Original Articles
Growth problems in a skeletal sample of children abandoned at Santa Casa da Misericórdia, Faro, Portugal (16th–19th centuries)
2015 Volume 123 Issue 3 Pages 149-159
Details
2015 Volume 123 Issue 3 Pages 149-159
In 2006, an excavation at Santa Casa da Misericórdia de Faro, Portugal, revealed a cemetery dating from the 16th–19th centuries with several phases of use, one of which presented 51 sub-adult individuals that had been abandoned at the institution through the ‘foundling wheel’—a device that allowed people to leave unwanted newborns anonymously. Of the 46 individuals for whom it was possible to estimate age-at-death, more than 80% were under two years. Skeletal samples from individuals of these ages are not commonly found in the archaeological record. This Faro sample is also unique in the sense that it is the first time it has been possible to study an osteological assemblage from abandoned children. Considering the assertion that dental development is buffered against environmental insults in comparison to skeletal development, and that discrepancies between dental and skeletal age estimations are suggestive of growth delay, this study aims to investigate if the immature individuals of this institution were exposed to severe environmental restrictions. Skeletal age was estimated according to long bone lengths and epiphyseal fusion. Dental age was calculated on the basis of dental development, namely dental calcification and the sequence of formation and eruption of teeth. Furthermore, a palaeopathological analysis of the sample was conducted. The age estimates obtained by the ossification and fusion of different skeletal elements resulted in too wide intervals and were therefore excluded from subsequent analysis. The age estimates obtained by the dental methods were consistent (100%). By contrast, the estimates obtained by osteometric and dental methods showed some disagreement (osteometric vs. dental calcification: 63.6%; osteometric vs. sequence of formation and eruption of teeth: 80.0%), the osteometric providing younger ages. Regarding the paleopathology, the high frequency of porotic lesions (60.8%) and new bone deposition (37.3%), especially in the individuals previously identified as small for their age, make evident the difficulties experienced by these individuals during their short lives.
In 2006, an archaeological survey done before the rehabilitation of the building of Santa Casa da Misericórdia de Faro, Faro, Portugal, brought to light a necropolis dating from the 16th–19th centuries with several phases of use: (1) three adult burials with an east–west orientation, with their heads to the west (in accordance with the Christian belief in the resurrection of the soul); (2) one large ossuary and 27 adult burials of both sexes and different ages-at-death; and (3) 51 inhumations of sub-adults, chronologically more recent, mainly with a south–north orientation. There were, however, a few exceptions in the orientation of the position of the skeletons: six individuals were buried with a southwest– northeast orientation, one with a southeast–northwest orientation and one with a north–south orientation. None of the sub-adults had a typical Christian orientation (Corga and Ferreira, 2010). Although this third phase is stratigraphically more recent than the previous, the absence of archaeological material prevented us from deducing the time span of these burials. Unfortunately, it was not possible to perform a radiocarbon analysis of the remains that would allow for a more precise chronology of the series. Santa Casa da Misericórdia is a group of confraternities based on religious values and with charitable purposes. In those days, these were divided into fourteen works of mercy (Sá, 1997; Abreu, 2000), two of which were the burial of the dead and the care of abandoned children (Sá, 1997; Araújo, 2000; Fonte, 2004). The abandoned newborns were received at the institution through the ‘foundling wheel’ (roda dos expostos). This device, built into the wall with a revolving door, was designed with the aim of making it possible to abandon (‘expose’ in the language of the time) newborns, while maintaining the anonymity of the abandoner (Fonte, 2004; Dias, 2008; Leandro, 2011). The anonymity was in fact a crucial issue because in several instances these women were young and single mothers from all socioeconomic strata for whom motherhood was not socially accepted. In other cases, poverty was the reason behind child abandonment (Dias, 2008; Leandro, 2011). A strong positive correlation has been observed between periods of subsistence crisis and the number of abandoned children. But there was also the possibility of socially unacceptable motherhood resulting in poverty. After arriving at the institution, the child was frequently delivered to a nanny, usually a wet nurse. Most of these women were illiterate and lived in rural areas, working in agriculture (Fonte, 2004).
Skeletal samples from such young individuals are found rarely in the archaeological record, and to date this Faro sample is the only osteological assemblage from abandoned children to be described and studied. Therefore, it provides a unique opportunity to investigate whether the immature individuals of this institution were exposed to severe environmental constraints. This objective may be attained by the evaluation of their growth and health.
Growth is a graded process that involves various changes related to age, characterized mainly by the increase in size and the differentiation in shape of each part of the body (Ribot and Roberts, 1996; Scheuer and Black, 2000). Its most important aspect is variation due to the complex interaction between genetic and environmental factors (Ribot and Roberts, 1996; Bogin, 1999; King and Ulijaszek, 1999; Scheuer and Black, 2000; Lewis, 2007). Modern studies have shown that growth is considerably adaptable, and can be affected by several factors, including nutrition, infection, socioeconomic status, urbanization, migration, physical activity, physiological stress, climate, altitude and parasitism (Lewis, 2007). However, the two main groups of environmental stressors that may lead to growth disruption are infection and nutrition (King and Ulijaszek, 1999; Saunders and Barrans, 1999). These two factors present a complex and intricate interaction. Poor nutritional status induces a decline in immunocompetence, while exposure to infectious agents leads to anorexia, malabsorption, elevated basal metabolic rate as a consequence of fever, and protein catabolism needed for the immune response (King and Ulijaszek, 1999). During a stressful period, the growth rate slows until normal conditions are resumed, with the magnitude of decelerated growth reflecting the severity of the episode (Bogin, 1999; Lewis, 2007). Therefore, individuals reach developmental stages, or biological ages, along the maturity continuum, at different chronological ages (Scheuer and Black, 2000). Given that growth disruptions are cumulative, the older the individual, the larger may be the discrepancy between physiological and chronological ages. Once the child’s health condition has been re-established, it has the ability to ‘catch-up’ with its peers and resume growth at an increased rate. Such ability is impaired with age, and if growth is decelerated for a long period of time, the increased growth velocity necessary for ‘catch-up’ cannot be realized (Lewis, 2007).
Before cessation of growth in height, age may be estimated from several dental and/or skeletal markers.
The dental age estimation is derived from dental development, namely tooth formation and emergence (Moorrees et al., 1963; Ubelaker, 1989). According to several authors (Moorrees et al., 1963; Demirjian et al., 1973), tooth formation is more informative than tooth emergence because the majority of the teeth can be analysed at each examination. Conversely, emergence represents only a specific stage of short duration in the continuous process of dental eruption, or migration to reach the occlusal level. Moreover, gingival emergence may be influenced by local and/or environmental factors, such as caries, loss of deciduous predecessors, lack of space in the dental arch, and malnutrition (Moorrees et al., 1963; Demirjian et al., 1973; Alvarez and Navia, 1989). In contrast, the formation rate of the permanent teeth is not affected by early loss of the deciduous teeth and is less prone to be influenced by environmental stressors (Moorrees et al., 1963; Demirjian et al., 1973; Meinl, 2007; Pinhasi, 2008). Another potential problem with the tooth eruption charts is that, in most clinical research, ‘eruption’ is defined by the emergence of teeth through the gingiva or gum, whereas in skeletal remains eruption is considered in terms of alveolar emergence (Halcrow et al., 2007). However, according to some authors, alveolar emergence in deciduous teeth probably occurs a short time before gingival emergence (Konigsberg and Holman, 1997) and only two months before the occlusion of teeth (Hulland et al., 2000). Regarding permanent teeth, the reported differences in timing between alveolar and gingival emergence are greater: 0.6 years for teeth with deciduous predecessors, and 1.4 years for first and second molars (Haavikko, 1970). Therefore, if eruption standards are used in archaeological remains, researchers should be aware of this issue.
The skeletal age may be estimated from the developmental state of ossification centres and/or from the dimensions of numerous osteological elements (Hoppa and FitzGerald, 1999; Scheuer and Black, 2000; Scheuer, 2002). It is generally observed that dental age is the most accurate indicator in the postnatal period (Hoppa and FitzGerald, 1999; Scheuer and Black, 2000; Scheuer, 2002; Brickley, 2004; Cardoso, 2007; Boccone et al., 2010; Elamin and Liversidge, 2013; García-Mancuso, 2014). A possible reason is that the development of part of the dentition takes place before birth, in a protected environment, whereas skeletal growth and development are exposed for a longer period of time to external factors such as nutritional quality, sanitary conditions, socioeconomic status, and possibly climate and altitude (Bogin, 1999; Hoppa and FitzGerald, 1999; Scheuer, 2002; Cardoso and Garcia, 2009). This difference is the basis of growthrelated studies where the children’s development is believed to be the reflection of health and nutritional status better than any other index (Lampl and Johnston, 1996; Hoppa and Fitzgerald, 1999; Saunders and Barrans, 1999; Larsen, 2002; Lewis, 2002). Moreover, growth of the various skeletal elements may vary in velocity and growth faltering due to differing susceptibilities to environmental insults (Pinhasi, 2008).
Non-specific stress indicators (e.g. dental enamel hypoplasias and porotic lesions such as cribra orbitalia, cribra humeralis, cribra femoralis, and abnormal porosity) may provide indirect information regarding the health of past populations (Larsen, 2002; Lewis, 2007). These lesions can arise in the developing child but, with the exception of dental enamel hypoplasias, will disappear if the child recovers (Lewis, 2007). The frequency of these stress indicators by age, the relation between them, and the comparison of their frequencies with both demographic and growth curves may provide more information about the sub-adults’ health (Ribot and Roberts, 1996; Larsen, 2002).
The aim of the present study is to compare the pattern of skeletal and dental development of the immature individuals of Santa Casa da Misericórdia de Faro in order to know if they were affected by poor living conditions and/or limited access to proper nutrition and healthcare. Furthermore, the paleopathological profile of these children, namely concerning enamel defects, cribra orbitalia, cribra humeralis, cribra femoralis, porosity, woven bone, osteomyelitis, trauma, endocranial lesions, and other osteolytic and/or osteoblastic lesions, will be drawn to help in the interpretation of their health status.
The 51 subjects studied were found in a poor state of preservation and completeness (Table 1), which hampered the examination.
| Individual no. | Preservation | Completeness |
|---|---|---|
| 1 | Fair | Relatively complete |
| 2 | Poor | Very incomplete |
| 3 | Fair | Very incomplete |
| 4 | Poor | Very incomplete |
| 5 | Poor | Very incomplete |
| 6 | Poor | Very incomplete |
| 7 | Poor | Very incomplete |
| 8 | Poor | Very incomplete |
| 9 | Poor | Very incomplete |
| 10 | Poor | Very Incomplete |
| 11 | Poor | Very incomplete |
| 12 | Poor | Very incomplete |
| 13 | Fair | Very incomplete |
| 15 | Fair | Very incomplete |
| 16 | Poor | Very incomplete |
| 17 | Fair | Very incomplete |
| 19 | Fair | Very incomplete |
| 20 | Poor | Very incomplete |
| 21 | Fair | Relatively complete |
| 25 | Poor | Very incomplete |
| 26 | Fair | Relatively complete |
| 27 | Fair | Relatively complete |
| 28 | Fair | Relatively complete |
| 29 | Poor | Very incomplete |
| 30 | Fair | Very incomplete |
| 31 | Fair | Very incomplete |
| 32 | Poor | Very incomplete |
| 33 | Poor | Very incomplete |
| 34 | Fair | Very incomplete |
| 36 | Fair | Very incomplete |
| 38 | Fair | Very incomplete |
| 40 | Poor | Very incomplete |
| 42 | Poor | Very incomplete |
| 43 | Fair | Very incomplete |
| 45 | Poor | Very incomplete |
| 46 | Poor | Very incomplete |
| 47 | Poor | Very incomplete |
| 48 | Poor | Very incomplete |
| 49 | Poor | Very incomplete |
| 50 | Poor | Very incomplete |
| 51 | Poor | Very incomplete |
| 52 | Poor | Very incomplete |
| 53 | Poor | Very incomplete |
| 54 | Poor | Very incomplete |
| 55 | Poor | Very incomplete |
| 56 | Poor | Very incomplete |
| 57 | Poor | Very incomplete |
| 58 | Fair | Relatively complete |
| 63 | Poor | Very incomplete |
| 64 | Poor | Very incomplete |
| 65 | Poor | Very incomplete |
Poor = severe bone damage, metaphyseal loss with long bones and cancellous exposure of the vertebrae.
Fair = slight bone damage, erosion of bone surface and/or bone fragmentation.
Good = no significant damage of the bone surface or bone fragmentation.
Very incomplete = more than 50% of the skeleton is absent.
Relatively complete = 25–50% of the skeleton is absent.
Almost complete = less than 25% of the skeleton is absent.
The following methods were used to estimate the age-at-death: Moorrees et al. (1963) for dental calcification; Ubelaker (1989) for the sequence of formation and eruption of teeth; Schaefer et al. (2009) for the skeletal dimensions; Fazekas and Kósa (1978) for the prenatal long bone lengths; Maresh (1970, reviewed by Cardoso, 2005) for the postnatal long bone lengths; and Scheuer and Black (2000) for development and bony fusion. As already highlighted by Ribot and Roberts (1996), the accuracy of these odontological and osteological methods has been tested by several authors on various American and European samples. Regarding Ubelaker’s (1989) chart, although it was originally created for anthropological work with American Indians, it has been approved by the Workshop of European Anthropologists (Ferembach et al., 1980).
Although tooth formation is more informative than tooth emergence (as explained in the Introduction) both approaches were used to estimate age-at-death because the macroscopic evaluation of the morphological changes occurring during the formation of teeth is not completely possible unless the walls of the tooth crypts have been damaged, exposing the tooth root and/or crown for observation. Unfortunately, it was not possible to perform a radiographic analysis of the jaws. On the other hand, it will be interesting to see if these two methods of dental age estimation give similar results.
Skeletal measures (in millimeters) were obtained from the left side of the skeleton (or from the right side if the left bone was missing or severely damaged), using a sliding calliper and an osteometric board, following Redfield (1970), Fazekas and Kosá (1978), Buikstra and Ubelaker (1994), and Schaefer et al. (2009).
All data were collected blindly by one observer (J.P.) on two different occasions. In the case of metric data, the intraevaluator error was assessed by the technical error of measurement (TEM). This index allows researchers to verify the accuracy when performing and repeating anthropometrical measurements (intra-observer) (Perini et al., 2005). For non-metric data, the intra-evaluator error was assessed through the percentage of agreement between observations and also with Cohen’s kappa coefficient.
All skeletal dimensions obtained for each individual were used to estimate age-at-death. In this case, the age estimation is given by an interval whose lower limit corresponds to the youngest age obtained and whose upper limit corresponds to the oldest age.
In order to compare the estimates of age-at-death obtained by the four methodological approaches, the percentages of agreement between the different ages were compared two by two.
Following Goode et al. (1993), Maresh’s data was used to standardize the long bone lengths with the following modification: the long bone data of Fazekas and Kósa (1978) for 10 lunar month individuals to represent the birth (0 year) age class was used.
Several pathological lesions were scrutinized according to Ortner (2003) and Buikstra and Ubelaker (1994): cribra orbitalia, cribra humeralis, cribra femoralis, porosity, woven bone, osteomyelitis, trauma, endocranial lesions, and other osteolytic and/or osteoblastic lesions. Cribra humeralis and cribra femoralis are similar in appearance to cribra orbitalia, but are located in the medial anterior region of the proximal end, near the metaphysis, of the humerus and femur, respectively. Given the high number of bone elements whose poor preservation hampered the observation and since the macroscopic characteristics and aetiology of these lesions are identical, the three types of cribra were grouped with the same general designation: cribra (Miquel-Feucht et al., 2000; Garcia et al., 2002). Following Ortner et al. (1999, 2001), abnormal porosity was defined as a localized condition, in which fine holes, typically less than 1 mm in diameter, penetrate a compact bone surface. Although vascular holes in cortical bone are a normal anatomical feature, these are much fewer in number in a given area and more variable in size, frequently exceeding 1 mm in diameter (Ortner et al., 2001). Abnormal porosity of methaphyses was distinguished from porosity associated with normal remodelling on the basis of the size of the affected area, extending more than a centimeter from the end of the bone (Krenz-Niedbała, 2015). For all available anatomical sites, the lesions were recorded as absent, present, or non-observable.
Enamel defects (hypoplasia or others) were investigated following Hillson (2001).
The statistical significance of the recorded values was tested with independent samples chi-square test. Statistical analysis was conducted by using the IBM SPSS statistics® program (version 22.0).
The age estimates obtained by the ossification and fusion of different skeletal elements resulted in too wide intervals and were therefore excluded from the subsequent analysis.
Table 2 presents age-at-death estimations for each individual by the three remaining methodologies. As can be seen, in five individuals it was not possible to estimate age-at-death due to the high level of fragmentation and incompleteness of the osteological elements (Table 1). Therefore, the sample studied hereinafter is composed of 46 sub-adults. These were all under 15 years of age and more than 80% were under 2 years of age.
| Individual no. | Dental calcification | Sequence of formation and eruption of teeth | Osteometric Data |
|---|---|---|---|
| 1 | 0.4–0.5 y | 6 ± 3 m | 4–9 m |
| 2 | — | — | 38 fw–9 m |
| 3 | — | 0 ± 2 m | 34 fw–2 w |
| 4 | — | 6 ± 3 m | 36 fw–2 m |
| 5 | — | — | 30–38 fw |
| 6 | — | — | 32–36 fw |
| 7 | — | 1 y ± 4 m | — |
| 8 | — | — | 0–6 m |
| 9 | — | — | 38 fw–6 m |
| 10 | — | 9 ± 3 m | 6 m–2 y |
| 11 | — | 0 ± 2 m | 36 fw–6 m |
| 12 | — | — | 36–38 fw |
| 13 | — | — | 32 fw–2 m |
| 15 | — | 7 ± 2 fm | 32–40 fw |
| 16 | — | — | 0–6 m |
| 17 | — | — | 36 fw–6 m |
| 19 | — | 0 ± 2 m | 36 fw–6 m |
| 20 | — | — | — |
| 21 | 7.0–10.2 y | 9 y ± 24 m | 6–10 y |
| 25 | 2.4–4.3 y | 3 y ± 12 m | 1–3 y |
| 26 | — | 0 ± 2 m | 34 fw–6 m |
| 27 | 1.8–1.9 y | — | 1–2 y |
| 28 | 1.3–1.7 y | 1 y ± 4 m | 0–2 y |
| 29 | — | 0 ± 2 m | 36 fw–6 m |
| 30 | — | — | 36 fw–6 m |
| 31 | — | 0 ± 2 m | 34–40 fw |
| 32 | — | — | 14–15 y |
| 33 | — | — | — |
| 34 | 1.8–3.5 y | 18 ± 6 m | 7–24 m |
| 36 | — | — | 0–9 m |
| 38 | — | — | 4–12 m |
| 40 | — | 6 ± 3 m | 2–4 m |
| 42 | — | — | 0–6 m |
| 43 | — | — | 0–9 m |
| 45 | — | — | 28–32 fw |
| 46 | 7.4–7.8 y | 8 y ± 24 m | 7–8 y |
| 47 | — | — | 6–18 m |
| 48 | 4.5–4.8 y | 5 y ± 16 m | 2.5–3 y |
| 49 | — | — | — |
| 50 | — | — | 2–5 y |
| 51 | — | — | 0–2 m |
| 52 | — | — | 2.5–3 y |
| 53 | — | 6 ± 3 m | — |
| 54 | — | — | — |
| 55 | — | — | 0.5–1 y |
| 56 | — | — | 6–24 m |
| 57 | 1.8–3.5 y | 18 ± 6 m | 2–3 y |
| 58 | 2.4–4.3 y | 3 y ± 12 m | 1.5–2 y |
| 63 | 4.4–5.8 y | 6 y ± 24 m | 7–12 y |
| 64 | — | — | 0–6 m |
| 65 | — | — | — |
y = years; m = months; fm = fetal months; fw = fetal weeks.
Regarding the intra-evaluator error for metric data, all relative TEM values were lower than 7.5%, and were thus considered acceptable (Table 3). Therefore, no osteological measurement was excluded from the remaining analysis. For non-metric data, although the values of N were small, both the percentages of agreement between the observations and Cohen’s kappa coefficients were high for the two approaches to dental age estimation, and the results were considered acceptable (Table 4). The same occurred in the paleopathological analysis (Table 5).
| Bone | Measurement | n | Relative TEM (%) |
|---|---|---|---|
| Pars basilaris | Maximum length | 10 | 0.73 |
| Saggital length | 11 | 0.25 | |
| Maximum width | 10 | 0.04 | |
| Pars lateralis | Maximum length | 7 | 6.18 |
| Width | 5 | 0.22 | |
| Pars petrosa | Maximum length | 7 | 0.13 |
| Width | 11 | 6.18 | |
| Mandible | Body length | 2 | 4.98 |
| Clavicle | Maximum length | 9 | 0.96 |
| Scapula | Height | 2 | 0.27 |
| Breath | 8 | 0.26 | |
| Ilium | Maximum length | 18 | 0.05 |
| Width | 19 | 1.08 | |
| Ischium | Maximum length | 15 | 0.59 |
| Width | 15 | 0.11 | |
| Pubis | Maximum length | 19 | 0.71 |
| Humerus | Maximum length | 16 | 1.14 |
| Radius | Maximum length | 9 | 0.07 |
| Ulna | Maximum length | 11 | 0.01 |
| Femur | Maximum length | 13 | 0.04 |
| Distal width | 9 | 0.01 | |
| Tibia | Maximum length | 10 | 0.06 |
| Fibula | Maximum length | 14 | 0.14 |
| 1st Metatarsal | Maximum length | 10 | 0.07 |
| Age-at-death estimation method | n | Error (n) | Error (%) | Agreement (%) | Cohen’s kappa |
|---|---|---|---|---|---|
| Moorrees et al. (1963) | 11 | 1 | 9.09 | 90.91 | 0.889 |
| Ubelaker (1989) | 23 | 1 | 4.35 | 95.65 | 0.902 |
| Type of lesion | n | Error (n) | Error (%) | Agreement (%) | Cohen’s kappa |
|---|---|---|---|---|---|
| Porosity | 551 | 65 | 11.98 | 88.02 | 0.712 |
| Woven bone | 547 | 24 | 4.39 | 95.61 | 0.788 |
| Cribra (orbitalia, humeralis, femoralis) | 52 | 3 | 5.77 | 94.23 | 0.885 |
Table 6 shows the percentage of agreement between the three methods for estimating age-at-death (dental calcification, sequence of formation and eruption of teeth, and osteometric data). The age estimates obtained by the methods of dental calcification and sequence of formation and eruption of teeth were absolutely concordant (100%). By contrast, the estimates obtained by osteometric and dental methods showed some disagreement (tooth calcification vs. osteometric: 63.6%; sequence of formation and eruption of teeth vs. osteometric: 80.0%), i.e. the osteometric methods indicate younger ages.
| Age-at-death estimation method | n | Error (n) | Error (%) | Agreement (%) |
|---|---|---|---|---|
| Dental calcification vs. sequence of formation and eruption of teeth | 10 | 0 | 0.00 | 100.00 |
| Dental calcification vs. osteometric data | 11 | 4 | 36.36 | 63.64 |
| Sequence of formation and eruption of teeth vs. osteometric data | 20 | 4 | 20.00 | 80.00 |
Standardized long bone lengths are presented in Table 7, following Goode et al. (1993). A strong trend was observed for fetuses and newborns to present measures very close to or slightly higher than those obtained by Fazekas and Kósa (1978) for the same dental age. On the other hand, older individuals had lower measures that the ones tabulated by Maresh (1970, values corrected by Cardoso, 2005). When considering the total sample’s average, a slight discrepancy between the upper and lower limbs was recognized, with the lower limbs presenting slightly lower values.
| Humerus | Radius | Ulna | Femur | Tibia | Fibula | Bones average | |
|---|---|---|---|---|---|---|---|
| Individuals aged ≤0 years | 1.02 | 1.02 | 0.98 | 1.034 | 1.05 | 1.03 | 1.00 |
| Individuals aged >6 months | 0.87 | 0.88 | 0.85 | 0.83 | 0.81 | 0.82 | 0.84 |
| All sample | 0.94 | 0.96 | 0.92 | 0.92 | 0.90 | 0.91 | 0.91 |
The most frequent pathological conditions observed were porosity (Figure 1), woven bone (Figure 2), and cribra (Figure 3). The chi-square test showed that the presence of cribra is distributed similarly in the different bone elements analysed (chi-square test: 0.345; d.f. = 2; P = 0.841), i.e. in the orbital cavity (cribra orbitalia: n observed = 5; 40% affected), humerus (cribra humeralis: n observed = 25; 52.0% affected) and femur (cribra femoralis: n observed = 22; 45.5% affected). In contrast, the presence of porosity (chi-square test: 161.538; d.f. = 21; P < 0.000) and woven bone varies (chi-square test: 61.035; d.f. = 21; P < 0.000) according to the bone element under analysis. Regarding porosity, the bone elements most affected were: sphenoid; occipital; pubis, ilia, and ischia; temporal; mandible; and humerus (Table 8). The woven bone was more frequently observed in the following bone elements: pubis, ilia, and ischia; humerus; femur; occipital; temporal; mandible; frontal; tibia; and scapula (Table 8).
Inferior view of the pars basilaris (individual no. 43; 0–9 months old) with porosity.
Medial view of the right tibia diaphysis (individual no. 43; 0–9 months old) presenting woven bone.
Anterior view of the proximal end of the left femur diaphysis (individual no. 34; 1.8–3.5 years old) where cribra femoralis is visible.
| Bone element | Porosity | Woven bone | ||
|---|---|---|---|---|
| n observed | % present | n observed | % present | |
| Occipital | 25 | 68.0 | 25 | 24.0 |
| Parietal | 24 | 29.2 | 24 | 8.3 |
| Temporal | 22 | 59.1 | 22 | 22.7 |
| Frontal | 16 | 43.8 | 16 | 18.9 |
| Sphenoid | 17 | 70.6 | 17 | 5.9 |
| Zygomatic | 13 | 53.8 | 13 | 7.7 |
| Maxilla | 8 | 25.0 | 8 | 0.0 |
| Mandible | 22 | 59.1 | 22 | 22.7 |
| Clavicle | 28 | 7.1 | 28 | 3.7 |
| Scapula | 35 | 8.6 | 35 | 17.1 |
| Sternum | 24 | 4.2 | 24 | 0.0 |
| Vertebra | 43 | 7.0 | 43 | 0.0 |
| Ribs | 42 | 0.0 | 42 | 0.0 |
| Humerus | 33 | 36.4 | 33 | 27.3 |
| Radius | 26 | 15.4 | 26 | 11.6 |
| Ulna | 27 | 14.8 | 27 | 7.4 |
| Hand bones | 31 | 0.0 | 31 | 0.0 |
| Pubis, ilia, and ischia | 29 | 65.5 | 29 | 31.0 |
| Femur | 26 | 42.3 | 26 | 26.9 |
| Tibia | 22 | 50.0 | 22 | 18.2 |
| Fibula | 22 | 4.5 | 22 | 4.5 |
| Foot bones | 16 | 18.8 | 16 | 0.0 |
When evaluating the frequencies of pathological evidence between age-groups (Table 9), a close relation was observed between the estimated age-at-death and the presence of porosity (chi-square test: 12.846; d.f. = 5; P = 0.025) and woven bone (chi-square test: 14.888; d.f. = 5; P = 0.011), with both conditions more frequently observed between the fetal period and the age of 3 years. Conversely, there was no relation between the presence of cribra and age (chi-square test: 7.269; d.f. = 5; P = 0.201).
| Age-group | Porosity | Woven bone | Cribra (orbitalia, humeralis, femoralis) | |||
|---|---|---|---|---|---|---|
| n observed | % present | n observed | % present | n observed | % present | |
| 7th lunar month–birth | 6 | 83.3 | 6 | 100 | 4 | 25.0 |
| Birth–28 days | 10 | 100 | 10 | 40.0 | 10 | 70.0 |
| 2nd month–3 years | 21 | 61.9 | 21 | 42.9 | 15 | 73.3 |
| 3–6 years | 6 | 50.0 | 6 | 0.0 | 5 | 40.0 |
| 7–10/12 years | 2 | 0.0 | 2 | 0.0 | 1 | 0.0 |
| Adolescence | 1 | 0.0 | 1 | 0.0 | 1 | 0.0 |
The diaphyseal lengths of the humerus and femur were plotted against dental ages for each sub-adult, in order to obtain a growth curve (Figure 4, Figure 5). The dental age estimation method by Ubelaker (1989) was the one that allowed more estimations and consequently was the most appropriate to use. In both graphics (Figure 4, Figure 5), two lines were drawn in order to compare the stressed children (with several stress indicators) with the less or unstressed children (with only one stress indicator or none). Although the sample size is small, it is possible to observe in both graphs that the stressed children had smaller diaphyseal lengths than less or unstressed children.
Growth curve with humeral diaphyseal lengths plotted against dental age estimation (Ubelaker’s method). —○: with no or one stress indicator; ---□: with more than one stress indicator.
Growth curve with femoral diaphyseal lengths plotted against dental age estimation (Ubelaker’s method). —○: with no or one stress indicator; ---□: with more than one stress indicator.
The paleopathological analysis of the sample also revealed two cases that should be mentioned here. One is a 2-year-old child (individual no. 27) who presented unusual intracranial marks consistent with bacterial meningitis (for more details about the differential diagnosis see Paredes et al., 2013). The other is an individual of about 3 years (individual no. 58) who presented plane-form enamel defects in the occlusal region (in which layers of enamel matrix are missing and a prominent step runs around the edge (Hillson, 2001)), above the contact area, of the labial surface of deciduous lateral incisors. Enamel hypoplasias were also found in the permanent teeth of three other children: individual no. 7 (1 year ± 4 months of age) had a linear hypoplasia in the upper right canine; individual no. 46 (7.4–7.8 years of age) had one linear defect in the left lower first premolar; and individual no. 63 (4.4–5.8 years of age) had two linear enamel hypoplasias in the left lower canine.
No further significant pathological lesions were found in the analysed sample.
During growth, each individual has a genetic potential for increasing in size and shape, which depends upon different environmental conditions (Ribot and Roberts, 1996; Lewis, 2007). As highlighted by many authors (e.g. Bogin, 1999; Lewis, 2007; Pinhasi, 2008), several studies on modern populations have shown that low socioeconomic status can significantly affect the growth pattern of a given population. Dental development is less influenced by environment than skeletal development, and is considered the best indicator of chronological age in skeletal remains (Saunders et al., 1993; Cardoso, 2007). Therefore, the difference between skeletal and dental development may provide a measure of growth and health in a population. However, there are few studies comparing age estimations based on dental and skeletal development in fetal and infant periods in the osteoarchaeological literature (Lewis, 2000; García-Mancuso, 2014). In this sense, this Faro sample, comprising immature abandoned individuals, offers a unique opportunity to investigate whether these children were exposed to severe environmental constraints.
In the present study, the age estimates obtained by the ossification and fusion of different skeletal elements resulted in too wide intervals, as previously postulated by Pinhasi (2008).
When applying the methods of dental calcification and sequence of formation and eruption of teeth to this Faro sample, the results were concordant. Given that, according to Pinhasi (2008), dental eruption patterns may be more affected by environmental stressors than dental calcification (i.e. delayed eruption may be indicative of high levels of stress), the present results could be interpreted as suggesting that these Faro children did not suffer high levels of physiological stress. However, in contrast to these results, the ages obtained by osteometric and dental methods were different. The disagreement between dental calcification and osteometric data was 63.6% and between the sequence of formation and eruption of teeth and osteometric data was 80% (these different percentages are probably related to the wider age intervals given by the Ubelaker’s chart that more easily comprise the age estimation obtained by osteometric data). The fact that linear size measurements of different osteological elements provided younger ages when compared with observations of dental development indicates that growth in the Faro orphans deviated from the general path, a pattern that tends to be associated with malnutrition or disease. Moreover, it should be noted that in Ubelaker’s chart (1989: 64) “Eruption refers to emergence through the gum, not to emergence from the bone” and as the gums are not present in skeletons, these estimates could be slightly higher for deciduous teeth and even higher for permanent teeth. Taking into account the origin and age span of these individuals, such results are not surprising. High rates of infection and malnourishment (related to weaning practices and/or poor dietary quality) affect growth more severely during the period of maximal growth velocity, i.e. between 6 and 35 months of life (Pinhasi, 2008). Moreover, the difference between dental and skeletal ages increased with age which is consistent with a cumulative effect of growth constraints. In this sample, it was possible to identify a strong trend for fetuses and newborns to present very close or slightly higher skeletal measurements than the ones obtained by Fazekas and Kósa (1978) for the same dental ages. The closeness of the results may be related to a more protective environment in utero in comparison to postnatal life. In contrast, older individuals have lower skeletal measurements than the ones tabulated in Maresh (1970, values corrected by Cardoso, 2005). However, unlike the Faro sample, Maresh’s tables were developed based on live sub-adults of average to high social status. The differences observed suggest that for the Faro individuals, childhood probably took place with hygiene and food restrictions. Moreover, the older the child, the longer he/she would have lived in the institution, resulting in a cumulative effect of the environmental constraints. Despite not being possible to be certain about how long these children lived in the institution it is believed that it would depend on their age-at-death, since the ‘foundling wheel’ was used mainly to abandon newborns (Fonte, 2004).
According to some authors (Grauer, 1993; Lewis, 2007), if a stress period lasts less than 3 months, the child can compensate the delay by catch-up growth, but if the problem lasts longer, normal growth may be irreversibly compromised. In other words, catch-up growth is related to the recovery of normal growth rate, by improved environmental conditions (Binns, 1998), or a process of habituation of the body to environmental restrictions. Although, at first sight, the Faro results suggest that growth disturbances lasted long enough to impede catch-up growth, as older individuals are shorter for their ages, it should not be ignored that the sample size for this age cohort may not be sufficiently large for this phenomenon to be noted.
Observing the total sample’s average, a slight discrepancy between the upper and the lower limbs can be recognized, the latter being more affected by environmental constraints experienced during development. The same result was also obtained by Sciulli (1994). This may be due to the fact that bones of the lower limb are larger, have greater growth velocity, and are more variable in terms of both size and growth than those of the upper limb (Smith and Buschang, 2004). Greater relative lower than upper limb growth has already been reported for fetal (Moss et al., 1955) and postnatal (Maresh, 1959) periods. According to Sciulli (1994), the most rapidly growing long bones are more susceptible to nutritional and disease stress.
Concerning paleopathology, a higher frequency of porosity and woven bone in younger individuals was noted, particularly those under 3 years of age. It is still possible to see the same tendency in cribra despite the lack of statistically significant differences. Cribra, porosity, and woven bone are frequently related to anaemia, i.e. iron deficiency (Lewis, 2007). However, these lesions can also appear in other deficiency diseases, such as rickets and scurvy, as well as due to non-specific infections (Ribot and Roberts, 1996; King and Ulijaszek, 1999; Lewis, 2007). In fact, conditions such as malnutrition, dietary deficiency, and increased pathogen load can be on the origin of anaemia (Lewis, 2000). Abundant blood supply in the red bone marrow, especially at the metaphysis of the long bones, makes young sub-adults particularly susceptible to bone infections. Rapid growth in these areas requires a good blood supply which also transports pathogens that rely on iron to achieve their replication potential (Lewis, 2000, 2007). However, the extremely rapid growth during fetal and infancy periods may cause bone porosity, striation, and remodelling, making it difficult to distinguish between this normal process and a pathological aetiology (Ribot and Roberts, 1996; Lewis, 2000, 2007). In the present study, great care was taken when distinguishing normal and pathological porosity, following the criteria recommended by Ortner et al. (1999, 2001). In fact, when diaphyseal lengths of humerus and femurs were plotted against dental age estimations for each sub-adult, and taking into account the level of stress, it was possible to observe that the very stressed children had smaller diaphyseal lengths than less or unstressed children.
The skeletal parts identified as the most affected by porosity were the sphenoid, occipital, pubis, ilia, and ischia, temporal, mandible, and humerus. Regarding the deposition of woven bone, this was more observed in the pubis, ilia, and ischia, humerus, femur, occipital, temporal, mandible, frontal, tibia, and scapula. In both cases, the most affected elements are some of the largest bones in the skeleton, which are more resistant to post-mortem alterations and, on the other hand, more easily analysed.
Apparently there are no differences in the susceptibility of the various bone elements to cribra. According to Miquel-Feucht et al. (2000), the three forms of cribra (orbitalia, humeralis, and femoralis) are similar lesions, sharing macroscopic, microscopic, and radiographic characteristics, related to pathologies with the same aetiology. One of the possible aetiologies is anaemia, either of genetic, infectious, or nutritional origin. However, it should be noted that the high number of bone elements exhibiting poor preservation made observation impossible.
The paleopathological analysis of the Faro sample also revealed plane-form enamel defects in an individual about 3 years of age. Dental enamel hypoplasias are a non-specific stress indicator. Among the conditions that have been shown to cause these enamel defects are fever and starvation (Hillson and Bond, 1997; Lewis, 2000). Exposed-plane-form defects are more commonly found in the occlusal part of the crown, and are the most noticeable manifestation of enamel hypoplasia. They presumably represent a pronounced growth disruption (Hillson and Bond, 1997). Although several studies have tried to assess the age at which a defect was formed by measuring the distance between the defect and the cemento-enamel junction, Hillson and Bond (1997) argued that the position of the hypoplasia is more an expression of the pattern of enamel layers than the timing of a particular stress episode. In addition, the extent of an exposed-plane-form defect frequently bears no relation to the duration of the stress episode that caused it (Hillson and Bond, 1997).
Child growth standards and non-specific stress indicators are frequently used as measures of the general health status of a community since their growth, development, and diseases are sensitive indicators of the quality of the socioeconomic environment in which they live (Johnston and Zimmer, 1989; Bogin, 1999; Lewis, 2007). Child survivability depends greatly on the surrounding environment, in particular, the level of sanitation and availability of healthcare facilities (Lewis, 2007).
In the Faro skeletal sample of abandoned children both growth and paleopathological results suggest unfavourable conditions, as expected when food and hygiene were scarce as mentioned in historical documents (Fonte, 2004; Dias, 2008). The exposed children were subject to more risks than other children. These risks were diverse, namely malnutrition, a high probability of contracting infectious diseases either by contact with wet nurses or with other children, neglect and maltreatment (Dias, 2008). In the first months/years of life, malnutrition was mainly due to weaning. Sometimes, newborns were breastfed by their wet nurses, but in many cases, the care-givers were themselves malnourished or had no milk. In these cases, young babies were fed with animal milk (cow, goat, or donkey) (Fonte, 2004). In addition to the lack of the most basic health and hygiene practices and the absence of prophylactic measures and alternative forms of artificial feeding, the institutions and the family care homes did not meet minimum living conditions. Without water and sanitation, most of these homes would consist of unhealthy spaces, sometimes contiguous with the stables. All this obviously could encourage infectious diseases and lead to premature death of the exposed children (Fonte, 2004). In the present study, and regarding paleopathology, the relatively high frequency of porotic injuries and new bone deposition corroborate the poor conditions experienced by the Santa Casa’s orphans during their short lives.
This Faro sample provides the first opportunity to learn more about children who were abandoned in the south of Portugal via a ‘foundling wheel’ in the 16th–19th centuries. Not only are there few skeletal samples from such young individuals in the archaeological record, but to date this sample is the only osteological assemblage from abandoned children in the world to be described and studied. Despite the impossibility of inferring a specific aetiology for the observed paleopathological lesions or even to understand how the analysed skeletal stress indicators are related to each other, the detailed study of growth and health of the individuals who make up this sample is extremely important for a better knowledge of the living conditions and the degree of stress imposed on the children delivered to Santa Casa da Misericórdia de Faro at that time. In sum, the fact that age estimates obtained by osteometric gave younger ages than the dental methods suggests that these individuals experienced great difficulties during their short lives. Moreover, the paleopathological data, namely the high frequency of porotic lesions and new bone deposition, especially in the individuals previously identified as small for their age, corroborate this conclusion. Hence, we believe this study enriches the scanty osteoarchaeological documentation of both growth and health in abandoned children, and can contribute to reduce the discrepancy between historical and biological evidence.
The authors thank the Centro de Ciências Forenses, Centro de Investigação em Antropologia e Saúde, Grupo Dryas Octopetala, and Mónica Corga. The authors also acknowledge the three anonymous reviewers and the handling editor whose valuable comments and suggestions allowed us to improve the manuscript. This work was financed by national funds by FCT—Fundação para a Ciência e Tecnologia, under the project with the reference PEst-OE/SADG/UI0283/2013.
