Soils consumed by chimpanzees in the Kibale Forest, Uganda

International Journal of Primatology, Vol. 26, No. 6, December 2005 ( C© 2005)DOI: 10.1007/s10764-005-8857-7

Soils Consumed by Chimpanzees of the KanyawaraCommunity in the Kibale Forest, Uganda

William C. Mahaney,1,5 Michael W. Milner,1 Susanne Aufreiter,2

R. G. V. Hancock,2 Richard Wrangham,3 and Sean Campbell4

Received February 11, 2004; revision May 17, 2004; second revision October 29, 2004;accepted November 17, 2004

We previously reported on a study of 4 soils that chimpanzees of theKanyawara community in the Kibale National Park, Uganda consumed on anear-daily basis. We suggested that iron was a possible chemical stimulus inassociation with high quantities of Si:Al = 1:1-dominated clay minerals in theconsumed material. To test our initial findings, we analyzed 18 samples fromthe same general area including 7 samples that the chimpanzees did not eat.Among the chemical elements, As, Au, Br, Ca, Cl, Dy, Mg, Ni, Sb, Sr, and Iare below detection limits. Only Fe stands out as a potentially important nu-tritional element present in sufficient quantity to provide a physiological stim-ulus for chimpanzees living at high elevations near the flanks of the Ruwen-zori Mountains. Along with Fe, metahalloysite is present in high amounts inthese soils. In its pure crystalline form as a pharmaceutical grade clay mineralmetahalloysite may well counteract the debilitating effects of diarrhea, with aneffect similar to what is achieved with kaolinite (cf. KaopectateTM). An unex-pected result, the relatively high nitrogen and carbon in the eaten samplesrelative to the uneaten group, indicates the chimpanzees may have a higherthreshold for organic-rich material than previously believed. Contrarily, thecolor of the ingested material, depicts a material with less humus than in theuneaten group, a finding that is compatible with previous work reported at

1Geomporphology and Pedology Laboratory, York University, 4700 Keele St., North York,Ontario, Canada, M3J 1P3.

2Department of Chemical Engineering and Applied Chemistry, University of Toronto.Toronto, Ontario, Canada, M5S 3E5.

3Department of Anthropology, Harvard University, Cambridge, Massachusetts 02138.4Department of Geography, University of Kentucky, Lexington, Kentucky 40506.5To whom correspondence should be addressed; e-mail: [email protected].

1375

0164-0291/05/1200-1375/0 C© 2005 Springer Science+Business Media, Inc.

1376 Mahaney, Milner, Aufreiter, Hancock, Wrangham, and Campbell

other geophagy sites in Africa. Of all the choices of soil available to them,the chimpanzees appear to be selecting highly homogeneous chemical nat-ural earths with well-leached and uniform mineralogical material similar tothe uneaten group, but with higher relative amounts of clay size material.

KEY WORDS: chimpanzee geophagy; clay minerals; secondary plant compound.

INTRODUCTION

Geophagy, a common behavior of primates and other animals, is welldocumented in the literature (Krishnamani and Mahaney, 2000; Laufer,1930; and Mahaney and Krishnamani, 2003). Studies of geophagy soils fromdifferent geographical locations worldwide have shown that they are oftencomprised of clay-rich material with similar mineralogies and chemical el-emental contents (Aufreiter et al., 1999; Browman and Gunderson, 1993;Krishnamani and Mahaney, 2000). Most of such soils analyzed contain ahigh proportion of Si:Al = 1:1 secondary clay minerals, with lesser amountsof Si:Al = 2:1 minerals, some quartz, and, often, high amounts of iron. Theclay fraction of the soils is a physically and chemically active material thatmay absorb or adsorb (or conversely release) water, inorganic elements,or organic compounds to varying degrees, depending on the amounts ofspecific clay minerals and depending on variations of such conditions aspH and solute concentration, either to individual chemical elements or tomineralogy.

The adaptive value of soil ingestion is likely linked to soil composition.Many studies have considered a benefit of geophagy to be the adsorption oftoxic organic components such as alkaloids in plant foods, as Gilardi (1996)and Gilardi et al. (1999) have shown in vivo for parrots and Wink et al.(1993) for geese. Clay minerals are also capable of adsorbing bacterial orparasite-generated toxins (Said et al., 1980). Researchers have consideredgeophagy to be important in ungulates for maintaining constant conditionsof factors such as pH in the rumen, particularly when dietary conditions arevariable (Jones and Hansen, 1985).

We report on the chemistry, geochemistry, and mineralogy of a seriesof soils wild chimpanzees of the Kanyawara community ingested across aportion (35 km2); of the Kibale National Park (795 km2).

The Kibale Forest in Uganda is a well-documented (Ghiglieri, 1984;Wrangham et al., 1996) research area for chimpanzee (Pan troglodytes)studies. It lies east of the Ruwenzori Mountains on the eastern side of theWestern Rift Valley in southwestern Uganda (Fig. 1), 50 km north of theequator.

Geophagy by Chimpanzees 1377

Fig. 1. Location of Kibale sites in the Kibale Forest, Uganda.

Previously, Mahaney et al. (1997) studied 4 paleosols (ancient soils)from sites where chimpanzees had eaten soil . They examined samples AJ,SY, ST, and sample K14-E14 for physical attributes such as color, struc-ture, and particle size distribution, and later prepared subsamples for min-eralogical analysis by X-ray diffraction (XRD) and for chemical analysis byatomic absorption spectrophotometry (AAS) and instrumental neutron ac-tivation analysis (INAA). The analyzed material is ingested by chimpanzeesof the Kanyawara community on close to a daily basis with unknown ef-fects, and is similar in chemistry and mineralogy to other soils analyzed formountain gorillas in Rwanda (Mahaney et al., 1990, 1995b), orangutans inIndonesia (Stambolic-Robb, 1997), chimpanzees in the Mahale Mountains


(Aufreiter et al., 2001; Mahaney et al., 1996), and rhesus macaques in CayoSantiago (Mahaney et al., 1995a) and Japan (Mahaney et al., 1993; Wakibaraet al., 2001). To test and extend the conclusions reached in the 4-samplestudy (Mahaney et al., 1997), Wroangham collected 18 additional samplesas part of studies carried out with Makerere University, Uganda. They in-clude 10 samples from sites where chimpanzees ate soil, and 8 from siteswhere they had not eaten soil. We report the results of tests of the physical,mineralogical, and chemical attributes of ingested soils vs. uneaten samples,so as to assess the extent to which chimpanzees select soils with specificproperties.

The research questions we address include the following:

Is clay mineralogy a controlling factor in geophagy, at least within theKanyawara community of chimpanzees?

Does it alone control the selection of material for consumption?Does comestible material size matter?Is there ample proof that chimpanzees seek clay-rich material regardless of

composition?Does Fe, or other elements, supply a secondary controlling factor in

geophagy?Is the selection of geophagy soils based on color and/or carbon/nitrogen

content?

MATERIALS AND METHODS

Sites

The Kanyawara chimpanzee geophagy soil sites are ca. 50 km north ofthe equator and 10 km southeast of Fort Portal (Fig. 1). They are located inthe northern headwaters of the Nyakagera, a left bank tributary of the DuraRiver, which drains south to Lake George. Bedrock geology is quartzite butsignificant surficial material derives from both a basement gneiss complexin the northern part of the Dura in the Fort Portal Volcanic field and a car-bonatite (igneous rock composed mainly of calcite) complex immediatelywest of the Dura Drainage. The sites are relatively close to one another andaccessible to the Kanyawara chimpanzees that inhabit an area rich in vol-canic ejecta and an abundance of loessic (wind blown) minerals that havecreated a fertile soil within the Kibale National Park (Mahaney et al., 1997).The mean elevation is 1500 m a.s.l. and the local relief is ca. 300 m. From1984 to 1990 the mean monthly maximum temperature was 22◦C (range19.8–24.7◦C) and the mean monthly minimum temperature was 15◦C (range13.9–17◦C). Mean annual precipitation is about ca. 140 cm. The prevailing


trade winds are from the northeast in winter and southeast in summer; localgravity winds from the Ruwenzori are common any time of the year. Grav-ity winds from the expanded alpine glaciers of the Pleistocene would haveled to the incorporation of fine-grained material (Mahaney, 1990a) derivedfrom the uplifted basement complex of granite and gneiss in the nearbyRuwenzori Mountains, and may explain the prevalence of quartz in manyof the soils under investigation.

Laboratory Analyses

We based sample selection and the number of eaten samples (n = 10)on observed geophagy; the 8 controls are from nearby sites refused by chim-panzees. We air-dried the samples in the laboratory and subsampled forvarious analyses including water retention, particle size, mineralogy, chem-istry, and geochemistry. Sufficient mass of sample was available to com-plete particle size analysis with the air-dried equivalent of 50 g oven-driedsoil (Mahaney and Krishnamani, 2003). The particle size analysis followedDay (1965), with the coarse fractions (63–2000 µm) determined by dry siev-ing and the fine fractions (<63 µm) by hydrometer. For XRD analysis wefollowed procedures established by Whittig (1965), Mahaney (1981), andWilson (2003). We used centrifugation to prepare oriented clay mounts(Jackson, 1956) and analyzed them on a Toshiba ADG-301H X-ray diffrac-tometer using CuK-alpha radiation. Geochemical analysis by neutron ac-tivation follows procedures outlined by Hancock (1984). We carried outchemical analysis by electron microprobe at Arizona State University ona JEOL JXA 8600 Superprobe with a 10-µm beam and 15 keV accelerat-ing voltage. We vacuum-impregnated the samples with epoxy resin beforepolishing on diamond lapping film. A carbon coating 20 A thick preventedsurface charging. We conducted an examination and analysis by scanningelectron microscopy (SEM) and energy dispersive spectrometry analysis ofsilt and sand grains at the University of Toronto using a JEOL 840 SEM-Link EDS system.

RESULTS

Kanyawara chimpanzees eat soil often, mostly between 1500 and1900 h, usually no more than once per day. Selected soils are mainly dry,colored red or brown, and taken from below the humus layer. Two kindsof site are preferred. Both are so abundant that chimpanzees appear to relyon finding them by chance, rather than returning to known sites (as they do


with fruit trees). Indeed, they rarely divert >5 m from a travel path to finda site for geophagy.

The more common type is the root mass of a fallen tree, from whichchimpanzees take the compacted soil (74% of 50 recorded visits). The otheris a natural hole or low cliff where the sub-humus soil is exposed (26% ofvisits). A chimpanzee normally breaks off a small chunk (visually estimatedat about 5–10 ml), and holds it manually while nibbling small pieces off it.It may take the chimpanzee as long as 2 min to eat this amount. Individ-uals traveling together may eat soil within the same 30-min period eitherfrom the same, or from nearby, sites (≤7 recorded to do so), but some-times only one chimpanzee will eat soil. These behaviors indicate that chim-panzees follow simple selection rules. Instead of choosing a particular typeof soil, they appear to eat soil from anywhere within their range where itcan be located below the humus layer, but not necessarily lacking in organicmatter.

Wrangham tasted the soil from a low cliff, which elephants had tuskedto obtain soil (April 1996), and which appeared typical of the kinds of soilthat chimpanzees eat. He described it as “chewy, sticky, pasty, smooth;some milk-of-magnesia quality, (like) milky glue; (it contains) a little grit,but only a few grains (remain in the mouth after swallowing); it coats theinside of the mouth and takes taste away.”

Particle Size

To determine the texture of the ingested and uneaten samples, we cal-culated the particle size curves (Figs. 2 and 3). Each particle size curve de-picts the amount of sand, silt, and clay in the sample. For example, the per-cent sand is read directly from the 63 µm or +4 phi line; the silt plus clayfrom the +9 phi line ([sand + silt] – sand = silt); and the clay from the+9 phi value – 100%.

In general, the uneaten samples are coarser than the ingested sampleswith a difference between them of ca. 20%. The ingested samples are closeto a claystone in particle size, with clay dominating the 3 major grade sizes,the slightly parabolic curves ranging from 39% to 91% clay, which includesthe textural grades of clay and clay loam. Samples 2 and 11 appear anoma-lous: sample 2 contains ca. 6% clay with ca. 60% sand, the coarsest samplein the ingested group; sample 11 contains almost no sand, 85% silt, and15% clay and is almost certainly of wind-blown origin (discussed previ-ously). The lower sand and higher clay in the ingested samples show thatthe chimpanzees, from the material available to them, exhibit a clear pref-erence (excepting samples 2 and 11) for eating finer textured samples.


Fig. 2. Particle size curves for samples; nos. 1,3, 5, 7, 8, and 10 ingested; nos. 2,4, 6, 9, 11, and13 are uneaten samples.

Soil Chemistry

The soil pH ranges from 5.4 to 6.3 in the ingested samples and from5.4 to 6.4 in the uneaten group, compatible with recent findings by Gilardi(1996) and Gilardi et al. (1999) with respect to geophagy practiced byparrots in South America. With some overlap, the electrical conductivity(Table I) shows the ingested samples to have slightly lower concentrationsof salts relative to the uneaten samples. Differences in inorganic salts areprobably relatively insignificant in determining which soils are selected foringestion.

The carbon content in the ingested samples is similar to, or lower than,in the uneaten group which follows from the color differences that suggest


Fig. 3. Particle size curves for samples A–E; B, C, and E are ingested; A and D are uneatensamples.

higher humus in the uneaten group. This range of carbon/nitrogen in theingested samples suggests the chimpanzees have a threshold level for car-bon in soils ranging up to 2.7%. This concentration of carbon is less thanhalf an order of magnitude higher than previously reported for carbon ingeophagic soils studied at other sites/areas, such as the Mahale Mountains(Aufreiter et al., 2001; Mahaney et al., 1996), and with respect to othernonhuman primates (Mahaney et al., 1993, 1995a,b, 1996) and to humans(Aufreiter et al., 1997).

The nitrogen trends follow the distribution of carbon (1:10) in the2 groups of samples, with somewhat lower concentrations in the uneatengroup (1:8). Despite minor overlap between the ingested and uneaten groupof soils the high N and H concentrtions in the ingested group, relative to


Table I. Selected physical and chemical parameters of theingested (i) and uneaten (u) samples, Kibale Forest, Uganda

pH E.C.b C H NSample Colora (1:5) S/cm2) (%) (%) (%)

1i 7.5YR 4/4 — — 1.3 0.96 0.182u 10YR 4/3 — — 14.0 2.40 1.303i 2.5YR 4/6 6.12 107.9 1.0 1.50 0.134u 5YR 4/6 6.10 184.0 8.7 2.20 0.865i 5YR 5/6 — — 1.1 1.50 0.146u 5YR 3/6 6.16 148.5 5.2 1.90 0.557i 2.5YR 4/6 5.44 100.0 2.2 1.60 0.278i 2.5YR 4/6 6.00 75.0 1.9 1.50 0.229u 2.5YR 4/4 5.40 84.0 5.5 1.90 0.55

10i 2.5YR 4/6 5.84 35.5 1.7 1.50 0.2211u 5YR 3/3 5.86 197.6 14.0 2.40 1.0912i 5YR 3/6 6.32 95.7 2.7 1.60 0.3313u 7.5YR 3/4 — — 12.0 2.50 1.27Au 5YR 4/4 5.29 70.0 4.7 1.50 0.51Bi 5YR 4/2 5.61 67.4 2.8 1.30 0.33Ci 5YR 5/4 5.41 74.0 0.7 1.00 0.10Du 7.5YR 4/3 6.42 106.8 2.7 1.10 0.31Ei 5YR 4/1 5.78 60.9 1.0 0.99 0.16

aAfter Oyama and Takehara (1970).bElectrical conductivity.

that reported elsewhere in Tanzania (Aufreiter et al., 2001), could representa greater range and possible concentrations of biologically useful organicssuch as amino acids; it appears that the chimpanzees selected soils with sim-ilar to in one case but generally with lower concentrations of carbon than inthe uneaten group.

Colorimetry

Often overlooked in the analysis of soils is the relationship between thesoil color (on a moist or dry basis) and the color of the samples in beakers orhydrometer jars, which are more similar to soil colors in the field. Soil colorin the field (moist color) is a useful guide to the amount of organic matterand iron in soils; black to dark brown indicates higher to lower amounts ofhumus; variations in reddish colors reflect concentrations of Fe2O3; yellowindicate goethite, or brown hydrous iron oxide. Because levels of organicmatter and presence of iron often dictate the geologic age of a material andwhether or not it will be eaten, color is an important soil property worthconsidering in geophagy investigations (Mahaney and Khrisnamani, 2003).As shown in Fig. 4, there are distinct variations in color among the samples


Fig. 4. Colorimetry of selected ingested and uneaten samples. The 2 dark samples on the rightof the frame, last and third from the end of the group, are organic-rich A horizons of soilavoided by the chimpanzees. The other samples show varying amounts of iron based on thestrength of the reddish colors and also the presence of organic matter in rather unusual largeconcentrations compared with other studies discussed in the text.

in the moist state, ranging from darker colors (lower chroma) in the uneatengroup to brighter brown and reddish brown colors in the eaten samples. Theuneaten samples are clearly darker in the moist state; however, in the caseof samples B and E the colors are grayish brown to brownish gray accordingto the soil color charts of Oyama and Takehara (1970). Uneaten samples A,C, and D, while not as dark as B and E, seem to have released a little Feand their darker color is indicated by lower chroma.

Ingested samples 7 through 10 have the strong yellow reddish hues typ-ical of the eaten samples, but with a C/N content similar to the uneaten sam-ples. A cursory examination of the data in Table I for C and N shows consid-erable overlap between the eaten and uneaten groups, so that while coloris darker for the uneaten group, this does not necessarily indicate higherconcentrations of nitrogen and carbon.

We took all colors in Table I in an air-dry state, which means overallthe colors are lighter than under field (or wet lab) conditions where theyare moister.


Mineralogy

We analyzed the clay mineralogy of the <2 µm fraction to determineif differences are discernible between the ingested and uneaten samples.Within the ingested sample group (Table II), primary minerals are scarceand we observed only residual quartz and minor plagioclase with no cleartrend. It is possible to find all the minerals present in the eaten samples,though in somewhat lesser quantities within the uneaten samples.

The clay mineralogy of both ingested and uneaten soils is dominatedby metahalloysite, the main clay mineral present, with minor amounts ofillite in 2 samples. The remainder of the <2 µm fraction contains only mi-nor amounts of quartz, plagioclase feldspar, mica, and orthoclase in smallto trace amounts. Minor amounts of allophane and goethite are present in 3of the ingested samples and absent in the uneaten samples. As a weatheringproduct in volcanic soils metahalloysite requires ≥100 ka to form. Theamounts present in the eaten and uneaten samples indicate relatively oldand well-weathered paleosols (ancient soils: (Mahaney, 1990a,b)).

In the samples, metahalloysite dominates in all the samples, and in theeaten samples a highly refined pharmaceutical grade of metahalloysite ispresent.

Table II. Mineralogya of the <2 µm fraction of the ingested (i)and uneaten (u) soils

Sample MH I Q P M O

1i X — — — — —2u XX tr X tr tr —3i XXX — — — — —4u XXX — — — — —5i XXX — — — — —6u XX — — — — —7i tr — tr tr tr —8i tr — tr — — —9u XXX — — — — —

10i X — tr tr tr —11u XX — tr X tr X12i XX — tr — — —13u X — — — — —Au XX — — — — —Bi XXX — — — — —Ci XXX — tr tr — —Du XXX — tr — — —Ei tr tr tr tr tr —

aMinerals are metahalloysite (MH), illite (I), quartz (Q), plagioclasefeldspar (P), mica (M), and orthoclase (O). The semiquantitativeamounts are given as: trace (tr), small (X), moderate (XX) and abundant(XXX).


Geochemistry

The results of geochemical analysis by INAA are in Tables IIIa andIIIb. The eaten samples are 1, 3, 5, 7, 9, 11, 13, B, C, E; the uneaten samplesare 2, 4, 6, 8, 10, 12, A, D. We took most sediment samples in pairs fromeach location, with the exception of samples 7 and E, which we gatheredin the north of the study area (Fig. 1). The eaten-uneaten sample pairs are1-2, 3-4, 5-6, 8-9, 10-11, 12-13, C-D. The single eaten-eaten sample pair isA-B.

The analytical data are relatively heterogeneous. Three samples (5, 6,and 7) have anomalously high K and Rb concentrations, high Th concen-trations, together with very low Sc, Co, Cr, and Fe concentrations. Similarto them are samples 3 and 4, which do not have high K and Rb contents,but similar high Th contents and low Sc, Co, Cr, and Fe contents.

Samples 1 and 2 have similar Th, Ta, Hf, and Na, but much higher Sc,Co, Cr, and Fe concentrations. The final large group (samples 8-13 and A-E) contains the lowest K, Rb and Th and the highest Sc, Co, Cr, and Feconcentrations.

These overall groupings are possibly caused by heterogeneous mixingof a number of distinct major mineral phases: one Th-rich, one Rb-rich, andone Fe-rich. There is no clear geographical distinction among the sites ofdifferent chemistry. While the bulk of the Fe-rich samples are from sites

Table IIIa. Geochemistry of the ingested (i) and uneaten(u) samples, Kibale Forest, Uganda

Sample K% Rb Th Ta Hf Na

1i 0.51 49 30.7 7.5 7.7 8102u 0.53 49 32.2 7.8 7.9 7003i 0.29 38 29.9 5.8 8.5 2004u 0.32 46 30.7 6.8 8.0 7505i 1.85 218 41.4 5.9 8.7 6206u 1.19 99 36.8 8.6 8.6 7407i 0.85 102 29.5 5.3 5.5 5808i 0.14 26 15.0 5.1 6.0 4609u 0.24 28 12.0 4.0 5.3 640

10i 0.18 31 13.6 4.6 5.3 34011u 0.27 24 12.6 4.2 5.5 60012i 0.18 38 20.0 6.6 6.7 42013u 0.26 25 19.8 6.0 6.0 510Au 0.20 28 17.5 5.7 5.8 600Bi 0.32 36 21.5 7.0 8.6 670Ci 0.21 39 12.9 3.6 4.0 510Du 0.25 42 19.8 6.6 7.4 500Ei 0.19 22 13.9 4.8 5.8 460

Note. All concentrations are in ppm by mass unless noted.


Table IIIb. Geochemistry of the ingested (i) and uneaten (u)samples, Kibale Forest, Uganda

Sample Sc Co Cr Fe% La Ce

1i 28.7 49.6 81 7.23 121 1822u 30.5 51.1 84 7.66 130 1943i 15.4 6.3 73 5.10 68 1354u 13.9 12.0 55 4.09 57 1315i 10.9 13.0 46 3.81 68 1686u 14.4 19.7 46 3.99 111 1857i 10.8 14.8 37 3.19 105 1678i 58.9 68.6 163 14.4 68 1349u 46.8 46.9 125 11.0 91 103

10i 54.9 54.1 157 13.8 70 11911u 49.1 77.3 154 11.9 93 10812i 50.5 39.3 204 12.8 99 24213u 58.2 91.5 219 13.9 102 166Au 51.1 87.7 192 12.0 117 147Bi 59.0 64.1 155 14.8 138 194Ci 29.5 33.6 92 8.0 85 115Du 52.5 46.9 251 12.9 110 163Ei 38.5 32.5 185 9.4 104 120

Note. As <3.0 ppm; Au <0.015 ppm; Ca <1%; Cl <200 ppm;Dy <4 ppm; I <40 ppm; Mg <1 %; Ni <60 ppm; Sb <0.33ppm; Sr <150 ppm. All concentration are in ppm by mass un-less noted.

to the west and north of the Nyakatojo River (samples A-D, 8, 9, 12, and13; Fig. 1), samples 10, 11, and E are from the east of the Nyakatojo. Thecomparatively similar samples 3, 4, 5, and 6 are from tributaries to the eastof the Nyakatojo River, in the southeastern quadrant of the study area,but geographical coherence fails with sample 7, which is the most northerlysample ca. 3 km away from others of similar chemistry.

The eaten-uneaten sample pair 1-2 is the only one that is remarkablychemically consistent. All other eaten-uneaten pairs of samples are lucky tohave elemental concentrations that match to within 20–50% relative. This,of course, makes sense when one is dealing with sediment samples, but doesnot help in data interpretation as it pertains to the potential geophagic be-haviors of the residents that live above the samples. Overall, our data donot suggest that the elemental content of these soils, with the notable ex-ception of elements integral to the clay matrices, is likely to explain theirbeing chosen for ingestion.

Electron Microprobe

The analysis of eaten sample no. 13 (Table I) (transects onFigs. 5 and 6) shows a fine-grained matrix of clay surrounding larger grains


Fig. 5. Electron microprobe transects 1–10.

of silt and fine sand. Each point on each transect is in clay material, andthe results in Table IV show a remarkably homogeneous material with aSi:Al ratio nearly always at 1:1, which is supported by the XRD results.An exception occurs at point analysis (Table IV) where Al is anomalouslyhigh, presumably owing to the incorporation of gibbsite or other clay sizealuminous minerals. The data also show the range of P across the plug of


Fig. 6. Electron microprobe transects 11–14.

eaten material, and indicate it may have some physiological importanceprovided it can be absorbed into the blood stream. Chromium across theplug compares closely with the geochemical data reported above and in thechimp diet may play a role in reducing body fat, an important nutritionalsupplement. The Cr available for absorption by the chimpanzees is close


Tab

leIV

.E

lem

enta

lche

mis

try

onm

icro

-tra

nsec

tsof

inge

sted

sam

ple

12

Poi

ntµ

mN

a 2O

MgO

Al 2

O3

SiO

2P

2O5

SO3

K2O

CaO

TIO

2C

r 2O

3M

nOF

eOB

aOT

otal

10

0.01

0.27

34.8

225

.69

0.37

0.17

0.13

0.63

1.17

0.09

0.19

15.6

80.

0479

.26

242

.60.

110.

3325

.13

26.5

50.

390.

120.

171.

500.

900

0.14

15.2

40.

1270

.70

385

.20

0.22

24.3

621

.31

0.18

0.15

0.14

0.49

8.02

0.04

0.70

20.3

50.

0876

.04

412

7.8

0.03

0.22

22.4

925

.39

0.41

0.10

0.16

0.64

1.08

0.01

0.06

15.5

20.

0866

.19

517

0.4

00.

2336

.35

21.4

60.

230.

220.

120.

761.

400

0.28

12.7

60.

0873

.89

621

3.1

00.

2226

.81

25.9

30.

250.

170.

201.

021.

050.

030.

1514

.85

0.10

70.7

87

255.

70.

040.

2821

.48

24.1

30.

230.

150.

161.

060.

950.

030.

1313

.35

0.07

62.0

68

298.

30.

040.

2019

.99

28.2

00.

110.

150.

271.

051.

050.

030

14.2

70.

0665

.42

934

0.9

00.

2327

.13

29.2

20.

410.

150.

110.

711.

270

0.13

16.4

00.

0475

.80

1038

3.5

0.08

0.33

24.7

133

.69

0.18

0.12

0.45

0.64

0.83

0.01

0.21

15.4

00.

0476

.69

1142

6.1

00.

2723

.05

31.3

60.

320.

100.

250.

701.

250.

030.

4314

.41

0.03

72.2

012

468.

70.

130.

3321

.48

26.4

00.

270.

170.

201.

231.

520

015

.97

0.02

67.7

213

511.

30

0.25

25.0

027

.38

0.23

0.12

0.16

0.77

1.05

01.

5016

.71

0.19

73.3

614

553.

90

0.27

23.8

126

.36

0.60

0.12

0.16

1.01

1.02

00.

2515

.17

0.09

68.8

615

596.

50

0.30

25.7

929

.78

0.34

0.12

0.23

0.77

1.08

00.

0914

.63

0.07

73.2

016

639.

10.

040.

3022

.07

28.0

30.

270.

200.

200.

590.

920.

030.

1314

.60

067

.38

1768

1.8

0.13

0.30

24.7

130

.31

0.30

0.17

0.17

0.70

1.05

0.04

0.12

15.5

80.

0473

.62

1872

4.4

0.01

0.25

27.5

729

.50

0.32

0.15

0.18

0.78

0.83

00.

2611

.96

0.03

71.8

419

767

0.04

0.22

25.8

528

.69

0.39

0.15

0.13

0.53

0.98

00.

2715

.49

0.12

72.8

620

809.

60.

150.

2825

.87

30.8

30.

320.

150.

240.

671.

100.

040.

3915

.98

0.13

76.1

5

Not

e.Se

eT

able

1fo

rco

lor,

pH,e

lect

rica

lcon

duct

ivit

yan

dC

,N,a

ndH

ofth

ew

hole

sam

ple.


to the estimated dietary intake for humans, ca. 0.020 to 0.035 mg/day (Foodand Nutrition Board, 2001; Mertz, 1981). Silicon, which is between 21% and33%, may also be important over and above its clay mineral significance, asit is a necessary ingredient in bone development.

DISCUSSION

The physical attributes of the soils, especially color (Table I), indi-cate that the chimpanzees consume a reddish material with hues of 2.5YRand 5YR (YR = strong yellow red hue). Indeed color and smell (the in-gested material has a potters’ clay aroma) may guide the chimpanzeesin soil selection, but the chimpanzees actually find soil by observing up-turned tree boles from a few m away. Color may be important in select-ing soils, and chimpanzees and humans have identical color vision (Jacobset al., 1996), but upturned tree-root masses have nearly the same reddishbrown color. Smell seems very unlikely to guide chimpanzees at the timeof soil selection, because their decision to engage in geophagy is clearlymade before they get close to the soil and they are not observed to sniffsoils. The smell of clay may nevertheless confirm to them that they havefound soil suitable for ingestion, if the adsorptive properties of high-claysoils are their crucial characteristics in geophagy. The chimpanzees clearlyprefer a reddish brown and yellowish brown material to the brownishblack and gray Ah (surface) horizons. The particle size data show theconsumed soils are very clay-rich with on average ≥20% clay in the con-sumed samples (Figs. 2 and 3). Small particle size increases surface areaand will increase the adsorptive nature of the soils, conferring a greatercapacity for both carrying materials into the body and for adsorption ofmolecules within the gastrointestinal (GI) tract. Small particle size likelycauses the effects of coating the mucous surfaces of the mouth, and mask-ing of taste noted by Wrangham, which may broaden the range of plantseaten by geophageous animals by masking unpleasant tastes of beneficialplants.

As Bolton et al. (1998) postulated for Kowloon monkeys and Mahaneyet al. (1996, 1997) for chimpanzees, the ingested soils are usually clay-richrelative to the uneaten samples. Also, the clayey soils are often possessedof blocky peds, structures that may merit study for possible inorganic ororganic components that might be beneficial to the chimpanzees. As re-ported elsewhere (Mahaney, 1993), soils ingested by mountain gorillas con-tain sands with appreciable clay coatings.

The pH of the ingested material is moderately acidic, suggesting thatthe humus-free soil might play a role in moderately lowering pH, which


in turn may slightly affect pH in the GI tract. Chimpanzees of the MahaleMountains of Tanzania (Mahaney et al., 1997) showed another pattern, withthe uneaten soils having higher humus compared with ingested soils. Whileoften ignored in studies of geophagy soils, the carbon and nitrogen concen-tration in the Kibale soils shows a different pattern, with the chimpanzeesconsuming soils with a color similar to and slightly lighter than the uneatengroup, but with a C/N content as high or higher than the uneaten group,which implies that chimpanzees prefer, or can beneficially withstand a min-imum amount of humus, its population of microorganisms notwithstanding(Krishnamani and Mahaney, 2000).

As Ketch et al. (2001) indicated, chimpanzees in the Mahale Mountainsmay seek soils carrying concentrations of various beneficial microorganisms(including actinomycetes, non-filamentous bacteria and fungi), a relation-ship that may well be significant here. The chimpanzees often choose soil toingest from among the roots of upended trees. Tropical soils from aroundthe roots of trees are richer in nutrient elements than surrounding soils(Stark and Jordan, 1978), but mycorhyzal fungi, and enzymes and othersecreted molecules from organisms associated with tree roots, are likelyto be present in the soils. Others have observed ingestion of soils fromamong plant roots: researchers described a favorite source of geophagysoils of mountain gorillas (Gorilla gorilla beringei) in the Parc des Volcans(Rwanda) as a cave with walls and a roof supported by tree roots (Fossey,1983; Mahaney et al., 1995b). Johns (1990) and Oates (1978) have pho-tographed excavation of geophagy soils from among the roots of trees andshrubs. We unfortunately had no opportunity to culture the samples col-lected at Kibale. Indications of organic content suggest the possibility ofbiologically active molecules such as microbial enzymes that are known tobe associated with the clay fraction of soils, and to adsorb onto clay sur-faces, retaining activity (McLaren, 1975; McLaren and Estermann, 1956;Theng and Orchard, 1995). Such enzymes have the potential, if protectedfrom digestion by their association with clays, to act in the GI tract to aid inthe breakdown of foodstuffs in the lumen of the GI tract or may simply beadditional minor sources of protein.

The soil mineralogy is notable in that metahalloysite is the predomi-nant clay in both the ingested and uneaten samples. Unlike previous stud-ies (Mahaney et al., 1995a, 1995b, 1996, 1997) in which the ingested materialis high in halloysite or metahalloysite and/or kaolinite relative to uneatensamples, in this case it is apparent the chimpanzees could consume any soilin the area to obtain high clay mineral content.

The pharmaceutical properties of metahalloysite are closely linked tothose of kaolinite as a 1:1 (Si:Al = 1:1) clay mineral (Brouillard and Rateau,1989). With the chemical composition Al2Si2O5· nH2O it is a principal


ingredient in KaopectateTM, widely used to treat gastric upsets and diarrheain humans (Mahaney et al., 1995b, 1997; Vermeer and Ferrell, 1985). Theonly other clay mineral present is illite but in concentrations that probablyplay no role in geophagy other than to release Fe or K or both in the gut.Other weathering products, including goethite, boehmite, and allophane,are only randomly present in trace quantities, probably to little effect.

We did not test the soils for the release of elements in acid conditionssuch as occur in the digestive system. However, clay minerals dissolve tovarying degrees in extremes of pH (Jozefaciuk and Bowanko, 2002; Ko-retsky, 2000; Rufe and Hochella, 1999). Acid attack on the clay structureincreases the surface area available for adsorption, and acid activation ofsoils is commonly used in industry to render soils more adsorptive (Hui,1996). Indeed, Dominy et al. (2004) showed evidence for acid activation andhigher levels of adsorption in the small intestine (where pH is higher thanin the stomach). This effect may be important if the main adaptive advan-tage of geophagy is its adsorption of toxins. Geophagy soils analyzed by ourgroup often show high amounts of finely divided material, and they could beexpected to have even greater surface areas after passage through the stom-ach. In this regard, acid-activated clays are used industrially to adsorb plantpigments such as carotenes, which are biologically active, and with foodoils to decolorize them (Christidis and Kosiari, 2003). Researchers have notstudied the adsorption of nutrients such as vitamins on ingested soils, whichcould be detrimental if it prevents the absorption of nutrients. Possibly, theadsorption of compounds on clay surfaces could protect them from degra-dation in the stomach, with subsequent release with changing pH for uptakein the small intestine.

Acid weathering of ingested clay minerals in the acid environment ofthe stomach may release elements adsorbed onto the clay surfaces, but mayalso solubilize part of the matrix elements of the minerals by dissolution atthe edges of clay platelets. We analyzed termite mound soils chimpanzeesof the Mahale Mountains of Tanzania ingest, shaking the samples in acidsolutions to simulate conditions in the stomach, and found that they did notrelease significant amounts of elements such as Fe that were present at highlevels in the solid soils. However, we detected and measured aluminum inthe acid soil extracts, suggestive of some clay mineral dissolution (Aufreiteret al., 2001). The release of various metals in the GI tract may have a nutri-tional benefit, long a hypothesized effect of geophagy in humans (Hunter,1973). Silicon release may be a significant benefit of geophagy. It is an es-sential element (Carlisle, 1982), and in the form of silicic acid is readilyabsorbed by cells (Birchall, 1994). Silicon is present in this form in soil so-lutions (Bennet and Casey, 1994) and in water, especially hard water, butits content is low in water of highly weathered areas (Birchall, 1994). While


humans are highly dependent on the availability of drinking water, manyanimals depend on the water content of foods for maintenance of fluid in-take and ingested soils may be a source of silicic acid for them. Silicon isnecessary for bone formation and growth (Carlisle, 1975, 1982).

The second major element of the clay mineral matrix, aluminum,is considered detrimental to health and has been associated with en-cephalopathies and osteomalacia. Birchall (1994) suggested that silicic acidhelps to exclude aluminum from absorption. Comparison of geochemicaldata here did not reveal element patterns indicating a greater potentialnutritional benefit from the eaten soils compared to those uneaten, andgiven the large geochemical differences among even ingested samples, itis not clear that there is a geochemical driving force for the choice of soils.However, the release of matrix and adsorbed elements from the soils maynonetheless occur, sufficient to benefit the chimpanzees. It seems likely thata number of benefits accrue from soil ingestion for the chimpanzees, includ-ing stabilization of the gut environment by the intake of mildly acidic soilswith a consistently high content of finely divided clay particles, which maybeneficially dilute the food contents of the digestive tract and hold waterwithin the tract, and may adsorb bacteria and toxins and carry them outwith the feces. Geophagy may also provide some release of beneficial nutri-ent elements such as iron and silicon.

The influence of carbonatite tuffs in soil mineralogy is not relevant aspart of this investigation, because INAA did not detect the Ca concentra-tion in the soils. Both the color factor and the pH in the soils argue forminimal carbonatite content, because carbonatite would generate high pH,high Ca values (we found <0.2% in the previous study), and light coloredsoils. Carbon (C) content values should also show a similar variation as Cavalues are on the average 34% for carbonatites and the CO2 component ofcarbonatites averages 29%.

CONCLUSIONS

In the soils ingested by Kanyawera chimpanzees, the clay mineralogy isdominated by highly refined and very well crystallized metahalloysite withno other significant clay species. Both the ingested and uneaten sampleshave high amounts of pharmaceutical grade metahalloysite, suggesting thatwhile the chimpanzees are most likely benefiting from eating the ingestedsamples, they could likely achieve similar mineralogical benefit from theavailable clay content by ingesting the uneaten samples. They may avoidingesting pathogens by choosing reddish soils, with high clay content. More-over, particle size matters and the consuming individuals select samples


with high clay content, material slightly elevated in Fe, but with mineral-ogy similar to that of the uneaten group of samples.

With respect to particle size, the chimpanzees ingest a material with aclayey texture, and relative to the uneaten samples the increase in clay is ca.20%. If the clay in these soils is a measure of the material available to coatthe mucous membranes, to adsorb toxins or to release available elementsin the GI tract, then the higher clay may well be beneficial nutritionally orpharmaceutically to the chimpanzees.

The soil chemistry suggests that neither pH nor total salts as deter-mined by electrical conductivity are likely important in determining thechoice of soils. The distribution of H+ follows the pH trend and is likelylinked to organic matter differences. Conversely the distributions of carbonand nitrogen show that the chimpanzees favor clay-rich material but withlower C and N compared with the uneaten soils.

The geochemistry shows that the geographical variations in the geo-chemistries tend to be far greater than the differences between eaten anduneaten samples at each soil ingestion location.

ACKNOWLEDGMENTS

The Natural Sciences and Engineering Research Council of Canada(NSERC) and minor research grants from Atkinson College of YorkUniversity partially funded this research. An infrastructure grant to theSLOWPOKE Reactor Facility of the University of Toronto made the geo-chemistry analysis possible. This material is based upon work in Kibale Na-tional Park supported by the National Science Foundation under Grant No.0416125. Caitlin Mahaney assisted with the chemical analysis.

REFERENCES

Aufreiter, S., Mahaney, W. C., Milner, M. W., Huffman, M. A., Hancock, R. G. V., Wink, M.,Reich, M., and Sodhi, R. (2001). Mineralogical and chemical interactions of soils eatenby chimpanzees of the Mahale Mountains and Gombe Stream National Parks, Tanzania.J. Chem. Ecol. 27: 285–311.

Aufreiter, S., Hancock, R. G. V., Mahaney, W. C., Stambolic-Robb, A., and Sanmugadas, K.(1997). Geochemistry and mineralogy of soils eaten by humans. Int. J. Food Sci. Nutr. 48:293–305.

Bennett, P. C., and Casey, W. (1994). Chemistry and mechanisms of low-temperature disso-lution of silicates by organic acids. In Pittman E. D., and Lewan, M. D. (eds.), OrganicAcids in Geological Processes. Springer-Verlag, pp. 162–200.

Birchall, J. D. (1994). Silicon-aluminum interactions and biology. In Bergna, H. E. (ed.), TheColloid Chemistry of Silica, Advances in Chemistry Series 234. American Chemical Soci-ety, Washington, DC.


Bolton, K. A., Campbell, V. M., and Burton, F. D. (1998). Chemical analysis of soils ofKowloon (Hong Kong) eaten by hybrid rhesus macacques. J. Chem. Ecol. 24: 195–205.

Brouillard, M.-Y., and Rateau, J.-G. (1989). Pouvoir d’asorption de deux argiles, la smec-tite et le kaolin sur des enterotoxines bacteriennes. Gastroenterol. Clin. Biol. 13: 18–24.

Browman, D. L., and Gunderson, J. N. (1993). Altiplano comestible earths: Prehistoric andhistoric geophagy of highland Peru and Bolivia. Geoarchaeology 8: 413–425.

Carlisle, E. M. (1975). Silicon. Nutr. Rev. 33: 257–261.Carlisle, E. M. (1982). The nutritional essentiality of silicon. Nutr. Rev. 40 (7): 193–198.Christidis, G. E., and Kosiari, S. (2003). Decolorization of vegetable oils: A study of the mech-

anism of adsorption of β-carotene by an acid-activated bentonite from Cyprus. Clays ClayMiner. 51: 327–333.

Day, P. E. (1965). Particle fractionation and particle size analysis. In Black, C. A. (ed.), Meth-ods of Soil Analysis, American Society of Agronomy, Madison, WI, pp. 545–567.

Dominy, N. J., Cavoust, E., and Minekus, M. (2004). Adaptive function of soil consumption:an in vitro study modelling the human stomach and small intestine. J. Exp. Biol. 207: 319–324.

Food and Nutrition Board, Institute of Medicine (2001). Chromium. Dietary Reference Intakesfor Vitamin A, Vitamin K, Boron, Chromium, Copper, Iodine, Iron, Manganese, Molyb-denum, Nickel, Silicon, Vanadium, and Zinc, National Academy Press, Washington, DC,pp. 197–223.

Fossey, D. (1983). Gorillas in the Mist, Houghton Mifflin, London.Ghiglieri, M. P. (1984). The Chimpanzees of Kibale Forest, Columbia University Press, New

York.Gilardi, J. D., Duffey, S. S., Munn, C. A., and Tell, L. A. (1999). Biochemical functions of

geophagy in parrots: Detoxification of dietary toxins and cytoprotective effects. J. Chem.Ecol. 25: 897–922.

Gilardi, J. D. (1996). Ecology of Parrots in the Peruvian Amazon: Habitat Use, Nutrition andGeophagy, Ph.D. Thesis, University of California, 103 pp.

Hancock, R. G. V. (1984). On the source of clay used for Cologne Roman pottery. Archaeom-etry 26: 210–217.

Hui, Y. H. (1996). Baileys’s Industrial Oil and Fat Products, 5th ed., Vols. 4 and 5, Edible Oiland Fat Products: Processing Technology, John Wiley & Sons, New York.

Hunter, J. M. (1973). Geophagy in Africa and in the United States: A Culture-NutritionHypothesis. Geograph. Rev. 63: 170–195.

Jackson, M. L. (1956). Soil Chemical Analysis-Advanced Course, published by the author,Dept. of Soil Science, University of Wisconsin, Madison, WI.

Jacobs, G. H., Deegan, J. F. II, and Moran, J. L. (1996). ERG measurements of the spectralsensitivity of common chimpanzee (Pan troglodytes). Vis. Res. 36: 2587–2594.

Johns, T. (1990). With Bitter Herbs They Shall Eat It: Chemical Ecology and the Origins ofHuman Diet and Medicine, University of Arizona Press, Tucson.

Jones, R. L., and Hanson, H. C. (1985). Mineral Licks, Geophagy, and Biogeochemistry ofNorth American Ungulates, Iowa State University Press, Ames.

Jozefaciuk, G., and Bowanko, G. (2002). Effect of acid and alkali treatments on surface areasand adsorption energies of selected minerals. Clay Clay Miner. 50: 771–783.

Ketch, L., Malloch, D., Mahaney, W. C., and Huffman, M. A. (2001). Comparative microbialanalysis and clay mineralogy of soils eaten by chimpanzees (Pan trogolodytes schwein-furthii) in Tanzania. Soil Biol. Biochem. 33: 199–203.

Koretsky, C. (2000). The significance of surface complexation reactions in hydrologic systems:a geochemist’s perspective. J. Hydrol. 230: 127–171.

Krishnamani, R., and Mahaney, W. C. (2000). Geophagy among primates: Adaptive signifi-cance and ecological consequences. Anim. Behav. 59: 899–915.

Laufer, B. (1930). Geophagy, Field Museum of Natural History, Anthropology Series,18 (2),97–198, Publication 280.


Mahaney, W. C. (1981). Paleoclimate reconstructed from paleosols: Evidence from EastAfrica and the Rocky Mountains. In. Mahaney, W. C. (ed.), Quaternary Paleoclimate.Geobooks, Norwich, UK, pp. 227–247.

Mahaney, W. C. (1990a). Glacially-crushed quartz grains in late Quaternary deposits in theVirunga Mountains, Rwanda - indicators of wind transport from the north. Boreas 19:81–89.

Mahaney, W. C. (1990b). Ice on the Equator, Wm Caxton, Ellison Bay, WI, 386 pp.Mahaney, W. C. (1993). Scanning electron microscopy of earth mined and eaten by mountain

gorillas in the Virunga Mountains, Rwanda. Primates 34: 311–319.Mahaney, W. C., and Krishnamani, R. (2003). Understanding geophagy in animals: standard

procedures for sampling soils. J. Geochem. Ecol. 29: 1503–1523.Mahaney, W. C., Watts, D. P., and Hancock, R. G. V. (1990). Geophagia by mountain go-

rillas (Gorilla gorilla beringei) in the Virunga Mountains, Rwanda. Primates 31: 113–120.

Mahaney, W. C., Hancock, R. G. V., and Inoue, M. (1993). Geochemistry and clay mineralogyof soils eaten by Japanese macaques. Primates 34: 85–91.

Mahaney, W. C., Stambolic, A., Knezevich, M., Hancock, R. G. V., Aufreiter, S., Sanmugadas,K., Kessler, M., and Grynpas, M. (1995a). Geophagy amongst Rhesus macaques on CayoSantiago, Puerto Rico. Primates 36: 323–334.

Mahaney, W. C., Aufreiter, S., and Hancock, R. G. V. (1995b). Mountain gorilla geophagy: Apossible seasonal behavior for dealing with the effects of dietary changes. Int. J. Primatol.16: 475–488.

Mahaney, W. C., Hancock, R. G. V., Aufreiter, S., and Huffman, M. A. (1996). Geochemistryand clay mineralogy of termite mound soil and the role of geophagy in chimpanzees ofthe Mahale Mountains, Tanzania. Primates 37: 121–134.

Mahaney, W. C., Milner, M. W., Sanmugadas, K., Hancock, R. G. V., Aufreiter, S., Wrangham,R. W., Pier, H. W. (1997). Analysis of geophagy soils in Kibale Forest, Uganda. Primates38: 159–176.

McLaren, A. D. (1975). Soils as a system of humus and clay immobilized enzymes. Chem.Script. 8: 97–99.

McLaren, A. D., and Estermann, E. F. (1956). The adsorption and reactions of enzymes andproteins on kaolinite: III. The isolation of enzyme-substrate complexes. Arch. Biochem.Biophys. 61: 158.

Mertz, W. (1981). The essential trace elements. Science 213: 1332–1338.Oates, J. F. (1978). Water plant and soil consumption by guereza monkeys (Colobus guereza):

a relationship with minerals and toxins in the diet? Biotropica 10: 241–253.Oyama, M., and Takehara, H. (1970). Standard Soil Color Charts. Japan Research Council for

Agriculture, Forestry and Fisheries.Rufe, E., and Hochella, F., Jr. (1999). Quantitative assessment of reactive surface area of phl-

ogopite during acid dissolution. Science 285: 874–876.Said, S. A., Shibl, A. M., and Abdullah, M. E. (1980). Influence of various agents on ad-

sorption capacity of kaolin for Pseudomonas aeruginosa toxin. J. Pharm. Sci. 69: 1238–1239.

Stambolic-Robb, A. (1997). Geophagy amongst free-ranging Sumatran orang-utans (Pongopygmaeus abelii) of Gunung Leuser National Park and ex-captive Bornean orang-utans(Pongo pygmaeus pygmaeus) of Sungai Wain Forest, Indonesia. MSc. Thesis, YorkUniversity, Toronto, p. 143.

Stark, N. M., and Jordan, C. F. (1978). Nutrient retention by the root mat of an Amazonianrain forest. Ecology 59: 434–437.

Theng, B. K. G., and Orchard, V. A. (1995). Interactions of clays with microorganisms andbacterial survival in soil: A physicochemical perspective. In Huang, P. M., Berthelin, J.,Bollag, J.-M., McGill, W. B., and Page, L. (eds.), Environmental Impact of Soil ComponentInteractions, CRC Press, Lewis Publishers, Boca Raton, FL.

Vermeer, D. E., and Ferrell, R. E., Jr. (1985). Nigerian geophagical clay: a traditional antidiar-rheal pharmaceutical. Science 227: 634–636.


Wakibara, J. V., Huffman, M. A., Wink, M., Reich, S., Aufreiter, S., Hancock, R. G. V., Sodhi,R., Mahaney, W. C., and Russell, S. (2001). The adaptive significance of geophagy forJapanese Macaques (Macaca fuscata) at Arashiyama, Japan. Int. J. Primatol. 22: 495–520.

Whittig, L. D. (1965). X-ray diffraction techniques for numerical identification and mineralog-ical composition, In Black, C. A. (ed.), Methods of Soil Analysis, Madison, WI, AmericanSociety of Agronomy, pp. 671–696.

Wilson, M. J. (2003). Clay mineralogical and related characteristics of geophagic materials.J. Chem. Ecol. 29: 1525–1547.

Wink, M., Hofer, A., Bilfinger, M., Englert, E., Martin, M., and Schneider, D. (1993). Geeseand plant dietary allelochemicals-food palatability and geophagy. Chemoecology 4: 93–107.

Wrangham, R. W., Chapman, C. A., Clark-Arcadi, A. P., and Isabirye-Basuta, G. (1996). So-cial ecology of Kanyawara chimpanzees; implications for understanding the costs of greatape groups. In McGrew, W. C., Marchant, L. F., and Nishida, T. (eds.), Great Ape Soci-eties. Cambridge University Press, Cambridge, UK, pp. 45–57.

Soils consumed by chimpanzees in the Kibale Forest, Uganda

Documents