Printer Friendly

Brazilian latosols and their B horizon microstructure as long-term biotic constructs.


The age of the landscape and soils

In 1868, after visiting Brazil, Louis Agassiz described with great astonishment the widespread presence of a clayey solum overlying a deep saprolite, commonly separated by stone-lines, that he considered to be a `drift' deposit, supposedly a proof of glaciation at tropical latitudes (Agassiz and Agassiz 1868). Hartt (1870) also indicated the presence of this pseudo-drift. Woodworth (1912) carried out the first detailed study of these stony pavements and overlying soils, pointing to a cyclic, morphoclimatic control, involving relief inversion. Later, these materials, then recognised as soils, were called latosols, following the proposal of American pedologist Charles Kellogg in 1948 (Kellogg 1949). Soils apparently very similar to the Brazilian latosols are now classified as oxisols according to Soil Taxonomy concept, ferrosols in the Australian system, and ferralsols in the FAO scheme.

If we consider that most, if not not all, latosols, began to develop in Tertiary epochs, then [is greater than] 60% of the Brazilian territory would present a `Tertiary paleoweathering mantle' (Fig. 1)--a soil (paleosol) which displays long-term, inherited features.


In a country which has not suffered any intense orogeny since the end of the Palaeozoic, the land-surfaces of most of Brazil are thought to be very old. Surfaces dating back to the Tertiary, Cretaceous, or even older, often displaced by tectonic movements, form high-level remmants in old, stable cratonic areas in Brazil (e.g. Mabesoone et al. 1977; Schaefer and Dalrymple 1995). Hence, much of the material of present-day soils has been subjected to repeated biological activity, pedogenesis, erosion, and redeposition under varying climatic regimes over long periods of time. Few soils, therefore, are the products of pedogenetic and biotic processes seen to be operating today.

If the concept of paleosol put forward by Valentine and Dalrymple (1976) is to be applied in the case of latosols, then most, if not all, latosols would be considered Paleosols, i.e. a `soil formed in a landscape of the past'. Against this background, we discuss some evidence of long-term biotic processes responsible for the most important feature in present-day latosols: their highly stable microstructure.

Latosols: genesis, characteristics, and history

Some general observations on Brazilian latosols are listed below.

1. Latosols are deeply weathered soils, dominated by kaolinite, Fe-oxides (goethite and hematite), gibbsite, and quartz, with traces of other minerals. Their nutrient status is generally very low. They are frequently acidic and saturated with exchangeable aluminium.

2. Latosols occur irrespective of present-day climatic regimes, being found in subarid north-east Brazil or in very humid areas elsewhere in the country. In the subarid northeast, latosols are found extensively at the tops of planation remnants, and are considered to be relict late Tertiary profiles, formed during, and inherited from, periods of high humidity and stable climates (Mabesoone et al. 1977; Mabesoone and Neumann 1994).

3. Latosols display widespread present-day biological activity and reworking by ants and termites, as well as by other animals (armadillos, insects, earthworms). One of the most outstanding features of latosols and associated landscape are the `murundus' (termite mounds) which cover many seasonally waterlogged areas and even some currently well-drained uplands (Fig. 2). The microstructure of these mounds is similar to those of latosols (Correa 1989), especially where the abandoned mounds have been eroded and dismantled.


4. Latosols are usually very deep soils, reaching depths of 30 m or even more, including the Cr horizon or saprolite. A common feature of deep latosols is biological channels of unknown origin, which occur deep in the saprolite and are usually infilled with soil material from the upper horizons. Some are called pedotubules (Rezende 1980; Resende et al. 1995), similar to those described by Brewer (1964). These tubules are often filled with gibbsite, and can be exploited as bauxites. Others are mainly kaolinitic, stained with Fe-oxides. Some are considered to be traces of roots, `palaeoworms', or termite tunnels.

The recognition of the biotic effects on long-term development of tropical soils has long been reported in the literature. In a recent account on the effects of faunal and floral disturbances in the topsoil, Johnson (1990) has proposed `biomantle evolution' as a process largely responsible for the redistribution of earth materials and artifacts. Thus, biomantles are differentiated zones in the upper part of soils, produced largely by the mixing action of biota, often aided by subsidiary processes. It is particularly difficult for those who live in temperate latitudes to appreciate the importance of termites and other animals in tropical soil formation.

In this work, I examine the micropedology of selected Brazilian latosols (oxisols), against the extensive literature on latosols and termites. Connections between the pedogenesis of latosols and the development and ecology of termites are discussed. Consideration of the general role of differing functional-biological classes of termites, e.g. humus feeders, tree dwellers, soil dwellers, and soil feeders, is beyond the scope of this paper. Extensive studies on neotropical termite communities, in both cerrado and tropical rainforest, can be found elsewhere (e.g. Wilson 1992; Martius 1994; Constantino 1995).

Materials and methods

Samples from the B horizon of latosols and related soils from Brazil were described and their properties quantified. Most of these selected pedons are well known, classically recognised soils, which were studied during the VIII International Soil Classification Workshop. Complementary data can be found in EMBRAPA (1988). We examined thin sections of a wide range of contrasting latosols developed from a variety of substrates (basalt, sandstone, shales, unconsolidated sediments, and crystalline rocks). All of the samples displayed some degree of microaggregation. Thin sections were prepared for optical microscope examination at high magnification. In addition, chemical analyses of selected features were carried out on selected spots of the thin sections, using a JEOL JX-840 Scanning Micro-Analyzer. This was operated under standard conditions of 20 kV and fitted with an energy dispersive X-ray detector (EDX) in conjunction with a Link AN 1000 Analyzer, fitted with ZAF/PB/FLS and ZAF/4 software, following the recommendations of Norton et al. (1983). Results enabled the characterisation of pedofeatures at different microscopic scales (e.g. whole soil and inner microaggregate).

The assessment of particle size was based on wet sieving and ultrasonic dispersion, adapted from EMBRAPA (1997). The clay and silt fractions were separated by sedimentation, followed by siphoning the [is less than] 0.002 mm fraction (clay). Subsamples of clay were floculated by adding 1 M Mg[Cl.sub.2] (1:50). One gram of clay was subsequently digested with 20 mL [H.sub.2][SO.sub.4] (15:1) and the relative proportions of Si[O.sub.2], [Al.sub.2][O.sub.3], [Fe.sub.2][O.sub.3], MnO, and Ti[O.sub.2] were measured in extract (Lim and Jackson 1982) by atomic absorption spectrometry. The percentage of oxides was corrected to 100% by taking the weight loss value between 105 [degrees] C and 1000 [degrees] C.

The mineralogy of the whole soil and the clay fraction was determined for each horizon, using X-ray diffraction analysis (XRD). XRD was carried out using monochromatic CuK[Alpha] radiation on oriented clay samples previously Mg-saturated, treated with ethylene glycol and heated at 350 [degrees] C and 550 [degrees] C. In order to clarify some of the results, clay samples were examined after K-saturation or after treatment with dithionitecitrate-bicarbonate to remove Al[(OH).sub.n] polymers and free iron oxides, before K-saturation. In the latter treatments, samples were subsequently heated at 100 [degrees] C, 350 [degrees] C, and 550 [degrees] C overnight. All routine chemical and physical measurements, such as determining bulk density, were done following standard procedures (Klute 1986). Exchangeable Al, Ca, and Mg were extracted by shaking with 1 M KCl (1:5 soil: extractant). Exchangeable K and Na were extracted with 0.05 N HCl + 0.025 N [H.sub.2][SO.sub.4] (EMBRAPA 1997).


Micropedology of selected Brazilian soils

Analytical results are summarised in Table 1. Corresponding values of percentages of Si[O.sub.2], [Al.sub.2][O.sub.3], and [Fe.sub.2][O.sub.3] for the whole soil, and for single microaggregates, are shown in Table 2.
Table 1. Chemical, physical, and mineralogical characteristics
of selected latosols from Brazil

Reference to the parent materials is made below (EMBRAPA 1988). 1,
mafic tuffites--Mata da Corda Formation (Upper Cretaceous); 2, clayey
sediments (Plio-Pleistoscene) over Pre-Cambrian granitic rocks; 3,
clayey sediments Pre-Cambrian granitic rocks; 4, sandy-clay sediments
with mafic contribution--Bauru Group (Upper Cretaceous); 5, clayey
sediments over metabasic/metapelitic rocks--Araxa and Bambui Groups;
6, colluvium derived from shales--Corumbatai Formation (Permian);
7, clayey sediments over Pre-Cambrian gneiss; 8, colluvium derived
from calcareous shales--Corumbatai Formation (Permian); Symbols: kaol,
kaolinite, gibb, gibbsite; x, present; xx, very common; xxx, abundant;
Tr, trace

Soil Dry colour Particle size

 Clay Silt Sand


P1, Anionic Acrudox Bw2 4YR 5/6 86 10 4
P2, Typic Kandiudox Bt2 2.5YR 3/6 50 16 34
P3, Humic Hapludox Bw1 6.5YR 5/6 56 18 26
P4, Typic Acrudox Bw1 9YR 4/4 40 7 53
P5, Anionic Acrudox Bw2 10R 4/7 81 8 11
P6, Typic Kandiudox Bt2 2.5YR 4/6 78 11 11
P7, Typic Haplortox Bw2 10YR 7/6 67 16 17
P8, Typic Kandiudox Bw2 1.5YR 4/6 76 14 10

Soil Clay Chemical composition

 mineralogy [Fe.sub.2] [Al.sub.2]
 [O.sub.3] [O.sub.3]

 Kaol. Gibb. (%)

P1, Anionic Acrudox Bw2 xx xxx 17.4 28.7
P2, Typic Kandiudox Bt2 xxx xx 11.4 18.9
P3, Humic Hapludox Bw1 xxx Tr 8.6 21.1
P4, Typic Acrudox Bw1 xxx xx 8.0 19.0
P5, Anionic Acrudox Bw2 xx xxx 14.0 36.3
P6, Typic Kandiudox Bt2 xxx x 9.2 25.1
P7, Typic Haplortox Bw2 xxx Tr 14.2 25.0
P8, Typic Kandiudox Bw2 xxx x 8.9 25.2

Soil Chemical Org.C Bulk
 composition (g/kg) density
 Si[O.sub.2] (g/[cm.sub.3])

P1, Anionic Acrudox Bw2 6.4 12.3 1.03
P2, Typic Kandiudox Bt2 11.3 2.0 1.30
P3, Humic Hapludox Bw1 24.8 3.7 1.32
P4, Typic Acrudox Bw1 5.4 6.4 1.39
P5, Anionic Acrudox Bw2 16.2 6.7 0.90
P6, Typic Kandiudox Bt2 25.0 8.2 1.16
P7, Typic Haplortox Bw2 29.4 6.9 1.23
P8, Typic Kandiudox Bw2 28.6 3.8 1.25

Soil Exchangeable cations
 Ca Mg K Al H ([H.sub.2]0)


P1, Anionic Acrudox Bw2 0.1 0.1 0.01 0 2.6 5.7
P2, Typic Kandiudox Bt2 1.1 0.2 0.12 0 0.8 6.3
P3, Humic Hapludox Bw1 0.1 0.1 0.02 0 2.3 5.4
P4, Typic Acrudox Bw1 0.1 0.1 0.01 0 1.9 5.4
P5, Anionic Acrudox Bw2 0.2 0.2 0.01 0 1.4 6.0
P6, Typic Kandiudox Bt2 0.9 0.4 0.03 0.9 2.9 4.6
P7, Typic Haplortox Bw2 0.1 0.1 0.01 0 2.5 5.3
P8, Typic Kandiudox Bw2 1.7 2.3 0.1 0 1.5 6.3
Table 2. Semi-quantitative chemical analyses of thin sections,
using SEM/EDX of the whole area and the inner aggregate

Type of microstructure: P1, very fine granular; P2, composite single
grain and granular; P3, composite granular and subangular blocky; P4,
granular; P5, granular; P6, incomplete, composite granular and sub-
angular blocky; P7, composite granular and subangular blocky; P8,

Micro- Si[O.sub.2] [Al.sub.2][O.sub.3] [Fe.sub.2][O.sub.3]
 % (whole area, 100 [mm.sup.2])

P1 15.1 35.2 28.4
P2 19.2 21.2 18.2
P3 28.5 20.1 10.2
P4 13.2 34.6 22.3
P5 16.5 43.2 17.5
P6 n.d. n.d. n.d.
P7 31.2 38.4 20.3
P8 29.4 27.2 12.5

Micro- Si[O.sub.2] [Al.sub.2][O.sub.3] [Fe.sub.2][O.sub.3]
 % (inner microaggreate, 1 [mm.sup.2])

P1 3.1 51.2 29.3
P2 5.5 35.5 25.2
P3 16.5 19.2 8.2
P4 6.5 41.5 29.5
P5 8.2 48.6 23.7
P6 n.d. n.d. n.d.
P7 30.4 32.1 26.8
P8 28.2 31.5 15.1

n.d., not determined.

Despite the great variability in parent materials, these soils display an essentially uniform microstructure with varying degrees of roundness (Fig. 3, Table 2), thus showing little lithodependence. Furthermore, these microstructures show no apparent relationship to the different proportions of minerals and particle sizes (Table 1). However, latosols developed from mafic rocks showed a trend of better microaggregation, with well-rounded forms (P1 and P5).


The most important micropedological evidence for the biological origin of the microstructure found in these soils can be summarised as follows: (1) lithorelicts of oval pellets of 100-1000 [micro]m diameter observed in the upper zone of the saprolite; (2) the unequivocal presence of microparticles of charcoal ([is less than] 100 [micro]m) within the inner microaggregates (Fig. 3); (3) the skeleton of the microaggregate formed of quartz grains of regular sizes not more than 100 mm in diameter, whilst quartz grains of the soil matrix as a whole range between 30 and 5000 [micro]m; (4) the strong similarity between microaggregates found in the soil and those derived from termite activity on shallow saprolite. The latter can be structures either produced by the dismantled arboreal nests or eroded hypogeic and endogeic soil nests.

With intense leaching of soluble products of weathering (bases and silica), latosols in Brazil are kaolinitic and gibbsitic with varying amounts of Fe-oxides. Together with colour, these are important criteria for the classification of latosols in the Brazilian System of Soil Classification (Camargo et al. 1988). Ferreira (1988) and Schaefer (1995) have postulated that gibbsite, primarily, and Fe-oxides (goethite and hematite), secondarily; are factors that favour the long-term stability of biotic microaggregates.

Due to its isodimensional crystaline structure, gibbsite has been considered a crucial factor in assembling microparticles by acting as nuclei (Varajao 1988; Varajao et al. 1990), and thus favouring the formation/conservation of microaggregates. Gomes (1988) found gibbsite as minute inclusions in the silt/sand fractions of latosols from the Vicosa Plateau. He considered it to be derived from kaolinite pseudomorphs after biotite, through desilification promoted by acid solutions. However, gibbsite alone cannot be the sole factor in forming microaggregates. This can be illustrated by Fig. 4, which shows (a, b) a bauxitic regolith from Cataguases and (c, d) a kaolinitic regolith from Vicosa (both from Minas Gerais State). A porous, reticulate, gibbsite-rich zone in the bauxitic regolith (Fig. 4a) is overlain by well-structured red-yellow latosol (Fig. 4b). The gibbsite occurs mainly as pseudomorphs after phenocrystals of feldspar. The overlying latosol shows strong microaggregation, of microgranular type (Fig. 4b). The evidence here shows that, in spite of the high gibbsite content of the saprolite, only in the overlying B horizon is a pattern of microstructure present. By implication, the B horizon, with its distinct microaggregates, is the product of long-term biotic reworking of bauxitic saprolite.


In the kaolinitc regolith from Vicosa, kaolinite flakes occur as pseudomorphs of biotite (Fig. 4c). In the overlying red-yellow latosol occurs a pattern of coalesced microgranular structure (Fig. 4d). Channels filled with termite pellets are commonly observed in the upper part of the saprolite, indicating a per descendum process of termite action down the profile, as suggested by Nunes et al. (2000).


I investigated whether a better microaggregation in latosols (P1-P8) may be related to the presence of gibbsite, associated with Fe-oxides. Evidence from semi-quantitative SEM/ EDX data of selected spots on thin sections (P1-P8) revealed varying increase in [Al.sub.2][O.sub.3] and [Fe.sub.2][O.sub.3] contents in the inner microaggregate (Table 2) compared with whole soil matrix. However, these differences can be attributed to the presence of greater quantities of quartz in the matrix, as previously reported. Thus, a higher concentration of gibbsite within the microaggregate acting as nuclei organising the aggregate phenomenon is unlikely, but cannot be ruled out. On the other hand, soils developed from mafic rocks display a better microaggregation, in terms of roundness and individuality, compared with those derived from crystalline or sedimentary rocks. The high Si[O.sub.2] content of P7, which showed a composite granular and subangular blocky structure, suggests that better microaggregation was prevented by Si[O.sub.2]-enriched solutions, derived from the abundant quartz grains and kaolinite flakes present in this saprolite (Fig. 4c; Schaefer 1995). According to a recent detailed chemical characterisation of microaggregates in latosols (Melo Marques 2000), no evidence of chemical gradients was found from the surface to the core aggregate. The author also suggests a possible biological origin for the aggregates.


Latosols as soils created by long-term faunal activities

The suggestion that termites play a role in the formation of laterite and latosols was initially made by Erhart (1951), followed by others (e.g. Sys 1955; Cailleux 1957; Watson 1972; Yakushev 1968; Lee and Wood 1971; Boyer 1973; Lepage et al. 1974; Wood and Sands 1978). More recently, this was further suggested by Wielemaker (1984), Eschenbrenner (1986), Correa (1989), Tardy and Roquin (1992), Lobry de Bruyn and Conacher (1995), and Schaefer (1996). Nevertheless, apart from some preliminary observations by Stoops (1964), Lee and Wood (1971), Correa (1989), Miklos (1992), and Schaefer (1996), little has been published about the relationship between the microstructure of latosols and termite mounds that could be interpreted as genetically related features. Recently, Garnier-Sillam and Harry (1995), Jungerius et al. (1999), and Nunes et al. (2000) demonstrated the importance of termites to the development of granular structure in tropical soils.

The structure and stability of latosols were found to be related to the `pseudo-sand' aggregates (Ahn 1970) in Africa, South America, and elsewhere in the tropics, with varying aggregate stability over contrasting parent materials. On the other hand, as early as 1968, preliminary Portuguese surveys in Africa (Missao de Pedologia de Angola e Mocambique 1968) proposed soil classes which were entirely or partially formed by reworking promoted by termite and ant activity (for example, vermisols in Angola). Barros Machado (1982) found termite remains within microaggregates of bauxitic deposits in France, Africa, Australia, and Central America.

In summary, termite mounds of varying shapes and sizes are commonly found on the top of old plains associated with latosols and laterites in Brazil and elsewhere, as well as on the slopes of different ages. Mounds as high as 6-7 m of wood-feeding termites have been reported from Amazonia and elsewhere (Oliveira 1929; Lee and Wood 1971). Termite activity results in an upward transport of fine-textured particles (sand, silt, and clay), deposited on the topsoil, and disturbed and disrupted by erosion. The depth of uptake can reach 12 or 13 m (Tardy 1992), although recently, Tertiary regoliths in Brazil (Barros Machado 1982) and Australia (Thiry et al. 1995) were found to contain termite channels reaching depths of 30 m. Lee and Wood (1971) reported structures at a depth of 70 m, which they attributed to soil-feeding termites.

In most cases, both the surficial soil materials and the underlying horizons have similar chemical and mineralogical properties to the underlying saprolite, despite their very contrasting structures at microscopic level. The solum and the saprolite are commonly separated by a stone-line pavement, consisting of rounded pebbles of resistant rocks. Such is the case of the soils investigated in the present study.

Termite activity can also lead to a downward movement of materials, either stone/gravel fragments or clayey materials. In part, these may account for the nature of some of the stone-lines, and the high degree of homogeneity displayed by the solum. Lateral surface and subsurface movements of lateritic materials down the slopes are also possible, associated with a vertical lowering of latosols and lateritic landscapes.

The microaggregate stability of termite pellets is much higher than of pellets produced by ants (Miklos 1992). Hence, the former are likely to persist longer. Leaf-cutter ants, with nests reaching deep in the profile, may be important for some aspects of the pedogenesis of latosols, especially the high degree of uniformity, but are unlikely to contribute to the formation of microaggregates.

Termites and nutrient cycling

The role of termites in nutrient cycling has been the subject of recent studies (Nutting et al. 1987; Coventry et al. 1988; Holt 1990, 1996; Lobry de Bruyn and Conacher 1990). Termite activity results in turnover of both soil and nutrients. As much as 62 tons of subsoil per hectare translocated to the surface by mound-building termites has been reported (Lee and Wood 1971). Salik et al. (1983) reported a turnover rate of 0.78 ton/ha.year of nutrients in north Amazonia. Soil-eating termites, which make up over 50% of all termite species (Brune and Kuhl 1995), show pH values in the hindgut in the alkaline range between 11 and 12.5, the highest pH in the biological world. Such high pH combined with widespread soil ingestion can account for increasing P bioavailability by solubilisation of P-Al/Fe-oxides, possible digestion of polyphenolic compounds, and higher silica losses in tropical ecosystems.

Taking the `Gaian' view of the Earth as a superorganism (Lovelock 1979), in which the action of living organisms maintains the Earth's surface at a comfortable state for life, the ecological implications of these postulated, termite long-term effects on physical properties can now be tentatively established:

(i) The large-scale activity of termites in the tropics led to a conversion of fresh saprolite into a well-structured soil matrix, which forms the overlying materials of the solum.

(ii) The favourable physical properties usually displayed by latosols (so-called physical fertility) can be closely associated with long-term termite activity. It is postulated that without this favourable physical status, the chemically poor latosols would rapidly degrade, becoming unsuitable for maintaining life.

Feedback processes involving organisms in soils may be involved in the development of properties that favour net primary productivity (van Breemen 1993). Jones et al. (1994) called these modulating organisms `ecosystems engineers'. I consider that the favourable structure of latosols evolved in close association with biotic factors. In his account of soil porosity as faunal-related, Jenny (1980) put it in vigorous words as follows: `It takes animal workers and their lifting, digging, and tunnelling activities to improve the soil fabric by rearranging the platelets into open structures of low density that have channel, crevasses, and large pores padded with dragged-in humus molecules. The little, busy porositors do mechanical work that imparts potential energy to soil particles, enhances water percolation, and renders the substrate a more livable abode for microbes and fine roots'.

Termites: origin and implications

Some striking evidence has come to light, with broad implications for the postulated model of the biotic genesis of latosols. Emerson (1955, 1965) discussed the origin and dispersal of numerous taxa of termites, showing that in mid-Cretaceous times fairly advanced types were already present. These subsequentially became widespread in the early to mid Tertiary. Ecologically, these taxa were associated with warm climates in America, Europe, Africa, and Asia. It is likely that all present-day termite groups existed in the Palaeocene-Eocene (Grasse 1986).

In view of the origin of termites, the biotic effects of long-term termite activity on tropical soils can be assumed to have been constant since mid Tertiary times. Thus, the superimposed effects triggered in the early Tertiary by the development of termites are coincident with the development of latosols and lateritic weathering, and their associated microstructures.

Despite all these lines of evidence, Brazilian pedologists have traditionally placed greater emphasis on morphoclimatic control in the genesis of latosols and soils overlying saprolite and stone-lines (Setzer 1949; Ab'Saber 1962). I suggest, however, that in view of the assembled data, a far greater importance should be placed on biotic factors, whilst not disregarding the morphoclimatic hypothesis as a subsidiary force (Cailleux 1957, 1966).

Latosols, plate tectonics, and angiosperm evolution: convergence of data

The break-up of the western flank of the Gondwana landmass, and the opening of the south Atlantic Ocean, started in the Jurassic period, with active rifting along reactivated Precambrian faultlines (PETROBRAS 1994). As a consequence, the Jurassic landscape began to be extensively eroded, and sediments accumulated in the tectonic depressions, forming several sedimentary basins along the newly created South American and West African continental shelves. Today, although Tertiary landforms are frequent, very few remnants of Jurassic/Cretaceous landforms still exist on land, having been almost totally eroded away by subsequent Cenozoic denudation. In deep-sea sediments, however, correlatable deposits can be found preserved, being the best witnesses of environmental conditions on the continents. The sedimentary record points to a trend of decreasing aridity and seasonality, from the Juro-Cretaceous to the Tertiary (PETROBRAS 1994; Schaefer and Dalrymple 1995).

As rifting progressed, lower base levels favoured increasing erosional rates in the South American and African borders. Hence, inland areas, which were subjected to aridity in the Jurassic/Cretaceous, became more equable and humid in the early Tertiary onwards. These conditions, in turn, were accompanied by periods of increasing tectonic stability, as subsidence slowed down in the Tertiary. Thenceforth, deeply weathered soils developed, and can still be found resting on the inland plains of both Brazil and Africa (Fig. 5). Their development is strikingly contemporaneous with the expansion of termites, as discussed earlier. The presence of similar types of latosols and planation surfaces in both sides of the southern Atlantic is a strong indication of a long-term biotic action, converging to the same end-product: the Brazilian and African latosols. This is further corroborated by the general observation that the older the planation surface, the deeper the latosols on them. In addition, latosols and regoliths have been intensively displaced by late Tertiary/Quaternary neotectonics, even in areas thought to be `tectonically stable', like the Vicosa Plateau in Minas Gerais (Fig. 6).


On the other hand, angiosperms were contributing to the fossil record as far back as the middle Jurassic period, some 50-70 million years prior to their assumption of prominence on land (Axelrod 1961). In their early history, angiosperms inhabited seasonally dry, open country in tropical uplands in pre-Cretaceous Gondwana. Their rapid radiation from the Cretaceous to the Tertiary coincides with a newly created humid tropical climate with increasingly equable rainfall (Axelrod and Bailey 1969). In both Brazil and Africa, this was produced by ocean-floor spreading. As the broadening South Atlantic Ocean reached maximum extent at the end of the Albian and early Cenomanian, angiosperms surged to their greatest diversity and dominance.

The advent and diversification of angiosperms was a crucial factor for greater mineral weathering through the Cenozoic (Knoll and James 1987), and so favoured the radiation of termites as `concentrators--recyclers' of nutrient and organic matter. It is considered that there are generally greater net nutrient losses in deciduous (angiosperm) dominated ecosystems than in evergreen (conifers) dominated ecosystems (Knoll and James 1987). Hence, the appearance of angiosperms and the synchronous decline in evergreens is likely to have increased nutrient losses from Tertiary ecosystems, all other factors being equal. This may also explain why deep-weathered profiles are not common in the pre-Cretaceous record. Furthermore, the diversification of grasses in the Tertiary (Eocene) may have contributed further to such an increase in weathering, helping to create latosols and their microstructure. Being synchronous with the radiation of angiosperms, termites have a role in cycling and releasing nutrients. I postulate that their storage strategy and consequent promotion of microstructure have been the main factors in latosol genesis. The ties linking termite activity and latosol genesis can now recognise the rise to dominance of angiosperms. I suggest that, without their evolution, the enormous diversity of termites could not have evolved, and nor could the typical deep-weathered, microgranular latosols have formed (Fig. 7).


Termites and the Quaternary forest expansion

The expansion of forest vegetation in Brazil during the Holocene, transgressing areas formerly under savannah, was accompanied by a varying degree of drainage incision, through which the trees found a suitable medium to predate savannah environs. Areas that became flooded when the rainfall build-up began witnessed the widespread development of mound-building termites, creating flood-free islands on which forest established more easily. Recent accounts of forest islands created by termites have shown its striking role in maintaining diversity and promoting forest invasion (Oliveira-Filho and Furley 1990; Abbadie et al. 1992; Dubs 1992). However, pedologists seldom recognise the key role of contemporary termites in maintaining tropical vegetation-efficient nutrient cycling, enhanced nitrogen fixation, and soil displacement.

Deeply weathered paleosols: are they latosols?

Elsewhere in the stratigraphic record, reference is made to very old, deeply weathered `soil-like' materials (so-called oxisols), dominated by aluminous and ferruginous materials since the Pre-Cambrian (Salop 1983; Morris 1985; Reimer 1986; Retallack 1992). However, none of these occurrences display the kind of microstructure found in present-day latosols, although some may have similar mineralogical compositions (e.g. Reimer 1986; Knauer and Schrank 1993). Laterites in Sri Lanka, dating to late Cretaceous/early Tertiary have extensive channeling, attributed to roots and burrowing animals (Vermatt and Bentley 1955). Similarly, Schaefer (1995) considered the formation of a Tertiary ironstone in the north-east of Brazil to be a former oxic B horizon with strong biological activity, cemented by iron-rich solutions and subjected to relief inversion.

Thus, it is likely that a true latosolic, microgranular structure began to develop in the late Cretaceous, reaching a climax from the mid-Tertiary onwards. Earlier deeply weathered soils, although of similar mineralogical composition, were shallower and structurally very different, as discussed early in this paper.

Whether the present conclusions apply to similar latosols elsewhere (e.g. ferrosols in Australia or oxisols in Africa) depends on further investigations with emphasis on their microstructure and ecological relationship with termite distribution and pedogenetic role.


1. Brazilian latosols display different degrees and types of microgranular microstructure.

2. This type of microstructure is a long-term biotic product, and soil termites are among the major organisms involved in their genesis.

3. The formation of latosols is synchronous with the appearance of termites and the associated rise to prominence of angiosperms. The widespread expansion of termites and angiosperms in the Tertiary was accompanied by a similar trend in latosol genesis in the tropics. It is likely that the greater net nutrient loss associated with the angiosperm-dominated vegetation favoured the widespread formation of gibbsite in the regolith (Fig. 7).

4. The microaggregates' stability and their preservation are favoured primarily by gibbsite, and secondarily by Fe-oxides. However, oxi-hydroxides alone cannot account for the microstructure of latosols.

5. Although the presence of the microgranular structure in not lithodependent, the degree of microgranular structure is related to the underlying lithology. Latosols developed on mafic rocks, like basalts, usually show better microaggregation, with well-rounded grains. They are also the most gibbsitic of all. Kaolinitic latosols on sedimentary deposits are normally the least microaggregated, and the granular particles are less rounded and more blocky.

6. Latosols senso stricto can be found only from the Cretaceous onwards in the stratigraphic record. Earlier candidates do not have a similar microstructure.


I wish to thank coleagues Dr John Dalrymple, Prof. Bob Gilkes, Prof. Mauro Resende, Dr Antonio E Mello Marques, Dr Cesar Varajao, Dr Joao Ker, and Dr Ian Fordyce for many stimulating discussions on tropical pedology and the role of termites on latosols genesis. Some thin sections of soils from the Eighth International Workshop on Soil Classification were kindly provided by the Late friend Mariza Duarte (EMBRAPA). I also thank the assistance provided by Mr Franz Street for the SEM/EDX analyses (PRIS, University of Reading, UK). Preliminary identification of termite genera was kindly provided by the Entomologist, Dr Og F. F. de Souza (Universidade Federal de Vicosa).


Abbadie L, Lepage M, Le Roux X (1992) Soil Fauna at the forest-savanna boundary: role of termite mounds in nutrient cycling. In `Nature and dynamics of forest-savanna boundaries'. (Eds PA Furley, J Proctor, J Ratter) pp. 471-482. (Chapman and Hall: London)

Ab'Saber AN (1962) Revisao sobre o horizonte subsuperficial de cascalhos inumados do Brasil Oriental. Boletim Universidade do Parana -Geografia Fisica 2, 1-32.

Agassiz L, Agassiz EC (1868) `A journey in Brazil.' (Ticknor and Fields Co.: Boston, MA)

Ahn PM (1970) `West African soils.' (Oxford University Press: Oxford)

Axelrod DI (1961) How old are the Angiosperms? American Journal of Science 259, 447-459.

Axelrod DI, Bailey HP (1969) Paleotemperature analysis of Tertiary floras. Paleogeography, Paleoclimatology, Paleoecology 6, 163-195

Barros Machado A (1982) Termitic remains in some bauxites. In `Annals of the II Seminar on Lateritisation Processes'. pp. 251-254. (USP Press: Sao Paulo)

Boyer P (1973) Action de certains termites constructeurs sur l'evolution des sols tropicaux. I: les termites et le sol. Ann. Sci. Nat. Zool. 15, 329-498

Brewer R (1964) `Fabric and mineral analysis of soils.' (Wiley: New York)

Brune A, Kuhl M (1995) pH profiles of the extremely alkaline hindguts of soil-feeding termites (isoptera: Termitidae) determined by microelectrodes. Journal of Insect Physiology 42, 1121-1127.

Cailleux A (1957) La ligne de cailloutis a la base de sols jaune. Zeitschrift fur Geomorphology I, 312-313.

Cailleux A (1966) Os depositos detriticos, a linha de cascalhos enterrados e os cupins. Noticia Geomorfologica 4, 43-50.

Camargo MN, Jacomine PKT, Carvalho AP, Olmos J (1988) The Brazilian classification of latosols. In `Proceedings of the 8th International Soil Classification Workshop'. pp. 190-202. (EMBRAPA: Rio de Janeiro)

Constantino R (1995) Revision of the neotropical termite genus Syntermes Holmgren (Isoptera: Termitidae). University of Kansas Science Bulletin 55, 455-518.

Correa GF (1989) Les microreliefs `Murundus' et leur environnment pedologique dans l'ouetat du Minas Gerais, region du Plateau Central Bresilien. PhD thesis, Universite de Nancy I.

Coventry R J, Holt JA, Sinclair DF (1988) Nutrient cycling by mound-building termites in low-fertility soils of semi-arid tropical Australia. Australian Journal of Soil Research 26, 375-390.

Dubs B (1992) Observations on the differentiation of woodland and wet savanna habitats in the Pantanal of Mato Grosso, Brazil. In `Nature and dynamics of forest-savanna boundaries'. (Eds PA Furley, J Proctor, J Ratter) pp. 431-449. (Chapman and Hall: London)

EMBRAPA (1988) Classification, characterization and utilization of Oxisols. In `Proceedings of the 8th International Soil Classification Workshop'. p. 160. (Ministerio da Agricultura: Rio de Janeiro)

EMBRAPA (1997) `Manual de Metodos de Analises do Solo.' (Ministerio da Agricultura: Rio de Janeiro)

Emerson AE (1955) Geographical origin and dispersal of termite genera. Fieldiana, Zoology 37, 465-521.

Emerson AE (1965) A review of the Mastotermitidae (Isoptera), including a new fossil genus from Brazil. American Museum Novitates 2236, 1-46.

Erhart H (1951) Sur le role des cuirasses termitiques dans la geographie des regions tropicales. Cahiers Recherche Academic Scientifique Paris 233, 966-968.

Eschenbrenner V (1986) Contribution des termites a la microagregation des sols tropicaux. Cahiers ORSTOM, Series Pedologie 22, 397-408.

Ferreira MM (1988) Influencia da mineralogia da fracao argila nas propriedades fisicas de latossolos Brasileiros. DSc thesis, Universidade Federal de Vicosa, Brazil.

Garnier-Sillam E, Harry M (1995) Distribution of humic compounds in mounds of some soil-feeding termite species of tropical rain forest: its influence on soil structure stability. Insectes Sociaux 42, 167-185.

Gomes MAF (1988) Intemperismo de anfibolitos e sua contribuicao a genese de solos de duas provincias metamorficas do Estado de Minas Gerais. MSc thesis, Universidade Federal de Vicosa, Brazil.

Grasse PP (1986) `Termitologia. 3: comportment, sociabilite, ecologic, evolution, systematique.' (Singer-Polignac: Paris)

Hartt CF (1870) `Geology and Physical Geography of Brazil.' (Fields Osgood and Co.: Boston, MA)

Holt JA (1990) Observations on the relationships between meat ants and termites in tropical Australia. Journal of Tropical Ecology 6, 379-382.

Holt JA (1996) Termites and tropical landscapes: implications for sustainable production. In `Annals of the Latin American Congress of Soil Science'. pp. 1-11. (Biology section, CD-ROM, Brazilian Society of Soil Science: Piracicaba)

Jenny H (1980) `The soil resource: origin and behaviour.' (Springer-Verlag: New York)

Johnson DL (1990) Biomantle evolution and the redistribution of earth materials and artifacts. Soil Science 149, 84-101.

Jones CG, Lawton JH, Shachak M (1994) Organisms as ecosystems engineers. Oikos 69, 373-386.

Jungerius PD, van der Ancker JAM, Mucher HJ (1999) The contribution of termites to the microgranular structure of soils on the Uasin Gishu Plateau, Kenya. Catena 34, 349-363.

Kellogg CE (1949) Preliminary suggestions for the classification and nomenclature of great soil groups in the tropical and equatorial regions. Commonwealth Bureau Soil Science Technical Communication No. 46, pp. 76-85.

Klute A (1986) `Methods of soil analysis. Part I--Physical and mineralogical methods.' (SSSA Press: Madison, WI)

Knauer LG, Schrank A (1993) A origem dos filitos hematiticos da Serra do Espinhaco Meridional, Minas Gerais. Geonomos 1, 33-38.

Knoll MA, James WC (1987) Effect of the advent and diversification of vascular land plants on mineral weathering through geologic time. Geology 15, 1099-1102.

Lavelle P (1983) The soil fauna of tropical savannas. In `Tropical Savannas'. pp. 485-504. (Ed. F Bourbiere) (Elsevier: Amsterdam)

Lee KE, Wood TG (1971) `Termites and soils.' (Academic Press: London)

Lepage M, Morel G, Resplendido C (1974) Decouverte de galeries de termites atteignant la nappe phreatique profonde dans le nord du Senegal Cahiers Recherche Academie Scientifique 278, 1855-1858.

Lim CH, Jackson ML (1982) Dissolution for total elemental analysis. In `Methods of soil analysis, chemical and microbiological properties'. (Soil Science Society of America: Madison, WI)

Lobry de Bruyn LA, Conacher AJ (1990) The role of termites and ants in soil modification: a review Australian Journal of Soil Research 28, 55-93.

Lobry de Bruyn LA, Conacher AJ (1995) Soil modification by termites in the central wheatbelt of Western Australia. Australian Journal of Soil Research 33, 179-193.

Lovelock JE (1979) Gaia, a new look at life on Earth (Oxford University Press: New York)

Mabesoone JM, Neumann VH (1994) Cenozoic relief and soil development in north-east Brazil. In `Proceedings of the 14th International Sedimentological Congress'. Vol. 8, pp. 10-11. (ISS: Recife)

Mabesoone JM, Rolim JL, Castro C (1977) Late Cretaceous and Cenozoic history of north-eastern Brazil. Geologie in Mijnbouw 56, 129-139.

Martius C (1994) Diversity and ecology of termites in amazonian forests Pedobiologia 38, 407-428.

Melo Marques JJGS (2000) Trace elements distribution in Brazilian cerrado soils at the landscape and micrometer scales. PhD thesis, Purdue University, West Lafaiette, USA.

Miklos AAW (1992) Biodynamique d'une couverture pedologique dans la region de Botucatu, Bresil. PhD thesis, Universite Paris VI, France.

Missao de Pedologia de Angola e Mocambique (1968) Esboco Pedologico do Distrito de Cuando-Cubango (Junta de Investigacoes do Ultramar: Lisboa)

Morris RC (1985) Genesis of iron ore in iron-banded formation by supergene and supergene-metamorphic processes. In `Handbook of strata-bound and stratiform ore deposits. Regional studies and specific deposits'. (Ed. KH Wolf) pp.73-235. (Elsevier: Amsterdam)

Norton LD, Bigham JM, Hall GF, Smeck NE (1983) Etched thin sections for coupled optical and electron microscopy and microanalysis. In `Submicroscopic studies of soils'. (Eds EBA Bisdom, J Ducloux) (Elsevier: Amsterdam)

Nunes WAGA, Schaefer CER, Ker JC, Fernandes-Filho EI (2000) Micropedological characterization of some soils of the Zona da Mata, Minas Gerais, Brazil. Brazilian Journal of Soil Science 24, 103-116.

Nutting WL, Haverty MI, Le Fage JP (1987) Physical and Chemical alteration of soil by two subterranean termite species in Sonoran desert grassland Journal of Arid Environments 12, 233-239.

Oliveira AI (1929) `A bacia do Rio Branco.' (Servico Geologico e Minealogico do Brasil: Rio de Janeiro)

Oliveira-Filho AT, Furley PA (1990) Monchao, Cocurutu, Murundu. Ciencia Hoje 11, 30-37.

PETROBRAS (1994) `Estratigrafia das bacias sedimentares Brasileiras.' pp. 5-245. (Special volume Boletim Geociencias Petrobrtis: Rio de Janeiro)

Reimer TO (1986) Alumina-rich rocks from the Early Precambrian of the Kaapvaal Craton as indicators of paleosols and as products of other decompositional reactions. Precambrian Research 32, 155-179.

Resende M, Curi N, Rezende SB, Correa GF (1995) `Pedologia: base para distincao de ambientes.' (NEPUT: Vicosa)

Retallack GJ (1992) Paleozoic paleosols. In `Weathering, soils and paleosols'. (Eds IP Martini, W Chesworth) pp. 543-560. (Elsevier: Amsterdam)

Rezende SB (1980) Geomorphology, mineralogy and genesis of four soils on gneiss in south-eastern Brazil. PhD thesis, Purdue University, USA.

Salik J, Herrera R, Jordan CF (1983) Termitaria nutrient patchiness in nutrient deficient rainforest Biotropica 15, 1-7.

Salop LJ (1983) `Geological evolution of the Earth during the Precambrian.' (Springer: Berlin)

Schaefer CER (1995) Micropedologia de uma sequencia Latossolo-Podzolico de Vicosa, MG: quimica e morfologia da superficie de agregados dos horizontes Bw e Bt. In `Annals of XXV Brazilian Congress of Soil Science'. pp. 372-374. (SBCS: Vicosa)

Schaefer CER, Dalrymple JB (1995) Landscape evolution in Roraima, north Amazonia: planation, paleosols and paleoclimates. Zeitschrift fur Geomorphology 39, 1-28.

Setzer J (1949) `Os solos do Estado de Sao Paulo.' (Conselho Nacional de Geografia: Rio de Janeiro)

Stoops G (1964) Application of some pedological methods to the analysis of termite mounds. In `Etudes sur les Termites Africains'. (Ed. A Bouillon) pp. 379-398. (Universite Lovanne: Leopoldville)

Sys C (1955) The importance of termites in the formation of latosols. Sols Africains 3, 392-395.

Tardy Y (1992) Diversity and terminology of lateritic profiles. In `Weathering, soils and paleosols'. (Eds IP Martini, W Chesworth) pp. 379-406. (Elsevier: Amsterdam)

Tardy Y, Roquin C (1992) Geochemistry and evolution of lateritic landscapes. In `Weathering, soils and paleosols'. (Eds IP Martini, W Chesworth) pp. 407-444. (Elsevier: Amsterdam)

Thiry M, Schmitt J-M, Rayot V, Milnes AR (1995) Geochemie des alterations des profils blanchis du regolithe tertiaire de l'interieur de l'Australie. Cahiers Recherche Academic Scientifique Paris 320, 279-285.

Valentine KWG, Dalrymple JB (1976) Quaternary buried paleosols: a critical review. Quaternary Research 6, 209-222.

van Breemen N (1993) Soils as biotic constructs favouring net primary productivity. Geoderma 57, 183-211.

Varajao AFDC (1988) Genese das Jazidas de Caulinita e Bauxita de Vargem dos Oculos, Quadrilatero Ferrifero, MG. MSc thesis, Universidade de Sao Paulo, Brazil.

Varajao AFDC, Melfi A J, Boulange B (1990) Morphological, mineralogical and chemical characterization of structural facies of kaolinite deposit of Vargem dos Oculos, Quadrilatero Ferrifero, MG. Revista Brasileira Geociencias 20, 75-82.

Vermatt JG, Bentley CF (1955) The age and channeling of Ceylon laterite. Soil Science 79, 239-247.

Watson JP (1972) The distribution of gold in termite mounds and soils at a gold anomaly in Kalahari sands. Soil Science 113, 317-321.

Wielemaker WG (1984) Soil formation by termites, a study in the Kisii area, Kenya. PhD Thesis, Wageningen University, The Netherlands.

Wilson EO (1992) Social insects as dominant organisms. In `Biology and evolution of social insects'. (Ed. J Billen) pp. 1-7. (Leuven University Press: Belgium)

Wood TG, Sands WA (1978) The role of termites in ecosystems. In `Production ecology of ants and termites'. (Ed. MV Brian) pp. 245-292. (Cambridge University Press: Cambridge)

Woodworth JB (1912) Geological expedition to Brazil and Chile, 1908-1909. Bulletin of the Museum of Comparative Zoology, Harvard University, LVI(1), Cambridge, MA

Yakushev VM (1968) Influence of termite activity on the development of laterite soil. Soviet Soil Science 1, 109-111.

Manuscript received 26 October 2000, accepted 12 March 2001

Carlos E. R. Schaefer

Universidade Federal de Vicosa, 36571-000 Minas Gerais, Brazil. E-mail:
COPYRIGHT 2001 CSIRO Publishing
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2001 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Schaefer, Carlos E. R.
Publication:Australian Journal of Soil Research
Article Type:Statistical Data Included
Geographic Code:3BRAZ
Date:Sep 1, 2001
Previous Article:Differences in sorption behaviour of carbaryl and phosalone in soils from Australia, Pakistan, and the United Kingdom.
Next Article:Application of mineral magnetism to describe profile development of toposequences of a sedimentary soil in south-eastern Australia.

Related Articles
Micromorphology and electron microprobe analysis of phosphorus and potassium forms of an Indian Black Earth (IBE) anthrosol from Western Amazonia.

Terms of use | Privacy policy | Copyright © 2019 Farlex, Inc. | Feedback | For webmasters