Clays and Clay Minerals, Vol. 42. No. 2, 180-186, 1994.

hk-ORDERING IN ALUMINOUS NONTRONITE A N D SAPONITE SYNTHESIZED NEAR 90~ EFFECTS OF SYNTHESIS

CONDITIONS ON NONTRONITE COMPOSITION A N D ORDERING

V. C. FARMER, 1 W. J. MCHARDY, 2 F. ELSASS, 2 AND M. ROBERT 2

The Macaulay Land Use Research Institute Craigiebuckler, Aberdeen AB9 2Q J, United Kingdom

2 Station de Science du Sol INRA Route de Saint-Cyr, 78026 Versailles, France

Abstract--In studies on the fate of aluminium in the environment, nontronite and saponite have been obtained by synthesis in reducing alkaline conditions close to those prevailing in poorly drained soils developed from limestones. The two minerals obtained have different structures and organizations cor- responding to two different growth and/or maturation mechanisms. High-resolution transmission electron microscopy of ultrathin sections of a synthetic aluminous nontronite embedded in resin showed the presence of crystallites consisting of two to ten co-terminating parallel layers, indicating synchronous growth. Electron diffraction showed that the individual crystallites had hk-ordering, i.e., orientation of layers with respect to the six-fold pseudosymmetry of the unit cell. Deposits of a synthetic saponite included hk-ordered crystallites and crinkled films with turbostratic stacking. The two saponite phases had slightly different b dimensions. Lattice fringe images of sections of saponite embedded in resin showed a high angular disorientation of the layers in the stacking direction, suggesting multiple nucleation and growth of individual layers, subsequently aggregated with imperfect parallelism.

Exploration of the synthetic conditions of the aluminous nontronite indicated that calcium was essential for an hk-ordered product. Syntheses using potassium or sodium hydroxides and carbonates for pH control gave poorly organized nontronites. Hydrazine was not essential for nontronite formation, but better crystallized products--judging by their IR spectra--were obtained in its presence by maintaining reducing conditions in the early stages of synthesis. Attempts to prepare ferruginous beidellites under similar conditions to those in which aluminous nontronites formed were unsuccessful.

Key Words--Beidellite, Electron diffraction, High-resolution electron microscopy, hk-order, Lattice im- aging, Layer stacking, Nontronite, Nontronite synthesis, Saponite.

I N T R O D U C T I O N

The fate o f a l u m i n i u m liberated by weather ing is o f impor tance for both agriculture and fresh-water fish- eries, since easily solubil ized amorphous precipitates readily absorb or react with phosphate and can be a source o f toxic levels o f a l u m i n i u m in soils and surface waters (Paterson et al., 1991). A l u m i n i u m liberated by weather ing f rom pr imary and secondary minera ls is finally incorpora ted into var ious hydrated crystall ine species, usually o f low solubility, whose exact nature depends on the chemical env i ronment . Often, how- ever, more react ive metas table amorphous and para- crystall ine species, such as a l lophane and imogoli te , are the i m m e d i a t e weather ing products, and these can persist for some thousands o f years (Farmer and Rus- sell, 1990; Wada, 1989). React ion between a l u m i n i u m and silicic acid in the laboratory generally yields only a l lophanes at r o o m tempera ture , or paracrystal l ine species at t empera tures o f 90~176 ( imogoli te at pH < 5.0 and a pseudohal loysi te at pH 8-9 under suitable condit ions: see F a r m e r et al. 1991a, and references therein).

Recently, Fa rmer et al. (1991b) showed that alu- rn inium can be incorpora ted into poorly crystall ine

layer silicate species at 23~ when precipi ta ted and incubated in a calcareous e n v i r o n m e n t in the presence o f selected concentra t ions o f Fe(II), Si(OH)4, and hy- drazine. Two of these systems gave wel l -developed alu- ruinous nontroni tes when incubated at 89~ for 8 weeks. The presence o f 1 m M Mg in these solutions did not affect the products formed at 23~ but gave rise at 89~ to a wen-deve loped saponi te in the absence o f Fe(II) and to a layer silicate (possibly an interstratified chlori te-saponite) with unusual X-ray diffraction char- acteristics when Fe(II) was present.

In this paper, we have used t ransmiss ion electron microscopy (TEM), high-resolut ion t ransmiss ion elec- tron microscopy (HRTEM) , and selected area electron diffraction (SAED) to de te rmine the organizat ion o f the growing material , i.e., the lateral extension o f sil- icate layers and crystallites as well as the n u m b e r and regularity o f stacking of the layers within the crystallites present in the better-crystal l ized products obta ined at 89~ Because o f its six-fold pseudo-symmet ry , a single sheet o f a smect i te gives a weak six-fold symmet r i c spot diffraction pattern. In turbostrat ic stacking, suc- cessive layers are super imposed in arbitrary orienta- t ion to produce ring diffraction patterns f rom thick

Copyright �9 1994, The Clay Minerals Society 180

Vol. 42, No. 2, 1994 hk-ordered synthetic nontronite and saponite 181

deposits, or several weak non-aligned six-fold patterns of equal intensity from thin deposits. In hk-ordered stacking, successive layers are superimpsoed with their a and b axes aligned, or rotated relative to each other by multiples of 60 ~ so that the diffraction patterns of the individual layers are superimposed to give a single strong six-fold symmetric spot pattern. Such hk-or- dered structures consist ofcrystallites with well-defined boundaries, whereas turbostratic material appears filmy with irregular boundaries.

The materials examined are identified by the num- bers 6220 (aluminous nontronite), 6204 (saponite), and 6224 (possibly a chlorite-saponite) used by Farmer et al. (199 lb), where the digits indicate the approximate relative initial molar concentrations of Si, A1, Fe, and Mg, in that order. The digit 4 corresponds to a con- centration of 1 mM.

In addition, we have re-examined the conditions of synthesis of the aluminous nontronite to determine the possible range of compositions and whether the pres- ence of hydrazine and calcium were essential for its formation.

EXPERIMENTAL METHODS

Syntheses

The products 6220, 6204, and 6224 were those whose syntheses, infrared (IR) absorption spectra, and X-ray diffraction (XRD) characteristics were described by Farmer et al. (1991b). Their statement that "no pre- cautions were taken to exclude air after initial pH ad- justment" is perhaps misleading. Although not work- ing under anaerobic condi t ions , exposure of the synthetic systems to air was very limited.

To determine the roles of the exchangeable cation and hydrazine in the formation of the aluminous non- tronite 6220, syntheses were attempted with the same ratios of Si:Al:Fe(II) as before, but with the following variations:

1) In the presence of zero, 1 and 4 mM hydrazine using saturated Ca(OH)2 solution to adjust the pH to 8.5 and incubating in the presence of solid CaCO3 for 18 days at 92 _+ I~ The procedure was close to that of Farmer et al. (1991b), who used 1 mM hy- drazine and incubated for 8 weeks.

2) In the presence of 4 mM hydrazine, using 0.1M KOH to adjust the pH to 7.0 and then with 0.05M K2CO3 to bring the pH to 8.5 before incubating at 92 _+ l~ for 18 days.

In addition, attempts were made to synthesize prod- ucts with lower Fe:A1 ratios, possibly in the ferruginous beidellite range.

The products were characterized by IR spectroscopy and X-ray diffraction. Assessments ofcrystallinity were based on the presence and degree of development of characteristic IR absorption bands and XRD maxima.

Electron microscopy and diffraction

The three best-crystallized products, 6220, 6204, and 6224 prepared by Farmer et al. ( 1991 b), were subjected to TEM and SAED as deposits on carbon support films and to HRTEM as randomly oriented clay pastes em- bedded in resin and sectioned by ultramicrotomy using a diamond knife. The preparation of such sections and their examination by high-resolution electron micros- copy have been described by Srod6n et al. (1990). The clay paste is water saturated before substitution of hy- dration water by dehydrating agents and then by resin. This treatment allows observation of the clay organi- zation in a state which is quite comparable to the wet state. As a matter of fact, the replacement of water by solvents does not usually disturb the general arrange- ment of layers and particles except by changing the swelling interlayer spacings, which are reduced to 1.36 nm in smectites.

The periodicity of lattice fringes was measured on negatives taken at 105,000 • magnification using a bin- ocular microscope with a graticule scale. The magni- fication was calibrated with thick crystals of muscovite. The estimated precision of the measured periods is _+ 5% on individual measurements.

RESULTS AND INTERPRETATION

Electron microscopy and diffraction

The aluminous nontronite (6220), saponite (6204), and possible chlorite-saponite (6224) prepared by Farmer et al. (1991 b) had been isolated by freeze-dry- ing the washed and centrifuged precipitates and were difficult to disperse into individual crystallites for ex- amination as deposits in the electron microscope. However, after ultrasonic dispersion and allowing larg- er aggregates to settle out over 48 hr, the clear super- natants gave deposits in which individual crystallites were clearly present in the 6220 and 6204 products, although still partially aggregated.

In the case of the aluminous nontronite (6220, elec- tron diffraction from the smaller particles showed only sharp diffraction spots, indicating that the individual crystallites consisted of hk-ordered regularly stacked layers. Some regions could be isolated in which one or two crystals dominated the diffraction (Figure 1), but most particles consisted of several superimposed crys- tals.

Seen in section, the crystallites were found to consist mostly of two to seven co-terminating parallel layers, indicating synchronous growth. Some particles exhib- ited up to 30 or more parallel layers, but these tended to break clown into groups of four to ten layers at the particle edges (Figure 2). It is uncertain whether these are associations of primary crystallites or single crystals that separate into smaller crystallites as they grow. The spacing between layers was mostly near 1.00 nm, in-

182 Farmer, McHardy, Elsass, and Robert Clays and Clay Minerals

Figure 2. HRTEM micrograph ofaluminous nontronite 6220, showing lattice-fringe images of layers in parallel orientation.

Figure 1. A is a TEM micrograph of aluminous nontronite 6220, deposited from dilute suspension; B is an Electron dif- fraction pattern from the region arrowed in A.

dicating that the organic components of the resin had not penetrated between the layers. A few groups of two to four layers within thicker crystals exhibited wider spacings of around 1.23 nm, but this is still too narrow to indicate the entry of resin between the layers and probably indicates incomplete dehydration of the ma- terial. The resin enters between smectite layers to give a 1.36 nm spacing, but does not enter between ver- miculite layers (Srod6n el al., 1990). X-ray diffraction also indicated that the aluminous nontronite has ex- pansion properties similar to vermiculite, as Ca-satu- rated material gave a spacing of only 1.51 nm with glycerol (Farmer et al., 1991b). A somewhat larger spacing, 1.66 nm, was given with ethylene glycol.

Unlike the aluminous nontronite, the saponite par- ticles gave both diffraction spots indicative of hk-or- dered crystallites and continuous rings indicative of turbostratic stacking of individual layers (Figure 3). The hk-ordered crystallites consisted of platy material with well-defined edges, whereas the turbostratic ma- terial consisted of crumpled films. The diffraction spots did not coincide with the continuous rings, indicating that the hk-ordered crystallites had slightly smaller b spacings (0.904 nm compared with 0.916 nm for the turbostratic material). It is uncertain whether this is

due to their different stacking or to a difference in com- position between the two phases.

Examination by HRTEM of sections exhibiting lay- ers aligned parallel to the beam showed that material with irregular stacking is the dominant phase. This phase displays an unambiguous high level of stacking faults (Figure 4). The stacks contain layers of uneven planar extension and the 3-D organization is similar to the quasicrystals that form by the coalescence of monolayers ofsmectite when they are in a Ca-saturated state (Tessier and Pedro, 1987). Individual crystallites with co-terminating layers were not seen; these appear to be concentrated in the finer particles examined in deposits giving the SAED pattern shown in Figure 3.

Figure 3. Electron diffraction pattern from saponite 6204, deposited from dilute suspension.

Vol. 42, No. 2, 1994 hk-ordered synthetic nontronite and saponite 183

Figure 4. HRTEM micrograph of saponite 6204, showing lattice fringe images of layers in parallel orientation.

Sections of preparation 6224 incorporting Mg, Fe, and AI showed a complex structure with groups of strongly scattering layers separated by weakly scatter- ing layers. Pairs of strongly scattering layers separated by single faint layers was a common pattern (Figure 5). As all layers appeared to be separated by 1.0 nm, the layer organization does not support a mixed-layer chlorite-saponite. On the other hand, we have not been able to reconcile the layer image with the observed X-ray diffraction traces (Farmer et al., 199 l b), which showed a strong diffraction peak at 0.91 nm when air dried, which shifted to 0.84 nm on glycerol treatment.

Conditions of synthesis o f aluminous nontronites

The original synthesis of the aluminous nontronite, 6220, was obtained in systems neutralized with Ca(OH)2 and buffered with CaCO3, in the presence of 1 mM hydrazine. As it was uncertain whether calcium ions, hydrazine, or hydrazinium ions were necessary in the synthesis, we have carried out exploratory syntheses omitting or increasing the hydrazine content, and also substituting K for Ca in the pH adjustment. The sys- tems were kept at room temperature for 24 hr after adjusting the pH and before incubating at 92 _+ I~

In the calcium system, aluminous nontronite was obtained whether hydrazine was present or not. But the product without hydrazine was less well crystal- lized, judging by its IR spectrum, and incorporated some free iron oxides, giving an orange product. Ex- clusion of oxygen during pH adjustment was essential for nontronite formation in the absence of hydrazine. Those with hydrazine gave yellow products, indicating that all or nearly all the iron was incorporated into the layer silicate. Judging by the persistence of blue-green colors, the presence of hydrazine maintained iron in the Fe(II) condition for longer. In the absence of hy-

Figure 5. HRTEM micrograph of synthetic layer silicate 6224, showing lattice-fringe images of a particle in parallel orien- tation.

drazine, little Fe(II) remained after 1 day at room tem- perature. With 1 mM hydrazine, oxidation appeared complete after 1 day at 92~ With 4 mM hydrazine, blue-green colors persisted for over 7 days at 92~ However, the final products after 18 days with 1 and 4 mM hydrazine were very similar. A preparation con- taining 3 mM hydrazine in which iron was maintained largely in the reduced condition by digestion in a glass bottle for four weeks at 70~ appeared from its infrared spectrum to be largely amorphous, while a similar preparation digested in a high-density polyethylene bottle, where an oxidized color was fully developed in 3 weeks, gave a fairly well-crystallized aluminous non- tronite without associated amorphous products. It ap- pears that the main crystallization process occurs after or during oxidation of iron.

When KOH/K2CO3 was used for pH adjustment, a highly dispersed reddish product was obtained, re- quiring addition of KC1 (0.01M) to precipitate it. Its IR spectrum (Figure 6A) indicated that it was a poorly crystalline nontronite, as it exhibited a typical non- tronite OH stretching band at 3560 cm -~, although the OH bending bands of nontronite at 875 cm -~ (A1- FeOH) and 817 cm- ~ (Fe2OH) were poorly developed. These bending bands were clearly resolved in the spec- trum of the product from the CaCO3 system (Figure 6B). XRD also reflected the poor crystallinity of the K2CO3 product. Its trace (Figure 7C) showed weak hk diffraction bands, but the basal reflection near 1.25 nm was not resolved from the background. In contrast, the X R D trace of the CaCO3 product (Figure 7B) showed most of the features exhibited by a natural nontronite (Figure 7A). The near absence of a basal reflection from the K2CO3 product could be due to the presence of irregular iron oxide coatings on the basal surfaces. The use of NaOH/Na2CO3 in pH adjustment gave a product with an IR spectrum and XRD pattern similar to the potassium product.

It appears, therefore, that calcium ions may be es-

184 Farmer, McHardy, Elsass, and Robert Clays and Clay Minerals

I l I I

I I I I I I I I t I I l 40o0 ~ 2000 1600 1200 ~ 4~

WQvenumbef/cm-1

Figure 6. IR spectra of nontronites synthesized at 92~ for 18 days: A) in a K2CO 3 buffered system, and B) in a CaCO3 buffered system. (1.0 mg samples in 13 mm KBr disks, dried at 150"C).

I

1 . 5 1

A_j 1 .43

e, l

. j

sentiat to the formation of a well-developed nontronite structure, displaying hk ordering together with basal reflections. Planarity and close contact of surfaces and compact aggregation of layers are well displayed by HRTEM (Figure 2). The presence of hydrazine gives improved crystallinity by delaying the oxidation of Fe(II).

Range of nontronite compositions

On the basis of chemical analyses by ICP of a so- lution prepared from a lithium metaborate fusion, the aluminous nontronite 6220 had a structural formula:

(Si6.66Al~.34)(A1L~gFe2.sl)O2o(OH)4(Cao.64Mgo.o3)

This confirms the high layer charge suggested by its expansion properties with glycerol (Farmer et aL, 199 lb). Its infrared spectrum showed bands due to OH bending frequencies at 817 cm -1 (Fe23+OH) and 875 cm -1 (Fe 3+ A1OH), but no band near 930 cm -~ as- signable to A12OH groupings. The OH stretching fre- quency lay at 3550-3560 cm- l .

We have attempted to increase the proportion of a luminium in the products, keeping the starting con- centration of silicic acid constant near 1.5 mM, but varying the concentrations of A1 and Fe from the ap- proximately 0.5 mM concentrations used for prepa- ration 6220.

With hydrazine present, poorly crystallized non- tronites, as judged by their IR spectra, were obtained for concentration pairs of (AI 0.6 mM, Fe 0.4 mM) and (A1 0.5 mM, Fe 0.2 mM) in which the OH stretch- ing frequency lay at 3550 cm -1, indicating no signifi- cant shift in composition of the octahedral layer from that in the (A1 0.5 mM, Fe 0.5 raM) system. Products from the concentration pairs (A1 0.7 raM, Fe 0.3 raM) and (A1 0.8 mM, Fe 0.2 raM) were amorphous and

/ t,- 1.25 f

I I I I I I I I i 9 0 7 0 5 0 3 0 10

D e g r e e s 2 0 C o K a

Figure 7. XRD traces of randomly oriented air-dried non- tronite powders: A) natural (Washington, CMS Source Clay Mineral), B) synthesized in the CaCO3 system, C) synthesized in the K2CO3 system. Peak positions in nm.

showed no OH stretching frequency assignable to non- tronite.

In the absence of hydrazine, higher OH stretching frequencies were obtained in the range 3575 cm -1 to 3590 cm-1), indicating higher octahedral AI to Fe(III) ratios. Well-developed products with sharp OH bend- ing absorption bands were obtained for the concentra- tion pairs (A1 0.5 mM, Fe 0.5 mM) and (A1 0.5 mM, Fe 0.4 mM), but these still showed no indication of an AI2OH bending frequency in the 900-940 cm- 1 region. Systems with the concentration pairs (AI 0.6 mM, Fe 0.4 raM) and (A1 0.5 mM, Fe 0.3 mM) required longer times of digestion with two additions of hydrazine to give good products, but these also showed no A12OH bending frequency.

From these results, it appears that ferruginous bei- dellite compositions may not be achieved under the synthetic conditions explored.

DISCUSSION

Natural nontronite and saponite can take the form of laths which produce single crystal diffraction pat-

Vol. 42, No. 2, 1994 hk-ordered synthetic nontronite and saponite 185

Figure 8. TEM micrograph of laths formed in the 6220 sys- tem after 8 weeks at 40~

terns (Nadeau and Tait, 1987). This paper presents the first evidence that such hk-ordered crystals can be syn- thesized in the laboratory at temperatures below 100~ In the case of the nontronite, it appears that calcium, rather than potassium or sodium, is essential to non- tronite formation with hk-ordering. Its role in saponite formation has not been examined. Structural infor- mation must be transmitted through the hydrated cal- cium ion from one sheet to the next during crystalli- zation to allow hk-ordering to occur. Presumably, this structural information cannot be transmitted by hy- drated sodium or potassium ions.

The presence of hydrazine promotes the crystalli- zation of nontronite by maintaining reducing condi- tions during the early stages of the synthesis. The tech- niques used did not prevent the final oxidation of Fe(II), as the high-density polyethylene bottles used were slightly porous to oxygen. However, polypropylene bottles appeared to be more porous and were found to be unsuitable for the synthesis.

It seems plausible that the presence of Fe(II) allows the formation ofa trioctahedral layer silicate with Fe(II) and A1 in the octahedral sheets. On subsequent oxi- dation of Fe(II) to Fe(III), nontronite would form by ejection of the excess octahedral cations. The ejected Fe(III) may in turn be reduced by hydrazine in the interlayer space and used to extend the layer structure. Nevertheless, the main crystallization process seems to proceed during or after oxidation of Fe(II).

It is consistent with an intermediate trioctahedral structure in nontronite synthesis that none of the prod- ucts obtained appear to contain A12OH groupings, since formation of the necessary precursors, Fe(II)A12OH groupings, would he inhibited by the high local positive charge. Crystallization ofa luminous nontronites is in- hibited with increasing starting ratios of Al to Fe. Poor- ly ordered nontronite-like materials (hisingerite) were obtained even at 23"C in systems containing only Fe(II) or with starting concentrations of A1 0.5 raM, Fe 1.0

raM, but only an amorphous product for A1 0.5 mM, Fe 0.5 mM (Farmer, 1992; Farmer et al., 199 lb). This last ratio has been found to produce a few well-formed crystalline laths associated with amorphous material at 40*(2 in 8 weeks (Figure 8), but requires higher tem- peratures (70~176 to achieve total crystallization of the gel (Figures 1 and 2). A well-developed nontronite has been obtained at 90~ with the starting concentra- tions A1 0.5 raM, Fe 0.3 mM, but crystallization was slow and required repeated oxidation-reduction cycles.

The conditions for successful nontronite synthesis found in the present and an earlier investigation (Farm- er et al., 199 lb) suggest that nontronite should form readily in an alkaline calcium-rich environment sub- ject to oxidation-reduction cycles with soluble silicon in excess of about 200 #M. These conditions are likely to prevail near the weathering front in saprolites on basic and ultrabasic rocks, where nontronite is a com- mon weathering product (Velde, 1985). Calcium ap- pears to be necessary for the formation of a well-de- veloped nontronite. Glauconite or celadonite might have been expected to form in potassium-rich envi- ronments, but no evidence for such products was ob- tained when calcium was replaced by potassium in the synthetic systems.

Poorly drained calcareous soils, such as vertisols, seem also a favorable environment for nontronite neo- formation, but nontronite in soils is rare (Borchardt, 1989). Nontronite is susceptible to attack by complex- ing acids (Farmer, 1992) and so may not survive in a biologically active zone where such acids are released by roots and microorganisms. Ferruginous beidellites are a widespread component ofvertisols (Wilson, 1987), but we have been unable to synthesize such materials under the conditions used here. It may be, however, that nontronite once formed may act as a precursor of or template for more aluminous layer silicates, such as beidellite, montmorillonite, and kaolinite.

REFERENCES

Borchardt, G. (1989) Smectites: in Minerals in Soil Envi- ronments: 2nd ed., J. B. Dixon and S. B. Weed, eds., Soil Science Society of America, Madison, 675-727.

Farmer, V. C. (1992) Possible confusion between so-called ferrihydrites and hisingerites: Clay Miner. 27, 373-378.

Farmer, V. C., Krishnamurti, G. S. R., and Huang, P. M. (1991b) Synthetic-allophane and layer-silicate formation in SiO:AlzO:FeO-Fe20:MgO-H20 systems at 23~ and 89"C in a calcareous environment: Clays & Clay Minerals 39, 561-570.

Farmer, V. C., McHardy, W. J., Palmieri, F., Violante, A., and Violante, P. (199 la) Synthetic allophanes formed in calcareous environments: Nature, conditions of formation, and transformations: Soil Sci. Soc. Am. J. 55, 1162-1166.

Farmer, V. C. and Russell, J. D. (1990) Structures and gen- esis of allophanes and imogolite, and their distribution in non-volcanic soils: in Soil Colloids and their Association in Aggregates, M. F. de Boodt, M. H. B. Hayes, and A. Her- billon, eds., Plenum Press, New York, 165-178.

Nadeau, P. H. and Tait, J. M. (1987) Transmission electron

186 Farmer, McHardy, Elsass, and Robert Clays and Clay Minerals

microscopy: in A Handbook of Determinative Methods in Clay Mineralogy, M. J. Wilson, ed., Blackie, Glasgow and London, 209-247.

Paterson, E., Goodman, B. A., and Farmer, V. C. (1991) The chemistry of alurninium, iron and manganese oxides in acid soils: in Soil Acidity, B. Ulrich and M. E. Sumner, eds., Springer-Verlag, Heidelberg, 97-124.

Srod6n, J., Andreoli, C., Elsass, F., and Robert, M. (1990) Direct high-resolution transmission electron microscopic measurement of expandability of mixed-layer illite/smec- tire in bentonite rock: Clays & Clay Minerals 38, 373-379.

Tessier, D. and Pedro, G. (1987) Mineralogical character- ization of2:1 clays in soils: Importance of the clay texture: in Proceedings of the International Clay Conference, Den- ver, 1985, L. G. Schultz, H. van Olphen, and F. A. Mump- ton, eds., The Clay Minerals Society, Bloomington, Indi- ana, 78-84.

Velde, B. (1985) Clay Minerals: A Physico-chemical Er nation of Their Occurrence. Elsevier, Amsterdam, 427 pp.

Wada, K. (1989) Allophane and imogolite: in Minerals in Soil Environments: 2rid ed., J. B. Dixon and S. B. Weed, eds., Soil Science Society of America, Madison, Wisconsin, 603-638.

Wilson, M.J. (1987) Soil smectites and related interstratified minerals: Recent developments: in Proceedings of the In- ternational Clay Conference, Denver, 1985, L. G. Schultz, H. van Olphen, and F. A. Mumpton, eds., The Clay Min- erals Society, Bloomington, Indiana, 167-173.

(Received 27 ,~Iay 1993; accepted 22 October 1993," ..l/Is. 2381)