AMERICANJOURNALOFPHYSICALANTHROPOLOGY143:285–297(2010) Sex Differentials in Frailty in Medieval England Sharon N. DeWitte* Department of Anthropology,University atAlbany,Albany,NY 12222 KEY WORDS paleodemography; paleoepidemiology; selective mortality;Black Death ABSTRACT In most modern populations, there are Asampleof299adults(173males,126females)fromthe sex differentials in morbidity and mortality that favor EastSmithfield cemetery wasanalyzed.The results indi- women.Thisstudyaddresseswhethersuchfemaleadvan- cate that the excess mortality associated with several tages existed to any appreciable degree in medieval osteological stress markers was higher for men than for Europe. The analyses presented here examine whether women. This suggests that in this medieval population, men and women with osteological stress markers faced previous physiological stress increased the risk of death the same risks of death in medieval London. The sample for men during the Black Death to a greater extent than used forthis studycomesfromthe EastSmithfieldBlack wastrueforwomen.Alternatively,theresultsmightindi- DeathcemeteryinLondon.Thebenefitofusingthisceme- cate that the Black Death discriminated less strongly tery is that most, if not all, individuals interred in East betweenwomenwithandwithoutpre-existinghealthcon- Smithfield died from the same cause within a very short ditionsthanwastrueformen.Theseresultsareexamined periodoftime.Thisallowsfortheanalysisofthedifferen- in light of previous analyses of East Smithfield and what ces between men and women in the risks of mortality is known about diet and sexually mediated access to associated with osteological stress markers without the resources in medieval England. Am J Phys Anthropol potential confounding effects of different causes of death. 143:285–297,2010. VVC2010Wiley-Liss,Inc. Inthemajorityofmodernpopulations,womentendtoex- leishmaniasis, listeriosis, and toxoplasmosis (Alexander, perience lower age-specific mortality rates at most (if not 1988; Roberts etal., 1995,2001;Pasche et al.,2005). all)agesandlivelongerthanmen(Heligman,1983;Coale, There are also important differences between men and 1991; Hill and Upchurch, 1995). For example, in 2008 the women in terms of the morbidity and mortality associ- lifeexpectancyatbirthforU.S.femaleswasnearly6years ated with degenerative diseases. In modern populations, longer than that of males, and in many countries, the dis- men suffer more severe symptoms or are at higher risks parity in life expectancies was even greater. In that same of mortality from suchchronic diseases as cardiovascular year, male life expectancies at birth exceeded those diseases, respiratory tuberculosis, malignant neoplasms, of females only in a very small minority of countries for renal disease, and cirrhosis of the liver (e.g., Lopez, which data were available (https://www.cia.gov/library/ 1984; Liu et al., 2003; Cerfolio et al., 2006; Kaminsky publications/the-world-factbook/fields/2102.html). Mortality et al., 2006; Ng, 2007; Pilote et al., 2007; Kalra et al., differentialsfavoringfemalesoccurbecausetheyappearto 2008; Silbiger and Neugarten, 2008; Farinati et al., be more resistant to many diseases and generally more 2009; McGovern et al., 2009). However, some diseases highlybufferedagainstenvironmentalstressorsthanmales disproportionately affect women. For example, there is a are(Stinson,1985). higher prevalence of diseases such as chronic obstructive Females are often atlower risks thanmales ofmorbid- pulmonary disease and autoimmune disorders among ity and mortality from many causes in modern popula- women (Fairweather et al., 2008; Cote and Chapman, tions, i.e., females often face lower probabilities than 2009), and women are more likely to suffer more severe males of contracting or developing many different types symptoms and higher mortality from stroke and diabetes of diseases and lower probabilities of dying from various causes. Many studies have found that males are more susceptible than females to a wide range of diseases caused by viruses, bacteria, parasites, and fungi (e.g., Grant sponsor: National Science Foundation; Grant number: Hoff et al., 1979; Oliviera et al., 1981; Kirkwood et al., BCS-0406252; Grant sponsor: the Wenner-Gren Foundation for 1983; Brabin and Brabin, 1992; Acuna-Soto et al., 2000; AnthropologicalResearch;Grantnumber:7142;Grantsponsors:the Klein, 2000; Noymer and Garenne, 2000; Wells, 2000; AmericanAssociationofUniversityWomen,theAmerican-Scandina- Blessmann et al., 2002; Owens, 2002; Leone et al., 2004; vianFoundation,UniversityatAlbanyCenterforSocialandDemo- Jansen et al., 2007; Noymer, 2008; Taylor et al., 2009). graphic Analysis (CSDA), the Universityat Albany Research Foun- dation. Males often suffer more severe symptoms or are at ele- vated risks of mortality from a variety of parasitic and *Correspondence to: Sharon DeWitte, Department of Anthropol- infectious diseases, such as staph infection, trypanoso- ogy,AS237,UniversityatAlbany,Albany,NY12222. miasis, leptospirosis, and respiratory infections (Hoff E-mail:[email protected] et al., 1979; Laupland et al., 2003; Jansen et al., 2007; Falagas et al., 2008). However, males are not always at Received23September2009;accepted8March2010 a disadvantage with respect to infectious and parasitic diseases. Many studies in human populations and exper- DOI10.1002/ajpa.21316 imental studies with animal models have found diseases Publishedonline27May2010inWileyOnlineLibrary that disproportionately affect females, such as malaria, (wileyonlinelibrary.com). VVC2010WILEY-LISS,INC. 286 S.N.DEWITTE (Moriyama, 1984; Reeves et al., 2008; Turtzo and ning in Scandinavia and followed decades later by other McCullough, 2008;Appelros etal., 2009). European countries(Gage, 2005).Thereis someevidence Even though males might fare relatively well with from historical documents that a sex differential in favor respect to some causes of morbidity and mortality, over- of women existed in some populations by at least several all it appears that women in modern populations are hundredyearsago.In17th-centuryLondon,forexample, generally less frail than men, particularly given the mortality rates for women were apparently lower than near-universal pattern of longer life expectancy or lower thoseofmen,despitereportsfromphysiciansthatwomen age-specific mortality rates at most ages for women sufferedmorethanmenfromdiseasesanduniquelyfrom (Heligman, 1983). Frailty refers to an individual’s rela- complications associated with pregnancy and childbirth tive risk of death compared to other members of the pop- (Graunt,1975).AccordingtoCoale(1991),mortalityrates ulation (Vaupel et al., 1979), or as Vaupel (1988: p. 277) havebeenlowerforwomeninEuropeanpopulationssince describes it, ‘‘a set of susceptibilities and risk factors at least the mid-19th century. At that time, women lived that alters their chances of death at different ages.’’ Dif- on average 2–3 years longer than men, and the disparity ferences in frailty among individuals within a population has increased since then such that women now live as (i.e., heterogeneous frailty) exist because of differences manyas7–8yearslonger,onaverage,thanmeninseveral in susceptibility to disease and death that may have Europeanpopulations(Preston,1977). genetic, biological, environmental, socioeconomic, or Investigators have attempted to address questions other causes (Wood et al., 1992; Aalen, 1994). The sex aboutmortalitydifferentialsinthepastbyexamininglife differentials in morbidity and mortality observed in mod- expectancy or mortality rates using skeletal data, with ern populationshave variouscauses.Some ofthesex dif- varying results. It is difficult to estimate sex-specific life ferences in favor of females might be genetically deter- expectancyfrombirthusingskeletaldatagiventheprob- mined. For example, diseases that are caused by reces- lems associated with determining sex for subadult skele- sive X-linked genes, such as X-linked immunodeficiency tal material (Gage, 2000). Some researchers thus only syndromes, disproportionately affect males (Waldron, estimate life expectancies at age 15 (or some other late 1984). Some of the observed sex differences in infectious teen/early adult age) or they estimate average ages at and parasitic disease patterns are attributed to sex hor- deathamongadults.Resultsfromthesestudieshavesug- mones, which play an important role in the immune sys- gestedthatseveraldifferentscenariosexistedinpastpop- tem, as estrogens generally enhance immunocompetence, ulations: longer life expectancies or higher average ages whereas androgens reduce it (Grossman, 1985; Ansar at death for females (e.g., among the prehistoric Pueblo Ahmed et al., 1999; Klein, 2000; Roberts et al., 2001; andin14th-centuryJapan:Bennett,1973;Nagaokaetal., Janele et al., 2006). Sex hormones also affect behaviors, 2006), longer life expectancies or higher average ages at such as aggression, which influence an individual’s risk deathformales(e.g.,inseveralearlyagriculturalpopula- of exposure to infectious disease (Klein, 2000). Sex hor- tions and late medieval Croatia: Heligman, 1983; Sˇlaus, mones may also influence the risks of some degenerative 2000),andcrossoverssuchthatlifeexpectancywashigher diseases. For example, sex hormones may affect athero- in one sex at early ages but lower at older ages (e.g., genesis (plaque formation within arteries) and thereby among 18th-century Arikara and in medieval Sweden: influence risks of cardiovascular disease (Choi and OwsleyandBass,1979;Ho¨gbergetal.,1987). McLaughlin, 2007). The sex differentials associated with In addition to estimating life expectancies from skeletal some degenerative diseases are also the result of behav- samples,biologicalanthropologistscanexamineosteological ioral differences, such as higher rates of cigarette smok- stressmarkersthatforminresponsetoepisodesofdisease, ing and alcohol consumption among men that are linked malnutrition, or other physiological stressors to determine to excess male mortality from such causes as coronary how sex affected risks of morbidity and mortality in past heart disease, certain cancers, and cirrhosis of the liver populations.Severalstudiesthathavelookedatdifferences (Hetzel, 1984;Lopez, 1984). between men and women in the frequency of particular The pervasiveness of morbidity and mortality differen- osteological stress markers have reported higher frequen- tialswithrespecttosexinmodernpopulationsraisesthe cies among men in past populations. For example, Ortner questionofwhethersuchdifferentialsexistedinthepast. (1998), Larsen (1998), and Cassidy (1984) found higher Researchers have examined mortality differentials in the frequencies of periostitis in men in prehistoric Native pastusingdocumentarydata,butsuchstudiesaregener- Americansamples.AnalysisofanearlyArchaicpopulation ally limited to the past several hundred years. According fromFloridafoundhigherfrequenciesoflinearenamelhy- to Bullough and Campbell (1980), men were believed to poplasia in males than in females (Berbesque and Doran, live longer than women in the ancient and Medieval 2008). According to Guatelli-Steinberg and Lukacs (1999), world,butbeginninginthe14thcentury,somedocuments when significant differences in enamel hypoplasia exist in indicate that women lived longer than men. Unfortu- skeletal samples, there is generally a higher frequency in nately, there are few empirical data to confirm this. Mor- males than females. In an early medieval skeletal sample tality data based on historical documents are available from Croatia, the frequency of periostitis in males was fromnorthernItalyasearlyasthe14thcenturyandfrom almosttwicethatoffemales(Sˇlaus,2008).Suchpatternsof other European countries (e.g., England, Wales, and higher frequencies of stress markers in men, however, are France)beginninginthe16thcentury,butrelativelygood not universal,and many studies have observedhigher fre- data on ages at death for both sexes, which would allow quencies in women. Larsen (1997) summarizes several forcomparisonsofbothlongevityandage-specificmortal- studies reporting higher frequencies or greater severity of ityratesbetweenthesexes,donotgenerallyappearuntil periostitisamongfemalesinprehistoricSouthwesternU.S. much later, during the 18th century (Russell, 1948; samples. Cucina et al. (2006) found a higher frequency of Hollingsworth and Hollingsworth, 1971; Herlihy, 1977; cribra orbitalia in females in a 2nd–3rd century A.D. Wrigley and Schofield, 1981; Willigan and Lynch, 1982; Roman site, and according to Cohen and Bennett (1993) Ell, 1985; Bonneuil, 1993; Gage, 2005). National data on thereisgenerallyahigherfrequencyofbothporotichyper- mortalitywerenotcollecteduntilthe18thcentury,begin- ostosis and cribra orbitalia among women in prehistoric American JournalofPhysical Anthropology SEXAND FRAILTY 287 and historic skeletal samples. Sˇlaus (2000) found that the sample from the East Smithfield cemetery in London. frequencyofperiostitisinwomenwastwicethatinmenin The East Smithfield cemetery was established in the alatemedievalskeletalsamplefromCroatia,althoughthe 14thcenturyforthe exclusivepurposeofburying victims results were not statistically significant. Studies of maxil- of the Black Death in London (Hawkins, 1990). Given lary sinusitis have found that the frequency of the pathol- that all individuals in East Smithfield died during the ogyissignificantlyhigheramongwomeninvarioussamples Black Death, most (if not all) individuals in the East from Native American, English, and Nubian cemeteries Smithfield cemetery died from the same cause. With the (Robertsetal.,1998;Roberts,2007).Kingetal.(2005)found East Smithfield sample it is therefore possible to control a higher frequency of linear enamel hypoplasia among for cause of death to a greater extent than is possible females in samples from 18th and 19th century London. with normal (i.e., nonepidemic) mortality samples. This Therearealsoseveralstudiesthathavefailedtofinddiffer- allows for the analysis of the differences between men encesinthefrequenciesofstressmarkersbetweenmenand and women in the risks of mortality associated with women.Astudyofstressmarkersina1750–1500B.C.Nu- osteological stress markers without the potential con- biansiterevealednosignificantdifferencesbetweenmales founding effectsof different causes of death. andfemales(BuzonandJudd,2008).Storey(1998)didnot Previous studies have examined the East Smithfield find any differences in the frequency of childhood stress cemetery to determine the selectivity of the Black Death markersbetweenmenandwomenofthesamesocialstatus with respect to sex and pre-existing health conditions (or amongtheLateClassicMayaofCopan.BuikstraandCook frailty) (DeWitte and Wood, 2008; DeWitte, 2009). Such (1981)andPowell(1988)foundnosexdifferencesinthefre- studies were done in part to evaluate the assumption quencyoftuberculosisinseveralMiddleandLateWoodland that because Black Death mortality was very high (i.e., and Mississippian sites. Møller-Christensen (1978) found 30–50% of affected populations died during the epi- no sex differences in the frequency of skeletalsigns of lep- demic), the epidemic killed people indiscriminately, rosyinamedievalDanishsite.Graueretal.(1998)foundin unlike normal causes of mortality, which tend to be a sample from a 19th century Chicago poorhouse that fre- selective. Selective mortality, or selectivity, refers to the quenciesofperiostitis,linearenamelhypoplasia,andother fact that most normal causes of mortality generally tar- lesions potentially indicative of infectious diseases did not get (and thus select out of the population) those individ- differsignificantlybetweenthesexes.Thisisnotanexhaus- uals at the highest risks of death in the population tivesurveyofallthestudiesthathaveexaminedsexdiffer- rather than killing all individuals at the same rate encesinosteologicalstressmarkers.However,thevariation (Wood et al., 1992). Skeletal samples, which are obvi- eveninthisbriefoverviewdemonstratesthatnoconsistent ously samples of dead individuals, are typically not rep- patternwhichwouldindicatesexdifferencesinunderlying resentative of the original living population because of frailtyhasyetbeenresolvedusingskeletaldata. selective mortality. Individuals who die at a given age, Rather than examine frequencies of osteological stress and thus comprise a skeletal sample, are usually those markers to investigate sex differentials in morbidity and with the highest frailty within the population and thus mortality in the past, the study presented here uses a are those most likely to die and be selected out of the hazards model to determine if the excess mortality asso- population (Wood et al., 1992). Previously, it was not ciated with stress markers differs between the sexes in a clear whether the Black Death was selective or if every- medieval skeletal sample from London. Such differences one exposed to the disease was at equal risk of dying. would be informative about whether previous exposure Some contemporary chronicles of the Black Death to stressors affected risks of dying equally for men and include statements that suggest that factors such as age, women.Theultimate goalofthisstudyistoexaminedif- sex, and social status had no effect on risk of death, but ferences in frailty between men and women in a medie- rather, all individuals were equally likely to succumb to val population, and the approach taken here is poten- the disease (e.g., see excerpts from and translations of tially more informative than comparing frequencies of chronicles in Horrox, 1994; Cohn, 2002). DeWitte and stress markers because it does not assume that such Wood (2008) used a multistate model of health and mor- markers indicate identical levels of frailty for all individ- tality (Usher, 2000), which was also used for the current uals in a sample. Rather, this study allows for (but does study and is described below in Materials and Methods, not require) variation between the sexes with respect to to determine whether the Black Death was selective the excess mortality associated with osteological stress with respect to frailty. The study examined the relation- markers. According to Cohen and Bennett (1993: p. 284), ship between osteological stress markers and risk of there may be inherent differences between males and death in East Smithfield to determine if pre-existing females in the body’s ability to buffer stress or the ways health conditions affected risks of death during the in which stress leaves visible signs on the skeleton. If Black Death. Estimates of excess mortality associated females were less frail than males in past populations, with osteological stress markers within the East Smith- they might have developed osteological stress markers field sample were compared to those estimated for a pre- but nonetheless been more resistant to a variety of Black Death normal mortality sample from Denmark to causes of deaththan males.Inother words, the effectsof evaluate whether the patterns of selectivity during the pre-existing health conditions on risk of dying might Black Death cemetery were similar to those that existed have been stronger and thus increased risks of mortality under conditions of normal mortality. Analysis of the to a greater extent among men than among women. If normal mortality Danish sample was necessary to estab- this was the case, the excess mortality associated with lish that the presence of stress markers really was indic- osteological stress markers should be higher among men ative of poor health under normal mortality conditions. than women in skeletal samples; i.e., stress markers All of the stress markers included in the study were should be associated with more highly elevated risks of associated with excess mortality in the normal mortality death inmalesthan females. Danish cemeteries. These results confirmed earlier find- This study examines sex differences in the risk of mor- ings thattheseparticular stress markers are informative tality associated with osteological stress markers using a about health condition or frailty; i.e., individuals with American JournalofPhysical Anthropology 288 S.N.DEWITTE stress markers in Denmark were in poor health and at ous physiological stress, this study examines differences higher risks of mortality than their peers without stress between menand women inthe risks of death associated markers. Results from the East Smithfield Black Death with osteological stress markers in the East Smithfield cemetery suggest that the Black Death was in fact selec- cemetery. Such differences might be informative about tive with respect to pre-existing health, such that indi- underlying differences in frailty between men and viduals who were in relatively poor health before the women inthe medievalpopulation ofLondon. epidemic were more likely than their healthier peers to die during the Black Death. Comparison of the results MATERIALS AND METHODS from the Black Death cemetery to those from a normal Skeletal sample mortality sample revealed a consistent pattern of higher East Smithfield Black Death cemetery. The skeletal estimated values of the risk of death associated with material for this study comes from the East Smithfield stress markers within the normal mortality sampled Black Death cemetery in London, a site in northeast compared to East Smithfield. These differences suggest London located near the Tower of London. The East that though the Black Death was selective with respect Smithfield cemetery is an exclusively Black Death ceme- to pre-existing health, it was perhaps not as strongly tery that is one of only a few excavated cemeteries with selective as normal mortality, presumably because many both documentary and archaeological evidence linking it more otherwise healthy individuals were victims of the to the 14th-century epidemic. The East Smithfield ceme- Black Death than would have died under conditions of tery was established in late 1348 or early 1349 to con- normal, medievalmortality. tain thousands of individuals killed during the Black Another previous study of the East Smithfield ceme- Death in London, and it was used until the epidemic tery examined the effect of sex on risk of death during ended in 1350 (Hawkins, 1990). The cemetery was exca- the Black Death (DeWitte, 2009). For the previous study, vated as part of the larger Royal Mint site by the Mu- a sample from East Smithfield was used to model sex seum of London Department of Greater London Archae- as a covariate acting upon the parameters of the ology (now the Museum of London Archaeology Service: Gompertz–Makeham model of adult mortality; no http://www.molas.org.uk) from 1986 to 1988. The Royal attempt was made to include children given the diffi- Mint site also includes the Cistercian Abbey of St. Mary culty of determining sex for subadults. The estimated Graces (circa 1350–1538), a victualling yard for the value of the parameter representing the effect of the sex Royal Navy (circa 1560–1785), and remains of the Royal covariate did not significantly differ from zero, which Mint itself (circa 1800s–1960s) (Grainger and Hawkins, suggests that there was no substantial difference in the 1988). Excavation of the East Smithfield cemetery risk of death between men and women in London during revealed several hundred skeletons, and (cid:1)600 individu- the Black Death. The results from that study suggested als were recovered and are currently curated at the Mu- that Black Death mortality was not selective with seum of London. The East Smithfield cemetery provides respect to sex among adults; i.e., men and women were a unique sample of the population of mid-14th century at approximately equal risks of dying during the epi- London. Individuals of all ages and both sexes are repre- demic. However, in that previous study, sex was modeled sented in the cemetery. Importantly, for the purposes of as a covariate affecting the entire Gompertz–Makeham this study, all individuals in the cemetery died within a model; i.e., sex was modeled as proportional to the entire short period of time and most, if not all, were victims of hazard,independentofage(DeWitte,2009).Itispossible medieval plague. As the goal of this study is to evaluate that differences in risk of death existed between men sex differences in the risk of mortality associated with and women at different adult ages, but such age pat- osteological stress markers, it is beneficial to be able to terns were not captured using the proportional hazard control, asmuch as possible,for causeof death,as differ- assumption. As discussed below, the results from this ent causes of death might vary in their selectivity with previous study might inform the interpretation of the respect to pre-existing health conditions (DeWitte and results from thecurrent study. Wood, 2008). The East Smithfield cemetery provides an Given the results from the DeWitte and Wood (2008) almost ideal sample for an evaluation of how pre-exist- study, i.e., that the Black Death was selective with ing health conditions differentially affected male and respect to frailty, the current study examines whether femalerisks of succumbingto asingle cause ofdeath. the strength of that selectivity was the same for both For this study, a sample of 299 adults (173 males, 126 sexes or if there were differences between men and females) was selected from the East Smithfield cemetery women. This study addresses the question of whether and analyzed by the author at the Museum of London previous exposure to stressors had the same effect on Centre for Human Bioarchaeology. This sample com- risk of death for men and women. The previous studies prises all of the excavated adults from East Smithfield described above did not address potential sex differences who were preserved well enough to provide sufficient in the excess mortality associated with osteological data on age and sex and the presence of certain osteolog- stress markers. However, it is possible that such differ- ical stress markers. Because of the difficulties associated ences exist. As mentioned above, there may be differen- with determining sex in juveniles, this study only exam- cesbetweenmalesandfemaleswithrespecttothebody’s ines patterns ofmortality among adults. ability tobuffer stressor thewaysinwhich physiological stress leaves visible signs on the skeleton (Cohen and Age and sex estimation Bennett, 1993). It is possible that there are important sex differences in the relationship between osteological Age estimation. Ages were estimated using the method stress markers and mortality in past populations and of transition analysis (Boldsen et al., 2002). This age- that stress markers are therefore not equally informa- estimation method uses statistical methods to avoid the tiveabouthealthcondition,orfrailty,inmenandwomen problem of age mimicry associated with traditional in those populations. To evaluate differences in the approaches (age mimicry refers to estimated ages that strength of Black Death selectivity with respect to previ- are biased toward the age distribution of a known-age American JournalofPhysical Anthropology SEXAND FRAILTY 289 reference sample) (Bocquet-Appel and Masset, 1982; covered by muscle markings that can interfere with Boldsen et al., 2002). For transition analysis, the condi- identification of subtle pathologies. Periostitis was iden- tional probability that a skeleton exhibits a certain age tified macroscopically and was scored as present if the indicator stage given the individual’s known age is esti- there was at least one distinct patch of woven or scle- mated from a known-age reference sample. This condi- rotic bone (or a combination of the two) laid down on the tional probability is then combined with a prior distribu- surface of the diaphysis. tion of ages at death (either a uniform prior or a prior Porotichyperostosisisandcribraorbitaliaarelesionson distribution based on documentary information) using the cranial vault bones and orbital roofs, respectively, Bayes’ theorem to determine the posterior probability which are characterized by a porous appearance of the that a skeleton died at a certain age given that it dis- outertableoftheaffectedbonethatisoftenassociatedwith plays particular age indicator stages. For this study, expansionoftheunderlyingdiplo¨icbone(Mensforthetal., transition analysis was applied to skeletal age indicators 1978; Ortner, 2003). Both porotic hyperostosis and cribra present on the pubic symphysis and iliac auricular sur- orbitalia are usually associated with anemia or other face and to cranial suture closure as described by Bold- causesduringchildhoodthatresultinanexpansionofthe sen et al. (2002), and the ADBOU (Anthropological Data- bone marrow and thus an expansion of the surrounding base, Odense University) Age Estimation software was diplo¨ic bone. Both stress markers can be retained into used to determine individual ages-at-death. The ADBOU adulthood(Walkeretal.,2009).Forthecurrentstudy,the program uses a conditional probability estimated from cranial vault was scored for porotic hyperostosis, and the the Smithsonian Institution’s Terry Collection that an roofs of both orbits were scored for cribra orbitalia. Both individual with certain age indicators will be a given stress markers were identified macroscopically. Porotic age. Theprogram also uses apriorage-at-death distribu- hyperostosis and cribra orbitaliawere scored as present if tion based on data from 17th-century Danish rural par- atleastonesquarecentimeterofporositywasvisible. ish records to provide estimates for the age-at-death dis- Linear enamel hypoplasia is a tooth enamel defect tribution in the target sample. Using Bayes’ theorem, caused by the disruption of enamel formation as a the ADBOU program combines the informative prior and result of infection or malnutrition (Huss-Ashmore et al., the conditional probability to estimate the highest poste- 1982; Dahlberg, 1991; Roberts and Manchester, 2005). rior pointestimateofageforeachindividualinthesample. Enamel hypoplasia occurs only while the teeth are developing during childhood, but it is not subject to Sex determination. Sexwasdeterminedfromdimorphic remodeling and thus can be retained well into adult- features of the skull and pelvis using the standards hood (Roberts and Manchester, 2005). Linear enamel described in Buikstra and Ubelaker (1994). The following hypoplasias appear as horizontal lines of varying width features of the skull were scored: glabella/supraorbital on the surface of the affected tooth. For the current ridge,supraorbitalmargin,mastoidprocess,externalocci- study, linear enamel hypoplasias were identified macro- pitalprotruberance/nuchalcrest,andthementaleminence. scopically on the buccal surface of the mandibular Thefollowingfeaturesofthepelviswerealsoevaluatedfor canines. The mandibular canine has a relatively long sex determination: the ventral arc of the pubis, subpubic developmental time-span ((cid:1)10 years) and is highly sen- concavity, ischiopubic ramusridge, and the greater sciatic sitive to physiological stress (Goodman et al., 1980; notch.Researchershavedemonstratedthattheaccuracyof Huss-Ashmore et al., 1982; Santos and Coimbra, 1999). theseindividualtraits,orcombinationsthereof,forthepur- Only permanent dentition with very little or no wear poses of sex determination ranges from 68 to over 96% were scored. Linear enamel hypoplasia were scored as (Phenice,1969;SutherlandandSuchey,1991;Grawetal., ‘‘present,’’ if one or more depressions on the surface of 1999;UbelakerandVolk,2002;Rogers,2005;Walker,2005; the tooth were palpable and were visible to the naked WilliamsandRogers,2006).Multipleskeletalindicatorsof eye under good lighting. sex were used for this study, given that including more These osteological stress markers are used in this thanoneindicatorimprovestheaccuracyofsexdetermina- study only as nonspecific indicators of general health. It tion (Meindl et al., 1985; Rogers, 2005; Williams and should be emphasized that the Black Death killed too Rogers,2006;Walker,2008). quicklyto leavevisiblesigns onthe skeleton,soosteolog- ical stress markers in the East Smithfield sample were Osteological stress markers not caused by the Black Death itself. For this study, the stress markers indicate an individual’s general level of Using the model described below, this study analyzes health (i.e., pre-existing health condition or previous ex- the risk of death associated with the following nonspe- posure to physiological stressors) before ultimately dying cific osteological markers of physiological stress: linear during the BlackDeath. enamel hypoplasia, porotic hyperostosis, cribra orbitalia, and tibial periostitis. Periostitis is part of an inflamma- tory response to trauma or infection, and it results in Model excess bone formation. It can occur at any age (i.e., it is not generally restricted to childhood as are the other The risk of death associated with osteological stress stress markers used in this study) (Larsen, 1997; Ortner, markers among adults in the East Smithfield cemetery 2003). For the current study, periostitis was scored on was assessed using a multistate model of health and the tibia because paleopathological studies have repeat- mortality that was developed for paleodemography by edly demonstrated that the tibia is commonly affected by Usher (2000) and was previously used to examine periostitis and because the tibia is a robust bone that is selective mortality in the East Smithfield cemetery by therefore often well-preserved in skeletal samples the author (DeWitte and Wood, 2008). The model, (Eisenberg, 1991; Milner, 1991; Larsen, 1997; Roberts shown in Figure 1, has three, nonoverlapping states: and Manchester, 2005). Only the anterior and medial State 1 includes all those individuals in the sample surfaces of the diaphysis of the tibia were scored for without any osteological stress markers that are visible periostitis, as the posterior surface of the tibia is often to the naked eye, State 2 includes those with visible American JournalofPhysical Anthropology 290 S.N.DEWITTE senescent risk of mortality, where a is the overall level 2 of mortality among adults and b is the rate at which this risk increases with age (Gage, 1988). The Gompertz– Makeham function is a parsimonious model of adult mortality that fits the general human pattern of rela- tively low mortality during the young adult ages and an increasing risk of death with senescence (Wood et al., 2002). The hazard of moving from State 1 to State 2, i.e., of developing stress markers, h (a), was estimated as a 12 constant k For this study, an age-specific hazard of 1. moving from State 1 to State 2 was not included in the model because for most osteological stress markers, one does not know the age at which an individual became ill or suffered some other physiological stress sufficient to cause a stress marker, nor does one know precisely how long it took for a stress marker to develop. For simplic- ity, in this study, the age of onset of stress markers was Fig. 1. Three-state model of morbidity and mortality. Indi- modeled as an exponential random variable. The hazard viduals are born into State 1, and transitions between states i of dying from State 2, h (a), was modeled as propor- and j occurat age-specifichazardratesh (a), wherea is age in 23 years. ThetransitionbetweenStates1andij3followsthebaseline tional to the baseline age-specific risk of dying from Gompertz–Makeham hazard function h13(a) 5 a1 1 a2 exp(ba), State 1. Under the proportional hazards specification, k2 and the proportional term associated with a particular stress is a proportional term on the Gompertz–Makeham func- marker, k , acts to modify this baseline rate for those who die tion and is thus independent of age. The k parameter 2 2 from State 2. The hazard of developing a detectable stress value indicates the proportional difference in risk of marker is set equal to the constant k1 (Redrawn from Usher, death between individuals with and without stress 2000.). markers. Estimated k values significantly greater than 2 one indicate that individuals with stress markers were osteological stress markers, and State 3 is death. athigherrisks ofdying compared topeers withoutstress Everyone in the skeletal sample used for this study is, markers. Estimated k values significantly lower than 2 of course, dead and thus observed in State 3, so States one indicate that individuals with stress markers face 1 and 2 represent the possible living states the indi- decreased risks of death compared to their peers. Esti- viduals could have been in right before they died. In mated k values equal to one indicate that individuals 2 the Usher model, individuals can move from State 1 to with and without stress markers were at approximately State 2 (i.e., suffer some physiological stress and thus the same riskofdeath. develop an osteological stress marker), and individuals For this study, the Usher model was fit to data on age, can die from either of the two living states (i.e., with sex, and the presence of stress markers. To evaluate sex or without stress markers). The transitions from either differences in the excess mortality associated with osteo- of the two living states to death are determined by logical stress markers within the East Smithfield sam- age-specific hazard rates. The model allows for (but ple, and thus whether previous exposure to stressors does not require) variation in the hazard rates had the same effect onrisk ofdeath for menand women, between each of the two living states and death; i.e., sex was modeled as a covariate affecting the k parame- 2 the hazard of making the transition from State 2 to ter in the Usher model. In this analysis, females were death can be higher or lower than, or the same as the coded as 0, and males were coded as 1. A significant pos- hazard of making the transition from State 1 to death. itiveor negative estimate forthe parameter representing Thus, the model can be used to estimate the differen- the effect of the sex covariate would suggest that the tial risk of death associated with the living states. The excess mortality associated with stress markers was model allows one to address the question of whether higher or lower, respectively, for men compared to certain osteological stress markers really are associated women. The model was fit separately to data on the with increased risks of death—i.e., whether individuals presence of each of the four following stress markers: po- with osteological stress markers are at elevated risks rotic hyperostosis, cribra orbitalia, linear enamel hypo- of dying compared to their peers without such stress plasia, and tibial periostitis. Maximum likelihood analy- markers (Ortner, 1991; Wood et al., 1992). Even sis was used to estimate the parameters for this study, though everyone in a cemetery sample is observed in and estimation was done with Holman’s mle program State 3, data on age-at-death and the presence of (2005). Likelihood ratio tests (LRT) were used to evalu- osteological stress markers can be used to estimate all ate the fit of the full model, which included all the pa- of the parameters of the model described below. rameters of the Usher model described above plus the For this study, the baseline risk of death from State 1, parameter representing the effect of the sex covariate, h13(a), was specified as a three-parameter Gompertz– compared to the reduced model that did not include sex Makeham model: as a covariate. The LRT therefore tests the null hypothe- sis that the risk of death associated with stress markers hðaÞ¼a þa eba was the same for men and women. The LRT was com- i 1 2 puted as follows: In thismodel, a is the age ofthe ith skeleton inyears, LRT 5 22[ln(L ) 2 ln(L )], where LRTapproxi- i reduced full a is the constant age-independent risk of mortality that mates a v2 distribution with df 5 1 (the degrees of free- 1 everyone within the population faces (i.e., the chance of dom are equal to the number of additional parameters in dying from accidents and other causes that are unre- the full model). It should be noted that the errors associ- lated to aging), and a eba is the exponentially increasing ated with age estimates were not taken into account 2 American JournalofPhysical Anthropology SEXAND FRAILTY 291 when estimating the Usher model. Therefore, the esti- TABLE1. Maximumlikelihoodestimatesofk andtheeffectof 2 mated values of the k parameter and the parameter thesexcovariatewithassociatedstandarderrors,andthe 2 representing the effect of the sex covariate should be resultsofthelikelihoodratiotestofthenullhypothesis viewed as general, qualitative measures of the excess H0:effectofsexcovariate50 mortality associated with stress markers and the sex dif- Stressmarker k^ ðs:e:Þ Sex(s.e.) LRT 2 ferences thereof. Tibialperiostitis 1.74(0.30) 0.47(0.28) 50.8 A hazards model was used in this study rather than PoroticHyperostosis 1.72(0.27) 1.71(0.33) 15.6 the more traditional life-table approach to avoid the CribraOrbitalia 1.90(0.60) 0.84(0.36) 5.45 potential problems associated with paleodemographic life LEHmandibularcanine 2.90(1.39) 1.3(0.54) 12.3 tables. Life tables, which display age-specific mortality rates (and often other, related measures), are an impor- LEH5linearenamelhypoplasia. tant tool in demographic research, and there is a long history of their use in paleodemography (Wood et al., field, which thereby suggested that the Black Death was 2002). The paleodemographic use of life tables is based selective with respect to frailty (DeWitte and Wood, on the idea that the distribution of ages-at-death within 2008). For the current study, though the estimated val- a skeletal sample is equivalent to the cohort age-at- ues of k suggest that exposure to stressors prior to the 2 death column in a life table (Milner et al., 2008). How- Black Death increased the risk of death for both men ever, this is true only when the population that gave rise and women during the epidemic, the consistently posi- to a cemetery sample was stationary (i.e., closed to tive estimated values of the parameter representing the migration, an intrinsic rate of increase equal to zero, effect of the sex covariate suggests that the excess mor- and age-specific mortality and fertility rates that do not tality associated with stress markers was higher for men change over time), and the assumption that a past popu- thanfor women. lation was stationary is not necessarily valid (Milner et It should be noted that the k estimates and the esti- 2 al., 2008). Furthermore, life tables are considered by mated values of the parameter representing the effect of many researchers to be an inefficient way to deal with the sex covariate have relatively large standard errors paleodemographic age-at-death data (Buikstra, 1997; associated with them as shown in Table 1, and these Hoppa and Vaupel, 2002; Konigsberg and Frankenberg, standard errors are probably underestimated. As men- 2002; Wood et al., 2002; Milner et al., 2008). Life tables tioned above, to estimate the parameters of the model require the estimation of one parameter (the central used in this study, point estimates of age were used mortality rate) for every age interval used, and there is without their corresponding errors. Because the errors no way to do this reliably without huge samples and associated with ages were not incorporated when the knowledge of the original population at risk, both of model was fit, the reported standard errors might be which are often unavailable for most paleodemographic underestimated to an unknown degree, and the standard investigations (Wood et al., 2002). Some paleodemogra- error estimates should thus be viewed with caution. phers have tried to avoid these and other problems asso- Nonetheless, there is a consistent pattern, as the k esti- 2 ciated with life tables by using a model life table mates for all stress markers are greater than one and approach, i.e., the application of a theoretical life table the estimated values of the parameter representing the that matches the age-at-death distribution observed in a effect of the sex covariate are positive. The consistent cemetery sample (Weiss, 1973; Wood et al., 2002). pattern suggeststhatthesestress markers are indicative Unfortunately, it is not always clear which model life ta- of higher frailty and that the risk of death associated ble is most appropriate to use in paleodemographic stud- with such markers were higher for men. However, the ies (Gage, 1988; Milner et al., 2008). Many researchers numerical value of estimates of the k parameter and 2 argue that rather than using life tables, some form of the parameter representing the effect of the sex covari- parametric or semiparametric hazards analysis is the ate should not necessarily be taken at face value given most powerful way to derive information from the small the uncertainty associated with the standard error esti- samples typical of paleodemography (Gage, 1988; mates. The results of the LRT further suggest that the Konigsberg and Frankenberg, 1992, 2002; Buikstra, excess mortality associated with stress markers was 1997; Hoppa and Vaupel, 2002; Wood et al., 2002). higher for men than women in the East Smithfield sam- According to Holman et al. (2002), most human age-at- ple, as they indicate that including the sex covariate death distributions can be described using five or fewer improved the fit of the model with respect to all of the parameters, and parsimonious parametric models pro- stress markers considered. vide a useful alternative to life tables in paleodemo- graphic studies. DISCUSSION RESULTS The estimated values of the excess mortality associ- ated with osteological stress markers from this study The maximum likelihood estimates of k , the excess suggest that exposure to physiological stressors prior to 2 mortality associated with the osteological stress markers the Black Death increased the risk of death for both used in this study, and the estimated values of the pa- men and women during the 14th-century epidemic. Men rameter representing the effect of the sex covariate, and women with osteological stress markers were more along with the standard errors associated with these likely than their peers without such markers to die dur- estimates and the results from the LRT, are shown in ing the Black Death. As mentioned above, these results Table 1. These results indicate that all stress markers are not particularly surprising, given similar results are associated with increased risks of death in the East from a previous study of East Smithfield (DeWitte and Smithfield sample, as all the estimated k values are Wood, 2008). However, even though it appears that prior 2 greater than one. These results are not surprising given exposure to stressors elevated risks of death for both similar findings from a previous study of East Smith- sexes during the Black Death, the estimated values for American JournalofPhysical Anthropology 292 S.N.DEWITTE the parameter representing the effect of the sex covari- Esimai et al., 2001; Zhou et al., 2005; Sakisaka et al., ate suggest that the excess mortality associated with the 2006; Dey and Chaudhuri, 2008). According to Chen stress markers considered in this study was higher for et al. (1981), because of preference for sons in rural Ban- men than for women. These differences between men gladesh, malnutrition is much more common in female and women might suggest that the effect of pre-existing children and is at least partly due to unbalanced provi- health condition on risk of dying during the Black Death sioning of children within families. The health effects of was stronger among men than among women and that such disparities are exacerbated because male children males were, on average, frailer than women in the popu- are also much more likely to receive medical care, and lation of medieval London. That is, pre-existing health they therefore are more likely to recover from infectious conditions made all individuals more vulnerable to death diseases. Kehoe and Giletti (1981) and Rosenberg (1980) during the Black Death, but the risk for men relative to summarize several studies that have demonstrated cul- their healthier peers might have been even greater than tural dietary rules that prevent females from obtaining that among women. Women in medieval London with a adequate high quality foods in modern stratified soci- history of physiological stress might have been better eties, such as restrictions on female consumption of able to resist dying from the Black Death, perhaps meat, certain vegetables, or dairy products. According to because of superior immune competence. An alternative Ortner (1998), in many developing countries, men have explanation of the observed data is that, rather than priority access to higher quality and larger quantities of suggesting sex differences in frailty in general, these food than women, and he argues that such differential results might suggest that the Black Death was not access to food has probably existed for a long time in quite as strongly selective with respect to frailty among human populations. Unequal distribution of food and women and killed more otherwise healthy women than other resources between the sexes might have important healthy men. consequences for observed patterns of frailty. Indeed, The possible interpretation that women were less frail many of the examples of preferential treatment of males than men in medieval London is consistent with the occur (or occurred in the past) in areas in which males widespread perspective that female buffering or female experience longer life expectancies and lower age-specific advantages in morbidity and mortality are biological mortality rates. norms for humans and other species. Numerous studies Given that malnutrition increases vulnerability to in- have revealed female morbidity and mortality advan- fectious disease (Scrimshaw, 1987; Ortner, 1998), did sex tages in humans and in many other animal species determine access to food in medieval England in ways (Promislow, 1992; Klein, 2000; Moore and Wilson, 2002). that might have influenced differences in frailty? In As described above, in most modern human populations, many parts of medieval Europe, anything a woman women tend to experience lower age-specific mortality owned became the legal property of her husband, so sons rates at most ages and live longer than men. There are were generally preferred over daughters because of the exceptions, and in some countries (e.g., in South Asia), desire to keep heritable properties within the family life expectancy at birth is lower for women and age-spe- (Bennett, 1987; Wiesner, 2000). This preference might cific mortality is higher for females at many ages have resulted in reduced care for or provisioning of (Bhatia, 1984). However, fetal and neonatal mortality daughters. However, the data from historical records on rates are higher for males in those countries (Chen differential treatment of sons and daughters are ambigu- et al., 1981; Bhatia, 1984). Furthermore, studies of the ous (Bennett, 1987). In early medieval religious com- effects of stress on prenatal growth generally show that munities, women were given smaller rations of food males are more vulnerable to stress in utero than are because of their smaller size (Pearson, 1997), and among females (Stinson, 1985). The sex differentials that occur the laity and in religious communities, women were during the prenatal or neonatal period are the result of encouraged to fast as a means of reducing sexual urges endogenous factors (i.e., rather than being the result of (Bynum, 1987). According to Pearson (1997), the nutri- preferential treatment or heterogeneity in exposure to tional needs of early medieval women, particularly given risk factors) and thus, according to some researchers, the strains of pregnancy, childbirth, and lactation, were reflect the ‘‘innate frailty of the male’’ (Bhatia, 1984: probably not adequately met for many women, given p. 167), or the ‘‘higher biological risk of male children’’ what is known about the foods available at that time. (Chen et al., 1981: p. 57). When postnatal morbidity and Some historical records indicate that women did not live mortality is lower in males than females, researchers as long as menin the early Middle Ages, and theirappa- often attribute the differential to preferential treatment rent reduced longevity might have been the result of of males (Chen et al., 1981; Stinson, 1985). Differentials iron-deficiency anemia caused by a lack of sufficient iron favoring males are also often attributed to the complica- in the diet coupled with iron depletion through menstru- tions associated with pregnancy and childbirth that ulti- ation, pregnancy, and lactation (Bullough and Cameron, matelyact to balance the natural femaleadvantages (see 1980). Premenopausal women would have been at ele- Lopez and Ruzicka, 1984;Gage, 1994). vated risks of anemia because of their higher iron If, as the results of this study might suggest, women requirements compared to men, and anemia might have were less frail than men in medieval London, could this rendered women more susceptible to morbidity and mor- be potentially informative about female access to resour- tality from a variety of other causes. According to ces in medieval London? This is important because Bullough and Campbell (1980), several historical sources nutritional status can affect immune functioning. If indicate that woman were living longer than men begin- females do not have adequate nutrition, they might be ning in the 14th century, and they argue that the differ- immunocompromised despite any innate advantages ential in favor of women was the result of dietary associated with being female. Higher levels of malnutri- changes, particularly an increase in the amount of pro- tion among females compared to males have been tein and iron available in the diet. Following the adop- observed in many studies within modern populations tion of the three-field system of agriculture in the 9th (e.g., Chen et al., 1981; Bhatia, 1984; Khadi et al., 1996; century, more protein-rich plant foods, such as peas and American JournalofPhysical Anthropology SEXAND FRAILTY 293 broadbeans, were included in the typical diet, and at the the Black Death killed women totally irrespective of pre- same time, more meat was also apparently being con- existing health condition. However, the lower excess mor- sumed (Bullough and Cameron, 1980). These dietary tality associated with stress markers among women com- changes might have benefitted women more than men pared to men as indicated by the estimated values of the because of an increase in iron-rich foods in the diet that parameter representing the effect of the sex covariate lowered the prevalence of anemia among women. The mightindicatethattheBlackDeathkilledmoreotherwise results of the current study might provide additional healthy women than healthy men. This interpretation of support forthis argument. the results in turn suggests that the Black Death dispro- Studies of dietary differences between medieval men portionately affected women. Alternatively, if the results and women have been done using stable isotope analysis indicate that men in medieval London were frailer than of skeletal material. Such studies are somewhat limited, women,areasonableinferenceisthatmeningeneralwere however, as they are not capable of revealing everything at a higher risk of death than women during the Black that was consumed by individuals in the past. Recon- Death. Thus, the two alternative explanations for the struction of diets, based on strontium and barium analy- results observed in this study lead to very different infer- ses, in an early medieval site in Poland does not reveal encesaboutthesexpatternofBlackDeathmortality.There any differences between males and females with respect are numerous modern examples of diseases that differen- to the consumption of seafood and milk products tially affect either women or men, some of which are (Szostek et al., 2009). At the site of St. Andrew, Fisher- describedabove,soeitherscenarioisplausible. gate in York, northern England, dated to the 13th–16th Is either inference about the sex pattern of Black centuries, nitrogen and carbon isotope ratios were gener- Death mortality consistent with existing evidence ally lower among females than males, suggesting that regarding sex differences in medieval plague mortality? women consumed less marine foods than men (Muldner Several contemporary chroniclers believed that more and Richards, 2007). However, analysis of males and women than men were killed by the 14th-century Black females from the same household and status revealed no Death, while others claimed that the epidemic killed significant differences between the sexes. The lower indiscriminately (Cohn, 2002). Previous investigations of ratios for women within the site might reflect different the sex patterns of Black Death mortality using skeletal migratory patterns or differences between members of remains from the East Smithfield cemetery suggest that the local monastic order and the laity rather than differ- neither sex was at a disproportionately elevated risk of ential access to food based on sex. Mays (1997) found no dying during the Black Death relative to the patterns differences in isotopes indicative of marine resources observed under conditions of normal mortality (Waldron, between males and females in two 13–16th century sites 2001; DeWitte, 2009). However, some have interpreted in northern England (Hartpoole and Newcastle). Studies the higher number of males within East Smithfield as of sites dating to the 11–14th centuries in Orkney indicating female advantages during the Black Death revealed significant difference between the sexes, with (Waldron, 2006). Burial records and several historical more marine resources consumed by men, perhaps accountsindicatethatinsomeofthe outbreaksof medie- reflecting differences in activities (e.g., greater involve- val plague that followed the 1347–1351 epidemic, men ment of men in fishing or seal hunting) (Barrett and were disproportionately affected. However, several other Richards, 2004;Richards et al.,2006). chroniclers also noted that mortality was the same for The existing historical and stable isotope data provide men and women in some of these subsequent outbreaks useful information about medieval dietary patterns, but (Cohn, 2002). For example, burial records for the London they do not clarify whether men and women in medieval parish of St. Botolph’s without Bishopsgate reveal an London were equally well (or poorly) nourished or if one excess of male deaths in the 1603 and 1625 epidemics sex had preferential access to food. If the results of this (Hollingsworth and Hollingsworth, 1971). But this pat- study really do indicate lower frailty among women in ternwasneitherconsistentacrossall parishesinLondon medieval London, they, in turn, suggest that women in nor across the 1593, 1603, and 1625 outbreaks (Finlay, general were not more highly malnourished than men. If 1979). Bradley (1977) did not find a difference in mortal- women in medieval London had lower average frailty ity between men and women during the 1665-6 plague than men, this might have been the result of a lack of in Eyam, England, and Schofield (1977) similarly found disparities in access to food resources. Approximately no evidence that either sex was disproportionately equal access to food perhaps enabled these women to affected in the town of Colyton, England that same year. achieve their biological potential of lower frailty relative Thus, the existing evidence regarding mortality during to men. Alternatively, men might have eaten more and the Black Death and subsequent outbreaks of plague better foods than women, but given that medieval diets does not clearly indicate that one sex was disproportion- in general might not have been very high in nutritional ately affected. This evidence therefore neither strongly quality (Dyer, 1983), preferential access might not have supports the conclusion that the results of this study made any appreciable difference with respect to immune indicate higher male frailty in medieval England nor functioning andfrailty. that the results suggest the Black Death was less The alternative explanation for the results from this strongly selective for frailty among women. Future stud- studymustbealsoconsidered.Asmentionedabove,rather ies of the differences in the risks of death associated than demonstrating lower frailty in women compared to with osteological stress markers might help determine men, the values for the parameter representing the effect which of these two possible interpretations of the data of the sex covariate shown in Table 1 might indicate that from this study is more likely. For example, finding that the Black Death discriminated less strongly between the differences with respect to the risks of death associ- women with and without pre-existing health conditions ated with stress markers between men and women in thanwastrueformen.Thereisexcessmortalityassociated noncatastrophic cemeteries are similar to those observed withosteologicalstressmarkersforbothmenandwomen, in East Smithfield might further support the idea that sothe results should not be interpreted asindicating that women werelessfrailthan meninpastpopulations. American JournalofPhysical Anthropology 294 S.N.DEWITTE CONCLUSION Berbesque JC, Doran GH. 2008. Brief Communication: physio- logical stress in the Florida archaic—enamel hypoplasia and The results of this study suggest differences between patterns of developmental insult in early North American men and women with respect to the excess mortality hunter-gatherers. Am J Phys Anthropol 136:351–356. associated with osteological stress markers. These Bhatia S. 1984. Traditional practices affecting female health results suggest that in medieval London, a difference and survival: evidence from countries of South Asia. In: Lopez AD, Ruzicka LT, editors. Sex differentials in mortal- existed in the average frailty of men and women, such ity: trends, determinants, and consequences. Canberra: that exposure to physiologic stressors increased risks of Dept. of Demography, Australian National University. p 165– death to a greater extent among men than women. If 191. this is true, it might mean that the modern pattern of Blessmann J, Van Linh P, Nu PA, Thi HD, Muller-Myhsok B, excess female longevity has underlying causes that are Buss H, Tannich E. 2002. Epidemiology of amebiasis in a not unique to modern populations and that the morbid- region of high incidence of amebic liver abscess in central ity and the mortality advantages experienced by women Vietnam.AmJTropMedHyg66:578–583. in modern populations also existed in the past. An alter- Bocquet-Appel JP, Masset C. 1982. Farewell to paleodemogra- native explanation for the results of this study is that phy.JHumEvol11:321–333. the Black Death was selective with respect to frailty BoldsenJL,MilnerGR,KonigsbergLW,WoodJW.2002.Transi- tion analysis: a new method for estimating age from skele- among both men and women, but that it discriminated tons. In: Hoppa RD, Vaupel JW, editors. Paleodemography: less strongly among women. If this was the case, the age distributions from skeletal samples. Cambridge: Cam- lower risk of mortality associated with stress markers bridgeUniversityPress.p73–106. among women compared to men was the result of the Bonneuil N. 1993. The trend method applied to English data. Black Death killing a greater proportion of healthy In: Reher DS, Schofield R, editors. Old and new methods in women than healthy men, rather than reflecting lower historicaldemography.Oxford:Clarendon.p57–65. average frailty among women. Additional studies of the Brabin L, Brabin BJ. 1992. Parasitic infections in women and sex differences in the risk of mortality associated with theirconsequences.AdvParasitol31:1–60. stress markers in a variety of cemetery samples might BradleyL.1977.SomemedicalaspectsofPlague.In:Cambridge yield data that further support the hypothesis that Group for the History of Population and Social Structure, editors. The Plague reconsidered: a new look at its origins women were less frail on average than men in past pop- and effects in 16th and 17th century England. Matlock, ulations. Derbyshire:LocalPopulationStudies.p11–24. Buikstra JE. 1997. Paleodemography: context and promise. In: Paine RR, editor. Integrating archaeological demography: ACKNOWLEDGMENTS multidisciplinary approaches to prehistoric population. Car- bondale, Ill: Center for Archaeological Investigations, South- The author thanks Christopher Ruff, the Associate ernIllinoisUniversityatCarbondale.p367–380. Editor,andtwoanonymousreviewersfortheirverycare- Buikstra JE, Cook DC. 1981. Pre-Columbian tuberculosis in ful reading of and helpful comments on this manuscript. West-Central Illinois: prehistoric disease in biocultural per- She is very grateful to Bill White, Jelena Bekvalac, and spective. Northwestern University Archaeological Program Rebecca Redfern at the Museum of London Centre for ScientificPapersNo.5.In:BuikstraJE,editor.Prehistorictu- Human Bioarchaeology for providing access to the East berculosis in the Americas. Evanston, IL: Northwestern Uni- Smithfield skeletons and for generously providing the versityArchaeologicalProgram.p115–139. physical facilities for this work. She also thanks Dr. Eric Buikstra JE, Ubelaker DH, editors. 1994. Standards for data Jones forhis helpfulcomments andcontinued support. collection from human skeletal remains: proceedings of a seminar at the Field Museum of Natural History (Arkansas ArchaeologyResearchSeries44).Fayetteville,Ark.:Arkansas ArcheologicalSurveyPress. LITERATURE CITED BulloughV,CampbellC.1980.Femalelongevityanddietinthe MiddleAges.Speculum55:317–325. Aalen OO. 1994. Effects of frailty in survival analysis. Stat BuzonMR,JuddMA.2008.InvestigatinghealthatKerma:sac- MethodsMedRes3:227–243. rificialversusnonsacrificialindividuals.AmJPhysAnthropol Acuna-SotoR,MaguireJH,WirthDF.2000.Genderdistribution 136:93–99. in asymptomatic and invasive amebiasis. Am J Gastroenterol BynumCW.1987.Holyfeastandholyfast:thereligioussignifi- 95:1277–1283. cance of food to medieval women, Vol. 16. Berkeley: Univer- AlexanderJ. 1988. Sex differences and cross-immunity in DBA/ sityofCaliforniaPress. 2 mice infected with L. mexicana and L. major. Parasitology Cassidy CM. 1984. Skeletal evidence for prehistoric subsistence 96(Part2):297–302. adaptation in the Central Ohio River Valley. In: Cohen MN, Ansar Ahmed SA, Hissong BD, Verthelyi D, Donner K, Becker Armelagos GJ, editors. Paleopathology at the origins of agri- K, Karpuzoglu-Sahin E. 1999. Gender and risk of autoim- culture.NewYork:AcademicPress.p307–345. munediseases:possibleroleofestrogeniccompounds.Environ Cerfolio RJ, Bryant AS, Scott E, Sharma M, Robert F, Spencer HealthPerspect107(Suppl5):681–686. SA, Garver RI. 2006. Women with pathologic stage I, II, and Appelros P, Stegmayr B, Terent A. 2009. Sex differences in III non-small cell lung cancer have better survival than men. stroke epidemiology: a systematic review. Stroke 40:1082– Chest130:1796–1802. 1090. ChenLC,HuqE,D’souzaS.1981.Sexbiasinthefamilyalloca- Barrett JH, Richards MP. 2004. Identity, gender, religion and tion of food and health care in rural Bangladesh. Popul Dev economy: new isotope and radiocarbon evidence for marine Rev7:55–70. resource intensification in Early Historic Orkney, Scotland, Choi BG, McLaughlin MA. 2007. Why men’s hearts break: car- UK.EurJArchaeol7:249–271. diovascular effects of sex steroids. Endocrinol Metab Clin BennettJM.1987.WomeninthemedievalEnglishcountryside: NorthAm36:365–377. genderandhouseholdin Brigstock beforethe plague, Vol.15. Coale AJ. 1991. Excess female mortality and the balance of the NewYork:OxfordUniversityPress.322p. sexes in the population: an estimate of the number of Bennett KA. 1973. On the estimation of some demographic ‘‘missingfemales’’.PopulDevRev17:517–523. characteristics on a prehistoric population fromthe American CohenMN,BennettS.1993.Skeletalevidenceforsexrolesand Southwest.AmJPhysAnthropol39:223–231. gender hierarchies in prehistory. In: Miller BD, editor. Sex American JournalofPhysical Anthropology