2016 Volume 124 Issue 2 Pages 117-133
This study examines the health and lifestyle of some of Taiwan’s earliest Neolithic inhabitants using skeletons from the Nankuanli East site (c. 5000–4200 BP) from the Tainan Science Park in southwestern Taiwan. Two indicators of health, cribra orbitalia and adult stature, and evidence of dental staining are reported for the first time. Comparisons between males and females, and between the early Neolithic Taiwan, Iron Age Taiwan (Shihsanhang site), and with skeletal series from surrounding regions of East and Southeast Asia are made. The estimated average stature, using non-ethnic formulae, is 160.1 cm for males and 154.7 cm for females from the Nankuanli East site, statures that are similar to those of indigenous Taiwanese and other prehistoric inhabitants from surrounding regions. Twenty-five percent of the adult individuals from Nankuanli East exhibit cribra orbitalia, a childhood indicator of health that is not significantly different between males and females. Approximately 75% of male and 52% of female teeth exhibit some type of staining, a difference that is statistically significant. Although taphonomic processes, such as the mineral content of the soil, may account for the observed dental staining, the dental profile suggests that the staining may be due to chewing betel (areca) nut. Comparisons with Iron Age skeletons from the Shihsanhang site, once again, do not support the prediction of a decline in health during the transition from early Neolithic to Iron Age Taiwan. Overall, the similarities in health between the early Neolithic and later Iron Age inhabitants of Taiwan indicate similar subsistence economies based on marine and terrestrial resources. Further studies of the skeletons from the Nankuanli East site and other sites in the Tainan Science Park will improve our understanding of the health and lifestyle of Taiwan prehistoric inhabitants.
A previous investigation that examined several indicators of health (e.g. linear enamel hypoplasia and dental pathology) in skeletons from the Nankuanli East (NKLE) site in southwestern Taiwan provided some of the first glimpses of the health and lifestyle of Taiwan’s earliest farmers (Pietrusewsky et al., 2013). While the early Neolithic inhabitants of Taiwan enjoyed relatively good dental health, an elevated frequency of linear enamel hypoplasia (LEH) suggested they may also have experienced more nonspecific physiological stress, including the possibility of infectious diseases and/or metabolic disorders, during early childhood. Because the culture associated with the skeletons from NKLE is believed to have been introduced from outside Taiwan, studies of the early inhabitants also provide insights into the challenges of adapting to a new environment. Initial comparisons of the early Neolithic and later Iron Age skeletons from the Shihsanhang (SSH) site in northwestern Taiwan suggested only two indicators of dental health (alveolar resorption and dental calculus) were significantly different, consistent with a transition to more advanced food-processing techniques associated with the Iron Age cultures. Further, the reduced amount of dental attrition observed in the Iron Age skeletons compared to those of the early Neolithic, although not statistically significant, was expected given an increased reliance on softer processed foods and the use of metal tools during the Iron Age.
This earlier study also found that, with the exception of higher frequencies of dental attrition in males, there were few significant differences between sexes for these indicators of health in the NKLE and SSH skeletons. Overall there were overwhelming similarities in dental health between the Neolithic and Iron Age inhabitants of Taiwan, suggesting that, despite being separated by several thousand years, these prehistoric people shared very similar subsistence bases, including farming, fishing, and the hunting and gathering of marine and terrestrial resources. Regional comparisons indicated that the frequencies of several indicators of dental infection in the Neolithic and Iron Age skeletons were among the lowest reported, suggesting that Taiwan’s prehistoric inhabitants enjoyed good dental health. Explanation for the observed differences and similarities focused on possible differences in subsistence economies, diet, hygiene, and cultural behaviors.
In this paper we expand our examination of the health of Taiwan’s earliest Neolithic inhabitants. We report, for the first time, two additional indicators of physiological health, adult stature and cribra orbitalia (CO). We also document dental staining observed in the NKLE skeletons and evaluate whether the observed staining is due to chewing areca (betel) nut or some other process and how this correlates with the dental health indicators of these early inhabitants. Given the general observation that differences in diet may explain differences in male and female health, we again make comparisons between the male and female NKLE skeletons. Further, we investigate the biocultural implications of changes in subsistence in the earliest Neolithic and later Iron Age Taiwan and whether differences in health support, or refute, the observation that health generally deteriorates with the transition from Neolithic to Iron Age. Finally, limited comparisons of the NKLE skeletons and those from surrounding regions are made that provide broader regional context for these data.
The NKLE site, located on the flood plains of Shanhua District, Tainan City, in southwestern Taiwan (Figure 1), represents one of the earliest Neolithic sites from Taiwan. Approximately 82 burials and an impressive archaeological record that includes pottery, ornaments, shell reaping knives, fishing implements, ground adzes and projectile points, an extensive inventory of faunal remains (dog, deer, wild boar, fish, etc.), and the presence of foxtail millet are documented for this site (Tsang, 2005; Li, 2013).
Map showing the approximate locations of the archaeological skeletal series examined in this study and the distribution of some of Taiwan’s present-day indigenous groups.
Twenty-four of the most complete adult burials representing 14 males and 10 females are used in the present study (Table 1). The majority of the individuals (n = 18) died as young (20–35 years of age) adults and six survived to middle age (35–50 years).
| Age/sex | NKLE | SSH | ||||
|---|---|---|---|---|---|---|
| Male1 | Female2 | Total | Male3 | Female4 | Total | |
| Young adult (20–35 yrs.) | 9 | 9 | 18 | 13 | 5 | 18 |
| Middle-aged (35–50 yrs.) | 5 | 1 | 6 | 2 | 3 | 5 |
| Total | 14 | 10 | 24 | 15 | 8 | 23 |
The SSH site (1800–500 years BP) is located on the northwestern coast of Taiwan near the mouth of the Danshui River (Yang, 1961; Tsang, 2000). More than 200 well-preserved human skeletons, many in the flexed position with heads facing southwest, were excavated from this site. Twenty-three of the best-preserved and most complete adult skeletons (Pietrusewsky and Tsang, 2003) are used in the present study. Eighteen individuals died as young adults (20–35 years of age) and five were middle-aged adults (35–50 years) at the time of their death (Table 1).
Comparative skeletal assemblagesThe skeletal series from China, Japan, and Southeast Asia, used for comparisons in this paper, are summarized in Table 2 and Figure 1.
| Site | Location | Approx. dates (years BP) | References |
|---|---|---|---|
| Taiwan | |||
| Nankuanli East | SW Taiwan | 5000–4200 | Tsang (2005) |
| Shihsanhang | NW Taiwan | 1500–1000 | Tsang (2000) |
| China | |||
| Jiahu | China | 9000–7800 | Pechenkina et al. (2013) |
| Early Yangshao1 | China | 6900–6000 | Pechenkina et al. (2013) |
| Middle Yangshao2 | China | 6000–5000 | Pechenkina et al. (2013) |
| Anyang | N China | 3385–3112 | Li (1977) |
| Japan | |||
| Middle/Late Jomon | Japan | 5000–3000 | Temple (2007), Temple and Larsen (2013) |
| Late to Final Jomon | Japan | 3000–2500 | Temple (2007), Temple and Larsen (2013) |
| Yayoi | Japan | 2500–1700 | Temple and Larsen (2007) |
| Thailand | |||
| Early Non Nok Tha3 | NE Thailand | 5000–4000 | Douglas and Pietrusewsky (2007) |
| Khok Phanom Di | C Thailand | 4000–3500 | Tayles (1999) |
| Early Ban Chiang | NE Thailand | 4100–2900 | Douglas (1996), Douglas and Pietrusewsky (2007), Pietrusewsky and Douglas (2002) |
| Ban Lum Khao | NE Thailand | 3400–2500 | Domett (2001, 2004), Domett and Tayles (2006) |
| Late Non Nok Tha4 | NE Thailand | 4100–1800 | Douglas and Pietrusewsky (2007) |
| Late Ban Chiang | NE Thailand | 2900–1800 | Douglas (1996), Douglas and Pietrusewsky (2007), Pietrusewsky and Douglas (2002) |
| Ban Na Di | NE Thailand | 2600–2400 | Pietrusewsky and Douglas (2002) |
| Nong Nor | C Thailand | 3100–2500 | Tayles et al. (1998) |
| Noen U-Loke | NE Thailand | 2300–1600 | Domett and Tayles (2006), Higham (2002) |
| Vietnam | |||
| Con Co Ngua | N Vietnam | 6000–5500 | Oxenham (2000), Oxenham et al. (2002a), Oxenham et al. (2006) |
| Man Bac | N Vietnam | 3500–3800 | Oxenham et al. (2011) |
| Nui Nap | N Vietnam | 3000–1700 | Oxenham et al. (2002b) |
| Metal Age/Dong Son Period | N Vietnam | 2500–1700 | Oxenham (2000), Oxenham et al. (2002a), Oxenham et al. (2006) |
| Cambodia | |||
| Vat Komnou | S Cambodia | 2500–1500 | Ikehara-Quebral (2010), Pietrusewsky and Ikehara-Quebral (2006) |
| Phum Snay | NW Cambodia | 2350–1800 | O’Reilly et al. (2004), Domett and O’Reilly (2009) |
The methods for determining sex and estimating age-at-death are discussed in Pietrusewsky et al. (2013). Previously investigated skeletal and dental indicators of health attributable to non-specific systemic stress (e.g. LEH) and specific stress (e.g. dental pathology) are presented in Pietrusewsky et al. (2013). Two new non-specific indicators of health (adult stature and CO) and evidence of tooth staining are introduced here. Fisher’s exact test (FET) was used to test for significant differences in the discontinuous data and Student’s t-test for continuous data (Thomas, 1986).
StatureA substantial literature, based largely on studies of living and recent historical populations, demonstrates a strong correlation between terminal adult height and the suppression of growth in childhood due to poor environmental conditions such as malnutrition and disease (Larsen, 2015). Individuals who are adequately nourished tend to achieve their growth potential while those who are poorly nourished do not. Generally, if the stress experienced during childhood is relieved, the child will resume growing and likely experience a ‘catch-up’ period of growth. Regardless of this mechanism, their terminal adult stature will still likely exhibit evidence of the stress experienced during development. Likewise, although not universal, differences in adult stature have also been correlated with changes in subsistence patterns, especially as seen in the transition from foraging to farming societies in the New World (Larsen, 2006). Although generally not as important as environmental influences, genetic factors can also affect adult stature (e.g. Bogin, 2001).
Given the manner in which the skeletons were excavated (see Pietrusewsky et al., 2013: 5), the lengths of the major long limb bones were recorded in situ using a measuring tape and sliding caliper while the bones were embedded in an artificial matrix that preserved their natural state of articulation (Figure 2, Figure 3). Although this method may not be as accurate as recording measurements in bones freed from the surrounding matrix using an osteometric board, this was the only option available at the time of our examination. Although the bones appear to be intact and solid, many were fragmented. Freeing them from the surrounding artificial matrix, applied at the time of excavation, would result in severe damage. The maximum length of the femur, the distance from the most superior point on the head of the femur to the most inferior point on the distal condyles, was used to estimate stature for adult males and females. Because of the lack of population-specific stature formulae for estimating adult stature in prehistoric skeletons from Taiwan, estimates using several different stature formulae for modern Asian (e.g. Trotter and Gleser, 1958; Sangvichien et al., 1985, n.d.; Shao, 1989) and non-ethnic formulae (Sjøvold, 1990) were used. Female stature was estimated using comparable formulae (Pearson, 1899; Sangvichien et al., 1985; Sjøvold, 1990).
Inferior view of burials F3-B3, a middle-aged male, and F3-B4, an adolescent ?male, from the Nankuanli East site. In addition to associated ceramic and shell artifacts, a bone spear point is embedded in the right scapula of F3-B3.
The skeleton of burial E3-B5, a young adult male from Nankuanli East, showing intact long limb bones, which were measured to estimate stature.
CO lesions, and the related condition known as porotic hyperostosis, result from bone marrow expansion of the cranial vault bones that produces sieve-like lesions in the orbital roofs (Figure 4) in response to increased red blood cell production (Stuart-Macadam, 1985, 1989, 1991). Both conditions have been linked to iron deficiency anemia as a result of nutritional deficiencies, especially during early childhood, infectious diseases, gastrointestinal parasitic infections leading to infant diarrheal disease, as well as hereditary hemolytic anemias (Stuart-Macadam, 1985, 1991; Walker et al., 2009; Larsen, 2015).
Lesions attributed to cribra orbitalia are visible in the superior orbits of burial F5-B40, a young adult male from Nankuanli East.
Active, unhealed, CO is more commonly observed in sub-adults, while in adults the condition is usually remodeled and healed. CO, healed or active, was scored in the left and right orbits, and per individual.
Dental stainingChewing the seed of the areca palm (Areca catechu L.)— commonly but incorrectly referred to as ‘betel nut’—typically accompanied with pepper leaf (Piper betle L.) and slaked lime, is a common cultural behavior documented among many indigenous peoples of South Asia, Southeast Asia, Taiwan, and the Western Pacific (Zumbroich, 2007–2008). Teeth with reddish-brown staining have been offered as evidence for the regular use of the areca nut with lime provided taphonomic and other components of the diet can be eliminated as unlikely causes (Hanson and Butler, 1997: 280; Zumbroich, 2007–2008: 98–99). The staining observed in the NKLE teeth is usually light to dark brown/reddish brown in color.
There is now a substantial epidemiological literature demonstrating a link between chewing betel nut and cariostasis (e.g. Howden, 1984; Möller et al., 2009; Schamschula et al., 1977) and other oral-dental pathologies including periodontal disease and dental calculus in living people (e.g. Trivedy et al., 1997, 2002; Chatrchaiwiwatana, 2006; Anand et al., 2014), and a documented increase in dental attrition (International Agency for Research on Cancer, 2004; Kumar et al., 2004). We report the frequency of stained teeth in skeletal series on a ‘per tooth’ and ‘per individual’ basis.
Estimates of living stature based on maximum femoral lengths, for 13 males and 8 females from the NKLE site using Asian, Chinese, and non-ethnic formulae are presented in Table 3 and Table 4, respectively. Using estimates based on non-ethnic formulae (Sjøvold, 1990), male statures range from 154.2 to 167.8 cm (mean = 160.1 cm) and female statures from 146.1 to 162.4 cm (mean = 154.7 cm). Using formulae for Thai, the average male stature is 161.0 cm and the average female stature is 152.9 cm. Comparisons of these stature estimates with indigenous Taiwanese and other skeletal series within and outside Taiwan are discussed later.
| Burial | Stature (cm) | Regression formula | Side/bone length (mm) | |
|---|---|---|---|---|
| ID | Age1 | |||
| F5B25 | YA | 167.8 ± 4.49 | Sjøvold (1990)2 | L femur (450) |
| 169.32 ± 3.8 | Trotter and Gleser (1958)3 | |||
| 167.86 ± 3.48 | Shao (1989)4 | |||
| 165.9 ± 5.4 | Sangvichien et al. (1985, n.d.)5 | |||
| F5B40 | YA | 154.2 ± 4.49 | Sjøvold (1990) | L femur (400) |
| 158.57 ± 3.8 | Trotter and Gleser (1958) | |||
| 156.36 ± 3.48 | Shao (1989) | |||
| 157.29 ± 5.4 | Sangvichien et al. (1985, n.d.) | |||
| F5B39 | MA | 159.68 ± 4.49 | Sjøvold (1990) | L femur (420) |
| 162.87 ± 3.12 | Trotter and Gleser (1958) | |||
| 160.87 ± 3.8 | Shao (1989) | |||
| 160.75 ± 5.4 | Sangvichien et al. (1985, n.d.) | |||
| F5B29 | MA+ | 162.39 ± 4.49 | Sjøvold (1990) | L femur (430) |
| 165.02 ± 3.8 | Trotter and Gleser (1958) | |||
| 163.23 ± 3.12 | Shao (1989) | |||
| 162.48 ± 5.4 | Sangvichien et al. (1985, n.d.) | |||
| F5B27 | MA | 166.46 ± 4.49 | Sjøvold (1990) | L femur (445) |
| 168.25 ± 3.8 | Trotter and Gleser (1958) | |||
| 166.77 ± 3.12 | Shao (1989) | |||
| 165.07 ± 5.4 | Sangvichien et al. (1985, n.d.) | |||
| F5B19 | MA | 165.10 ± 4.49 | Sjøvold (1990) | L femur (440) |
| 167.17 ± 3.8 | Trotter and Gleser (1958) | |||
| 165.69 ± 3.12 | Shao (1989) | |||
| 164.2 ± 5.4 | Sangvichien et al. (1985, n.d.) | |||
| F5B11 | YA+ | 157.0 ± 4.49 | Sjøvold (1990) | L femur (410) |
| 160.72 ± 3.8 | Trotter and Gleser (1958) | |||
| 158.66 ± 3.48 | Shao (1989) | |||
| 159.02 ± 5.4 | Sangvichien et al. (1985, n.d.) | |||
| F3B3 | YA | 155.6 ± 4.49 | Sjøvold (1990) | L femur (405) |
| 159.65 ± 3.8 | Trotter and Gleser (1958) | |||
| 157.51 ± 3.48 | Shao (1989) | |||
| 158.15 ± 5.4 | Sangvichien et al. (1985, n.d.) | |||
| E4B4 | YA | 158.33 ± 4.49 | Sjøvold (1990) | L femur (415) |
| 161.8 ± 3.8 | Trotter and Gleser (1958) | |||
| 159.81 ± 3.48 | Shao (1989) | |||
| 159.88 ± 5.4 | Sangvichien et al. (1985, n.d.) | |||
| E3B5 | YA | 157.0 ± 4.49 | Sjøvold (1990) | L femur (410) |
| 160.72 ± 3.8 | Trotter and Gleser (1958) | |||
| 158.66 ± 3.48 | Shao (1989) | |||
| 159.02 ± 5.4 | Sangvichien et al. (1985, n.d.) | |||
| E3B2 | YA+ | 155.62 ± 4.49 | Sjøvold (1990) | L femur (405) |
| 159.65 ± 3.8 | Trotter and Gleser (1958) | |||
| 157.51 ± 3.48 | Shao (1989) | |||
| 158.15 ± 5.4 | Sangvichien et al. (1985, n.d.) | |||
| E3B1 | YA | 156.4 ± 4.49 | Sjøvold (1990) | L femur (408) |
| 160.29 ± 3.8 | Trotter and Gleser (1958) | |||
| 158.2 ± 3.48 | Shao (1989) | |||
| 158.67 ± 5.4 | Sangvichien et al. (1985, n.d.) | |||
| D3B2 | YA+ | 165.10 ± 4.49 | Sjøvold (1990) | L femur (440) |
| 167.17 ± 3.8 | Trotter and Gleser (1958) | |||
| 165.56 ± 3.48 | Shao (1989) | |||
| 164.2 ± 5.4 | Sangvichien et al. (1985, n.d.) | |||
| Burial | Stature (cm) | Regression formula2 | Side/bone length (mm) | |
|---|---|---|---|---|
| ID. | Age1 | |||
| D3B3 | YA | 162.39 ± 4.49 | Sjøvold (1990) | L femur (430) |
| 160.25 ± 3.0 | Sangvichien et al. (1985, n.d.) | |||
| 156.48 | Pearson (1899) | |||
| D4B1 | YA | 149.65 ± 4.49 | Sjøvold (1990) | R femur (383) |
| 148.11 ± 3.0 | Sangvichien et al. (1985, n.d.) | |||
| 147.34 | Pearson (1899) | |||
| F3B1 | MA | 156.97 ± 4.49 | Sjøvold (1990) | L femur (410) |
| 155.08 ± 3.0 | Sangvichien et al. (1985, n.d.) | |||
| 152.59 | Pearson (1899) | |||
| E6B1 | YA | 156.97 ± 4.49 | Sjøvold (1990) | L femur (410) |
| 155.08 ± 3.0 | Sangvichien et al. (1985, n.d.) | |||
| 152.59 | Pearson (1899) | |||
| F5B33 | YA+ | 152.91 ± 4.49 | Sjøvold (1990) | L femur (395) |
| 151.21 ± 3.0 | Sangvichien et al. (1985, n.d.) | |||
| 149.67 | Pearson (1899) | |||
| F5B32 | YA | 146.13 ± 4.49 | Sjøvold (1990) | L femur (370) |
| 144.76 ± 3.0 | Sangvichien et al. (1985, n.d.) | |||
| 144.81 | Pearson (1899) | |||
| F5B20 | YA | 161.04 ± 4.49 | Sjøvold (1990) | L femur (425) |
| 159 ± 3.0 | Sangvichien et al. (1985, n.d.) | |||
| 155.51 | Pearson (1899) | |||
| F5B9 | YA | 151.55 ± 4.49 | Sjøvold (1990) | L femur (390) |
| 149.91 ± 3.0 | Sangvichien et al. (1985, n.d.) | |||
| 148.7 | Pearson (1899) | |||
Based on the total number of orbits available, there is a slightly higher frequency of CO in females (40.0%) than males (25.0%), although this difference is not statistically significant. The combined frequency of CO in the NKLE skeletons is 30.8% (Table 5).
| Per orbit | |||
| A/O1 | % | P value2 | |
| Male | 4/16 | 25.0 | |
| Female | 4/10 | 40.0 | 0.6645 |
| Total | 8/26 | 30.8 | |
| Per individual | |||
| A/O | % | P value | |
| Male | 2/10 | 20.0 | |
| Female | 2/6 | 33.3 | 0.6044 |
| Total | 4/16 | 25.0 | |
The frequency of dental staining observed in the NKLE skeletons is recorded by tooth, right and left sides combined for males and females, and by individual (Table 6). Staining is more frequently observed in male teeth (75.5%) than in female teeth (52.3%), a difference that is statistically significant (two-tailed P < 0.0001 for FET test). However, the frequency of tooth staining by individual is not statistically significant (two-tailed P = 1.0000 for FET test). In males, the mandibular teeth are more frequently stained (82.7%) than the maxillary teeth (66.0%). This difference is not replicated in the female dentitions but the number of teeth available is considerably reduced.
| Tooth | Male (n1 = 14) | Female (n1 = 9) | Total (n1 = 23) | |||
|---|---|---|---|---|---|---|
| A/O2 | % | A/O2 | % | A/O2 | % | |
| Maxillary teeth | ||||||
| M3 | 12/21 | 57.1 | 5/9 | 55.6 | 17/30 | 56.7 |
| M2 | 17/26 | 65.4 | 7/11 | 63.6 | 24/37 | 64.9 |
| M1 | 19/24 | 79.2 | 7/12 | 58.3 | 26/36 | 72.2 |
| P4 | 16/26 | 61.5 | 7/11 | 63.6 | 23/37 | 62.2 |
| P3 | 13/24 | 54.2 | 6/9 | 66.7 | 19/33 | 57.6 |
| C | 3/3 | 100.0 | — | — | 3/3 | 100.0 |
| I2 | — | — | — | — | — | — |
| I1 | 17/23 | 73.9 | 15/17 | 88.2 | 32/40 | 80.0 |
| Total maxillary teeth | 97/147 | 66.0 | 47/69 | 71.0 | 144/216 | 66.7 |
| Mandibular teeth | ||||||
| M3 | 14/20 | 70.0 | 7/14 | 50.0 | 21/34 | 61.8 |
| M2 | 21/26 | 80.8 | 9/16 | 56.3 | 30/42 | 71.4 |
| M1 | 21/25 | 84.0 | 8/16 | 50.0 | 29/41 | 70.7 |
| P4 | 21/26 | 80.8 | 6/16 | 32.5 | 27/42 | 64.3 |
| P3 | 20/24 | 83.3 | 5/15 | 33.3 | 25/39 | 64.1 |
| C | 22/26 | 84.6 | 8/17 | 47.1 | 30/43 | 69.8 |
| I2 | 21/24 | 87.5 | 6/16 | 37.5 | 27/40 | 67.5 |
| I1 | 22/25 | 88.0 | 6/16 | 37.5 | 28/41 | 68.3 |
| Total mandibular teeth | 162/196 | 82.7 | 55/126 | 43.7 | 217/322 | 67.4 |
| Total maxillary and mandibular | ||||||
| Molars | 104/147 | 70.7 | 43/78 | 55.1 | 147/220 | 66.8 |
| Premolars | 70/100 | 70.0 | 24/51 | 47.1 | 94/151 | 62.3 |
| Canines | 25/29 | 86.2 | 8/17 | 47.1 | 33/46 | 71.7 |
| Incisors | 60/72 | 83.3 | 27/49 | 35.1 | 87/121 | 71.9 |
| Total (all teeth) | 259/343 | 75.5 | 102/195 | 52.3 | 361/538 | 67.1 |
| By individual | 13/14 | 92.9 | 8/9 | 88.9 | 21/23 | 91.3 |
Two indicators of health, adult stature and CO, allow us to expand on an earlier examination of health of the earliest Neolithic inhabitants of Taiwan (Pietrusewsky et al., 2013), an examination that focused on dental indicators of health. The discussion is organized into three sections. First, we examine differences between male and female skeletons from the NKLE site. Second, comparisons between the earliest Neolithic inhabitants of Taiwan (NKLE) and the Iron Age skeletons from the SSH site are examined. Lastly, we compare the early inhabitants of Taiwan with more recent indigenous Taiwanese, and archaeological skeletal series from surrounding regions.
Sex differences in NKLE skeletons StatureGiven that differences in the timing and velocity of the adolescent growth spurt in humans result in sexual dimorphism, it is not surprising that the average male stature in NKLE males (depending on the stature estimation formula used) is 4–8 cm greater than the average NKLE female stature. Comparisons with Iron Age and recent indigenous Taiwanese, and other skeletal series from surrounding regions, are discussed later.
Cribra orbitaliaThe combined frequency of CO in the NKLE adults is relatively low (25.0%) and the difference between male and female frequencies for this indicator is not statistically significant (Table 5). Similarly, as reported in Pietrusewsky et al. (2013), no significant sex difference was found for LEH (males = 57.1%, females = 44.1%), another indicator of early childhood stress. However, the relatively high frequency of LEH (51.3%) in the NKLE skeletons suggests the existence of childhood physiological stress, conditions associated with a variety of anemic conditions, and other disorders such as rickets and scurvy as well as nutritional deficiencies associated with parasitic infections and diarrheal disease (Walker, 1986; Walker et al., 2009). Despite possible interruption to growth during childhood, the stature of the NKLE adults does not appear to have been adversely affected. This discussion will be broadened when comparisons with Iron Age skeletons from Taiwan and other skeletal series from surrounding regions are introduced.
Tooth stainingThe majority of the teeth in the NKLE skeletons exhibit variable reddish-brown discoloration. As reported for other Neolithic skeletons from Taiwan (e.g. Lien, 1989), we previously reported that the dental staining observed in the NKLE skeletons was likely the result of chewing Areca catechu L. (betel) nut (Pietrusewsky et al., 2013, 2014). However, upon closer inspection, the staining of the tooth enamel is neither uniform nor entirely restricted to the enamel surfaces of the teeth (Figure 5, Figure 6). Often, the most intense staining occurs where the dentin is exposed, in the pulp, and at sites where carious lesions are observed. The staining is also visible on the root portions of the teeth and along postmortem cracks and fissures. Large areas of the tooth enamel are not stained. While dental staining of a similar nature is observed in modern betel nut chewers (e.g. Giri et al., 2014), the bones of the NKLE skeletons were observed to exhibit a similar reddish-brown stain, suggesting that contact with various minerals and organic substances in the soil may be responsible for at some of the observed staining. Reddish-brown staining observed in skeletal remains has been linked to well-drained soils, tannins in the soil solution, and iron oxides (Schultz et al., 2003; Dupras and Schultz, 2013: 324; Pokines and Baker, 2013).
Burial F5-B11, an adult male, from Nankuanli East: (a) upper portion of burial showing matching color of bones and teeth; (b) discoloration of the maxillary central incisors including staining of postmortem vertical cracks in the tooth enamel; (c) mandibular teeth showing staining of the occlusal surfaces, especially of the dentin.
Burial D3-B3, adult female, from Nankuanli East: (a) upper portion of burial showing reddish brown color of bones and teeth; (b) discoloration of the maxillary teeth; (c) showing stained mandibular teeth.
As reported earlier, the low frequencies of dental pathology observed in the NKLE skeletons indicate good dental health for these early inhabitants of Taiwan (Pietrusewsky et al., 2013). Overall, with the exception of tooth staining, dental attrition, and dental calculus, no statistically significant differences were observed between male and female frequencies of dental pathology.
Although contradictory evidence exists (e.g. Williams et al., 1996), numerous researchers have reported that regular chewing of betel nut has a cariostatic effect (e.g. Howden, 1984; Kelley et al., 1991; Larsen et al., 1991; Möller et al., 2009; Schamschula et al., 1977). Other studies have linked betel nut chewing with periodontal disease (alveolar resorption and calculus formation) (e.g. Amarasena et al., 2002; Anand et al., 2014; Chatrchaiwiwatana, 2006; Trivedy et al., 2002). As previously reported (Pietrusewsky et al., 2013), a significantly higher frequency of advanced dental attrition (males: 11.7%, females: 2.2%) and lower rates of caries (males: 1.7%, females: 2.1%) and antemortem tooth loss (AMTL) (males: 0.2%, females: 0.4%) observed in the NKLE males correlates with significantly higher frequency of staining observed in males. This correlation strengthens the argument that the staining observed in the NKLE teeth may be due to chewing betel (areca) nut. Further, compared to females, the slightly lower rates of dental caries and AMTL observed in males, although not statistically significant, compare favorably with the known cariostatic effects of chewing betel nut.
Several studies have shown that betel nut chewers experience greater loss of periodontal attachment and increased calculus formation than those who do not chew betel nut (e.g. Ånerud et al., 1991; Amarasena et al., 2002; Trivedy et al., 2002; Chatrchaiwiwatana, 2006; Parmar et al., 2008; Anand et al., 2014). As previously reported (Pietrusewsky et al., 2013), although not statistically significant, the frequency of advanced alveolar resorption in males (12.4%) was greater than the frequency observed in the NKLE females (6.1%). Unexpectedly, the frequency of advanced dental calculus was observed to be significantly greater in females (13.9%) compared to males (4.4%). However, as noted by others (e.g. Anand et al., 2014), confounding variables such as oral hygiene, dietary factors, general health, and dental status may have a significant influence on periodontal status, which may explain the lower frequency of dental calculus in the NKLE males. Also, the use of high-pressure air guns to clean the NKLE skeletons may have inadvertently removed some of the calculus deposits in these remains.
Comparisons of Early Neolithic and Iron Age skeletons from TaiwanBased on previous research in bioarchaeology (Cohen and Armelagos, 1984; Larsen, 2006; Cohen and Crane-Kramer, 2007; Temple and Larsen, 2013), it is predicted that the subsistence economies of Iron Age Taiwan will be associated with an increase in systematic stress and certain dental indicators of health when compared to the early Neolithic communities in Taiwan. Thus, frequencies of childhood stress (CO and LEH) and dental pathology are expected to increase from early Neolithic to later Iron Age Taiwan. Likewise, is it expected that adult stature should decrease with the transition to more intensified agriculture.
Contrary to expectations, average adult statures for the Iron Age SSH skeletons exceed, by approximately 5 cm, the adult statures for the Neolithic NKLE skeletons, a difference that is statistically significant for males and not quite statistically significant for females (Table 7). As expected, CO is slightly greater in the Iron Age series than in the NKLE skeletons, although the difference is not significant (Table 8). However, as demonstrated earlier and contrary to expectations, the frequency of LEH is significantly lower in the Iron Age skeletons (37.1%) (Pietrusewsky et al., 2013) and the expected increase in dental pathology (AMTL, caries, and alveolar defects) in the Iron Age skeletons was not confirmed. Only the frequencies of alveolar resorption and dental calculus, indicators of periodontal disease, increased in the Iron Age series as expected (Pietrusewsky et al., 2013). The differences in these two indicators of periodontal disease may also reflect differences in oral hygiene, diet, and the prophylactic effects of chewing betel nut between the two series. The overall frequency of tooth staining (based on the number of teeth) in the SSH remains is 2.8% (Table 9), with no difference between males and females.
| Sample | A/O | % | P-value1 | References |
|---|---|---|---|---|
| Taiwan | ||||
| Nankuanli East | 4/16 | 25.0 | this study | |
| Shihsanhang | 8/18 | 44.4 | 0.2966 | Pietrusewsky and Tsang (2003) |
| China | ||||
| Jiahu | 7/28 | 25.0 | 1.0000 | Pechenkina et al. (2013) |
| Early Yangshao | 6/66 | 9.1 | 0.0985 | Pechenkina et al. (2013) |
| Middle Yangshao | 13/95 | 13.7 | 0.2646 | Pechenkina et al. (2013) |
| Thailand | ||||
| Early Non Nok Tha | 6/31 | 19.4 | 0.7156 | Douglas and Pietrusewsky (2007) |
| Early Ban Chiang | 2/23 | 8.7 | 0.2053 | Douglas and Pietrusewsky (2007) Pietrusewsky and Douglas (2002) |
| Late Non Nok Tha | 1/27 | 3.7 | 0.0556 | Douglas and Pietrusewsky (2007) |
| Khok Phanom Di | 31/57 | 54.4 | 0.0492 | Tayles (1999) |
| Late Ban Chiang | 4/10 | 40.0 | 0.6645 | Douglas and Pietrusewsky (2007) Pietrusewsky and Douglas (2002) |
| Vietnam | ||||
| Con Co Ngua | 46/57 | 80.7 | 0.0008 | Oxenham (2000) |
| Man Bac | 19/26 | 73.1 | 0.0039 | Oxenham et al. (2011) |
| Metal Age/Dong Song | 39/53 | 73.6 | 0.0035 | Oxenham (2000) |
| Cambodia | ||||
| Vat Komnou | 5/12 | 41.7 | 0.4319 | Ikehara-Quebral (2010) |
| Japan | ||||
| Jomon2 | 18/209 | 8.6 | 0.0569 | Temple (2010) |
| Yayoi3 | 8/91 | 8.8 | 0.0790 | Temple (2010) |
| Site | Male | Female | Total | P-value1 | References | |||
|---|---|---|---|---|---|---|---|---|
| A/O2 | % | A/O | % | A/O | % | |||
| Nankuanli East | 259/343 | 75.5 | 102/195 | 52.3 | 361/538 | 67.1 | this study | |
| Ban Chiang, Thailand3 | 47/569 | 8.3 | 31/464 | 6.7 | 78/1033 | 7.6 | <0.0001 | Pietrusewsky and Douglas (2002) |
| Vat Komnou, Cambodia | 131/174 | 75.3 | 98/131 | 74.8 | 229/305 | 75.1 | 0.0155 | Pietrusewsky and Ikehara-Quebral (2006) |
| Nui Nap, Vietnam | 140/275 | 50.9 | 120/218 | 55.0 | 281/558 | 50.4 | <0.0001 | Oxenham et al. (2002b) |
| Shihsanhang, Taiwan | 10/360 | 2.9 | 4/145 | 2.8 | 14/505 | 2.8 | <0.0001 | this study |
Further, as predicted, the frequency of advanced dental attrition is lower in the Iron Age skeletons, a change that is consistent with an increased reliance on softer, more processed foods and the use of metal tools in the Iron Age (Pietrusewsky et al., 2013).
Overall, very few differences in the frequencies of systemic stress indicators between the Neolithic and Iron Age Taiwan series were observed, suggesting there is no temporal increase in reliance on cereals or change in emphasis between millet and rice; rather, the subsistence base remained very broad, including marine, riverine and terrestrial resources. If anything, there was a decline in childhood stress over time that may be linked to improvements in food-processing techniques, better weaning foods, and less contamination of food/water in later Iron Age Taiwan (Pietrusewsky et al., 2013).
Except for mostly minor differences, which may be attributable to differences in diet and oral hygiene, the dental health of the people represented by the Neolithic skeletons from NKLE and later Iron Age inhabitants of SSH was remarkably similar. While the early Neolithic inhabitants at NKLE were among Taiwan’s first farmers who likely cultivated millet, they also relied extensively on marine and terrestrial resources (Tsang, 1995; Li, 2013). Likewise, the subsistence of the Iron Age settlers at SSH, included not only cultivated crops such as rice, but also marine and terrestrial resources (Tsang, 2000). Reliance on marine resources represented an important part of the subsistence strategies of Taiwan’s Neolithic and later Iron Age inhabitants for at least several millennia. The overall similarities in the indicators of health for NKLE and SSH revealed in this study are consistent with this interpretation.
Regional comparisonsUnderscoring the uniqueness of these samples, there are limited comparative data for skeletal series outside of Taiwan that overlap temporally and spatially with those from NKLE and SSH (Table 2). This is especially true for skeletal series across the Taiwan Strait in southern China. Another problem is the lack of standardization in reporting data in the studies conducted by different researchers. Where possible, we have selected sites with similar subsistence economies and ones that temporally overlap, at least partially, the NKLE and SSH skeletons.
StatureAlthough dependent on the stature estimation formula used, with the exception of some of the indigenous groups in Taiwan and the Jomon skeletons, the average adult male stature for NKLE generally falls below the average statures for most of the skeletal series from China and Southeast Asia investigated (Table 7). Comparing the average statures of Taiwan’s indigenous groups, the NKLE adult male stature is most similar to Atayal, Saisiyat, and Puyuma groups. In contrast, the average stature of NKLE adult females is comparable or exceeds female adult stature for many of the skeletal series from East and Southeast Asia. Among the indigenous groups, the stature of NKLE females is most similar to the Amis and Tsou indigenous groups. Some of these differences may be due to the use of different stature formula and how the lengths of femur were recorded in the NKLE skeletal remains.
Cribra orbitaliaThere is considerable variability in the prevalence of CO among the comparative series (Table 8), suggesting the possibility of differences in the methods used to record this indicator. Some of the highest frequencies of this early childhood stress indicator are those reported for early skeletal series from Vietnam and the Khok Phanom Di series from Thailand, all significantly different when compared to the relatively lower frequency reported for NKLE. The early inhabitants of these sites, based on these and other data, were no doubt experiencing compromised health. Some of the lowest frequencies of this indicator are those for Jomon and Yayoi, skeletons from sites in Northeast Thailand, and early and middle Yangshao skeletons from China. The frequencies of CO in the NKLE skeletons are most similar to the skeletal series from China and Thailand. As reported in Pietrusewsky et al. (2013) the frequency of another indicator of early health, LEH, is moderately elevated in the NKLE series, suggesting that the early Neolithic inhabitants of Taiwan experienced relatively high levels of physiological stress during childhood.
Dental staining and dental pathologyWhile the cultural practice of chewing of betel nut (Areca catechu L.) is fairly prevalent in contemporary indigenous groups throughout Southeast Asia, Taiwan, and the Pacific, the lack of comparative data recorded in archaeological remains from these regions impedes comparisons. Limited information (Table 9) indicates that betel-stained teeth were reported for Vat Komnou (Cambodia) and Nui Nap (Vietnam) but were relatively infrequent in the Ban Chiang (Thailand) and SSH skeletal series. The betel staining observed in the Bronze Age site Nui Nap in Thanh Hoa Province in Vietnam is likely due to the intentional application of betel nut residues to the teeth rather than as a result of chewing betel nut (Oxenham et al., 2002b). Although high compared to the other groups the frequency of dental staining observed in the NKLE skeletons, as discussed earlier, causes other than chewing betel nut may be contributing to the observed staining. If we include other sites that report, at least anecdotally, dental staining attributed to betel nut chewing (Table 10), the number of sites increases. Based on our observations of the staining observed in the NKLE teeth, some of these reported cases might require further evaluation.
| Site/Country | Location | Cultural period | Dates | Reference |
|---|---|---|---|---|
| Taiwan | ||||
| Nankuanli East | SW Taiwan | Early Neolithic | 5000–4200 BP | this study, Pietrusewsky et al. (2013) |
| Shihsanhang | NW Taiwan | Iron Age | 1500–1000 | Pietrusewsky and Tsang (2003), Chang (1993) |
| Peinan | SE Taiwan | Neolithic | 3500–2800 BP | Lien (1989) |
| Kenting | S Taiwan | Neolithic | ~4000 BP | Ohashi and Matsumura (1933), Cited in Lien (1989: 181) |
| Oluanpi | S Taiwan | Neolithic | 5000–2000 BP | Sung et al. (1967: 37), Cited in Lien (1989: 181) |
| Southeast Asia | ||||
| Duyong Cave | SW Palawan, Philippines | Early Neolithic | 4630 ± 250 BP | Fox (1970: 60–65), Bellwood (1997: 221–222), Barreto-Tesoro (2003) |
| Ubujan | Bohol Island, C Visayas, Philippines | Metal Age | 1000 BP | Yankowski (2005) |
| Leang Buidane | Talaud Islands, N Indonesia | Early Metal | 2000 BP | Bellwood (1997: 297) |
| Ban Chiang* | NE Thailand | pre-Metal to Bronze Age | 4000–1700 BP | Pietrusewsky and Douglas (2002) |
| Non Pa Klauy* | NE Thailand | 4000–1800 BP | Pietrusewsky (1988) | |
| Vat Komnou | S Cambodia | Iron Age | 2500–1500 BP | Pietrusewsky and Ikehara-Quebral (2006) |
| Nui Nap | N Vietnam | Bronze Age | 3000–1700 BP | Oxenham et al. (2002b) |
| Red, Ma, and Ca River sites | N Vietnam | Metal Age | 3300–2600 BP | Oxenham et al. (2002a) |
| Giong Ca Vo | S Vietnam | Sa Huynh | 2500 BP | Oxenham et al. (2002b: 914) |
| Pacific | ||||
| Chelechol ra Orrak | Palau | 3000 BP | Fitzpatrick (2003), Fitzpatrick et al. (2003) | |
| Aptoguan | Guam, Mariana Islands | Latte Period | 1000–1521 BP | Douglas et al. (1997) |
| Alaguan | Rota, Mariana Islands | Latte Period | 900–300 BP | Hocart and Fankhauser (1996) |
| St Matthias Island (Mussau Island) | Bismarck Archipelago, Papua New Guinea | Lapita | 1600–500 BCE | Kirch et al. (1989) |
| Motupore Island | S Papua New Guinea | 800–350 BP | Allen et al. (1997) | |
| Nebria | S Papua New Guinea | 1000–400 BP | Pietrusewsky (1976) | |
As reported previously (Pietrusewsky et al., 2013), the frequencies of dental pathology in the NKLE and SSH series are among the lowest reported, suggesting generally good dental health for these prehistoric inhabitants of Taiwan. Low frequencies of dental caries have been linked to rice-dominated agricultural communities (as opposed to other grains such as corn) (Tayles et al., 2000), non-agricultural subsistence economies, marine diets, diets low in starches and sugars, as well as the effects of chewing betel nut. The coastal locations of NKLE and SSH and archaeological evidence suggests that the early inhabitants of both sites were involved in fishing and the exploitation of marine resources.
The average stature of Taiwan’s early Neolithic people from the NKLE site is estimated to be approximately 160 cm for adult males and 155 cm for adult females, similar to the stature of indigenous Taiwanese and other prehistoric inhabitants from surrounding regions. The frequencies of CO, a childhood indicator of health, are similar in male and female skeletons from NKLE. The relatively low frequency of this indicator suggests the early Neolithic inhabitants experienced some physiological stress as children, most likely attributable to diseases and malnutrition, this did not have an adverse effect on adult stature. The significantly higher frequencies of stained teeth, advanced attrition, and lower frequencies of dental caries, dental calculus, and AMTL in the NKLE males lends support that chewing of betel (areca) nut is a likely cause of the observed dental staining in these remains.
As demonstrated previously (Pietrusewsky et al., 2013), the prediction that there was a decline in health corresponding to the transition from early Neolithic to Iron Age in Taiwan is not fully supported by the results of this study. While the frequency of CO is significantly greater in the Iron Age SSH series, as expected, a reduction in stature is not observed in the Iron Age series. The similarities in dental health between the early Neolithic skeletons from NKLE and the later Iron Age skeletons from SSH suggest they shared a similar subsistence economy, one that was likely based on fishing and the hunting and gathering of marine and terrestrial resources.
Comparisons of the NKLE skeletons with those from surrounding regions demonstrate that the average stature of NKLE males generally falls below many of the comparative series, while the average stature of NKLE females is generally greater than most of the series from East and Southeast Asia. Overall, the stature of NKLE adults is similar to modern indigenous groups in Taiwan. The frequency of CO in the NKLE skeletons is relatively low and similar to skeletal series from China and Thailand. Additional studies of the skeletons from the NKLE site and other sites in the Tainan Science Park will improve our understanding of the health and lifestyle of Taiwan prehistoric inhabitants.
Our thanks to Marween Yagin, Graphics Specialist, Center for Instructional Support, University of Hawaii at Manoa, who made the map for this paper. The National Research Council of Taiwan supported this research.
