778 I CURRENT ANTHROPOLOGY

Together these results point in one direction: Variablesmust be missing. Consequently, Gaulin and Boster'stheoretical reasoning needs to be supplemented. On themethodological level it is apparent that in research situ-ations such as this one conventional statistical tools ob-scure both the significant and the insignificant but Bool-ean analysis does not.

References CitedDECENNE, ALAIN, AND MARIE-ODILE LEBEAUX. 1992. Ana-

lyse booieenne des questionnaires: Programme BOOLEEN. MS.DICKEMANN, MILDRED. 1991. Women, class, and dowry. Amer-

ican Anthropologist 93:946-48.FLAMENT, CLAUDE. 1976. L'analyse hooleenne de question-

naire. Paris: Mouton.FREUNDLICH, YEHUDAH. 1976. Resurrecting the ravens. Syn-

these 33-34i-,S4'GAULIN, STEVEN ]. G., AND TAMES S. BOSTER. I99O. Dowry

and female competition- American Anthropologist 92:994-1005.

. 1991- Dowry and female competition: A reply to Dicke-mann. American Anthropologist 93:946-48.

GLYMOUR, CLABK. 1975. Relevant evidence, foumal of Philoso-phy 122:403-26.

HARTUNG, lOACHlM, BARBEL ELPELT, AND KA&L-HE1N2KLOSENER. 1985. Statistik. Miinchen: R. Oldenburg.

HEMPEL, CARL G. 1965 ii94S). "Studies in the logic of confir-mation," in Aspects of scientific explanation. New York: FreePress.

LAMBERT, KAREL, AND GORDON G. BRITTAN. I99 I . Eine Ein-fiihrung in die Wissenschaftsphilosophie. Berlin: de Gruyter.

MILLER, STEVEN I., AND MARCEL FREDERICKS. I99 I . A casefor "qualitative confirmation" for the social and behavioral sci-ences. Philosophy of Science ^8:452-67.

RAGIN, CHARLES. 1989. The logic of the comparative methodand the algehra of logic- journal of Quantitative Anthropology1:373-98-

WHITE, D. R., M. L. BURTON, AND L. A. BRUDNER. 1977- En-taiiment theory and method: A cross-cultural analysis of thesexual division of labor. Behavior Science Research 12:1-24.

Is the Fertility ofAgrieulturalists Higher ThanThat of Nonagriculturalists?'

GILLIAN R. BENTLETf, GRAZ"yNA 7AS1ENSKA, ANDTONY GOLDBERGDepartment of Anthropology. NorthwesternUniversity, Evanston. III. 60208. U.S.A. (Bentley)/Department of Anthropology, Harvard University,Peabody Museum, Cambridge, Mass. 02138, U.S.A.(Jasieriska and Goldberg). 28 v 93

Gampbell and Wood (1988) have presented fertility datafrom a large sample of subsistence societies that suggest

r. © 1993 by The Wenner-Gren Foundation for AnthropologicalResearch. All rights reserved OOII-3204/93/3405-ODO7$I.OO. Weare grateful to Jim Wood for providing us with unpublished data

there are no significant differences in mean total fertilityrates^ (TFRs) among foragers, horticulturalists, and agri-culturalists.^ Their main purpose was simply to charac-terize the levels of fertility typical of traditional, natu-ral-fertility populations'" and to attempt to explain thevariability observed within the framework of a proxi-mate-determinants model.^ Their paper has often beencited, however, for its bearing on anthropological anddemographic paradigms concerning fertility transitions(see, e.g., Blurton-)ones et al. 1992; Borgerboff Mulder1992; Hewlett 199I; Pennington and Harpending 1992:158, 161). Specifically, their results suggest that tech-nological developments such as agricultural intensifi-cation have a greater impact on mortality than on fertil-ity, thus fueling the continuing debate in anthropologyabout the causes of population change in prehistory (e.g.,Armelagos, Goodman, and Jacobs 1991, Handwerker1983, Harris and Ross 1987).

Our more recent study, stimulated by Campbell andWood's intriguing findings, has produced different re-sults (Bentley, Goldberg, and [asieriska 1993). Using asample of 57 populations (12 foraging, 14 horticultural,and 31 agriculturall, we found a mean TFR of 6.1 ± 0.2S.E., while the mean TFR for foragers was 5.6 ± 0.4,for horticulturalists 54 ± 0.2, and for agriculturalists

and for several critical comments and stimulating discussions.Bachtiar Alam, Mariam Aii, Nicole Crepeau, Julie Goldman, SarahHemphill, Lida lunghans, Sharon Lang, and Elizabeth Sarkodie-Mensah helped in the eariy stages of compiling the data. We thatikRobert Aunger, Monique Borgerhoff Mulder, Henry Harpending,Barry Hewlett, Chris Himes, Dennis Hogan, George Milner, Na-dine Peacock, Virginia Vitzthum, and Steven Weeks for readingmanuscript drafts, offering data, and providing various insights.This work was partially supported by a National Institute of AgingPostdoctoral Fellowship at Pennsylvania State University to thefirst author.2. The total fertility rate is the total number of live births towomen during their reproductive careers in a given popuiation.There are two metbods ot computing TFRs (see Newell 1988): pe-riod rates are computed cross-sectionally, usually from a group ofwomen aged 15-45 years, while cohort rates examine the fertilityhistories of women who have ceased reproduction |aged 4'; + ]. Inmicrodemographic studies of small populations typical tor anthro-pology, cohort rates are more generally used.3. Foragers are defined here as populations that either collect, hunt,or scavenge necessary resources from their immediate environ-ment; horticutturalists are defined as shifting |extensive| cultiva-tors using hoes, with cultivated produce providing most nutritionalneeds; intensive agriculturalists are distinguished by their repeatedcultivation of the same plots, often involving crop rotation and,generally, the use of the plow. Information for these subsistencecategories was obtained either from the original articles containingdemographic data or from supplementary papers that expanded on agroup's subsistence practices (see Bentley, Goldberg, and Jasieriska1993 for further details).4. Although the term "natural tertility" is problematic, we use ithere to refer to populations recorded as not using any form of mod-em contraceptives such as steroid or barrier methods. We recognizethat it is possible that any group in our sample may have beenusing culturally based tamily planning practices such as inten-tional prolongation of lactation or coitus interruptus, but thesemethods were rarely, if ever, mentioned in the literature.

.̂ Proximate determinants are the intermediate behavioral and bio-logical factors through which social, economic, or environmentalvariables affect fertility (Bongaarts and Potter 1983:1).

Volume 34, Number 5, December 1993 \ 779

6.6 ± 0.3.^ There were no significant differences in fer-tility rates between foragers and horticulturalistsInonagriculturalists}, but there were significant differ-ences between nonagriculturalists and intensive agricul-turalists (Mann-Whitney U, p = 0.004). In the courseof reanalyzing Campbell and Wood's data, we found anumber of minor errors in their data, as well as inclusionof groups that we label here as transitional, that is, un-dergoing acculturation and modernization. The purposeof this report is to present the results of our reevaluationof Campbell and Wood's sample (hereafter the CW sam-ple) in order to clarify how we came to substantiallydifferent conclusions. Since the analytical goals thatgovemed our research were ultimately different fromCampbell and Wood's, this critique of part of their papershould not detract from their other important contribu-tions. For example, they point out that there is a greatdeal of variability and overlap among the fertility ratesof foragers, horticulturalists, and agriculturalists. Ourreanalysis confirms this basic finding and the fact thatit is therefore impossible to predict fertility on the basisof subsistence alone (Bentley, Goldberg, and Jasiehska1993)-

Table i and Figure i compare our findings with thoseof Campbell and Wood. The mean TFR for the CW sam-ple matches that for our sample, but mean TFRs in theirsample are higher for nonagriculturalists and lower foragriculturalists.^ From figure i it is clear, however, thatthe two samples overlap considerably, the tnaior differ-ence being that ours has higher variance. As Campbelland Wood report, there are no significant differences intheir sample among the three subsistence groups(Kruskal-Wallis, p = 0.2). However, when they com-pared the TFRs of agriculturalists against the other twosubsistence groups the results were marginally signifi-cant using analysis of variance (p = 0.08) but not sig-nificant when a multiple analysis of variance controlledfor the overrepresentation of European groups in the ag-ricultural category [F = 0.18, df = 2, p > 0.8) (Campbelland Wood 1988:63 n. 4).

Campbell and Wood employed the following criteriafor the inclusion of populations in their sample (p. 41):A population had to have a low degree of acculturation,no recent or major drop in fertility rates, and a minimumnumber of 50 women (where cohort fertility rates wereused) or a minimum of 200 registered births per annum

TABLE I

Statistical Analyses of Total Fertility Rates forVarious Suhsistence Groups in the Campbell andWood and Bentley et al. Samples

ForagersHorticulturalistsAgriculturalistsAll groups

Campbell andWood Sampie

Mean

5-7S-96.3(S.I

S.E.

0 .40.20.2O.I

n

102 6

347 0

Bentley et al.Sample

Mean

5.6S.46.66.1

S.E.

0 .40.2

0 .3D.2

n

12

143157

NOTE: Mann-Whitney U test for foragers and horticulturalists vs.agriculturalists, p = 0.08 for CW sample, 0.004 tor Bentley et al.sample. Kruskal-Wallis test for foragers, horticulturatists, and ag-riculturalists, p = 0.2 for CW sample, 0.02 for Bentley et al. sam-ple.

(where period fertility rates were used).^ The study inquestion had to have demographic analysis as its pri-mary intent and (if contemporary) had to indicate atten-tion to possible sources of missing data and other errors.Data based on unreliable methods were excluded, aswere populations with high rates of primary infertility(0.15 or greater), which might be due to sexually trans-mitted diseases.

Reviewing the original sources, we found several in-stances in which the data do not fully conform to thesecriteria (table 2). For example, there are two populations(Pahira and Bhoksa) for which the sample of women isless than 50. Campbell and Wood frequently relied onsecondary sources for TFRs [specifically, Henry 1961:84,table I; Leridon 1977:107-9, table 7.I; and Spuhler1976:192-93, table 45)- In some instances, the TFR inthe secondary source differs from that given in the origi-nal article' or more than one is reported (e.g., AustralianAborigines in Kirk 1981). In one of the secondary sources(Henry 1961:84, table i), it is clear that the author modi-fied [smoothed) the data for some populations for hisparticular analysis [Wilson, Ocppen, and Pardoe 1988).Sometimes the TFRs cited in the secondary publications

6. This sample includes 32 populations derived from Campbell andWood's 70 groups. Seven of these 32, however, correspond to 23separate listings in Campbell and Wood. The data for the rest oiour sample either are not represented in Campbell and Wood'spaper or were unavailable at the time it was published. Readersare referred to our earlier paper for a more extensive discussionoi the implications oi these findings [Bentley, Gojdberg, and Jasieri-ska 1993). Campbell and Wood inadvertently omitted one of thepopulations included in their analyses from their article, namely,Bilheres-d'Ossau in France.7. We are very grateful to Jim Wood for providing us with theoriginal data from which we could calculate these and other re-sults.

8. In our study we relaxed the criterion tor sample sizes in orderto be able to include a larger number of foraging groups than wouldotherwise be available (Bentley, Goldberg, and lasieriska 1993).Only ID groups have a sample size below 50 women, and 6 of theseare above 30. When we compared this sample with one that onlyincluded groups with 50 or more women, we tound that there wereno significant differences in the statistical results.9. For exampie, a TFR of s.4 is given for the Black Carib in Spuhler(1976) and is repeated by Campbell and Wood. Spuhler actuallyobtained his tigure from a personal communication hy Firschein;an article published in 1984 by Firsehein gives a TFR of 5.8. Simi-larly, the Caingang are cited as having a TFR of 6.1 in Spuhler'stahle 45, but his primary source oniy gives a "mean number ofchildren," which is 2.2 (Salzano 1961I.

78O I CURRENT ANTHROPOLOGY

251

20-

a 15E

10-

om

min

mo

Total Fertility Rates

FIG. I. Percentage distributions of total fertility rates for CW sample [solid bars) and Bentley et al. sample.

are themselves summaries, making them tertiarysources of information. This reliance on secondary andtertiary sources has led to various problems that mayhave biased Campbell and Wood's results:

Unreliable data. For some populations included in theCW sample—Iranians, Hindus from Bengal, Fouta-Djalon, and Lapps—the data are unreliable [see Coaleand Trussell 1974; Wilson, Oeppen, and Pardoe 1988)."*The Sioux-Ojibwa, Dhurwa, and Australian Aboriginesshould be excluded on similar grounds (see table 2). TheLapps, Dusun, and Thule Eskimos are reported to havehigh rates of infertility (11.7%, 10%, and 16%, respec-tively] which raise the suspicion of exposure to sexuallytransmitted diseases."

Duplication. The Punjabi Chamars and the Khannaare in fact the same population, and Anhausen is in-cluded in Knodei's (1978) group of Bavarian villages.̂ ^

10. The Iranian data were originally drawn from Mashayekhi,Mead, and Hayes (1953] and cited in Henry (195 3]. Data for theHindus were cited in Lorimer [1954), come from a small sample otwomen, and were extracted trom an unpublished survey that can-not be verified. The information for Fouta-Djaion was derived troma government report that is not readily obtainable and is cited inHenry ii96i). The TFRs tor the Lapps weie taken irom Fraecaro(i9i;9:92), who originally derived his information trom Wahlund(1932)-11. We adopted a more conservative figure of o.i for infertilityrates hecause most natural-fertility populations rarely have rates ofprimary sterility above 0.05 unless there is a problem with sexuallytransmitted diseases (Bongaarts and Potter 1983:41-42).12. Data for the Chamars are given in Potter et al. I1965) and forthe Khanna in Wyon and Gordon [1971). The Kharma are, in fact.

Questionable subsistence categories. The Pahira areidentified by Campbell and Wood as "tribal" horticul-turalists but described as foragers in the original sources;the Nasik, considered foragers, are agriculturalists whoperform some wage labor; the Bhoksa, called "tribal,"are agriculturalists; and the Dusun, identified as "tribal"horticulturalists, are agriculturalists with some cashcrops. These misassignments seem due to Campbell andWood's grouping of horticulturalists and pastoralists un-der the kinship designation "tribal." It is also not en-tirely clear how their "peasant" agriculture is distin-guished from horticulture. In three cases—Dhurwa,Hindu villages, and Fouta-D)alon—it was im.possible toascertain the subsistence hase. In addition, while Camp-bell and Wood claim that their sample includes semi-nomadic and nomadic pastoralists, these subsistencegroups are in fact not represented.

Transitional populations. Eight hunting and gatheringpopulations [Asmat, Australian Aborigines, Lapps, Nu-namiut, Ramah Navajo, Sioux, Thule Eskimos, andTiwi) and one horticultural group [Caingang) in the CWsample were undergoing significant acculturation and/or demographic change at the time of investigation. Inaddition, two horticultural groups (Karkar and Makin)

a suhsample of the Chamars. In addition, the Khanna study wasspecifically designed to assess the impact of contraceptive useamong select groups in the Punjah. Data for Anhausen are derivedtrom Knodei [1970), and these data are repeated in his analyses ofBavarian villages I1978).

Volume 34. Number |, December 1993

TABLE 2Campbell and Wood's Sample with Total Fertility Rates and Notes on Problems with the Data

Population Data Source

Original OriginalCW TFR TFR CW^ Revised^TFR [Cohort! iPeriod! Subsistence Subsistence Problems

EuropeNonnan, CrulaiNorman, SottevilleFrench, BoulavFrench, QuercyFrench, Bilheres-d'OsiauFrench, TourouvreFrench, northwesternFrench, southwesternEnglish, 14 parishesGerman, AnhausenGerman. BavariaGerman, East FrieslandGerman, HesseSwiss, GenevaSWISS, Geneva

Lapps, SwedenNorwegians, NorwaySwedes, Estonia

European, North AmericaAmisi, OhioCanadian, QuehecHutterites, U.S.A.Hutterites, U.S.A.Mormons, UtahMormons, Utah

AfricaEuropeans, TunisiaFouta-Dialon, GuineaDobe IKung, BotswanaNgbaka, Central African RepublicSerer, SenegalYao, Malawi

Southwest AsiaIndians, BombayBhoksa, Uttar PradeshChamais, PunjabDhurwa, IndiaIndians, Hindu villagesIranians, IranKhanna, PunjabKota, KeralaBangladeshis, Matiah ThanaIndians, NasikIndians, South PahiraDusuii, Peoampang, Borneo

Semai Senoi, MalaysiaTaiwanese, Yunlin

PacificAborigines, AustraliaTiwi, Australia

Asimt,. Iiian |aya

Enga, New GuineaMoeioe, Irian lay a

Fak-Fak, liiao {aya

Ninboran, Irian lay a

Noemtoor, Irian iaya

Schouten, Irian Iaya

Waropen, hian [aya

Gainj, New GuineaKarkar, New Gumea

Lufa, New Guinea

1674-1741 Gautier and Henry (1958)1760-90 Girard liy^y)pre-i7So Houdailk 1:967)1700-1791 Valmary I1965I1740-79 Fresei-Losey (1969!1665-176S Charbonneau ii97ol1670-1769 Henrv and Houdaille (1975)1720-69 Hetiry 11971!i6oo-r799 Wilson I1984)r629-i7fJ9 Knodei [1970Ir75O-i8^c Knodel (1978Ir7SO-i85o Rnodel I197HI1750-1850 Knodei I197SIpre-1600 Henry 11956)1600-1649 Hetiry (1956I1791-1890 Fraccaro 11959!1871-1900 Henry I1970I1841-1900 Hyrenius (1958)

1900-1910 Crass and McKusick (1970!1700-1750 HenripJn (19S4Ipre-i9ii Eaton and Mayer (i9SilI92I--3O Eaton and Mayer |i9s}l[S20-45 SkoJnick et aJ. I1978)1846-80 Skolmckii97fl|

1840-59i9S4-iS1967-691948-571963-6;1946-47

Ganiage I1960IHenry |r96i)Howell 11979!Thomas !i963lCantrelie [1969Miccheli [1949!

6.9

7.2

6.3

6.06.55.95.57-57.6i.86.45.75-06.34.9

8.07-S9.8

7.0474.65-56.74-9

S-55.14.7

yO

6-3

6.J

7.68.2

4.26.7

6.8

6.4

7-55.3

6.1

6.1

5.65.0

SJS-3

6->5.0

—

6.1

6.3S-J

5.7

7.1

6.5—

1954-s; Dandekar [1959!1975 Garg, Tyagi, and Sankhyan (191900-1914 Potter et al. (1965!1959 Rakshit J1972J1941-46 Lorimer (1954)1950 Heniy 11953!1960-6]; Wyon and Gordon (1971)1963-68 Ghosh (1976,11969-71 Chen et al. (1974)l95i-!;3 Dandekarand Sovam (1955)1963 Basu [1967!

Glyn-Jones figs3)

1968-69 Fix (1977!I9>2 Tuan I1958)

1970 Kiik I1981)1952-61 lones I1963!

1972 Van Arsdale (1978)

1966 Sinnett and Whyte 11973)1959-64 Groenewegen and van de Kaa

I1964I1959-64 Groenewegen and van de Kaa

(1964I1959-64 Groenewegen and van de Kaa

I1964I1959-64 Groenewegen and van de Kaa

(1964)1959-64 Groenewegen and van de Kaa

11964)1959-64 Groeneweg«i and van dcKaa

t977-78 Wood et al. [1985!, Wood 11991) 4-31968-69 Stanhope and Homabrook !i974) 6.3

1968-69 Stanhope and Homabrook (1974) 4.6 4.6

6.9

5.6^6.8"7.1''y '6.3'6.0=

7.7°6.9"

S-5'7.5'7.6"i.8^h.i'S.I'1.1'

6.5"

7.48.0"

7.0"

4.7'

6.7

6.3

6.3

5-64.6

7.0

7.3

74

7.8

6.4

4-36.3

5-94.4

4.6

6.7

7.1

6.3

5.7

4-36.4

4-7

7.1

7.1

7.4

7.8

6.4

ppppppppppppppp

HGPP

PPPPPP

PT

HGTTT

PTPIPPPPP

HGTT

TP

HGHG

HG

TT

T

T

T

T

T

TT

Some iamilv limitaiion.'

9 Duplicated.'^

999

12

12

transitionaf Sterility rate fr.7%.9 Unreliable.'^9

y

n.a.n.a.

7

Unrehable.Exposed to STDs.'

ContiaceptJng and duplicated.''Unreliable.Unreliable.Unreliablt.Contraceptin;; and duplicated.'

Sterility tate ic%; demographynot primary aim.

transitional Unreliable.transitional Eiving on mission and accultu-

transitional Acculturated with recent popula-tion increase.

4

5 Acculturated with massive popula-tion increase.

782 I CURRENT ANTHROPOLOGY

TABLE 2

(Continued)

Population Data Source

Original OngmalCW TFR TFR CW^ Revised^TFR (Cohonl (Periodl Subsistence Subsistence Problems

Makin, Gilbert Isles

Maring, New GuineaOntone lava. Pacific isles

Native North America, GreenlandNunamiut, Alaska

Navaio, RamahSioux-Ojibwa, Great PlainsEskimos, Thule

South America, CaribbeanAymara, Chile

Black Carib, St. VincentCamgang, Brazil

Martmique, Piench AntillesTerena, BrazilYanomama, Venezuela

1947-71

1966-761920-71

193S-6H

rS44-i9441894

r961-66

1946-471958

1914-181955-601964-72

Lamben 1197si

Buchbinder and Wood (19B4)Bayliss-Smith (1975)

Binford and Chasko (1976I

Morgan I1968, 1973IBoas (1894)Malaurie, Tabah, and Sutter

Cruz-Coke et al. (1966]

Spiihler(i976|, Pirschein I1984)Salzano (1961I

tendon [1971ISalzano and de Ohveira (1970)Neel and Weiss Ii97sl, Early and

Peters (1990I

6.4

4-763

6.9

6.55-9

6.9

S-46,1

545-58.2

6.3

_

6.3

—

6,5

—

7.0

5.82.2

—

8,2

4-7

6.9

-

—

—

5.4

—

T

TT

HG

HGHGHC

P

THG

PTT

4

38

transitional

transitionalaansitional

t

ro

transitional

11

31

Recently subject to depopulationand disease.

Recently subject to depopulationand disease.

Accuiturated with population in-crease.

Accuiturated.Unreliable,Sterility rate i6,c%.

TFR only for women aged 3 5 -1-and unreliable.

Accuiturated on reservations andfarms.

'Calculated trom TMFRs.''P: peasant agnculturahsts^ T, tribal honiculturalists; HG, hunter-gatberers.' 1 , subsistence toragers; 2, forager-honiculturalists;.!, subsistence horticulturahsts; 4, horticultural 1sts with cash crops, i, honicukuralists with wage labor, 6. honiculturalistswith berds, 7, subsistence agriculturalists, S. agriculturahsts with cash cropS; 9, agnculturalists with wage labor, JO, agropastoralists, 11, wage laborers in agricultural communi-ties^ IZ, merchants and wage laborers in agricultural communities; titmsitional, undergoing acculturation.•̂ See Wilson, Oeppen, and Pardoe I19H8I'See Knodel (1970).'We could find no matching figure in the odginal source."See Pennington and Harpending |t99i].''See Wyon and Gordon I1971I.'See Potter et al. I1965I.

and one agricultural group (Ontong Java) had sufferedstriking demographic upheaval, mostly caused hy intro-duced changes in health conditions.^^

Galton's problem. Pseudoreplication arising from thecultural (and statistical) interdependence of groups witha common geography or ethnographic history can seri-ously hias results. Camphell and Wood's sample of 70populations includes 19 historical demographic cases[27%) that are either European or European-derived. Insome of these cases, particular regions and periods oftime are represented more than once (e.g., the Svi'iss andGerman villages); in others, the infomiation from Euro-pean villages is aggregated [e.g., the British parishes).The historical particularity of European marital and fer-tility patterns has been pointed out in several studies(Hajnal 1965, Spagnoli 1977), and the inclusion of allthese groups seems to represent a striking example ofGalton's problem. In addition, six different villages inIrian laya are represented separately (Groenewegen andvan de Kaa 1964). Similarly, the duplication of particularpopulations by including TFRs for more than one time

13. For a discussion of some of these traditional groups includedby Campbell and Wood see Harris |i977l-

period (e.g., Hutterites and Mormons) is problematic.Campbell and Wood included these duplicate cases be-cause they represent historical fluctuations in fertilitythat have been noted hy demographers as important fortheir particular analyses.

Campbell and Wood did, as we have seen, control forgeographical and historical overrepresentation in one oftheir statistical analyses. They noted that when agricul-turalists are compared with all other groups, the differ-ence in TFRs approaches statistical significance, butthey surmised that this resulted from the inclusion inthe agricultural sample of a large number of historicalEuropean populations generally characterized hy higherTFRs. A multiple analysis of variance to control for theeffects of subsistence and region removed the apparentnear-significance, validating the notion that the resultswere biased towards these specific historical groups.We repeated the statistical comparisons after removingthe historical European and European-derived groups(French, British, Norman, German, Swiss, Swede, Cana-dian, and Tunisian). Contrary to Camphell and Wood'sprediction, removal of these groups increases the meanfor agriculturalists (6.8 ± 0.3 S.E.) and the significancelevel for the comparison of nonagriculturalists and agri-culturalists (Mann-Whitney U, p = 0.003).

Volume 34, Number 5, December 1993 \ 783

TABLE 3Analysis of the Campbell and Wood Sample Modified to Remove Individual Sources of Error

Unreliable Questionable Transitiona]Data Subsistence Populations

Gallon'sProblem

Duplicate No HistoricaJ All ChangesPopulations Populations Together

TFR S.E. n TFR S.E. n TFR S.E. n TFR S.E. n TFR S.E. n TFK S.E. n TFR S.E. n

ForagersFlorticulturalistsAgriculturalists

All groups

6.16.06.46.2

0.30.20.2

Q.2

723

61

5.76.06.36,1

0.40.30.20.1

9

20

39

54

6.1

0.80,20.20.2

2

21

34

5-6

5.7

6.2

5.9

0.30.2

0.30.2

II2121

5 96.36.1

0.40.2

0.2O.I

1026

68

5.75-96.46,0

0.40.20.40.2

1026

51

5-45.76.25.9

0.80.3

21416

NOTE: Mann-Whitney U test for foragers and horticulturaiists vs. agriculturalists, left to right, p = 0.2, 0.2, 0.06, 0.1, 0.1, 0.3, 0.2.Kruskal-Wallis test for foragers, horticulturaiists, and agriculturalists, p = 0.4, 0.3, 0.2, 0.3, 0,3, 0.6, 0.4.

TABLE 4Total Fertility Rates of Transitional Populations

Population

Navajo, RamahLapps, SwedenSioux-Ojibwa, Great PlainsNunamiut, AlaskaAsmat, Irian JayaKarkar, Papua New GuineaTiwi, Bathurst IslandAborigines, AustraliaEskimos, Flail BeachEskimos, WainwrightEskimos, KuskokwimAborigines, AustraliaShipibo-Conibo, Ucayali RiverAthapascan, Old CrowAborigines, Northern Territory

Source

Morgan |j973lFraccaro (1959!Boas (1894!Binford and Chasko (1976)Van Arsdale (1978)Stanhope and Fiornabrook I1974!Jones I1963]Kirk (1981)McAlpine and Simpson (i97()lMilan (1970)Hrdlicka I1936)[ones (1972)Hem I1977, 1990)Roth (1981)(ones [1963)

Date

1844-941791-189018941935-6819731968-691952-611959-801969-711910-681930s1967-681968-69post-19001958-60

TFR

6,5S.O

5.96.96.96.4S.O

6-SII.I

9.95-86.6

IO.S6.64.2

The calculation of total fertility rates from total mari-tal fertility rates. In several of the primary sources, thedata given are total marital fertility rates [TMFRs) ratherthan TFRs. In order to calculate the latter, Campbelland Wood used an average ratio of 0.762 (TFR/TMFR)derived from the estimates of TFR and TMFR by Leridon(1977) for a sample of 25 natural-fertility populations.Leridon was able to calculate a TFR from the originalsources for just 10 cases. Of these 10, only 3 were de-rived from non-European populations {Senegal, Marti-nique, and Bombay). Five of the remaining 7 were de-rived from historical European examples subject to lateages at marriage, low rates of remarriage, and high celi-bacy rates (Hajnal 1965I. Thus, while this procedureseems appropriate for historical populations, it is uncer-tain whether one can legitimately apply it to non-European *̂

14. We retained the cohort TFR for Yunlin Region without anymodification hecause the data indicate almost universal marriage

To reanalyze Campbell and Wood's data, we derivednew TFRs for modified samples by removing step-by-step the data we have identified as problematic (table 3).Most of the corrections in fact make Uttle difference.The most significant effect appears to be the removal ofthe transitional societies, which effectively hiased theTFRs upward for nonagricultural groups. When we ex-amined the mean TFR for 15 transitional nonagricul-tural societies {table 4), including some from the CWsample,, we found an average TFR of 6.9 ± 0.5, whichis significantly higher than the TFR for nonagriculturalgroups in our sample (Mann-Whitney U, p ^ o.oi) andhigher but not significantly so than the mean TFR foragriculturalists {p ~ 0.8). It is therefore likely that theinclusion of these transitional societies is the primary

hy age 20-24. Data from rural northem China confirm a pictureof universal marriage for the same time period, with a celibacy rateof less than o.ooi for women ovei 25 (Barclay et al. 1976).

784 I CURRENT ANTHROPOLOGY

reason Campbell and Wood did not find significant dif-ferences between subsistence categories. The finding ofhigher fertility among these mostly transitional foraginggroups underscores the impact that acculturation mayhave on traditional mechanisms of fertility control,whether these are deliberate (such as postpartum absti-nence) or involuntary (such as the long periods of lacta-tion often essential for infant nutrition in environmentswith poor weaning foods).

We conclude from these analyses that, althoughCampbell and Wood's sample of traditional societies andtheir fertility rates remains impressive in terms of size,it should not be used as the sole basis for anthropologicalor demographic inferences that rely on comparisons ofdifferent subsistence groups. This is not to suggest thatour own analysis is perfect. Our readers will doubtlessfind grounds on which to criticize our methodology andchoice of data. We urge them to compare Campbell andWood's data with ours and welcome further discussionabout the comparative fertility of traditional suhsistencecommunities.

References CitedARMELAGOS, G. ^., A. H. GOODMAN, AND K. H. 1ACOBS.

1991, The origins of agriculture: Population growth during a pe-riod of declining health. Population and Environment 13:9-22.

BARCLAY, G. W., A. ]. COALE, M. A. STOTO, AND T. I. TRUS-SELL. 1976. A reassessment of the demography of traditionalrural China. Population Index 42:606-35.

BASU, A. 1967. Selection intensity in the Pahiras. Eugenics Quar-ieily 14:241-42.

BAYLISS-SMITH, T. P. 1975- "Ontong-[ava: Depopulation and re-population," m Pacific atoll populations. Edited hy V, Carroll,pp. 417-84. Honolulu: University Press of Flawaii.

BENTLEY, G. R., T. GOLDBERG, AND G. lASIENSKA. I993.The fenility of agricultural and non-agricultural traditional so-cieties. Population Studies. In press.

BINFORD, L. R., AND w. 1. CHASKO, TR. i97('. " N u n a m i u t de-mographic histor}': A provocative case," m Demngiaphic an-thropology: Quantitative approaches. Edited hy E. B. W. Zu-hrow, pp. 63-143. Alhuquerque: University of New MexicoPress.

BLURTON-TONES, N. G., L. C. SMITH, I. F. O'CONNELL, K.HAWKES, AND c. L. KAMU2ORA. 1992. Demography of thcHadza, an increasing and high-density population of savannaforagers. American journal of Physical Anthropoiogy Sf):js9-8i.

BOAS, F. 1894. The half-blood Indian: An anthropometric study.Popular Science Monthly 4^:761-70.

BONGAARTS, ] . , AND R. G. POTTER. 1983, Fertility, biology.and behavioi: An analysis of the proximate determinanta.New York: Academic Press.

BORGERHOFF MULDER, M. 1993, Demography of pastoialists:New data on the Datoga of Tanzania. Human Ecology 20. Inpress.

BUCHBINDER, G., AND T. w. WOOD. 1984. Population struc-ture of the Simhai Valley Manng, New Guinea, i. Liie tablesand stahie population analysis. MS.

CAMPBELL, K. L., AND T. w. WOOD. 1988. "Fertility in tradi-tional societies," in Natural human fertility: Social and biolog-ical mechanisms. Edited hy P. Diggory, S. T'eper, and M. Potts,pp. 39-69. London: Macmilian.

CANTRELLE, P. 1969. Etude demographique dans la region duSme-Saloum (Senegal): Etal civil et observation demogra-phique 196J-1965. Paris: Office de la Recherche Scientifiqueet Technique Outre-mer.

CHARBONNEAU, H. 1970. Touiouvie-au-Perche aux XVIIeme etXVIlIewe siecles: Etude de deinographie historique. InstitutNational d'EtuiJes Demographiques Cahier 55.

CHEN, L. C , S. AHMED, M. GESCHE, AND W. H. MOSLEY.1974- Prospective study of birth interval dynamics in rural Ban-gladesh. Population Studies 28:277-96.

COALE, A. ^., AND T. ) . TRUSSELL. 1974. Model fertility sched-ules: Variations m the age structure of childhearing in humanpopulations. Population Index 40:185-218.

CROSS, H. E,, AND V. A. MCKusicR. 1970, Amish demogra-phy. Social Biology 17:83-101.

CRUZ-COKE, R., A. P. CRJSTOFFANINI, M. ASPILLAGA, ANDF. BiANGANi. 1966. Evolutionary forces in human populationsin an environmental gradient in Arica, Chile. Human Biology38:421-38.

DANDEKAR, K. 1959. Demographic survey of six rural communi-ties. Gokhale Institute of Politics and Economic Studies 37,

DANDEKAR, K., AND N. V. sovANi. 1955. Fertility Survey ofNasik. Koiaba and Satara (North) Districts. Poona: SangamPress,

EARLY, ). D., AND ) . F. PETERS. 1990. The population dynam-ics of the Mucaiai Yanomama. New York: Academic Press.

EATON, I. w., AND A. ). MAYER. 1953. The social hiology ofvery high fertility among the Hutterites: The demography of aunique population. Human Biology 2^:206-64.

FIRSCHEIN, 1. L. 1984. "Demogiaphic pattems of the Garifuna(Black Cahhs) of Belize," in Black Caribs: A case study in bio-culturai adaptation. Edited hy M. H. Crawford, pp. 67-94,New York: Plenum Press.

FIX, A. G. 1977. The demography of the Semai Senoi, AnnArbor: Museum of Anthropology.

FRACCARO, M. 1959. Fertility differential in two Lappish popula-tions. American journal of Human Genetics 11:92.

FRESEL-LOSEY, M. 1969- Histoire demographique d'un villageen Beam: Bilheres-d'Ossau au XVIIIeme-XIXeme siecles. Bor-deaux: Biscaye Freres.

GANiAGE, i. i960. La population europeenne de Tunis au mi-lieu du XlXtme siecle: Etude demographique- Universite deTunis Publications de la Faculte des Lettres, 4eme serie (His-toire], 2.

GARG, s. K,, D. TYACi, AND A. R. SANKHYAN, 1981. Repro-ductive life of Bhoksa. Annals of Human Biology 8:361-70.

GAUTIER, E., AND L. HENRY. 1958. La population de Crulaiparoisse normande: Etude historique. Institut Nationald'Etudes Demographiques Travaux et Documents 33.

GHOSH, A. K. 1976- The Kota oi the Njlgiri Hills: A demo-graphic study, fournal of Biosocial Science 8:17-26.

GIRARD, p. 1959- Aper^us de la demographie de Sotteville-les-Rouen vers la fin du XVIIIe siecle. Population 14:485-508.

GLYN-jONEs, M. i9S3- The Dusun of the Penampang Plains inNorth Borneo. London: Colonial Social Science Research Coun-cil.

GROENEWEGEN, K., AND D. J. VAN DE KAA. 1964. ReSultS ofthe demographic research project Western New Guinea. 6vols. The Hague: Government Printing Office.

HAiNAL, 7. 196^. "European marriage pattems in perspective,"in Population in history: Essays in historical demography. Ed-ited by D. V. Glass and D. E. C. Eversley, pp. 101-43. Chicago:Aldine.

HANDWERKER, P. 19^3- The first demographic transition: Ananalysis of subsistence choices and reproductive consequences.American Anthropologist 85:5-27.

HARRIS, D. R. 1977- "Settling down: An evolutionary model forthe transformation of mobile hands into sedentary communi-ties," in The evolution of social systems. Edited by J. Friedmanand M. J. Rowlands, pp. 401-17. Pittsburgh: University of Pitts-burgh Press.

HARRIS, M., AND E. B. ROSS. i^Sj- Death, sex. and fertility:Population regulation in preindustrial and developing societ-ies. New York: Columbia University Press.

HENRiPiN, (. 1954. La fecondite des menages canadiens au de-but du XVIIIe siecle. Population 9:61-84.

Volume 34, Number 5, December 1993 \ 785

HENRY, L, 1953. Aspects demogtaphiqucs d'une region rurale deI'lran. Population 8:590-92.

. 1956. Anciennes families genevoises: Etude demogra-phique. XVIe-XXe siecie. Paris: Presses Universitaires deFrance.

-. 1961. Some data on natural fertiiity. Eugenics Quarterly8:81-91.

-. 1970. La population de la Norvege depuis deux siecles.Population 25:^43-57.

-. 1972, Fecondite des manages dans le quart sud-ouest dela France de 1720 a 1829. Annaies Economies. Societes. Civi-lisations 27:612-1023.

HENRY, L., AND ]. HOUDAiLLE. 1973- Fccondite des mariagesdans le quart nord-ouest de la France de 1670-1829. Popula-tion 28:873-924.

HERN, w. M. 1977. High fertility in a Peruvian Amazon Indianvillage. Human Ecology 5:355-68.

. 1990. Individual fertility rate: A new individual fertilitymeasure for small populations. Social Biology 37:102-9.

HEWLETT, B. s. 1991. Demography and childcare in preindus-trial societies, journal of Anthropological Research 47:1-39.

HOUDAILLE, J. 1967. La population de Boulay (Mosellej. Popuia-tion 22:1055-84.

HOWELL, N. 1979- Demography oi the Dobe IKung. New York:Academic Press.

HRDLICKA, A. 1936- Fecundity of Eskimo women. Americanjournal of Physical Anthropology 22:91-95.

HVK£Nius, H. 195S. FerdJity and reproduction in a Swedish pop-ulation group without family limitation. Population Studies12:121-30.

TONES, F, 1. 196^, A demographic survey of the Aboriginal popu-lation of the Noithern Teriitory. with special reference to Bath-urst island Mission. Australian Institute of Aboriginal Studies,Occasional Papers in Aboriginal Studies i.

. 1972. Fertility patterns among Aboriginal Australians. Hu-man Biology in Oceania 1:245-54.

KIRK, R. L. 1981. Aboriginal man adapting: The human biologyof Australian Aborigines. Oxford: Clarendon Press.

KNODEL, T. 1970. Two and a half centuries of demographic his-tory in a Bavarian village. Population Studies 24:353-76.

. 1978. Natural fertility in prc-industrial Germany. Popula-tion Studies 32:481-510.

LAMBERT, B. 1975. "Makin and the outside world," in Pacificatoll populations. Edited by V, Carroli, pp. 212-85. Associa-tion for Social Anthropology in Oceania Series 3.

LERIDON, H. 1971. Les facteurs de !a fecondite en Martinique.Population 26:277-^00.

. 1977. Human fertility: The basic components. Chicago:University of Chicago Press.

LORiMER, FRANK. 1^54. Culture and human feitiJitv. PsTis:UNESCO.

MGALPiNE, P. ] , , AND N. E. SIMPSON, 1976. Fertility andother demograjiihic aspects of thc Canadian Eskimo communi-ties of Igloolik and Hall Beach, Human Biology 48:113-38.

MALAURIE, ]., L. TABAH, AND |. SUTTER. I952. L'isolat Esqui-mau de Thule (Groenland). Population 7:675-92.

MASHAYEKHl, M. B., P. A. MEAD, AND G. S. HAYES. I953.Some demographic aspects of a rural area in Iran. Mi]bani< Me-morial Eund Quarterly 31:149-65.

MILAN, F. A. 1970. The demography of an Alaskan Eskimo vil-lage. Arctic Anthropology 7:26-43.

MiTGHELL, I. G. 1949. An estimate of fertility in some Yaohamlets in Liwonde district of southem Nyasaland. Africa29:293-308.

MORGAN, K. 1968. The genetic demography of a small Navaiocommtmity. Ann Arbor: University Microfilms International.

. 1973. "Historical demography of a Navajo community,"in Methods and theories of anthropological genetics. Edited by

M. H. Crawford and P. L. Workman, pp. 263-314. Albuquer-que: University of New Mexico Press.

;. V., AND K. M. wEiSS. 1975- The genetic Structure of atribal population, the Yanomama Indians XII: Biodemographicstudies. American journal of Physical Anthropology 42:25-52.

N£wELL, c. J988. Methods and models in demography. NewYork: Guilford Press.

P O T T E R , R. G. , M. L. N E W , ]. B. W Y O N , A N D I. E. G O R D O N .1965. A fertility differentia! in eleven Punjab villages, MilbankMemorial Fund Quarterly 43:185-201.

RAKSHiT, H. K, 1972. The Dorla and the Dhurwa of Bastar: Ademographic profile, journal of the Indian Anthropological So-ciety 7'-ii s-28,

ROTH, E. A. 1981. Sedentism and changing fertility pattems in anorthem Athapascan isolate, fournal of Human Evolution10:413-25.

SALZANO, F. M. 1961. Studies on the Caingang Indians, i. De-mography. Human Biology 33:110-30.

SALZANO, F, M., AND R. C. DE OLIVEIRA. r97O. Getietic aS-pects of the demography of Brazilian Terena Indians. Social Bi-ology 17:217-23.

sf^fN£TT, p. F., AND H. M. WHITE. 1973. Epidemiologics] stud-ies in a highland population of New Guinea: Environment, cul-ture, and health status. Human Ecology 1:245-77.

SKOLNICK, M., I , BEAN, D. MAY, V. AKBON, K. DE NEVERS,AND P. CARTWRiGHT. 1978- Mormon demographic histor>'.I. Nuptiality and fertility of once-married couples. PopulationStudies 32:5-19.

SPAGNOLI, p. G. 1977- Population history from parish mono-graphs: The problem of demographic variation, journal of Inter-disciplinary History 8:427-52.

SPUHLER, |. N. 1976. "The maximum opportunity for natural se-lection in some human populations," in Demographic anthro-pology: Quantitative approaches. Edited by E. B. W, Zubrow.Albuquerque: University of New Mexico Press.

STANHOPE, 1. M., AND R. W, HORNABROOK. 1974- Fertilitypattems of two New Guinea populations: Karkar and Lufa.fournal of Biosocial Science 6:4:59-52.

THOMAS, ]. M. c. 1963. Les Ngbaka de la Lobaye: Le depeuple-ment rural chez une population forestiere de la republique cen-trafricaine. Paris: Mouton.

TUAN, c, H. I9S8. Reproductive histories of Chinese women inrural Taiwan. Population Studies 12:40-50.

v'ALMAKY, p, 196;. Families paysannes au XVIIie siecle en Bas-Quercy: Etude demographique. Institut National d'Etudes De-mographiques Travaux et Documents, Cahier 45.

VAN ARSDALE, P. w, iy78. Population dynamics among Asmathunter-gatherers of New Guinea: Data, methods, comparisons.Human Ecology 6:435-67.

WAHLUND, s. G. w. 1932. Demographic studies in the nomadicand the settled populations of Northern Lapland. Uppsala:Almqvist and Wiksells.

WILSON, c. 1984. Natural fertility in pre-industrial Britain. Pop-ulation Studies 38:225-40.

WILSON, c , I. OEPPEN, AND M. PARDOE, 1988, What is natu-ral fertility? The modelling of a concept. Population Index54:4-20.

WOOD, T. w. 1993. "Fertility and reproductive biology," in Hu-man biology m Papua New Guinea. Edited by R. Attenboi-ough and M. Alpers. Oxford: Oxford University Press.

WOOD, 1. W., p. L. TOHNSON. AND K. L, CAMPBELL. I985.Demographic and endocrinological aspects of low natural fertil-ity in liighland New Guinea, /ournal of Biosocial Science17:57-79-

WYON, ^. B,, AND I. E. GORDON. 1971. The Khanna study:Population problems in the rural Punjab. Cambridge: HarvardUniversity Press,