European Journal of Soil Science, December 2002, 53, 521±527

Effect of termites on clay minerals in tropical soils: fungus-growing termites as weathering agents P. J O U Q U E T a, L. M A M O U b, M. L E P A G E a & B. V E L D E b a

Laboratoire d'Ecologie, UMR 7625, Ecole Normale SupeÂrieure, 46 rue d'Ulm, 75230 Paris 05, and bLaboratoire de GeÂologie, UMR 8538, Ecole Normale SupeÂrieure, 24 rue Lhomond, 75230 Paris 05, France

Summary Termites of the subfamily Macrotermitinae play an important role in tropical ecosystems: they modify the soil's physical properties and thereby make food available for other organisms. Clay is important in the architecture of Macrotermitinae termite nests, and it has been postulated that termites could modify the mineralogical properties of some clays. We have tested this hypothesis of clay transformation by termites in the laboratory under controlled conditions, using Odontotermes nr. pauperans termite species, one of the main fungus-growing species at Lamto Research Station (CoÃte d'Ivoire). Soil handled by termites in nest building was saturated with SrCl2, glycol or KCl and afterwards heated at 250 C for X-ray diffraction analyses. Termite handling led to an increase in the expandable layers of the component clay minerals. Heating and saturation by potassium of modified clays did not close the newly formed expandable clay layers. However, differences occurred between parts of the constructions built by termites, and the clays can be ranked according to their degree of alteration in the following order: unhandled soils < galleries < chamber walls. Consequently, termites can be seen as weathering agents of clay minerals, as previously shown for micro-organisms and plants.

Introduction Characteristics that affect the transformation of clay are considered to be mainly temperature, degree of humidity and the chemical environment. Soil organisms are considered to have little or only an indirect effect on the transformation. When organisms die, their decomposition produces acids which can affect the mineralogical properties of the clay by chemical action (Robert & Berthelin, 1986). However, micro-organisms and higher plants secrete organic compounds such as oxalic and citric acids which are strong complexing molecules and which could directly alter clay minerals by releasing protons or by selectively taking up interlayer potassium (Hinsinger, 1990). Some soil organisms could thus be considered as `weathering agents' which could have direct effects on pedogenesis and clay transformations. In fact, micro-organisms (bacteria and fungi) (Landeweert et al., 2001) and macro-organisms (plants) (Hinsinger, 1990; Hinsinger et al., 1992, 1993; Augusto et al., 2001) can modify the properties of the silicate clays in a weakly reversible or irreversible way. The modifications proposed involve an opening of mica or illite clays by ionic exchanges between interlayer potassium ions and hydrated cations, combined or not with more or less complete dissolution of the mica-clay framework. Biochemical and biological alterations Correspondence: P. Jouquet. E-mail: [email protected] Received 18 October 2001; revised version accepted 27 June 2002 # 2002 Blackwell Science Ltd

of clays are thought to be dominant in temperate ecosystems, but Duchaufour (1983) suggested that they are of little significance in tropical ecosystems where weathering is driven by geochemical processes (abiotic weathering). Among the soil macrofauna, fungus-growing termites (Isoptera, subfamily Macrotermitinae) play a primary role in tropical soils. Through their actions, they deeply modify their immediate environment by increasing the content of fine particles and organic matter of the soil there and consequently stimulating microbial activity. Conspicuous differences in the clay/sand ratio have been observed in the nests of fungusgrowing termites. The proportion of clay is always more than in the bulk soil and often maximal in the royal cell and minimal in the outer wall. Boyer (1982) and Mahaney et al. (1999) found that clay minerals of Macrotermitinae termite nests are different from those of the surface soil, although Lee & Wood (1971) observed no termite influence on clay mineralogy. Leprun & Roy-NoeÈl (1976) showed that termites are very sensitive to the type of clay; they found relations between soil clay mineralogy and the presence of some termite species. Boyer (1982) and Mahaney et al. (1999) speculated about the possibility for termites, either directly or indirectly, to modify clay mineralogy. However, this hypothesis is yet to be confirmed, because one major difficulty was to determine precisely the origin of the soil selected and modified by termites for building their nests.

521

522 P. Jouquet et al. We aimed to verify the hypothesis proposed by Boyer (1982). We manipulated worker termites from the Macrotermitinae termite subfamily in the laboratory under controlled conditions. We collected the soil handled by the termites and analysed its mineralogical properties by X-ray diffraction. The results are reported below.

Materials and methods Termite model and study site Termites were collected at the Lamto Research Station in CoÃte d'Ivoire (West Africa, 6 130 N, 5 020 W) at the edge of the rain forest (Menaut & CeÂsar, 1979), in the Guinean bioclimatic zone (rainfall  1200 mm year 1). The termite studied is Odontotermes nr. pauperans (Silvestri), one of the dominant Macrotermitinae species in the Lamto savanna (Josens, 1971). It has an aggregated distribution, and termitaria are a conspicuous component of the ecosystem (KonateÂ, 1998). Odontotermes nr. pauperans cultivates an exosymbiosis fungus in interconnected chambers of fungus-comb, and it also constructs covered runways (sheetings or galleries) on the soil surface so that it can collect plant materials.

Experiments in the laboratory Manipulations were done in a rearing room at the Lamto station (Figure 1): 125 worker termites were put in nine boxes (17.5 cm long, 11.5 cm wide and 6.5 cm high). The soil utilized was sampled at 15±20 cm depth. It was sieved to pass 2 mm and spread to cover the bottom of the boxes to a depth of about 3 mm. No particle selection is made by grain size (texture) by Odontotermes nr. pauperans termites for building

their constructions with this kind of soil (Jouquet et al., 2002). About 2 g of a fungus-comb was placed in the middle of each box (see Figure 1). Experiments lasted 20 days in order to need only a few grams of soil material. At the end of the period, the soil sheeting covering the fungus (the fungus-comb wall) was sampled in each box (n ˆ 9). Termites did not build vertical foraging galleries in each box, and consequently gallery soils were collected from only five boxes (n ˆ 5) (see Figure 1).

Measurements Clay mineralogy was determined by X-ray diffraction (XRD) with a Philips step scanning system with Cu-K radiation. Samples were scanned at a speed of 2 (2) minute 1 in the range of 2±13 . Samples were first prepared routinely by saturation with Sr and X-rayed in the air-dried and glycoltreated state. Samples were also saturated with K and run in the air-dried and glycol-saturated state. The same procedure was performed on K-saturated samples heated to 250 C. Spectra were decomposed using the DECOMPRX program of Lanson (1990) for a range of 4±10 2. Decomposition was done using Gaussian shapes for smectite, illite, smectitic or illitic interstratified illite±smectite clays, and Lorentzian shapes for mica clays. We determined the relative abundance of clay types after estimating the peak area of the different soils derived from peak decomposition methods (Lanson & Besson, 1992). This method is only relative and should give changes in the abundance of clay species present for similar materials (Lanson & Velde, 1992; Velde, 1995).

Statistical analyses We analysed the abundance of clay types through analysis of variance (ANOVA) with soil type (control, gallery and wall) and treatment (Sr, glycol, K saturation and heating) as factors. All tests were performed at the 0.05 significance level.

Chamber wall soil Gallery soil

Results Clay mineral properties of control soil Control soil

Figure 1 Termites were put in boxes with soil from Lamto savanna ecosystem (15±20 cm deep) and a piece of fungus-comb. Shortly after, the termites covered the fungus (chamber wall soil) with control soil and built galleries (gallery soil) for foraging and exploring the outside environment. At the end of the manipulation, both the soil covering the fungus-comb and soil from galleries was collected and their clay mineralogical properties determined.

Sr saturation, air-dried. The decomposition of the X-ray spectra of control soil in SrCl2 saturation air-dried state is shown in Figure 2. Kaolinite is present in major quantities (85%) in all samples. However, this mineral does not appear to be affected by termite activity. Therefore we report our findings for mica, illite and mixed layer illite±smectite (I±S) because the proportion of these minerals does change. Three peaks can be observed at different times in the 10±20-AÊ region: 10.2, 10.8 and 12.5 AÊ (Figure 3a). These peaks correspond to mica + illite, illitic interstratified illite±smectite (I±Sillite) and smectitic interstratified illite±smectite clays (I±Ssmectite) (Lanson, 1990;

# 2002 Blackwell Science Ltd, European Journal of Soil Science, 53, 521±527

Effect of termites on clay minerals in tropical soils 523

Kaolinite

Mica, illite

Intensity

Mixed layer illite–smectite

Control soil

Termite constructions

Figure 2 X-ray diffraction spectrum after Sr saturation (air-dried) of control or termite-handled soil from 2 to 13 AÊ, showing the proportion of kaolinite. Peaks are mixedlayer illite±smectite (I±S), mica + illite, and kaolinite.

2

13 Position, 2θ /°

(a) I–SI 10.8 Å

M 10.2 Å

(b) I–SI 10 Å

Intensity

I–SS 12.5 Å

I–SI 12 Å

5

10

5

10

(d)

(c)

10.2 Å

Figure 3 Decomposed spectra of control soil after: (a) Sr saturation (air-dried); (b) glycol saturation; (c) K saturation (air-dried); (d) heating at 250 C. XRD spectra are detailed from 5 to 10 AÊ. M, mica±illite; I±SI, illitic mixed layer; I±SS, smectitic mixed layer.

M < 10 Å

10.2 Å

10.8 Å

5

10.8 Å

5 10 Position, 2θ /°

Lanson & Besson, 1992). The proportions of mica + illite, I±Sillite and I±Ssmectite are, respectively, 28 (standard error, SE ˆ 5), 57 (SE ˆ 4) and 15 (SE ˆ 4) % in the control soil.

10

Sr saturation, glycol treatment. Results obtained after glycol saturation are presented in Figure 3(b). Glycol treatment of a disordered smectitic I±S displaces the 12-AÊ peak to 17 AÊ. If the

# 2002 Blackwell Science Ltd, European Journal of Soil Science, 53, 521±527

524 P. Jouquet et al. smectite content is small, near 50% (I±Sillite) the peak becomes very weak in a high background and is difficult to identify. Glycol treatment gives a peak at about 12 AÊ and a new wide 10-AÊ peak when about 30% smectite is present in an ordered I±S mineral (Moore & Reynolds, 1997). This appears to be the case in our control soil samples. The peak area proportions at < 10, 10 and 12 AÊ are, respectively, 2 (SE ˆ 1), 74 (SE ˆ 5) and 24 (SE ˆ 2) %.

K saturation and heating treatment. Saturation by potassium (Figure 3c) and subsequent heating at 250 C (Figure 3d) give similar spectra with two peaks in 10.8- and 10.2-AÊ positions. The peak proportions are, respectively, 22 (SE ˆ 2) and 78 (SE ˆ 1) %. Potassium collapses the expandable layers of the clay to yield illite±mica layers, indicating that most smectite has a large charge (Howard, 1981; Moore & Reynolds, 1997). Clay mineral properties of termite buildings Sr saturation, air-dried. Decomposed spectra of termite structures (chamber walls and galleries) after saturation with Chamber

I–SI 10.8 Å

I–SS 12.5 Å I–SSS 15 Å

Sr saturation, glycol treatment. Although we found no difference between chamber walls and galleries when we saturated the soil with SrCl2 and dried them, treatments with glycol, potassium and heating distinguished the two types of constructions. Examples of Sr-decomposed spectra after glycol treatment are presented in Figure 4(c,d), and the peak area proportions are shown in Figure 5(b). The peak area proportions at both 10 and < 10 AÊ of chamber walls and galleries are not significantly different (P > 0.05). However, the position of

Galleries

Air-dried

(a)

SrCl2 in the air-dried state are shown in Figure 4(a,b), and the proportions of each peak area appear in Figure 5(a). We found no significant difference between chamber wall and gallery materials (P > 0.05). In addition to the decrease in the peak proportions at 12.5, 10.8 and 10.2 AÊ compared with the control soil, a new peak at about 15 AÊ occurs. It corresponds to a smectite content I±S clay (I±SSS) greater than 50%. The values for the 10.2, 10.8, 12.5 and 15 AÊ positions are, respectively, 23 (SE ˆ 8), 47 (SE ˆ 14), 20 (SE ˆ 8) and 10 (SE ˆ 5) %.

Air-dried

(b) I–SS 12.5 Å

M 10.2 Å

I–SI 10.8 Å

M 10.2 Å

I–SSS 15 Å

5

10

5

10

Glycol

Glycol (d)

(c) < 10 Å

10 Å

Intensity

10 Å 17 Å

< 10 Å

11.5 Å

12 Å

4.5

10

5

10

K-saturated and air-dried

K-saturated and air-dried

(e)

(f) 10.8 Å

10.2 Å

10.8 Å

10.2 Å

15 Å 12.5 Å 12.5 Å

5

5 10 Position, 2θ /°

# 2002 Blackwell Science Ltd, European Journal of Soil Science, 53, 521±527

10

Figure 4 Decomposed spectra of termite construction: (a) Sr saturation (air-dried) of chamber wall clays; (b) Sr saturation of gallery clays; (c) glycol saturation of chamber wall clays; (d) glycol saturation of gallery clays; (e) K saturation of chamber wall clays; (f) K saturation in the air-dried state. M, mica±illite; I±SI, illitic mixed layer; I±SS, smectitic mixed layer; I±SSS, high smectitic mixed layer.

Effect of termites on clay minerals in tropical soils 525

Figure 5 Peak area proportions for the control clays (in black), for the gallery clays (in white) and for chamber wall clays (in grey): (a) Sr saturation (air-dried); (b) glycol saturation; (c) K saturation (air-dried); (d) after heating at 250 C. Peak positions are given as indicative data and are only approximate, because of the variation between the samples.

the I±Ssmectite is at 12 AÊ for soil from the chamber walls and at 11.5 AÊ for that from galleries. As in the case for the control soil, the peak at 10.8 AÊ disappears after treatment with glycol. A peak at 17 AÊ is present only in the soil of chamber walls. Proportions of peak areas for chamber walls are 41 (SE ˆ 14), 17 (SE ˆ 9), 44 (SE ˆ 13) and 2 (SE ˆ 4) %, respectively, in the following positions: < 10, 10, 12 and 17 AÊ. For the gallery material, the proportions are 41 (SE ˆ 18), 11 (SE ˆ 7) and 48 (SE ˆ 18) %, respectively, for the peaks at < 10, 10 and 11.5 AÊ. These results indicate that termites increased the proportion of I±S clay of large or small smectite content and produced a new 15-AÊ peak in the air-dried state and the appearance of a 17-AÊ peak in chamber material, indicating the presence of a new clay, with a large smectite content.

K saturation and heating treatment. Saturation with potassium (Figure 5c) shows that the clays in the chamber wall and the

gallery differ. Examples of XRD spectra after saturation with potassium of chamber wall or gallery soil are in Figure 4(e,f). Peak area proportions of chamber wall soils are 35 (SE ˆ 13), 42 (SE ˆ 12), 20 (SE ˆ 7) and 4 (SE ˆ 6) %, respectively, at 10.2, 10.8, 12.5 and 15 AÊ. On the other hand, peak area proportions of gallery soil are 65 (SE ˆ 17), 25 (SE ˆ 16) and 9 (SE ˆ 8) %, respectively, at 10.2, 10.8 and 12.5 AÊ. Peak area proportions after heating at 250 C are shown in Figure 5(c). This treatment always leads to four peaks, at 10.2, 10.8, 12.5 and 15 AÊ, in the chamber wall soil. The peak area proportions are, respectively, 41 (SE ˆ 16), 37 (SE ˆ 5), 14 (SE ˆ 9) and 1 (SE ˆ 2) %. As for the glycol and K saturation, the XRD spectra of the gallery soil show only three peaks, at 10.2, 10.8 and 12.5 AÊ. The peak area proportions are, respectively, 67 (SE ˆ 7), 31 (SE ˆ 6) and 2 (SE ˆ 3) %. Treatments show that a new I±S mineral with small charge and large smectite content is present (peak at 12.5 AÊ) as a result of the termites' activity (Howard, 1981; Moore & Reynolds, 1997).

# 2002 Blackwell Science Ltd, European Journal of Soil Science, 53, 521±527

526 P. Jouquet et al.

Discussion Termite actions on soil mineralogical properties While most of the studies on termites' influence on soil do not distinguish between constructions and control soils, our experiments were done using a control soil which was compared with the same material treated by termites. This experiment should indicate the effect on soil clays. Jouquet et al. (2002) observed that Odontotermes nr. pauperans do not select particle sizes, nor therefore clay minerals, for their buildings (fungus-comb chamber walls and galleries) when they use soil from 15 to 20 cm depth in the Lamto savanna. Consequently, the differences we found in our analyses of the clays were not the consequence of particle selection. Soils collected in the boxes are first saturated with SrCl2, which displaces all the exchangeable ions. The unhandled soils contained illite±mica and I±S minerals (I±Sillite) that expand little. The soil from the termite workings contained illite±mica, I±Sillite and two I±Ssmectite minerals, one with smectite layers of small charge and the second with large charge. This indicates an increase in the content of expandable layers of the clays. Our results can partly explain in situ observations of Boyer (1982) and Mahaney et al. (1999). The presence of different clays in Macrotermitinae termite nests relative to the control soils can be explained by two processes: the first is commonly explained by the bringing up of fine particles with different mineralogical clays into the nest (Lee & Wood, 1971; Holt & Lepage, 2000), and the second results from modifications of clay properties by the termites. We treated the clays in our experiment by saturating them with Sr, glycol and potassium and by heating them at 250 C. The results from these treatments suggest that initially two I±S minerals were present in the control soil, both with smectite layers of large charge (collapse on K saturation). In the clays handled by the termites, a new, smectite-rich I±S (> 50% smectite) was found with a smectite layer of small charge. Also, a smaller amount of smectite I±S mineral with layers of small charge was present. Thus the termites seemed to have produced smectite layers with small charge in I±S minerals that are stable when saturated with potassium (peak at 12.5 AÊ). In termite galleries, a strong 11.5-AÊ peak appeared upon glycol treatment. This peak is not present in the soil of the chambers, but is probably present at 12 AÊ (Figure 5b). On the other hand, a new smectite-rich I±S is present in the chamber material (peak at 17 AÊ). Differences were observed between chamber walls and galleries. Smectite layers of gallery clays collapsed to a larger extent after K saturation than did layers of the chamber wall clays (Figure 5b). The gallery soil evidently contains weakly non-reversibly expanded clays (smectite with large charge), whereas the soil of the chamber walls can be considered as strongly non-reversibly expanded clays (smectite with smaller charge). Differences between these two kinds of material can probably be explained by their role in termite biology: chamber

walls are permanent structures and have an important function in maintaining humidity and protecting fungus and termites, whereas galleries constitute only temporary structures for foraging. Consequently, the interaction of termites is not quite the same in the two, but in both cases the behaviour of the clay is modified. We can rank the soil materials according to their content in expandable and small-charge clays in the order: chamber walls > galleries > control soil.

Fungus-growing termites as agents of clay alteration: hypothetical processes Termites can decrease, in a more or less irreversible way, the charge of some clay layers, allowing an adsorption of hydrated or polar ions between the layers. This process changes the average interlayer distance in the clay, and the mineral can then swell. This process is often called `vermiculitization' (Hinsinger, 1990) when the smectite charge is large. Because termite action leads to an opening of clay layers, this suggests that the exchange equilibrium between inner (interlayer) K+ and other cations (Sr2+ in our case) might have been shifted by termite action. Interlayer potassium ions are probably released from the interlayer space to the outer solution. Results obtained in our control conditions can be extrapolated to in situ observations. For example, Mahaney et al. (1999) observed an exchange of potassium for calcium in Macrotermes and Pseudacanthotermes termite nests. Vermiculitization of clays has been observed to result from the activity of other organisms such as plants (e.g. Badraoui et al., 1992; Hinsinger et al., 1992, 1993; Mengel & Uhlenbecker, 1993; Mengel & Rahmatullah, 1994; Rahmatullah & Mengel, 2000) and fungi (Leyval & Berthelin, 1991; Landeweert et al., 2001). The mineralogical alterations such as we observed should first be explained by a direct effect of saliva supplied by the termites, by an indirect stimulation of microflora with saliva, or by an incorporation of fungi within their constructions, or by a combination of these. Second, termite handling increases the contact area with soil and thus enhances exchanges between clays and substrates incorporated by termites. Further study is necessary to investigate the role of the micro-organisms or termite saliva, or both, in this alteration of the clay.

Conclusion Our experiments indicate that Odontotermes nr. pauperans termites can modify in a more or less irreversible way the mineralogical properties of silicate clays. Therefore, phenomena observed in situ by Boyer (1982), where expanding minerals were found, could be explained by our experiments. Termites could be considered to be weathering agents of clays as the normal process of weathering is to create expandable clay minerals (Dixon & Weed, 1989). Although geochemical alterations in tropical soils are generally assumed to be the main processes by which clays are changed, our study

# 2002 Blackwell Science Ltd, European Journal of Soil Science, 53, 521±527

Effect of termites on clay minerals in tropical soils 527 illustrates a possible biological alteration of clays mediated by termites.

Acknowledgements We thank Souleymane KonateÂ, Director of the Lamto Ecological Station (Abobo-Adjame University of CoÃte d'Ivoire), for all the facilities offered to us in the field. This research was supported by the PNDBE core project. We are also grateful to J.B. Kouassi for his help in the field and to Dr Hinsinger and the Editor for the documentation they supplied and for the helpful suggestions on the draft script.

References Augusto, L., Ranger, J., Turpault, M.P. & Bonnaud, P. 2001. Experimental in situ transformation of vermiculites to study the weathering impact of tree species on the soil. European Journal of Soil Science, 52, 81±92. Badraoui, M., Bloom, P.R. & Delmaki, A. 1992. Mobilization of nonexchangeable K by ryegrass in five Moroccan soils with and without mica. Plant and Soil, 140, 55±63. Boyer, P. 1982. Quelques aspects de l'action des termites du sol sur les argiles. Clay Minerals, 17, 453±462. Dixon, J.B. & Weed, S.B. 1989. Mineral in Soil Environments, 2nd edn. Book Series 1, Soil Science Society of America, Madison, WI. Duchaufour, P. 1983. PeÂdologie. Masson, Paris. Hinsinger, P. 1990. Action des racines sur la libeÂration du potassium et l'alteÂration de mineÂraux silicateÂs Ð Incidences agronomiques. TheÁse de Doctorat, Ecole Nationale SupeÂrieure Agronomique de Montpellier. Hinsinger, P., Jaillard, B. & Dufey, J.E. 1992. Rapid weathering of a trioctahedral mica by the roots of ryegrass. Soil Science Society of America Journal, 56, 977±982. Hinsinger, P., Elsass, F., Jaillard, B. & Robert, M. 1993. Root-induced irreversible transformation of a trioctahedral mica in the rhizosphere of rape. Journal of Soil Science, 44, 535±545. Holt, J.A. & Lepage, M. 2000. Termite and soil properties. In: Termites: Evolution, Sociality, Symbiosis, Ecology (eds T. Abe, D.E. Bignell & M. Higashi), pp. 389±407. Kluwer Academic Publishers, Dordrecht. Howard, J.J. 1981. Lithium and potassium saturation of illite/smectite clays from interlaminated shales and sandstones. Clays and Clay Minerals, 29, 136±142. Josens, G. 1971. Recherches eÂcologiques dans la savane de Lamto (CoÃte d'Ivoire): donneÂes preÂliminaires sur le peuplement en termites. La Terre et la Vie, 2, 255±272. Jouquet, P., Lepage, M. & Velde, B. 2002. Termite soil preferences and particle selections: strategies related to ecological requirements. Insectes Sociaux, 49, 1±7.

KonateÂ, S. 1998. Structure, dynamique et roÃle des buttes termitiques dans le fonctionnement d'une savane preÂforestieÁre (Lamto, CoÃte d'Ivoire): Le termite champignonniste Odontotermes comme ingeÂnieur de l'eÂcosysteÁme. TheÁse de Doctorat, Universite de Paris VI. Landeweert, R., Hoffland, E., Finlay, R.D., Kuyper, T.W. & van Breemen, N. 2001. Linking plants to rocks: ectomycorrhizal fungi mobilize nutrients from minerals. Trends in Ecology and Evolution, 16, 248±254. Lanson, B. 1990. Mise en eÂvidence des meÂcanismes reÂactionnels des interstratifieÂs illite/smectite au cours de la diageÂneÁse. TheÁse de Doctorat, Universite de Paris VI. Lanson, B. & Besson, G. 1992. Characterization of the end of smectiteto-illite transformation: decomposition of X-ray patterns. Clays and Clays Minerals, 40, 40±52. Lanson, B. & Velde, B. 1992. Decomposition of X-ray diffraction patterns: a convenient way to describe complex I/S diagenetic evolution. Clays and Clay Minerals, 40, 629±643. Lee, K.E. & Wood, T.G. 1971. Physical and chemical effects on soils of some Australian termites, and their pedological significance. Pedobiologia, 11, 376±409. Leprun, J.C. & Roy-NoeÈl, J. 1976. MineÂralogie des argiles et reÂpartition des nids eÂpigeÂs de deux espeÁces du genre Macrotermes au SeÂneÂgal occidental (Presqu'õà le du Cap-Vert). Insectes Sociaux, 23, 535±547. Leyval, C. & Berthelin, J. 1991. Weathering of a mica by roots and rhizospheric microorganisms of pine. Soil Science Society of America Journal, 55, 1009±1016. Mahaney, W.C., Zippin, J., Milner, M.W., Sanmugadas, K., Hancock, R.G.V., Aufreiter, S. et al. 1999. Chemistry, mineralogy and microbiology of termite mound soil eaten by the chimpanzees of the Mahale Mountains, Western Tanzania. Journal of Tropical Ecology, 15, 565±588. Menaut, J.C. & CeÂsar, J. 1979. Structure and primary productivity of Lamto savannas, Ivory Coast. Ecology, 60, 1197±1210. Mengel, K. & Uhlenbecker, K. 1993. Determination of available interlayer potassium and its uptake by ryegrass. Soil Science Society of America Journal, 57, 761±766. Mengel, K. & Rahmatullah 1994. Exploitation of potassium by various crop species from primary minerals in soils rich in micas. Biology and Fertility of Soils, 17, 75±79. Moore, D.M. & Reynolds, R.C. 1997. X-ray Diffraction and the Identification and Analysis of Clay Minerals. Oxford University Press, Oxford. Rahmatullah & Mengel, K. 2000. Potassium release from mineral structures by H+ ion resin. Geoderma, 96, 291±305. Robert, M. & Berthelin, J. 1986. Role of biological and biochemical factors in soil mineral weathering. In: Interactions of Soil Mineral with Natural Organics and Microbes (eds P.M. Huang & M. Schnitzer), pp. 453±495. Special Publication No 17, Soil Science Society of America, Madison, WI. Velde, B. 1995. Origin and Mineralogy of Clays: Clays and the Environment. Springer-Verlag, Berlin.

# 2002 Blackwell Science Ltd, European Journal of Soil Science, 53, 521±527