Why On Earth?: Evaluating Hypotheses About The ...WHY ON EARTH?: EVALUATING HYPOTHESES ABOUT THE...

Why On Earth?: Evaluating Hypotheses About The Physiological Functions Of HumanGeophagyAuthor(s): Sera L. Young, Paul W. Sherman, Julius B. Lucks, Gretel H. PeltoSource: The Quarterly Review of Biology, Vol. 86, No. 2 (June 2011), pp. 97-120Published by: The University of Chicago PressStable URL: http://www.jstor.org/stable/10.1086/659884 .Accessed: 20/09/2011 15:14

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .http://www.jstor.org/page/info/about/policies/terms.jsp

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range ofcontent in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new formsof scholarship. For more information about JSTOR, please contact [email protected].

The University of Chicago Press is collaborating with JSTOR to digitize, preserve and extend access to TheQuarterly Review of Biology.

http://www.jstor.org

http://www.jstor.org/action/showPublisher?publisherCode=ucpress

http://www.jstor.org/stable/10.1086/659884?origin=JSTOR-pdf

http://www.jstor.org/page/info/about/policies/terms.jsp

WHY ON EARTH?: EVALUATING HYPOTHESES ABOUT THEPHYSIOLOGICAL FUNCTIONS OF HUMAN GEOPHAGY

Sera L. YoungDepartment of Obstetrics and Gynecology, University of California, San Francisco, California 94105 USA

Division of Nutritional Sciences, Cornell University, Ithaca, New York 14853 USA

e-mail: [email protected]

Paul W. ShermanDepartment of Neurobiology and Behavior, Cornell University, Ithaca, New York 14853 USA


Julius B. LucksSchool of Chemical and Biomolecular Engineering, Cornell University, Ithaca, New York 14853 USA


Gretel H. PeltoCollege of Human Ecology, Division of Nutritional Sciences, Cornell University, Ithaca, New York 14853

USA


keywordspica, vertebrates, mammals, pregnancy, protection, micronutrient supplement

abstractGeophagy has been hypothesized to be an adaptive behavior, either as a means to allay nutrient

deficiency or to protect against ingested pathogens and toxins. Others have proposed that geophagy isnon-adaptive, occurring either to allay hunger or as an epiphenomenon of nutrient deficiencies. Thispaper evaluates these hypotheses using 482 published cultural-level accounts of human geophagy and330 accounts of geophagy among 297 species of mammals, birds, and reptiles. Information wasextracted from reports of human geophagy to permit statistical analysis; reports of non-humangeophagy were tabulated. Human geophagy did not parallel changes in nutrient requirements,occurred most frequently among children and pregnant women and in tropical areas (where pathogendensities are highest), and was associated with ingestion of toxic substances and gastrointestinaldistress. Earth ingested by humans was craved and carefully selected and prepared; it had high clay

The Quarterly Review of Biology, June 2011, Vol. 86, No. 2

Copyright © 2011 by The University of Chicago Press. All rights reserved.

0033-5770/2011/8602-0002$15.00

Volume 86, No. 2 June 2011THE QUARTERLY REVIEW OF BIOLOGY

97

content, but few bioavailable mineral nutrients. In primates, geophagy was associated with bothprotection from toxins and obtaining nutrients, whereas in other vertebrates it was associated mainlywith obtaining nutrients. Our results indicate that human geophagy is best explained as providingprotection from dietary chemicals, parasites, and pathogens, whereas animal geophagy may involveboth micronutrient acquisition and protection.

Introduction

GEOPHAGY IS the intentional consump-tion of earth. It is a specific type of the

more general phenomenon that is commonlyreferred to as “pica,” the purposive consump-tion of non-food substances. Geophagy iswidely practiced: it has been observed in hun-dreds of cultures on all inhabited continents(Laufer 1930; Anell and Lagercrantz 1958).Geophagy was first documented by Hippocra-tes (460–380 bc) more than 2000 years ago(Hippocrates and Adams 1849). However, ar-cheological evidence suggests that the practiceof geophagy is thousands of years older andmay date back to Homo habilis (Baudouin 1924;Clark 2001; Brady and Rissolo 2006). Eartheating continues throughout the world today(Young 2010, 2011).

Scholars from diverse fields including an-thropology, behavioral ecology, biochemistry,ethology, geography, and medicine have of-fered a range of hypotheses for why earth isconsumed. Yet, even after much investigation,geophagy remains an enigma for many rea-sons (Young 2010). These include the scarcityof hypothesis-driven research as well as the typ-ically single-discipline approaches to its study. Itis also due to underreporting, firstly becauseresearchers often do not inquire about ge-ophagy and, secondly, even if they do, fearof being judged harshly leads some geoph-agists to conceal their behavior (Young etal. 2008).

Because of both the prevalence of geoph-agy and its association with positive and neg-ative health consequences, there is a clearlyacknowledged need to understand the eti-ology of the behavior. Indeed, calls for elu-cidation of the physiological functions ofthis fascinating and enigmatic phenome-non are a recurring theme in the geophagyliterature (e.g., Whiting 1947; Edwards etal. 1994; Geissler et al. 1997; Limpitlaw2010).

The HypothesesThree general physiological explana-

tions for the etiology of human geophagyhave been advanced. The first two sug-gest contexts in which the behaviorwould be adaptive, whereas the third sug-gests that it is not an adaptive behavior.

Hypothesis 1: Nutrient Deficiency. This hy-pothesis proposes that people eat earth inorder to make up for dietary deficiencies ofmineral micronutrients, particularly ironand zinc (Hunter 1973; Cavdar et al. 1983;Prasad 2001b) and the macromineral cal-cium (Wiley and Katz 1998). For brevity,these will collectively be referred to as nu-trients. Because anemia (the state of insuf-ficient red blood cells or hemoglobin) isfrequently associated with geophagy (Gei-ssler et al. 1999; Kawai et al. 2009), themost commonly reiterated of the nutrienthypotheses suggests that geophagy reme-dies iron deficiency. It should be noted,however, that anemia can be caused bymicronutrient deficiencies other thaniron, as well as infections and blood loss(Yip and Dallman 1988). Sodium defi-ciency has been proposed to motivate ge-ophagy in non-human primates and otheranimals (Jones and Hanson 1985; Krish-namani and Mahaney 2000). However,human geophagists do not attribute theirbehavior to a dearth of salt (Vermeer andFrate 1979; Kraemer 2002) and with only afew exceptions (Laufer 1930), geophagicearth is not typically salty (Young et al.2008).

Hypothesis 2: Protection. This hypothesisproposes that earth is eaten as a medica-ment, to reduce the short-term malaise andlong-term effects of harmful chemicals andparasites and pathogens. Many human foodplants produce toxic chemicals, such as tan-nins and glycoalkaloids to protect themselvesfrom biotic enemies (pathogens and herbi-vores). Other sources of harmful chemicals

98 Volume 86THE QUARTERLY REVIEW OF BIOLOGY

in the human diet are enterotoxins secretedby food- and waterborne bacteria such asEscherichia coli, Staphylococcus aureus, Salmo-nella enterica, and Listeria monocytogenes. Inges-tion of these toxins can cause gastrointestinaldistress, dizziness, and muscle pains; in suffi-cient quantities, they can be mutagenic,carcinogenic, or deadly (Hui et al. 2001b).Dangerous pathogens include food- andwaterborne bacteria as well as viruses andparasitic nematodes.

Under this hypothesis, there are twomechanisms by which geophagic earth maybe protective: by reducing the permeability ofthe gut wall to toxins and pathogens andby binding directly to toxins and pathogens(Young 2010). The intestinal mucosal layeracts as a physical barrier between ingestaand the bloodstream by filtering out largemolecules, as well as a chemical barrier bymaintaining a pH gradient. Geophagic earth,especially if it is clay-rich, can bind with andthereby reinforce the protective mucosal layerand/or enhance mucosal secretion, therebyreducing permeability of the intestinal walls(Gonzalez et al. 2004).

The second mechanism involves bindingdirectly to toxins, parasites, and other patho-gens. This can either render them unabsorb-able by the gut or inhibit their respiration.Hladik and Gueguen (1974) first proposedthat clays were protective against plant sec-ondary compounds consumed by primates,and Profet (1992) suggested that clays couldbe protective against human teratogens. Anumber of clays found in geophagic earthsare capable of binding pathogens, includingviruses (Lipson and Stotzky 1983; Rey 1989;Dornai et al. 1993), fungi (Smith and Carson1984; Lavie and Stotzky 1986a,b; Phillips etal. 2008), and bacteria (Maigetter and Pfister1975; Said et al. 1980; Ditter et al. 1983;Gardiner et al. 1993), as well as toxins, in-cluding poisonous herbicides (Okonek et al.1982; Lotan et al. 1983), pharmaceuticals(Tsakala et al. 1990), and plant secondary com-pounds (Johns 1986; Johns and Duquette1991; Gilardi et al. 1999; Houston et al. 2001;Dominy et al. 2004). It is also possible thatearth inhibits larger pathogens (e.g., geohel-minths) from colonizing hosts (Krishnamani

and Mahaney 2000) although the mechanismby which this occurs has not been elucidated.

Hypothesis 3: Non-Adaptive. This hypothesisproposes that there is no benefit to eatingearth. Instead, people do so either becausethey have no food to eat or because micronu-trient deficiencies have caused neurological orsensory problems. In the first case, earth is sup-posedly consumed to ease hunger pains whenno other food is available (e.g., La Billardiere1800; Mallory 1926; Wiley and Katz 1998). Inthe second case, cravings for earth are sug-gested to be epiphenomena of nutrient defi-ciencies that affect appetite-regulatingbrain enzymes (von Bonsdorff 1977; You-dim and Iancu 1977) or taste sensitivity(Chisholm and Martin 1981; Prasad2001a), causing non-food substances to be-come appealing.

AimsThis paper has two aims. First, we pro-

vide an updated review of the literature ongeophagy in humans and, for comparison,in other vertebrates. For the human litera-ture, our starting points were eight excellentmonographs and reviews (Laufer 1930; Anelland Lagercrantz 1958; Hochstein 1968; Dan-ford 1982; Sayetta 1986; Loveland et al. 1989;Horner et al. 1991; Reid 1992) which werethen augmented by primary sources. We com-piled this information into a database thatenabled us to quantitatively describe geoph-agy worldwide and investigate the circum-stances under which it occurs. Our secondaim was to use the database on human ge-ophagy to systematically evaluate the threehypotheses for the etiology of geophagy.This paper does not directly address thecultural forces and beliefs that shape ge-ophagy (the proximate determiners of thebehavior), but rather explores the ecolog-ical triggers and physiological underpin-nings of geophagy (the ultimate causes).Socioeconomic status may be an underly-ing cause of hunger, nutrient deficiency,or increased exposure to toxins and patho-gens, but it is not a direct cause of geoph-agy itself. Therefore, it is not an alternativeexplanation for geophagy and is not testedhere.

June 2011 99PHYSIOLOGICAL FUNCTIONS OF HUMAN GEOPHAGY

Additionally, we synthesized reports ofgeophagy by non-human animals to pro-vide further insights into the distributionof the behavior and its potential physiolog-ical causes. Our starting points were severalexcellent reviews of geophagy in animals(Jones and Hanson 1985; Kreulen 1985;Krishnamani and Mahaney 2000; Ferrari etal. 2008). Because this information was farless detailed and comprehensive than dataon humans, it was tabulated, but not ana-lyzed statistically.

Methodologyliterature searching

Initially, we searched online databases(Agricola, Dissertation Abstracts, GoogleScholar, Human Relation Area Files, ISIWeb of Science, JSTOR, Library of Con-gress, LexisNexis, OCLC, Proquest Histor-ical Newspapers, PubMed, and ZoologicalRecord) for entries containing “geophagy,”“geophagia,” “pica,” “clay eating,” “chalk eat-ing,” “cachexia Africana,” “mal d’estomac,”“malacia,” “citta,” “erde essen,” “aarde eten,”and “dirt eating.” We sought information ongeophagy in humans, non-human primates,and other vertebrates. We used the referencelists in each identified publication to locateadditional primary sources. This process wasiterated until no new references were found.Our search was not restricted by language,format of reference (e.g., microfiche andthesis, among others), or date of publication;the references span many languages andnearly 500 years.

criteria and rationale for inclusionWe defined geophagy as the regular,

purposive consumption of earth. With thisdefinition, the mouthing behavior ofyoung children is excluded (their earthconsumption may not be intentional, butrather part of a larger behavior of environ-mental exploration). Instances of geoph-agy that were described as purely symbolic,such as the ingestion of tiny amounts dur-ing solemn occasions (oaths, mourning,tests of innocence or for religious purposes)also were excluded. Finally, if a mineralfound in soil was used in the preparation of

food, but was not ultimately consumed, suchas during nixtamalizaton (the soaking ofcorn in a limestone solution before grind-ing), it was not considered geophagy.

The human geophagy literature can be clas-sified into five categories: 1) individual casereports—e.g., a Turkish woman living in Parisate chalk and clay every day (Henon et al.1975); 2) an enumerated population—e.g.,55% of anemic Namibian women eat earth(Thomson 1997); 3) a cultural group—e.g.,pregnant Otomacs regularly engage in clay-eating (von Humboldt et al. 1821); 4) soil anal-ysis studies—e.g., montmorillonite was amajor component of ch’aqu (Browmanand Gunderson 1993:415); and 5) litera-ture reviews in which no new data werepresented— e.g., “According to La Bil-lardiere and confirmed by the reports ofHekmeyer [clay] figures are crunched onby women and children” (Ferrand 1886:549).

In constructing our database on humangeophagy, we focused on firsthand reportsof geophagy among cultural groups (cate-gory 3). If an author referred to a report ofgeophagy by someone else (category 5), weobtained the original document. This in-sured that reports were not included morethan once, and that the translation was ofhigh quality. Thus, the unit of analysis forour study is a “cultural report.” We did notinclude reports of individual casesof geophagy (category 1) because of thelikelihood of bias regarding both healthconsequences (people not suffering illhealth are unlikely to visit a health careprovider) and geographical occurrence(health care providers are not equally avail-able worldwide). We also did not includestudies of groups among whom geophagywas studied because of a biological or be-havioral condition (e.g., anemia, dialysis,lead poisoning; category 2) because theywere not representative of the populationat large. Results from soil analyses (cate-gory 4) are drawn on throughout the pa-per.

In reviewing the animal literature, we fo-cused on geophagy and excluded lithophagy(ingesting of rocks or grit). Authors referredto animal geophagy using the terms “clay


licks,” “mineral licks,” “salt licks,” and “ge-ophagy”; all such descriptions were includedin our tables and analysis. The original doc-ument was obtained whenever possible; re-ports citing personal communication werenot included. We located 330 ethological ac-counts of geophagy among mammals, birds,and reptiles encompassing 297 species and26 orders.

constructing the database on humangeophagy

From each article on human geophagywe extracted information on as many of thefollowing variables as possible: year of ob-servation, geographic location, climate,non-food materials consumed (e.g., appear-ance, source, preparation), life stage of theconsumer (e.g., child, adolescent, preg-nant), and any associations with physio-logical conditions, such as gastrointestinaldistress, anemia, and hunger. If several re-ports were made about earth eaten by thesame group of people or in the same areawithin a 10-year span, but by different au-thors, these were combined into one culturalreport. A custom built Web-based form wasdeveloped specifically for this project us-ing Ruby on Rails software (http://www.rubyonrails.org/), and information from

each article was entered manually. Datafrom biological, epidemiological, and cul-tural sources were included, making a bi-ocultural analysis possible.

Our database on human geophagy in-cluded 482 publications, which contained367 separate cultural reports of geophagyfrom all over the world (in supplementarymaterial, available at The Quarterly Reviewof Biology homepage, http://journals.uchicago.edu/QRB). This database in-cludes every obtainable, written culture-level report of human geophagy. To ourknowledge, it is the most complete com-pilation of such information in existence.References for all sources are listed inthe supplementary materials; the entire“Pica Literature Database” will be madeavailable once all planned analyses havebeen completed.

Constructing a map of the frequency of oc-currence of geophagy worldwide (Figure 1)presented a challenge because observationsat the level of the cultural group do not“map” perfectly onto political maps, sincecultures sometimes are dispersed over sev-eral countries or located in just one sectionof a country. To achieve a reasonable ap-proximation, we selected the country inwhich the ethnic group was primarily situ-

Figure 1. Worldwide Distribution of Cultural Reports of Geophagy


ated at the historical time described in theoriginal paper(s).

We used Köppen’s classification systemto categorize the climatic regions in whichgeophagy occurs (Kottek et al. 2006). TheKöppen system separates climates into fivecategories: polar, cold, temperate, tropical,and dry (McKnight and Hess 2005). To testwhether the distribution of geophagy by cli-mate type was different from the distributionof cultural reports by climate type, we classi-fied each of the 186 cultures in the StandardCross-Cultural Sample (SCCS) (Murdockand White 1969) by climate type. Cultures inthe SCCS were chosen because of their inde-pendence from each other, so use of theSCCS minimizes “Galton’s problem” of lackof independence of cultures in close geo-graphical proximity to each other (Naroll1961). Additionally, we compared thegeophagy distribution to the world populationdistribution by climate region (Staszewski1963).

Because information on human geophagyspans 2000 years and includes reports frommany fields of study, the quality and detail ofobservations varies widely. Accounts were writ-ten by ethnographers, colonial explorers,government officials, missionaries, medicaldoctors, nurses, nutritionists, and journalists.Some reports of geophagy are lengthy (e.g.,more than 20 pages), and describe in detailthe characteristics of individuals who prac-tice geophagy, when in their lifetimes theydo so, sources and preparation of earth, andcosts, among other topics, whereas other re-ports are no more than brief mentions (e.g.,a single phrase in a 417-page ethnographicreport).

The literature on non-human geophagywas far less detailed and specific, e.g., therewere few descriptions of life stage, sex, orreproductive status of geophagists, nor thecomposition of earths consumed. The cre-ation of an analyzable database of thesereports could not yield similar insights, so wetabulated the ethological reports. Reports ongeophagy in non-human primates were typ-ically more detailed than those on geophagyin other vertebrates (see supplementary ma-terial, Tables 2 and 3). A “report” was de-fined as a description of geophagy by a

particular species in one geographic loca-tion, i.e., if there were multiple observationsof the same species in the same location,they were not tallied twice.

operationalizationTo create variables in the database that

could be analyzed, we first grouped the lifestages of geophagists into eight categories:(1) “infants” are children younger than twoyears or are still breastfeeding; (2) “preado-lescents” are children who have not reachedpuberty (3–12 years old); (3) “adolescents”are boys and girls who have reached puberty(13–18 years old); (4) “pregnant women” areadolescent or adult women who are gestat-ing; (5) “lactating women” are those who arebreastfeeding; (6) “women” are adults whosepregnancy or lactation status is unknown;(7) “men” are adult males; and (8) “elderly”are those described in reports as being “old”or no longer bearing children. Regardingcategory 6, if women’s reproductive statuswas not mentioned, we inferred that theywere not pregnant or lactating, although thismay have resulted in some misclassificationbecause in many traditional societies, womenof reproductive age typically were pregnantor lactating most of the time.

Data on proportions of populations orsubpopulations that engaged in geophagywere rarely given in the reports we com-piled. Rather, qualitative terms such as“some,” “all,” “frequently,” and “rarely” wereused instead. In an attempt to quantify thesedescriptive terms, we constructed a scoringsystem (Table 1) similar to that of Wiley andKatz (1998).

TABLE 1Geophagy scoring system used in the

operationalization of geophagy frequency

Score Terms Associated

0 Never1 Rarely, few2 Sometimes, occasionally3 Frequently, common, habit, very common, quite

general, many, endemic, widely, often4 Usually, typically5 Always


Only rarely did observers identify whichsubgroups within the population did not eatearth. More typically the observer wouldcomment only on positive occurrences (e.g.,“pregnant women and children eat clay”).Therefore, we standardized the frequenciesby the absolute number of cultural reportsmade for each life stage. To do this, we di-vided the frequency of geophagy for each lifestage by the number of cultural reports pre-senting information on that particular lifestage (Figure 2). To reduce the geophagyfrequency to a more intuitive value, we cre-ated a geophagy score, which is the meangeophagy frequency for that life stage.

When observers discussed the timing ofconsumption during pregnancy, terms suchas “early” and “late” were used more fre-quently than specific months or trimesters.We classified pregnancy timing described as“early” as first trimester and grouped thosedescribed as “mid” or “late” pregnancy into asingle category “second or third trimester.”

“Anemia” was not a term that was in usefor most of the time period encompassed by

our literature review, and biomedical tests ofanemia have become standard only relativelyrecently. Thus, associations between geoph-agy and mild anemia would rarely havebeen reported. However, we were able toidentify associations with severe anemiawhen certain symptoms were described,e.g., pallor of skin and/or mucosal mem-branes and thin “watery” blood. “Chloro-sis” is a term that was frequently used in thepast to describe weak, pale patients; it isnow considered to be synonymous withanemia (Hudson 1977). Thus, if geopha-gists were described as pale, having thinblood or exhibiting chlorosis, we recordedgeophagy as being associated with anemiain that report.

Textures and consistencies of earths varywidely due to different proportions of thefour major solid components: sand, silt,clay, and organic matter. Proportions ofthese components is important becauseeach has very different effects on the body(Wilson 2003). In particular, sand is theleast reactive inorganic fraction of earth,whereas clay is the most reactive (Saetherand Caritat 1997). We classified geophagicearths based on qualitative descriptions inthe original reports. Earth was categorizedas “claylike” if the author of the report de-scribed it in terms such as “clayey,” “plastic,”“moldable,” “used for pottery,” or as “marl.”

Our database includes reports of con-sumption of 402 different types of earth,ranging in color from bright white to lightyellow, orange, red, red-brown, purple, darkgrey, black, blue, and light green. If morethan one type was reported to be consumedin a culture, we used only the one that wasmost frequently consumed for each lifestage. We chose this approach becausecounting more than one geophagic earthwould lead to oversampling if one group ateseven types of clay, each one infrequently,whereas another group ate only one type ofclay, but did so frequently. For example, if areport stated that “pregnant women rarelyate red clay but usually ate grey clay,” preg-nant women were classified as “usually” con-sumers of “grey clay.”

The elemental constituents of only a fewgeophagic earths have been chemically an-

Figure 2. Standardized Geophagy Frequenciesand Mean Geophagy Score (DottedLine) by Life Stage Based onCulture-Level Reports of Geophagy


alyzed; in many cases only qualitative de-scriptions are available. In general, it is notpossible to determine the nutrient constitu-ents of earth by visual inspection, with oneexception: it can be inferred that soils con-tain iron if they are red in color (Jeff Wilson,personal communication). However, ironmay be present if the soils are not red, be-cause some iron-rich components result inpigmentation other than red. The calciumor zinc contents of most geophagic earthswere not quantified. Even if they had been,the total elemental composition of the earthdoes not indicate the bioavailability of itsconstituents, i.e., the proportion freely avail-able to cross an organism’s cellular mem-branes. Bioavailability is typically much lowerthan total amounts (Wilson 2003).

Statistical Analyses

All statistical analyses were performedusing STATA 9.2 for Macintosh (STATACorporation, College Station, Texas). Dif-ferences in geophagy scores among sexand life-stage categories were tested using amultilevel, mixed-effect model (to controlfor repeated measures within a culturalreport; Figure 2). Differences betweenobserved frequency of geophagy by circum-stances of consumption and the null hypothe-sis were tested by using Pearson’s chi-squareanalyses (see Figures 3 and 7). We testedwhether nutrient requirements could pre-

dict geophagy scores (Figure 4) using Spear-man’s test of non-parametric correlation.

Evaluating the Hypothesesadaptive hypothesis 1: nutrient

deficiencyAssociation with Nutrient DeficienciesIf geophagy were a response to a nutri-

ent deficiency, it should occur in conjunc-tion with such a deficiency. Indeed, anassociation between geophagy and ane-mia was recognized as early as 40 ad whenCornelius Celsus, a Roman physician,wrote “those that have a bad color for along time without jaundice, are either dis-trest with pains in the head, or labor undera malacia” (the term then used for crav-ings of non-food substances) (Celsus andGrieve 1756:59). The geophagy-anemia as-sociation has been confirmed repeatedlyworldwide since then. For example, in Zan-zibar, the Swahili term for anemia, safura,was mistranslated by Livingstone as “thedisease of . . . earth eating” (Livingstoneand Waller 1875:346). In 20th-century In-dia (e.g., Hooper and Mann 1906) and onslave plantations in the Americas (e.g.,Buckingham 1842), pallor was often usedas a symptom of the disease of earth eating.Indeed, in our database, anemia and ge-ophagy were associated significantly moreoften than would be expected under thenull hypothesis (p�0.001) (Figure 3a).Available data did not permit us to test anassociation between geophagy and calciumor zinc deficiencies.

Frequency of Geophagy and NutrientRequirements

If geophagy were a response to a defi-ciency in iron, zinc, or calcium, then wewould expect people with the greatest needsto practice geophagy most often. To evaluatethis corollary, we determined the daily refer-ence intakes for each of these elements bypeople in each life stage (Institute of Medi-cine 2002; Figure 4). These values were stan-dardized by dividing by energy requirementsat each life stage to capture the relative re-quirements, a standard consideration in eval-uating risk of nutrient deficiency.

0%

20%

40%

60%

80%

100%

(a) Anemia (b) GI distress (c) Hunger (d) Craving

Pro

porti

on o

f rep

orts

NeverSometimesAlways

(n=79) (n=50) (n=72) (n=77)

*

**

Figure 3. Reports of the Frequency ofAssociation of Geophagy with (A)Anemia, (B) Gi Distress, (C) Hunger,and (D) Craving

The proportion of reports associated with anemia,GI distress, and craving was significantly higher (*)than expected under the null hypothesis (p�0.005).


If geophagic earth were consumed to obtaincalcium, as Wiley and Katz (1998) proposed,one would expect preadolescents, adolescents,and the elderly, who have the highest calciumrequirements, to engage in geophagy most fre-quently. If zinc deficiency were the impetus forgeophagy, it should occur uniformly among allcategories of adults, because they have similarzinc requirements. And, if earth were con-sumed to obtain iron (Hunter 1973; Abra-hams 1997), we would expect infants andpregnant women to ingest earth most fre-quently. However, nutrient requirementswere not significant predictors of geophagyscores for calcium (Spearman’s rho 0.332,p�0.422), iron (Spearman’s rho 0.542,p�0.165), or zinc (Spearman’s rho 0.267,p�0.523). Thus, occurrences of geophagydo not parallel requirements for any of thesethree nutrients (Figure 4).

If geophagy were a response to nutrientdeficiency, then pregnant women shouldconsume earth most often late in gestation,when nutrient requirements are highest.Women need less iron in early pregnancy

than they do when not pregnant becausethey are not experiencing menstrual bloodloss, but later in pregnancy women’s ironrequirements are higher than when notpregnant because of the needs of the devel-oping fetus (Institute of Medicine 2002).Women also need less calcium early in preg-nancy compared to later, because fetal skel-etal growth accelerates in mid-pregnancy;most of the calcium used by the fetus is ac-cumulated during the third trimester (Insti-tute of Medicine 2002). Zinc requirementsdo not change markedly throughout preg-nancy (Institute of Medicine 2002). Our dataindicate that geophagy occurs nearly twice asfrequently in early pregnancy as in late preg-nancy (Figure 5). This pattern is not pre-dicted by the nutrient deficiency hypothesis.

Nutrient Content of Geophagic EarthIf geophagy were an adaptive response

to nutrient deficiencies, we would expectgeophagic earth to contain the nutrientsthat are in short supply. Unfortunately,

Figure 4. Geophagy Score and Selected Nutrient Requirements, by Life StageAll values have been standardized against those for pregnant women for ease of comparison


available information only enables us toevaluate this corollary for iron, and thenonly incompletely. Red color, indicative ofthe presence of iron, was reported in lessthan half the soils that were eaten (70 of158; Figure 6a). Strikingly, when there wasa choice between red clays and clays ofother colors, the non-red clays were pre-ferred (six of eight reports). For example,the Luo people of Kenya and Tanzania pre-ferred the white clay sold at the market tothe reddish clay that could be collected lo-cally (Geissler 2000) and in the U.S. (Ala-bama), white clay was preferred over redclay (Spencer 2002). According to Vermeer(1971), the Ewe people of Ghana actuallyremove iron from red clay soils before con-sumption.

In the geophagy literature, it has beencommon to measure only the total elementalcomposition of geophagic soils (typically us-ing acid digests). This is problematic becauseacid digests alone ignore much of the body’sbiochemistry, most critically, the pH of theintestine, the site of most elemental nutrientabsorption. Because intestinal pH is muchhigher than the stomach, and nutrients aremore soluble at low pH, equating the totalelemental composition with the amount avail-able to cross an organism’s cellular mem-branes—i.e., its bioavailability (Semple et al.2004)—vastly overestimates the usable nutrientcontent (Wilson 2003). Therefore, methodsthat involve only an acid digest can merelyindicate if there is any element of interest pres-ent; the establishment of bioavailability re-

quires more sophisticated techniques (Younget al. 2008).

Only five in vitro studies of geophagic sam-ples have considered intestinal biochemistryin their analyses of bioavailable nutrients.Two of these used the physiologically basedextraction test, which includes a phase thatmimics the pH and digestive enzymes in thegut, to study Ugandan (Smith et al. 2000)and Indian geophagic soils (Abrahams et al.2006), respectively. In these studies, less than5% of the total iron present was bioavailable.Negligible amounts of other biologically nec-essary minerals, including zinc, were avail-able.

Kikouama et al. (2009) investigated traceelements released by six West Africangeophagic clays under conditions that mim-icked the oral, gastric, or intestinal environ-ment (samples were not passed througheach stage consecutively). They found theavailability of both ferric and ferrous iron tobe lowest under intestinal pH, but did notcalculate potential iron contribution.

A fourth in vitro study that attemptedreplication of intestinal conditions wasconducted on two samples of South Afri-can geophagic soils (Dreyer et al. 2004).Earth was added to iron-enriched Ringer’slactate solution and the precipitation ofelements at gastric and intestinal pH wasmeasured. Black geophagic earth adsorbedsodium, potassium, and iron and liberatedcalcium and magnesium at pH 6.2. Ironfrom the Ringer solution was absorbed by

Figure 6. Description of Geophagic SoilsShowing Frequency of ReportsDiscussing (A) Color, (B) Texture,and (C) Preparation

Figure 5. Timing of Geophagy DuringPregnancy (n�15)


the red geophagic earth at pH 6.2; nochange in other elements was seen. In short,neither South African sample provided ironor zinc, but one might provide calcium.

Hooda et al. (2004) conducted the mostthorough study of bioavailability becausethey not only examined the nutrients thatgeophagic materials could contribute in vi-tro, but their capacity to bind nutrients insuspension, thus rendering them unavail-able. Results indicated that the fivegeophagic samples from around the worldprovided bioavailable calcium, magnesium,and manganese, but significantly reducedthe availability of iron and zinc in suspen-sion, suggesting that geophagic earth doesnot contribute these nutrients and is in factlikely to bind the iron and zinc available iningested foods.

Most of the in vivo studies of the effectsof geophagy on micronutrient absorptionused outdated methods, small sample sizes,and were not adequately statistically ana-lyzed (Young 2010). However, the limiteddata they do offer suggest that geophagyeither decreases or does not alter micronu-trient status, rather than increasing it.

Briefly, Minnich et al. (1968) demon-strated that the mean proportion of ironabsorbed by people who ingested 5g ofTurkish soil together with either radio-labeled iron sulfate or radio-labeled hemeiron decreased by 9%. These results weresubsequently replicated using other Turk-ish soils by members of the same researchgroup (Cavdar and Arcasoy 1972). Talking-ton et al. (1970) tested the impact of twopopular Texan geophagic clays on radio-labeled iron absorption and found a 1 to3% increase in iron absorption in the pres-ence of clay. This difference is unlikelystatistically significant. Sayers et al. (1974)studied iron absorption among five habit-ual geophagists in South Africa. 55Fe ascor-bate absorption was greatly decreasedwhen 250g of geophagic earth from partic-ipants’ own supplies was eaten (mean ab-sorption was 17.4% without earth versus5% with earth).

In a study of 17 Turkish children, the 12geophagists demonstrated impaired ironand zinc absorption compared to the five

non-geophagists (Arcasoy et al. 1978). Asecond study of zinc absorption, in the pres-ence and absence of 5g of geophagic clay,also indicated that clay impeded zinc absorp-tion (Cavdar et al. 1983). The authors sug-gested that earth might bind not just withdietary zinc, but also with endogenous zincreleased from the pancreas. In studies ofrats, Smith and Halsted (1970) determinedthat modified Iranian geophagic soil couldcontribute dietary zinc. These results con-trast with bioavailability data from in vitroanalyses of unadulterated geophagic earth(Dreyer et al. 2004; Hooda et al. 2004; Abra-hams et al. 2006), which indicated that littlezinc was available and that some geophagicearth samples bound dietary zinc, renderingit unavailable. Finally, in the most recent invivo study of the binding capacity of clays,pregnant rats were fed varying amounts ofclay in a nutritionally complete diet (Ed-wards et al. 1983). The rats as well as theirpups suffered skeletal and fur changes andslowed development, but exhibited no differ-ences in hemoglobin or red blood cell countafter 60 days; other nutrient indices were notevaluated. This suggests that some nutrientsmay have been chelated, but the data areinconclusive. Based on these few, small ex-perimental studies, we can conclude thatsome geophagic clays interfere with absorp-tion of cations, which can, in turn, result innutrient deficiencies.

In sum, few of the available data supportthe hypothesis that geophagy functions toameliorate mineral nutrient deficiencies. Infact, if clays bind dietary nutrients, this couldhelp to explain the association between ge-ophagy and anemia: eating certain earthmight actually cause nutrient deficiencies. Itis important to note, however, that presentlyavailable data on iron and zinc bioavailabilityare limited in quality and quantity, and in-formation on the bioavailability of calciumand other minerals is fragmentary. More re-search is needed in this area.

Geophagy After the Resolution of aDeficiency

If geophagy were a response to nutrientdeficiency, the resolution of that deficiency


should result in cessation of the behavior.Although the literature is not extensiveenough to rigorously test this corollary, therehave been a few studies that assessed theeffect of resolving a nutrient deficiency onpica behavior.

Three single-blinded studies suggestedthat iron supplementation resulted in ces-sation of pica behavior, including geoph-agy (McDonald and Marshall 1964; Mohanet al. 1968; Rogers 1972). However, therewere numerous problems with the studydesigns, including a lack of controls, smallsample sizes, and poor measurement of ironstatus, all of which make it impossible toattribute behavioral changes to iron sup-plementation alone (Reid 1992; Young2010).

There have been two controlled double-blind studies of iron supplementation andpica. In the first (Gutelius et al. 1962), nocorrelation was found between changes inhemoglobin concentration and changes inpica. In the second study (Nchito et al.2004), which focused specifically on geoph-agy rather than pica more generally, neitherrandomization to 10 months of iron supple-mentation nor 10 months of multivitaminssignificantly reduced geophagy among 402Zambian schoolchildren. Based on multivar-iate logistic models, the authors concludedthat neither iron supplementation nor multi-micronutrient supplementation were signif-icant predictors of geophagy (p�0.44,p�0.88, respectively). Thus, experimentaldata do not support the hypothesis thatchanges in iron status alter geophagic behav-ior.

There have been three studies investigat-ing the effects of zinc supplementation onpica behavior (Bhalla et al. 1983; Chen etal. 1985; Lofts et al. 1990). Pica decreasedafter administration of zinc in all three, butit is not clear that this was attributable tothe zinc supplementation because therewas no indication of other messages givento the subjects, no controls, and no evi-dence of increase in zinc levels of subjectsin two of the three studies (Bhalla et al.1983; Chen et al. 1985). In sum, availabledata are insufficient to permit conclusions

about the efficacy of zinc supplementationin causing cessation of pica (Young 2010).

In the sole study of the effects of calcium(Gutelius et al. 1963), experimental sup-plementation had no effect on pica behav-ior.

adaptive hypothesis 2: protectionAssociation with Gastrointestinal Distress

The protection hypothesis predicts thatgeophagy should often be associated withgastrointestinal distress. Indeed, geophagywas associated with symptoms of gastrointes-tinal malaise (e.g., diarrhea, stomach pain,flatulence) in 48 of the 50 (96%) reports inwhich the occurrence of gastrointestinal dis-tress was recorded (Figure 3b). This is signif-icantly higher than would be predicted bythe null hypothesis that there is no differ-ence in geophagy by gastrointestinal malaise(p�0.001).

Clay Content of EarthsConsistent with the protection hypothe-

sis, geophagists are highly selective aboutthe earth they eat. In 237 of 243 culturalreports (98%) with descriptions, geoph-agic earth was described as clay-like (Fig-ure 6b). Geophagists regularly expressedpreferences for earth that was clay-like orsmooth rather than gritty or sandy (e.g., vonHumboldt et al. 1821; Beccari 1904; Ver-meer and Frate 1979). Individuals sometimeswent to great lengths to obtain clay-rich earth.They were willing to walk many kilometers toreach a site where a deposit of appropriateclay occurred (e.g., Forsyth and Benoit1989). Even among clay-rich earths, therewere explicit favorites. For example, one hus-band who dug clay for his wife from a depositthat was closer to home and less public thanher preferred site; after she tasted it, she senthim back to get the exact clay she craved(Finger 1993).

Pathogen Content of Geophagic EarthsUnder the protection hypothesis,

geophagic earth should not be a vectorfor the transmission of parasites andother pathogens. However, parasitic in-


fections, especially by geohelminths,have sometimes been attributed to ge-ophagy (Hooper and Mann 1906; Anelland Lagercrantz 1958; Halsted 1968;Glickman et al. 1999). Although the par-asite and pathogen contents of ingestedearth were not quantified in the culturalreports in our database, there are threereasons to believe that geophagic soilstypically are not vectors of geohelminthtransmission.

First, geophagists typically select subsoilsthat are less likely to contain geohelmintheggs than earth closer to the surface, wheredefecation occurs (Young et al. 2007). Sec-ond, geophagists carefully prepare the earththey eat. In 118 of 120 cultural reports(98%), geophagic soils were heated, sifted,dried, or brushed off prior to consumption,rather than being excavated and immedi-ately consumed (Figure 6c). Indeed, in In-donesia, Ghana, India, and Guatemala, anindustry developed around geophagy that in-volved excavators, traders, and vendors(Anonymous 1881; Hooper and Mann 1906;Vermeer 1971; Hunter et al. 1989). In otherplaces, clay preparation is handled on anindividual basis, either by the consumer orsomeone else in the household. Whether ona large or small scale, the preparation ofgeophagic earth usually involves: (1) remov-ing impurities by crushing the earth andthen sifting out sand and small stones, pick-ing off the outer crust of earth, or sometimessieving it through cloth; and (2) baking, fry-ing, sun drying, or smoking the earth. Thesepreparation practices likely kill most endo-parasites and other pathogens.

Third, there is little evidence to supportthe transmission of hookworm by geophagy(Gelfand 1945; Heymann 2004), especiallysince hookworms are spread transdermally.There is conflicting evidence about whethergeophagy might be a mechanism of trans-mission of whipworms (Trichuris) or round-worms (Ascaris). None of the preparedgeophagic earths from Tanzania sampledby Young et al. (2007) contained live geo-helminths, but in two other studies (Wonget al. 1991; Geissler et al. 1998), viablehelminths were discovered in geophagicsoils. This difference may be explained by

the latter being conducted on soils eatenby children, who may have been more care-less than adults about preparing the soilbefore consuming it (Young et al. 2007).

Geophagy by Climate TypeThe protection hypothesis predicts that

individuals who are frequently exposed toharmful foodborne microbes should fre-quently engage in geophagy. Foodbornepathogens multiply rapidly in hot, humid,tropical climates (Hui et al. 2001a), andspecies of pathogens and infectious dis-eases are more diverse in equatorial areasthan in more northern latitudes (Guernieret al. 2004). Thus, we would expect geoph-agy to occur most frequently at low latitudesand altitudes. Although geophagy occursthroughout the world (Figure 1), it is espe-cially common in tropical climate zones andexceedingly rare in polar and cold climates(Figure 7). The proportion of cultures intropical areas that practice geophagy ismuch higher than would be predicted byeither the distribution of cultural groups inthe SCCS (p�0.0001) or the worldwidepopulation distribution by climate region(p�0.0001).

0%

25%

50%

75%

Polar Cold Dry Temperate Tropical

Pro

porti

on o

f cul

ture

s

Koeppen climate typeCultural reports in Pica Literature Database (n=361)Groups in Standard Cross-Cultural Sample (n=186)World population (Staszewski 1963)

*

Figure 7. Distribution of Geophagic Culturesin the Pica Literature Database (SeeAlso Figure 1), Standard Cross-Culture Sample, and WorldPopulation Distribution by ClimateType

The proportion of cultures in tropical areas thatpractice geophagy is significantly higher (*) thanwould be predicted by either the distribution of cul-tural groups in the SCCS or the worldwide populationdistribution by climate region (p�0.0001).


Geophagy and Susceptibility to Toxinsand Pathogens

Under the protection hypothesis, peo-ple should engage in geophagy whenthey are most susceptible to the harmfuleffects of toxins, parasites, and otherpathogens. Particularly susceptible lifestages are those during which rapidgrowth and cell division is occurring(i.e., embryogenesis and preadolescence;Bearer 1995). Furthermore, pregnantwomen are adaptively immunosup-pressed (to avoid rejecting the embryo), soavoidance of parasites and pathogens is espe-cially important for the woman’s own healthduring pregnancy (Flaxman and Sherman2000). Therefore, within a culture in whichgeophagy occurs, the hypothesis predicts that itshould occur more frequently among preg-nant women and children than among anyother age or sex groups.

Indeed, the data indicate that pregnantwomen and preadolescents consumedearth most frequently (Figures 2 and 4). Inmultilevel, mixed-effect regression modelsof geophagy score, in which life stage isone of the independent variables and preg-nant women is the reference level, the betacoefficients for all other life stages are sig-nificantly smaller. In other words, preg-nant women consumed earth significantlymore often than any other life stage group(p�0.0001). Thus we can reject the hy-pothesis that life stage cannot predict ge-ophagy behavior.

In fact, there are more total reports ofgeophagy in pregnant women than reportsof geophagy for all other life stages com-bined. Geophagy is so closely associatedwith pregnancy that the consumption ofearth has been termed “a sign of the com-mencement of pregnancy” (Hooper andMann 1906:254). In many countries, ge-ophagy is thought of as a behavior uniqueto pregnancy, e.g., “It would be very sur-prising if pregnant women in Malawi didnot eat clay. That’s how you know whenyou are pregnant!” (Hunter 1993:75).

The protection hypothesis further pre-dicts that geophagy should be more frequentearly in pregnancy when embryonic tissues

are most susceptible to damage from terato-gens (Moore and Persaud 1998; Flaxmanand Sherman 2000). Consistent with this, 10of the 15 accounts that discussed the timingof geophagy during pregnancy indicatedthat consumption occurred during the firsttrimester (Figure 5).

non-adaptive hypothesis 3a: hungerIf geophagy were a non-adaptive behav-

ior that occurs when a famished personattempts to fill an empty stomach, wewould expect it to occur most often intimes of food shortage or famine. We lo-cated data on hunger status of geophagistsin 72 cultural reports (Figure 3c). Amongthese reports, geophagy was attributedsolely to hunger in only 16 (22%). In con-trast, hunger was explicitly not associatedwith geophagy in 36 reports (50%). Insome of these reports, the adequacy of thefood supply of geophagists was commentedon directly by observers (Buckingham1842) or indicated indirectly through discus-sion of the obesity of geophagists (e.g., Ver-meer and Frate 1979) or the wealth of thosewho consumed earth (e.g., Gautier and Mc-Quoid 1853; Livingstone and Waller 1875).In the remaining 20 reports (28%), earthwas sometimes eaten out of hunger whileother times for “pleasure,” “custom,”“craving,” or “habit” (e.g., von Humboldt etal. 1821). This distribution is significantly dif-ferent (p�0.006) than would be expectedunder the null hypothesis.

The relative frequencies of geophagy byvarious sex and age groups (Figure 2) and itstiming within pregnancy (Figure 5) also of-fer no support for hunger motivating geoph-agy. For example, men and non-pregnantwomen have similar caloric requirementsand so would be equally likely to engage ingeophagy under this hypothesis. However,women practice geophagy much more oftenthan men. Pregnant women require morecalories (and are thus more likely to be hun-gry) late in pregnancy when the embryo islarge and growing rapidly than early in preg-nancy (Institute of Medicine 1990). Yet ge-ophagy occurs more commonly in early thanin late gestation (Figure 5). Finally, lactat-


ing women have the greatest caloric re-quirements of all the groups (Institute ofMedicine 1990) and are therefore likely tobe hungry most often. However, they arenot the most frequent geophagists.

Earth Selection and PreparationUnder the non-adaptive hypothesis, if

earth were being eaten simply to fill anempty stomach, any sort of earth (or anyother non-toxic substance) should do. How-ever, consumers were highly selective aboutthe earth they ate (Figure 6). Indeed, we didnot find a single report in which any type ofearth was desired. Of the 77 cultural reportsthat mention how people felt about the earththey were eating, 72 (93%) explicitly dis-cussed their desire for specific types ofearth (Figure 3d), usually clay-rich soils. Ifearth was consumed as a last resort in theface of hunger, we would not expect de-scriptions such as “a devouring passion”(Galt 1872:403), “enjoyed” (Walker 1910:220), and “great attachment” (Shannon1794:375).

Quantity ConsumedFinally, if hunger motivated geophagy,

we would expect that enough earth wouldbe eaten to sate the appetite, i.e., to fill thegeophagist’s stomach. Although most re-ports did not quantify the amount of earthconsumed, many described it qualitatively.The amount usually was small, for example:“a few morsels” (Maupetit 1911:179), “size ofa hazelnut” (Garnier 1871:283), and “lumpthe size of an egg” (Whiting 1947:611). In 11reports, the amount of earth was weighed.The modal amount that an individual con-sumed was approximately 30g, although inthree cases more than 100g was reportedlyeaten. Recently, more rigorous biomedicalstudies have recorded consumption of 30–50g of earth per individual (Geissler et al.1997; Saathoff et al. 2002; Luoba et al. 2005;Young et al. 2010). The implication is thatthe usual amount of earth consumed issmall, more like a medicament than a meal.Although clays can expand in volume in amoist environment, it is unlikely that the

small quantities ingested would quell hungerpains, and certainly not for very long sincegeophagic earth provides no energy.

non-adaptive hypothesis 3b:epiphenomenon of nutrient

deficiencyIf geophagy were a non-adaptive epi-

phenomenon of nutrient deficiencies, wewould expect the behavior to be associatedwith such deficiencies. In our database,anemia and geophagy are associated, butthe cross-sectional nature of the data donot permit determination of temporality.If, however, a deficiency caused geophagy,the cessation of geophagy should occurupon supplementation with deficient nu-trient. This is usually not the case (see theearlier section, Geophagy After the Reso-lution of a Deficiency). Research on theneurological and sensory consequences ofnutrient deficiencies is lacking, but withcurrent data, there is little support for thishypothesis.

Geophagy in Non-human AnimalsGeophagy also occurs in a wide range of

non-human vertebrates (see supplemen-tary material, Tables 2 and 3, available athttp://journals.uchicago.edu/QRB). We lo-cated 79 accounts of its occurrence in 57 spe-cies of primates (Table 2) and 251 accounts ofgeophagy in 240 species of other vertebrates(Table 3), including mammals (in 29 families),birds (in 13 families), and reptiles (in five fam-ilies). Geophagy is likely far more common,but has gone unnoticed because detailed, long-term observations of that species’ dietary habitshave not been made.

Among primates, multiple hypotheses forgeophagy have received empirical support.Krishnamani and Mahaney (2000:899) con-cluded that “mineral supplementation, ad-sorption of toxins, treatment of diarrhoeaand pH adjustment of the gut seem the mostplausible reasons why primates engage in ge-ophagy.” In our tabulation of primate geoph-agy (Table 2), 49 of 79 accounts describedgeophagy as an adaptive behavior, and in theother 30 accounts, the probable function ofgeophagy was not specified. Among the


adaptive reports, 32 (65%) attributed geoph-agy as probably or definitely motivated by thedetoxification of plant secondary com-pounds or protection from parasites andpathogens (including treatment of diar-rhea), and 32 (65%) attributed geophagy asprobably or definitely related to obtainingnutrients. The proportions sum to morethan 100% because some authors proposedmultiple explanations. The vast majority of pri-mates in which geophagy has been observedinhabit tropical areas, and detoxification wasinferred most commonly in leaf- and fruit-eating primates. Ingested earth most fre-quently came from the forest floor and termitemounds, and often was described as having aclay-like consistency.

Among vertebrates other than primates(Table 3), 176 of the 251 accounts (70%)indicated that geophagy was definitely orprobably an adaptive behavior, whereasonly 5 (2%) indicated that it was non-adaptive. In the other accounts the proba-ble function of geophagy was not specified.Among the adaptive reports, 88% attrib-uted geophagy to obtaining mineral nutri-ents (primarily salt and calcium) and 19%attributed geophagy to detoxification ofplant secondary compounds. The propor-tions sum to more than 100% becausesome authors proposed multiple explana-tions.

The possibility that geophagy providesprotection from parasites and pathogenswas rarely considered in non-primates, andthe sex and reproductive status of geopha-gists were infrequently mentioned. How-ever, geophagy was described by sex andpregnancy status among 17 species of neo-tropical, fruit-eating bats (Bravo et al. 2008,2010). Intriguingly, females engaged ingeophagy more often than males, andpregnant females did so more often thannon-pregnant females.

Studies with laboratory animals indicatethat geophagy can provide relief from gas-trointestinal distress. Rats cannot rid them-selves of toxins through emesis, but when theyare poisoned experimentally they preferen-tially ingest kaolin, which reduces poison-related morbidity and mortality (Mitchell et al.1976; Burchfield et al. 1977; Watson et al. 1987;

Takeda et al. 1993; Madden et al. 1999). Simi-lar results were observed in experiments onparrots (Gilardi et al. 1999).

Animal geophagy has been observed in awide range of climate and habitat types,although the majority of studies (especiallyof birds) were conducted in the tropics.Geophagy was most often detected duringobservations at traditionally used “claylicks,” “mineral licks,” and “salt licks,” andneed for nutrients (especially sodium andcalcium) was most commonly inferred asthe function of geophagy (especially in un-gulates). Detoxification was inferred pri-marily in tropical, fruit-eating birds such asparrots and pigeons. However, the soils atmany of the traditional mineral licks weredescribed as clay-rich in composition, so de-toxification and protection against patho-gens may be more common than is currentlyrecognized. Indeed, some studies (Tables 2and 3) attributed geophagy to both detoxifi-cation and micronutrient acquisition.

Challenges to the Interpretation ofData

There are several reasons for caution in in-terpreting our results. First is the danger ofunderreporting, which is inherent in any liter-ature review. Human and animal geophagy isunquestionably underreported because it iseasily missed, even by trained observers. Inhumans, geophagists and local informantsmay attempt to conceal the behavior for fearof being judged negatively or chastised, orbecause earth eating may be interpreted asan indication of pregnancy status (whichsome may want to keep private), poverty,or lack of self control (Hooper and Mann1906; Dickins and Ford 1942; Sayetta 1986).Even when geophagists are not furtive, inves-tigators may not know to inquire about eartheating specifically, and they “discover” ge-ophagy only by accident (e.g., Hooper andMann 1906; Vermeer 1966; Cooksey 1995;Rainville 1998; Grigsby et al. 1999).

Second, for human geophagy, there isconsiderable variability in objectivity andthoroughness among studies. Althoughjudgmental language is stripped from ourdatabase, we wonder if the revulsion ge-ophagy sometimes elicits has colored pub-


lished reports by limiting the amount ofinformation investigators pursued or weregiven in the course of fieldwork. Precon-ceived notions about who engages in ge-ophagy may have also biased some of thereports. For example, if geophagy was at-tributed to “the weaker sex” (Maler 1692),the observer might not try to find out howoften men practiced geophagy.

Third, there is an additional difficulty instudying human geophagy, and pica ingeneral: it does not easily fit into a specificcultural conceptual category. From a cul-tural perspective, the people being inter-viewed may think of pica substances asmedicines, food additives, or just cravings,and food recall questionnaires generallydo not probe these issues with appropriateprompts (Young and Ali 2005). Thus, be-cause we can only report on positive obser-vations of geophagy, there are likely to besome false negatives (Type 2 errors), i.e.,societies in which geophagy occurs, but hasnot been documented.

Fourth, in humans some misclassifica-tion errors probably occurred (see Meth-odology). However, classifying a pregnantwoman as non-pregnant or an anemic per-son as non-anemic would only weaken thetrends we discovered. And misclassification,underreporting, or false negatives would notexplain the significant differences we docu-mented among categories of geophagistswithin societies, including the relationshipsbetween geophagy and anemia (Figure 3a),gastrointestinal distress (Figure 3b), the pre-dominance of the behavior among pregnantwomen and children (Figure 2), or the oc-currence of geophagy in early pregnancy(Figure 5).

DiscussionThree hypotheses have been proposed

for the functional significance of humangeophagy. Of these, the non-adaptive hy-pothesis that geophagy is an attempt to filla hungry stomach explains few cases. Ge-ophagy occurs when food is plentifullyavailable. Moreover, that small quantitiesof earth are consumed, the age and sexbiases, and the frequent association withstrong cravings for specific types of earth

are not consistent with the hunger hypoth-esis. The second non-adaptive hypothesis,that geophagy is an epiphenomenon of nu-trient deficiencies that cause neurologicalor sensory problems, also is not supportedby available data. Nutrient supplementa-tion does not regularly cause the cessationof geophagy.

The first adaptive hypothesis is that ge-ophagy results from nutrient deficiencies.Seemingly consistent with this hypothesis, ge-ophagy is frequently associated with anemia.However, the timing of geophagy does notparallel the timing of changes in nutrientneeds through the life span, nor within preg-nancy. The irregular presence and low bio-availability of calcium, zinc, and iron ingeophagic earth, the fact that iron supple-mentation does not reduce geophagic be-havior, and the experimental data indicatingthat micronutrient absorption is limited afterthe consumption of earth cast doubt on thishypothesis.

The second adaptive hypothesis is thatgeophagy is a mechanism of protectionagainst plant toxins, parasites, and otherpathogens. Consistent with this hypothesisis the association of geophagy with gastro-intestinal distress and with consumption oftoxic substances, the high clay content ofmost geophagic soils (clay adsorbs danger-ous chemicals and pathogens), the occur-rence of geophagy in areas of the world withthe highest parasite and pathogen densities(the tropics), and the sex bias and timing ofgeophagy during periods of greatest suscep-tibility to harm from parasites, pathogens,and toxins (childhood and early in preg-nancy).

Use of clay in food preparation is a well-known means of neutralizing toxins (Johnsand Duquette 1991; Johns 1996). In 27reports in the Pica Literature Database,clay was used in the preparation of majorfood items, e.g., staple crops and fish (honey,salt, or oil were not considered major fooditems). In ten of these cultures (37%), claywas used in the preparation of or eaten withfoods that contain harmful substances, suchas Andean potatoes (which contain glycoal-kaloids; Johns 1996) and Sardinian acorns(high in tannins; Wagner and Cortes 1921;


Usaı and Mazzarella 1969). In WesternAustralia, aborigines used clay in the prep-aration of mene, a tuber known to causediarrhea when ingested raw (Grey 1841).Several of the cultural reports in our databasecontained explicit information about the asso-ciation of geophagy with exposure to toxins.For example, people in the Northern Territoryof Australia explained that they ate clay to “linethe stomach” before eating fish they knew tobe poisonous (Grey 1841).

The detoxifying properties of clay mayeven explain some of the geophagy that hasbeen observed in times of food shortages(Figure 3c). When people are forced to eatplant parts they would normally avoid due tothe secondary compounds they contain(e.g., weed stems, bark, and roots), con-sumption of small amounts of clay couldreduce the dangers associated with ingest-ing these marginal foods by binding withthe toxic chemicals that typically makethem unpalatable (Johns 1996).

Occurrences of geophagy in non-humanprimates and other vertebrates also sup-port the two adaptive hypotheses over thenon-adaptive alternative. However, no con-clusions can be drawn about the relativeimportance of micronutrient deficienciesversus protection against plant toxins inthe occurrence of non-human geophagy,primarily because the possibility that ge-ophagy provides protection against para-sites and pathogens was rarely consideredfor non-primates. In primates, geophagywas attributed to protection from toxinsand to micronutrient deficiencies with ap-proximately equal frequencies (Table 2),whereas in other vertebrates geophagy wastypically attributed to nutrient deficiencies(Table 3). Whether this apparent differ-ence is real is impossible to determine be-cause many studies of primates specificallyconsidered the protection function of ge-ophagy, whereas most studies of other ver-tebrates did not. The frequency with whichgeophagy was reported in tropical leaf- andfruit-eating birds and mammals suggeststhat detoxification of plant secondary com-pounds is a more important function ofthe behavior than presently is realized.

If protection from pathogens and detox-

ification of plant secondary compoundsare the primary functions of geophagy,what do we make of the strong and consis-tent associations between geophagy andanemia in humans? There are two possibil-ities, both of which pertain to the complexand delicate balance between iron statusand infection. Anemia can be an adaptivebodily response to infections, a nutritionaladaptation whereby the sequestration ofcertain nutrients can protect against patho-genic agents (Prentice et al. 2007; Wanderet al. 2009). Many foodborne bacteria re-quire iron to reproduce and iron sequestra-tion reduces bacterial growth rates. Underthis hypothesis, the relationship betweenanemia and geophagy is correlational butnot causal, that is, both the ingestive behav-ior and the physiological response are adap-tations to minimize the severity of foodbornebacterial infections.

The second possibility is that ingestion ofgeophagic earth not only inhibits parasites andother pathogens but also impedes iron absorp-tion, either by binding with dietary iron directly(Hooda et al. 2004) or with the mucin layer inthe small intestine (Leonard et al. 1994)thereby making it difficult for bound iron mol-ecules to pass through the brush border. Un-der this scenario, the relationship between ge-ophagy and anemia is causal—i.e., geophagycauses anemia as a side effect of its anti-parasite/pathogen benefits. Information thatis presently available is insufficient to decidebetween these alternatives. Regardless of whichis correct, the anemia-geophagy correlationcould be consistent with the protection hy-potheses.

Further tests of the two adaptive hypotheseswould be useful. In terms of the protectivehypothesis, it would be illuminating to com-pare the amounts and toxicities of plant sec-ondary compounds and foodborne parasitesand pathogens in the diets of geophagic andnon-geophagic human societies and animalspecies, and among individuals within thosesocieties or populations. Exposing laboratoryanimals to biotic enemies and toxins, and thenfeeding them geophagic earth or a placebowould also help quantify the protective effectsof geophagic soils. A third test of this hypoth-esis would be to establish the capacity of


geophagic earths to bind harmful toxins,pathogens, and endoparasites in in vivo con-ditions.

We hope this paper stimulates such re-search. More importantly, we hope readersagree that it is time to stop regarding geophagyas a bizarre, non-adaptive gustatory mistake.Our data indicate clearly that geophagy is awidespread behavior in humans and other ver-tebrates that occurs during both vulnerable lifestages and when facing ecological conditionsthat require protection.

acknowledgments

We thank Daniel Dykhuizen, Kathleen Rasmussen,Rebecca Stoltzfus, and several anonymous reviewers

for their insights and useful comments on the man-uscript. We would also like to acknowledge thetireless efforts of the team of library scientists andstaff at the Cornell libraries, especially the Interli-brary Loan Office; they were integral to trackingdown many obscure documents for the Pica Liter-ature Database. We greatly appreciate the transla-tions provided by Benedetta Bartali, Jen Baker,Urvashi Batra, Brian English, Tim Haupt, Jacque-line Kung’u, Helena Pachon, Rinat Ran-Ressler,Angelos Sidakalis, Owen Strijland, Vincenzo Vitelli,and Winthrop “Skip” Wetherbee. We are gratefulfor the following sources of financial support: theHertz Foundation (Julius B. Lucks), the NationalInstitutes of Health-TG #5 T32 HD007331 (Sera L.Young), The Weiss Presidential Fellowship Fund atCornell University (Paul W. Sherman), andWenner-Gren Foundation (Sera L. Young).

REFERENCES

Abrahams P. W. 1997. Geophagy (soil consumption)and iron supplementation in Uganda. TropicalMedicine and International Health 2:617–623.

Abrahams P. W., Follansbee M. H., Hunt A., Smith B.,Wragg J. 2006. Iron nutrition and possible leadtoxicity: an appraisal of geophagy undertaken bypregnant women of UK Asian communities. Ap-plied Geochemistry 21:98–108.

Anell B., Lagercrantz S. 1958. Geophagical Customs.Uppsala (Sweden): Studia Ethnographica Upsa-liensia.

Anonymous. 1881. Land- en volkenkunde van Neer-landsch Indie: Geophagie. Mededeelingen van wegehet Nederlandsche Zendelinggenootschap 25:293–296.

Arcasoy A., Cavdar A. O., Babacan E. 1978. Decreasediron and zinc absorption in Turkish children withiron deficiency and geophagia. Acta Haematologica60:76–84.

Baudouin M. 1924. A propos de l’abrasion des dents.La Semaine dentaire 6:82–84.

Bearer C. F. 1995. Environmental health hazards: howchildren are different from adults. The Future ofChildren 5:11–26.

Beccari O. 1904. Wanderings in the Great Forests of Bor-neo. Singapore: Oxford University Press.

Bhalla J. N., Khanna P. K., Srivastava J. R., Sur B. K.,Bhalla M. 1983. Serum zinc level in pica. IndianPediatrics 20:667–670.

Brady J. E., Rissolo D. 2006. A reappraisal of ancientMaya cave mining. Journal of Anthropological Re-search 62:471–490.

Bravo A., Harms K. E., Emmons L. H. 2010. Puddlescreated by geophagous mammals are potentialmineral sources for frugivorous bats (Stenoderma-tinae) in the Peruvian Amazon. Journal of TropicalEcology 26:173–184.

Bravo A., Harms K. E., Stevens R. D., Emmons L. H.2008. Collpas: activity hotspots for frugivorous bats(Phyllostomidae) in the Peruvian Amazon. Biotro-pica 40:203–210.

Browman D. L., Gundersen J. N. 1993. Altiplano co-mestible earths: prehistoric and historic geophagyof highland Peru and Bolivia. Geoarcheology 8:413–425.

Buckingham J. S. 1842. The Slave States of America:Volume I. London (UK): Fisher Son and Company.

Burchfield S. R., Elich M. S., Woods S. C. 1977.Geophagia in response to stress and arthritis. Phys-iology & Behavior 19:265–267.

Cavdar A. O., Arcasoy A. 1972. Hematologic and bio-chemical studies of Turkish children with pica: apresumptive explanation for the syndrome ofgeophagia, iron deficiency anemia, hepatospleno-megaly and hypogonadism. Clinical Pediatrics 11:215–227.

Cavdar A. O., Arcasoy A., Cin S., Babacan E., Gozda-soglu S. 1983. Geophagia in Turkey: iron and zincabsorption studies and response to treatment withzinc in geophagia cases. Pages 71–97 in Zinc Defi-ciency in Human Subjects: Proceedings of an Interna-tional Symposium held in Ankara, Turkey, April 29–30, 1982, edited by A. S. Prasad, A. O. Cavdar, G. J.Brewer, and P. J. Aggett. New York: Alan R. Liss,Inc.

Celsus A. C., Grieve J. 1756. Of Medicine. London(UK): D. Wilson and T. Durham.

Chen X. C., Yin T. A., He J. S., Ma Q. Y., Han Z. M.,Li L. X. 1985. Low levels of zinc in hair and blood,pica, anorexia, and poor growth in Chinese pre-school children. American Journal of Clinical Nutri-tion 42:694–700.

Chisholm J. C., Jr., Martin H. I. 1981. Hypozincemia,


ageusia, dysosmia, and toilet tissue pica. Journal ofthe National Medical Association 73:163–164.

Clark J. D. 2001. Kalambo Falls Prehistoric Site: VolumeIII. Cambridge (UK): Cambridge University Press.

Cooksey N. R. 1995. Pica and olfactory craving ofpregnancy: how deep are the secrets? Birth 22:129–137.

Danford D. E. 1982. Pica and nutrition. Annual Reviewof Nutrition 2:303–322.

Dickins D., Ford R. N. 1942. Geophagy (dirt eating)among Mississippi Negro school children. Ameri-can Sociological Review 7:59–65.

Ditter B., Urbaschek R., Urbaschek B. 1983. Ability ofvarious adsorbents to bind endotoxins in vitro andto prevent orally induced endotoxemia in mice.Gastroenterology 84:1547–1552.

Dominy N. J., Davoust E., Minekus M. 2004. Adaptivefunction of soil consumption: an in vitro studymodeling the human stomach and small intestine.Journal of Experimental Biology 207:319–324.

Dornai D., Mingelgrin U., Frenkel H., Bar-Joseph M.1993. Direct quantification of unadsorbed virusesin suspensions of adsorbing colloids with theenzyme-linked immunosorbent assay. Applied andEnvironmental Microbiology 59:3123–3125.

Dreyer M. J., Chaushev P. G., Gledhill R. F. 2004.Biochemical investigations in geophagia. Journal ofthe Royal Society of Medicine 97:48.

Edwards A. A., Mathura C. B., Edwards C. H. 1983.Effects of maternal geophagia on infant and juve-nile rats. Journal of the National Medical Association75:895–902.

Edwards C. H., Johnson A. A., Knight E. M., OyemadeU. J., Cole O. J., Westney O. E., Jones S., Laryea H.,Westney L. S. 1994. Pica in an urban environment.Journal of Nutrition 124:954S–962S.

Ferrand E. 1886. Terres comestibles de Java. Revued’Ethnographie 5:548–549.

Ferrari S. F., Veiga L. M., Urbani B. 2008. Geophagyin New World Monkeys (Platyrrhini): ecologicaland geographic patterns. Folia Primatologica 79:402–415.

Finger M. 11 December 1993. The clay eaters of Mem-phis. Memphis.

Flaxman S. M., Sherman P. W. 2000. Morning sick-ness: a mechanism for protecting mother and em-bryo. Quarterly Review of Biology 75:113–148.

Forsyth C. J., Benoit G. M. 1989. ‘Rare, old, dirtysnacks’: some research notes on dirt eating. Devi-ant Behavior 10:61–68.

Galt F. 1872. Medical notes on the upper Amazon.American Journal of Medical Sciences 63:395–416.

Gardiner K. R., Anderson N. H., McCaigue M. D.,Erwin P. J., Halliday M. I., Rowlands B. J. 1993.Adsorbents as antiendotoxin agents in experimen-tal colitis. Gut 34:51–55.

Garnier J. 1871. Voyage Autour du Monde. La Nouvelle-Caledonie (cote orientale). Paris (France): E. Plon.

Gautier T., McQuoid T. R. 1853. Wanderings in Spain.London (UK): Ingram, Cooke and Company.

Geissler P. W. 2000. The significance of earth-eating:social and cultural aspects of geophagy amongLuo children. Africa: The Journal of the InternationalAfrican Institute 70:653–682.

Geissler P. W., Mwaniki D. L., Thiong’o F., Friis H.1997. Geophagy among school children in west-ern Kenya. Tropical Medicine and InternationalHealth 2:624–630.

Geissler P. W., Mwaniki D. L., Thiong’o F., Friis H.1998. Geophagy as a risk factor for geohelminthinfections: a longitudinal study of Kenyan primaryschoolchildren. Transactions of the Royal Society ofTropical Medicine and Hygiene 92:7–11.

Geissler P. W., Prince R. J., Levene M., Poda C., Beck-erleg S. E., Mutemi W., Shulman C. E. 1999. Per-ceptions of soil-eating and anaemia among preg-nant women on the Kenyan coast. Social Scienceand Medicine 48:1069–1079.

Gelfand M. 1945. Geophagy and its relation to hook-worm disease. East African Medical Journal 22:98–103.

Gilardi J. D., Duffey S. S., Munn C. A., Tell L. A. 1999.Biochemical functions of geophagy in parrots: de-toxification of dietary toxins and cytoprotectiveeffects. Journal of Chemical Ecology 25:897–922.

Glickman L. T., Camara A. O., Glickman N. W.,McCabe G. P. 1999. Nematode intestinal parasitesof children in rural Guinea, Africa: prevalence andrelationship to geophagia. International Journal of Ep-idemiology 28:169–174.

Gonzalez R., de Medina F. S., Martınez-Augustin O.,Nieto A., Galvez J., Risco S., Zarzuelo A. 2004.Anti-inflammatory effect of diosmectite in hapten-induced colitis in the rat. British Journal of Pharma-cology 141:951–960.

Grey G. 1841. Journals of Two Expeditions of Discovery inNorth-west and Western Australia, During the Years1837, 38, and 39, Under the Authority of Her Majesty’sGovernment: Describing Many Newly Discovered, Impor-tant, and Fertile Districts, With Observations on theMoral and Physical Condition of the Aboriginal Inhab-itants. London (UK): T. and W. Boone.

Grigsby R. K., Thyer B. A., Waller R. J., Johnston G. A.,Jr. 1999. Chalk eating in middle Georgia: a cul-ture-bound syndrome of pica? Southern MedicalJournal 92:190–192.

Guernier V., Hochberg M. E., Guegan J.-F. 2004. Ecol-ogy drives the worldwide distribution of humandiseases. PLoS Biology 2:e141. doi:10.1371/journal.pbio.0020141.

Gutelius M. F., Millican F. K., Layman E. M., CohenG. J., Dublin C. C. 1962. Nutritional studies ofchildren with pica. Pediatrics 29:1012–1023.


Gutelius M. F., Millican F. K., Layman E. M., CohenG. J., Dublin C. C. 1963. Treatment of pica with avitamin and mineral supplement. American Journalof Clinical Nutrition 12:388–393.

Halsted J. A. 1968. Geophagia in man: its nature andnutritional effects. American Journal of Clinical Nu-trition 21:1384–1393.

Henon P., Gerota I., Caen J. 1975. Letter: one canremain a geophagist in Paris. La Nouvelle PresseMedicale 4:1431.

Heymann D. L. 2004. Control of Communicable DiseasesManual. Washington (DC): American Public HealthAssociation.

Hippocrates, Adams F. 1849. The Genuine Works ofHippocrates. London (UK): Sydenham Society.

Hladik C. M., Gueguen L. 1974. Geophagie et nutri-tion minerale chez les primates sauvages. ComptesRendus de l’Academie des Sciences, Serie III 279:1393–1396.

Hochstein G. 1968. Pica: a study in medical and an-thropological education. Pages 88–96 in Essays onMedical Anthropology, Southern Anthropological SocietyProceedings, No. 1, edited by T. Weaver. Athens(GA): University of Georgia Press.

Hooda P. S., Henry C. J. K., Seyoum T. A., ArmstrongL. D. M., Fowler M. B. 2004. The potential impactof soil ingestion on human mineral nutrition. Sci-ence of the Total Environment 333:75–87.

Hooper D., Mann H. H. 1906. Earth-eating and theearth-eating habit in India. Memoirs of the AsiaticSociety of Bengal 1:249–273.

Horner R. D., Lackey C. J., Kolasa K., Warren K. 1991.Pica practices of pregnant women. Journal of theAmerican Dietetic Association 91:34–38.

Houston D. C., Gilardi J. D., Hall A. J. 2001. Soilconsumption by elephants might help to minimizethe toxic effects of plant secondary compounds inforest browse. Mammal Review 31:249–254.

Hudson R. P. 1977. The biography of disease: lessonsfrom chlorosis. Bulletin of the History of Medicine51:448–463.

Hui Y. H., Smith R. A., Spoerke D. G., Jr. 2001a.Foodborne Disease Handbook, Volume 1: DiseasesCaused by Bacteria. Second Edition. New York: Mar-cel Dekker.

Hui Y. H., Smith R. A., Spoerke D. G., Jr. 2001b.Foodborne Disease Handbook, Volume 3: Plant Toxi-cants. Second Edition. New York: Marcel Dekker.

Hunter J. M. 1973. Geophagy in Africa and in theUnited States: a culture-nutrition hypothesis. Geo-graphical Review 63:170–195.

Hunter J. M. 1993. Macroterme geophagy and preg-nancy clays in southern Africa. Journal of CulturalGeography 14:69–92.

Hunter J. M., Horst O. H., Thomas R. N. 1989. Reli-gious geophagy as a cottage industry: the holy clay

tablet of Esquipulas, Guatemala. National Geo-graphic Research 5:281–295.

Institute of Medicine. 1990. Nutrition During Pregnan-cy: Part I: Weight Gain, Part II: Nutrient Supplements.Washington (DC): National Academy Press.

Institute of Medicine. 2002. DRI: Dietary ReferenceIntakes for Vitamin A, Vitamin K, Arsenic, Boron,Chromium, Copper, Iodine, Iron, Manganese, Molyb-denum, Nickel, Silicon, Vanadium, and Zinc. Wash-ington (DC): National Academy Press.

Johns T. 1986. Detoxification function of geophagyand domestication of the potato. Journal of Chemi-cal Ecology 12:635–646.

Johns T. 1996. The Origins of Human Diet and Medicine:Chemical Ecology. Tucson (AZ): University of Ari-zona Press.

Johns T., Duquette M. 1991. Traditional detoxifica-tion of acorn bread with clay. Ecology of Food andNutrition 25:221–228.

Jones R. L., Hanson H. C. 1985. Mineral Licks, Geoph-agy, and Biogeochemistry of North American Ungulates.Ames (IA): Iowa State University Press.

Kawai K., Saathoff E., Antelman G., Msamanga G.,Fawzi W. W. 2009. Geophagy (soil-eating) in rela-tion to anemia and helminth infection amongHIV-infected pregnant women in Tanzania. Amer-ican Journal of Tropical Medicine and Hygiene 80:36–43.

Kikouama J. R. O., Le Cornec F., Bouttier S., LaunayA., Balde L., Yagoubi N. 2009. Evaluation of traceelements released by edible clays in physicochem-ically simulated physiological media. InternationalJournal of Food Sciences and Nutrition 60:130–142.

Kottek M., Grieser J., Beck C., Rudolf B., Rubel F.2006. World map of the Köppen-Geiger climateclassification updated. Meteorologische Zeitschrift 15:259–263.

Kraemer S. 2002. Clay, vicks, and gold medal flour.Southern Medical Journal 95:1228–1229.

Kreulen D. A. 1985. Lick use by large herbivores: areview of benefits and banes of soil consumption.Mammalian Review 15:107–123.

Krishnamani R., Mahaney W. C. 2000. Geophagyamong primates: adaptive significance and ecolog-ical consequences. Animal Behavior 59:899–915.

La Billardiere M. 1800. An Account of a Voyage in Searchof La Perouse, Undertaken By Order of the ConstituentAssembly of France, and Performed in the Years 1791,1792, and 1793, in the Recherche and Esperance, Shipsof War Under the Command of Rear-Admiral Brunid’Entrecasteaux, Volume II. London (UK): J. De-brett.

Laufer B. 1930. Geophagy. Field Museum of NaturalHistory—Anthropological Series 18:97–198.

Lavie S., Stotzky G. 1986a. Adhesion of the clay min-erals montmorillonite, kaolinite, and attapulgite


reduces respiration of Histoplasma capsulatum. Ap-plied and Environmental Microbiology 51:65–73.

Lavie S., Stotzky G. 1986b. Interactions between clayminerals and siderophores affect the respirationof Histoplasma capsulatum. Applied and Environmen-tal Microbiology 51:74–79.

Leonard A., Droy-Lefaix M. T., Allen A. 1994. Pepsinhydrolysis of the adherent mucus barrier and sub-sequent gastric mucosal damage in the rat: effectof diosmectite and 16,16 dimethyl prostaglandinE2. Gastroenterologie clinique et biologique 18:609–616.

Limpitlaw U. G. 2010. Ingestion of earth materials forhealth by humans and animals. International Geol-ogy Review 52:726–744.

Lipson S. M., Stotzky G. 1983. Adsorption of reovirusto clay minerals: effects of cation-exchange capac-ity, cation saturation, and surface area. Applied andEnvironmental Microbiology 46:673–682.

Livingstone D., Waller H. 1875. The Last Journals ofDavid Livingstone, in Central Africa: From EighteenHundred and Sixty-Five to His Death. New York:Harpers and Brothers.

Lofts R. H., Schroeder S. R., Maier R. H. 1990. Effectsof serum zinc supplementation on pica behaviorof persons with mental retardation. American Jour-nal of Mental Retardation 95:103–109.

Lotan N., Siderman S., Tabak A., Taitelman U., Mi-hich H., Lupovich S. 1983. In vivo evaluation of acomposite sorbent for the treatment of paraquatintoxication by hemoperfusion. International Jour-nal of Artificial Organs 6:207–213.

Loveland C. J., Furst T. H., Lauritzen G. C. 1989.Geophagia in human populations. Food and Food-ways 3:333–356.

Luoba A. I., Geissler P. W., Estambale B., Ouma J. H.,Alusala D., Ayah R., Mwaniki D., Magnussen P.,Friis H. 2005. Earth-eating and reinfection withintestinal helminths among pregnant and lactat-ing women in western Kenya. Tropical Medicine andInternational Health 10:220–227.

Madden L. J., Seeley R. J., Woods S. C. 1999. Intraven-tricular neuropeptide Y decreases need-induced so-dium appetite and increases pica in rats. BehavioralNeuroscience 113:826–832.

Maigetter R. Z., Pfister R. M. 1975. A mixed bacterialpopulation in a continuous culture with and withoutkaolinite. Canadian Journal of Microbiology 21:173–180.

Maler E. C. F. 1692. Disputatio Medica Inauguralis dePica. Basel (Switzerland): Typis Johann RudolphiGenathii.

Mallory W. H. 1926. China: Land of Famine. New York:American Geographical Society.

Maupetit. 1911. Le geophagisme au Laos Siamois.Bulletin de la Societe medico-chirurgicale de l’IndochineApril:176–181.

McDonald R., Marshall S. R. 1964. The value of irontherapy in pica. Pediatrics 34:558–562.

McKnight T. L., Hess D. 2005. Physical Geography: ALandscape Appreciation. Upper Saddle River (NJ):Pearson Prentice Hall.

Minnich V., Okcuoglu A., Tarcon Y., Arcasoy A., CinS., Yörukoglu O., Renda F., Demirag B. 1968. Picain Turkey: II. Effect of clay upon iron absorption.American Journal of Clinical Nutrition 21:78–86.

Mitchell D., Wells C., Hoch N., Lind K., Woods S. C.,Mitchell L. K. 1976. Poison induced pica in rats.Physiology and Behavior 17:691–697.

Mohan M., Agarwal K. N., Bhutt I., Khanduja P. C.1968. Iron therapy in pica. Journal of the IndianMedical Association 51:16–18.

Moore K. L., Persaud T. V. N. 1998. The DevelopingHuman: Clinically Oriented Embryology. Philadelphia(PA): Saunders.

Murdock G. P., White D. R. 1969. Standard cross-cultural sample. Ethnology 8:329–369.

Naroll R. 1961. Two solutions to Galton’s problem.Philosophy of Science 28:15–39.

Nchito M., Geissler P. W., Mubila L., Friis H., Olsen A.2004. Effects of iron and multimicronutrient sup-plementation on geophagy: a two-by-two factorialstudy among Zambian schoolchildren in Lusaka.Transactions of the Royal Society of Tropical Medicineand Hygiene 98:218–227.

Okonek S., Setyadharma H., Borchert A., KrienkeE. G. 1982. Activated charcoal is as effective asfuller’s earth or bentonite in paraquat poisoning.Klinische Wochenschrift 60:207–210.

Phillips T. D., Afriyie-Gyawu E., Williams J., HuebnerH., Ankrah N.-A., Ofori-Adjei D., Jolly P., JohnsonN., Taylor J., Marroquin-Cardona A., Xu L., TangL., Wang J.-S. 2008. Reducing human exposure toaflatoxin through the use of clay: a review. FoodAdditives and Contaminants: Part A 25:134–145.

Prasad A. S. 2001a. Discovery of human zinc defi-ciency: impact on human health. Nutrition 17:685–687.

Prasad A. S. 2001b. Recognition of zinc-deficiencysyndrome. Nutrition 17:67–69.

Prentice A. M., Ghattas H., Cox S. E. 2007. Host-pathogen interactions: can micronutrients tip thebalance? Journal of Nutrition 137:1334–1337.

Profet M. 1992. Pregnancy sickness as adaptation: adeterrent to maternal ingestion of teratogens.Pages 327–366 in The Adapted Mind: EvolutionaryPsychology and the Generation of Culture, edited byJ. H. Barkow, L. Cosmides, and J. Tooby. Oxford(UK): Oxford University Press.

Rainville A. J. 1998. Pica practices of pregnant womenare associated with lower maternal hemoglobinlevel at delivery. Journal of the American Dietetic As-sociation 98:293–296.


Reid R. M. 1992. Cultural and medical perspectiveson geophagia. Medical Anthropology 13:337–351.

Rey C. 1989. Rotavirus viral diarrhoea: the advantagesof smectite. Annales Paediatrici 196:1–4.

Rogers M. E. 1972. Practice of pica among iron defi-cient pregnant women. [MS thesis]. Auburn (AL):Auburn University.

Saathoff E., Olsen A., Kvalsvig J. D., Geissler P. W. 2002.Geophagy and its association with geohelminth in-fection in rural schoolchildren from northernKwaZulu-Natal, South Africa. Transactions of the RoyalSociety of Tropical Medicine and Hygiene 96:485–490.

Saether O. M., de Caritat P. 1997. Geochemical Processes,Weathering, and Groundwater Recharge in Catchments.Rotterdam (The Netherlands): Balkema.

Said S. A., Shibl A. M., Abdullah M. E. 1980. Influenceof various agents on adsorption capacity of kaolinfor Pseudomonas aeruginosa toxin. Journal of Phar-maceutical Sciences 69:1238–1239.

Sayers G., Lipschitz D. A., Sayers M., Seftel H., BothwellT. H., Charlton R. W. 1974. Relationship betweenpica and iron nutrition in Johannesburg Blackadults. South African Medical Journal 48:1655–1660.

Sayetta R. B. 1986. Pica: an overview. American FamilyPhysician 33:181–185.

Semple K. T., Doick K. J., Jones K. C., Burauel P.,Craven A., Harms H. 2004. Defining bioavailabilityand bioaccessibility of contaminated soil and sed-iment is complicated. Environmental Science andTechnology 38:228A–231A.

Shannon R. 1794. Practical Observations on the Operationand Effects of Certain Medicines in the Prevention andCure of Diseases to which Europeans are Subject in HotClimates, and in these Kingdoms; Particularly Those ofthe Liver, Flux, and Yellow Fever: Applicable Also to thePrevention and Cure of the Scurvy. Written in a Famil-iar Style. Recommended to the Perusal of Every PersonGoing to Sea, and Residing Abroad. To which areAdded, Plain Directions for Private Use in the Absence ofa Physician; and Observations on the Diseases and Dietof Negroes. With a Copious Explanatory Index. London(UK): Vernor and Hood.

Smith B., Rawlins B. G., Cordeiro M. J. A. R., HutchinsM. G., Tiberindwa J. V., Sserunjogi L., TomkinsA. M. 2000. The bioaccessibility of essential andpotentially toxic trace elements in tropical soilsfrom Mukono District, Uganda. Journal of the Geo-logical Society 157:885–891.

Smith J. C., Jr., Halsted J. A. 1970. Clay ingestion(geophagia) as a source of zinc for rats. Journal ofNutrition 100:973–980.

Smith T. K., Carson M. S. 1984. Effect of diet on T-2toxicosis. Advances in Experimental Medicine and Bi-ology 177:153–167.

Spencer T. 19 December 2002. Dirt-eating persists inrural south. Newhouse News Service.

Staszewski J. 1963. Population distribution according

to the climate areas of W. Koppen. The ProfessionalGeographer 15:12–15.

Takeda N., Hasegawa S., Morita M., Matsunaga T.1993. Pica in rats is analogous to emesis: an animalmodel in emesis research. Pharmacology, Biochemis-try, and Behavior 45:817–821.

Talkington K. M., Gant N. F., Jr., Scott D. E., PritchardJ. A. 1970. Effect of ingestion of starch and someclays on iron absorption. American Journal of Obstet-rics and Gynecology 108:262–267.

Thomson J. 1997. Anaemia in pregnant women in east-ern Caprivi, Namibia. South African Medical Journal87:1544–1547.

Tsakala M., Tona L., Tamba V., Mawanda B., VielvoyeL., Dufey J., Gillard J. 1990. In vitro study of theadsorption of chloroquine by an antidiarrhealremedy traditionally used in Africa. Journal dePharmacie de Belgique 45:268–273.

Usaı A., Mazzarella S. 1969. Acorn Bread and Geophagy inSardinia. Cagliari (Italy): Editrice Sarda Fossataro.

Vermeer D. E. 1966. Geophagy among the Tiv ofNigeria. Annals of the Association of American Geog-raphers 56:197–204.

Vermeer D. E. 1971. Geophagy among the Ewe ofGhana. Ethnology 10:56–72.

Vermeer D. E., Frate D. 1979. Geophagia in ruralMississippi: environmental and cultural contextsand nutritional implications. American Journal ofClinical Nutrition 32:2129–2135.

von Bonsdorff B. 1977. Pica: a hypothesis. British Jour-nal of Haematology 35:476–477.

von Humboldt A., Bonpland A., Williams H. M. 1821.Personal Narrative of Travels to the Equinoctial Regionsof the New Continent, During the Years 1799–1804,Volume 5. Paris (France): Longman, Hurst, Rees,Orme, and Brown.

Wagner M. L., Cortes C. W. 1921. Das Landliche LebenSardiniens im Spiegel der Sprache: Kulturhistorisch-Sprachliche Untersuchungen. Heidelberg (Germany):C. Winter.

Walker H. W. 1910. Wanderings Among South Sea Sav-ages and in Borneo and the Philippines. London(UK): Witherby.

Wander K., Shell-Duncan B., McDade T. W. 2009. Eval-uation of iron deficiency as a nutritional adaptationto infectious disease: an evolutionary medicine per-spective. American Journal of Human Biology 21:172–179.

Watson P. J., Hawkins C., McKinney J., Beatey S.,Bartles R. R., Rhea K. 1987. Inhibited drinkingand pica in rats following 2-deoxy-D-glucose. Phys-iology and Behavior 39:745–752.

Whiting A. N. 1947. Clay, starch and soot eatingamong southern rural Negroes in North Carolina.Journal of Negro Education 16:610–612.

Wiley A. S., Katz S. H. 1998. Geophagy in pregnancy: atest of a hypothesis. Current Anthropology 39:532–545.


Wilson M. J. 2003. Clay mineralogical and relatedcharacteristics of geophagic materials. Journal ofChemical Ecology 29:1525–1547.

Wong M. S., Bundy D. A., Golden M. H. 1991. Therate of ingestion of Ascaris lumbricoides and Trichu-ris trichiura eggs in soil and its relationship toinfection in two children’s homes in Jamaica.Transactions of the Royal Society of Tropical Medicineand Hygiene 85:89–91.

Yip R., Dallman P. R. 1988. The roles of inflammationand iron deficiency as causes of anemia. AmericanJournal of Clinical Nutrition 48:1295–1300.

Youdim M. B. H., Iancu T. C. 1977. Pica hypothesis.British Journal of Haematology 36:298.

Young S. L. 2010. Pica in pregnancy: new ideas aboutan old condition. Annual Review of Nutrition 30:403–422.

Young S. L. 2011. Craving Earth: Understanding Pica—The Urge to Eat Clay, Starch, Ice, and Chalk. NewYork: Columbia University Press.

Young S. L., Ali S. M. 2005. Linking traditional treat-

ments of maternal anaemia to iron supplement use:an ethnographic case study from Pemba Island, Zan-zibar. Maternal and Child Nutrition 1:51–58.

Young S. L., Goodman D., Farag T. H., Ali S. M.,Khatib M. R., Khalfan S. S., Tielsch J. M., StoltzfusR. J. 2007. Geophagia is not associated with Tri-churis or hookworm transmission in Zanzibar,Tanzania. Transactions of the Royal Society of TropicalMedicine and Hygiene 101:766–772.

Young S. L., Wilson M. J., Miller D., Hillier S. 2008.Toward a comprehensive approach to the collec-tion and analysis of pica substances, with emphasison geophagic materials. PLoS ONE 3:e3147. doi:10.1371/journal.pone.0003147.

Young S. L., Wilson M. J., Hillier S., Delbos E., AliS. M., Stoltzfus R. J. 2010. Differences and com-monalities in physical, chemical and mineralogi-cal properties of Zanzibari geophagic soils. Journalof Chemical Ecology 36:129–140.

Associate Editor: Locke Rowe


Why On Earth?: Evaluating Hypotheses About The ...WHY ON EARTH?: EVALUATING HYPOTHESES ABOUT THE...

Documents

Transcript of Why On Earth?: Evaluating Hypotheses About The ...WHY ON EARTH?: EVALUATING HYPOTHESES ABOUT THE...