Skeletal Effects of Castration on Two Eunuchs of Ming China Jacqueline T

ANTHROPOLOGICAL SCIENCE Advance Publication Vol. advpub(0), 000–000, 2009

Skeletal effects of castration on two eunuchs of Ming China Jacqueline T. ENG1*, Quanchao ZHANG2, Hong ZHU2 1Department of Anthropology, Western Michigan University, Kalamozoo, MI 49008, USA 2Research Center for Chinese Frontier Archaeology, Jilin University, Changchun 130012, China

Received 30 April 2009; accepted 24 September 2009

Abstract The practice of castrating men is an ancient one. Eunuchs have served as guards to harems and as palace chamberlains for many early courts, but details about their lives are often hazy or shroud- ed in secrecy. Although the changes wrought to their physical appearance from castration are well- documented, little is known about the magnitude of the skeletal changes resulting from the loss of sex hormones associated with the procedure. Such a loss of hormones, especially before puberty, affects skeletal growth and development and may result in early osteoporosis as well as impacting quality of life. The burials of two eunuchs from the Ming Dynasty (1368–1644 AD) of imperial China provide an opportunity to examine the consequences of castration upon the human skeleton. These eunuchs may have been castrated at different periods in their lives. One eunuch appears to have been castrated before the development of secondary sexual characteristics; the delayed epiphyseal closure accompa- nying androgen deficiency may account for his long limbs. Skeletal evidence also sheds light on the lives of these eunuchs, including their oral health, history of childhood stress, and activity patterns.

Key words: castration, eunuch, height, paleopathology, China

Introduction firmly entrenched in the imperial culture and persisted for over 25 dynasties (4000 years) until the early 20th century Castration is any action—surgical, chemical, or other- when the last emperor was removed from power. wise—by which a biological male loses testicular function. Eunuchs (castrated men) were part of the societies of ancient Castration procedure in imperial China Mesopotamia, Assyria, Israel, Ethiopia, Egypt, Persia, While castration typically involves the removal or mutila- Greece, Rome, Byzantium, Korea, and China (Tsai, 1996; tion of the testes, the penis may also be removed (penecto- Scholz, 2001). In the 16th–19th centuries, castrati such as my), and this total removal of external genitalia was the Farinelli were castrated before puberty to preserve their so- practice in dynastic China (Mitamura, 1970). In the late 19th prano voices for opera and for liturgical singing in the Cath- century, Stent (1877) gathered data on the castration proce- olic Church (Jenkins, 1998). In Eastern Europe, members of dure of the Qing Dynasty (1644–1912 AD). More recent in- the Christian Skoptzy sect (18th–early 20th century) under- terviews with surviving eunuchs indicate that procedures went castration in response to Christ’s counsel to amputate were relatively similar in the later years of the dynasty. As body parts to attain purity through the avoidance of sin described by Anderson (1990), the operations were per- (Matthew 18: 8–9 and 19: 12 KJV). Even today, castrated formed by government-approved, unsalaried specialists out- men (hijra) exist in India and small numbers of eunuchs side the western gate of the imperial palace. During the pro- serve as guards in parts of the Middle East (Scholz, 2001, cedure, the abdomen and upper thighs were tightly bound p. 26), while countries such as the United States sometimes with strings or bandages, and then the genitalia were washed castrate men as social control measures for deviant aggres- in hot pepper-water as a local anesthetic. While semi- sive and sexual behavior, either surgically or via other means reclined and held down by assistants, the specialist used a such as chemical treatment (Scott and Holmberg, 2003). In slightly curved blade (Figure 1) to slice off the scrotum, pe- ancient cultures, eunuchs were entrusted to guard women’s nis, and testes with a single cut (Wong and Wu, 1932). A quarters or act as chamberlains because they were deemed plug was inserted into the urethra to prevent stricture forma- safe from the sexual appetites that tempted normal men. tion, and the wound was covered and bound with moistened Arguably the most famous and long lasting example of paper. The new eunuch was made to walk around for 2–3 eunuch service is in China, where the eunuch system was hours before being allowed to lie down. During the next three days after the operation the eunuch was not allowed to drink water or urinate. After this period, the plug was re- * Correspondence to: Jacqueline T. Eng, Department of Anthropol- moved, and if urine appeared, the operation was deemed a ogy, 1045 Moore Hall, Western Michigan University, Kalamozoo, MI 49008, USA. success, with full recovery of the wound expected after E-mail: [email protected] about 100 days. In addition to severe pain, however, many Published online 1 December 2009 died from hemorrhage, infection, urine blockage, or other in J-STAGE (www.jstage.jst.go.jp) DOI: 10.1537/ase090430 complications, and urinary incontinence appears to have

chemical castration as hormone therapy for prostate cancer. Both forms of castration result in osteoporosis as measured by femoral neck mineral density (Daniell et al., 2000), and appear to increase susceptibility to osteoporotic fractures (Daniell, 1997; Smith et al., 2006). It has been suggested that osteoporosis in castrated males is linked to: (1) the absence of protein-anabolic action of the male steroids, which leads to decreased osteogenesis during remodeling; and (2) decreased bulk and strength of the skeletal muscles (Heller and Shipley, 1951). Most medical studies have focused on hypogonadism as Figure 1. Implements used to castrate men of the Chinese impe- the cause of androgen loss in human males, although a few rial court. Redrawn from Wong and Wu (1936) by J. Eng. studies have been conducted on eunuchs. Read (1921) studied the metabolism of eunuchs by analyzing urine samples from several adult “Oriental” eunuchs. The eunuchs secreted been a common occurrence. abnormally high levels of ammonia and creatine, which The mortality rate for such an operation was reportedly were comparable to levels normally expected from a prepu- not higher than found for any other surgery at the time, with bescent boy or female (or an adult male afflicted by certain about 2% for eunuchs of the Qing Dynasty (Wu and Gu, types of illnesses or dietary problems). One eunuch castrated 1991). This rate is considerably lower than other accounts of at age 29, after the development of secondary sex character- high mortality from other regions, such as a survival rate of istics, secreted creatine levels comparable to normal males. only 25% for those castrated in Coptic monasteries in Egypt Read concluded that the removal of male sexual organs in for the slave trade, possibly owing to differences in sanitary prepubescent males results in the eventual chemical and conditions (Scholz, 2001). The highest documented mortali- physical development of secondary female characters in ty rate in China was during the early Ming Dynasty (1368– those castrated males. 1644 AD) when 329 of the 1565 boys (20%) from the de- More recent clinical studies are limited to men who have feated Miao tribe in southwest China died from the castra- been castrated for health issues, or surgical repair for those tion procedure (Tsai, 1996). castrated from assault or accidents, while studies of genital self-mutilation mainly focus on the psychological aspects Previous studies: physical effects of castration (Catalano et al., 2002; Brett et al., 2007; Wassersug and Little research has been published about the skeletal Johnson, 2007). Studies of the hijras (also known as hijada) consequences of castration in humans (see Zitzmann and in India mainly concentrate on the cultural aspects of their Nieschlag, 2001, for critique of studies), although many re- institutionalized third-gender role (e.g. Nanda, 1985; Patel, ports describe the endocrinological effects of castration on 1988), while issues of health often center on risky sexual be- laboratory animals (e.g. Wink and Felts, 1980; Schenck and haviors (e.g. Baqi et al., 2006), making generalizations of Slob, 1986; Zumpe et al., 1992; Dixson, 1993; Davis, 2000). their health status problematic. Investigations into the long- Castrated males lose function of their testicles, and are thus term health effects of castration have focused on three sterile and have greatly reduced production of sex hormones, groups of men: the Skoptzy; the court eunuchs of the Chi- testosterone in particular. Silberberg and Silberberg (1971) nese Empire; and those of the Ottoman Empire (Wilson and have summarized previous research on the effects that the Roehrborn, 1999). The numbers among all three groups absence or deficiency of sex steroids have on the skeletal de- have dwindled significantly, and the last known eunuch of velopment of humans and other mammals. In particular, tes- the Chinese imperial court died in 1996 (Seth, 1996). ticular hormones strongly influence the development of tu- Wilson and Roehrborn (1999) have summarized the most bular and flat bones, as well as the development of commonly documented long-term medical consequences of secondary sexual characteristics. Testicular deficiency caus- castration upon these three groups of men. The health effects es a delay in the closure of epiphyses and the age at which include: enlargement of the pituitary, breast enlargement they normally unite during puberty. The earlier that castra- (gynecomastia), reduction and/or disappearance of the pros- tion occurs, the longer the delay until epiphyseal closure. tate, and skeletal changes. Of the latter, osteoporosis in the This delay results in disproportionately long bone lengths form of bone thinning in the skull and kyphosis of the spine, between those elements that fuse before puberty, which are and the failure of epiphyseal closure were documented not at all or less affected by castration, and the long bones among Skoptzy and Chinese eunuchs. The effects were typ- whose epiphyses normally unite during puberty. Thus, dep- ically more pronounced in those eunuchs who had been cas- rivation of male steroid hormones delays closure of epiphy- trated at a younger age, before development of secondary ses and prolongs the growth period, and also influences the sexual characteristics. These physiological changes also ap- maintenance of bones. pear in modern clinical studies of males with hypogonadism In animals such as rats and macaques, osteoporosis has and androgen deprivation (Heller and Shipley, 1951; Wilson been linked to castration owing to the subsequent loss in et al., 1980; Flaig and Glode, 2008). male sex hormones (Wink and Felts, 1980; Schenck and Wu and Gu (1991) also studied the long-term effects of Slob, 1986). Other studies have been conducted on men who castration upon 26 surviving eunuchs of the Chinese Qing have undergone orchiectomy (removal of a testicle), or court. Seven became eunuchs between the ages of 10 and 14 Vol. 117, 2009 SKELETAL EFFECTS OF CASTRATION 3 years, and 18 between 16 and 26 years, while one could not documented in other cases that tomb inscriptions indicated recall. At time of their examination in 1960, the men com- the eunuch status of those interred, as such people were at prising this sample had been eunuchs for an average of 54 times accorded relatively high status and given elaborate years. Characteristics they shared with descriptions of other burial treatments reflecting that power (Mai, 1977). Thus, castrated men included: beardlessness; enlargement of based on archeological context, these individuals, whose breasts in nine eunuchs out of 25; tall stature in 12 (where tomb inscriptions labeled them as eunuchs, are indeed eu- height > 170 cm); and five displayed unusually small pros- nuchs, and the osteological evidence also suggests a life of tates. The average height was 167.7 cm (range 152– labor and/or abnormalities in epiphyseal closure that are as- 179 cm), and five of the tallest 12 had been castrated before sociated with eunuchism. The remains were examined while the age of 15 years. they were held at the Research Center for Chinese Frontier Physical characteristics also ascribed to court eunuchs in- Archaeology at Jilin University, Jilin Province. clude increase of body weight and voice change to a more Osteological analysis followed protocol outlined in Stan- falsetto pitch. The former may be attributed to the reduction dards for Data Collection of Human Skeletal Remains in physical activity and the ready availability of food as a (Buikstra and Ubelaker, 1994). Owing to limitations in time palace servant. Clinical studies of patients with orchiectomy and available equipment, osteological examination by one show an average weight gain of 10% that is generally local- author (J. Eng) were conducted within a day and consisted of ized in the abdominal and hip region, which increased pa- macroscopic observation and photographic documentation, tient risk of diabetes and cardiovascular disease (Zitzmann without aid of an osteometric board to measure long bone and Nieschlag, 2001; Keating et al., 2006). Moreover, length. The lengths of long bones were later estimated based changes in the lungs and cardiovascular systems from hor- on these photographs (taken with a scale) to approximate monal changes decreased the ability for efficient oxygen up- maximum lengths in centimeters. To determine the accuracy take and blood flow throughout the body (Tsai, 1996), also of these measurements estimated from photographs as com- likely resulting in decreased physical exertion. The retention pared to measurements from an osteometric board, J. Eng of a boyish, high voice stems from low levels of testosterone conducted a pilot study with sample adult long bones during puberty, which prevents male-type laryngeal devel- (n = 11). The same camera was used, replicating the staging opment (Jenkins, 1998). The dramatic decrease in sex ste- of the photographs (i.e. location of the scale by the bones, roids also leads to general loss of elasticity, with wrinkling angle of focus, and camera distance), and the procedure was of the skin, stiffening of joints, and decrease in muscle repeated three times. Maximum length measurements esti- strength (Tsai, 1996). There is no evidence that castration mated from the photographs of these sample bones were significantly affects male lifespan (Wilson and Roehrborn, then compared to similar measurements taken from an os- 1999). teometric board. The average standard deviation was a Studies of human castration have many limitations, not 2.4 mm over-measurement based on photographs, which the least of which is the relative rarity of modern cases and shows that this technique produces relatively little measure- cultural taboos. The long history of institutionalized eu- ment error. nuchism in ancient China provides the potential to study the The remains of the two individuals were well preserved health consequences of castration from a biocultural stand- with fully intact skulls (Figure 2), although not all postcrani- point. This paper describes the skeletal remains of two adult al elements had been recovered. They were clearly discrete eunuchs from pre-modern China, documenting the skeletal individuals, not only indicated by the fact that the excavators and dental changes that may be associated with castration, as had packed them separately and by the different colorations well as skeletal changes reflective of the imperial servant of the remains, attributable to differences in soil conditions lifestyle. for each burial, but also suggested by the body size and chronological age differences between the two skeletons Materials and Methods (discussed below). Observations were made for sex and age determinations, and pathological conditions and activity- In 2003 the skeletal remains of two adult males were ex- related skeletal changes were also recorded. cavated by a team led by Professor Chen Guang (then at the Peking Archaeological Institute). These burials were located Case 1: Burial M1 in the cemetery at Wutasi, Beijing, People’s Republic of Burial M1 is that of a relatively complete skeleton, but has China. This cemetery had over 60 burials, with several some postmortem damage and is missing the right humerus, tombs marked as interments of Ming Dynasty (1368–1644 radius, and ulna, hand and foot bones, as well as several ver- AD) palace eunuchs based on the associated tomb inscrip- tebrae. Age-related changes to pelvic morphology suggest tions (Di and Cao, 2003). Among these eunuch burials, the that this person was in his mid-30s at time of death. Other two studied in this paper include one eunuch we designated features support this age assessment, including the degree of Burial M1, while the second, given the designation of Burial cranial suture closure, tooth wear, and full epiphyseal clo- M2, had a tomb inscription and documentation in his burial sure of all long bones. Sexually dimorphic features of the indicating the eunuch’s name was ‘Huang Zhong,’ a eunuch skull and pelvis are clearly male. On the basicranium, just of the Ming period (C. Guang, personal communication). anterior to the right occipital condyle, the bone is thin and Unfortunately, there are no other publications on the analy- extremely porous, which may be the result of localized post- sis of the excavation that could help verify with certainty the mortem damage, although adjacent bony elements are not eunuch status of all burials labeled as such, but it has been affected (Figure 3). 4 J.T. ENG ET AL. ANTHROPOLOGICAL SCIENCE

Figure 2. Skulls of eunuchs, anterior and left lateral views. (Above) Burial M1. (Below) Burial M2.

The central maxillary incisors and right M1 have lesions from linear enamel hypoplasia, with three arrest lesions on left I1, two on right I1, and one on right M1. Crown formation of these affected teeth occurs at about the same time, so the lesions likely reflect the same stress events. There is evidence of an alveolar abscess around right M3, antemortem tooth loss of left I1, and slight alveolar resorption around left M1 and right M1. The left mandibular condyle displays slight proliferative growth on the superior portion, and is more bul- bous than the right, suggesting slight temporomandibular joint disease on the left, although the condylar fossa exhibits no change. Degenerative changes (pitting, marginal new bone prolif- eration, and/or eburnation) were observed in the following postcranial joints: the left shoulder, left elbow, left hip, and left knee, which has eburnation on the anterior portion of the Figure 3. Right basioccipital of Burial M1 displays porosity. lateral condyle (Figure 4). All of the five preserved vertebrae (one cervical, two thoracic, two lumbar) show moderate development of osteophytes on the margins of their bodies, Vol. 117, 2009 SKELETAL EFFECTS OF CASTRATION 5

to the right tibia (Figure 5). Associated with this healed fracture are the remnants of slight bony buildup of the callus located on the medial aspect, as well as osteoperiostitis in the surrounding area, extending 11.9 cm.

Case 2: Burial M2 (‘Huang Zhong’) This male, named ‘Huang Zhong,’ is also a relatively complete skeleton, missing only the smaller elements (hand and some foot bones), most vertebrae, clavicles, and ribs. Age estimation included assessment of multiple indicators of development. The epiphyses of several long bones were only partially fused or completely unfused, in the case of his distal ulnae. This pattern of fusion would normally indicate that he died in his mid-to-late teens (Table 1). However, castration is known to delay the rate of epiphysis closure in individuals castrated before maturity (Silberberg and Silberberg, 1971), while tooth formation is regulated more strongly by non-sex steroids such as growth hormone (Zhang et al., 1992). In his case, the third molars had fully erupted and have slight wear, and furthermore, the auricular surface re- Figure 4. Distal left femur of Burial M1 with osteoarthritis; note tains slight (youthful) transverse ridges. Thus, based on the the eburnation. dental (and pelvic auricular) information ‘Huang Zhong’ (Burial M2) was likely a young adult, about 20–24 years old and both lumbar vertebrae exhibit a Schmorl’s node on the at death. superior surface. Compared to Burial M1, who was in his 30s at time of The left tibia has a healed fracture just distal to midshaft death, Burial M2 has much less overall wear on teeth, that resulted in a shortening of about 6 mm in length relative consistent with a younger adult age determination. Several

Figure 5. Tibiae of Burial M2. (Above) Fracture just below midshaft of left tibia, seen as superior tibia. (Below) Left tibia, healed fracture. 6 J.T. ENG ET AL. ANTHROPOLOGICAL SCIENCE

Table 1. Age estimation (years) of Burial M2 based on rates of epiphyseal fusiona Bone Beginning Active Complete Proximal humerus 19–20 Distal humerus 16 Proximal ulna 18–19 Distal ulna 17b Proximal radius 18–19 Distal radius 17 Proximal femur (head) 17–18 Distal femur 16–21 Proximal tibia 16–22 Distal tibia 17–18 a From McKern (1970, n.d.), McKern and Stewart (1957) and Ubelaker (1989). Note, estimates from these sources derive from North American male samples. b Element completely unfused. sexually dimorphic features of the cranium (supraorbital Figure 6. Mandibular left incisors of Burial M2 display fusion. To the left is a labial view, and to the right is a lingual view. margin, glabella, mastoid process, and nuchal area) are am- biguous, but the mandible, subpubic region, and greater sciatic notch have a male morphology. The pelvic morphology Table 2. Maximum lengths and stature estimatesa (in cm) of Burial is relatively diagnostic considering his young age (e.g. M1 and Burial M2 Walker, 2005). M1 Side Length Stature ± There are two linear enamel hypoplasia lesions present on the lower right canine. The lower left central and lateral in- Femur Left 49.5 178.995 3.8 Tibia Left 37.4b 170.836b 3.27 cisors (Figure 6) have ‘fused’ crowns and roots (also known Tibia Right 38.0c 172.27c 3.27 as ‘double teeth’ or ‘twinning’), a result of primary develop- mental abnormalities of the teeth whereby two separate M2 tooth buds united at some time during their development Femur Left 53.6 187.81 3.8 (Aguilo et al., 1999). Femur Right 53.5 187.595 3.8 The postcranial elements do not have any marked patho- Tibia Left 44.5 187.805 3.27 logical conditions. There is no skeletal indication as to cause Tibia Right 42.0 181.83 3.27 Fibula Left 42.0 181.36 3.24 of death. As seen in Table 2, the length of Burial M2’s long Humerus Right 38.2 185.03 4.16 bones and associated stature estimates are substantially Radius Left 27.6d longer than those of Burial M1. Using femur length to calcu- Radius Right 23.8d late stature based on Trotter and Gleser (1958), Burial M1 Ulna Left 28.8d d has a stature of approximately 179 cm, while Burial M2 has Ulna Right 25.8 a stature of approximately 188 cm. In addition to his tall stat- a Stature estimate based on Trotter and Glessar (1958) formulae for ure, the bones from Burial M2 were relatively gracile. Mongoloid sample. b Fractured left tibia, which shortened maximum length. c Missing the tip of the medial malleolus and hence this is an esti- Long bone dimensions mate of maximum length. To determine whether body size and long bone lengths of d Unfused distal epiphysis, which precluded stature estimates. the two burials were affected by the documented hormonal effects of castration, these data were compared to measurements taken from other archeological skeletal samples col- regions (except for Burial M1’s comparison to the north- lected by the primary author (J. Eng). These other samples western sample). Of note are the unusually high Z-scores of derive from different time periods and varied sites along Burial M2, the eunuch who may have been castrated in his China’s northern region and from central China in Henan youth. Burial M2 has Z-scores of 5.00 SD or higher com- province (Eng, 2007; Walker and Eng, 2007). Mean mea- pared to all groups except the northwestern sample, where surements of the humerus, femur, and tibia from these sam- the Z-score was still above 2.00. On the other hand, the data ples were compared to the corresponding eunuch long bone on humerofemoral and crural limb proportions (respectively, measurement where available (the ulna and radius of Burial humerus: femur and tibia: femur) of these eunuchs fall with- M2 were not included owing to lack of distal epiphyseal fu- in the range of proportions found in the comparative samples sion) (Table 3). The measures of the eunuch long bones (Table 5). were then transformed to Z-scores to compare with each region (Table 4). Discussion In most instances, the long bones of both eunuchs were longer than the comparative groups (Z-score > 2.00). Both The earliest records of eunuchs in China consist of picto- eunuchs have longer femoral lengths than the comparative graphs on the oracle bones of the Bronze Age Shang Dynasty Vol. 117, 2009 SKELETAL EFFECTS OF CASTRATION 7

Table 3. Long bone measurements (in cm) of eunuch burialsa compared with archeological skeletal datab from China Humerus Femur Tibia Region Time n Avg. SD n Avg. SD n Avg. SD Northwestern 2000 BC–220 AD 35 31.9 1.560 52 45.1 2.412 40 36.8 2.506 North central 1000 BC–1368 AD 3 31.7 0.611 24 43.2 1.968 9 35.2 1.604 Northeastern 3800 BC–1368 AD 88 31.1 1.300 141 43.8 1.867 131 34.7 1.664 Central 476 BC–220 AD 21 30.8 1.396 18 42.8 1.756 16 34.6 1.648 Burial M1 n/a 49.5 38.0c Burial M2 38.2 53.6 43.3 a Measurements of eunuch long bone lengths were averaged if both sides of an element were present. b The comparative Asian data are recalculated from raw data collected by one author (J. Eng) in the northern provinces and from the centrally located Henan province (Eng, 2007; Walker and Eng, 2007). The left and right elements of an individual were averaged, and the average of the total sample region is presented here. c This is an estimate of maximum length of right tibia, missing the tip of the medial malleolus. The left tibia was shortened in life owing to a fracture and was not used here.

Table 4. Z-scores comparisona of eunuch (Burial M1 and Burial Table 5. Comparison of limb proportions of the eunuchs, Burial M1 M2) measurements with archeological skeletal data from China and Burial M2, with archeological skeletal data from China Humerus Femur Tibia Humerus: Femur Tibia: Femur M1 M2 M1 M2 M1 M2 Northwestern 0.71 0.82 North central 0.73 0.82 Northwestern NA 4.04 1.82 3.52 0.48 2.59 Northeastern 0.72 0.79 North central NA 10.64 3.20 5.28 1.75 5.05 Central 0.72 0.81 Northeastern NA 5.46 3.05 5.25 1.98 5.17 Central NA 5.30 3.82 6.15 2.06 5.28 Burial M1 NA 0.77 a Burial M2 0.71 0.81 Italics indicate Z-scores lower than 2.00 SD.

(17th century–1122 BC). Eunuchs were known generally as Chinese historians describe them as venal, greedy syco- huanguan, or taijian, with differentiation made between phants, perhaps in part because of jealousy over eunuchs’ in- those castrated in childhood (‘pure from birth’) and those timate access and influence over the emperor and imperial emasculated in adulthood (‘pure of body’). Males were cas- policy (Yi, 1951; Tsai, 1996). By the end of the Ming Dy- trated to make them ‘safer’ servants for the royal court and nasty, the number of eunuchs nationwide swelled to an esti- also as a form of punishment that was considered second mated 100000, and they were fixtures of nearly every gov- only to death in severity. The primary source for eunuchs ernmental agency. The cases of eunuchs who wielded was prisoners of war taken during campaigns against fron- extreme power were relatively rare, and most palace eu- tier and neighboring hostile groups. Another source came nuchs led routine lives that were restricted and dull (Tsai, from tributes: eunuchism was practically Pan-Asian, and 1996). In the present study, the burial treatment and lack of countries such as Korea and northern Vietnam regularly sent elaborate grave goods associated with the two burials ana- castrati as tribute to China (Tsai, 1996). There was also a do- lyzed here suggest they were ordinary eunuchs, though mestic supply of eunuchs from among the impoverished ‘Huang Zhong’ may have been a eunuch of some note as his peasants, often from poorer southern provinces. In these ar- name was inscribed for his tomb. eas that often suffered famine, the only recourse from starva- tion or sale into slavery was self-castration or castration of a Interpretation of Ming sample son, since imperial eunuchs were assured a life of relative Long bone lengths and height comfort with regular stipend, room and board within palace The skeletal effects of castration are more clearly evident walls. While self-castration was repeatedly prohibited, when in the long-limbed Burial M2, the young adult ‘Huang dynasties such as the Ming sank into financial decay, desper- Zhong.’ As noted previously, there is likely to be some slight ate males, often adults, took this course of action and clam- over-measurement (approximately 2.4 mm) using the photo- ored for work at the palace (Anderson, 1990). graphs to estimate long bone length, but even accounting for Eunuchs served the emperor, his queens and concubines, that difference, the limbs are still relatively long. When ob- and other members of the royal family. During the Ming Dy- serving the humerofemoral and crural limb proportions of nasty there were 24 official agencies charged with various these individuals to comparative ancient Asian samples, aspects of palace maintenance, including care of furniture, there is no marked difference, suggesting there was no un- gardens, kitchens, rituals, cloths, entertainment, and armory, usual disproportion in limb development in these eunuchs. so that their presence permeated all aspects of imperial life On the other hand, the maximum lengths of Burial M2’s hu- (Mitamura, 1970). Beginning in the Ming Dynasty, some merus, femur, and tibia are significantly greater than that of were educated and became imperial secretaries, furthering the comparative samples. Taken together, the longer limb di- eunuch infiltration into the inner circles of the court. Ancient mensions of Burial M2 and the retarded rate of epiphyseal 8 J.T. ENG ET AL. ANTHROPOLOGICAL SCIENCE closure relative to tooth development in this individual sug- Degenerative joint disease gest that Burial M2 was castrated before puberty. However, Burial M1 has signs of degenerative changes at several when compared to the average height (168 cm) of surviving major joints, while Burial M2 has none. Burial M1 was older 20th-century Chinese eunuchs in Wu and Gu’s (1991) study, than Burial M2 at age of death, which may account for the both Burial M1 and Burial M2 are taller than Wu and Gu’s higher observation of degenerative lesions. It should also be criterion for tall stature (height > 170 cm). Thus, with re- noted that castrated boys were often more favored by the spect to the marked differences between the long bone court ladies and treated like young girls, and thus had fewer lengths of these two eunuchs, this may be a factor of: (1) ge- manual labor activities than adult eunuchs (Mitamura, netic heterogeneity and height potential related to their ori- 1970). Thus, if Burial M2 had been castrated in youth, and gins from different populations; or (2) the different ages at died soon afterwards in his early adulthood, this may explain which they were each castrated, with the longer-limbed why it appears he led a relatively less active lifestyle than Burial M2 castrated in pre-puberty before epiphyseal closure Burial M1. of long bones; or (3) individual differences in frailty between them including influences from childhood health and Conclusion nutrition. These factors may not be mutually exclusive. The burials of two Ming palace eunuchs provide new in- Osteoporosis sight into the skeletal changes associated with castration, as Although circumstances prevented weighing and taking well as clues to eunuch life in the Chinese imperial court. radiographs of the skeletons, bones did not feel markedly These eunuchs may have been castrated at different periods lighter than ‘normal’ ancient bones of East and Central of growth and development, as suggested by the long limbs Asian populations that the authors have observed. Burial M1 and partially fused epiphyses of the young adult ‘Huang has some signs suggestive of osteoporosis, though these in- Zhong’ (Burial M2). Arthritic changes and a healed fracture dicators may be due to his more advanced age at death. His in Burial M1 suggests he led a harsher life than experienced skull has slight porosity in the basioccipital area, and al- by Burial M2, who may have had a more pampered life fa- though the localization of the visible porosity does not fully vored by court ladies, if he was castrated in youth. Dental support osteoporosis, there may have been further thinning health suggests both experienced childhood stress, possibly not immediately visible throughout the skull. Burial M1 had owing to poverty or wartime, which may have precipitated few vertebrae to study for possible kyphosis, but there are their entry into the financially secure eunuch service. While degenerative changes in all vertebrae present, which may this report has offered some new information on the lives of also be due to age-related changes or lifestyle. The healed two eunuchs from the Ming Dynasty, it is by no means a fracture on the left tibia may also be (very slight) evidence general life characterization of the health of eunuchs. Future for osteoporosis, though accidental injury is a much more studies of other eunuch burials will shed further light into likely cause. this once secretive, yet fascinating cultural phenomenon.

Oral health Acknowledgments The presence of enamel hypoplasia on both eunuchs indi- cates stress during childhood, which may be expected as This research was funded by support given to J. Eng by most eunuchs of this period were either castrated to escape the following granting agencies during an early visit to poverty, to serve as tributes, or as prisoners of war. Neither China (in 2003) and during the collection of her dissertation individual has dental caries, but Burial M1 has an aveolar data: Fulbright-Hays Doctoral Research Abroad Program abscess, antemortem loss of a tooth, and resorption of the (Award No. p022a040064); Pacific Rim Research Grant, alveolar bone—all indicative of poor dental health. The University of California (Project Reference No. 04TPRRP fused left incisors in Burial M2’s mandible marks a relative- 08-0011); Humanities and Social Sciences Research Grant, ly rare find, at least among people of European ancestry (less UCSB; and the National Science Foundation Graduate Re- than 1%, e.g. Jarvinen et al., 1980; Barac-Furtinovic and search Fellowship. The authors would like to extend deepest Skrinjaric, 1991), although it was found in slightly higher thanks to Professor Chen Guang for permission to analyze frequencies (~4%) in one modern Japanese sample (Yonezu the material. J. Eng would like to thank her dissertation ad- et al., 1997). When present, it is more prevalent in males and visor, Dr Phillip Walker, and several members of the Walker in the anterior teeth, usually of the mandible (Razak and Lab who read and commented on early drafts. Nik-Hussein, 1986; Duncan and Helpin, 1987; Yonezu et al., 1997). Although no environmental factors have been linked to the development of fused teeth, it is more frequent References among siblings (Razak and Nik-Hussein, 1986). Clinical Aguilo L., Gandia J.L., Cibrian R., and Catala M. (1999) Primary problems associated with fused teeth include increased prev- double teeth. A retrospective clinical study of their morpho- alence of caries, supernumerary teeth, impaction of succes- logical characteristics and associated anomalies. International sors, and aplasia (Brook and Winter, 1970; Ravn, 1971), but Journal of Paediatric Dentistry, 3: 175–183. Anderson M.M. (1990) Hidden Power: The Palace Eunuch of such was not the case for Burial M2, whose teeth appeared Imperial China. Prometheus Books, Buffalo. unaffected. Baqi S., Shah S.A., Baig M.A., Mujeeb S.A., and Memon A. (2006) Seroprevalence of HIV, HBV and syphilis and associated risk behaviours in male transvestites (hijras) in Karachi, Vol. 117, 2009 SKELETAL EFFECTS OF CASTRATION 9

Pakistan. Journal of Pakistan Medical Association, 56 (Sup- Age Identification. Technical Report EP-45. Environmental plement 1): S17–S21. Protection Research Division, Quartermaster Research and Barac-Furtinovic V. and Skrinjaric I. (1991) Double teeth in Development Center, US Army, Natick, MA. primary dentition and findings of permanent successor. Acta Mitamura T. (1970) Chinese Eunuchs: The Structure of Intimate Stomatological Croatian, 25: 39–43. Politics. Charles E. Tuttle Company, Rutlands, VT. Brett M., Roberts L., Johnson T., and Wassersug R. (2007) Nanda S. (1985) The hijras of India: cultural and individual dimen- Eunuchs in contemporary society: expectations, conse- sions of an institutionalized third gender role. Journal of quences, and adjustments to castration (part II). The Journal Homosexuality, 11: 35–54. of Sexual Medicine, 4: 946–955. Patel H.G. (1988) Human castration: a study of Hijada (eunuch) Brook A.H. and Winter G.B. (1970) A retrospective study of ‘ger- community of Gujarat in India. Man and Life, 14: 67–76. minated’ and ‘fused’ teeth in children. British Dental Journal, Ravn J.J. (1971) Aplasia, supernumerary teeth and fused teeth in 129: 123–130. the primary dentition. Scandinavian Journal of Dental Buikstra J.E. and Ubelaker D.H. (1994) Standards for Data Collec- Research, 79: 1–6. tion from Human Skeletal Remains. Arkansas Archaeological Razak I.A. and Nik-Hussein N.N. (1986) A retrospective study of Survey, Fayetteville. double teeth in the primary dentition. Annals of the Academy Catalano G., Catalano M.C., and Carroll K.M. (2002) Repetitive of Medicine, 15: 393–396. male genital self-mutilation: a case report and discussion of Read B.E. (1921) The metabolism of the eunuch. Journal of Bio- possible risk factors. Journal of Sex & Marital Therapy, 28: logical Chemistry, 56: 281–283. 27–37. Schenck P.E. and Slob A.K. (1986) Castration, sex steroids, and Daniell H.W. (1997) Osteoporosis after orchiectomy for prostate heterosexual behavior in adult male laboratory-housed stump- cancer. Journal of Urology, 157: 439–444. tailed macaques (Macaca acrtoides). Hormones and Behav- Daniell H.W., Dunn S.R., Ferguson D.W., Lomas G., Niazi Z., and ior, 20: 336–353. Stratte P.T. (2000) Progressive osteoporosis during androgen Scholz P.O. (2001) Eunuchs and Castrati: A Cultural History. deprivation therapy for prostate cancer. Journal of Urology, Marcus Wiener Publishing, New York. 163: 181–186. Scott C.L. and Holmberg T. (2003) Castration of sex offenders: Davis S.J.M. (2000) The effect of castration and age on the devel- prisoners’ rights versus public safety. Journal of the Ameri- opment of the shetland sheep skeleton and a metric compari- can Academy of Psychiatry and the Law, 31: 502–509. son between bones of males, females and castrates. Journal of Seth F. (1996) The death of the last emperor’s last eunuch. New Archaeological Science, 27: 373–390. York Times, Late edn, p. 3. Di C. and Cao B. (2003) Beijing Wutasi faxian guqun [Discovery Silberberg M. and Silberberg R. (1971) Steroid hormones and of ancient Wutasi Monastery in Beijing]. Beijing Qingnian bone. In: Bourne G.H. (ed.), The Biochemistry and Physiol- Bo, Beijing. ogy of Bone, VIII. Academic Press, New York, pp. 401–484. Dixson A.F. (1993) Observations on effects of neonatal castration Smith M.R., Boyce S., Moyneur E., Duh M., Raut M., and upon sexual and aggressive behavior in the male common Brandman J. (2006) Risk of clinical fractures after gonadotropin- marmoset (Callithrix jacchus). American Journal of Primatol- releasing hormone agonist therapy for prostate cancer. Journal ogy, 31: 1–10. of Urology, 175: 136–139. Duncan W.K. and Helpin M.L. (1987) Bilateral fusion and gemi- Stent G.C. (1877) Chinese eunuchs. Journal of North-China nation: a literature analysis and case report. Oral Surgery, Branch of the Royal Asiatic Society, 10: 166. Oral Medicine, Oral Pathology, 64: 82–87. Trotter M. and Gleser G.C. (1958) A re-evaluation of estimation of Eng J.T. (2007) Nomadic Pastoralists and the Chinese Empire: A stature based on measurements of stature taken during life Bioarchaeological Study of China’s Northern Frontier and of long bones after death. American Journal of Physical [Dissertation]. University of California, Santa Barbara, 298 p. Anthropology, 16: 79–123. Flaig T. and Glode L. (2008) Management of the side effects of Tsai S-sH. (1996) The Eunuchs in the Ming Dynasty. State Uni- androgen deprivation therapy in men with prostate cancer. versity of New York Press, New York. Expert Opinion on Pharmacotherapy, 9: 2829–2841. Ubelaker D.H. (1989) Human Skeletal Remains: Excavation, Heller A.L. and Shipley R.A. (1951) Endocrine studies in aging. Analysis, Interpretation, 2nd edn. Taraxacum, Washington. Journal of Clinical Endocrinology and Metabolism, 11: 945– Walker P.L. (2005) Greater sciatic notch morphology: sex, age, 962. and population differences. American Journal of Physical Jarvinen S., Lehtinen L., and Milen A. (1980) Epidemiologic study Anthropology, 127: 385–391. of joined primary teeth in Finnish children. Community Den- Walker P.L. and Eng J.T. (2007) Long bone dimensions as an tal Oral Epidemiology, 8: 201–202. index of socioeconomic change in ancient Asian populations. Jenkins J.S. (1998) The voice of the castrato. Lancet, 351: 1877– American Journal of Physical Anthropology, 132(Supple- 1880. ment 44): 241. Keating N.L., O’Malley A.J., and Smith M.R. (2006) Diabetes and Wassersug R.J. and Johnson T.W. (2007) Modern-day eunuchs. cardiovascular disease during androgen deprivation therapy Perspectives in Biology and Medicine, 50: 544–557. for prostate cancer. Journal of Clinical Oncology, 24: 4448– Wilson J., Aiman J., and MacDonald P. (1980) The pathogenesis 4456. of gynecomastia. Advances in Internal Medicine, 25: 1–32. Mai Y. (1977) Guangzhou Dongshan Ming taijian Wei Chuan mo Wilson J.D. and Roehrborn C. (1999) Commentary: Long-term qingli jianbo [Excavation of the tomb of Wei Chuan, a Ming consequences of castration in men: lessons from the Skoptzy eunuch in Dongshan county, Guangzhou province]. Kaogu, 4: and the eunuchs of the Chinese and Ottoman courts. Journal 280–283. of Clinical Endocrinology and Metabolism, 84: 4324–4331. McKern T.W. (1970) Estimation of skeletal age: from puberty to Wink C.S. and Felts W.J.L. (1980) Effects of castration on the about 30 years of age. In: Stewart T.D. (ed.), Personal Identi- bone structure of male rats: a model of osteoporosis. Calcified fication in Mass Disasters. National Museum of Natural His- Tissue International, 32: 77–82. tory, Washington, DC, pp. 41–56. Wong K.C. and Wu L. (1932) History of Chinese Medicine. The McKern T.W. (n.d.) Osteometry for Physical Anthropologists; Tientsin Press Ltd., Tientsin, China. With a Special Section on Age Identification of Human Skel- Wu C.P. and Gu F-L. (1991) The prostrate in eunuchs. In: Smith etal Remains. Unpublished laboratory manual. P.H. and Pavone-Macaluso M. (eds.), Urological Oncology: McKern T.W. and Stewart T.D. (1957) Skeletal Age Changes in Reconstructive Surgery, Organ Conservation, and Restoration Young American Males: Analysed from the Standpoint of of Function. Wiley-Liss, New York, pp. 249–255. 10 J.T. ENG ET AL. ANTHROPOLOGICAL SCIENCE

Yi D. (1951) Mingdai tewu zhengzhi [Politics of the Secret Police 546. During the Ming Dynasty]. Beijing. Zitzmann M. and Nieschlag E. (2001) Testosterone levels in Yonezu T., Hayashi Y., Sasaki J., and Machida Y. (1997) Preva- healthy men and the relation to behavioral and physical char- lence of congenital dental anomalies of the deciduous denti- acteristics: facts and constructs. European Journal of Endocri- tion in Japanese children. The Bulletin of Tokyo Dental nology, 144: 183–197. College, 38: 27–32. Zumpe D., Bonsall R.W., and Michael R.P. (1992) Some contrast- Zhang C., Young W., Li H., Clayden A., Garcia-Aragon J., and ing effects of surgical and ‘chemical’ castration of the behav- Waters M. (1992) Expression of growth hormone receptor by ior of male cynomolgus monkeys (Macaca fasicularis). immunocytochemistry in rat molar root formation and alveo- American Journal of Primatology, 26: 11–22. lar bone remodeling. Calcified Tissue International, 50: 541–