J. Zool., Lond. (1999) 247, 91±103 # 1999 The Zoological Society of London Printed in the United Kingdom Geophagy in the golden-faced saki monkey (Pithecia pithecia chrysocephala) in the Central Amazon E. Z. F. Setz1,2, J. Enzweiler3, V. N. Solferini4, M. P. AmeÃndola4 and R. S. Berton5 1 Depto. Zoologia, IB, Universidade Estadual de Campinas, CP 6109, 13083-970 Campinas, SaÄo Paulo, Brazil 2 Depto. Ecologia, IB, Universidade Estadual Paulista, Rio Claro, SaÄo Paulo, Brazil 3 Inst. de GeocieÃncias, Universidade Estadual de Campinas, Campinas, SaÄo Paulo, Brazil 4 Depto. GeneÂtica, IB,Universidade Estadual de Campinas, Campinas, SaÄo Paulo, Brazil 5 SecËaÄo de Fertilidade do Solo e NutricËaÄo de Plantas, Instituto AgronoÃmico, Campinas, SaÄo Paulo, Brazil (Accepted 29 April 1998) Abstract The golden-faced saki monkey Pithecia pithecia chrysocephala (Cebidae, Primates) was observed eating soil from termite nests during a long-term study of a family group in a Central Amazonian forest fragment. In this paper we describe the behaviour involved in the geophagy in these monkeys, and the results of geochemical and physical analyses of the termite nest material, as well as root mat and topsoil samples below the trees, in order to clarify the possible reasons for it. The sakis ate soil from nine arboreal termite nests on 26 soil feeding-bouts (in 853 observation hours); 25 soil feeding-bouts occurred in March 1987 (rainy season), during 19 days or 132 observation hours, and occupied 0.7% of the feeding time. Geophagy frequencies did not differ between sexes (17 feeding-bouts of four females and 8 for two males). Mineral composition was higher in arboreal termitaria than in the topsoil. Kaolinite was the major clay component. Tannin adsorptive capacity, tested through a modi®ed radial diffusion method of Hagerman, was around 10±20%, similar to a control with kaolin (10±20%), but lower than bentonite or celite (30±45%). The observations reported here, although inconclusive as to the function of geophagy in this species, indicate that it is not a mineral supplement during times of scarcity or high consumption of leaves, as has been reported for other primates, nor that it is related to fruit consumption (redressing possible mineral imbalance), as has been suggested for some other frugivorous mammals. Our results do not rule out tannin adsorptive hypothesis for the ingestion of clays, but, being an irregular habit, we argue that it is most likely related to rare and occasional dietary components. Key words: arboreal termitaria, Central Amazon, geophagy, Pithecia, rain forest, tannin adsorption, Trichilia INTRODUCTION Mahaney et al., 1993, 1995; Knezevich, 1998); (3) antacid action of clays (Morris, 1927 in Poirier, 1970) or The consumption of soil, geophagy or pedophagy, has adjustment of pH in the stomach forechamber (Oates, been reported for reptiles (Marlow & Tollestrup, 1982), 1978); (4) tactile sensations in the mouth (Hladik & birds (Emmons & Stark, 1979, Mayer & Brand, 1982; Gueguen, 1974); (5) tradition (Mahaney et al., 1990). In Izawa, 1993), and mammals (Weeks & Kirkpatrick, humans, geophagy is also related to certain parasitic 1976, 1978; Oates, 1978; Mayer & Brand, 1982; Izawa, infestations, as a source of iron to counteract anaemia 1993). The principal functions have been attributed to caused by ancylostomiasis (Neves, 1991). (1) mineral supplementation (Mayer & Brand, 1982; However, the function of this behaviour is still not Vermeer & Ferrell, 1985; Davies & Baillie, 1988; known, and may vary from species to species, and even Mahaney, Watts & Hancock, 1990; Heymann & within a species it may serve different functions at Hartmann, 1991; John & Duquette, 1991 in Mahaney, different times (Davies & Baillie, 1988; Heymann & Aufreiter & Hancock, 1995); (2) adsorption of plant Hartmann, 1991; Izawa, 1993). tannins and toxins (Hladik, 1977a, b; Oates, 1978) Mineral supplementation may be related to diets with and/or counteraction of gastric upsets or diarrhoea high foliage content year-round or at times of food (Said, Shibel & Abdullah, 1980; Vermeer & Ferrell scarcity. Sodium is most frequently suggested as the 1985; John & Duquette, 1991 in Mahaney et al., 1995; target element (see Weir, 1972; Mayer & Brand, 1982). 92 E. Z. F. Setz ET AL. At times of fruit abundance, however, potassium/ menting minerals (Neves, 1991) and counteracting sodium ratios can reach excessive values because of the diarrhoea and intestinal problems (Knezewich, 1998). high potassium content in fruits. In this situation, Therapeutic behaviour in non-human primates appears sodium would also be needed (Weeks & Kirkpatrick, also as charcoal consumption to reduce dietary toxins 1976, 1978; but see Hladik, 1977b, and Sick, 1949). In the by red colobus monkeys (Struhsaker, Cooney & Siex, tortoise, calcium appears to be the target for long-term 1997) and as medicinal plant use against a strongyle geophagy (Marlow & Tollestrup, 1982); while in parrots nematode by chimpanzees (Huffman et al., 1997). and macaws sodium and magnesium are also implicated New World primates (Saguinus mystax, Alouatta (Emmons & Stark, 1979). At high altitudes, iron is seniculus, A. caraya, Ateles belzebuth, Chiropotes albi- considered the main candidate for African buffalo nasus, Callicebus personatus melanochir) eat soil (Izawa, (Mahaney, 1987), and iron and sodium are considered 1975, 1993; Heymann & Hartmann, 1991, Bicca- the key elements for the mountain gorilla (Mahaney, Marques & Calegaro-Marques, 1994; Ferrari, 1995; Watts et al., 1990). Iron is also the main available MuÈller, Ahl & Hartmann, 1997). Heymann & element for chimpanzee in Kibale (Mahaney, Milner Hartmann (1991) suggest that mineral supplementation et al. 1997). For macaques Macaca mulatta, however, is the important factor, while Izawa (1993) arrived at no the iron content of leaves and fruits is believed to cover decisive conclusion and MuÈller et al., (1997) support their physiological requirements, and clay ingestion may adsorption of plant toxins. be important as an adsorbent of plant tannins Here we report on geophagy by the golden-faced saki (Lindburg, 1977). monkey Pithecia pithecia chrysocephala. We describe The tannin adsorption hypothesis originated from this behaviour and investigate its possible mineral observations on the consumption of clay with acorns supplementation and tannin adsorption functions with and potatoes in human populations (Hladik, 1978; regard to the monkeys' feeding ecology and life in a Johns, 1986). Both adsorption of plant tannins and forest fragment. We compare geochemistry and toxins and/or counteraction of gastric upsets or diar- mineralogy of two termite nests and nearby surface soil rhoea are related to kaolinite (Vermeer & Ferrell, 1985) samples, and test their tannin adsorptive capacity. Our or halloysite, a hydrated form of kaolinite clay (see aim is to determine if potential chemical or mineral Mahaney, Aufreiter et al., 1995). The lattice structure of stimuli are present in the samples studied, which can kaolinite is believed to function as an adsorbent for explain this aspect of the sakis' feeding behaviour. toxins and bacteria (Said et al., 1980) and has been reported to form a protective coat on the mucous membrane of the digestive tract (Swinyard, 1965 in METHODS Vermeer & Ferrell, 1985). Smectite (montmorillonite) swells when wet, and halloysite may have a similar effect Study site because of its capacity to adsorb molecular water The study was conducted in a 10 ha elongate forest (Mahaney, Aufreiter et al., 1995). fragment known as `Colosso', along a creek ravine The antacid function of ingested soil may derive from (2825'S, 59850'W) of the Biological Dynamics of Forest both the adsorptive properties of clay minerals and the Fragments Project of the Smithsonian Institution and buffering capacity of the clay fraction's exchange capa- the National Institute for Amazon Research (INPA), city, and from a higher pH of the soil. In some primates 80 km north of Manaus, Amazonas, Brazil (see Lovejoy with sacculated stomachs, the adsorption of fatty acids & Bierregaard, 1990). from rapid anaerobic fermentation can prevent fatal Soils from the region are clayish, and classi®ed as Red `acidosis' (Goltenboth, 1976, in Davies & Baillie, 1988). Yellow and Alic Yellow Podzolic with high clay content Also, a higher pH from the ingested earth helps maintain and poor in nutrients (Ranzani, 1980; Chauvel, 1982). a higher pH in the stomach forechamber, where fermen- The dominant clay fraction is comprised of kaolinite tation occurs (Oates, 1978; see also Morris, 1927, in (80%), gibbsite (10%) and goethite (Chauvel, 1982). Poirier, 1970). This higher pH also can enhance nutrient The vegetation is lowland terra ®rme tropical rain availability (Sanchez, 1976, in Davies & Baillie, 1988). forest. Mean annual rainfall is 2500 mm and mean Tactile sensation in the mouth was suggested by temperature is 24.8 8C. The rainy season extends from Hladik & Gueguen (1974) after investigating thoroughly November to May, with a dry period from June to either the soil or the diet for all mineral contents. October. Flowering peaks in October±November and The hypothesis of tradition appeared to explain the fruiting in January. Both ¯owers and fruits are scarce use of soil by one group of mountain gorillas (Mahaney, for the rest of the year (Setz, 1993). Watts et al., 1990). No other gorilla group studied ate soil at all. Geophagy correlation with endoparasitism is rooted Subjects in Brazilian (Afro-Brazilian?) folklore. People say that if someone eats soil, it is a sign that he has worms, but we The study group of golden-faced saki monkeys P. p. could ®nd nothing in the literature besides Vermeer & chrysocephala consisted of 6 or 7 individuals. Most were Frate's (1979) study in rural Mississippi. In this case, habituated to the presence of human observers in 1985, geophagy could be a therapeutic mediator supple- when a long-term study on their ecology was begun  Geophagy in golden-faced sakis 93 Table 1. Hours of observation on the ecology of golden-faced (1979), Ranzani (1980), Chauvel (1982), Davies & Baillie sakis in a forest fragment in Central Amazon (1988), Izawa (1993). 5 cm3 of soil were leached with 1985 1986 1987 1988 1989 1990 Total 50 ml of ammonium acetate at pH 7. P was extracted by ion exchange resin (Amberlite IRA-400 and IR-120, Jan 45 45 0.4 mm mesh; Raij & Quaggio, 1983), and organic Feb X 2 4 6 matter was determined by the colorimetric method. Mar 132 4 136 Two sample preparations were used for metal deter- Apr 15 15 30 minations. In one set, samples were dried for 2 h at May 4 4 Jun 4 4 1 9 110 8C then digested with 50 ml 2 M HCl at 30 8C. After Jul 133 14 2 149 4 h, samples were centrifuged and solutions were made Aug 2a 55 57 up to a ®nal volume of 50 ml with distilled and deio- Sep 1a 55 3 16 75 nized water. Another sample was ignited at 800 8C for 3 Oct 6a 28 37 33 104 h. A portion of the sample was digested by heating on a Nov 9 33 34 59 135 hotplate in 10 ml of concentrated HNO3. After centri- Dec 27 66 10 103 fugation samples were brought to 50 ml ®nal volume. Total 45 133 408 45 90 132 853 Na, K, Mg, Zn, Cu, Co and Ni were determined by a atomic absorption spectrometry using an air-acetylene Saki habituation. ¯ame, and Ca and Al using nitrous oxide-acetylene. Ionization of Na, K, Ca and Al was suppressed by (Setz, 1993). In 1987, the group consisted of 1 adult diluting samples and adding an excess of caesium male, 2 adult females, 1 young female and 2 one-year- chloride (®nal concentration 1000 mg ml71). A Varian old juveniles (a female and a male). Since golden-faced AA 1475 atomic absorption spectrometer was used for sakis are sexually dichromatic from birth (Hershkovitz, all determinations. 1987), sex was easily recognized. For X-ray diffractometry, samples were prepared according to a modi®ed procedure after Jackson (1969; Appendix). Data collection We followed the group throughout the day (from 06:00 Tannin adsorption analysis to 16:00: Setz, 1993). Systematic data on activity budgets and the diet were collected by scan sampling at The adsorptive capacity of tannins by earth samples was 10 min intervals. All occurrences of geophagy were also determined by comparing tannic acid (Ecibra 0180) recorded as soil feeding-bouts (sfb) or the length of time solutions with the same solutions added to earth, using a single animal was recorded eating soil from a single the radial diffusion method of Hagerman (1987). In this termite nest (de®nition adapted from Klein & Klein, assay, a visible precipitation ring develops as tannin 1974), including the identity of the individual, time, and from the solution being tested diffuses from a well into a location of the termite nest. Total sampling time from protein-containing agar slab. The amount of tannin in 1985 to 1991 was 853 h (Table 1). Information on diet, the solution is proportional to the area of the precipita- time budgets and use of space by the golden-faced sakis, tion ring. 1 ml of tannic acid solution (0.05 M, 0.1 M and as well as vegetation analyses of the forest fragment, are 0.2 M) was added to 625 mg of each sample. For presented elsewhere (Setz, 1993). comparative purposes, the procedure was repeated We collected 2 earth samples from each of 2 arboreal using kaolin (Reagen 10317), bentonite (Sigma B3378) termite nests (Nos 9A and 16AE; Fig. 3) where the sakis and celite (Merck 365). For bentonite and celite, a had been feeding, one from the upper part and the other sample of 39 mg with diluted tannic solutions (400 ml from the lower part of the nest. A sample was also tannic acid solution and 6 ml water) was used, because collected from the root mat in the surface (0±5 cm of the great liquid adsorption of these compounds. The depth) and another from the topsoil (5±15 cm) near the same procedure was used for kaolin as for the earth base of the tree with the nest. The samples were air- samples. After 48 h, 300 ml of the supernatant was dried in the ®eld, and oven-dried (< 40 8C) in the diluted with 700 ml of acetone. After 3 h the solution laboratory, ground up in a mortar and sieved (2 mm was applied to the plate wells (3 successive 8 ml mesh). Termites were identi®ed by Eliana M. Cancello, aliquots). Petri dishes were covered, sealed with and the ants by Carlos Roberto BrandaÄo, at the Para®lm and incubated at 30 8C. Precipitation ring Zoology Museum of the University of SaÄo Paulo diameters were measured after 48 h. (MZUSP ± col. no. 9550). RESULTS Geochemical and physical analyses Frequency and seasonality of geophagy Ammonium acetate-extractable cations (Na, K, Ca, Mg) were determined as in Stark (1970), Emmons & Stark Soil intake from nine arboreal termite nests was 94 E. Z. F. Setz ET AL. Behaviour during geophagy In all soil feeding-bouts observed, the saki descended to about 2 m from the ground, and using a tree or a liana near the nest, broke a bit off the termite nest and ate it (Fig. 1). Eighty per cent of the soil feeding-bouts lasted less than 3 min, and the longest took 6 min. On one soil feeding-bout, a female saki took two bits from the same termite nest (11 March, nest No. 2.5C: Figs 2 & 3). Only once (12 March, Fig. 2) did all six sakis eat soil in a 5-minute span, but on this occasion from two different termite nests (Nos. 16AE ± 4 sfb and W ± 2 sfb: Fig. 3). On 22 March six soil feeding-bouts occurred on four different occasions, over 3 h on three different termite nests (Fig. 2). Gender differences There were no gender differences in the frequency of geophagy. The four female sakis (66% of the six sakis) were responsible for 17 soil feeding-bouts (68% from 25). The average frequency of soil intake was 0.03 soil feeding-bouts per individual per observation hour, during March 1987 (n = 132 obs. h: Fig. 2). Although the young male was recorded more than the adult (6 vs 2 sfb), age considerations are problematic because of records of unrecognized females (11 sfb). 1m Association of geophagy with other foods Fifteen (60%) soil feeding-bouts were preceded by fruit ingestion (pulp or seeds were the last things they had fed on, 1±17 min earlier), from six different plant species: Trichilia cf. quadrijuga (Meliaceae±8 sfb), Inga sp. (Mi- mosaceae±2 sfb), Hevea guianensis (Euphorbiaceae±2 sfb), Licania apetala (Chrysobalanaceae±1 sfb), Du- Fig. 1. Arboreal termite nest used by golden-faced sakis guetia latifolia (Annonaceae±1 sfb), Vitex sp. (drawn from photographs). (Verbenaceae±1 sfb). The saki which ate a fruit of Vitex sp. 2 min before soil, had eaten Trichilia cf. quadrijuga observed during 26 soil feeding-bouts (Fig. 1) during 5 min earlier. Apart from Duguetia, which also 853 observation hours (0.03 sfb/obs. h). Twenty-®ve soil appeared in the diet of the sakis in April 1987, these feeding-bouts occurred in March 1987 (Fig. 2), com- fruit species were used only in March 1987. prising 0.7% of the feeding time (six scan records), Ten (40%) soil feeding-bouts occurred after ingesting estimated by scan sampling during 132 observation new leaves from at least three different species: Clarisia hours (0.19 sfb/obs. h) in nine (47%) from 19 obs. days. ilicifolia and Sorocea muriculata (both Moraceae, each Nine soil feeding-bouts took place from 08:00 to 09:00 1 sfb) and Eschweilera tessmannii (Lecythidaceae, 1 sfb). (36%) and nine others, from 11:00 to noon (36%). In The leaves of both Moraceae species were used by the November 1990, another soil feeding-bout was observed sakis in other months (December 1987 and July 1986, (D. A. Gaspar, pers. comm). respectively), but they were not followed by geophagy. March is in the late wet season when fruit is abundant Trichilia fruit pulp was one of the seven most con- (10.9% of the species bearing fruits vs 5.6% in the dry sumed items in the rainy season (49 scan records or season: Setz, 1993). Soil feeding was not observed in 5.5% of the feeding-time, about 80 fruit feeding-bouts: July 1986 (dry season, n = 133 obs. h, 19 days), when the Setz, 1993). On the Trichilia trees, three of the six sakis sakis occasionally ate mature leaves (only 2% of the would feed at the same time and alternate with the diet), nor during an another 10-month observation others, giving a considerably higher number of Trichilia period at the same site in 1987 (n = 408 obs. h, Table 1), fruit feeding-bouts than scan records. The high fre- nor later during periodic observations in 1989 and 1990 quency (36%, 9 from 25 sfb) of soil ingestion after (n = 222 obs. h: Table 1). Trichilia is probably also related to the fact that the  Geophagy in golden-faced sakis 95 7 6 Soil feeding-bouts per day 5 4 3 2 1 0 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 Day of March 1987 Fig. 2. Relation between soil feeding-bouts and sampling day for a habituated group of golden-faced saki monkeys near Manaus, Brazil. Apart from 13 March, when the sakis were not observed, 19 days summed 131.7 hours of observation. three Trichilia cf. quadrijuga trees used were near the were old. No termites were found when removing pieces four most visited termite nests (Nos. 9A, 2.5C, 16AE up to a depth of 3 cm. Termites were found only in the and 16AW: Fig. 3). area of direct nest/trunk contact. We compared days with and without ingestion of termite soil to days where each of the six most- consumed plant species (Helicostylis tomentosa, Termite species Moraceae ± pulp and immature seeds Theobroma subincanum, Sterculiaceae ± pulp only; Inga alba, Mimo- One termite nest (No. 9A) was occupied by Nasutitermes saceae ± pulp only; Fusaea longifolia, Annonaceae ± sp. (Termitidae, Nasutiterminae) and the ant Dolicho- seeds only; Trichilia cf. quadrijuga ± pulp and seeds; derus laminata (Mayr) (Formicidae, Dolichoderinae), Passi¯ora nitida, Passi¯oraceae ± pulp; seeds were ob- and the other (No. 16AE), by an undescribed species of served in faecal samples) were eaten or not through a Cavitermes (Termitidae, Termitinae). Nasutitermes chi-square test. Only Trichilia gave a signi®cant value builds its nests on or inside tree trunks, in the subsoil or (w2 = 4.23, P < 0.05, d.f. = 1), showing a positive daily at the soil surface, while Cavitermes occurs in nests built co-occurrence. Fruit from another Trichilia species by other species, where they feed on the organic soil of (T. micrantha) was eaten by the sakis in October 1990. the nest (Mathews, 1977). Faecal samples Topsoil Faecal samples collected opportunistically early in the In the study area, there were exposed soils at the bases morning in March 1987 (n = 9 from 5 days), as well as of trees uprooted by wind, and creek banks, which, during other observation periods (n = 7 from 4 days, although as accessible as the termite mounds, were July 1986; n = 10 from many days, April±July 1987), never exploited by the sakis during the long-term study showed no signs of diarrhoea (E. Setz, unpubl. data). period. Digging samples of soil near the trees with the No parasite investigation was performed. termitaria was dif®cult because of the dense, at least 5-cm-thick root mat under the leaf litter and above the soil surface. Termite nests The nine termite nests visited were on tree trunks at a Geochemistry of termitaria, root mat and topsoil height of 2 m. They were oval-shaped (about 40 cm width 6 100 cm length and with the long axis aligned All the earth samples were acid, with pH values (in to the tree trunk: Fig. 1) and dark brown, almost black water) between 4.0 and 4.8. Termite nest soils presented in colour. The termite-nest soils were hard and dry, and a total extractable cation and a cation exchange capacity judging by the washed-out look of their surface and the (CEC) considerably higher than the topsoil, and the absence of recently deposited soft soil, the termitaria root mat presented intermediate results (Table 2). The 96 E. Z. F. Setz ET AL. Termite nest site 18 A 16 AW 16 AE* 2.5 C 14.5 A 13.5 A 11.5 A 100 m 9A* 1 AW 0 1 2 3 4 5 6 Soil-feeding bouts Fig. 3. Relation between saki soil feeding-bouts and termite nest, showing also the geographical distribution of the nests in the forest fragment. Table 2. pH, extractable cations and cation exchange capacity of earth from arboreal termite nests consumed by Pithecia pithecia chrysocephala, and of root mat and topsoil samples under the termite nest trees. n, number of samples analysed (see text) Earth sample Termite nests (n = 4) Root mat (n = 2) Topsoil (n = 2) pH, H2O 4.3 ‹ 0.3 4.1 ‹ 0.0 4.1 ‹ 0.1 pH, CaCl2 3.5 ‹ 0.4 3.2 ‹ 0.0 3.6 ‹ 0.1 P mg/cm3a 38 ‹ 12 19 ‹ 3 12 ‹ 2 Na meq/100 cm3b 0.05 ‹ 0.04 0.03 ‹ 0.01 0.02 ‹ 0.01 K meq/100 cm3b 0.26 ‹ 0.06 0.13 ‹ 0.06 0.004 ‹ 0.003 Ca meq/100 cm3b 0.55 ‹ 0.13 0.07 ‹ 0.04 0.03 ‹ 0.03 Mg meq/100 cm3b 0.54 ‹ 0.19 0.20 ‹ 0.0 0.05 ‹ 0.0 Total extractable cations 1.7 ‹ 0.2 0.7 ‹ 0.3 0.2 ‹ 0.0 H + Al meq/100 cm3 37.1 ‹ 6.3 20.7 ‹ 3.0 10.5 ‹ 2.3 CEC 38.72 ‹ 6.3 21.3 ‹ 3.3 8.0 ‹ 2.3 Base saturation % 4.2 ‹ 1.1 2.6 ‹ 0.1 3.2 ‹ 3.1 a By ion exchange resin (Amberlite IRA-400 and IR-120; Raij & Quaggio, 1983). b By leaching 5 cm3 of soil with 50 ml ammonium acetate at pH 7. Table 3. Organic matter (colorimetric method; meq/100 cm3) termitaria pH was equally acid. Termite nest No. 9A of earth from two arboreal termite nests, consumed by had more organic matter than both root mats, which in Pithecia pithecia chrysocephala, and of root mat and topsoil samples under the respective termite nest trees turn had more than topsoil (Tables 2 & 3). Termite nest No. 16.5AE had an organic matter content similar to Earth sample the root mat (Table 3). On average, termite nest samples presented higher concentrations of P, K, Ca, Mg, Mn Nest No. Termite nests Root mat Topsoil and Fe, besides Al and H+, compared with topsoil 9A Superior 18.5 (Tables 2 & 4). Na, Co, Zn, Cu and Ni contents were 11.1 6.0 similar between termitaria and topsoil (Table 4). Inferior 18.5 Analyses with prior ignition showed similarly equal 16.5A E Superior 8.4 results between the three substrates (Table 4), with the 9.1 3.8 exception of Mg and Ca for the termite nest samples Inferior 8.4 (Table 4).  Geophagy in golden-faced sakis 97 Table 4. Mineral composition of earth from arboreal termite Mineralogical contents and tannin adsorbent properties nests consumed by Pithecia pithecia chrysocephala, and of root mat and topsoil samples under the termite nest trees. n, X-ray diffractograms showed that the clay mineral number of samples analysed (see text). Analysis by ¯ame atomic absorption spectrophotometer (Intralab AA12/1475) fraction is composed mainly of kaolinite, with traces of after digestion in HCl/HNO3. Numbers in parenthesis gibbsite, feldspar, plagioclase and quartz. included previous ignition at 800 8C for 3 h Ring diameter measures from the termite nest, the topsoil and root mat samples indicated an adsorptive Earth sample capacity around 10±20% (Table 5), similar to control Termite nests Root mat Topsoil with kaolin (10±20%), but lower than bentonite or celite (n = 4) (n = 2) (n = 2) (30 to 45%). In comparisons using ANOVA and Tukey (Wilkinson, 1991), with tannin concentration as a co- Element mg/g mg/g mg/g variable, ring measures among samples (F = 102.31, Na 48 ‹ 33 43 ‹ 18 16 ‹ 5 P < 0.0001) and tannin concentrations (F = 958.10, (93 ‹ 28) (70 ‹ 17) 80 ‹ 21) P < 0.0001) differed signi®cantly. Termite nest No. Mg 90 ‹ 41 45 ‹ 8 18 ‹ 3 16.5A E samples adsorbed signi®cantly less than (56 ‹ 16) (46 ‹ 6) (33 ‹ 3) samples from termite nest No. 9A (pairwise mean Al 916 ‹ 197 487 ‹ 62 447 ‹ 54 difference =70.391, P = 0.01), but neither differed (18158 ‹ 13950) (18399 ‹ 4544) (20677 ‹ 6930) signi®cantly from the topsoil and root mat samples or K 136 ‹ 27 84 ‹ 54 31 ‹ 16 kaolin. Rings from both termite nest samples were (156 ‹ 35) (129 ‹ 38) (103 ‹ 10) Ca 221 ‹ 41 61 ‹ 52 18 ‹ 4 signi®cantly smaller than the tannin control (differ- (60 ‹ 14) (33 ‹ 34) (7 ‹ 1) ence = 1.11 to 1.30, P < 0.0001), showing tannin Mn 9‹3 5‹1 1‹0 adsorptive capacity, but signi®cantly larger than those (13 ‹ 2) (9 ‹ 1) (7 ‹ 2) for bentonite and celite (difference = 1.41 to 2.08, Fe 1094 ‹ 90 565 ‹ 112 499 ‹ 75 P < 0.0001). Bentonite and celite were better tannin (4257 ‹ 419) (2679 ‹ 497) (3336 ‹ 906) adsorbents. Co 3‹1 3‹0 3‹1 (10 ‹ 2) (8 ‹ 3) (8 ‹ 1) Zn 26 ‹ 15 24 ‹ 25 18 ‹ 10 (17 ‹ 4) (12 ‹ 3) (17 ‹ 14) DISCUSSION Cu 0.3 ‹ 0.3 0.8 ‹ 1.0 0.4 ‹ 0.5 (6 ‹ 1) (4 ‹ 1) (6 ‹ 4) Whereas some primates consume soil exposed by up- Ni 5‹1 5‹1 4‹1 turned trunks (Indri indri ± Pollock, 1977), stream banks (11 ‹ 1) (10 ‹ 0) (10 ‹ 1) (Colobus guereza ± Oates, 1978), subsoil sediments Cr < 4 (< 4) < 4 (< 4) < 4 (< 4) (Gorilla gorilla beringei ± Mahaney, Watts et al., 1990; Table 5. Average precipitation ring diameter, standard deviation and percentage of ring reduction (red) for arboreal termite nests, topsoil and root mat (n = number of samples 6 number of replications), in relation to the precipitation ring observed for tannin control samples, for each tannic acid concentration used Tannic acid concentration (M) 0.05 0.1 0.2 Precipitation ring precipitation ring precipitation ring Diameter Red Diameter Red Diameter Red (mm) (%) (mm) (%) (mm) (%) Tannin 6.6 ‹ 0.24 ± 7.6 ‹ 0.33 ± 9.7 ‹ 0.55 ± (n = 24) (n = 24) (n = 20) Termite nests 5.8 ‹ 0.35 12 6.5 ‹ 0.51 14 8.2 ‹ 0.80 15 (n = 268) (n = 268) (n = 268) 5.3 ‹ 0.24 20 6.2 ‹ 0.43 18 7.9 ‹ 0.59 19 (n = 268) (n = 268) (n = 268) Topsoil 5.6 ‹ 0.42 15 6.5 ‹ 0.30 14 8.2 ‹ 0.36 15 (n = 268) (n = 268) (n = 268) Root mat 5.6 ‹ 0.55 15 6.0 ‹ 0.76 21 8.2 ‹ 0.55 14 (n = 8) (n = 7) (n = 8) Kaolin 5.8 ‹ 0.49 12 6.9 ‹ 0.32 9 7.6 ‹ 1.37 22 (n = 12) (n = 4) (n = 4) Celite 4.6 ‹ 0.25 30 4.9 ‹ 0.25 35 5.4 ‹ 0.32 44 (n = 4) (n = 4) (n = 8) Bentonite 4.8 ‹ 0.29 27 4.9 ‹ 0.25 35 5.3 ‹ 0.20 45 (n = 4) (n = 8) (n = 4) 98 E. Z. F. Setz ET AL. Pan troglodytes ± Mahaney, Milner et al., 1997), natural exchange capacity and a lower total extractable content salt licks (Alouatta seniculus and Ateles belzebuth ± than samples from 0 to 10 cm depth analysed by Izawa 1993), and plastic clayish soil after rains (Pan Chauvel (1982), average values for A horizon samples troglodytes ± Hladik, 1973), Pithecia used soil from analysed by Ranzani (1980), or 0±3 cm depth podzol termite nests, as has been reported for a number of sands analysed by Stark (1970). Cation content (Na, K, primates (Pan troglodytes ± Hladik, 1973 and Ca, Mg) was lower (Ranzani, 1980; Chauvel, 1982) or Wrangham, 1977, Macaca mulatta ± Lindburg, 1977, similar, except for potassium which is lower than in Alouatta seniculus and Ateles belzebuth ± Izawa, 1993). podzol sands (Stark, 1970). Our samples were similar to Other animals have also been seen to eat other types of the 10±20 cm depth samples of Chauvel (1982), except insect constructions such as homopteran larval `chim- for total extractable cations and Ca content (both lower neys' (Pan troglodytes ± Hladik, 1973) and leaf-cutter in our samples). ant mounds (Chacoan peccary Catagonus wagneri ± Mayer & Brand, 1982, Saguinus mystax ± Heymann & Hartmann, 1991, Callicebus personatus melanochir ± MuÈller et al., 1997). The `earth that has been moulded Evidence for geophagy as mineral supplement by insects' (Hladik, 1973) usually has a higher mineral content than topsoil (Salick, Herrera & Jordan, 1983), There was a consistent pattern of mineral concentration, as was observed here. with termitaria having the highest values, followed by As pointed out by MuÈller et al. (1997), even if soils the root mat, and the soil surface (Salick, Herrera & did not differ markedly in quality, it would be advanta- Jordan, 1983; Table 3). The relative cation concentra- geous for monkeys to obtain soil from an ant mound (or tion in termite nests tends to be greater on poor soils termite nest) rather from the forest ¯oor, because of the (Pomeroy, 1983 in Davies & Baillie, 1988; see also lower risk of predation and the soil's clod structure Izawa, 1993) and this clearly applies to `Colosso'. Both (cemented together by insects). termite nests examined, although arboreal, presented Unlike termitaria used by spider monkeys and localized mineral concentrations in an intensely leached howlers, which were built by Constrictotermes (Izawa, environment, as was also found by Salick et al. (1983). 1993), sakis consumed soil from Nasutitermes nests, Termite or ant nest soil consumption in these circum- indicating that these nests are not unpalatable, as has stances provides a concentrated source of scarce been proposed for Labiotermes (Izawa, 1993). Lower minerals (Emmons & Stark, 1979) in a generally oligo- organic matter of nest No. 16.5AE is probably related trophic environment (but see Mayer & Brand, 1982). to Cavitermes feeding on the organic soil of the host High iron (Ripley, 1970 in Hladik, 1977a; Lindburg, mound (Mathews, 1977). 1977; Mahaney et al., 1990, 1993, 1997), sodium and As was found for the lemurs Indri indri, where bromine content (Mahaney, Watts et al., 1990) may be Pollock (1977) recorded between ®ve and ten such important in primate geophagy. Calcium, potassium earth-feeding sites in each group (2±6 individuals) terri- and magnesium are all better supplied by the mound tory, sakis from our study group ate soil from nine soils, although there is no information on whether any termitaria. A rhesus macaque Macaca mulatta group of these minerals were de®cient in the sakis' highly (141 monkeys) used ten mine sites (Knezevich, 1998). diverse diet (Setz, 1993). Unlike chimpanzees and indris, however, saki group In three primate studies, analysis of soil samples members did not usually feed sequentially at a speci®c failed to show high concentrations of speci®c minerals spot or termitaria that they would return to several days (Pollock, 1977; Izawa, 1993; MuÈller et al., 1997). The later (Pollock, 1977; Hladik, 1977b). elements, which might have some nutritive value for The spatial association of geophagy and food (like chimpanzee, langur or gorilla, occur at lower levels in the three most used termite nests and Trichilia trees the earth samples than in many common food plants observed here) was also observed for rhesus macaques (Hladik, 1977a, b; Mahaney, Watts et al., 1990). Soils (preferred mines and monkey chow corrals: Knezevich, ingested by other primates had higher or much higher 1998). However, temporal association, like alternate concentrations of Na, K, Ca or Mg (Davies & Baillie, consumption of chow and soil in rhesus macaques 1988) and Mn, Fe, Zn or Cu (Hladik, 1977a, b; Oates, (Knezevich, 1998), was not observed in sakis. 1978; Davies & Baillie, 1988). Besides this, higher cation exchange capacities (not observed here) can interfere with the absorption of iron (Minnich et al., 1968 in Topsoil geochemistry Vermeer & Ferrell, 1985; but see Mahaney, Aufreiter et al., 1995), even contributing to the development of Typically in the Central Amazon, soil ¯oor samples are anaemia (Minnich et al., 1968 in Underwood, 1977). poor in soluble nutrients (Stark, 1970; Ranzani, 1980; Contrary to observations by Davies & Baillie (1988), Chauvel, 1982). Compared with other surface soil termitaria eaten by the sakis also had higher concentra- samples (0±20 cm) in the same region (30 km from our tions of aluminum than adjacent soil samples. Clay site), our soil samples had a lower (Stark, 1970), similar ingested by the South Indian elephant and the Nilgiri (Ranzani, 1980) or higher pH in water (Chauvel, 1982). langur also showed a high concentration of aluminum Our soil also had a lower to almost lower cation (see Poirier, 1970). Such high aluminum levels are  Geophagy in golden-faced sakis 99 Table 6. Reported geophagy frequency for primates and peccaries. Fo, folivores; Fr, frugivores; See, seed eaters; Om, omnivores. Common name Main diet Event frequency No. of study months Reference Nilgiri langur Fo 5 12 Poirier, 1970 Black colobus Fo Occasional 9 Harrison & Hladik, 1986 Red leaf monkey Fo See 9 13 Davies & Baillie, 1988 Moustached tamarin Fr 3 (dry season) 3 (345 h) Heymann & Hartmann, 1991 Titi monkey Fr 14 times (warm season) 11 (1030 h) MuÈller et al., 1997 Saki monkey Fr 26 times (rainy season) >15 (853 h) This study Howler monkey Fo 40% of obs. days 203 days Izawa, 1993 Spider monkey Fr ± ± Izawa, 1993 Indri lemur Fo Fr 75% of obs. days Pollock, 1977 Rhesus macaque Fr Fo Once a month 12 Lindburg, 1977 (dry season) Chow Common Sultana & Marriot, 1982 Routinely 3 Knezewich, 1998 Japanese macaque Considerable time Inoue, 1987 in Izawa, 1993; Mahaney, Hancock et al., 1993 Mandrill Fr Rare (incidental?) Lahm, 1986 Chimpanzee Om Once a day Wrangham, 1977 Twice a day Hladik, 1977b Mountain gorilla Fo 5 to 6 times/year Watts, 1984 Once a day (dry season) Mahaney et al., 1990, 1993,1995 Chacoan peccary Fr 43% of obs. days (winter) Mayer & Brand, 1982 poorly absorbed (Underwood, 1977) and may be fruiting season in P. p. chrysocephala could point to immobilized by the clay structure (Mahaney. Aufreiter excess potassium, but potassium content was higher et al., 1995). than sodium in the termite nest earth (Table 3). Sodium Unlike mountain goats, where females use natural content was lower in our samples than in soil ingested salt licks during postpartum stress (Hebert & Cowan, by other primates (Hladik, 1977a, 1977b, Mahaney, 1971), saki geophagy did not show any gender difference Watts et al., 1990) and soil samples from salt licks in frequency or relation to reproductive cycle (see Setz (Weeks & Kirkpatrick, 1976; Emmons & Stark, 1979). & Gaspar, 1997). This was expected since New World Soil samples ingested by chimpanzees in the wet season primates do not have menstrual cycles (Flowerdew, in Gabon had higher quantities of potassium (and 1987). Besides, adult females were not pregnant or calcium) than their diet (Hladik, 1977b; Uehara, 1982). lactating for dependent infants in March 1987 (feeding juveniles was observed sporadically: E. Setz, pers. obs.), when most soil feeding-bouts (96%) were observed. Evidence for geophagy as antacid Monotonous diets or species-poor diets might be expected to be de®cient or unbalanced for some nutri- As for regolith analysed and discussed by Mahaney, ents (Hladik, 1977a; Oates, 1978), but geophagy is Aufreiter et al. (1995), termite soil had 10% Al and 1% reported for primates with diverse diets (Davies & Mg, an elemental composition comparable to commer- Baillie, 1988). Sakis with their highly diverse diet (190 cial antacid preparations (10% Al, 1.1% Mg), its spp. eaten: Setz, 1993) are no exception. Although effectiveness depending on the chemical form of the geophagy is usually related to nutrient-poor and highly elements. folivorous diets, sakis, along with tamarins (Heymann Since sakis do not have sacculate stomachs and the & Hartman, 1991), spider monkeys (Izawa, 1993), termitaria pH was low (acid), we do not discuss here the bearded sakis (Ferrari, 1995), and masked titis (MuÈller soil function as pH buffer of the forestomach, as et al., 1997) are predominantly frugivorous (Setz, 1993; suggested for Colobus guereza (Oates, 1978). see Table 6). Seasonal folivory or mature leaf consumption (Harrison & Hladik, 1986) could favour geophagy in Evidence for geophagy as an adsorbent for tannins and the dry season (Table 6) when fruits are scarce, either as toxins a mineral supplement, or as a secondary compound adsorbent (see Mahaney, Aufreiter et al., 1995). Immature fruits or seeds were consumed equally by the However, in our study sakis ate soil in the rainy season, sakis in both the dry (19.6 and 18.4%, respectively) and when fruits are abundant and leaves are least prevalent rainy (23.4 and 21.7%) seasons (Setz, 1993). Fruit in their diet (4% vs. 18%: Setz, 1993). As in deer, consumption (64.2% dry season vs 95.8% wet season) woodchuck, and fox squirrels (Weeks & Kirkpatrick, and seed predation (33.3% vs 19.6%), however, were 1976, 1978), the correlation between geophagy and the more frequent in the rainy season (Setz, 1993). Our soil 100 E. Z. F. Setz ET AL. samples showed an adsorptive capacity for tannins golden-faced saki, although inconclusive on its function (Table 4). If sakis ingest soil because of high levels of in this species, demonstrate that it does not occur tannin in fruits and seeds in general (see Kinzey & always when food is scarce and is not associated with Norconk, 1993) they would eat soil more regularly in folivory, as reported for other primates, nor is it con- the fruiting season, which was not the case. In spite of sistently related to the fruiting season, as has been similar sample sizes in October, November and found for some other frugivorous mammals. The tannin December (Table 1), only one soil feeding-bout was adsorptive role of the clay cannot be ruled out. Supra- observed then (November, 1990). Abrupt seasonal annual observations point to an irregular habit of changes in the diet such as a sudden lack of ®bre, or geophagy, probably related to rare dietary components. increases in carbohydrates and soluble proteins, may lead to digestive disorders apparently ameliorated by geophagy (Kreulen, 1985; Mahaney, Aufreiter et al., Acknowledgements 1995). These would occur in a short period of dietary transition (such as the event observed in November We thank PeÂrsio S. Andrade, Armando M. Calouro, 1990, when fruits start to increase in the sakis' diet), but Jean Philip Boubli and Denise A. Gaspar for assistance would regularly also occur. in the ®eld. Eliana M. Cancello and Carlos Roberto Since supra-annual fruiting is common, or even pre- BrandaÄo identi®ed the termites and the ants, respec- dominant in Amazon forests (see Schulz, 1960; Setz, tively. Paulo R. Pedroso, Ruth Costa Lopes, Simone de 1993), geophagy related to species-speci®c detoxi®cation Arruda Leite and Sandra Midori Higashi assisted in could occur less regularly. The coincidental ingestion of conducting soil analyses and Wanderley A. Tremocoldi Trichilia spp. and soil favours this idea which, however, helped with X-ray diffractometry. Laboratory analyses needs to be veri®ed. were completed in the Soil Fertility Laboratory of the Agronomy Institute of Campinas (IAC) and the Chemical and the Geoscience Institutes of the State Evidence for geophagy as tactile sensations in the mouth University of Campinas (UNICAMP). Ivan Sazima, or as tradition Claude Gascon, Jose Roberto Trigo, M. Christina M. Amorozo and an anonymous reviewer made helpful Infrequent geophagy led Mahaney, Watts et al. (1990) comments on the manuscript. Anthony Rylands' com- to suggest that there is a real possibility that it has ments much improved this manuscript. Final drawings almost no essential nutritional role (but provides tactile were made by Esmeralda Zanchetta Borghi. The ®eld- sensations in the mouth, for example), but our data do work was supported by the Biological Dynamics of not rule out a relation to detoxi®cation of items which Forest Fragments Project (World Wildlife Fund-US, rarely appear in the diet. Smithsonian Institution and National Institute for Since we studied only one group of sakis we do not Amazon Research (INPA)) and its contribution number have data to discuss traditional factors. 206 of their technical series. Fieldwork was completed when the ®rst author was a doctoral candidate in Biology Program at UNICAMP, with travel support by Evidence for geophagy as a therapeutic mediator of the Brazilian Higher Education Authority (CAPES ± endoparasitism CoordenacËaÄo de AperfeicËoamento de Pessoal de Nõ vel Superior). In spite of heavy infection of enteric parasites, rhesus macaques exhibit low levels of diarrhoea and excellent REFERENCES physical condition, with a high reproductive rate and low mortality (Knezevich, 1998). Although we do not Bicca-Marques, J. C. & Calegaro-Marques, C. (1994). A case of have data on endoparasite prevalence, signs of diar- geophagy in the black howling monkeys Alouatta caraya. rhoea were not observed in the sakis. Adult female sakis Neotropical Primates 2: 7±8. reproduced every second year (see Setz & Gaspar, Chauvel, A. (1982). Os latossolos amarelos, aÂlicos, argilosos 1997), as expected for monkeys of their size (Dunbar, dentro dos ecossistemas das Bacias Experimentais do INPA e 1987; Setz, 1993). Only one saki death was registered for da regiaÄo vizinha. Acta Amazon. 12: 47±60. Davies, A. G. & Baillie, I. C. (1988). Soil-eating by red leaf the group in 6 years (a sub-adult male; see Setz & monkeys (Presbytis rubicunda) in Sabah, Northern Borneo. Gaspar, 1997). The young male was recorded eating soil Biotropica 20: 252±258. more often than the adult, and this is consistent with a Dunbar, R. I. M. (1987). Demography and reproduction. In higher prevalence of parasites in younger animals, such Primate societies: 240±249. Smuts et al. (Eds). Chicago: Uni- as occurs in rhesus macaques (Knezevich, 1998). But versity of Chicago Press. unlike Cayo Santiago macaques, which commonly Emmons, L. H. & Stark, N. M. (1979). Elemental composition of engage in geophagy (Sultana & Marriot, 1982; Kneze- a natural lick in Amazonia. Biotropica 11: 311±313. Ferrari, S. F. (1995). Observations on Chiropotes albinasus vich, 1998), saki geophagy was an irregular habit, and from the rio Marmelos, Amazonas, Brazil. Primates 36: 289± too infrequent to consider it further as a therapeutic 293. mediator without evidence of endoparasitism. Flowerdew, J. R. (1987). Mammals, their reproductive biology and The observations reported here on geophagy in the population ecology. London: Edward Arnold.  Geophagy in golden-faced sakis 101 Hagerman, A. E. (1987). Radial diffusion method for determining lemurs, monkeys and apes: 223±249. Clutton-Brock, T. H. (Ed.). tannin in plant extracts. J. Chem. Ecol. 13: 437±449. London: Academic Press. Harrison, M. J. S. & Hladik, C. M. (1986). Un primate granivore: Lovejoy, T. E. & Bierregaard, R. O. Jr (1990). Central Amazonian le colobe noir dans la ForeÃt du Gabon; potentialite d'eÂvolution forests and the Minimum Critical Size of Ecosystems Project. du comportement alimentaire. Rev. Ecol. (Terre Vie) 41: 281± In Four Neotropical forests: 60±71. Gentry, A. H. (Ed.). New 298. Haven: Yale University Press. Hebert, D. & Cowan, I. McT. (1971). Natural salt licks as a part Mahaney, W. C. (1987). Behaviour of the African buffalo on of the ecology of the mountain goat. Can. J. Zool. 49: 605±610. Mount Kenya. Afr. J. Ecol. 25: 199±202. Hershkovitz, P. (1987). The taxonomy of South American sakis, Mahaney, W. C., Aufreiter, S. & Hancock, R. G. V. (1995). genus Pithecia (Cebidae, Platyrrhini): a preliminary report and Mountain gorilla geophagy: a possible seasonal behavior for critical review with the description of a new species and a new dealing with effects of dietary changes. Int. J. Primatol. 16: subspecies. Am. J. Primatol. 12: 387±468. 475±487. Heymann, E. W. & Hartmann, G. (1991). Geophagy in mous- Mahaney, W. C., Hancock, R. G. V. & Inoue, M. (1993). tached tamarins, Saguinus mystax (Platyrrhini: Callitrichidae), Geochemistry and clay mineralogy of soils eaten by japanese at the Rõ o Blanco, Peruvian Amazonia. Primates 32: 533±537. macaques. Primates 34: 85±91. Hladik, C. M. (1973). Alimentation et activite d'un groupe de Mahaney, W. C., Milner, M. W., Sanmugadas, K., Hancock, chimpanzeÂs reÂintroduits en ForeÃt Gabonaise. Rev. Ecol. (Terre R. G. V., Aufreiter, S., Wrangham, R. & Pier, H. W. (1997). Vie) 27: 343-413. Analysis of geophagy soils in Kibale forest, Uganda. Primates Hladik, C. M. (1977a). A comparative study of the feeding 38: 159±176. strategies of two sympatric species of leaf monkeys: Presbytis Mahaney, W. C., Watts, D. P. & Hancock, R. G. V. (1990). senex and Presbytis entellus. In Primate ecology: studies of Geophagia by mountain gorillas (Gorilla gorilla beringei) in the feeding and ranging behaviour in lemurs, monkeys and apes: Virunga Mountains, Rwanda. Primates 31: 113±120. 324±353. Clutton-Brock, T. H. (Ed.). London: Academic Press. Marlow, R. W. & Tollestrup, K. (1982). Mining and exploitation Hladik, C. M. (1977b) Chimpanzees of Gabon and chimpazees of of natural mineral deposits by the desert tortoise, Gopherus Gombe: some comparative data on the diet. In Primate agassizii. Anim. Behav. 30: 475±478. ecology: studies of feeding and ranging behaviour in lemurs, Mathews, A. G. A. (1977). Studies on termites from the Mato monkeys and apes: 481±501. Clutton-Brock, T. H. (Ed.). Grosso State, Brazil. Rio de Janeiro: Academia Brasileira de London: Academic Press. CieÃncias. Hladik, C. M. (1978). Adaptive strategies of primates in relation Mayer, J. J. & Brand, P.N. (1982). Identity, distribution and to leaf-eating. In The ecology of arboreal folivores: 373±395. natural history of the peccaries, Tayassuidae. In Mammalian Montgomery, G. G. (Ed.). Washington, D.C.: Smithsonian biology in South America: 433±455. Mares, M. A. & Genoways, Institution Press. H. H. (Eds). Linesville: University of Pittsburgh. Hladik, C. M. & Gueguen, L. (1974). GeÂophagie et nutrition MuÈller, K. H., Ahl, C. & Hartmann, G. (1997). Geophagy in minerale chez les primates sauvages. C. R. Acad. Sc. Paris, (SeÂr. masked titi monkeys (Callicebus personatus melanochir) in D) 279: 1393±1396. Brazil. Primates 38: 69±77. Huffman, M. A., Gotoh, S., Turner, L. A., Hamai, M. & Neves, D. P. (1991). Parasitologia humana. 8th edn. Rio de Yoshida, K. (1997). Seasonal trends in intestinal nematode Janeiro: Atheneu. infection and medicinal plant use among chimpanzees in Oates, J. F. (1978). Water-plant and soil consumption by Guereza Mahale mountains, Tanzania. Primates 38: 111±125. monkeys (Colobus guereza): a relationship with minerals and Izawa, K. (1975). Foods and feeding behavior of monkeys in the toxins in the diet? Biotropica 10: 241±253. Upper Amazon Basin. Primates 16: 295±316. Poirier, F. E. (1970). The nilgiri langur (Presbytis johnii) of South Izawa, K. (1993). Soil-eating by Alouatta and Ateles. Int. J. India. In Primate behaviour: developments in ®eld and laboratory Primatol. 14: 229±242. research 1: 251±383. Rosenblum, L. A. (Ed.). New York: Jackson, M. L. (1969). Soil chemical analysis ± advanced course. Academic Press. Madison: University of Wisconsin. Pollock, J. I. (1977). The ecology and sociology of feeding in Indri Johns, T. (1986). Detoxi®cation function of geophagy and domes- indri. In Primate ecology: studies of feeding and ranging beha- tication of the potato. Chem. Ecol. 12: 635±646. viour in lemurs, monkeys and apes: 38±69. Clutton-Brock, T. H. Johns, T. & Duquette, M. (1991). Detoxi®cation and mineral (Ed.). London: Academic Press. supplementation as functions of geophagy. Am. J. Clin. Nutr. Raij, B. & Quaggio, J. A. (1983). MeÂtodos de anaÂlise de solo para 53: 448±456. ®ns de fertilidade. B. TeÂcn. Inst. Agron. 81: 1±31. Kinzey, W. G. & Norconk, M. A. (1993). Physical and chemical Ranzani, G. (1980). Identi®cacËaÄo e caracterizacËaÄo de alguns solos properties of fruit and seeds eaten by Pithecia and Chiropotes in da EstacËaÄo Experimental de Silvicultura Tropical do INPA. Surinam and Venezuela. Int. J. Primatol. 14: 207±227. Acta Amazon. 10: 7±41. Klein, L. L. & Klein, D.J. (1977). Feeding behaviour of the Said, S. A., Shibel, A. M. & Abdullah, M. E. (1980). In¯uence of Columbian spider monkey. In Primate ecology: studies of various agents on adsoption capacity of kaolin for Pseudo- feeding and ranging behaviour in lemurs, monkeys and apes: monas aeruginosa toxin. J. Pharm. Sci. 69: 1238±1239. 153±181. Clutton-Brock, T. H. (Ed.). London: Academic Press. Salick, J., Herrera, R. & Jordan, C. F. (1983). Termitaria: nutrient Knezevich, M. (1998). Geophagy as a therapeutic mediator of patchiness in nutrient-de®cient Rain Forest. Biotropica 15: 1±7. endoparasitism in a free-ranging group of rhesus macaques Schulz, J. P. (1960). Ecological studies on rain forest in northern (Macaca mulatta). Am. J. Primatol. 44: 71±82. Suriname. In The vegetation of Suriname II. Hulster, I. A. & Kreulen, D. A. (1985). Lick use by large herbivores: a review of Lanjouw, J. (Eds.). Amsterdam: Van Eedenfonds. bene®ts and banes of soil consumption. Mammal Rev. 15: 107± Setz, E. Z. F. (1993). Ecologia alimentar de um grupo de parauacus 123. (Pithecia pithecia chrysocephala) em um fragmento ¯orestal na Lahm, S. A. (1986). Diet and habitat preference of Mandrillus AmazoÃnia Central. Doctoral thesis, Universidade Estadual de sphinx in Gabon: implications of foraging strategy. Am. J. Campinas, Brazil. Primatol. 11: 9±26. Setz, E. Z. F. & Gaspar, D. A. (1997). Scent marking behaviour in Lindburg, D. G. (1977). Feeding behaviour and diet of rhesus the free-ranging golden-faced saki monkeys, Pithecia pithecia monkeys (Macaca mulatta) in a siwalik forest in north India. In chrysocephala: sex differences and context. J. Zool. (Lond.) Primate ecology: studies of feeding and ranging behaviour in 241: 603±611. 102 E. Z. F. Setz ET AL. Sick, H. (1949). Sobre a extracËaÄo do sal de cinzas vegetais pelos Vermeer, D.E. & Frate, D. A. (1979). Geophagia in rural Mis- õ ndios do Brasil Central. Revista do Museu Paulista, n.s. 3: 381± sissippi: environmental and cultural contexts and nutritional 390. implications. Am. J. Clin. Nutr. 32: 2129±2135. Stark, N. (1970). The nutrient content of plants and soils from Weeks, H. P., Jr & Kirkpatrick, C. M. (1976). Adaptation of Brazil and Surinam. Biotropica 2: 51±60. white-tailed deer to naturally occurring sodium de®ciencies. Strusaker, T. T., Cooney, D. O. & Siex, K. S. (1997). Charcoal J. Wildl. Mgmt. 40: 610±625. consumption by zanzibar red colobus monkeys: its function Weeks, H. P., Jr & Kirkpatrick, C. M. (1978). Salt preferences and and its ecological and demographic consequences. Int. J. sodium drive phenology in fox squirrels and woodchucks. Primatol. 18: 61±72. J. Mammal. 59: 531±542. Sultana, C. J. & Marriot, B. M. (1982). Geophagia and related Weir, J. S. (1972). Spatial distribution of elephants in an African behavior of rhesus monkeys (Macaca mulatta) on Cayo San- National Park in relation to environmental sodium. Oikos 23: tiago Island, Puerto Rico. Int. J. Primatol. 3: 338. 1±13. Uehara, S. (1982). Seasonal changes in the techniques employed Wilkinson, L. (1991). SYSTAT: the system for statistics. Evan- by wild chimpanzees in the Mahale Mountains, Tanzania, to ston: Systat Inc. feed on termites (Pseudacanthotermes spiniger). Folia Primato- Wrangham, R. W. (1977). Feeding behaviour of chimpanzees in logica 37: 44±76. Gombe National Park, Tanzania. In Primate ecology: studies Underwood, E. J. (1977). Trace elements in human and animal of feeding and ranging behaviour in lemurs, monkeys and apes: nutrition. New York: Academic Press. 504±538. Clutton-Brock, T. H. (Ed.). London: Academic Vermeer, D. E. & Ferrell, R. E. Jr (1985). Nigerian geophagical Press. clay: a traditional antidiarrheal pharmaceutical. Science 227: 634±636.  Geophagy in golden-faced sakis 103 Appendix. X-ray diffractometry sample preparation the sequence repeated. After cooling the mixture was procedure modi®ed from Jackson (1969): centrifuged, the supernatant discarded, the material A portion of the sample was transferred to a beaker re-suspended and centrifuged again for 2.45 min at 750 and 60 ml of sodium acetate/acetic acid (pH 5) was rpm to separate the silt fraction from clay minerals. The added, followed by 20 ml of hydrogen peroxide (30%). supernatant was divided in two fractions which were This mixture stood at room temperature overnight and treated with 1 M NaCl solution and 1 M MgCl2 respect- was then gently heated on a steam bath. More H2O2 ively, by suspending the clay fraction in the salt was added until reaction ceased completely. The excess solution, centrifuging and discarding the supernatant. of H2O2 was removed by boiling. The mixture was This step was repeated three times. The excess of salt centrifuged and the supernatant discarded. The sample was washed by successive steps of washing with de- was re-suspended with water and passed through a 270 ionized water, centrifugation until the supernatant mesh sieve. The ®ne fraction was transferred to a beaker solution was free of chlorides. and 50 ml of a combined sodium citrate/bicarbonate Suspended clay fraction was pipetted over a glass and (pH 7.3) mixture was added and the whole was heated left to dry. The K-saturated clay fraction was run on the to 75 8C on a steam bath. Sodium ditionite (c. 1 g) was diffractometer as derived and after heat treatment at added, the mixture was homogenized, left reacting, and 350 and 550 8C. The Mg-saturated fraction was run as the second portion of sodium ditionite was added and derived and after saturation with ethyleneglycol.