Human Skeletal Remains from the Mamilla cave, Jerusalem

Yossi Nagar


The present paper reports on the skeletal remains from a cave on Mamilla street (henceforth: the ‘Mamilla’ cave), a mass burial site near the old city of Jerusalem, excavated in 1992 by R. Reich.  It is considered as one of some thirty burial sites of the Christian-Byzantine population of Jerusalem, which was massacred by the Persian army in 614 AD (Milik 1961:­­182-183; Reich 1994:117-118).  According to ancient sources about 24,000 bodies were buried near the pool of Mamilla (Reich 1996:33).

Regrettably, much of the recovered skeletal material from the cave is fragmentary, leaving only the relatively small bones for examination.  After a thorough sorting of the intact bones, the following were investigated: vertebrae; teeth (loose or in jaw fragments); patellae; tarsal, carpal, metacarpal and metatarsal bones; proximal, middle and distal phalanges of both limbs; temporal bones (including petrus bones); distal ends of adult humeri; children’s humeri and sacrum.  These can be regarded as a representative sample of the burials at Mamilla.[1]

Due to the poor state of preservation, the best indicators of age and sex, namely skull and pelvic bones, are absent in the sample. Nevertheless, an anthropological profile of the population buried at the Mamilla cave is reconstructed, using factors such as age distribution, frequency of pathologies and epigenetic traits (Nagar and Verdene, internal IAA file, 1996). This provides us with a useful database for comparative purposes.

[1]  The bones were randomly broken in the course of excavation, regardless of their previous state of preservation; age/sex estimation bias is not expected to exceed the normal.


Minimum Number of Individuals

The large amount of fragmentary bones in the Mamilla cave suggests that thousands of people were buried there.  However, the poor state of preservation allowed the identification of a minimum number of 526 individuals (Nagar and Verdene, internal IAA file, 1996).


Estimation of Sex

Having no skulls or pelvic bones complicated the sexing of the adult individuals.  It was done using measurements of the epicondylar width of the distal end of the humerus, where sex identification reliability exceeds 85% in various populations (Bass 1987:150–151).

Measurement of the epicondylar width of 189 humeri, right and left sides separately, yielded two bimodal graphs in which two populations (male and female) are represented.  Sex ratio is 38/100 males/females (averaged from right and left measurements), respectively .


The number of females greatly exceeds that of males.  One can assume that the Mamilla population’s epicondylar widths are smaller in average than those of the reference population (following Bass 1987:150–51), hence the odd sex ratio.  However, the two peaks of the graph: at 56–58 mm and at 64–65 mm are similar to the male and female population means reported for the reference population (Bass 1987:150–151).  Even if the individuals in the intermediate range (59–62 mm) are chiefly male, we still have a majority of females represented in the Mamilla skeletal remains. Considering the tragic circumstances of these individuals’ demise, a probable explanation to this unusual situation could be that males died in battles while females were mostly slain in the city.


Age Distribution

Reconstructing the overall age distribution of the Mamilla population to the accuracy needed for paleodemographic analysis is impossible.  The relative absence of males in the adult sample and the large under-representation of children’s teeth preclude the combination of adult and children numbers.

Determination of age in individuals older than 10 years was based on tooth attrition stages (using tables modified from Lovejoy 1985:47-56 and Hillson 1986:176–201).  A sample of 541 lower first molars (teeth 36, 46), and upper first premolars (teeth 14, 24) was studied (456 loose, 85 in jaw fragments). Age at death distribution of the population, using 10-year age intervals, is presented in Table 2, which represents the age distribution of the living population as well, since all died at the same time. Elderly individuals are probably under-represented, due to antemortem tooth loss, hence, a factor of 10% was added to the two highest age categories. Calculating survivorship in the common life-table methodology yields a relatively straight line, atypical of a regular cemetery population (Nagar and Verdene, internal IAA file, 1996). Therefore, this data is graphically presented as the age pattern of the population rather than as the common life-table method .  The proportion of individuals in the first (0–10) age group is an estimate based on Coale and Demeny (1966:6, west 1 model r=5).



The tapering shape of the age pyramid is typical of historical populations.  Such age pattern characterizes recent populations in most of the developing countries (Swedlund and Armelagos 1976:8-17).  These populations are distinguished by a relatively low life expectancy (Swedlund and Armelagos 1976:8-17).  However, not having survivorship data for the first age group impedes on the estimation of life expectancy at birth of the Mamilla population.

Another indication of the average age for the Mamilla adult population is obtained through the inspection of the jugular process of the temporal bone and the study of pathologies of the vertebral column. The frequency of age-related pathologies of the vertebral column bears upon the average age of the adult population buried at Mamilla (Stewart 1958:144-151; Waldron 1991:109). The results (see discussion below) indicate a relatively young population, as is expected of a living, rather than a cemetery population. In the temporal bones, the petro-occipital articulation is grossly indicative of age. In the majority of the population, this articulation ossifies with age (Hershkovitz 1997:365–373).  Examination of the joint in the Mamilla sample reveals that 27.3% of the joints are still open (n=66), which is typical of a young population. The mean age of the adult population is estimated at 30–35 years.

Epigenetic Traits

Differences in the frequency of epigenetic traits are used to distinguish between populations (Hauser and DeStefano 1989). A battery of 11 epigenetic traits was studied in the Mamilla sample. The frequency of the traits is compared between the Mamilla population and contemporary Jewish (‘En Gedi) and Bedouin (Rehovot-in-the-Negev) populations (Nagar 1999:72-85).
A 2 correspondence analysis test shows statistically a significant dependence between population and trait (df=18; 2= 47.21; p=0.0002), i.e. the Mamilla population is distinct from the contemporary Jewish and Bedouin populations. It is characterized mainly by a relatively high frequency of incomplete fusion of the costal element of the transverse process of the axis and low frequency of posterior bridge of the atlas.
This result supports the assumption that the skeletal remains from the Mamilla cave represent the Christian population of Jerusalem.



All the bones of the sample were checked for pathologies; pathological cases were sorted out and diagnosed. Most of the pathologies within the Mamilla skeletal sample are those commonly found in historical populations: trauma, non-specific infections, arthritic lesions of various kinds and various congenital defects.  Nonetheless, extremely rare pathology of the lower jaw was found.  A selective description of the studied pathologies is hereby presented.


Two cases of advanced tuberculosis are evident in the skeletal remains.

1.  Pulmonary tuberculosis is represented by a 6 cm wide fibrotic calcified soft tissue, identified as a calcified pleura.

2.  Tuberculosis of the elbow joint is represented by an inflammated right ulna.  The olecranon process manifests hypertrophic growth, active periostitis, and cloacae.  X-ray photograph of the bone ruled out the possibility of trauma.

These types of tubercular lesions are relatively common (Ortner and Putschar 1985:141–176).  Moreover, non-specific inflammation of right and left fifth metacarpal bones might also be the result of tuberculosis (X-ray photograph of the bones rules out the possibility of trauma).

Tuberculosis is a contagious disease affecting the bones in only 5–7% of the infected individuals (Steinbock 1976).  Therefore, we assume that tuberculosis was widespread in this population.  This finding is significant, since tuberculosis is considered to be rare among Jewish populations (Fishberg 1916:245).  Arensburg et al. (1985:1975) found no cases of tuberculosis in Jewish skeletal remains from the Hellenistic, Roman, and Byzantine periods. This again supports the assumption that the skeletal remains from the Mamilla cave represent a Christian population.

Lesions of the Vertebral Column

The main findings are:

1.  Relatively low occurrence of osteophytes and arthritic lesions.  The occurrence of adult vertebrae having marked osteophytosis is approximately half that reported for Khan el-Ahmar, a nearby contemporary Christian site (Hershkovitz et al. 1993:378). This necessarily indicates a relatively young population (Stewart 1958:144-151; Waldron 1991:109).

2.  The conclusive prevalence of pathologies, anomalies and discrete traits in the vertebral column is presented by three comprehensive graphs (Nagar, Taitz, and Reich 1999:Figs. 5–7).  This data can serve as a basis for future studies of vertebral material in other populations.

Hemifacial Microsomia

An extremely rare condition of bilateral aplasia (agenesis) of the mandibular condyles was detected (Fig. 6).  The pathology is diagnosed as hemifacial microsomia (Nagar and Arensburg 2000:135-139).  The mandible, belonging to an elderly adult individual, is extremely short.  The mandibular foramen, the entrance for the inferior alveolar vessels, is extremely wide; the insertion areas of the adductor muscles (masseter, temporalis, medial pterygoid) are hypertrophied.  The presence of a medial pterygoid tubercle in this specimen, a trait characteristic of Neandertals, is of great interest (Nagar and Arensburg 2000:135-139).


The Mamilla skeletal remains are in a poor state of preservation. A sample of relatively small intact bones (teeth, patellae, vertebrae, hand and feet bones) has been studied, showing the following results:

1.  The Mamilla population is rather young compared with typical cemetery populations, as evidenced from the rate of tooth attrition, the frequency of age-related pathologies of the vertebral column, and closure stages of the jugular process of the temporal bone.

2.  The age distribution of the Mamilla skeletal remains, based upon tooth attrition stages, is expected in a population that perished in a fatal event. The age pattern of the population is characteristic of historical populations, as well as recent populations in developing countries.  However, infant mortality data is absent and life expectancy at birth cannot be estimated.

3.  Sex ratio in the adult population was calculated as 38/100 males/females.  The predominance of female versus male is explained by the circumstances in which men fell in battle, whereas women stayed behind and were indiscriminately slain.

4.  The kinds and patterns of pathologies, together with the specific distribution of epigenetic traits, distinguishes the Mamilla population from contemporary Jewish and Bedouin populations.

5.  An extremely rare pathology, hemifacial microsomia, is reported and discussed for the first time in archaeological populations (Nagar and Arensburg 2000:135-139).

The anthropological study of the skeletal remains from the Mamilla cave sustains the theory that this was the Christian population of Jerusalem that was massacred. Despite the fragmentary nature of the bones, two extensive studies of the skeletal sample were already completed (Nagar, Taitz, and Reich 1999; Nagar and Arensburg 2000). This excavation provided us with a large and important database of relative frequencies of pathologies and epigenetic traits, which is its major contribution to anthropological research in Israel.


Arensburg B., Goldstein M.S. and Rak Y. 1985. Observations on the Pathology of the Jewish Population in Israel (100 B.C. to 699 C.E.). Koroth  9:73–83.
Bass W.M. 1987. Human Osteology (3rd ed.). Columbia.
Coale A.J. and Demeny P. 1966. Regional Model Life Tables and Stable Populations. Princeton.
Fishberg M. 1916.  Consumption (Tuberculosis). In I. Singer ed. The Jewish Encyclopedia.  New-York. Pp. 245–248.
Hauser G. and De Stefano G.F. 1989. Epigenetic Variants of the Human Skull.  Stuttgart.
Hershkovitz I., Latimer B., Dutour O., Jellema L.M., Wish-Baratz S., Rothschild C. and Rothschild B.M. 1997. The Elusive Petrooccipital Articulation. American Journal of Physical Anthropology 103:365–373.
Hershkovitz I., Yakar R., Taitz C., Wish-Baratz S., Pinhasov A. and Ring B. 1993. The Human Remains from the Byzantine Monastery at Khan el-Ahmar. Liber Annuus 43:373–385.
Hillson S. 1986. Teeth. Cambridge.
Lovejoy C.O. 1985. Dental Wear in the Libben Population: Its Functional Pattern and Role in the Determination of Adult Skeletal Age at Death. American Journal of Physical Anthropology 68:47–56.
Milik J.T. 1961. La Topographie de Jérusalem vers la fin de L'Époque Byzantine. Melanges de l'Université Saint Joseph 37:127–189.
Nagar Y. 1999. The Anthropology of Rehovot-in-the-Negev as an Example of a Large Byzantine Settlement in the Negev. Ph.D. Dissertation. Tel Aviv University. Tel Aviv (Hebrew).
Nagar Y. and Arensburg A. 2000. Bilateral Aplasia of the Condyles in a Byzantine Mandible from Israel. American Journal of Physical Anthropology 111:135–139.
Nagar Y., Taitz C. and Reich R. 1999. What can we make of these fragments? Excavation at ‘Mamilla’ cave, Byzantine period, Jerusalem. International Journal of Osteoarchaeology 9:29–38.
Nathan H. 1962. Osteophytes of the Vertebral Column. An Anatomical Study of their Development According to Age, Race and Sex with Considerations as to their Etiology and Significance. Journal of Bone and Joint Surgery 44:243–268.
Ortner J.D. and Putschar W.G.J. 1985. Identification of Pathological Conditions in Human Skeletal Remains (2nd ed.). Washington.
Reich R. 1994.  The Ancient Burial Ground in the Mamilla Neighborhood, Jerusalem. In H. Geva ed. Ancient Jerusalem Revealed. Jerusalem. Pp. 111–118.
Reich R. 1996.  God knows their Names. Mass Christian grave revealed in Jerusalem. Biblical Archaeology Review 22:26–33.
Steinbock R.T. 1976. Paleopathological Diagnosis and Interpretation. Springfield.
Stewart T.D. 1958. The Rate of Development of Vertebral Osteoarthritis in American Whites and its Significance in Skeletal Age Identification. The Leech 28:144–151.
Swedlund A.C. and Armelagos G.J. 1976. Demographic Anthropology. Dubuque, IA.
Waldron T. 1991. The Prevalence of, and the Relationship between Some Spinal Diseases in a Human skeletal Population from London. International Journal of Osteoarchaeology 1:103–110.