P‘STASSHJM EELEfLSE CHARMTERESTECS 0F SELECTED PEBUVlkN SGELS “20st: Fem i'he {Jayme at; Dim D. ‘iiCQ'Cé‘i STATE UMVER if? Denaid El. Gelsiigie 29H .A.... a a”... This is to certify that the thesis entitled POTASSIUM RELEASE CHARACTERISTICS OF SELECTED PERUVIAN SOILS presented by Donald D. Oelsligle has been accepted towards fulfillment of the requirements for Ph.D. Soil Science degree in (\> } , Wt mew Majot professor ‘_._-_L__‘_.__— ABSTRACT POTASSIUM RELEASE CHARACTERISTICS OF SELECTED PERUVIAN SOILS BY Donald D. Oelsligle Results from soil fertility experiments in Peru have indicated that the ability to predict a response to added potassium (K) fertilizer is not adequate. To eval- uate the K status of some Peruvian soils, samples from 16 soils were shipped to Michigan State University where they were intensively cropped using the Stanford-DeMent tech— nique, and K release was studied using sodium tetraphenyl- boron (NaTPB) for varying time periods. The amounts of K removed and rates of K release were then related to the clay mineral composition of the soil, and the soil solu- tion, exchangeable, and total K contents of the soils. When 1 g soil samples were equilibrated with solu- tions containing NaTPB for 0.25, l, 10, 100, 1000 and 2000 hours, a linear relationship was found between K removed by NaTPB and the logarithm of time from 10 to 1000 hr. Rates of K release varied from -0.03 to 7.43 me K/lOO g l Donald D. Oelsligle soil/wk. Large differences in the total amount of K re- moved after 2000 hr (1.9 to 66.9 me K/100 9 soil) were well related to the presence and apparent degree of crys- tallinity of the illite in the clay fraction. Variations in rate of K release and sources of plant—available K appeared to explain why a poorer relationship between the K extracted with NaTPB for 15 min and K uptake by cropping (r = 0.643) was obtained than has previously been reported. Amounts of plant-available K and rates of K release as determined from cumulative K uptake from five successive, two-week oat crops varied from 0.35 to 3.24 me K/lOO g soil and 0.016 to 0.215 me K/lOO g soil/wk, respectively. Total K uptake was not well correlated with K extracted with ammonium acetate (r = 0.388); the correlation was improved by considering K uptake from only the first of the five crops (r = 0.693) and was much higher when only K uptake from those soils with low K supplying capabilities were used in computing the correlation coefficient (r = 0.979). Total K in the soils was better related with K uptake by plants as cropping progressed indicating that higher amounts of K were being removed from nonexchangeable forms. Potassium uptake from those soils which released high amounts of K by cropping and with NaTPB was correlated 2 Donald D. Oelsligle somewhat higher with the percent K in the sand than with the percent K in the silt, which suggests that these soils are relatively unweathered. The concentration of K in the soil solution was not correlated with K uptake by plants for any of the crOppings. POTASSIUM RELEASE CHARACTERISTICS OF SELECTED PERUVIAN SOILS BY Donald D. Oelsligle A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Cr0p and Soil Sciences 1972 . /"- .- TO JEAN This thesis is affectionately dedicated to my wife. . ii ACKNOWLEDGEMENT The author wishes to express appreciation to Dr. E. C. Doll for his special interest, patience and guidance throughout this study and to Dr. R. L. Donahue, Dr. B. G. Ellis, Dr. B. D. Knezek and Dr. C. J. Pollard for serving as members of the author's guidance committee. Grateful acknowledgement is given the Department of Crop and Soil Sciences at Michigan State University for financial assistance and to the faculty members, fel- low graduate students, and personnel of the Soil Testing Laboratory who assisted during this study go my profound thanks. Appreciation is given to the North Carolina State University USAID mission to Peru which provided financial aid for a part of this project and to its members in Peru who provided constant help in logistics during the author's stay there. Special thanks is given to Dr. R. E. McCollum for his assiStance and encouragement in the initial stages of this study. The author also wishes to thank the personnel of the Soil Science Department at the La Molina Agricultural iii Experiment Station for their help in the laboratory work conducted there. Special thanks is due Dr. Carlos Val- verde S. for his continued assistance during the initial stages of the project in Peru. Assistance by members of the International Soil Fertility Evaluation and Improvement Program is gratefully acknowledged. iv TABLE OF CONTENTS Page LIST OF TABLES O 0 O O O O O O O O 0 O O O O O O 0 O Viii LIST OF FIGURES O O O O O O O O O O O O O O O O O O Xi LIST OF PI‘ATES. O O O O O O O O O O O O O O O I O O Xii INTRODUCTION. . . . . . . . . . . . . . . . . . . . 1 LITERATURE REVIEW . . . . . . . . . . . . . . . . . 3 The Soils of Peru . . . . . . . . . . . . . . . 3 Arid western coast. . . . . . . . . . . . . 3 Central Andean highlands (sierra) . . . . . 4 Eastern tropical lowlands . . . . . . . . . 5 Forms of Soil Potassium . . . . . . . . . . . . 6 Framework silicates . . . . . . . . . . . . 6 Structure . . . . . . . . . . . . . . . 6 Methods of potassium release. . . . . . 7 Layer silicates . . . . . . . . . . . . . . 8 Structures. . . . . . . . . . . . . . . 8 Potassium release mechanisms. . . . . . 9 Factors affecting weathering rates. . . 11 Release of Soil Potassium to Plants . . . . . . 15 METHODS AND MATERIALS 0 O O O O O O O O O O O O I O l 9 Sampling Procedure and Supporting Data. . . . . 19 Soil Franction Analysis . . . . . . . . . . . . 22 TABLE OF CONTENTS (Cont.) Separation of soil fractions. . . . . . X-ray analysis. . . . . . . . . . . . . Total potassium . . . . . . . . . . . . Short-term Intensive Cropping . . . . . . . Cropping studies in Peru. . . . . . . . Cropping study in Michigan. . . . . . . Sodium Tetraphenylboron Extractions . . . . Soil Solution Potassium Determinations. . . RESULTS . . . . . . . . . . . . . . . . . . . . Soil Chemical and Mineralogical Properties. Chemical measurements on soil horizons. Total potassium in surface samples. . . Clay mineralogy of surface samples. . . Sodium fluoride test for allophane. X-ray analYSiS. o o o o o o o o o o Short-term Intensive Cropping . . . . . . . Cropping studies in Peru. . . . . . . . Cropping study in Michigan. . . . . . . Sodium Tetraphenylboron Extractions . . . . Soil Solution Potassium Determinations. . . DISCUSSION. . . . . . . . . . . . . . . . . . . Potassium Release Patterns of Soil Groups . Soils with low potassium supplying capabilities. . . . . . . . . . . . Soils 3 and 29. . . . . . . . . . . Soil 4. . . . . . . . . . . . . . . Soils 12 and 31 . . . . . . . . . . Soils 5, 8 and 21 . . . . . . . . . Soils 6 and 10. . . . . . . . . . . vi Page 22 24 25 26 26 27 30 31 32 35 35 39 42 42 43 47 48 50 67 77 80 80 80 80 82 83 84 85 TABLE OF CONTENTS (cont.) Page Soils with medium potassium supplying capabilities. I I O O O O O O O O O O O 86 Soils 27, 28 and 34 . . . . . . . . . . 86 Soils 16, 17 and 24 . . . . . . . . . . 88 Soils with high potassium supplying capabilities. O O O O O O O O O O O O O 89 Soils 14 and 26 . . . . . . . . . . . . 89 Potassium Release Characterization. . . 90 SUMMARY AND CONCLUSIONS . . . . . . . . . . . . . . 99 LITEMTURE CITED 0 O O O O O O O O O O O O O . C O O 105 APPENDIX. 0 O O O O O O O O O O O O O O O O O O O O 113 vii Table LIST OF TABLES Location, altitude, stoniness, and slope of the soils at the 30 sites sampled in Peru . Chemical and physical determinations made in Peru on soil samples from 30 locations. Chemical measurements made in Michigan on 16 soil profile samples from Peru and 2 Michigan surface samples. . . . . Total potassium content of the entire soil and of the sand, silt and clay fractions before crOpping and of the silt and clay fractions after cropping of 16 Peruvian and 2 Michigan soils. . . . . . . . . . . . pH values in a 1:2 soil solution ratio of E NaF (pH 7.9) after 0.5, l, 2, 4 and 8 minutes for 16 Peruvian and 2 Michigan soils . . . . . . . . . . . . . . . . Qualitative estimates of the clay minerals in 16 Peruvian and 2 Michigan soil clays . . Yields (g/100 g soil), concentration of K in plant material (%), and K uptake (me K/lOO 9 soil) of sorghum grown at La Molina and oats grown at Huancayo in a short-term cropping study of 30 Peruvian soils . . . . Correlation coefficients showing the relation- ship between various K extractants and K up- take by crOpping. . . . . . . . . . . . . . Yields of each of five successive two-week oat crOps grown on716 Peruvian and 2 Michigan soils. . . . . . . . . . . . . . . viii Page 33 34 36 40 44 45 49 50 52 LIST OF TABLES (cont.) Table 10. 11. 12. 13. 14. 15. 16. Page Concentration of potassium in each of five successive two-week oat crops grown on 16 Peruvian and 2 Michigan soils . . . . . . . 53 Potassium uptake from five successive two- week oat crOps grown on 16 Peruvian and 2 Michigan soils. . . . . . . . . . . . . . 55 Linear regression coefficients expressed as rate of K release and correSponding coeffi- cients of determination computed from accumulated uptake data for the 2nd, 3rd, 4th, and 5th oat harvests from 16 Peruvian and 2 Michigan soils. . . . . . . . . . . . 61 Exchangeable K levels before and after 5 successive croppings and uptake of soil K expressed as total K uptake, % of total K in soils and % of K uptake derived from exchangeable and nonexchangeable sources when 16 Peruvian and 2 Michigan soils were intensively cropped in a growth chamber . . 63 Correlation coefficients between accumulated net uptake of K after each of 5 successive crOps and exchangeable soil K before and after cropping, and uptake of exchangeable and nonexchangeable soil K by oats grown on 16 Peruvian and 2 Michigan soils . . . . 66 Potassium extracted (me/100 g) from 16 Peruvian and 2 Michigan soils by solutions containing sodium tetraphenylboron for different time periods. . . . . . . . . . . 68 Rate indices, rate of K release, and coeffi- cients of determination for linear regres- sion equations describing removal of K from 16 Peruvian and 2 Michigan soils by solutions containing sodium tetraphylboron in contact with soil samples for 10, 100, and 1000 hr periods . . . . . . . . . . . . 76 ix LIST OF TABLES (cont.) Table Page 17. The concentration of potassium (ppm K) in soil extracts after equilibration with solutions of varying potassium concentra- tions and the soil solution potassium concentration calculated from these data for 16 Peruvian and 2 Michigan soils . . . 79 18. Potassium status and K release characteris- tics éf l6 Peruvian and 2 Michigan soils grouped according to release capabilities and clay minerals present in the clay fraction . . . . . . . . . . . . . . . . . 81 LIST OF FIGURES Figure 1. Cumulative uptake of potassium by five suc- cessive two-week oat crOps grown on soils from the northern and southern mountain areas of Peru . . . . . . . . . . . . . 2. Cumulative uptake of potassium by five suc- cessive two-week oat crOps grown on soils from the central mountain region of Peru. 3. Cumulative uptake of potassium by five suc- cessive two-week oat creps grown on soils from the coastal (27 and 28) and jungle (24 and 29) regions of Peru . . . . . . . 4. Cumulative uptake of potassium by five suc- cessive two-week oat crops grown on two MiChigan SOilS O O O . I O I O O I O O O O O 5. Potassium extracted from soils 17, 27, 28 and 34 (curve A), 6, 8, 10 and 12 (curve B), and 3, 4, S, 21, 29, and 31 (curve C) with sodium tetraphenylboron for different time periods . . . . . . . . . . . . . . . . . 6. Potassium extracted from soils 14, 16, 24 and 26 with sodium tetraphenylboron for differ- ent time periods. . . . . . . . . . . . . 7. X-ray diffraction patterns of magnesium- saturated glycerol-solvated clay from soil 26 before and after intensive crOpping. . xi Page 56 57 58 59 73 74 94 Plate LIST OF PLATES X-ray diffraction patterns of the clay frac- tions from soils 3 and 4 (northern mountainS) O O O O I o O O O O O O O I O O O 0 X-ray diffraction patterns of the clay frac- tions from soils 5 and 6 (northern mountains). . . . . . . . . . . . . . . . . X-ray diffraction patterns of the clay frac- tions from soils 8 and 10 (southern mountains). . . . . . . . . . . . . . . . . X-ray diffraction patterns of the clay frac- tions from soils 12 and 14 (southern mountains). . . . . . . . . . . . . . . . . X-ray diffraction patterns of the clay frac- tions from soils l6 and 17 (central mountains). . . . . . . . . . . . . . . . X-ray diffraction patterns of the clay frac- tions from soils 21 (central mountains) and 24 (central jungle) . . . . . . . . . . . . X-ray diffraction patterns of the clay frac- tions from soils 26 (central mountains) and 27 (coaSt) o o o o o o o o o o o o o o o o 0 X—ray diffraction patterns of the clay frac- tions from soils 28 (coast) and 29 (central jungle) . . . . . . . . . . . . . . . . . . X-ray diffraction patterns of the clay frac- tions from soils 31 and 34 (Michigan) . . xii Page 113 115 117 119 121 123 125 127 129 INTRODUCTION The diversity in climate, agricultural crops and soils in Peru and the scarcity of arable land present a challenge for those workers whose goal is increasing food production. Only 8.6% of the land area of Peru is suitable for intensive agriculture, 53.5% of the land should be used for only permanent types of agriculture and the remaining 37.9% is unusable for agricultural purposes (Zamora, 1969). Land which can be used for agricultural purposes is found in ecological zones which include tropical, desert and sub- alpine as well as the transitional areas between these dif- ferent zones (Tosi, 1960). If food production in Peru is to keep pace with the population growth, increased production must be obtained on those soils now under cultivation, and new lands which are capable of producing food must be brought under cultivation. The use of chemical fertilizers is one of the necessary com- ponents in an increased food production program; however, fertilizers in a developing country are costly and thus must be used as efficiently as possible. Field experimental work in recent years has shown that the ability to predict a response to potassium (K) fertilizers has not been satisfactory and a recommendation as recent as 1970 was that studies be made which would lead to a better evaluation of the availability of nonexchange- able K to crops in Latin America (Fitts, 1970). A study was initiated in Peru and completed at Michigan State University (MSU) to evaluate the K-supplying ability of selected soils from the mountains (sierra), jungle and coast of Peru. Two Michigan soils were included in the study to make comparisons in amounts of K removed and rates of K release with the Peruvian soils. LITERATURE REVIEW The Soils of Peru The soils of Peru are probably as variable as in any country in the world due to the extreme variability in the five soil-forming factors: climate, vegetation, relief, par— ent material and time. The few soil maps which are available are general in their soil descriptions (Zamora, 1969). When describing these soils the country is usually divided into three geographical regions: the arid western coast, the central Andean mountains and the eastern tropical lowlands (Zavaleta and Arca, 1963; Drosdoff et al., 1960). Arid western coast The soils of this region are found in a narrow band from 50 to 100 kilometers wide along the Pacific Coast. They have received essentially no rainfall due to the pres- ence of the cool Humboldt current and the Andean mountains and thus are relatively unweathered. The four main soil groups recognized by Drosdoff g£_al. (1960) are alluvial, desert lithosol, desert regosol and red and gray desert. 3 Central Andean highlands (sierra) Jan Beek and Bramao (1968) describe this region as follows: The Andean mountain ranges are composed of a western range of marine Mesozoic and more recent volcanic rocks, with an elongated batholitic core, and an eastern chain of folded sedimentary, often Palaeozoic, rocks with scattered older granite and schist outcrops. The high variability of the soils is due principally to in- tensive erosion, sedimentation and accumulation of colluvial material. Zavaleta and Arca (1963) recognized alluvial, lithosol, brown, Andean meadow, chernozem and rendzina, humic gley, and organic soils as the major groups of this area. This was similar to the groupings described by Dros- doff §E_§l. (1960) who listed lithosols of the arid western slopes, the Andean valley association, puna association, titicaca association, lithosols and hydromorphs of the very high mountains and soils of the frigid region as the major groups. There is limited information available regarding the clay mineralogy of these soils. Miller and Coleman (1952) identified montmorillonite and illite in the dark- colored Paramo (probably puna association) soils. There was also some evidence for the presence of non-crystalline minerals. Amorphous materials could be expected to be pres- ent in the soils of the north and south of Peru if they are volcanic in origin (Zavaleta, 1969). The soils in the moist regions of the sierra also contained kaolinite as well as expanding lattice and illite clay minerals. Eastern tr0pical lowlands This area, dominated by the Amazon Basin, consists mainly of Tertiary and Pleistocene unconsolidated sediments that contain mainly kaolinitic clays and quartz sands (Jan Beek, and Bramao, 1968). The acid trOpical soils of the eastern forests as outlined by Estrada (1971) include spo- dosols, ultisols and oxisols. The clay mineralogy of these soils suggest one group which has a predominance of 1:1 type lattice clay minerals and another group that contains equal or greater amounts of 2:1 type clay minerals. Similar results were obtained by Sanchez and Buol (1971) who re- ported the presence of Paleudalts which were predominantly kaolinitic with smaller amounts of montmorillonite in some cases and poorly drained Paleudalfs, Tropaquepts and TrOpa- qualfs in which the clays were predominantly montmorillon- itic or of mixed mineralogy. Forms of Soil Potassium Total K levels in soils are among the highest of any of the soil-derived plant nutrients. Bowen (1966) re- ported an average total K value for soils of 1.40% with a range from 0.03 to 3.00%. Other values given for the United States range from less than 0.3% in the southeast and south- ern coastal plain to 2.5% in the midwest and west (Tisdale and Nelson, 1966). Failyer, et_31. (1908) reported a ten- dency for the K concentration in soil fractions to increase as the particle size decreased. However, a wide variation would be expected due to different degrees of weathering and kinds of parent material. Native soil K sources are the feldspars, micas and micaceous clay minerals. Jackson (1948) classifies the feldspars as framework silicates and the micas and micaceous clay minerals as layer silicates. Framework silicates Structure: The framework silicates are character- ized by a lattice formed by a sharing of four oxygens (Jackson, 1948). When aluminum and potassium are partially substituted for the silica in the oxygen tetrahedra, the formula for the potassium feldspars, orthoclase and microc1ine, is obtained (K(AlSi)408). In general, the feldSpars are found in the sand and silt fraction and release only small amounts of K to a plant-available form during any one crOpping season. Methods of potassium release: By evaluating the amounts of different feldspars present in a podzol profile from the Netherlands, Van der Marel (1949) estimated that about 8-15 kg KZO/Ha (300m) became available each year from feldspars. Biological agents appear to be one of the more effective mechanisms of feldspar weathering. The decomposition of feldspars with the subsequent form- ation of SiO2 and A1203 was demonstrated by Novorossova gt_al. (1947) by placing feldspars in contact with a mixed microflora in a medium containing glucose and organic N. Bassalick (1912), working with selective bacterial cul- tures, observed appreciable decomposition of feldspar where the micro-organisms formed a continuous film on the mineral particles. The stability of this weathering film and the particle size are factors which Arnold (1960) gives as controlling K release from feldspars. Layer silicates Structures: The major native sources of K for plant growth are the layer silicates of which the micas and illites predominate in importance. The two basic structures in the layer silicates are the silica tetrahedron and the aluminum octahedron. A unit cell is comprised of two tetrahedral layers with their apices pointed in towards an octahedral layer between them. If there are two aluminum ions in the octahedral layer the mineral is considered to be diocta- hedral, and if there are three divalent cations such as iron and magnesium instead of two aluminum ions in the oc- tahedral layer, the mineral.is said to be trioctahedral. General forumulas for muscovite (dioctahedral) and biotite (trioctahedral) micas are (OH)4K2(Si 'A12)Al40 and 6 20 (OH)4K2(816'A12)(Mg'Fe)602 pre8pectively (Grim, 1968). 0 V The term illite, used for the major K bearing min- eral in the clay fraction, has not always been well defined. Gandette gt_gl. (1964) analyzed a series of clay samples and concluded that illites have less K than the well- crystallized micas and are not necessarily mixed-layer materials. Less aluminum substitution for silica in the tetrahedral layers resulting in a lower net charge and partial replacement of K from the interlayer spaces by other cations are possible variations of the illites from the true micas given by Grim (1968). Thus illite is prob- ably secondary in origin and is found mainly in the clay fraction, while the more crystalline micas which may be found either primary or secondary in origin are usually found in the silt and sand fractions (Black, 1968). Potassium release mechanisms: For the feldspars, complete destruction of the structure is necessary for K to be released. However, the.1ayered structure of micas and illite permits K to diffuse out with a minimal disrup- tion of the mineral. Theories describing K weathering have been preposed by Jackson gt_§l. (1952), Reed and Scott (1962), Mortland (1958, 1961), Newman and Brown (1966) and Tucker (1964) and excellent reviews are given by Reitemeier (1951) and Rich (1968). In the weathering sequence of the layer silicates as proposed by Jackson gt_al. (1952), depotassication, hydroxyla- tion, dealumination and disilication are essential chemical reactions. The removal of K,.or depotassication, is the necessary first step and is based on the preferential weathering plane theory, i.e. micas weather along a given 10 potassium plane at a rate order of magnitude faster (once the weathering is initiated) than the rate of initiation of weathering of such a plane, and alternate mica inter- planes weather more easily than the remaining interplanes (particularly dioctrahedral micas). The driving force for the above process is the entrance of water and hydrated cations between the layers. Reed and Scott (1962) essentially expanded this theory when they prOposed a circular disc as a model for a mica particle and determined rates of release of K by pre- cipitating it in solution with sodium tetraphenylboron (NaTPB). In this system, in contrast to Jackson's theory, the initial rate of K release was higher than that after more K had been removed. They concluded that diffusion within the particle was the rate-limiting step in K release (Scott and Reed, 1962). When samples of biotite were leached with 0.1 N NaCl, Mortland (1958) found that the rate of K release was independent of the amount of K in the mineral until 75% of the K had been removed after which time there was a loga- rithmic relation between the rate of K release and time. It was later concluded that diffusion of K and Na ions through the film surrounding the particle was the 11 rate-limiting process (Mortland and Ellis, 1959). Although there appears to be a discrepancy in the above two theories, both appear to be valid conclusions in view of the experi- mental conditions which existed (TPB precipitating K in so- lution in one case and K being removed by leaching in the other case). Factors affecting weathering rates: The weathering of micas and illites is more easily understood when some of the factors which affect it are considered. Type of mica, particle size and pH of the environment are among the major factors which influence K weathering rates. There is a general consensus among researchers that the K in dioctahedral micas and illites is more resistant to release than K in trioctahedral micas and illites. Weed §£_gl. (1969) removed 63% of the K from biotite and 20% of the K from muscovite by placing the minerals in contact with fungi in a buffered nutrient solution. The fungi pro- vided a sink for K and kept the soil solution K at a low level. Scott and Smith (1966) used NaTPB as a precipitat— ing agent for.K to keep the soil solution K level low. They removed essentially all of the K from vermiculite, biotite and muscovite, but only 66% of the K from illite. 12 They had postulated earlier that interlayer material made the remaining K inaccessible to exchange (Smith and Scott, 1966). Differential K release due to type of mica has been explained by differences in the orientations of the O-H bonds (Quirk and Chute, 1968; Leonard and Weed, 1970), length of the K-0 bond (Leonard and Weed, 1970; 1970) and charge densities as indicated by calculated Si/Al ratios (Ross and Kodama, 1970). The release of both fixed K and native K to plants was investigated by Mortland §E_gl. (1957). They noted a comparatively rapid release of-K fixed by montmorillonite and.vermiculite and native biotite. The release of K from illite and muscovite was more difficult. Further investiga- tion by Ellis and Mortland (1959) showed that the release of K from native biotite was at a constant rate while approximately 50% of the K was removed whereas the rate of release from fixed forms of K decreased with time. Amount and rate of release of K from micas as af- fected by particle size appear to be related to the amount of stress applied to the particle. Initially the smaller biotite particles released more K when they were leached with 0.1 N NaCl (Mortland and-Lawton, 1961). However, after 50% of the total K had been removed, the larger 13 particles had lost as much K as the finer particles (1-250 u). Reichenbach and Rich-(1969), using 0.1 N BaCl2 at 120 C in a repeated batch technique, found that 92% of the K in larger muscovite particles (5-20 u) was released while in the smaller particles (0.08-27 u) only 57% was released. Similar results were obtained by Scott (1968) by equilibrating different particle sizes of muscovite with NaTPB; however, the smaller-particles (0.2-0.7 u) initially released more K than the larger particles (50- 60 u). He proposed that although both edge and layer weathering were taking place the predominant mechanism changed from edge to layer weathering with.a decrease in particle size. Thus if there.is an initial high loss of K due to layer weathering in the small particles, the re- maining K in the next interlayer will be more tightly bound as was proposed by Bassett.(l959). This agrees with results found by Doll et;al. (1965) when soil clays were cropped in the greenhouse.._ The effect of pH on the release of K from micas and illites is not always clear., When NaTPB was used as the potassium sink in a buffered.solution, Scott and Smith (1966) found that muscovite (50 u), biotite (50 u) and il- lite (2 u) all released more K at solution pH values of 14 4.6 than at 7.2 or 9.2. The differences were greater for biotite than for muscovite or illite. These same workers have also reported that solution pH had essentially no ef- fect on the amount of K removed from a grundite illite (Smith and Scott, 1966). Huang et_al. (1968) found that muscovite and biotite released-more K where they were equi- librated in dilute HCl solutions than in distilled water. A much wider range in pH values was used by Tucker (1964) with soil illites. The K release decreased as pH increased to pH 11 and then began to increase again. He concluded that below pH 11 K release is due to instability of the clay mineral structure towards hydrogen ions or to dis- placement of K by other cations.acting together with hydro- gen ions; above pH 11 the release is due to breakdown of the clay mineral. Based on studies on the action of dilute HCl and NaCl solutions on a lepidomelane-biotite flake, Wells and Norrish (1968) concluded that the hydrogen ion can replace interlayer K by the same general processes as do other cations and in addition it can at high concentra- tions dissolve the mica. 15 Release of Soil Potassium to Plants Equilibrium reactions which determine the K found in the soil solution, on exchange sites, and in the clay lattice will govern the ability of a soil to release K to plants. Thomas and Hipp (1968) stated that the equilibrium between the K in the soil solution and exchangeable K is of major importance for K availability to plants for time periods shorter than a several years while the nonexchange— able-exchangeable K equilibrium is an important considera- tion for sustained plant growth.for longer periods of time. Much of the current information regarding the weathering of K from pure minerals is based on establishing an efficient sink for K. The low K contents in the soil solution provide an excellent.concentration gradient caus- ing increased diffusion of K from the mineral. Scott and Smith (1966) using NaTPB increased the amount of K removed from biotite from 30% to 100% by decreasing the K in solu- tion from 10 to 7 ppm. They also found that soil solution K levels as low as 0.1 ppm restricted the removal of K from muscovite to 17% of the total K. The level of K in the actual soil solution is usually much higher than this. Asher and Ozanne (1967) reported that for 14 plant species, a K 16 concentration between 24 and 95 uM in solution was necessary for normal plant growth. Barber-(l962) reported soil solu- tion K values from the literature which ranged from 3 to 156 ppm. For soils with high concentrations of K in the soil solution, mass flow and not diffusion may be the predominant mechanism supplying K for plant growth. It would seem, therefore, that solution K measurements would be a valid criteria if they were considered with exchangeable K de- terminations. Hipp and Thomas (1967) found this to be true. They obtained a good correlation (r.= 0.890) between the percent K in sorghum leaves at full bloom and solution K + log (1 + exchangeable K). The amount of K which.is released from lattice forms to a plant-available form wi11.be determined by.the combina- tion of clay minerals present.in-the soil, their size dis- tribution and exchangeable K levels as.well as the concen- tration of K in the soil solution.. Very few soils have a_. homogeneous clay mineral content nor uniform particle size, and thus one specific K release rate.is very.un1ikely. This was recognized by Mortland (1961) and Smith gt_gl. (1971). Mortland found that K release and fixation can occur simultaneously in a system-containing a heterogeneous group of 2:1 minerals not in equilibrium with each other and 17 Smith and co-workers determined that curves for soil K re- lease merely represent the combined effects of the various particle sizes in the soil and not a distinctive mode of release. When cropping Mg-saturated soil clays Doll gt_gl. (1965) found that larger amounts of K were removed from the-. coarse clay fractions (0.2-2-0.u) than from the finer clay fraction (<0.08 u) with intermediate amounts from the medium fraction (0.08-0.2 u). In contrast, by relating the K con- tent of the various clay size fractions and the relative amount of that fraction to the amounts of nonexchangeable K released, Arnold and Close (1961) concluded that neither the amount of the coarse clay (0.3—2.0 u) nor its K content was necessarily related to the K.supplying power of the soils. By using a Mg-saturated clay, Doll gt_gl. were able to put much more stress on the-lattice K and, as has been discussed earlier, the coarse fractions will continue to release K under higher stress conditions while the finer fractions will not. In support Evans and Attoe (1948) found that up to three times.more nonexchangeable K was re- _ leased from soils when the exchangeable K level was dropped -solution before cropping. by leaching with a CaCl -MgCl 2 2 It seems that the amounts of K released depend upon the 18 exchangeable K level which acts as a "buffer" between the soil solution K and the K in the lattice of the clay minerals. Although ammonium acetate extractions in the past have, in most cases, adequately-provided a good prediction of the K release to plants, other methods of estimating K release for plant growth have been used. Potassium removal with successive salt extractions using 1 N NH OAc and l N 4 NaCl-0.1 N HCl and a single NaOAc-NaTPB extraction were compared with K uptake by the Stanford and Dement cropping technique by Scott and Welch (1961). The concluded that the NaOAc-NaTPB solution removes the easily released K very- quickly, that a crOp can remove as much or more K if given enough time, and that successive NaCl-HCl extractions will also remove as much or more K due to the acid nature of the solution. WOrking with 9 soils, Schulte and.Corey (1965) found that the best estimate of plant available K was the amount of K extracted from 2 g samples of nondried soil with 10 m1 of 0.3 N NaTPB in 15 minutes (r.= 0.991). The amounts of K extracted with NaTPB related to the K removed by 8 successive harvests-of.ryegrass over a 54 week period was nearly a 1:1 relationship. METHODS AND MATERIALS In August, 1970, a study was initiated in Peru to determine the relative availability of native soil K to plants and to study their K release patterns. Soils were screened in Peru by a cropping study and various laboratory measurements, and samples of selected soils were sent to Michigan State University (MSU) for more detailed studies. Sampling Procedure and Supporting Data Results from recent soil fertility experiments and from soil samples sent to the central soil testing laboratory in Peru were used to select sites which best represented the Peruvian sierra (mountains) with respect to area, agri- culture and K content. A small number of jungle and coastal soils were included to provide a comparison of the K status of these different climatic regions with those of the moun- tains. Profile samples in which horizons were separated on the basis of color, texture and structure were obtained from 30 locations (Table l, p. 33) distributed as follows: 19 20 6 from the northern sierra, 9 from the central sierra, 9 from the southern sierra, 4 from the jungle and 2 from the coast. In Peru, soil pH was.determined, using a 1:2.5 soil:solution ratio, with a glass electrode potentiome- ter. The Bouyoucus hydrometer method (Day, 1965) was used for texture measurements. A 1:5.soil:solution ratio and a loudnute shaking period with 6 N H2804 was used to extract soil potassium, and K in solution was determined with a flame photometer. Organic matter estimates were obtained using the Walkley-Black wet combustion method with H3PO4 additions to sharpen the titration endpoint (Allison, 1965). Total exchangeable acidity was determined by extracting with 1 N KCl (1:10 soilzsolution ratio) and titrating with 0.01 N NaOH (McLean, 1965). This information (Table 2, p.34) along with results from a short-term intensive crOpping study was used to select 16 of the 30 soils for more de- tailed work at MSU. Five kg surface samples and 500 g sub- surface horizon samples were steam sterilized by the Plant Quarantine Division of the USDA at Miami, Florida enroute to East Lansing. At MSU the profile samples were routinely analyzed (Table 3, p.36) for pH in;a 1:1:soil:water-suspension with a 21 glass electrode potentiometer, and exchangeable K, Ca and Mg were extracted for 1 minute with 1.0 N NH OAc (pH 7.0), using 4 a 1:8 soi1:solution ratio. Potassium was determined on a Coleman flame photometer and Ca.and Mg with a Perkin-Elmer 290 atomic absorption unit. Cation exchange capacities were determined by saturating the soil with Na using 1 N NaOAc pH 8.2, and then replacing the adsorbed Na with NH 4 as 1 N NH OAc pH 7.0 (Chapman, 1965). Sodium concentrations 4 were determined using a Coleman flame photometer. Total carbon measurements on the surface samples were made with a Leco carbon analyzer. Data from 2 Michigan soils, a Kalamazoo sandy loam and a Sims clay loam, which were in- cluded in the study, are also given in Table 3, p. 36. Soils with allophane or amorphous material present would be expected to have high exchangeable K levels, but a low capacity factor for K release. A test for allophane prOposed by Fieldes anngerrott(1966) is based on the fol- lowing.reaction: 6 NaF + A1(OH)X-————9 Na AlF + 3Na + on' 3 6 A production of hydroxyl ions causes axrapid.pH rise.above that of NaF (pH 7.9) which is good evidence for the presence of allophane. Twenty m1 of l N NaF (saturated solution) was 22 added to 10 9 soil samples, the mixture stirred, and pH de- termined after 30 sec, 1 min,.2 min, 4 min and 8 min time periods with a glass electrode potentiometer. Soil Fraction Analysis The sand, silt and clay fractions were separated for total K and X-ray analysis both during the initial phase of this study in Peru and after cropping at MSU. X-ray analyses of the clay fraction and total K contents of the clay and silt fractions.befbre and after intensive crOpping at MSU were used to evaluate changes in the clay mineralogy and the relative K contributions of the differ- ent soil fractions to plant growth. Separation of soil fractions Soil samples containing approximately 5 9 (Peru) and 2 g (Michigan) of clay were exposed to various pre- treatments to obtain a better separation of the soil frac- tions (Kunze, 1965). The soi1.samples were treated with 50 ml of 1 N NaOAc buffered to pH 5 with acetic acid and digested for 30 minutes at low heat on a hot plate to 23 dissolve carbonates and remove soluble salts. The suspen- sion was filtered under suction.and the filtrate discarded. The soil samples were then wetted with the NaOAc buffer and treated with successive increments of 30% H202 to destroy the organic matter. When there was no longer any apparent reaction with H202, the suspensions were centrifuged, the supernatant liquid discarded, and the sample warmed on a steam bath after addition of a solution of 40 ml of 0.2 M sodium citrate and 5 ml of 1 M NaHCO in order to remove 31 iron oxides. The suspension was heated to 75-80 C and 1 g of Na25204 added to reduce and-solubilize the iron. After a 15-minute digestion period with frequent stirrings, the sample was centrifuged and washed 2 times with alcohol, each time discarding the supernatant. The samples were then dispersed by shaking 6 hours with either 200 ml Na2CO3 pH 9.5 (Peru) or 50 m1 of a 5% Calgon solution (MSU). The sand fraction was removed by wet seiving with a 47 micron sieve before the clay and silt were separated in sedimen- tation cylinders, with settling time and sampling depth determined from a nomograph by Tanner and Jackson (1947). After separation the clays were flocculated with l NMgCl2 and dried in a forced air oven at 80 C in Peru, and left in solution in Michigan. 24 X-ray analysis X-ray analyses were made-on the clay fractions of the soils before and after the.intensive cropping at MSU to obtain estimates of the clay mineralogy of the soils and to determine any changes in the mineralogy by Cropping. The Peruvian clays were treated for periods of less than 1 minute with an ultrasonic dispersion apparatus to obtain clay suspensions. To prepare oriented clay specimens, small amounts of clay suspension were placed on a porous plaster of paris block in a holder to which suction was applied. After each of the following treatments in the holder the clay specimens were X-rayed with a Norelco dif- fraction unit using Cu radiation: Mg saturated, glycerol solvated; K saturated, air dried; K saturated, heated to 300 C for 2 hr; and K saturated, heated to 550 C for 2 hr. The soil clays which were X-rayed after intensive cropping received only the Mg saturation, glycerol solvation treat- ment. Powder samples of several of the soil clays were prepared to obtain 060 spacings which were used to deter- mine if the illite present was dioctahedral or trioctar hedral. 25 Total potassium Total K content of the whole soil and of the sand, silt and clay fractions was determined to estimate their relative contributions to the total. After intensive crop- ping, the silt and clay fractions.were analyzed for total K to evaluate the relative K supplying ability of the two fractions. Hydrogen peroxide (30%) was added to samples of .soi1.(0.4 g), clay (0.1-0.3 g), Silt (0.2 g) and sand (1.0 g) in Teflonl beakers to oxidize the organic matter. The samples were then treated with 0.5 ml concentrated H2804 and 5 ml HF, and placed on a sand bath at 200-220 C. After the samples were destroyed (in some cases more than one treatment was necessary) a 1:1 mixture of 6 N HCl and di- stilled water was added and heated to boiling to dissolve the residue. The samples were then taken to volume, and K in the solutions determined with a Coleman flame pho- tometer . lTeflon is a DuPont trademark for polyfluoro- olefin polymers. 26 Short-term Intensive CrOpping Intensive crOpping studies were established in Peru. "at Huancayo (sierra) and La Molina (coast) with all 30 soils .which were sampled to evaluate their short-term K supplying. ability, and the 16 soils selected for further study were- exhaustively cropped at MSU to determine K release patterns. Cropping studies in Peru For the studies in Peru, an amount of air—dry soil equivalent to 100 g of oven-dry soil was weighed in trip- licate and mixed with amounts ofCaCO3 sufficient to neu- tralize the exchangeable acidity.as measured by 14N KCl (1.5 me Ca applied for each l.0.me.TEA). The samples were equilibrated at approximate moisture equivalent for 24 hr. Approximate moisture equivalent was.determined by measur- ing the water content of the soil behind the wetting front” in a beaker with an aeration tube to which water had been applied 24 hr previously.. After the incubation.period, 100 g of acid-washed silica sand was added to each pot, the seeds planted (45 sorghum seeds at La Molina and 40 oat seeds at Huancayo), 27 covered with an additional 200 g.of silica sand, and brought to approximate moisture equivalent with distilled water. The pots were maintained as closely as possible to this weight with daily additions of distilled water. Regular additions of Hoagland's minus K nutrient solution (Hoagland and Arnon, 1950) were added until a total of 400 ml had been added to each pot. After cropping for 46.days at La Molina and 45 days at Huancayo, the plants were cut at sand level, dried at. 70 C, weighed and ground to pass a.20—mesh sieve. Two- tenths gram of dry plant material was weighed and shaken. for one hour with 50 ml 0.1 N.HC1, filtered and diluted. when necessary, and K in the filtrate determined with a flame photometer. Crgpping study in Michigan At MSU, the 16 selected-Peruvian.soils and 2 Mich— igan soils were intensively cropped under controlled en- . vironmental conditions to.better evaluate their K supply- ing ability and to determine-K.release patterns. Five separate croppings were made using .a cropping procedure adapted from that described by Stanford and DeMent (1957). 28 Amounts of CaCO3 to neutralize the milliequivalents of hydrogen present as determined by the SMP buffer method (Shoemaker eE_al., 1961) were equilibrated with triplicate. 100 9 samples of ovendry-equivalent soil. The soils were incubated in 16 oz wax cartons-for three weeks during which. time they were wetted to one-third.bar moisture levels.and allowed to air-dry four times. .After incubation, the soils were placed in the growth chamber.for crOpping. The growth chamber was set to operate on a day period of 16 hr at 24 C and a night period of 8 hr.at-18.C. The light intensity was. 2,000 ft-candles at one-half canopy level. The relative . humidity was maintained as nearly as possible above 50 per- cent. Approximately 500 g of silica sand were added to 16 oz wax cartons into which another bottomless 16 oz carton had been inserted.. Thirty oats-seeds were evenly spread . on the surface, all cartons brought to.750 g with addi- tional amounts of silica sand, 80 ml distilled water added-” for germination and grown for 10 days in the growth chamber. During this time two 50 ml-increments of nutrient solution which did not contain K were added.. The inside carton con-.. taining the cats seedlings with a.mat of roots at the bote.. tom was then removed and placed on tOp of the soils, also 29 in 16 oz waxed cartons, for.a l4eday.cropping period. .After the transfer an amount of.water was added equivalent to the one-third bar value plus-75 ml.for the sand, and all cartons were adjusted to a.uniform weight by adding more silica sand to facilitate.daily.moisture control. During the 24-day growing period, each carton received 200 ml of Hoagland's nutrient solution without K for the first and second cropping, and 250 m1.of the same solution but with extra nitrogen.for the third, fourth, and fifth cr0pping. Iron was added separately as FeEDDHA in 4 increments of 50 ml of a 5 ppm Fe solution to each crop. After the mat of cats roots.had been in contact with the soil for 14 days,.the.oats were harvested by cut- ting.above the seed. No attempt.was.made to recover.the. roots from the soil. The.1ayer.of.soil was separated from the sand, air dried and passed.through.a 4emesh screen. Each soil carton was weighed.so.compensations for soil loss and sand and root gains could be made and the pre-. viously described steps repeated for the second, third, fourth.and fifth oats crops. ..... The plant material.was dried in a forced air oven at 65 C for at least 48 hr, weighed and ground to pass a 30 20-mesh sieve. Four-tenths g of plant material was shaken with.100 ml 0.1 N HCl for 1 hr, filtered and the concen- tration of K in the extract determined with a Coleman flame photometer. On completion of the fifth harvest the soils were allowed to air dry and passed.through a 2 mm sieve. Equal volumes of each replication were combined and mixed for total K and mineralogical analyses.. Exchangeable K was determined on samples from each replicate. Sodium Tetraphenylboron Extractions Relative K removal rates.and amounts of K which can eventually be removed by exchange.reactions were determined using NaTPB as a precipitating.agent.for K. The method was essentially that of Smith and Scott (1966) in which 10 m1. of 1.7 N NaCl was added to duplicate l g (O.D. equivalent) soil samples, and 1.02 g of solid NaTPB.. The mixture was swirled and placed in a constant.temperature cabinet (25.C)- for 15 min, 1 hr, 10 hr, 100.hr,.1,000.hr and 2,000 hr time periods. When the time corresponding to the prescribed treatment had elapsed, 500 me of NH Cl and 6 mmole of HgCl 4 2 31 were immediately added. The sample was boiled for at least 20 minutes, cooled and filtered by suction. The solution. was diluted to 1 liter and Klin solution determined with a Coleman flame photometer. Soil Solution Potassium Determination Estimates of the concentration of K in the soil solution were made using a method described by Beckett (1964) for determining potassium.activity ratios. Dupli- cate 5 g soil samples were equilibrated with 50 ml of so- lutions which were 0, 0.00025,.0.0005,.0.001 and 0.002. molar with.respect to potassium.(KCl) and 0.002 molar with respect to calcium (CaClz)...The samples were shaken for 1 hr, allowed to stand for.12.hr,-shaken for 30 min, filtered, and K in solution determined by flame.photometry.. The concentration of-K in the initial solution which core responded to no change after equilibration (AK =.0) was used as an estimate of the concentration of K in the soil solution. RESULTS Soil fertility work in Peru has shown that the ability to predict responses to K fertilization is poor (McCollum and Valverde, 1968). To better characterize the potassium status of Peruvian soils, particularly those from the mountains, 30 soils were sampled and eval- uated for K by crOpping studies and laboratory measure- ments in Peru; 16 of these soils were selected and taken to MSU for more intensive studies. Of the 30 soils originally sampled, 24 were from the mountains, with all except one at altitudes above 10,000 ft; slopes varied from O to 90% (Table 1). The two coastal soils were at altitudes of 300 and 400 ft with slopes of less than 1%, and the four-jungle soils were at altitudes of 500 to 2,600 ft. with slopes vary- ing from 0 to 100 percent. The pH values of the soils sampled varied from 4.5 to 8.5 in the mountains, from 4.1 to 6.2 in the jungle, and from 8.2 to 8.4 on the coast; textures varied from clay to loamy sand (Table 2). Organic matter levels in these soils varied from 1.3 to 8.5% with no consistent 32 .uiit‘q- . .5.» ~.n 5 in. t... 33 .auamum>flco mucum cmmflzofiz um htsum umnuusm How owuomawm mafloms mno mcoc oom Coastsm .m Honouou mmaawozm cm <9H>H om mlv mcoc oom.H oosnmsm .m,Honouoo mmHHmosm ha ¢BH>H «mm Hno filo oov mafia msflq was 24mm «mm Huo mac: 00m mafia mumamo aneumdsH oueamH new .omm «am omnom muo ooo.HH cease couodmucoo mmaoo mmsoo «mm Mlo HIo oom.~ cacah omwumm omHDMm mmcmuw mg mm coauom muo ooo.~ cease manna cause: an moanmo new .omm .vm mao omImH oow.oa cacsh :ofiomwocou coaomoocoo OHHDH on w>osz mm mlo mcoc oom.OH sense smash cooono mmemmmnmoo mm mum mcon oon.oa cacao o>mocmnm oncomnm mumaawuomq ma saw muo Hlo oom.oa GMGSH owwocmam mamoflm cm>momno ON muo sum oom.on cause enema sameness «on 42H ma mlo ococ oom.oa Cacao smash cumucmz Hm muzo ma mmnom ococ oom.HH cflcsn coaommocou .9 mmcflouom mmaaa mcmw «ha mlo mco: oon.oa sense o>mocmsm oases Hm vmx and modem «0H calm HIO oov.ma Odom ocsm moanommmo OHflsmooooa ma mum maloa oom.mH ocnm mcmocmsm commas ooosmcmw «ca mum muo omv.mH ocsm onflsnosno ossoas» oomsnms ma mlo mno oov.ma ocsm ocsm ocnm snoH «NH mno ococ ooo.HH oondu mflcocmu muses moumsoumso Ha omuma ococ oom.aa oonsu mucd cocoooumsm taaoou momno «0H oanm mcoc oov.HH oousu muss tuna ammsnouumfim m mno mum oom.oH oouso must «use m>msz mnsmnmocoo «m calm mcoc oom.HH condo sued muHMSN moHonmuflz e mlo mno oon.m monmsmmmu nonmemnmu moumsmflmu camou0H> MA 40 ovIOm mlo ooa.HH moumsmmmu monmsmnmu moumsmnmu connomsnnmu «m omnma mcoc 00>.HH moumEmmmo moumsmwmo mousfimnmo coouom «e emuom ococ ooa.HH omuuman MA connomamsm oonnomfimsm cmomsnmu «m omuom mac: oom.HH nmunmnnq an oousuo mnsmmmflamma anemone an m ovIOM 0H ooo.ma omuumneq ma oounpo. cmoasn ouamm occasm H va va Aumv .ucmfiuummoa wond>oum uoauumaa pamam .0: macaw monoum mosuflua< . . . . . . . . Hwom .summ cw oonEMm mouflm om may at wHHom on» no mmoam was smoocwcoum .oosuauam .coaumooqul.a manta 34 Table 2.--Chemica1 and physical determinations made in Peru on soil samples from 30.locations. Texture Total ex- changeable acidity (me/lOO g) Potassrum (me/100 g) 0 M. (percent) (percent) 8011 number Sand Silt Clay 2.40 3 0.92 0.37 0.57 1 0 25 39 15 22 28 10 24 22 53 5.1 .80 33 75 63 4.5 5. 1.65 5 2.3 0 0 9 8.1 .55 .03 13 5. 0.10 .43 0.62 23 55 .00 .00 0.00 0.10 0.20 0.00 0.05 0.15 0.05 0.10 0.20 1.25 0.10 31 37 25 26 30 20 28 38 34 43 30 1.65 0.83 0.57 1 33 3 55 43 6.9 29 31 57 2 10 6.0 31 11 12 .00 32 58 0.37 0. 3 14 22 79 63 2 15 5.9 13 1.7 4 15 21 33 37 69 26 44 28 36 20 32 34 30 29 20 20 22 24 34 26 26 33 59 37 14 15 . 16 63 0.22 0.50 0.40 0.50 0.75 0.45 0.33 0.40 0. 0. 5.6 5.8 39 . 27 6.4 17 3 11 18 19 20 21 22 23 24 25 26 27 .00 0.00 0.10 1.70 0.05 29 35 29 28 25 17 13 39 31 41 2 6.0 43 2 55 .10 0.20 0.05 0.00 0.00 5.60 18 1 63 2 6.0 0.47 0.75 1.38 0.77 3 65 21 55 6.8 2.0 21 29 27 40 45 8.2 45 28 29 30 20 0.38 4.1 47 3 4.1 10 27 35 differences apparent between the different areas. Amounts of K extracted with 6-N H SO varied from 0.18 to 1.65 2 4 me/100 g; 12 of the 30 soils tested below 0.45 me/lOO g. In 1969, (Pitts, 1969) K fertilizers were recommended for Peruvian soils testing below 0.41 me/100 g if the pH was 6.4 or lower and below 0.51 me/lOO 9. if the pH was 6.5 or higher. Potassium chloride extractions for total ex- changeable acidity ranged from 0 to 8.10 me/lOO g with those soils having pH values at 5.0 or below containing the higher amounts. Soil Chemical and Mineralogical Properties Chemical measurements on soil horizons To evaluate the uniformity of parent material and obtain an estimate of weathering, the pH, exchangeable K, Ca and Mg and cation exchange capacities were determined on profile samples from the 16 Peruvian soils (Table 3) selected for intensive studies in Michigan. Soil pH varied less than 0.5 units throughout the profile for the mountain soils 3, 4, 6, 10, 14 and 26, the coastal 36 Table 3.--Chemica1 measurements made in Michigan on 16 soil profile samples from Peru and 2 Michigan surface samples 4‘ AA Horizon Exchangeable cations Cation Base Organic Soil depth pH (me/100 g) exchange sat. matter no. ‘ capacity (cm) K Ca Mg '(me/100 g) (%) (%) ..... ....................... Peruvian soils............................. 3 0-20 4.4 0.35 0.88 0.16 9.46 14.7 1.77 20-40 4.2 0.24 0.43 0.13 10.65 7.5 40-160 4.3 0.21 0.43 0.11 9.56 7.8 160-170 4.4 0.13 0.65 0.16 15.22 6.2 4 0-35 4.7 0.94 6.07 1.29 70.11 11.8 13.27 35-50 4.7 1.05 5.16 1.89 66.96 12.1 50+ 5.1 0.93 2.45 0.77 37.83 11.0 5 0710 6.4 0.34 21.70 5.45 43.48 63.2 1.72 10-50 6.1 0.16 19.33 4.52 46.96 51.1 50-100 6.2 0.18 21.71 4.99 51.74 51.9 i100-120 6.6 0.18 23.86 4.92 43.48 66.6 120-140 6.8 0.22 31.88 5.86 42.60 89.1 6 0-10 7.9 0.44 25.06 1.41. 42.71 63.0 4.16 10-20 7.8 0.38 27.73 1.11 39.56 73.9 20-40 7.9 0.26 22.66 0.94 26.96 88.5 40-120 8.0 0.32 21.71 0.88 21.96 104.3 8 0-20 5.9 0.33 12.29 1.41 30.76 45.6 3.47 20-40 5.7 0.27 12.52 1.41 32.39 43.8 40-75 6.7 0.40 12.52 1.89 28.48 52.0 75-125 6.7 0.26 13.45 2.11 30.87 51.2 10 0-20 6.0 0.45 6.30 1.41 21.41 38.1 2.43 20-80 6.0 ' 0.53 9.51 3.88 36.63 38.0 80-140 6.4 0.56 12.52 4.84 37.39 47.9 12 0-20 6.4 0.13 6.08 0.77 9.78 71.4 1.11 20-40 6.4 0.16 6.76 1.09 15.22 52.6 40-90 6.9 0.09 2.45 0.41 6.52 45.2 90-170 7.3 0.09 2.22 0.32 5.43 48.4 170-220 5.7 0.38 12.75 1.92 37.61 40.0 14 0-30 7.0 0.13 7.22 1.18 15.00 56.9 1.55 30-45 7.5 0.11 7.68 0.66 15.87 53.2 45-60 7.4 0.10 6.08 0.60 14.35 47.2 60-110 7.4 0.14 7.67 1.14 16.52 54.2 110-160 7.1 0.14 8.36 2.20 18.70 57.2 Table 3 (continued) 37 ‘ - 3 Horizon Exchangeable cations Cation Base Organic Soil depth (me/100 9) exchange sat. matter no. pH capacity (cm) K Ca Mg (me/100 g) (%) (%) ............................ Peruvian soils.............................. 16 0-30 5.6 0.15 9.51 1.62 23.59 47.8 2.65 30-55 6.7 0.20 11.82 2.33 24.56 58.4 55-110 6.6 0.17 10.66 1.95 24.13 53.0 110-145 6.9 0.21 11.82 2.14 30.43 44.8 17 0715 5.0 0.42 2.90 0.72 23.91 16.9 4.29 15-35 5.1 0.35 2.23 0.74 24.78 13.4 35-70 5.3 0.35 3.13 1.20 22.17 21.1 70-120 5.4 0.30 4.48 1.71 18.70 34.7 120-180 5.6 0.23 4.71 1.98 21.96 31.5 21 0-15 6.1 0.40 9.05 2.23 25.54 45.7 1.90 15r50 6.5 0.68 17.43 8.12 43.91 59.7 50-110 7.7 0.58 13.68 4.52 29.78 63.1 110-155 7.7 0.87 21.71 10.16 31.74 103.1 155—185 7.9 0.37 19.37 '3.01 18.15 125.1 185-190 7.7 0.54 22.19 3.14 26.96 95.9 24 0-30 6.3 0.10 9.51 0.86 23.91 43.8 3.69 30-140 7.1 0.09 9.28 1.44 17.93 60.3 140+ 7.1 0.09 10.43 2.68 23.48 56.3 26 0-30 6.5 0.21 6.08 0.66 15.65 44.4 2.95 30-65 6.3 0.15 3.80 0.60 12.61 36.1 65-125 6.5 0.17 3.80 0.57 11.74 38.7 125-185 6.5 0.14 4.71 0.80 13.48 41.9 27 0-25 8.0 0.71 13.45 1.14 18.70 81.8 1.64 25-45 8.1 0.37 12.52 0.63 12.93 104.6 45-75 8.0 0.36 19.33 1.95 29.13 74.3 75-90 8.0 0.37 13.22 1.98 19.56 79.6 90-180 8.1 0.16 3.80 0.66 6.63 69.7 28 0-20 7.9 0.42 13.68 0.77 26.30 56.5 1.72 20-40 7.9 0.32 10.90 0.54 20.00 58.8 40-95 7.5 0.25 11.13 0.83 20.65 59.1 95-125 7.4 0.23 11.59 0.88 22.83 55.6 125-185 7.4 0.17 9.05 0.68 17.17 57.6 29 0-5 4.0 0.23 0.88 0.35 20.65 7.1 1.69 5-18 4.2 0.16 0.65 0.16 18.59 5.2 18-29 4.2 0.14 0.43 0.11 19.24 3.5 29-45 4.5 0.14 1.33 0.13 19.35 8.3 45-90 4.3 0.17 0.43 0.11 18.59 3.8 38 Table 3 (continued) 1 — I Horizon Exchangeable cations Cation Base Organic Soil depth H (meZlOO g). exchange sat. matter no. p capacity (cm) K Ca Mg (me/100 g) (%) (%) .............................. Michigan Soils............................ 31* 0-20 7.1 0.42 2.21 1.47 12.50 32.8 1.18 34 0-20 6.0 0.48 11.90 3.17 40.11 38.8 6.37 *31, Kalamazoo sandy loam; 34, Sims clay loam. soils 27 and 28, and the jungle soil 29. Soil pH increased with depth in soils 5, 8, 16, 17 and 21 (mountain) and 24 (jungle), and first increased and then decreased for soil 12 (mountain). Exchangeable K levels decreased with depth in soils 3, 5, 6, 17 and 26 from the mountains, 27 and 28 from the coast and 29 from the jungle, increased with depth for the mountain soil 10, remained relatively constant in soils 14 and 16 (mountain) and 24 (jungle) and were var- iable throughout the profile in soils 4, 8, 12 and 21 (mountain). Intensive weathering as indicated by a low percent base saturation could be inferred for northern mountain soils 3 and 4 and jungle soil 29; however, the cation ex- change capacity was much higher in soil 4 than in soils 3 and 29 (Table 3). An increase in pH and percent base 39 saturation with depth in the soil profile would indicate moderate weathering, which probably is the case for moun- tain soils 5, 8, 10, 16, 17 and 21 and the jungle soil 24. The remaining soils: 6, 12, 14 and 26 from the mountains and 27 and 28 from the coast appear to be relatively un- weathered, however, only 12 and 26 were not calcareous. Total potassium in surface samples Soil samples and samples of the sand, silt and clay fractions were treated with H2804 and HF to destroy silicates to determine total K in the 16 Peruvian and 2 Michigan soils. The clay and silt fractions were also analyzed for total K after intensive cropping to measure any changes in the K content of these fractions during cropping. The amounts of total K in the Peruvian soils varied from 0.15 to 3.97 percent K (Table 4). In general, the K contents of soils from the northern mountains were lower than those of other soils, with a K range from 0.15 to 0.86 percent. These soils are thought to be of vol- canic origin which may explain the low values (Zavaleta, 1969). Of the 18 soils studied, 5 soils (3, 4, 5 and 6 40 Table 4.-—Tota1 potassium content of the entire soil and of the sand, silt and clay fractions before crOpping and of the silt and clay fractions after.cropping of 16 Peruvian and 2 Michigan soils. Potassium concentration in percent Soil no. After cropping Soil Sand Silt Clay Silt Clay 3 0.15 0.02 0.10 0.95 0.11 0.65 4 0.47 0.80 0.68 0.34 0.68 0.26 5 0.86 1.21 1.19 0.34 1.02 0.37 6 0.70 0.65 0.50 1.59 0.36 1.36 8 1.46 1.52 1.61 1.33 1.28 1.33 10 2.22 1.66 2.44 2.97 2.32 2.23 12 1.45 1.20 2.04 2.18 1.95 2.15 14 3.02 3.29 2.39 4.16 2.19 4.05 16 2.04 1.67 1.84 3.36 1.45 3.15 17 1.71 1.32 1.35 2.46 1.13 2.56 21 1.17 0.69 1.29 1.58 1.19 1.32 24 2.29 2.38 2.19 2.59 1.90 2.41 26 3.97 4.03 2.50 3.25 3.61 3.55 27 2.01 1.82 1.91 2.07 1.71 1.85 28 1.91 1.64 1.73 2.82 1.65 2.49 29 0.27 0.03 0.07 1.00 0.07 0.82 31 1.15 1.08 1.86 1.19 1.75 1.25 34 1.92 0.51 2.01 2.41 1.98 2.93 from the mountains and 29 from the jungle) had total K contents lower than 1%, 7 soils (8, 12, 17 and 21 from the mountains, 28 from the coast and 31 and 34 from Mich- igan) contained between 1 and 2% K, and 6 soils (10, 14, 16 and 26 from the mountains, 24 from the jungle, and 27 from the coast) had amounts of K greater than 2%. Ten 41 of the 18 soils studied had total K levels higher than those reported by Olivera et_a£. (1971) who found total K contents ranging from 1,780 to 14,200 ppm in soils from southeastern Brazil which were classified as Oxisols, U1- tisols, Alfisols, Mollisols and Inceptisols. Total K contents of the sand, silt and clay frac- tion were variable with many soils containing different relative amounts in the various fractions. The amounts of K in the sand were the most variable with a range from 0.02 to 4.03 percent. The ranges of total K in the silt and clay were 0.07 to 2.50% and 0.34 to 4.16%, reSpec- tively. The K content increased as the particle size de- creased for mountain soils 3, 10, 12, 16, 17 and 21, coastal soils 27 and 28, jungle soil 29, and Michigan soil 34, indicating some weathering has taken place on these soils. For mountain soils 4 and 5 the K content decreased as the particle size decreased which infers that the K present in these soils is mainly in primary minerals in the larger particles and not in the clay min- erals. The K in the remaining soils was well distributed among the 3 fractions except for soil 6 which had substan- tially more K in the clays than in the sand and silt fraction. 42 The K content of the clay and silt fractions was in most cases lower after cropping than before cropping. Soils 3, 10 and 21 (mountain), 28 (coast), and 29 (jungle) appeared to have lost more K from the clay than the silt fraction, while mountain soils 5, 8, 14, 16 and 17 and jungle soil 24 had more K loss from the silt than the clay. These are trends only and statistically there was no overall significant difference at the 0.05 level of significance between K content in the silt or clay before and after cropping. Clay mineralogy of surface samples Sodium fluoride test for allophane: The presence of allophane or amorphous material was estimated by react- ing 10 9 soil samples with 20 ml 1 N NaF (pH 7.9) and measuring the increase in pH after 30 sec, 1 min, 2 min, 4 min and 8 min time periods (Fieldes and Perrott, 1966). A rapid pH rise was considered to be a positive test for the presence of allophane. The above authors state that the only likely exceptions of the above test are soils initially high in alkali or containing free CaCO3 in 43 which a reaction occurs with NaF to precipitate CaF2 and yield Na2CO3. The pH values of the NaF-soil mixtures at differ- ent times after contact are given in Table 5. Soils 4, 6, 27, 28 and 31 all had pH values greater or equal to 9.3 at 2 min; however, the presence of free carbonates was noted in soils 6, 27, and 28 by effervescence with 3 N HCl. Positive evidence for the presence of allophane exists only for soils 4 (mountain) and 31 (Michigan). The pres- ence of allophane is suspected in the mountain soils 5, 10 and 21 since the pH did increase to 9.0 at the 2 min time period. X-ray analysis: The X-ray diffraction patterns of the oriented clay samples after various treatments are given in the Appendix. .The interpretation of these patterns and the results of the NaF quick test for amorphous ma— terial are presented in Table 6. There is very little illite present in the soils from the northern sierra (3, 4, 5 and 6). Soil 3, a red soil which appeared to be highly weathered, contains mostly metahalloysite and amorphous material. The dominating clay mineral in soil 5 is montmorillonite. Soils 6fand 12 did not show a 44 Table 5.--pH values in a 1:2 soil:solution ration 0f N‘ NaF (pH 7.9) after 0.5, 1, 2, 4 and 8 minutes for 16 Peruvian and 2 Michigan soils. Soil pH n°° 0.5 1 2 4 8 min - mm mm mm min ~3 8.4 8.6 8.7 8.8 8.9 4 9.3 9.4 9.5 9.6 9.6 s 8.9 9.0 9.0 9.1 9.1 6 10.0 10.1 10.1 10.2 10.2 8 8.2 8.2 8.2 8.4 8.4 10 8.8 8.8 9.0 9.1 9.2 12 8.5 8.5 8.6 8.7 8.8 14 8.7 8.8 8.8 8.9 9.0 16 8.0 8&1 8.1 8.2 8.3 17 8.4 8.6 8.7 8.9 9.0 21 8.9 9.0 9.1 9.2 9.3 24 7.9 7.9 8.0 8.1 8.3 26 8.4 8.5 8.6 8.7 8.7 27 10.0 10.1 10.1 10.2 10.2 28 9.9 10.0 10.0 10.1 10.1 29 8.4 8.5 8.6 8.7 8.8 31 9.2 9.3 9.4 9.5 9.6 34 7.9 7.9 8.0 8.1 8.2 45 Table 6.--Qua1itative estimates of the clay minerals in 16 Peruvian and 2 Michigan soil clays. W Soil Clay minerals no. .0...I...0.0......OOOOOOOOOOOPeruV-ian SOilSOOOOODOOIOOOOOrOOIOOOOOOOOOO 3 Metahalloysite; some 2:1 intergrade material; quartz 4 Metahalloysite; allophane; small amounts of illite and vermiculite 5 Dioctahedral montmorillonite; vermiculite; illite; small amounts of quartz 6 Randomly interstratified montmorillonite, illite, and vermiculite; small amounts of kaolinite and quartz 8 Small amounts of montmorillonite; 2:1 intergrades; quartz 10 Illite; metahalloysite; quartz; feldspar 12 Randomly interstratified illite, chlorite, and montmorillonite; metahalloysite; small amounts of quartz and feldspars 14 Illite; randomdy interstratified chlorite and vermiculite; meta- halloysite; quartz 16 Illite; metahalloysite; vermiculite; quartz; feldspar l7 Metahalloysite; illite; quartz; feldspar 21 Small amounts of illite and metahalloysite; amorphous material; quartz; feldspar 24 Illite; metahalloysite; vermiculite; montmorillonite; quartz; feldspar 26 Dioctahedral and trioctahedral illite; metahalloysite; vermiculite; quartz; feldspar 27, Illite; kaolinite; montmorillonite; chlorite; randomly interstrati- "“fied chlorite and montmorillonite; quartz; feldspar 28 Illite; kaolinite; chlorite; randomly interstratified chlorite, vermiculite, and montmorillonite; quartz; feldspar 29 Metahalloysite; randomly interstratified montmorillonite, illite and chlorite; quartz OOOOOOOOQOOOOOOOOOOOOOOOOOOOOOOMiChigan 80118....OOOOOOOQOOOODOOOOOOOO. 31 Kaolinite; vermiculite and chlorite randomly interstratified with smaller amounts of illite; quartz; amorphous material 34 Illite; kaolinite; randomly interstratified vermiculite, chlorite and montmorillonite; quartz 46 dominating clay mineral species, but rather some 1:1 ma- terial and a random interstratification of montmorillon- ite, illite and either vermiculite (6), or chlorite (12). Metahalloysite, illite, quartz and some feldspars were present in mountain soils 10, 17 and 21; soils 10 and 21 contain some amorphous material, and soil 8 probably also contains some amorphous material even though it did not give a positive NaF test. Soils 14, 16 and 26 from the mountains and soil 24 from the high jungle all contained metahalloysite, illite, quartz and some vermiculite; however, 14 appeared to have smaller amounts of chlorite randomly interstratified with the vermiculite. Soil 24 also contained montmorillonite. The illites in soil 26 appear to be predominantly diocta- hedral; however, 060 spacing indicates the presence of some trioctahedral species. The presence of feldspars was indicated by peaks in‘the 3.18 A range for soils 16, 24, and 26. The mineralogy of the two coastal soils, 27 and 28 is heterogeneous but similar with illite, kaolinite, chlorite, quartz, and feldspars present in both soils. Soil 27 also has a distinct montmorillonite component as well as montmorillonite randomly interstratified with chlorite while soil 28 has montmorillonite randomly . 47 interstratified with chlorite and vermiculite. Metahalloy- site is the dominant clay mineral with small amounts of randomly interstratified montmorillonite, illite and chlorite for the jungle soil number 29. Soil 31, a Michigan soil, contains kaolinite, amorphous material, quartz, and a random interstratification of chlorite, with smaller amounts of illite. Illite, kaolinite, quartz and randomly interstratified vermiculite, chorite and montmorillonite are the components of the clay frac- tion of soil 34, the other Michigan soil included in this study. Short-term Intensive Cropping Adaptations of the cropping method of Stanford and DeMent (1957) were used to evaluate the K supplying ability of Peruvian soils. Thirty soils were sampled and crOpped in Peru in pot studies at 2 locations differ- ing in climate and altitude. Results of these studies together with laboratory measurements were used to select 16 of the 30 soils for more rigorous crOpping and labora- tory studies at MSU. 48 Cropping studies in Peru Dry matter yields, K concentration, and K uptake data from small pot intensive cropping experiments con- ducted on the coast at La Molina and in the mountains at ’Huancayo are given in Table 7. The dry matter yields were consistently lower, K concentrations higher, and K uptake generally higher in the mountains than on the coast, due partly to unusually cool and cloudy weather on the coast. At both locations the pots were outside; however, the pots at Huancayo were below a corrugated, clear plastic roof. The relationship between K uptake at the two locations and K extracted by neutral, normal ammonium acetate and 6 E H2804 are given as correlation coefficients in Table 8. With the more desirable growing conditions at Huancayo it made little difference if ammon- ium acetate (r = 0.625) or sulfuric acid (r = 0.627) was used to predict K uptake. Potassium extracted with ammon— ium acetate (r = 0.532) related better than that extracted with sulfuric acid (r = 0.394) to K uptake in the experi- ment on the coast. Of the above correlations none can be considered particularly good. 49 Table 7.--Yie1ds (g/100 9 soil), concentration of K in plant material (%), and Kuptake (me K/100 g soil) of sorghum grown at LaEMolina and oats grown at Huan- cayo in a short-term cropping study of 30 Peruvian soils. La Molina Huancayo Soil no. Yield K 'K Yield K K concentration uptake concentration uptake (9) (%) (me) (9) (%) (me) 1 2.21 1.24 0.69 ‘“0.99 2.91 0.73 2 1.83 0.52 0.25 0.91 1.59 0.37 3* 0.94 0.70 0.17 0.90 1.67 0.38 4* 2.24 0.86 0.49 1.42 2.68 0.97 5* 1.58 0.62 0.25 1.12 1.66 0.47 6* 1.76 0.49 0.22 1.18 1.59 0.48 7 0.99 0.46 0.11 1.26 0.92 0.29 8* 1.55 0.60 0.24 1.09 1.48 0.41 9 2.23 0.75 0.43 0.99 2.06 0.52 10* 2.82 0.58 0.42 1.10 1.77 0.50 11 1.92 0.62 0.30 1.14 2.46 0.72 12* 0.88 0.67 0.15 0.81 1.09 0.23 13 1.80 0.54 0.25 0.97 1.84 0.45 14* 1.71 0.66 0.29 0.98 2.79 0.71 15 1.52 0.76 0.29 1.17 2.42 0.72 16* 1.91 0.72 0.35 0.70 1.61 0.29 17* 1.92 1.05 0.52 1.04 1.50 0.40 18 2.18 1.05 0.59 0.62 1.89 0.30 19 1.09 0.65 0.18 1.08 1.40 0.39 20 2.86 0.62 0.46 1.21 2.24 0.69 21* 2.39 0.30 0.18 1.08 1.43 0.39 22 2.06 0.34 0.18 1.07 1.27 0.35 23 1.36 0.36 0.12 0.80 1.33 0.26 24* 1.85 0.31 0.15 0.97 1.20 0.30 25 2.53 0.47 0.30 1.16 1.87 0.55 26* 2.54 1.04 0.68 1.16 4.60 1.22 27* 1.69 1.05 0.45 1.34 3.78 1.28 28* 1.59 0.63 0.25 1.15 1.75 0.52 29* 1.24 0.31 0.08 0.99 0.83 0.21 30 1.65 0.39 0.16 0.51 1.41 0.18 Ave 1.83 0.64 0.31 1.03 1.90 0.51 *Soils selected for further study at MSU. 50 Table 8.—-Corre1ation coefficients showing the relationship between various K extractants and K uptake by crOpping. Uptake Uptake NH4OA (Huancayo) (La Molina) (Peru)C H2804 (Peru) 0.627** 0.394 0.718** NH4OAc (Peru) 0.626** 0.532* --- Uptake (La Molina) 0.644** --— .___ - --- 0.953** _NH4OAc (M1Ch) - *Significant at 0.05 level. **Significant at 0.01 level. Sixteen of the 30 soils were selected on the basis of variations in K uptake, location, exchangeable K level, and distribution of K in the profile for further cropping and laboratory studies at MSU. Cropping study in Michigan The 16 Peruvian tensively cropped using to evaluate the release oat crops were grown on soils and 2 Michigan soils were in- the Stanford and DeMent technique of K to plants. Five successive the soils with the mat of oat roots being in contact with the soil for 14 days for each crOp- ping period. 51 Dry matter yields of the oats tops (Table 9) varied from 1.79 to 2.33 g for the lst harvest and from 0.94 to 1.54 g for the 5th harvest. When the dry matter yields from the sand checks without soil were subtracted, the total accumulated dry matter yields varied from 1.54 to 3.98 g per 100 g soil, which corresponds to 15.4 to 39.8 tons per Ha. Soil 29, a weathered jungle soil, produced the least and soil 27, a relatively unweathered soil from the coast, produced the most dry matter. Dry matter yields from each soil, although still slowly decreasing, appeared to be similar for harvests 3, 4 and 5. Concentrations of K on a dry matter basis in the oats tOps from 5 successive crops (Table 10) varied from 0.59 to 2.80% in the first harvest and from 0.38 to 1.16% for the 5th harvest. DeMent gt_al. (1959) considered young oat plants containing 0.58% or less K to be extremely K deficient. Broeshart and Van Schouwenburg (1961) also reported K contents of 0.54% in 18 day old oat plants as being in the low range. There were only 2 soils, 26 (moun- tain) and 27 (coast), which produced oats plants in the 5th crop that had K concentrations substantially higher (1.16 and 0.87%, respectively) than these deficient levels. The 52 Table 9.--Yields of each of five successive two-week oat crOps grown5, 2 4 6 8 IO 5 TIME (WEEKS) h] 1% [.— 3 I4 6O'J SOUTHERN MOUNTAIN SOILS . m 0 2 2 o s ' '0 2 . 8 3o. . . I M ' l2 /: ‘J + #4:) E A is 'e lo TIME (WEEKS) Fig. I.--Cumu1ative uptake of potassium by five successive two-week oat crOps grown on soils from the northern and southern mountain areas of Peru. 57 26 IZO‘ to 9 o: ‘3 2|, l7, l6 A CUMULATIVE UPTAKE (mg K/IOOg SOIL) CD CD ‘1 I 6 TIME (WEEKS) N" .5 ma 5 Fig. 2.--CumuIative uptake of potassium by five successive two-week oat crops grown on soils from the central mountain region of Peru. I20" :0 C? CUMULATIVE UPTAKE (mg K/ IO 09 SOIL) (D (3 tn ‘3 58 N b L L -4 ach c) l l l 2 4 6 8 TIME (WEEKS) 5 Fig. 3.--Cumulative uptake of potassium by five successive two-week oat craps grown on soils from the coastal (27 and 28) and jungle (24 and 29) regions of Peru. 59 m o 34 m c? CUMULATIVE UPTAKE (mg K/lOOg SOIL) 30" in 2 4 6 8 IO TIME (WEEKS) Fig. 4.--Cumulative uptake of potassium by five successive two-week oat crOps grown on two Michigan soils. 60 from the northern and southern mountain regions are given in Figure l, the central mountain region in Figure 2, the coast and jungle areas in Figure 3, and from Michigan, Figure 4. Since there appeared to be a linear relationship between cumulative uptake and time between 4 and 10 weeks, the slopes of the linear regres- sions were calculated for each soil during this period. The slopes or rate of K release as me K per 100 g soil per week are given in Table 12, together with coefficients of determination for each slope (linear regression coeffi— cient). With one half of the coefficients of determina- tion above 0.90 (90% of the variability in K uptake is explained by increasing the time of cropping), it appears that these slopes may give an estimate of the rate of re- lease of K to plants over time. The highest K release rates of 0.1501 and 0.2146 me K/lOO g/week were obtained from soils 27 (coast) and 26 (mountain) respectively while 0.0157 and 0.0159 me K/100 g/week from soils 3 (mountain) and 29 (jungle) were the lowest. More than a tenfold dif- ference in ratessof K release from soils crOpped inten— sively with oats may be expected from these soils. 61 Table 12.--Linear regression coefficients expressed as rate of K release and corresponding coeffi- cients of determination computed from accum- ulated uptake data for the 2nd, 3rd, 4th, and 5th oat harvests from 16 Peruvian and 2 Michigan soils. Soil Rate of Coefficient of no. K release determigatlon (me/100 g/week) (R ) 3 0.0157 0.689 4 0.0321 0.633 5 0.0365 0.962 6 0.0676 0.863 8 0.0410 0.961 10 0.0452 0.973 12 0.0218 0.863 14 0.0746 0.965 16 0.0292 0.640 17 0.0291 0.969 21 0.0377 0.817 24 0.0347 0.901 26 0.2146 0.941 27 0.1501 0.949 28 0.0715 0.970 29 0.0159 0.624 31 0.0186 0.604 34 0.0787 0.882 62 The exact mechanism controlling K release is diffi- cult to determine, since mass flow, diffusion and root in- terception are probably not occurring in the same relative proportions as was outlined by Barber (1962). In this case mass flow is probably more important in supplying K to plants since daily additions of water were made. Since the roots are in constant contact with the soil, root in- terception may be more important in supplying K to the plants and the roots can act as a more efficient sink for K removal. However, diffusion within the soil particles will probably ultimately determine K uptake. More discus— sion of these rates of release will be given in a later section. Exchangeable K levels before and after cropping, K removal by cropping from exchangeable and nonexchange— able forms and K uptake as a percent of the total K are given in Table 13. Potassium removal from exchangeable forms was calculated as the difference between the ammon- ium acetate extraction before and after cropping; nonex- changeable K removal was determined as the difference between total K uptake by plants and the decrease in ex- changeable K due to cropping. Exchangeable K levels of the different soils were decreased from 0.04 to 0.82 63 00 mv.a 0N 50.0 0.0 00.H HH.0 0v.0 vm 00 0N.0 00 00.0 N.N 00.0 00.0 «0.0 Hm H0 0H.0 mv 5H.0 H.0 00.0 00.0 mm.0 mm mm 00.H Hm bN.0 0.N hN.H 0H.0 «0.0 mm mm h0.~ 0H 50.0 m.0 vm.m va.0 Hr.0 hm hm ma.m m HH.0 m.m vN.m 0H.0 HN.0 0m 00 H0.0 5 v0.0 0.0 00.0 00.0 0H.0 vm >0 00.0 mm bN.0 h.~ Hm.0 ma.0 00.0 Hm 0v Hm.0 N0 v0.0 0.H 00.0 00.0 Nv.0 ha 00 00.0 0H 00.0 N.H 00.0 00.0 0H.0 0H 50 mv.H m 00.0 0.N mm.a 00.0 ma.0 0H 00 00.0 NH 00.0 N.H 00.0 00.0 mH.0 NH 00 00.0 Hm 00.0 h.H 00.0 0H.0 0v.0 0H 05 00.0 mm 0H.0 N.N 00.0 0H.0 mm.0 0 00 00.0 0H hH.0 0.0 0H.H hm.0 00.0 0 NS 00.0 mm mm.0 0.0. mm.0 HH.0 vm.0 0 vv 00.0 00 mm.0 N.NH nv.a NH.0 00.0 0 av 0N.0 mm 0N.0 0.NH 00.0 00.0 00.0 m oxmums A000H\oEv oxmums A000H\mfiv M A000H\ofiv 0swmmouo mcflmmouo M H33 M 138 you: 3003 mo ucoowmm M mo usoouom M Hmuou oxmumn .0: manmommmnochoz manmumdmsoxm mo unwoumm Hmuoe A000H\ofiv Hwom M manmomcmnoxm Edwmmmuom Hwom mo mxwums - . .Honsmgo buzowm m ca 0mmmowo mam>fimcmusfl mums manm cmwflnoflz N can cmw>suom 0H 20:3 mooHsOm oaamomcmgoxoco: 0cm manm nomcmnoxo Eowm 00>Auow wxmums M mo w 0cm mawom :0 M Hmuou 00 a .oxdwmd M Hmuou mm commowm Ixo M Hfiom mo mxmum: 0cm macammowo o>flmmooosm 0 woumm 0cm ouomon mao>oa M manmomcmnoxmul.ma manna 64 me/100 g by crOpping, with the greatest decrease in ex- changeable ngith soil 4 (mountain) and least decrease with soil 24 (jungle). From 3 to 59% of the total K up- take by plants was from exchangeable K. For 8 of the 18 soils'cropped, 20% or less‘of the total uptake came from exchangeable sources and for 14 of the 18 soils cropped, more than 50% of the net K uptake was from nonexchange- able K sources. Potassium extracted by ammonium acetate would be a poor indication of K release to plants for those soils with low exchangeable K levels, but which released sub- stantial amounts of K to plants. Mountain soils 14 and 26 had exchangeable K levels before cropping of 0.13 and 0.21 me/100 9 soil respectively, but released 1.53 and 3.24 me/100 g to oats plants in 5 successive croppings with 97% of the net K uptake coming from nonexchangeable sources. Of those soils from which more than one me of K was removed by cropping, soil 4 (mountain) was the only soil in which more than 50% of the total K uptake was from the exchangeable form. This soil had the highest initial exchangeable K level (0.94 me/100 g), but it de- creased to 0.12 me/100 9 after crOpping. Soils 3 and 17 65 from the mountains, 29 from the jungle, and 31 from Mich- igan also furnished substantial amounts of K to oats plants from exchangeable forms. From mountain soils 3, 4 and 6, coastal soil 27 and jungle soil 29, more than 5% of the total K was removed by intensive cropping, with 12.5 and 12.2 percent being removed from soils 3 and 4, respectively. Soils such as these will require frequent K fertilizer additions to avoid K deficiencies. The K extracted by neutral, normal ammonium ace- tate was correlated with the total uptake of K by oats plants for each of the five harvests. As the soils were successively crOpped, the correlations between initially exchangeable K and the K taken up by plants decreased as is shown by a decrease in the correlation coefficients from 0.693 to 0.388 (Table 14). The importance of non- exchangeable K when intensively cropping soils is shown by the correlation coefficient of 0.973 between total up- take of K by plants and the uptake of nonexchangeable K. 'The relationship between K uptake from the first cropping and.K:remova1 from nonexchangeable forms was higher (r = (3.793) than that for K removed from exchangeable sources (r == 0.667), indicating the importance of K from nonex- changeable forms even at the beginning of intensive 66 Table 14.--Correlation coefficients between accumulated net uptake of K after each of 5 successive crOps and exchangeable soil K before and after cropping, and uptake of exchange- able and nonexchangeable soil K by oats grown on 16 Peru- vian and 2 Michigan soils. l I E Accumulated xchangeable K Uptake Sources t tak ne up e Before After Exchangeable Nonexchangeable by harvest . . crOpping crOpping Ki Kit 1 .693** .263 .667** .793** 1,2 .535* .264 .498* .914** 1,2,3 .457* .258 .417 .950** 1,2,3,4 .423 .268 .377 .963** 1,2,3,4,5 .388 .265 .342 .973** + Uptake of exchangeable K = Exchangeable K before cropping - Ex- changeable K after crOpping. ++Uptake of nonexchangeable K = Total K uptake - Uptake of exchange- able K. *Significant at 0.05 level. **Significant at 0.01 level. cropping. Indications from this study are that for soils which will be intensively crOpped, ammonium acetate will not suffice to extract amounts of K which will be related to uptake. Most field crOps, however, will not remove the large amounts of K from a soil that was removed here by intensive cropping and ammonium acetate in general has adequately predicted K uptake under field conditions. 67 Sodium Tetraphenylboron Extractions Sodium tetraphenylboron (NaTPB) when employed as a precipitating agent for K insures that the K in solution will be at a low level (Ksp KTPB = 2.24 X 10-8; Geilmann and Gebauhr, 1953). With low K levels in solution, diffu- sion of K from the interlayers of illites and micas in- creases in an attempt to establish equilibrium. Sodium ions are added to the mixture so that they may replace the K ions diffusing out and thus maintain electrical neutrality in the clay mineral. One 9 soil samples were allowed to stand with 10 m1 of solutions which were 0.3 M NaTPB and 1.7 M NaCl at constant temperature (25 C) for 0.25, 1, 10, 100, 1,000 and 2,000 hr time periods to de- termine rates of K release and the total amount of K which could be removed by exchange reactions. The amounts of K which were precipitated as KTPB from 16 Peruvian and 2 Michigan soils for different time periods are given in Table 15. After the 15 min exposure time, the amounts of K released from the soils ranged from 0.8 to 4.4 me/100 g with the unweathered coastal soil (27) releasing the most and 2 soils from the northern :mountains (3 and 5) releasing the least. Compared to 15 68 Table 15.--Potassium extracted (me/100 g) from 16 Peruvian and 2 Michigan soils by solutions containing sodium tetraphenylboron for different time periods. Extraction time (hours) Percent Soil ‘ of total no. 0.25 1 10 100 1000 2000 soil K removed ...... ..... ...me K/100 g..... ...... .... (2000 hr) 3 0.8 1.0 1.7 2.7 2.9 2.8 73.2 4 1.9 2.2 2.5 4.1 3.7 3.8 31.9 5 0.8 1.1 1.7 1.8 1.5 1.9 8.7 6 2.3 3.2 4.3 5.9 6.2 6.9 38.5 8 4.1 4.1 5.6 6.6 7.3 7.4 19.9 10 1.9 2.5 4.3 7.5 9.0 9.5 16.7 12 1.2 1.4 3.1 5.9 6.0 6.4 17.2 14 2.7 4.9 20.5 48.3 64.3 66.9 86.6 16 1.8 2.5 5.0 17.6 36.2 38.5 73.8 17 1.7 1.7 4.9 12.8 20.3 21.0 47.9 21 1.8 1.9 2.9 3.6 4 3 4.3 14.5 24 1.4 1.4 3.7 8.2 14.1 16.4 27.9 26 2.5 4.2 11.7 26.2 45 0 52.2 51.4 27 4.4 5.4 8.9 14.3 18.8 18.9 36.8 28 3.3 5.4 11.4 17.5 18 7 19.6 40.1 29 1.1 1.0 1.6 2.7 3 7 3.3 48.0 31 1.8 1 3 l 9 2.9 4.8 4.9 16.5 34 3.9 5.9 11 5 18.7 21.2 23.7 48 2 69 min there was a small increase in the range of amounts of K extracted during the 1 hr time period (1.0 to 5.9 me/ 100 g). After 10 hr most of the soils which were to even- tually release large amounts of K could be recognized, namely mountain soils 14 (20.5 me/100 g) and 26 (11.7), coastal soils 27 (8.9) and 28 (11.4) and the Michigan soil 34 (11.5). At 10 hr soil 29, a weathered jungle soil, had released the least amount of K (1.7 me/100 g). Soils 16 and 17 (mountain) and 24 (jungle) also were to release substantial amounts of K but did not show an in- dication of their releasing ability until the 100 hr time period when they released 17.6, 12.8 and 8.2 me/100 g, respectively. At this time period the mountain soil 14 had released the most (48.3 me/100 g) and soil 5 the least K (1.8 me/100 g). Soils 3, 4, 5, 6, 8, 12 and 21 (mountain) and 29 (jungle) released less than 1.0 me/100 g from 100 hr to 1000 hr and thus during this time period had essentially reached their maximum capability to supply K. Amounts of K which could be expected from these soils by exchange reactions varied from 1.5 to 6.2 me/100 9 soil. After 1000 hr as after 100 hr, mountain soil 14 released the greatest amount of K (64.3 me/100 g). For the mountain soils 10 and 17, the Coastal soils 27 and 7O 28, and the Michigan soil 31, the release of K after 2000 hr as compared to 1000 hr was less than 1.0 me/100 g; thus after 2000 hr extraction time, 13 of the 18 soils had essentially stopped releasing K. Of the remaining soils 14, 16 and 26 (mountain), 24 (jungle), and 34 (Michigan), only soil 34 re- leased as much K from 1000 to 2000 hr as it had from 100 to ‘1000 hr, and this was a low amount, 2.5 me/100 g. Soils 24 and 26 from the mountains released the most K of any of the soils, 66.9 and 52.2 me/100 g respectively, and soils 3 and 5, also from the mountains, released the least, 2.8 and 1.9 me/100 g respectively. The percentages of the total soil K removed by precipi- tation with NaTPB from each soil after the 2000 hr extraction period are given in Table 15. Comparisons between the amount of K extracted after 2000 hr with NATPB and its percent of the total K indicate variations in K sources and release mechanisms. Mountain soils 3 and 16 both released approximately 73% of their total K after 2000 hr; however, soil 3 released only 2.8 me/100 9 while 38.5 me/100 g was removed from soil 16. Although these soils both released a high percentage of their total K, there was a greater than tenfold difference in the amounts of K removed. Soils 5, 8, 10, 12 and 21 from the mountains, and 31 from Michigan were more consistent in that they all 71 released less than 10 me K/100 g which amounted to less than 20% of their total K. The low amounts of K removed and corresponding low percent of the total K indicates the presence of framework silicates which would release only small amounts of K by this method. There was con- siderable variation in the amounts of K (3.8, 6.9, 16.4, and 18.9 me/100 g) removed from the respective mountain (4 and 6), jungle (24), and coastal (27) soils even though the corresponding percentages of total K (31.9, 38.5, 27.9 and 36.8) were similar. Soils 17, 28, 29 and 34 whose re- spective locations were the mountains, coast, jungle and Michigan all released between 40 and 50% of their total K, and except for the jungle soil, the corresponding amounts of K removed (21.0, 19.6, 3.3 and 23.7 me/100 g) were also about the same. Thus, a substantial amount of the K found in these soils can be removed by exchange reactions. Those soils which released the largest amounts of their total K (soils 3, 14, 16 and 26) were all from the mountains of Peru. However, the 73.2 percent of the total K released from soil 3 was only 2.8 me/100 g. The 66.9 me K/100 g from soil 14 was the largest amount and the highest per- cent of the total K (86.6) removed of any of the soils in the study. The maximum K removed from soils by this 72 method reported by Smith et_al. (1968) was 24.6 me/100 g which corresponded to 47% of the total K. The relationships between amount of K extracted with NaTPB and loqarithm of contact time are illustrated in figures 5 and 6. The three curves in Figure 5 repre- sent 14 of the 18 soils which had similar release pat- terns. The mountain (3, 4, 5 and 21), jungle (29), and Michigan (31) soils represented by curve C, released K in low amounts and at a low rate. Larger amounts of K were released by mountain soils 6, 8, 10 and 12 as is shown by curve B; however, these soils also were slow releasers of K. Curve A represents soils 17 (mountain), 27 and 28 (coastal) and 34 (Michigan) and shows an over- all faster release rate than either B or C. The soils represented by curve A would be expected to release mod- erate amounts of K until a maximum had been reached. There were no common release patterns noted for the re- maining soils presented in Figure 6. It is apparent that soils 24 (jungle) and 26 (mountain) are still releasing K even after 2000 hr although the rate of K release is higher for soil 26 than for soil 24. The rate of K re- lease from the mountain soils 14 and 16 is decreasing rapidly after high release rates which removed substan- tial amounts of K from these soils. 73 N O L - I I K EXTRACTED (me/I009 SOIL) a 5. I I r I I0 I00 IOOO EXTRACTION TIME (HOURS) T Fig. 5.-—Potassium extracted from soils I7, 27, 28 and 34 (curve A), 6, 8, l0 and l2 (curve B) and 3, 4, 5, 21, 29 and 3l (curve C) with sodium tetraphenylboron for diffrent time periods. 74 70' I4 60' 26 3 . ° 550 o (I) on O o . l6 <40 0 O o S 3 830 q ‘o a: t: 1.120 o 24 x o :0- ° . °~fl駗l—’o l IO I00 IOOO EXTRACTION TIMEIHOURS) ig. 6.--Potassium extracted from soils l4, I6, 24 and 26 with sodium tetraphenylboron for different time periods. 75 An extractable K rate index as suggested by Smith 3531. (1968) was computed by dividing the difference in K removed from 10 to 1000 hr by the difference of the loga- rithms of the respective times. Although the units of these numbers would be me/100 g, they represent an average over a logarithmic scale and thus should be used only for comparative purposes among soils. It can be inferred from these numbers that soils 3, 4, 5, 6, 8, 12 and 21 from the mountains, 29 from the jungle, and 31 from Michigan are relatively slow K releasers, soils 10 and 17 from the mountains, 24 from the jungle, 27 and 28 from the coast, and 34 from Michigan will release K at a medium rate and K can be removed at a rapid rate from mountain soils 14, 16 and 26. The K release during this same time period was also calculated as an average release rate of K (me K/100 g/wk) so that these rates (Table 16) could be compared it}: those obtained by crOpping (Table 12) . The coeffi- ients of determination computed for linear regression nations (Table 16) computed with no logarithmic trans- rmation of the time variable are indicators of the lidity in using this average release rate for each par- :ular soil. In most cases the fit of the regression 76 Tflfle]fi.--Rate indices, rate of K release, and coeffi- cients of determination for linear regression equations describing removal of K from 16 Peruvian and 2 Michigan soils by solutions containing sodium tetraphenylboron in'contact with soil samples for 10, 100 and 1000 hr periods. m . . Average . Linear 8011 Rate index rate release Linear R2 _. * *** no. 10 1000 10_1000** R Log**** ..(me K/lOOg/wk).. 3 0.60 0.20 0.50 0.87 4 0.60 0.20 0.14 0.54 5 -0.10 —0.03 0.72 0.34 6 0.95 0.32 0.47 0.79 8 0.85 0.29 0.67 0.93 10 2.35 0.80 0.61 0.93 12 1.45 0.49 0.35 0.75 14 21.90 7.43 0.68 0.97 16 15.60 5.30 0.89 0.99 17 7.70 2.61 0.81 0.99 21 0.70 0.23 0.85 0.99 24 5.20 1.76 0.87 0.99 .26 16.65 5.65 0.87 0.99 .27 4.95 1.68 0.77 0.99 428 3.65 1.24 0.46 0.85 29 1.05 0.36 0.82 1.00 31 1.45 0.49 0.91 0.95 34 4.85 1.65 0.57 0.92 K extracted (1000 hr) - K extracted (10 hr) log 1000 - log 10 K extracted (lOOO hr) - K extracted (10 hr) 1000 - 10 *Rate index = **Average K release rate = ***R2 values for linear regression equation: me K = a+b(time) with time ranging from 10 to 1000 hr. r"*"fi‘RZ values for linear regression equation: me K = a+b(1og time) with time ranging from 10 to 1000 hr. 77 line to the data was much better with. a logarithmic trans- formation of time. Smith et a1. (1968) have reported a linear relationship between the amount of K extracted with NaTPB and the logarithm of the contact time for 16 Iowa soils. More discussion of these release rates will follow in another section. Soil Solution Potassium Determinations By equilibrating soil samples with solutions of varying K concentrations, an estimate of the soil solu- tion K concentration can be obtained for that particular time. After equilibration the concentrations of K in so- lution are obtained and compared to the original concen- tration of the solution which was equilibrated with the soil sample. When plotting the difference in these two concentrations (AK) against the concentration of K orig- inally in the solution, a linear relationship should be obtained. The point or K concentration where K is neither adsorbed nor released by the soil (AK = 0) is considered to be the concentration of K in the soil solution. A value similar to this called the activity ratio was 78 computed by Beckett (1964) using the activities of K, Ca and Mg as AK/VA rather than the K concentration (Ca + Mg) in the soil solution. Values obtained by determining only the K concen- tration in solution are given in Table 17. The range in K concentrations from 3.7 to 72.0 ppm is in accord with values from 3 to 156 ppm found by Barber (1962) in the literature. Nine of the eighteen samples had values less than 10 ppm and fifteen were below 30 ppm. This is simi- lar to the frequency distribution of the K concentration in the saturation extracts of 142 soils from the midwest of the United States (Barber et al., 1962). 79 Table l7.--The concentration of potassium (ppm K) in soil extracts after equilibration with solutions of varying potassium concentrations and the soil solution potassium concentration calculated from these data for 16 Peruvian and 2 Michigan soils. . Potassium concentration of initial ppm K So11 . . . no. solution (ppm K) in soil 0 10.4 20.5 40.7 84.4 solution ,:.ppm K in solution after equilibration... 3 11.2 18.6 27.5 44.7 82.8 72.0 4 19.1 23.0 27.3 36.3 61.5 32.8 5 5.5 8.7 12.2 19.5 36.4 8.0 6 2.7 5.5 7.6 14.3 30.5 3.7 8 4.1 8.3 11.6 20.3 38.9 6.9 10 5.1 10 0 15.0 25.7 50.3 9.6 12 2.2 8.6 15.0 28.4 64.0 5.8 14 2.5 7.7 12.5 22.3 45.0 5.0 16 2.2 8.0 11.3 20.4 36.3 5.0 17 8.2 13.8 20.7 33.5 68.5 21 0 21 4.0 7.5 12.7 22.3 46.1 6.0 24 2.1 7.7 14.7 26.9 59.5 4.5 26 6.0 11.9 19.3 33.6 70.5 16.0 27 15.6 19.7 25.5 37.1 68.0 32.2 28 5.9 10.5 15.8 26.3 50.6 10.6 29 7.2 14.4 22.3 36.9 76.0 27.0 31 7.5 14.0 21.0 34.9 72.5 27.2 34 5.7 11.0 15.6 24.8 47.2 11.5 DISCUSSION The amounts and rates of release of K from the different soils are extremely variable as are the K re- moval rates within a soil when comparing short-term K release to the total K supplying power of a soil. The clay mineralogy and various chemical measurements are oftentimes helpful in explaining many of these differ- ences. Potassium Release Patterns of Soil Groups Similarities in clay mineralogy and K release by cropping and NaTPB extractions were used to place the soils in several groups (Table 18). The soils which re- leased comparatively low, medium, and high amounts of K will be considered first, second, and third, respectively. Soils with-low potassium supplying capabilities Soils 3 and 29: Soil 29, a red, weathered jungle soil, and soil 3, a red, highly weathered mountain soil, 80 81. 0.0a vm.H Hm.o 0.HOH 00.0 v.H0 . N.N0 0HN.0 vm.m 0N 0.m 50.0 MH.0 N.hh M¢.h 0.00 0.00 th.0 mm.H vH cootoo.noooouocooaoocooooooooococoommfluflqflnmmmu mqflmfiqmsm sawmmmug gmufic cuUH} “Hwomoooocoooooooooococoa-000000000 m.v Nv.0 0H.0 0.0m OF.H 0.5N ¢.0H mm0.0 mm.0 GN 0.HN 05.H N¢.O h.mv H0.N 0.5v 0.HN 0N0.0 no.0 5H 0.m m0.0 mH.0 N.Nm 0m.m 0.Mh m.mm 0N0.0 $0.0 0H m.HH ON.H mv.0 H.0v m0.H N.m¢ h.MN 050.0 mm.H vm 0.0H 00.H Nv.0 w.mv vN.H H.0v 0.0M Hh0.0 NN.H 0N N.Nm 0m.m Hh.0 ¢.Hm m®.H 0.0m 0.0H OmH.0 vN.m hm Ono-00.acoo00.000000ooooooooooooomwfluflflflnamo unwmfimqsm Eaflmmmug ESNBE cum”; mNuflomlbuoonoooooooooooooooooooooooo 0.0 OH.N mv.0 0.0m 00.0 F.0H m.0 mV0.0 00.0 CH h.m m0.H vv.0 0.0H Nm.0 m.mm 0.0 000.0 ©H.H 0 0.0 hm.H 06.0 0.0N VN.0 m.¢H m.v mm0.0 Hm.0 HN 0.0 hO.H mm.0 m.hm 0N.0 m.mH v.5 H60.0 mm.0 m 0.0 05.0 vm.0 0.NN m0.0l fi.m ®.H 0M0.0 mm.0 m N.bN ©m.m Nv.0 v.0N mv.0 m.®H 0.V 0H0.0 v0.0 HM m.m MM.H MH.0 H.0m mv.0 N.hH V.® NN0.0 Mv.0 NH m.Nm VM.H V0.0 0.NH 0N.0 m.Hm m.m NM0.0 h¢.H ¢ 0.0N HH.H MN.0 0.0 0m.0 O.mv m.m ©H0.0 mm.0 0N 0.Nh Oh.m mm.0 m.m 0N.0 N.Mh 0.N 0H0.0 0v.0 m I0000o.o.co00000onocoo.cocoa-onloocommfluufiflwnamo mew-afimmsm sflmmmug 30H guflk mNNOWOIOOIcocoooouoocooooooooooono. lemme iomo lacesxmec imooH\mec ixzxoooa\x mes x Manon Amooaxmec 1x3\mooH\x wee Aoooa\x mes M M0 #0 M HS OOOHIOH MO ucmoumm .OC COHUSHOW Hunom EOHW wmmmHmN mmmwde Vm meumg Ho HHOW M wHDMQWMWSUxm HMUOB MO wumm COH#UMHUX0 HE OOON NO wflmm HMHOB H. m wucwsmuswmwa x anew maemz an Hm>oEmu x manaaouo Mn Hm>oswu x I, .mH magma 82 were very similar in their mineralogical and chemical prOperties. The K in exchangeable form in soil 3, al- though higher than that in soil 29 (0.35 and 0.23 me/ 100 g respectively) is not as strongly adsorbed as that in soil 29 as is indicated by soil solution K concentra- tions of 72 and 27 ppm and K saturation percentages of 3.70 and 1.11, respectively. Unpublished results of field experiments on soil 3 have shown an apparent re- sponse to Mg fertilizers; the high concentrations of solution K measured in this soil may have been a con- tributing factor to the development of Mg deficiency. This substantiates the need for cation balance studies for these weathered soils with low cation exchange cab pacities. Long term K release will be low, as is indi- cated by an extremely low K rate removal by crOpping (0.016 and 0.016 me K/lOO g/wk) and NaTPB (0.20 and 0.36 me K/lOO g/wk) for soils 3 and 29, respectively, and is substantiated by low amounts of K bearing minerals in the clay fraction. Soil 4: Low amounts of K would be expected to be :released from this soil once the initial exchangeable K luas been depleted. High soil solution K concentration ‘ In" 83 (32.8 ppm), a relatively high K saturation of the CEC (1.34%), and high exchangeable K (0.94 me/100 g) are re- flected in a high initial K release to crOpping which decreased rapidly after the second crop. Only small amounts of illite were present in the clay fraction and exchangeable K accounted for 7.8% of the total K in the soil. Regular monitoring of exchangeable K levels will probably be necessary and although this was not a liming study, adding lime to neutralize acidity due to Al (TEA 5.5 me/100 9) should enable this soil to adsorb more exchangeable K since K has been found to compete more favorably with Ca than with Al for exchange sites (Mahilum et a1. , 1970) . Soils 12 and 31: Although there were differences in the clay mineralogy of these soils the chemical prOp- the K sources in soil 12, a mountain erties were similar: were interstratified illite, montmorillonite, and soil, feldspar and the K sources in 31, a Michigan soil, were vermiculite and amorphous material. Rates of K removal by cropping were similar for soils 12 (0.022 me/100 g/wk) (0.019 me/100 g/wk) , although the K uptake varied and 31 (0.43 and 0.64 me/100 g) as did the exchangeable K (0.13 84 and 0.42 me/100 g) and the K in the soil solution (5.8 and 27.2 ppm), respectively. It appears that once the initial K levels are decreased by cropping as predicted by exchangeable and soil solution K the different K sources as indicated above would release K at similar rates. This is substantiated by the NaTPB extractable K which was 6.4 and 4.9 me/100 g and 17.2 and 16.5 per- cent of the total K for soils 12 and 31, respectively. Soils 5, 8 and 21: Montmorillonite (5) and amor- phous material (8 and 21) with their similar cation ex- change capacities (%100 me/100 g) appear to be the major contributors of exchange sites for K in these mountain soils. Short-term rates of release as indicated by crop- ping were low and almost identical for these 3 soils (0.036, 0.038, 0.041 me K/100 g/wk for soils 5, 21 and 8 respectively), although further K release would be ex- pected to be much less from soil 5 than from soils 8 and 21 as is shown by the NaTPB extractions (-0.03, 0.29 and 0.24 me/100 g/wk). Low soil solution K levels of 8.0, 6.9 and 6.0 ppm and low to medium exchangeable K levels of 0.34, 0.33, and 0.40 me K/100 g for soils 5, 8 and 21, respectively, indicate that the amount of K which is 85 present on the exchange sites is tightly sorbed but that the equilibrium is rapid enough to supply K for low to moderate crop production. The sand and silt fractions contain more K than does the clay in soils 5 and 8 sug- gesting that major amounts of K in the soils is probably in primary minerals that would be expected to release only low amounts of K as is indicated by the above data. The K, although present in higher amounts in the clay in soil 21, is probably tightly held in the poorly struc- tured illite. There was no response to added K fertil- izer in a field experiment on soil 21 indicating that the amount of K removed by ammonium acetate (0.40 me/100 g) is an adequate level for crop production for this soil. A K saturation percentage of 1.57 which is about average for the soils in this study would indicate that K is com- peting favorably for the exchange sites. Thirty-three percent of the K uptake by crOpping was due to exchange- able K forms. It would seem that a response to K ferti- lizer could be expected in the near future if crop pro- duction was maintained at a high level. Soils 6 and 10: Potassium release patterns of the mountain soils 6 and 10 were similar as is noted by the K 86 release by crOpping (0.068 and 0.045 me K/100 g/wk) and the NaTPB extractions (0.32 and 0.80 me K/lOO g/wk, re- Spectively). Similar exchangeable K values of 0.44 and 0.45 me/100 g corresponded to similar total uptake values of 1.16 and 0.96 me K/100 g even though the soil solution K values of 3.7 and 9.6 ppm from soils 6 and 10, respec- tively, were more variable. The 6.9 and 9.5 me of K re- moved with NaTPB accounted for 38.5 and 16.7 percent of the total soil K from soils 6 and 10, respectively. This variation can probably be explained by the feldspar de— tected in the clay fraction of soil 10. Both soils con- tained poorly structured illite which is probably the main source of the smaller amounts of K which diffused from the interlayers and this structural disorder may be such that the K-0 bonds have been shortened which would make the K more difficult to remove. Soils with mediumppotassium supplying capabilities Soils 27, 28, and 34: Although the clay mineral makeup of these soils is very similar, a lower base sat- uration in the surface of the Michigan soil 34 (38.8%) indicates that this soil has been more intensely weathered 87 than the 2 Peruvian coastal soils 27 and 28, with their respective base saturations of 81.8 and 56.5%. Although soils 27, 28 and 34 all released similar amounts of K when treated with NaTPB (18.9, 19.6 and 23.7 me/100 g, respectively) which was 36.8, 40.1 and 48.2 percent of the total K, the soil solution and exchangeable K values were much higher for soil 27 (32.2 ppm and 0.71 me/100 g respectively) than for soils 28 (10.6 ppm and 0.42 me/ 100 g) and 34 (11.5 ppm and 0.48 me/100 g). This was reflected in the total K uptake and K release rates which were 3.24, 1.27 and 1.85 me/100 g and 0.150, 0.071 and 0.079 me K/100 g/wk for soils 27, 28 and 34, respectively. A liberal K fertilizer application to soil 27 in recent years would explain the above results. Since some of this K could have been fixed by the vermiculite, it would be more easily released as was shown by Mortland ep_al. (1957) than if the K had to come from illite. These soils all have a fairly well ordered illite component and K contents in the clay fraction above 2 per- cent which would support the high amounts of K removed by NaTPB. 88 Soils 16, 17 and 24: Except for no vermiculite in soil 17 (central mountains) and the presence of mont- morillonite in soil 24 (high central jungle) the clay mineralogy of these 3 soils is very similar. The percent K in the clay fraction (3.36%) was substantially higher, however, in soil 16 (central mountains) than in soils 17 (2.46%) and 24 (2.59%). Even though similar amounts of K were removed by cropping (0.64, 0.64 and 0.55 me/100 g from soils 16, 17 and 24 respectively) and the K release rates by cropping were similar (0.029, 0.029 and 0.035 me K/100 g/wk) the exchangeable (0.15, 0.42 and 0.10 me/ 100 g) and soil solution K (5.0, 21.0 and 4.5 ppm) levels were somewhat different. It may be that K fixed in ver- miculite in soils l6 and 24 was not measured by the ammon— ium acetate due to a blocking effect by the ammonium ion but this K was available to plants when a concentration gradient was established by the plant root. There was evidently some trioctahedral illite present in soil 16 as this soil released 38.5 me K/100 g which was 73.8% of its total K. This is substantiated somewhat by a 060 x-ray Spacing of 1.54 A and a low second order (002) i1- 1ite peak at 4.98 A. Vermiculite, illite and feldspars were probably responsible for the lower, but substantial, 89 amounts of K removed by NaTPB from soils 17 (0.65 me/100 g) and 24 (0.55 me/100 g). Soils with high potassium supplying capabilities Soils 14 and 26: Potassium was released at high rates from these mountain soils which were similar in their clay mineralogy. NaTPB removed 66.9 me/100 g (86.6% of the total K) from soil l4 and 52.2 me/100 g (51.4% of the total K) from soil 26. When these soils were cropped, soil 14 released 1.53 me K/100 g and soil 26 released 3.24 me K/100 g. The rates of K removal were 0.075 and 0.215 me/lOO g/wk by cropping and 7.43 and 5.65 me/100 g/wk with NaTPB for soils l4 and 26 respectively. The high amounts of K released by cropping were not indicated by the K con- centration in the soil solution (5.0 and 16.0 ppm) nor the low exchangeable K levels of 0.13 and 0.21 me/100 g re- spectively, from soils 14 and 26. Both of these soils have good crystalline illite peaks and the presence of dioctahedral and trioctahedral species are indicated by a high second order illite peak and the 060 X-ray diffrac— tion Spacing, respectively. They appear to be relatively unweathered as is shown by base saturation percentages in 90 most cases above 50% throughout the profile. For soils such as these, ammonium acetate as a K extractant is not adequate to predict a yield response to applied K ferti- lizer. Potassium Release Characterization Most of the soils in this study released amounts of K that represented maximum quantities which could be expected to be released from the soil by surface and in- terlayer exchange reactions. Soils 24 and 26 were the only soils which were still releasing substantial amounts of K at the 2000 hr extraction period. In general, the amounts of K which these soils were capable of releasing was related to the clay minerals, specifically illite, present in the clay fraction. Of the soils classified as having low K supplying capabilities (3, 4, 5, 6, 8, 10, 12, 21, 29 and 31) only soils 10 and 12 had X-ray diffraction patterns showing any illite present. This illite was very poorly structured as was indicated by low, broad 10 A peaks. Thus, very little interlayer K is available for release and most of 91 the K must come from either primary or secondary minerals in the larger size fractions or from surface exchange sites. Smith gp_§1. (1968) removed equal amounts of K (2.3 me/100 g) from <50u orthoclase feldspar with NaTPB extractions for 1 day and 15 wk time periods which indi- cates that small amounts of K can be expected to be re- leased from feldspar minerals. The soils in the low K supplying group, however, only released from 1.9 to 9.5 me/100 g at low rates varying from -0.03 to 0.80 me K/100 g/wk indicating that feldspars may be contributing sub- stantial amounts of the total K removed, especially for the lower K releasers. A good relationship exists be- tween the exchangeable K and the uptake of K from the first cr0p of oats for these 10 soils (r = 0.979). Thus for soils with none or low amounts of illite present, the surface-sorbed K is more important as a source of K for plant growth, and ammonium acetate should be an adequate extractant for removing K Which is related to that taken up by plants. Soils 16, 17, 24, 27, 28 and 34 were considered to have medium potassium supplying capabilities although soil 16 released amounts of K in the 2000 hr NaTPB extraction comparable to the amounts of K removed from the soils 92 considered to be high K releasers. The clay mineral i1- 1ite present in all of these soils was much more crystal- line as indicated by sharper and narrower peaks than that found in soils 10 and 12 of the previous group. In addi- tion to better structured illite, vermiculite was also present in all of the soils except number 17 which also had the poorest illite structure. It is possible that this vermiculite is slowly forming as K in the illite is removed. After 1500 hr, Scott and Reed (1962) were able to remove from a Grundite illite 74 me/100 g with NaTPB which amounted to 68% of the total K present. The soils classified as medium in K supplying power released from 16.4 to 38.5 me K/100 g which amounted to 27.9 and 73.8% of the total K, respectively. This would suggest that K in the illite in the above soils was a major source of K removed by NaTPB. Due to the presence of this illite, these soils would probably be able to supply more K at a given exchangeable K level. In this group are 2 of 4 soils from the central mountains, the 2 coastal soils, and 1 soil each from the jungle and Michigan. The two soils whibh were considered to have high K releasing capabilities had higher and sharper 10 A peaks than any of the previous soils. Vermiculite was 93 also present, but in smaller amounts as was indicated by peak heights. These soils are probably much younger in age than the other soils and appear to be relatively un- weathered. The K in this crystalline material, which may be mica, is much easier to remove as is shown by both cropping and NaTPB extractions. It appears that rela- tively small amounts of K need be removed to change some of this 10 A material to vermiculite in soil 26 (Figure 7). Supporting evidence such as K fixing ability and cation exchange capacities are needed to substantiate this since only 3 me K/100 g were removed in the cropping experiment. This soil should release much larger amounts of K by cropping if more intensive methods such as those of Mort- land gE_al. (1956) and Doll §E_§1. (1965) were used. When comparing the K removal by NaTPB and the appearance of the illite X-ray diffraction peaks, it appears that the more crystalline illites release K in larger amounts and at higher rates than those which are poorly structured. The diffusion of K out of the in- terlayers should be related to the crystallinity of the illite. Rich (1968) states that the type of bonding within the crystal structure, the extent and type of disorder within the crystal or at crystal terminations, 94 SOIL CLAY 26 BEFORE CROPPING MG-GLY AFTER CROPPING MG-GLY mama 3 3.3 3.5 4 s c 7 I: I0 I2 I4 no 24 as 1 l 1 l l l l l L J 1 l L 1 l l T T l 29 24 20 I6 I2 0 4 beam-:63 20 Fig. 7.—-X-ray diffraction patterns of magnesium saturated, glycerol * solvated clay from soil 26 before and after intensive crop- ping. 95 and crystal size are mineral properties that determine the rate at which weathering occurs and K is released. There were no good relationships between soil solution, exchangeable, total or NaTB-15 min extractable K and the total K uptake by oats in 5 successive crop- pings. The lowest K removal of any of the soils was by jungle soil 29 (276 Kg/Ha) and in most cases the K re- moved was substantially higher than amounts which most field crOps would be expected to remove in a cropping season. The relationship of exchangeable K was best with K uptake from the first crop. This would be ex- pected since the exchangeable K is an important K source in the earlier stages of cropping. The concentration of K in the soil solution was not well related with either K uptake from the first crOp or the total K uptake indi- cating the level of K in the soil solution is not a de- termining factor for uptake for these soils, but more important how fast the K in solution can be replenished. This is substantiated somewhat by a correlation between rate of K release by cropping and K uptake from the first crop of 0.751 and suggests that even at the first cropping, the rate of release or diffusion of K was influencing K uptake. The concentration of K in the soil solution was 96 somewhat related to the percent K saturation (r = 0.717) and to the total K in the soil (r = 0.786) indicating that the surface and inter-particle exchange sites for K do influence the level of K in solution to some extent. The total K contents of the soils was not related to the K uptake by the first crop, but there appeared to be a relationship between total K in soils and K uptake by plants for the 2nd, 3rd, 4th and 5th crOps, as shown by correlation coefficients of 0.662, 0.688, 0.687 and 0.714, respectively. Thus the total amount of K in the soil, most of which would be in the mineral form is re- lated to K removal under intensive crOpping conditions. This further substantiates the use of slopes of cumula- tive K uptake curves as indicators of the rate of K re- lease by cropping from these soils. The total K uptake was not as well related to the amount of K removed with NaTPB by 15 min extractions as has previously been shown by Schulte and Corey (1965) who reported a correlation coefficient of 0.991; however, their reported relationship between exchangeable K and K uptake by intensive cropping was also very good (r = 0.971). They also reported a 1:1 relationship between K uptake and the K removed with NaTPB after 15 min while 97 NaTPB in this study removed in many cases 3 times as much K from the Peruvian and Michigan soils as was removed by the 5 successive oats crOps. The intensive cropping experiment thus removed more K than what a field crop would be expected to re- move and less than that with the 15 min NaTPB extraction. Thus some of these soils will release substantial amounts of K to NaTPB even at short time periods which would sug- gest that the surface-sorbed K is not always readily available to plant roots. As an example of this, soil 8 released 4.1 me/100 g to NaTPB in 15 min. but only 0.83 me to 5 oats crops. X-ray diffraction patterns for this soil indicated essentially no clay minerals present which suggests that much of the K available for plants must be surface adsorbed K. In contrast, soil 26 released less K with the 15 min‘NaTPB extraction (2.5 me/100 9) than was removed by the 5 oats crops (3.2 me/100 g). In this case well-crystallized illite, maybe mica, was indicated by X-ray diffraction analysis and the K removed probably was mainly lattice K since the exchangeable K was very low (0.21 me/100 9). With a 15 min extraction period the amount of K removed was limited by the rate of dif- fusion.which was not the case for cropping since a larger 98 time period permitted more K to diffuse from the clay mineral and thus the plant root could remove it. With soils varying in mineralogy and therefore in K release mechanisms as these soils do, a simple ex- tractant may not be adequate to evaluate the K status of these soils, with respect to plant-available K. Po- tassium extracted with ammonium acetate seems to relate well with the K release by soils which have small amounts of poorly structured illite present such as those in the first grouping; however, for these soils with well crys- tallized illite present, the K releasing capacity is not adequately evaluated. SUMMARY AND CONCLUS IONS Soil profile data, kinds of clay minerals, extrac- tions with NaTPB, intensive cropping studies and exchange- able, soil solution, and total K determinations were used to evaluate the K status of 16 Peruvian and 2 Michigan soils. Variation in pH, exchangeable K, and percent base saturation through the soil profiles indicated that very few of the soils studied have been intensively weathered, On 1 mountain soil and 1 jungle soil the base saturation was near 10% and the pH values ranged from 4.0 to 4.5 through the profile indicating that they were the more intensely weathered soils of the group. In a few cases exchangeable K levels were higher in the subsurface than in the surface horizons suggesting that K measurements of the surface horizon may not always give a true indication of the level of plant-available K. Using soil test recom- mendations for potatoes in Peru published in 1969 (Fitts, 1969), K fertilizer would have been recommended for all but 2 of the 16 Peruvian soils. Metahalloysite was found in most of the mountain and jungle soils while kaolinite appeared to be present 99 100 in the coastal soils; however, the presence of any kao- linite in the jungle and mountain soils would have been masked by the metahalloysite peaks in the X-ray diffrac- tion patterns. Some illite was found in many of the soils, although in about one half of the soils the i1- lite was poorly structured and present in small amounts. If these soils are as young as is believed, the illite could just be forming from primary minerals. In most cases the chlorite and vermiculite present was in inter- stratified systems. There was a noticeable absence of feldspar in the clay fraction of the 4 northern mountain soils although it may be present in small amounts in the larger fractions, since total soil K levels for the soils varied from 0.15 to 0.86% K. One soil from the northern mountains which con- tained predominantly montmorillonite and another soil from the northern mountains did not contain quartz, sug- gesting that the areas in which these soils are located is not intensively weathered. Allophane was detected in one of the Peruvian soils and was suspected to be present in several others. To evaluate the susceptibility of interlayer K to exchange and determine how fast it will diffuse from the 101 clay lattice, soil samples were equilibrated with solu- tions containing NaTPB for different time periods. After 2000 hr, from 1.9 to 66.9 me K/100 g of K were removed which represented 8.7 and 86.6% of the total K present in the soils. Rates of K release as calculated by the K removed from 10 to 1000 hr, where the rate of K removed appeared to be decreasing at a logarithmic rate, varied from -0.03 to 7.43 me K/100 g/wk demonstrating the wide variability in the ability of these soils to release K. The amounts of K released and the rates of K release were related to the amount and type of clay minerals present in the clay fraction. The soils with low K supplying capacities contained none or very small amounts of i1- 1ite, and when illite was present, it was poorly struc- tured as indicated by low, broad 10 A peaks. Larger amounts of more crystalline illite, as indicated by higher, sharper 10 A peaks, was present in all of the soils which had medium K supplying capabilities. Illite in the soils with high K supplying capabilities was well crystallized and may possibly be muscovite and biotite micas. The amounts of available K and rates of K release from plant-available forms were determined by growing 102 five successive oat crops for two weeks each on 100 9 soil samples. Amounts of K removed by the 5 oat crops varied from 0.35 to 3.24 me K/100 9 soil and rates of K release, as determined by the linear regression slopes of cumula- tive K uptake from the 2nd to the 5th crop, varied from 0.016 to 0.215 me K/100 g/wk. The amounts of K removed by intensive cropping were not well correlated with K ex- tracted with ammonium acetate (r = 0.388); however, if only the uptake data from the first of the five crops were considered, the correlation was higher (r = 0.693), and was much higher when only those soils with low potas- sium supplying capabilities were used for computing the correlation coefficient (r = 0.979). This would be ex- pected since a major source of K for plants from these soils would be from exchangeable forms. In contrast those soils with well crystallized illite released substantial amounts of K during cropping, but had low exchangeable K levels indicating that rela- tively large amounts of K can diffuse from the clay min- eral lattice during cropping. Total K was extremely variable in the soil as was the total K content of the sand, silt and clay fractions. For those soils from which high amounts of K were removed 103 with NaTPB and by intensive crOpping, there was a higher correlation between K removed and the percent K in the sand than with the percent K in the silt, which further indicates that these soils are relatively unweathered. Total K was much better related with K uptake as the cropping progressed indicating that higher amounts of K were being removed from nonexchangeable sources and that the higher the total soil K, the easier it can be removed. The concentration of K in solution was not related to K uptake by plants for any of the croppings which sug- gests that for the soils studied, K capacity factors are much more important than K intensity factors in supplying K to plants. This may be misleading due to the conditions under which the crOps were grown, i.e. daily additions of water and a high concentration of roots in a small soil value. From this study, the following conclusions can be made: 1. The soils in Peru vary in degree of weathering as was indicated by variation in the chemical prop- erties of the soil profile and the clay minerals present. 104 Although there are wide differences in the ability of these soils to release K under stress condi- tions (NaTPB), variations in K release could fre- quently be explained by the presence and crystal- linity of illite in the clay fraction; using these two criteria the soils could be grouped as having low, medium and high K supplying capabilities. The poor relationship between K extracted with NaTPB for 15 min and K uptake by plants is prob- ably due to different sources of plant available K and the mechanism supplying K to plants. Total K uptake by short-term intensive cropping was not well related to K extracted with ammonium acetate although this relation can be improved by considering uptake from only the first harvest or considering only those soils classified as having low K supplying capacities. Due to differences in mineralogy and K release mechanisms, a single extractant may not be ade- quate to evaluate the ability of these soils to supply K to plants. LI'I'ERA'IUIE CITED LITERATURE CITED Allison, L. E. 1965. Organic carbon. £3 C. A. Black (ed) Methods of Soil Analysis, PartrZ,Amer. Soc. Agron., Madison. Pp. 1367-1378. Arnold, P. W. 1960. Nature and mode of weathering of soil-potassium reserves. J. Sci. Food Agri. 6:285-292. Arnold, P. W. and Close, B. M. 1961. Release of nonex- changeable potassium from some British soils crOpped in the glasshouse. J. Agri. Sci. 57:295-304. Asher, C. J. and Ozanne, P. G. 1967. Growth and potas- sium content of plants in solution culture main- tained at constant potassium concentrations. Soil Sci. 103(3):155-161. Barber, S. A. 1962. A diffusion and mass flow concept of soil nutrient availability. Soil Sci. 93: 39-49. Barber, S. A., Walker, J. M., and Vasey, E. H. 1962. Principles of ion movement through the soil to the plant root. Trans. Comm. IV and V, Intern. Soil Conf., New Zealand. A3:3-6. Bassalick, K. 1912. Uber silikatzersetzung durch boden- bakterien. Z. Garungsphysiol 2:1. £2_M. M. Kononova, Soil Organic Matter. P. 186. Bassett, W. A. 1959. The origin of the vermiculite de- posit at Libby, Montana. Am. Mineralogist 44: Beckett, P. H. T. 1964. Studies on soil potassium. I. Confirmation of the ratio law: Measurement of potassium potential. J. of Soil Sci. 15:1-8. 105 106 Black, C. A. 1968. Soil-Plant Relationships. John Wiley Sons Inc., New York, New York, pp. 654-773. Bowen, H. J. B. 1966. Trace Elements in Biochemistry, Academic Press, London and New York. P. 39. Broeshart, H. and Van Schouwenburg, J. C. 1961. Early diagnosis of mineral deficiencies by means of plant analysis. Netherlands Journ. Agri. Sci. 9:108-117. £2 Homer Chapman (ed) Diagnostic Criteria for Plants and Soils, University of California, Division of Agricultural Sciences, 1966. P. 663. Chapman, H. D. 1965. Cation exchange capacity. lg C. A. Black (ed) Methods of Soil Analysis, Part 2. Amer. Soc. Agron., Madison. Pp. 891-901. Day, Paul R. 1965. Particle fractionation and particle- size analysis. £2 C. A. Black (ed) Methods of Soil Analysis, Part 1. Amer. Soc. Agron., Madison. Pp. 545—567. DeMent, Jack D., Stanford, George, and Bradford, B. N. 1959. A method for measuring short-term nutrient absorption by plants: II. Potassium. SSSAP 23:47-51. Doll, E. C., Mortland, M. M., Lawton, K., and Ellis, B. G. 1965. Release of potassium from soil fractions during cropping. SSSAP 29:699-702. Drosdoff, Matthew, Quevedo, Felix, and Zamora, Carlos. 1960. Soils of Peru. 7th Intern. Congress of Soil Sci., Madison. V.13:97-104. Ellis, B. G. and Mortland, M. M. 1959. Rate of potassium release from fixed and native forms. SSSAP 23:451- 453. Estrada, Jose A. 1971. Mineralogical and chemical prop- erties of Peruvian acid tropical soils. Ph.D. Thesis, University of California, Riverside. 107 Evans, C. E. and Attoe, O. J. 1948. Potassium supplying power of virgin and cropped soils. Soil Sci. 66 (5):323-334. Failyer, G. H., Smith, J. G., and Wade, H. R. 1908. The mineral composition of soil particles. U.S. Dept. Agr. Bur. Soils Bul. 54. £2 C. A. Black, Soil-Plant Relationships, 2nd ed., Wiley, p. 655. Fieldes, M. and Perrott, K. W. 1966. The nature of allo- phane in soils. III. Rapid field and laboratory test for allophane. New Zealand Journal of Science 9:623-9. Fitts, J. W. 1969. Annual Report, International Soil Testing, Contract AID csd-287, North Carolina State University at Raleigh. Fitts, J. W. 1970. Annual Report, International Soil Testing, Contract AID la-646, North Carolina State University at Raleigh. Gandette, H. E., Eades, J. L., and Grim, R. E. 1964. The nature of illite. Clays and Clay Minerals, 13 conf. Pergamon Press. Pp. 33-48. Geilmann, W. and Gebauhr, W. 1953. Precipitation of alkali metals as tetraphenylboron compounds. Z. Anal. Chem. 139:161-181. £2 Schulte, E. E. and Corey, R. B. 1965. Extraction of potassium from soil with sodium tetraphenylboron. SSSAP 29:33-35. Grim, Ralph E. 1968. Clay Mineralogy. McGraw-Hill Book Company, New York, New York. Hipp, Billy W. and Thomas, Grant W. 1967. Method for predicting potassium uptake by grain sorghum. Agronomy Journal 60:467-469. Hoagland, D. R. and Arnon, D. I. 1950. The water-culture method for growing plants without soil. Calif. Agr. Expt. Sta. Cir. 347. 108 Huang, P. M., Crosson, L. S., and Rennie, D. A. 1968. Chemical dynamics of potassium release from po— tassium minerals common in soils. Trans. 9th Intern. Cong. Soil Sci., Adelaide. Vol 2:705- 711. Jackson, M. L. Pennington, R. P., and Mackie, W. Z. 1948. Crystal chemistry of soils. I. The fundamental structural groups and families of silicate minerals. SSSAP13:139-l45. Jackson, M. L., Hseung, Y., Corey, R. B., Evans, E. J. , and Van den Heuvel, R. C. 1952. Weathering se- quence of clay-size minerals in soils and sedi- ments: 11. Chemical weathering of layer sili- cates. SSSAP 16:3-6. Jan Beek, Klaas and Bramao, D. Luis. 1968. Nature and geography of South American soils. Ip E. J. Fittkau, J. Illies, H. Klinge, G. H. Schwabe, and H. Sioli (eds) Biogeography and Ecology in South America. Dr. W. Junk N.V., Publishers, The Hague. Pp. 82-112. Kunze, G. W. 1965. Pretreatment for mineralogical analy- sis. Ip C. A. Black (ed) Methods of Soil Analysis, Part 2. Amer. Soc. Agron., Madison. Pp. 568-577. Leonard, R. A. and Weed, S. B. 1970. Mica weathering rates as related to mica type and composition. Clays and Clay Minerals 18:187-195. Leonard, R. A., and Weed, S. B. 1970. Effects of potas- sium removal on the b-dimension of phlogopite. Clays and Clay Minerals 18:197-202. Mahilum, B. C., Fox, R. L., and Silva, J. A. 1970. Residual effects of liming volcanic ash soils in the humid tropics. Soil Sci. 109:102-9. McCollum, Robert E. and Valverde, Carlos. 1968. The fertilization of potatoes in Peru. 1. A summary and interpretation of data from field experiments completed from 1959 through 1964 in the sierra. North Carolina Agricultural Experiment Station Tech. Bul. No. 185. 109 McLean, E. O. 1965. Aluminum. £2 C. A. Black (ed) Methods of Soil Analysis, Part 2. Amer. Soc. Agron., Madison. Pp. 978-998. Miller, E. V. and Coleman, N. T. 1952. Colloidal prop- erties of soils from western equatorial South America. SSSAP 16:239-244. Mortland, M. M., Lawton, K., and Uehara, G. 1956. Al- teration of biotite to vermiculite by plant growth. Soil Sci. 82(6):477-481. Mortland, M. M., Lawton, K., and Uehara, G. 1957. Fix- ation and release of potassium by some clay min- erals. SSSAP 21:381-384. Mortland, M. M. 1958. Kinetics of potassium release from biotite. SSSAP 22:503-508. Mortland, M. M. and Ellis, Boyd. 1959. Release of fixed potassium as a diffusion controlled process. SSSAP 23:363-364. Mortland, M. M. 1961. The dynamic character of potassium release and fixation. Soil Sci. 91(1):11-13. Mortland, M. M. and Lawton, K. 1961. Relationship between particle size and potassium release from biotite and its analogues. SSSAP 25:473—476. Newman, A. C. D. and Brown, G. 1966. Chemical changes during alteration of micas. Clay minerals 6:297- 309. Novorossova, L. E., Remezov, N. P., and Suskina, N. H. 1947. The decomposition of aluminosilicates by soil bacteria. Dokl. Akad. Nauk. SSSR 58.4. I2 M. M. Kononova, Soil Organic Matter. P. 187. Olivera, V., Ludwick, A. E., and Beatty, M. T. 1971. Potassium removed from some southern Brazilian soils by exhaustive crOpping and chemical extrac- tion methods. SSSAP 35:763-767. 110 Quirk, J. P. and Chute, J. H. 1968. Potassium release from mica-like clay minerals. 9th Intern. Cong. Soil Sci., Adelaide. Vol. 2:671-680. Reed, M. G. and Scott, A. D. 1962. Kinetics of potas- sium release from biotite and muscovite in sodium tetraphenylboron solutions. SSSAP 26:437-440. Reichenbach, H. Graf Von and Rich, C. I. 1969. Potassium release from muscovite as influenced by particle size. Clays and Clay Minerals 17:23-29. Reitemeier, R. F. 1951. Soil potassium. Adv. in Agr. 3:113-164. Rich, Charles I. 1968. Mineralogy of soil potassium. 52 V. S. Kilmer, S. E. Younts, and N. C. Brady (eds) The Role of Potassium in Agriculture, Amer. Soc. of Agron., Madison. Pp. 79-108. Ross, G. C. and Kodama, H. 1970. Differential release of potassium from interstratified mica clay min- erals as related to probable differences in their mica layer components. Clays and Clay Minerals 18:151-6. Sanchez, P. A. and Buol, S. W. 1971. Characteristicas morfologicas, quimicas y mineralogicas de algunas suelos principales de la selva amazonica Peruana. Informe No. 56, Programa Nacional de Arroz, Mini- sterio de Agricultura, Zona Agraria II, Lambayeque, Peru. Schulte, E. E. and Corey, R. B. 1965. Extraction of po- tassium from soils with sodium tetraphenylboron. SSSAP 29:33-35. Scott, A. D. and Welch, L. F. 1961. Release of nonex- changeable soil potassium during short periods of cropping and sodium tetraphenylboron extrac- tion. SSSAP 25:128-131. Scott, A. D. and Reed, M. G. 1962. Chemical extraction; of potassium from soils and micaceous minerals I with solutions containing sodium tetraphenylboron: II. Biotite. SSSAP 26:41-45. 111 Scott, A. D. and Reed, M. G. 1962. Chemical extraction of potassium from soils and micaceous minerals with solutions containing sodium tetraphenylboron: III. Illite. SSSAP 26:45-48. Scott, A. D. and Smith, S. J. 1966. Susceptibility of interlayer potassium in micas to exchange with sodium. Clays and Clay Minerals, Proc. 14th Conf., Pergamon Press, New York. Pp. 69-81. Scott, A. D. 1968. Effect of particle size on inter- layer potassium exchange. 9th Intern. Cong. Soil Sci., Adelaide. Vol. 2:649-659. Shoemaker, H. E., McLean, E. 0., and Pratt, P. F. 1961. Buffer methods for determining lime requirements of soils with appreciable amounts of extractable aluminum. SSAP 25:274-277. Smith, S. J. and Scott, A. D. 1966. Extractable potas- sium in Grundite illite: 1. Method of extrac- tion. Soil Sci. 102(2):115-121. Smith, S. J., Clark, L. J., and Scott, A. D., 1968. Exchangeability of potassium in soils, Trans. 9th Intern. Cong. Soil Sci., Adelaide. Vol. 2: 661-669. Smith, S. J., Lieu, M., Clark, L. J., and Scott, A. D., 1971. Interpretation of curves describing soil potassium exchange. SSSAP 35:718-721. Stanford, George and DeMent, J. D. 1957. A method for measuring short-term nutrient absorption by plants: I. Phosphorus. SSSAP 21:612-617. Tanner, C. B. and Jackson, M. L. 1947. Nomographs of sedimentation times for soil particles under grav- ity or centrifugal acceleration. SSSAP 12:60-5. Thomas, Grant W. and Hipp, Billy W. 1968. Soil factors affecting potassium availability. £2: U. J. Kilmer, S. E. Younts, and N. C. Brady (eds.) The Role of Potassium in Agriculture. Amer. Soc. Agron., Madison. Pp. 269-291. 112 Tisdale, Samual L. and Nelson, Werner L. 1966. Soil Fer- tility and Fertilizers. The Macmillan Company, Collier-Macmillan Limited, London. Pp. 252-287. Tosi, Joseph A. 1960. Zonas de vida natural en el Peru. Boletin Tecnico No. 5 IICA, OEA. Tucker, B. M. 1964. The solubility of potassium from soil illites: I. The dependence of solubility on pH. Aust. J. Soil Res. 2:56-66. Tucker, B. M. 1964. The solubility of potassium from soil illites: II. Mechanisms of potassium re- lease. Aust. J. Soil Res. 2:67-75. Van der Marel, H. W. 1949. Mineralogical composition of a heath podzol profile. Soil Sci. 67:193-207. Weed, S. B., Davey, C. B., and Cook, M. G. 1969. Weather- ing of mica by fungi. SSSAP 33:702-706. Wells, C. B. and Norrish, K. 1968. Accelerated rates of release of interlayer potassium from micas. 9th Intern. Cong. Soil Sci., Adelaide. 2:683-694. Zamora, C. 1969. Inventario de estudios de suelos del Peru (segunda approximacion), Oficina Nacional de Evaluacion de Recursos Naturales, Lima. P. 25. Zavaleta Garcia, Amaro and Arca Bielick, Manuel. 1963. Grandes grupos de suelos identificados en forma generalizada en el Peru. Agronomia 30(3):37-53. Zavaleta, Amaro. 1969. Distribucion geographica y charac- teristicas de los suelos derivados de cenizas vol- canicas del Peru. £2 Panel sobre Suelos Derivados de Cenizas Volcanicas de America Latina. Centro de Ensenanza e Investigacion del IICA, Turrialba, Costa Rica. APPENDIX 113 .Plate 1.--X-ray diffraction patterns of the clay fractions from soils 3 and 4 (northern mountains). 114 SOILS M6 GLY 25‘ K 25’ K 300' K 550' MG 'GLY 26‘ muons 3 3.3 3.5 4 5 6 7 8 IO I2 I4 I8 24 36 .l 1 1 l l L L 1 l i l L 1 1 I I I I I I I 20 24 20 I6 I2 8 4 115 Plate 2.--X-ray diffraction patterns of the clay fractions from soils 5 and 6 (northern mountains). 116 SOIL 5 MG-GLY 250 I SOIL 6 MG-OLY 25' K 550' I j I I 20 l6 l2 6 4 DEGREES 20 117 Plate 3.--X-ray diffraction patterns of the clay fractions from soils 8 and 10 (southern mountains. 118 SOIL 8 MG GLY 25‘ K 25' K 300‘ K 550' SOIL IO M66LY25' KZS’ K 300' K 350’ 7 8 I0 I2 l4 I8 24 36 £3 13 4 5 6 l I l L-u _ _ h.- _ N _ _ a- 4 119 Plate 4.--X-ray diffraction patterns of the clay fractions from soils 12 and 14 (southern mountains). 925$ L I—u I. 120 SOIL I2 MG'GLY 23' K 25' K 300' K 350' SOIL I4 MG-GLY 23° K 25’ K 300' K 350‘ ANGSTROMS 6 8 IO I2 I4 I8 24 36 I I I 41 I T I I 20 I6 I2 8 4 DEGREES 26 “'0 I-‘I _ )— PI r 121 Plate 5.--X-ray diffraction patterns of the clay fractions from soils 16 and 17 (central mountains). 5 122 SOIL I6 M6 GLY 25’ K 25' K 300' K 530' SOIL I7 MO-GLY 23' K 25' K 300' K 550' 8 2 4 I8 24 36 5 6 7 8 I0 I I 1 31.5 “I J I I I I I L J I I 1 T I j I I 24 20 l6 l2 0 4 123 Plate 6.-—X-ray diffraction patterns of the clay fractions from soils 21 (central mountains) and 24 (central jungle). r—U 1J24 MG-GLY 25° K300° KSSO' MG'BLY 25’ K 300‘ K 350‘ uncannous 5 c I I I I 24 20 " accuses 20 fi‘ --0 SOIL 2| SCNL 24» 125 Plate 7.—-X-ray diffraction patterns of the clay fractions from soils 26 (central mountain) and 27 (coast). SOIL 26 J MG'GLY 25. j I A 25. SOIL 27 MG-GLY 25° fl ‘N ”K 25. Ii” mesrnous 3 33 3.3 4 s 7 8 I0 I2 I4 Is 24 3s 1 I I I I I g I I I I I I L I ji I r I I I 23 24 2 I2 8 4 I6 DEGREES 26 127 Plate 8.--X-ray diffraction patterns of the clay fractions from soils 28 (coast) and 29 (central jungle). I—O 128 MG-GLY 25' K 300' K 350‘ MG-GLY 25' K 25' K 550' I I I 20 l6 l2 8 DECKEES 2D SOIL 28 SOIL 29 129 Plate 9.--X-ray diffraction patterns of the clay fractions from soils 31 and 34 (Michigan). 130 SOIL 3| us our 23° K 25° K 300° K 330° SOIL 34 MG-GLY 25' IIIIIJIIIIIIII IIIIIIIIIIIIIIIIIIIII 9 6 1 3 o 3 9 2 1 3 IIIIIIIIIIIII