abc dn ab cdn SSS(SSY)2 _ SS(SSSY)2 ssSlide PN In Y ff2+nffDCC+ ndffCCB within Clay Sep. dn cdn 18 MsFieldZSMe p F r i I Among Field/Slide within Slides 54 a b c d n a b c d n SSSS(SY)2 _ SSS(SSY)2 n dn ssFieldZSlide 54 MsFieldZSMe MSRead/Field ff2+nffDCC prp e Within Field/Slide (Read/Field) TOTAL 162 242 abcdn abed n SSSSSY2 - SSSS(SY)2 n abcdn SSSSSY2 - CT ssReadZField 162 a2 Soils = 3 Clay Sep. = 3 Slide = 3 Field/Slide = 3 Read/Slide = 3 ab cdn C.T. = (SSSSSY)2 abcdn Table 15. Peak Area Measurements (m2) for the Preliminary Study Sample No. 17A Mg-EG IOA Mg-EG 14A Mg-EG 14A K350 7A K350 7A MgEG 3.5A MgEG 3.3A MgEG 11111 5.10 0.21 0.19 0.50 1.35 0.94 0.46 0.50 11112 7.92 0.24 0.36 0.56 1.30 0.95 0.46 0.42 11113 5.00 0.18 0.36 0.43 1.31 0.93 0.46 0.44 11121 3.19 0.14 0.15 0.23 1.16 1.08 0.42 0.66 11122 3.24 0.28 0.25 0.22 1.16 1.08 0.44 0.66 11123 3.70 0.23 0.27 0.31 1.19 0.93 0.44 0.65 11131 4.23 0.16 0.29 0.36 1.41 0.75 0.40 0.55 11132 4.21 0.17 0.15 0.40 1.45 0.75 0.42 0.49 11133 4.13 0.14 0.28 0.32 1.49 0.75 0.40 0.61 11211 4.70 0.22 0.27 0.36 1.23 1.05 0.58 0.53 11212 4.70 0.23 0.23 0.45 1.24 1.05 0.57 0.44 11213 4.70 0.16 0.25 0.39 1.24 1.05 0.60 0.44 11221 4.05 0.27 0.26 0.24 0.90 0.95 0.46 0.61 11222 4.55 0.18 0.25 0.23 0.88 0.95 0.46 0.66 11223 4.63 0.23 0.19 0.17 0.91 0.90 0.46 0.72 11231 5.39 0.24 0.19 0.62 1.33 0.89 0.40 0.60 11232 ' 6.12 0.18 0.28 0.74 1.31 0.87 0.40 0.59 11233 5.69 0.18 . 0.28 0.77 1.29 1.07 0.50 0.59 11311 5.00 0.24 0.28 0.47 1.30 1.19 0.66 0.63 11312 4.50 0.22 0.24 0:47 1.35 1.20 0.66 0.56 11313 5.05 0.20 0.28 0.43 1.35 1.21 0.68 0.60 11321 3.84 0.20 0.21 0.21 0.79 1.11 0.62 0.65 11322 4.37 0.22 0.15 0.30 0.79 1.13 0.60 0.63 11323 4.37 0.22 0.25 0.27 0.79 1.16 Q.60 0.70 11331 6.53 0.25 0.35 0.78 1.27 1.25 0.60 0.60 11332 7.25 0.28 0.40 0.67 1.26 . 1.27 0.58 0.63 11333 6.47 0.26 0.40 0.76 1.25 1.27 0.60 0.70 12111 4.60 0.17 0.18 0.47 1.33 1.03 0.52 0.50 12112 4.14 0.18 0.21 0.58 1.38 1.04 0.50 0.50 12113 4.23 0.16 0.18 0.37 1.31 1.05 0.48 0.55 12121 3.78 0.14 0.18 0.27 1.28 0.97 .0.54 0.60 12122 4.10 0.14 0.20 0.23 1.31 0.96 0.50 . 0.55 12123 3.36 0.16 0.24 0.41 1.31 1.00 0.52 0.55 12131 2.42 0.19 0.21 0.60 1.30 0.75 0.33 0.38 12132 2.42 0.15 0.20 0.51 1.30 0.77 0.32 0.34 12211 4.50 0.18 0.31 0.29 2.07 1.17 0.61 0.55 12212 4.79 0.25 0.28 0.45 2.08 1.16 0.62 0.55 12213 5.30 0.21 0.31 0.28 2.07 1.20 0.58 0.55 12221 4.85 0.18 0.14 0.21 1.19 1.09 0.61 . 0.60 12222 3.84 0.17 0.21 0.21 1.26 1.09 0.64 0.63 12223 4.23 0.18 0.21 0.13 1.25 1.15 0.60 0.65 Table 15 (continued) 70 Sample No. 17A Mg-EG IOA Mg-EG 14A Mg-EG 14A K350 7A K350 7A MgEG 3.5A MgEG 3.3 A MgEG 12231 6.13 0.20 0.32 0.51 1.87 0.97 0.38 0.50 12232 6.13 0.23 0.28 0.39 1.86 0.94 0.40 0.49 12233 6.25 0.20 0.32 0.57 1.86 0.97 0.38 0.52 12311 4.55 0.20 0.30 0.35 1.29 0.96 0.52 0.55 12312 4.27 0.19 0.21 0.49 1.25 0.97 0.52 0.48 12313 4.19 0.17 0.24 0.36 1.22 0.99 0.48 0.55 12321 3.85 0.18 0.27 0.13 0.95 0.86 0.40 0.50 12322 3.47 0.15 0.16 0.12 0.97 0.93 0.44 0.50 12323 3.28 0.12 0.14 0.23 1.01 0.83 0.42 0.50 12331 6.46 0.13 0.25 0.54 1.38 . 0.87 0.48 0.50 12332 6.00 0.15 0.28 0.56 1.38 0.89 0.46 0.55 12333 6.25 0.18 0.32 0.45 1.37 0.94 0.44 0.55 13111 4.73 0.16 0.21 '0.34 1.41 0.81 0.28 0.27 13112 4.67 0.14 0.23 0.39 1.45 0.77 0.33 0.27 13113 4.73 0.20 0.22 0.45 1.44 0.77 0.45 0.39 13121 3.11 0.09 0.13 0.16 1.24 0.83 0.40 0.48 13122 2.65 0.10 0.09 0.19 1.25 0.74 0.44 0.36 13123 2.29 0.08 0.11 0.19 1.23 0.70 0.42 0.40 13131 4.23 0.14 0.23 0.67 1.63 0.43 0.20 0.20 13132 4.50 0.12 0.13 0.47 1.59 0.49 0.25 0.24 13133 4.74 . 0.11 0.10 0.45 1.61 0.44 0.37 0.16 13211 4.41 0.14 0.28 0.52 1.26 1.03 0.52 0.57 13212 4.55 0.11 0.21 0.45 1.23 1.01 0.48 0.50 13213 4.50 0.13 0.26 0.39 . 1.26 1.01 0.48 0.54 13221 3.35 0.10 0.25 0.12 0.86 0.73 0.36 0.34 13222 2.88 0.09 0.06 0.15 0.86 0.73 0.36 0.32 13223 2.88 0.09 0.20 0.20 0.89 0.74 0.38 0.34 13231 5.40 0.09 0.06 0.64 1.46 0.60 0.29 0.30 13232 5.52 0.09 0.15 0.49 1.46 0.65 0.31 0.28 13233 . 5.58 0.13 0.08 0.64 1.39 0.61 0.31 0.28 13311 4.30 0.12 0.17 0.62 1.57 0.99 0.50 0.50 13312 4.16 0.12 0.16 0.45 1.65 0.90 .0.53 6.54 13313 4.50 0.17 0.21 . 0.65 1.57 0.96 0.54 0.56 13321 4.05 0.12 0.18 0.25 1.53 1.09 0.64 0.48 13322 3.96 0.13 0.20 0.19 1.50 1.12 0.64 0.52 13323 3.87 0.14 0.19 0.17 1.45 1.10 0.66 0.56 13331 7.25 0.21 0.26 0.68 1.69 1.10 0.56 0.44 13332 7.97 0.16 0.28 0.60 1.70 1.14 0.55 0.50 13333 7.84 0.14 0.25 0.54 1.73 1.09 0.53 0.48 21111 0.00 0.40 0.15 . 0.51 0.71 1.11 0.47 0.47 21112 0.00 0.37 0.21 0.45 0.67 1.13 0.49 0.54 71 Table 15 (continued) Sample No. ' 17A Mg-EG IOA Mg-EG . 14A Mg-EG 14A K350 Ik ■ K35'0 ,A . MgEG 3.5A MgEG 3.3A MgEG 21113 0.00 0.33 0.15 0.41 0.69 1.16 0.59 0.47 21121 0.00 0.27 0.06 0.22 0.36 • 1.21 0.65 0.70 211.22 0.00 0.28 0.10 0.22 . 0.44 1.19 0.72 0.68 21123 0.00 . 0.28 0.08 0.22 0.44 1.15 0:77 0.61 21131 0.00 0.36 0.15 0.50 0.64 1.19 0.53 0.58 21132 0.00 0.35 0.16 . 0.57 0.60 1.17 0.60 0.58 21133 ti.00 0.45 0.17 . 0.62 . 0.59 1.15 0.51 0.58 21211 0.00 0.35 0.20 0.42 0.50 1.27 0.73 0.70 21212. 0,00 0.39 0:24 0.41 0.44 1.31 0.69 0.61 21213 0.00 0:41 0.28 . 0.44 0.53 1.31 .0.63 0.67 21221 0.00. 0.51 0.17 0.25 0.25 1.74 0.99 . 0.89 21222 0.00 . 0.49 0.19 0.27 0.25 ■ 1.73 0.99 . 0.87 . 21223 0.00 0.48 0,15 0.23 0.25 1.71 0.85 1.00 21231 ■ 0.00 0.38 0.26 0.42 0.86 1.35 0.92 0.68 21232 0.00 . 0.49 0.16 0.47 0.89 1.32 ■ 0.71 0.68 . 21233 0.00 0.39 0.24 0.41 0.92 1.33 0.94 0.58 21311 0.00 0.40 0.14 0.52 0.69 2.04 1.13 0.96 21312 0.00 0.48 0.19 0.44 0.69 2.05 0.99 1,08 21313 0.00 0.52 0.17 ’■ 0.44 . 0.80 2.07 1.03 1.06 21321 0.00 0.43 0.14 0.25 . .0.35 . 1.75 0.90 0.84 21322 0.00 0.44 0.20 0.31 0.32 1.73 0.93 0.92 21323 0.00 0.43 0.13 0.25 ■ 0.33 1.73 1.07 0.94 21331 0:00 . 0.89 0.29 0.65 0.92 2.16 1.29 1.04 21332 0.00 0.87 0.25 0.57 1.07 2.23 1.25 1,08 21333 0.00 0.71 0.32 0.47 0.91 2.22 1.05 1.08 22111 0.00 0.73 0.21 0.36 . 1.03 1.41 0.64 0.59 . 22112 0.00 • 0.44 0.21 0.39 0.97 1.39 0.51 0.68 . 22113. 0.00 0.44 0.21 0.29 1.03 1.49 0.60 . 0.72 22121 0.00 0.49 ' 0116 0.20 0.60 . 1.99 0.86 0.79 22122 0.00 0.51 0.14 . 0.25 0.62 2.04 0.71 0.96 22123 0.00 0.51 0.17 0.20 0.64 2.09 • 0.73 . 0.94 22131 0.00 0.47 0.41 0.43 1,02 1.53 0.65 0.59 22132 0.00 0.47 • 0.37 ■ 0.47 .0.90 1.51 0.53 0.68 22133 0.00 0.49 0.30 0.44 1.01 L43 0.51 0.72 22211 0.00 0.47 0.16 0.47 0.81 1.36 0.67 . 0.68 22212 0.00 0.40 0.21 0.39 0.85 1.56 0.66 0.76 22213 0.00 0.47 0.18 0.51 0.84 1.39 0.65 0.76 22221 0.00 0.55 0.17 0.25 0.57 2.09 1.03 1.10 22222 0.00 . 0.53 0.18 . 0.25 0.53. . 2.01 0.99 1.10 22223 • 0.00 0.52 0.20 0.28 0.48 2.05 0.96 1.06 22231 0.00 0.68 0.30 0.50 1.09 2.04 0.77 0.88 72 Table 15 (continued) t Sample No. 17A Mg-EG IOA Mg-EG ., 14A : Mg-EG 14A K350 7A K350 ■ 7A MgEG 3.5A MgEG 3.3A MgEG 22232 0.00 0.77 0.25 0.56 1.09 2.09 0.81 0.84 22233: 0.00 0.61 0.31 . 0.51 1.07 2.09 0.86 0.88 22311 0.00 0.53 0.21 0.46 0.79 1.73 .. 0.81 . 1:00 22312 0.00 0.42 0.20 0.41 0.80 . 1.74 . 0.83 0.87 22313 0.00 0.44 0.17 0.47 0.80 1.68 1.17 1.02 22321 0.00 0.51 0.19 . 0:28 0.40 1.92 1.01 1.12 22322 0.00 0.64 0.20 0.26 0.40 1.91 1.04 1.16 22323 0.00 0.49 0.21 0.25 0.41 1.92 1.21 1.01 22331 0.00 0.46 0.36 0.38 1.10 1.77 0.74 0.73 22332 0.00 . 0.48 0.34 0.44 1.13 1.69 0,75 0.82 22333 0.00 0.50 0.26 0.44 1.10 1.67 0.75 0.88 . 23111 ■ 0.00 0.49 0.24 0.40 0.52 1,43 0.80 0.88 23112 0.00 0.49 0.19 0.35 0.54 1.47 0.91 0.82 23113. 0.00 . 0.51 0.21 0.44 0.50 1.50 0.93 0.94 23121 0.00 . 0.49 0.17 . 0.31 0.25 1.66 0.94 1.12 23122 0.00 0.51 0.18 0.25 0.26 1.69 . 0.95 1.12 23123 0.00 0.50 0.17 0.29 0.25 1.63 0.99 1.00 23131 0.00 0.51 0.35 . 0.45 .1.00 1.35 0.67 0.76. 23132 0.00 0.48 0.31 0.45 1.03 1:33 0.67 0.78 23133 0.00 0.50 0.30 0.53 0.97 1.35 0.64 0.68 23211 0.00 0.51 0.29 0.45 0:68 1.53 0.73 0.70 23212 0.00 0.58 0.22 0:46 0.70 1.55 0.76 0.86 23213 0.00 0.60 0.27 0.39 0.68 1.50 0.76 0.77 23221 0.00 0.54 0.18 0.24 0.16 1.75 0.97 1.14 23222 0.00 0.58 0.19 0.21 . 0.21 1.73 ' 1.01 1.16 23223 0.00 0.56 0.18 0.31 0.18 1.73 0.99 1.03 23231 0.00 0.43 0.28 0A4. . 0.73 1.43 1.07 1.14 23232 0.00 0.46 0.19 0.67 0.86 1.47 0.93 . 1.12 23233 0.00 0.47 0.19 0.65 0.84 1.45 1.26 1.14 23311 0.00 0.43 0.24 0.34 0.30 1.37 1.10 1.06 23312 0.00 0.40 0.25 0.29 0.30 1.15 0.97 . 1.04 23313 0.00 0.40 0.21 0,32 0.28 1.19 0.99 1.02 23321. . 0.00 0.32 ■ . 0.11 . 0.26 0.18 1.26 1.04 0.84 23322 0.00 0.33 0.16 0.22 0.16 1.27 1.04 0.88 23323 .0.00 . 0.33 0.11 0.21 0.15 1.07 0.89 0.86 23331 0.00 0.65 0.29 0.36 0.44 1.53 0.74. 0.99 23332 0.00 0.63 0.27 0.47 0.47 1.53 0.71 0.75 23333 0.00 0.60 0.29 0.35 0.47 1.51 0.83 0.75 31111 0.86 0.43 . 0.45 0.34 . 1.23 1.70 0.83 0.82 31112 0.91 0.45 0.40 0.37 1.20 1.77 0.84 0.76 31113 1.03 0.45 0.50 0.37 . 1.20 1.45 0.83 0.78 Table 15 (continued) 73 Sample No. 17A Mg-EG IOA Mg-EG 14A . Mg-EG 14A K350 7A . K350 - 7 A MgEG 3.5A MgEG 3.3A MgEG 31121 0.67 0.41 0.36 0.22 0.90 1.38 0.87 0.84 '31122 0.63 0.43 0.27 0.19 0.93 1.36 1.09 0.86 31123 0.73 0.42 0.20 0.19 0.91 1.39 0.85 0.84 31131 1.14 0.77 1.02 0.36 1.08 1.40 1.03 0.78 31132 0.99 0.47 0.84 0.23 1.11 1.79 0.82 0.80 31133 1.13 0.67 0.66 0.30 1.10 1.75 0.83 0.72 31211 1.14 0.46 0.40 0.39 1.54 1.44 0.84 0.80 31212. 0.97 0.48 0.46 0.39 1.55 1.46 1.05 0:84 31213 1.10 0.48 0.60 0.39 1.55 . 1.42 .0.85 0.84 31221 0.78 0.51 0.40 0.14, 1.07 1.44 0.89 .0.84 31222 . 0.64 0.50 0.34 0.16 1.26 1.52 0.91 0.84 31223 0.65 0.50 0.43 0.37 1.04 1.53 0.89 0.84 31231 ' 1.27 0.50. 0.50 0.48 1.56 1.44 0.84 0,78 31232 1.56 . 0.54 0.65 0.40 L49 1.80 0.86 0.82 31233 1.59 0.48 0.70 0.43 1.50 1.44 0.83 0.80 31311 0.75 0.48 0.36 0.47 1.09 1.83 0.86 0.76 31312 0.84 0.48 0.49 0.36 1.11 1.83 0.89 0.84 .31313 0.90 0.50 ■ 0.37 0.43 1.12 1.85 0.85 0:82 31321 0.71 0.45 0.24 0.15 0.83 1.46 0.88 0.82 . 31322 0.69 0.46 0.25 0.28 0.85 1.48 0.89 0.82 31323 0.60 0.49 0.55 0.16 0:87 1.85 0.93 0.84 31331 1.28 0.49 0.63 0.29 0,95 1.48 0.84 0.76 31332 1.38 0.50 0.66 0.25 0.95 1.80 0.85 0.76 31333 1.36 0.53 0.55 0.22 0.97 1.80 0.83 6.74 32111 0.82 0.50 0.38 0.25 1.33 1.73 0.79 0.76 . 32112 0.94 0.75 0.49 0.41 1.28 1.73 0.85 0.80 32113 0.82 0.63 0.35 0.27 1.35 ' 1.77 - 1.00 0.80 32121 0.49 0.43 0.26 0.19 0.96 1.50 : 0.90 0.76 32122 0.46 0.43 0.28 0.17 0.84 1.50 0.93 0.76 32123 0.58 0.41 0,27 0.18 0.84 1.68 0.93 0.78 32131 1.10 0.69 0.52 0.27. 1.35 1.70 0.91 . 0.70 32132 1.24 . 0.52 0.36 0.29 1.37 1.65 0.93 0.72 32133 1.24 0.63 ■ 0.67 0.41 1.38 1.70 0.91 0.74 32211 0.64 0.55 0.36 0.30 1.30 ■ 1.57 0.85 0.66 32212 . 0.70 0.63 0.27 0.25 1.33 1.50 0.69 0.64 32213 0.82 0.66 0.40 0.22 1.31 1.57 0.85 0.64 32221 , 0.56 0.40 0.24 0.16 0.85 1.43 0.71 0.58 32222 0.50 0.40 0.23 0.12 0.83 1.47 0.61 0.56 32223 0.42 0.53 0.19 0.16 '0.88 1.45 . 0.59 0.58 32231 1.25 0.63 0.55 0.28 1.50 1.70 0.96 0.76 32232' 1.18 0.52 0.40 0.28 1.50 1.67 0.78 0.78 . Table 15 (continued) Sample ' No. 17A Mg-EG , IOA Mg-EG 14A Mg-EG 14A K350 7A K350 7A MgEG 3.5A MgEG 3.3 A MgEG 32233 1.19 0.61 0.55 0.30 1.50 1.70 0.99 0.80 32311 0.80 0.66 0.37 0.39 . 1.38 1.67 0.99 0.80 32312 0.70 0.61 0.38 0.30 ,1.41 1.70 0.81 0.78 32313 0.85 0.63 . 0.39 .0.47 ■ 1.40 1.75 1.05 0.82 32321 0.58 0.40 0.24 0.25 1.15 1.50 0.94 0.74 32322 0.51 0.41 0.24 0.16 . 1.18 1.55 0.75 0.74 32323 0.60 0.44 . 0.17 . 0.19 1.18 1:53 0.95 0.74 . 32331 1.22 .0.65 0.46 0.30 1.47 . 2.13 1.01 0.78 32332 1.24 0.66 0.66. 0.37 1.45 . 1.75 1.04 ’ 0.82 32333 1.20 0.63 0.69 0.30 1.47 1.75 1.05 0.84 33111 0.77 0.59 0.36 0.31 0.49 1.13 0.61 0.82 33112 0.85 0.63 0.28 0:42 0.52 1.15 0.74 0.86 33113. 1.04 0.51 0.44 0.39 0.52 1.15 0.76 0.84 33121 0.53 0.49 . 0.23 0.17 0.43 1.17 0.80 0.92 33122 0.74 0.49 0.26 0.24 0.43 1.21 0.80 0.87 33123 0.64 . 0.40 0.20 0,21 0.42 1.50 0.83 0.96 33131 1.62 • . 0.53 0.41 : 0.32 0.49 1.19 0.74 0.88 33132 1.57 ' 0.46 • 0.33 0.28 0.49 1.21 0.76 0.79 33133 1.42 0.47 0.58 0.34 0.49 1.45 ‘ 0.75 0.77 33211 0.79 0.63 0.45 0.35 0.71 1.20 0.75 0.82 33212 0.84 0.56 0.36 0.33 0.71 1.41 0.71 0.82 33213 0.77 0.55 0.45 0.43 0.67 1.21 0.79 0.82 ■33221 0.62 0.44 0.20 0.27 0.56 1.50 0.81 0.92 33222 0.62 0.55 0.30 0.26 0.55 1.25 . 0.85 0.90 33223 0.53 0.57 0.21 0.24 0:59 1.25 0.83 0.86 33231 1.29 0.59 0.38 0.72 ' 0.70 1.40 0.67 0.96 33232 1.33 0.61 0.48 0.56 , 0.67 1.37 0.85 0.98 33233 1.15 0.64 0.55 0.54 0.70 1.40 0.83 0.96 33311 . 0.73 0.61 0.33 0.26 0.50 1.57 0.87. 0.82 33312 0.75 0.63 0.40 0.21 0.54 1.55 0.70 0.80 33313 0.72 . 0.53 0.38 0.22 0.55 1.55 0.89 0.86 33321 0.56 0.49 0.24 0.11 0.34 1.74 0.81 0.76 33322 0.59 0.49 0.25 0.09 0.32 1.45 0.81 0.76 33323 0.59 0.45 0.24 0.17 0.29 1.40 0.83 0.74 33331- • 1.49 0.68 . 0.58 0.34 0.60 1.60 0.94 0.88 33332 1.29 0.65 0.42 0.35 0.59 1.89 0.95 0.86 33333 1.17 0.65 0.63 0.31 0.56 1.55 0.97 0.88 , 75 Table 16. Percent Clay Mineral Composition for the Preliminary Study. Sample No. SMECT ILL VERM CHLOR KAOL QUARTZ 11111 62 10 00 12 11 04 11112 71 09 00 10 09 02 11113 64 09. 00 11 12 04 11121 . 55 ■ 10 00 08 19 09 11122 51 17 ' 01 07 17 07 11123 56 14 00 09 14 07 . 11131 63 09 00 11 11 06 11132 62 10 00 12 11 OS 11133 63 . 09 00 10 11 08 11211 61 11 00 09 14 05 11212 60. 12 00 11 13 . 03 11213 63 09 00 10 14 .04 11221 58 15 01 07 14 06 11222 63 10 01 06 13 07 11223 63 13 01 05 12 07 . 11231 6.1 11 00 14 10 05 11232 64 07 00 . 15 09 05 11233 60 08 00 16 11 05 11311 : 59 . 11 00 11 14 OS 11312 57 . 11 00 12 15 05 11313 60 10 00 10 . 15 05 11321 58 ■ ■ 12 00 06 17 07 11322 59 . 12 00 08 . 15 06 11323 58 12 00 07 15 07 11331 61 09 00 15 12 04 11332 64 10 00 12 11. 04 11333 60 10 00 14 12 05 12111 60 09 00 12 13 OS 12)12 56 10 00 . 16 14 05 12113 60 09 00 10 15 06 12121 60 09 00 09 . 15 08 12122 63 . 09 00 07 15 07 12123 54 10 00 13 16 . 07 12131 . 45 14 00 22 14 04 12132 48 12 00 20 . 15 05 12133 45 14 00 20 17 04 12211 . 61 10 01 08 16 06 • 12212 58 12 00 . 11 14 04, 12213 63 10 01 07 14 . 05 76 Table 16 (continued) Sample No. SMECT ILL VERM CHLOR KAOL QUARTZ 12221 64 09 00 . 06 14 06 12222 59 . . 10 00 06 .17 08 12223' 60 10. 02 04 16 07 .12231 66 09 00 11 11 04 12232 67 10 00 09 10 03 12233 . 66 08 00 12 10 04 12311 61 11 00 09. 13 05 12312 58 10 00 13 13 05 12313 60 . io 00 10 14 06 12321 61 11 04 04 14 06 12322 61 ■ 11 01 04 16 . 07 12323 60 09 00 08 15 07 12331 69 06 00 12 . 09 . .04 12332 66 . 07 00 12 10 05 12333 68 . 08 00 10 10 05 13111 67 . 09 00 10 12 . 02 . 13112 67 08 00 Tl 11 02 - 13113 64 11 00 . 1 2 10 03 13121 62 . 07 00 06 17 08 13122 59 09 00 09 17 06 13123 57 08 00 09 17. 08 13131 64 08 ob .20, 07 01 13132 69 07 00 14 07 02 13133 72 07 00 14 07 01 13211 59 07 00 14 14 06 13212 62 06 00 12 14 06 13213 62 . 07 00 11 14 06 13221 64 08 .01 05 14 05 13222 64 • . 08 00 07 16 06 13223 62 08 00 09 16 .0 6 13231 69 05 00 16 08 03 13232 71 ' 05 00 13 08 03 13233 68 06 00 16 07 02 13311 58 07 00 17 • 13 05 13312 60 07 00 13 13 07 . 13313 • 57 09 00 17 12 05 . 13321 62 07 00 08 17 06 13322 62 08 00 06 . 17 07 13323 61 09 01 05 17 0.7 77 Table 16 (continued) Sample No. SMECT ILL VERM CHLOR KAOL QUARTZ 13331 67 08 00 13 10 03 13332 70 06 00 11 10 03 13333 72 OS 00 10 10 03 21111 00 41 00 26 29 04 21112 00 39 00 24 30 07 21113 00 37 00 23 33 06 21121 00 33 00 14 37 15 21122 00 35 00 14 37 15 21123 00 36 00 14 37 13 21131 00 37 00 25 30 08 21132 00 35 00 28 29 08 21133 00 41 00 28 26 05 21211 00 35 00 21 32 11 21212 00 39 00 21 33 08 21213 00 39 00 21 31 09 21221 00 43 00 10 36 11 21222 00 41 00 11 37 11 21223 00 41 00 10 36 13 21231 00 37 00 21 33 10 21232 00 43 00 21 29 07 21233 00 39 00 21 33 07 21311 00 30 00 19 38 12 21312 00 35 00 16 37 13 21313 00 37 00 15 36 12 21321 00 38 00 11 39 11 21322 00 37 00 13 37 13 21323 00 38 00 11 . 38. 13 21331 00 48 00 18 29 OS 21332 00 48 00 16 31 OS 21333 00 43 00 14 34 08 22111 00 57 00 14 28 01 22112 00 41 00 18 33 08 22113 00 42 00 14 35 09 22121 00 41 00 08 42 09 22122 00 40 00 10 40 11 22123 00 40 00 08 41 11 22131 00 . 42 00 19 34 05 22132 00 40 00 20 32 07 22133 00 42 00 19 31 08 78 Table 16 (continued) Sample No. SMECT ILL VERM CHLOR KAOL QUARTZ 22211 00 42 00 21 30 07 22212 00 36 00 18 35 11 22213 00 40 00 , 22 30 09 22221 • 00 40 00 09 38 13 22222 00 40 00 09 38 13 22223 00 39 00 10 38 13 22231 00 44 00 16 33 06 22232 00 47 00 17 32 04 22233 00 41 00 17 35 07 22311 00 39 00 17 32 11 22312 00 35 00 17 36 12 22313 00 35 00 19 33 14 22321 00 39 00 11 37 14 22322 00 45 00 09 34 12 22323 00 39 00 10 38 13 22331 00 39 00 16 37 08 22332 00 39 00 18 34 09 22333 00 40 00 17 33 10 23111 00 42 00 17 30 11 23112 00 43 00 15 32 10 23113 00 41 00 18 30 11 23121 00 39 00 12 33 15 23122 00 41 00 10 34 15 23123 00 41 00 12 34 13 23131 00 44 00 19 29 08 23132 00 . 42 00 20 29 09 23133 00 42, 00 23 29 07 23211 00 43: 00 19 32 07 23212 00 45 00 18 30 08 23213 00 48 00 16 30 06 23221 00 42 00 09 34 14 23222 00 45 00 08 33 14 23223 00 43 00 12 33 12 23231 00 35 00 18 29 17 23232 00 34 00 25 27 14 23233 00 35 00 24 27 15 23311 00 ' 38 00 15 30 16 23312 00 39 00 14 28 18 23313 00 39 00 15 29 17 79 Table 16 (continued) Sample No. SMECT ILL VERM CHLOR KAOL QUARTZ 23321 00 35 00 14 34 16 23322 00 36 00 12 35 17 23323 00 39 00 12 31 18 23331 00 49 00 13 29 09 23332 00 48 00 18 29 05 23333 00 49 00 14 31 06 31111 '15 30 04 12 30 09 31112 16 32 01 13 31 07 31113 18 31 OS 13 25 08 31121 14 33 06 09 28 11 . 31122 13 ' 36 03 08 28 11 31123 15 36 00 08 29 11 31131 15 39 17 09 18 03 31132 15 28 18 07 26 07 31133 16. 38 10 09 25 03 31211 20 32 00 14 25 08 31212 17 33 02 14 25 08 31213 18 31 07 13 23 08 31221 14 37 09 OS 26 08 31222 12 38 07 06 29 09 31223 12 36 02 13 28 08 31231 21 33 01 16 23 07 31232 22 30 07 11 . 25 06 31233 24 27 08 13 21 07 31311 13 33 00 16 31 07 31312 14 32 04 12 30 08 31313 15 33 00 14 31 07 31321 14 37 04 06 30 10 31322 14 37 00 11 29 09 31323 10 33 13 05 .31 . 08 31331 20 30 11 09 23. .06 31332 20 29 12 07 26 06 31333 20 31 10 07 27 OS 32111 14 35 . os 09 30 07. 32112 14 44 02 12 ' 25 03. 32113 13 41 03 09 29 OS 32121 11 37 03 08 32 09 . 32122 10 37 , OS 07 32 09 32123 12 33 . 04 07 34 10 80 Table 16 (continued) Sample No. SMECT ILL VERM CHLOR KAOL QUARTZ 32131 16 41 07 08 25 03 32132 21 35 02 10 27 05 32133 17 36 07 12 24 04 32211 12 41 02 11 29 05 32212 13 46 01 09 28 03 32213 14 44 06 07 26 02 32221 13 37 04 07 33 06. 32222 12 37 05 06 34 06 32223 09 47 01 07 32 04 32231 18 37 08 08 25 04 32232 19 34 04 09 27 06 32233 18 36 07 09 25 . 05 32311 13 43 00 13 27 05 32312 12 41 03 10 29 05 32313 13 39 00 15 27 OS 32321 13 35 00 11 33 09 32322 11 36 03 07 34 09 32323 13 38 00 08 33 .09 32331 17 36 05 08 30 04 32332 17 36 08 10 24 OS 32333 17 35 11 08 24 OS 33111 14 44 02 12 21 07 33112 15 44 00 15 20 07 33113 19 37 02 14 21 08 33121 11 42 03 07 25 12 33122 05 40 01 10 25 10 33123 13 33 00 09 31 14 33131 . 26 34 03 10 19 08 33132 27 32 02 10 21 08 33133 22 30 08 11 23 07 33211 14 44 03 12 21 06 33212 15 40 01 12 25 07 33213 14 40 01 16 22 07 33221 12 35, 00 11 30 12 33222 12 43 ; 02 10 24 09 . 33223 .11 46 00 10 . 2 5 , 09 33231 18 34 00 21 20 07 33232 20 36 00 17 20 08 33233 17 38 00 16 21 07 81 Table 16 (continued) Sample No. SMECT ILL VERM CHLOR KAOL QUARTZ 33311 13 42 02 09 27 06 33312 13 42 06 07 26 05 33313 13 38 06 08 28 08 33321 11 38 05 04 34 08 33322 12 40 07 04 30 08 33323 13 39 03 07 38 09 33331 20 37. 07 07 . 22 09 33332 18 37 02 09 27 05 33333 17 37 09 10 22 06 82 Table 17. ANOVA for Main Effects on the 17A Peak (MgEG) for All Three Soils Studied- Preliminary Study Soxirce Degrees of Freedom Sum of Squares Mean Squares Variance Component F-Value Total 242 1108.160 4.579 6.635 Soils 2 972.767 . 486.383 5.997 780.43* Clay Sep. 6 3.739 0.623 -0.028 0.45 Slide 18 24.946 1.386 -0.045 0.77 Field/Slide 54 96.770 1.792 0.577 29.21* Error 162 9.937 - 0.061 0.061 * - significant at a = 0.05 Table 18. ANOVA for Main Effects on the IOA Peak (MgEG) for AU Three Soils Studied- Preliminary Study Source Degrees of Freedom Sumof Squares Mean Squares Variance Component F-Value Total 242 8.243 0.034 0.047 Soils 2 6.294 3.147 0.038 81.14* Clay Sep. 6 0.233 0.039 0.001 2.08 Slide . 18 0.336 0.019 0.000 0.97 Field/Slide 54 1.028 0.019 0.006 8.77* Error 162 0.353 0.002 0.002 * - significant at a = 0.05 I 83 Table 19. ANOVA for Main Effectsonthe 14A Peak (MgEG) for AU Three SoUs Studied— PreUminary Study Source Degrees of Freedom Sum of Squares Mean Squares Variance Component F-Value Total 242 4.991 0.021 0.028 Soils 2 2.043 1.021 0.012 16.34* Clay Sep. 6 0.375 0.063 0.002 10.64* SUde 18 0.106 0.006 -0.003 0.17 Field/SUde 54 1.854 0.034 0.010 9.07* Enor 162 0.613 . 0,004 0.004 * - significant at a = 0.05 Table 20. ANOVA for Main Effects on the 14A Peak (K350) for AU Three SoUs Studied— PreUminary Study Source Degrees of Freedom Sumof Squares Mean Squares Variance Component F-Value Total 242 5.170 0.021 0.027 Soils 2 0.550 * 0.275 0.003 18.45* Clay Sep. 6 0.089 0.015 0.000 0.59 SUde 18 0.451 0.025 -0.004 0.38 Field/SUde 54 3.534 0.065 0.021 19.45* Enor 162 0.545 0.003 0.003 * - significant at a = 0.05 84 Table 21. ANOVA for Main Effects on the 7A Peak (MgEG) for All Three Soils Studied- Preliminary Study Source Degrees of Freedom Sum of Squares Mean Squares Variance Component F-Value Total 242 134.659 0.143 0.185 Soils 2 20.615 10.307 0.121 21.82* Clay Sep. 6 2.835 0.472 0.005 1.44 Slide 18 5.903 0.328 0.028 4.16* Field/Slide 54 4.259 0.079 0.024 12.20* Error 162 1.047 0.006 0.006 * - significant at a = 0.05 Table 22. ANOVA for Main Effects on the 7A Peak (K350) for All Three Soils Studied- Preliminary Study Source Degrees of Freedom Sum of Squares Mean Squares Variance Component F-Value Total 242 44.365 0.183 0.224 Soils 2 20.073 10.036 0.102 5.76* Clay Sep. 6 10.450 1.742 0.056 7.67* Slide 18 4.087 0.227 0.005 1.28 Field/Slide 54 9.594 0.178 0.059 177.52* Error 162 0.162 0.001 0.001 * - significant at a = 0.05 85 Table 23. ANOVA for Main Effects on the 3.5 A Peak (MgEG) for All Three Soils Studied- Preliminary Study Source Degrees of Freedom Sum of Squares Mean Squares Variance Component F-Value Total 242 12.311 0.051 0.068 Soils 2 7.397 3.698 0.045 63.14* Clay Sep. 6 0.351 0.059 -0.002 0.49 Slide 18 2.137 0.119 0.100 3.75* Field/Slide 54 1.711 0.032 0.009 7.19* Error 162 0.714 0.004 0.004 * - significant at a = 0.05 Table 24. ANOVA for Main Effects on the 3.3 A Peak (MgEG) for All Three Soils Studied- Preliminary Study Source Degrees of Freedom Sum of Squares Mean Squares Variance Component F-Value Total 242 10.449 0.043 0.055 Soils 2 5.913 2,956 0.034 16.96* Clay Sep. 6 1.046 0.174 0.003 2.18 Slide 18 1.441 0.080 0.005 2.49* Field/Slide 54 1.734 0.032 0.010 16.49* Error . 162 0.315 0.002 . 0.002 * „ significant at a = 0.05 86 Table 25. ANOVA of Percent Smectite Composition for All Three Soils Studied—Prelimi­ nary Study Source Degrees of Freedom Sum of Squares Mean Squares Variance Component F-Value Total 242 16.925 0.070 0.104 Soil 2 16.596 8.298 0.102 1463.49* Clay Sep. 6 0.034 0.006 0.000 1.85 Slide 18 0.055 0.003 0.000 0.82 Field/Slide 54 0.201 0.004 0.001 15.28* Error 162 0.039 0.000 0.000 * - significant at a = 0.05 Table 26. ANOVA of Percent Illite Composition for All Three Soils Studied—Preliminary Study Source Degrees of Freedom Sum of Squares Mean Squares Variance Component F-Value Total 242 5.003 0.021 0.030 Soil 2 4.656 2.328 0.029 180.88* Clay Sep. 6 0.077 0.013 0.000 11.92* Slide 18 0.019 0.001 0.000 0.37 Field/Slide 54 0.158 0.003 0.001 5.09* Error 162 0.093 0.001 0.001 * - significant at a = 0.05 87 Table 27. ANOVA of Percent Vermiculite Composition for All Three Soils Studied—Pre­ liminary Study Source Degrees of Freedom Sum of Squares Mean Square . Variance Component F-Value Total 242 0.219 0.001 0.001 Soil 2 0.090 0.045 0.001 18.58* Clay Sep. 6 0.015 0.002 0.000 3.55* Slide 18 0.012 0.001 0.000 0.63 Field/Slide 54 0.059 0.001 0.000 4.06* Error 162 0.043 0.000 0.000 * - significant at a = 0.05 Table 28. ANOVA of Percent Chlorite Composition for All Three Soils Studied—Prelimi­ nary Study Source Degrees of Freedom Sum of Squares Mean Squares Variance Component F-Value Total 242 0.579 0.002 0.003 SoU 2 0.185 0.093 0.001 29.77* Clay Sep. 6 . 0.019 0.003 0.000 0.83 SUde 18 0.067 0.004 0.000 0.77 Field/SUde 54 0.261 0.005 0.001 16.66* Error 162 0.047 0.000 0.000 * - significant at a = 0.05 88 Table 29. ANOVA of Percent Kaolinite Composition for All Three Soils Studied—Prelimi­ nary Study Source Degrees of Freedom Sum of Squares Mean Square Variance Component F-Value Total 242 1.987 0.008 0.012 Soil 2 1.693 0.847 0.010 99.01* Clay Sep. 6 0.051 0.009 0.000 5.65* Slide 18 0.027 0.001 0.000 0.47 Field/Slide 54 0.172 0.003 0.001 11.68* Error 162 ' 0.044 0.000 0.000 * - significant at a = 0.05 Table 30. ANOVA of Percent Quartz Composition for AU Soils Studied—Preliminary Study Source Degrees of Freedom Sum of Squares Mean Squares Variance Component F-Value Total 242 0.291 0.001 0.001 Soil 2 ■ 0.114 0.057 0.001 16.06* Clay Sep. 6 0.021 0.003 0.000 4,06* Slide 18 . 0.016 0.001 0.000 0.40 Field/Slide 54 0.119 0.002 0.001 17.03* Error 162 0.021 0.000 0.000 * - significant at a = 0.05 Table 3 1. Complete Peak Area Measurements (in.2 ) for the Main Study- Sample No. 17AMgEG IOAMgEG HAMgEG I4AK350 7AK350 7AK500 7AMgEG 3.5 A MgEG 3.3AMgEG 11 4.45 0.35 0.52 0.59 0.18 0.00 0.73 0.47 0.47 12 5.65 0.38 0.60 0.62 0.26 0.00 0.65 0.41 0.44 21 2.40 0.39 0.59 0.41 0.27 0.00 0.63 0.43 0.67 22 2.50 0.33 0.34 0.49 0.48 0.00 0.36 0.37 0.58 31 1.39 0.23 0.19 0.42 0.25 0.00 0.36 0.38 0.61 32 2.09 0.28 0.19 0.54 0.27 0.00 0.48 0.54 41 2.07 0.14 0.21 0.44 0.39 0.00 0.43 0.23 0.36 42 1.85 0.12 0.18 0.52 0.18 0.00 0.59 0.50 SI 5.17 0.09 0.25 0.33 0.73 0.00 0.61 0.33 0.33 52 2.89 0.06 0.17 0.48 0.21 0.00 0.60 0.42 61 8.56 . 0.22 0.72 0.66 0.39 0.00 1.38 0.59 0.56 62 4.27 0.14 0.30 0.62 0.47 0.00 0.76 0.55 71 15.53 0.28 0.00 0.75 0.22 0.00 1.01 0.52 0.73 72 7.20 0.16 0.35 0.59 0.13 0.00 0.66 0.37 0.65 SI 5.04 0.81 1.29 0.90 0.29 0.00 2.00 1.20 1.10 82 3.19 0.46 - 0.63 0.63 0.28 0.00 1.32 0.71 0.72 91 0.93 0.79 0.42 0.80 0.45 0.00 1.98 0.91 0.62 92 0.91 0.68 0.60 0.88 0.53 0.00 1.85 1.31 0.78 IOl 0.35 0.53 0.20 0.69 0.46 0.00 1.32 1.05 1.08 102 1.21 0.84 0.69 0.61 0.44 0.00 1.50 1.01 0.91 II I 0.22 0.99 0.31 0.89 0.57 0.00 2.51 1.35 0.87 .112 0.51 0.89 0.49 0.60 2.11 0.00 2.73 1.40 0.85 121 0.13 1.21 0.31 0.59 1.83 0.00 3.20 1.76 1.04 122 0.31 0.72 0.49 0.69 1.54 0.00 1.80 1.20 0.74 131 0.00 1,01 0.35 0.72 1.64 0.00 2.81 2.24 1.32 132 0.00 0.71 0.35 0.58 1.65 0.14 2.55 1.15 0.75 141 4.61 0.40 0.35 0.42 0.36 0.00 0.72 0.50 0.53 142 4.10 0.38 0.22 0.47 0.44 0.00 0.55 0.36 0.38 151 5.52 0.14 0.35 0.55 0.14 0.00 0.44 0.27 0.44 Table 31 (continued) Sample No. 17AMgEG IOAMgEG 14AMgEG 14AK350 7AK350 7AK500 7AMgEG 3.5 A MgEG 3.3A MgEG 152 6.80 0.17 O.O0 0.41 0.12 0.00 0.41 0.27 0.37 161 4.51 0.11 0.00 1.17 0.19 0.00. 0.29 0.30 0.70 162 14.61 0.19 0.00 0:52 0.24 0.00 0.42 0.21 0.54 171 6.30 0.11 0.00 0.65 0.21 0.00 0.22 0.21 0.55 -172 15.93 0.28 O.O0 0.79 0.22 0.00 0.55 0.35 1.06 181 2.68 0.16 0.00 - 0.78 0.16 0.00 0.15 . 0.27 0.46 182 13.30 0.35 0.00 0.70 0.11 0.00 0.48 0.34 1.03 191 2.57 . 0.19 0.00 0.68 0.13 0.00 0.23 0.68 192 16.40 0.38 0.00 0.84 0.13 o.oo 0.51 0.35 0.88 201 8.30 0.57 0.00 1.15 0.00 0.00 0.87 0.61 1.03 202 7.57 0.63 0.00 1.31 0.00 0.00 0.93 0.43 0.65 211 6.25 0.60 0.00 0.91 0.12 0.00 0.97 0.73 1.17 212 9.71 . 0.95 0.00 1.24 0.00 0.00 1.01 0.65 0.87 221 2.59 0.28 0.00 1.08 0.60 0.00 0.49 0.45 0.72 222 11.00 0.53 0.00 0.68 0.14 0.00 1.01 . 0.67 0.99 231 8.50 0.22 0.00 1.13 0.53 0.00 0.52 0.75 232 4.59 0.16 0.00 0.68 0.16 0.00 0.38 0.63 241 11.00 0.12 0.00 0.79 0.34 0.00 0.39 0.77 . 242 3.22 0.11 0.00 0.73 0.16 0.00 0.29 0.74 251 8.50 0.34 0.00 0.99 0.19 0.00 0.79 0.62 252 6.70 0.30 0.00 1.01 0.08 0.00 0.95 0.47 0.62 261 8.90 0.50 0.00 1.20 0.13 0.00 1.02 0.46 0.44 262 8.28 0.52 0.00 1.07 0.15 0.00 1.33 0.71 0.71 271 8.87 0.87 0.00 0.80 3.60 0.00 4.13 2;03 1.21 272 8.16 0.88 0.00 0.74 2.84 0.00 4.16 2.11 0.91 281 1.22 1.51 1.18 0.92 0.82 0.00 3.30 2.36 2.10 282 1.57 1.33 0.63 0.69 1.05 0.00 3.24 2.62 1.90 291 0.00 1.59 0.41 0.13 4.72 0.00 5.59 3.78 1.47 292 0.00 1.50 0.23 0.21 5.63 0.00 5.06 3.13 1.15 Table 31 (continued) Sample No. 17AMgEG IOAMgEG 14AMgEG 14AK350 7AK350 7AK500 TAMgEG. 3.5AMgEG 3.3 A MgEG 301 0.00 LlO 0.00 0.00 4.63 0.00 6.06 3.19 0.73 302 0.00 0.57 0.00 . 0.16 2.46 0.00 2.44 1.38 0.35 311 0.00 1.37 0.00 0.23 0.94 0.00 5.03 2.59 1.20 312 0.00 1.61 0.00 0.21 1.79 0.00 4.59 2.64 1.09 321 0.00 . 2.03 0.44 0.49 2.10 0.00 3.03 2.15 1.91 322 0.00 2.05 0:62 0.61 2.97 0.00 3.56 2.05 1.61 331 0.00 1.65 0.39 0.58 2.03 0.00 3.63 2.31 1.60 332 0.00 1.91 0.49 0.71 2.53 0.00 3.75 2.07 1.54 341 0.00 1.97 0.56 0.56 3.25 0.00 3.47 1.91 1.53 342 0.00 1.92 0.54 0.66 1.24 0.00 3.22 2.11 1.33 351 0.00 1.15 0.00 0.00 1.67 0.00 2.23 1.42 0.89 352 0.00 „ 0.76 0.00 0.00 0.68 0.00 1.93 1.03 0.71 361 0.00' 1.69 0.00 0.00 3.97 0.00 5.22 2.75 1.00 362 0.00 1.40 0.44 0.28 1.12 0.00 4.69 2.69 1.11 371 0.00 1.20 0.00 . 0.00 1.04 0.00 4.78 2.49 0.89 372 0.00 1.29 0.00 0.22 • 0.81 0.00 4.95 2.41 0.90 381 8.53 0.62 0.00 0.55 0.79 0.00 1.44 1.03 1.40 382 8.45 0.58 0.00 0.49 0.61 0.00 1.33 0.94 1.13 391 6.60 0.88 0.95 0.54 0.42 0.00 1.65 1.49 1.51 392 5.00 0.69 1.05 0.81 0.49 0.00 1.60 1.31 1.67 401 5.85 0.44 0.00 0.45 0.20 0.00 0.82 0.55 0.47 402 5.12 0.53 0.00 0.31 . 0.23 0.00 0.92 0.48 0.56 411 5.82 0.53 0.80 0.34 0.11 0.00 0.89 0.49 0.58 412 6.24 0.63 0.96 0.30 0.32 0.00 0.97 0.73 0.69 421 0.95 1.26 0.45 0.49 1.15 0.00 2.40 1.43 1.19 422 . 1.27 1.11 0.45 0.34 1.69 0.00 2.43 1.33 1.10 431 0.00 1.30 0.48 0.40 3.36 0.00 2.66 1.61 1.24 432 0.00 1.58 0.41 0.41 0.96 0.00 2.60 1.39 1.00 441 ODQ 1.34 0.43 0.57 3.13 0.00 3.97 2.15 1.07 Table 31 (continued) Sample No. 17AMgEG IOAMgEG 14AMgEG 14AK350 7AK350 7AK500 7AMgEG 3.5A MgEG 3.3A MgEG 442 0.00 1.40 0.48 0.54 2.63 0.00 3.81 2.06 1.08 451 0.00 0.64 0.00 0.00 0.29 0.00 1.90 1.24 0.95 452 0.00 0.68 0.00 0.18 0.24 0.00 1.90 1.21 1.01 461 0.00 0.84 0.57 0.81 1.13 0.00 7.40 4.63 0.86 462 0.00 0:77 0.55 0.71 1.56 0.00 7.60 4.03 0:80 471 0.00 1.21 0.62 0.49 4.25 0.00 7.50 5.37 1.45 472 0.00 1.00 0.58 0.54 4.06 0.00 7.25 3.80 1.10 481 0.00 1.30 0.63 0.56 4.06 0.00 6.55 4.59 1.39 482 0.00 1.28 0.54 0.41 2.33 0.00 6.35 3.75 1.20 491 0.00 1.35 0.54 0.45 1.14 0.00 3.59 2.81 1.37 492 0.00 1.39 0.54 0.59 0.87 0.00 4.25 2.81 1.52 501 1.53 0.70 0.00 0.39 0.10 0.00 1.75 1.65 1.69 502 2.27 0.90 0.00 0.27 0.14 0.00 2.21 1.57 1.76 VOto 93 Table 32. Complete Data Table for Modification IV PSM PIL PVM PCA PKA PQU ECEC LCEC TKIL8 I 54.3 15.7 0.0 13.1 15.0 2.0 62.70 56.25 15 2 39.7 22.6 0.0 14.1 17.9 5.6 50.43 47.50 15 3 35.2 21.2 0.0 18.3 17.1 8.2 46.62' 42.50 14 4 43.3 10.7 0.0 17.1 21.7 7.1 52,18 42.50 13 5 58.9 4.7 0.0 12.7 18.7 5.0 64.88 51.25 8 6 57.3 6.7 0.0 12.4 19.3 4.1 63.71 46.25 11 7 70.1 5.7 0.0 9.2 11.1 3.7 74.85 62.50 8 8 34.4 21.0 0.0 13.2 27.8 3.7 45.18 32.50 20 9 9.6 30.9 0.0 17.7 40:2 1.6 25.03 21.25 27 10 9.0 33.3 0.0 16.2 35.0 6.3 24.36 20.00 30 11 3.2 34.3 0.0 13.4 47.6 1.4 19.02 20.00 27 12 1.6 36.7 0.0 . 12.7 47.3 . 1.7 17.76 17.50 27 13 0.0 33.0 0.0 12.6 50.7 3.6 15.52 17.50 29 14 51.3 19.6 0.0 11.3 15.8 1.9 60.36 40.00 19 15 67.4 7.6 0.0 11.6 10.2 3.5 73.02 58.75 . 14 16 75.6 4.3 0.0 13.1 4.7 2.2 80.37 72.50 11 17 72.3 5.5 0.0 11.8 5.4 4.9 77.17 62.50 11 18 72.8 8.1 0.0 9.1 5.6 4.5 77.64 63.75 17 19 37.7 15.2 0.0 27.2 9.2 10.7 49.25 42.50 19 20 48.7 17.5 0.0 17.9 13.1 2.8 58.65 48.75 20 21 46.8 21.0 0.0 14.8 13.9 3.4 56.98 43.75 23 22 59.0 12.6 0.0 12.8 12.0 3.5 66.38 48.75 22 23 57.6 8.1 0.0 18.8 9.6 5.9 65.20 46.25 17 24 47.6 6.6 0.0 24.3 10.7 10.8 56.38 53.75 15 25 54.5 11.1 0.0 16.7 14.6 3.1 62.69 47.50 18 26 52.0 14.2 0.0 15.8 16.4 1.5 60.85 37.50 26 27 36.0 16.3 0.0 7.2 38.6 1.8 45.05 28.75 22 28 12.2 35.0 0.0 10.1 37.3 5.4 26.51 22.50 29 29 0.0 35.1 1.9 1.7 60.4 0.8 17.30 17.50 24 30 0.0 28.6 0.0 2.2 69.2 0.0 13.24 12.50 21 31 0.0 37.0 0.0 2.7 59.7 0.5 14.71 17.50 23 32 0.0 50.2 0.0 7.4 40.5 1.9 17.68 12.50 37 33 0.0 44.3 0.0 8.0 46.1 1.5 16.81 12.50 38 34 0.0 49.5 0.0 7.7 42.5 0.2 17.72 10.00 33 35 0.0 46.9 0.0 0.0 51.9 1.1 15.90 23.75 27 36 0.0 36.0 3.5 2.4 57.9 0.2 20.46 17.50 24 37 0.0 33.0 0.0 1.4 65.2 0.0 13.91 17.50 24 38 53.2 16.0 0.0 6.9 18.4 5.4 60.52 47,40 19 39 33.7 20.5 0.0 7.6 31.5 6.5 43.43 42.50 22 40 52.6 20.1 0.0 7.8 18.0 1.5 61.08 47.50 20 Table 32 (continued) PSM PIL PVM PCA PKA PQU ECEC LCEC TKIL8 41 49.6 19.0 9.1 5.3 15.3 1.6 72.95 38.75 21 42 9.1 40.3 0.9 6.3 41.7 1.6 25.71 17.50 30 43 0.0 47.6 0.0 7.2 44.1 1.1 17.24 12.50 33 44 0.0 38.0 0.0 7.7 53.9 0.3 15.76 12.50 33 45 0.0 35.9 0.0 6.0 51.6 6.5 14.72 31.25 23 46 0.0 16.1 0.0 7.6 75.1 1.1 11.97 20.00 19 47 0.0 20.6 0.6 5.2 71.1 2.5 13.24 17.50 20 48 0.0 26.4 1.0 4.9 66.0 1.7 14.99 17.50 20 49 0.0 37.2 0.0 6.9 53.0 2.8 15.33 12.50 34 50 14.3 30.2 0.0 7.0 37.6 10.9 26.85 30.00 24 95 Table 33. Complete Data Table for Modification V PSM PIL PVM PCA PKA PQU ECEC LCEC TKIL8 I 47.2 13.7 0.0 11.3 26.0 1.7 55.58 56.25 15 2 33.7 19.2 0.0 12.0 30.3 4.7 44.07 47.50 15 3 30.0 18.2 0.0 15.6 29.2 7.0 40.93 42.50 14 4 35.6 8.8 0:0 14.1 35.6 5.8 44.33 42.50 13 5 49.8 3.9 0.0 10.6 31.3 4.2 56.04 51.25 8 6 48.0 5.6 0.0 10.4 32.4 3.5 54.68 46.25 11 7 63.2 5.2 0.0 8.2 20.1 3.3 68.23 62.50 8 8 26.9 16.4 o:o 10.3 43.5 2.8 37.12 32.50 20 9 6.9 22.0 0.0 12.6 57.4 LI 20.17 21.25 27 10 6.7 24.8 . 0.0 11.9 51.9 4.7 20.17 20.00 30 11 2.2 23.2 o:o 9.1 64.5 1.0 15.47 20.00 27 12 1.1 24.9 0.0 8.6 64.2 1.2 14.63 17.50 27 13 0.0 22.0 0.0 8.3 67.3 2.4 13.01 17.50 29 14 44.3 16.9 o.o . 9.7 27.2 1.7 53.23 40.00 19 15 61.2 6.7 0.0 10.5 18.5 3.1 67.04 58.75 14 16 72.2 4.1 0.0 12.6 9.0 2.1 77.14 72.50 11 17 68.5 5.2 0.0 11.2 10.3 4.7 73.58 62.50 11 18 68.9 7.7 0.0 8.6 10.6 4.2 73.91 63.75 17 19 34.6 13.9 0.0 24.9 16.8 9.8 45.84 42.50 19 20 43.0 15.5 0.0 15.8 23.1 2.5 52.76 48.75 20 21 41.2 18.5 0.0 13.0 24.3 3.0 51.08 43.75 23 22 52.7 11.3 0.0 11.5 21.5 3.1 60.18 48:75 22 23 52.5 . 7.4 0.0 17.1 17.4 5.4 60.19 46.25 17 24 42.9 6.0 0.0 21.9 19.3 9.8 51.67 53.75 15 25 47.6 9.8 0.0 14.6 25.4 2.6 55.81 47.50 18 26 44.7 12.3 o,o 13.6 28.1 1.3 53.49 37.50 26 27 26.0 11.8 0.0 5.1 55.7 1.3 34.71 28.75 22 28 8.9 25.4 0.0 7.3 54.3 3.9 21.52 22.50 29 29 0.0 21.9 1.2 1.0 75.3 0.5 13.79 17.50 24 30 0.0 16.9 . 0.0 1.3 81.8 0.0 11.11 12.50 21 31 0.0 23.2 0.0 1.7 74.7 0.3 12.22 17.50 23 32 0.0 35.7 0.0 5.2 57.6 1.3 14.89 12.50 37 33 0.0 30.3 0.0 5.5 63.1 1.0 14.03 12.50 38 34 0.0 34.7 0.0 5.4 59.7 0.1 14.83 10.00 33 35 0.0 30.9 0.0 0.0 68.3 0.7 13.22 23.75 27 36 0:0 22.8 2.3 1.5 73.4 0.1 15.88 17.50 24 37 0.0 20.2 0.0 0.8 78.9 0.0 11.58 17.50 24 38 44.9 13.5 0.0 5.8 31.2 4.5 52.32 47.40 19 39 25.7 15,6 0.0 5.7 47.9 5.0 34.97 42.50 .22 40 44.6 16.9 0.0 6.6 30.4 1.3 52.95 47.50 20 96 Table 33 (continued) PSM PIL PVM PCA PKA PQU ECEC LCEC TKIL8 41 43.1 16.5 7.9 4.6 26.6 1.4 64.27 38.75 21 42 6.4 28.5 0.7 4.5 58.9 1.1 20.60 17.50 30 43 0.0 33.1 0.0 4.9 61.2 0.8 14.41 12.50 33 44 0.0 24.7 0.0 5.0 70.1 0.2 13.04 12.50 33 45 0.0 23.6 0.0 3.9 68.0 4.3 12.42 31.25 23 46 0.0 9.2 0.0 4.3 85.8 0.6 10.28 20.00 19 47 0.0 12.0 0.3 3.0 83.1 1.4 11.07 17.50 20 48 0.0 15.9 0.6 3.0 79.5 1.0 12.16 17.50 20 49 0.0 24.3 0.0 4.5 69.3 1.9 12.81 12.50 34 50 10.4 22.0 0.0 5.1 54.5 . 7.9 21.71 30.00 . 24 97 Table 34. Complete Data Table for Modification IX. PSM PIL PVM PCA PKA PQU ECEC LCEC TKIL8 I 48.6 17.7 0.0 14.6 16.8 2.2 58.11 56.25 15 2 34.5 ' 24.6 0.0 15.4 19.4 6.1 46.18 47.50 15 3 30.3 22.8 0.0 19.7 18.4 8.8 42.52 42.50 14 4 37.9 11.7 0.0 18.8 23.8 7.8 47.62 42.50 13 5 59.3 5.7 0.0 13.8 17.5 3.7 65.65 51.25 8 6 51.8 7.5 0.0 14.0 21.9 4.7 59.08 46.25 11 7 65.3 6.7 0.0 10.6 12.9 4.3 70.81 62.50 8 8 29.5 22.5 0.0 14.1 29.9 3.9 41.12 32.50 20 9 7.8 31.4 0.0 18.0 41.0 1.6 23.54 21.25 27 10 7.4 34.0 0.0 16.5 35.7 6.4 23.00 20.00 30 11 2.6 34.5 0.0 13.5 47.9 1.4 18.52 20.00 27 12 1.3 36.8 0.0 12.7 47.5 1.7 17.51 17.50 27 13 0.0 33.0 0.0 12.6 50.7 3.6 15.52 17.50 29 14 45.8 21.8 0.0 12.5 17.6 2.1 55.86 40.00 19 15 62.3 8.6 0.0 13.4 11.8 4.0 68.81 58.75 14 16 71.3 5.0 0.0 15.5 5.6 2.6 76.93 72.50 11 17 67.6 6.4 0.0 13.7 6.3 5.8 73.31 62.50 11 18 68.1 9.5 0.0 10.6 6.5 5.2 73.75 63.75 17 19 32.7 16.4 0.0 29.4 9.9 11.6 45.17 42.50 19 20 43.2 19.3 0.0 19.8 14.6 3.1 54.18 48.75 20 21 41.4 23.2 0.0 16.3 . 15.4 3.7 52.58 43.75 23 22 53.5 14.3 0.0 14.6 13.6 4.0 61.89 48.75 22 23 52.1 9.1 0.0 21.2 10.9 6.7 60.70 46.25 17 24 42.1 , 7.3 0.0 26.8 . 11.8 11.9 51.81 53.75 15 25 49.0 12.5 0.0 18.7 16.3 3.4 58.18 . 47.50 18 26 46.5 15.9 0.0 17.7 18.3 1.7 56.33 .37.50 26 27 31.0 17.5 0.0 7.7 41.6 2.0 40.75 28.75 22 28 10.0 35.9 0.0 10.4 38.2 5.6 24.72 22.50 29 29 0.0 35.1 1.9 1.7 60.4 0.8 17.30 17.50 24 30 0.0 28.6 0.0 2.2 69.2 0.0 13.24 12.50 21 31 0.0 37.0 0.0 2.7 59.7 0.5 14.71 17.50 23 32 0.0 50.2 0.0 7.4 40.5 1.9 17.68 12.50 37; 33 0.0 44.2 0.0 8.0 46.1 1.5 16.81 12.50 38 34 0.0 49.5 0.0 . 7.7 42.5 0.2 17.72 10.00 33 35 0.0 46.9 0.0 0.0 51.9 1.1 15.90 23.75 27 36 0.0 36.0 3.5 2.4 57.9 0.2 20.46 17.50 24 37 OO 33.3 0.0 1.4 65.2 0.0 13.91 17.50 24 38 47.6 17.9 0.0 7.7 20.6 6.0 55.79 47.40 19 39 28.9 22.0 0.0 8.1 33.8 7.0 39.34 42.50 22 40 47.0 22.4 0.0 817 20.0 1.7 56.48 47.50 20 98 Table 34 (continued) PSM PIL PVM PCA PKA PQU ECEC LCEC TKIL8 41 44.0 21.2 10.1 5.9 16.9 1.9 69.95 38.75 21 42 7.9 41.0 1.0 6.5 42.5 1.6 24.47 17.50 30 43 0.0 47.6 0.0 7.2 44.1 1.1 17.24 12.50 33 44 0.0 38.0 0.0 7.7 53.9 0.3 15.76 12.50 33 45 0.0 35.9 0.0 6.0 51.6 6.5 14.72 31.25 23 46 0.0 16.1 0.0 7.6 75.1 1.1 11.97 20.00 19 47 0.0 20.6 0.6 5.2 71.1 2.5 13.24 17.50 20 48 0.0 26.4 1.0 4.9 66.0 1.7 14.99 17.50 20 49 0.0 37.2 0.0 6.9 53.0 2.8 15.33 12.50 34 50 11.8 31.2 0.0 7.2 38.6 11.1 24.72 30.00 24 99 Table 35.. Complete Data Table for Modification X PSM PIL PVM PCA PKA PQU ECEC LCEC TKIL8 I 41.7 15.1 0.0 12.5 28.7 1.9 50.95 56.25 15 2 28.9 20.6 0.0 12.9 32.5 5.0 40.01 47.50 15 3 25.6 19.3 0.0 16.6 31.0 7.4 37.17 42.50 14 4 30.8 9.5 0.0 15.2 38.3 6.3 40.1 i 42.50 13 5 50.5 4.9 0.0 11.7 29.8 3.1 57.10 51.25 8 6 42.5 6.1 0.0 11.5 35.9 3.8 49.92 46.25 11 7 57.9 5.9 0.0 . 9.5 22.9 3.8 63.66 62.50 8 8 22.7 17.3 0.0 10.9 46.0 . 3.0 33.49 32.50 20 9 5.6 22.3 0.0 12.8 58.2 1.1 19.05 21.25 27 10 5.5 25.0 0.0 12.1 ... 52.6 4.7 19.09 20.00 30 11 1.7 23.3 0.0 9.1 64.8 1.0 15.07 20.00 27 12 0.8 25.0 0.0 8.6 64.2 1.2 14.43 17.50 27 13 0.0 22.0 0.0 8.3 67.3 2.4 13.01 17.50 29 14 38.9 18.5 0.0 10.7 29.9 1.8 48.65 40.00 19 15 55.8 7.6 0.0 11.9 21.1 3.6 62.44 58.75 14 16 67.5 4.8 0.0 14.7 10.5 2.5 73.27 72.50 11 17 63.6 . 6.0 0.0 13.0 11.9 5.4 69.43 62.50 11. 18 63.9 8.9 0.0 10.0 12.3 4.9 69.71 63.75 17 19 29.7 14.9 0.0 26.7 . 18.1 , 10.6 41.76 42.50 19 20 37.7 .16.9 0.0 17.4 25.3 2.7 48.39 48.75 20 21 . 35.9 20.1 0.0 14.1 26.5 3.2 . 46.71 43.75 23 22 47.1 12.6 0.0 12.8 ' 24.0 3.5 . 55.44 48.75 22 23 47.0 8.2 0.0 19.2 19.5 6.0 55.57 46.25 17 24 37.7 6.5 0.0 23.9 21.1 10.7 47.18 53.75 15 25 42.2 10.8 0.0 16.0 28.0 3.0 51.20 47.50 18 26 39.3 13.4 0.0 15.0 30.8 1.4 48.91 37.50 26 27 22.0 12.4 0.0 5.4 58.7 1.4 31.15 28.75 22 28 7.2 25.9 0.0 7.5 55.3 4.0 20.05 22.50 29 29 0.0 21.9 1,2 1.0 75.3 0.5 13.79 17.50 24 30 0.0 16.9 6,0 1.3 81.8 0.0 11.11 12.50 21 31 0.0 23.2 0.0 1.7 74.7 0.3 12.22 17.50 23 32 0.0 35.7 0.0 5.2 57.6 1.3 14.89 12.50 37 33 0.0 30.3 0.0 5.5 63.1 1.0 . 14.03 12.50 38 34 0.0 34.7 0.0 5.4 59.7 0.1 14.83 10.00 33 35 0.0 30.9 0.0 0.0 68.3 0.7 13.22 23.75 27 36 0.0 22.8 2.3 1.5 73.4 0.1 15.88 17.50 24 37 0.0 20.2 0.0 * 0.8 78.9 0.0 11.58 17.50 24 38 39.5 14.8 0.0 6.4 34.2 5.0 47.60 47.40 19 39 21.7 16.4 0.0 6.1 50.5 5.3 31.48 42.50 22 40 39.2 18.6 0.0 7.3 33.4 1.4 48.38 47.50 20 100 Table 35 (continued) PSM PIL PVM PCA PKA PQU ECEC LCEC TKIL8 41 37.7 18.0 8.7 5.0 29.0 1.5 60.99 38.75 21 42 5.2 28.8 0.7 4.5 59.6 1.1 19.55 . 17.50 30 43 0.0 33.1 0.0 4.9 61.2 0.8 14.41 12.50 33 44 0.0 24.7 0.0 5.0 70.1 0.2 13.04 12.50 33 45 0.0 23.6 0.0 3.9 68.0 4.3 12.42 31.25 23 46 0.0 9.2 0.0 4.3 85.8 0.6 10.28 20.00 19 47 0.0 12.0 0.3 3.0 83.1 1.4 11.07 17.50 20 48 0.0 15.9 0.6 3.0 79.5 1.0 12.16 17.50 20 49 0.0 24.3 0.0 4.5 69.3 1.9 12.81 12.50 34 50 8.5 22.5 0.0 5.2 55.7 8.1 20.05 30.00 24 MONTANA STATE UNIVERSITY LIBRARIES 3 1762 00116369 8 N378 H778 Hopper, Roger W. E. cop .2 A sem i-quantita tive x-ray d if fr a c t io n technique fo r e stim a tic o f sm ectite , i l l i t e , . D A T E I S S U E D T O $