The Late Roman Infant Cemetery near Lugnano in Teverina, Italy: some implications
by
David Soren,
Todd Fenton,
Walter
Birkby
From 1988 to 1992 excavations by the University of Arizona near the town of Lugnano in Teverina in the province of Terni yielded a Roman villa in the ruins of which was placed the largest infant cemetery yet found in Italy. The excavations revealed that the cemetery was installed at one time or over a very short period, perhaps due to an epidemic. A case based on circumstantial evidence was made that the infants died from Plasmodium falciparum malaria, although this is conclusion is now supported by DNA testing by Robert Sallares at the University of Manchester.
In Southern Umbria, about seventy miles north of Rome along the Tiber River, near the town of Lugnano in Tevernia, archaeologists and forensic anthropologists investigated the ruins of a large Roman villa originally constructed at the end of the first century B.C. (Fig.1) (Soren & Borghetti, 1992; Soren & Aylward, 1994; Soren & Soren, 1995). After having fallen into ruin by the third century A.D., the five rooms of the southwest section of the villa, built up against a small hill, were reused as an infant cemetery close to the middle of the fifth century A.D. (Fig. 2, Fig. 3, Fig. 4, Fig. 5, Fig. 6, Fig. 10a, Fig. 10b, Fig. 10c). Forty-seven infants have been discovered interred in these rooms, and more, perhaps as many as a dozen, await recovery.
Evidence suggests that all the bodies were interred at about the same date during a brief period: a day, a week or perhaps a month. The fill used for the burials raised the ground level of the rooms about three meters, and contained many diverse small lenses of earth. The highest most recent strata were composed largely of ash. Pottery finds from this area are relatively uniform, and joins have been found linking sherds from upper levels to those from lower levels.
An individual burial was found in the lowest level of the cemetery, while paired burials are found in more recent levels. The most recent strata contain multiple burials of up to seven infants interred at the same time. This is not a normal burial pattern found in Roman cemeteries; it strongly suggest that a rapidly worsening epidemic which killed a few at first and many more as its severity increased was responsible for these deaths (for the arrangement of Roman cemeteries, see Reese, 1982).
Twenty-two of the infants were premature, and probably aborted fetuses. Eighteen were newborns, and the others were aged up to 5 to 6 months, with the exception of one 2-3 year old. Six burial techniques were used; simple inhumation; burial under or within reused roof cover-tiles from the villa; interment within a gabled triangle of pan tiles; burial inside a crude rectangular “house” of tiles or under fragments of an amphora (transport jar) or within an amphora (Fig. 11, Fig. 12, Fig. 13).
With some exceptions the eldest infants were confined to the three north rooms where they were given the most elaborate tombs (See Table 1 and Fig. 4, Fig. 5). Neonates and premature babies in the simple tombs or uncommemorated inhumations were found in the two south rooms. Partially overlapping the latter tombs was a large refuse dump filled with debris from the villa. It continued south of the infant cemetery and was principally located in Room 9 (Fig.2).
Scattered among the tombs, especially those of the neonates and premature infants, were several finds suggesting that black magic and witchcraft may have been employed to protect the community and deliver it from an epidemic. These finds include a single raven’s talon which can be interpreted as chthonic symbol and talisman against evil (Fig.19; Green, 1992; Virgil, Eclogues IX. 15). Found with another child was the partial skeleton of a toad, an apotropaic and cure-all commonly associated with magic and witchcraft (Fig. 18; Pliny, XXXII. 114; Horace, Sat. 1. 17-24). Finally, a cooking pot was found buried upside down towards the earth, a possible reference to chthonic power (Fig. 16; Greenewalt, 1978). It contained a glass pouring vessel and mammal bone scrap. Another offering with possible magic connections was a bone doll (Fig. 15; Lück 1985; Virgil, Eclogue VIII; Horace, Sat. I. 8). Ritual doll offerings are often found with arms (Gager, 1992) but here the separately made arms and legs are missing. A pair of bronze cauldrons (Fig. 17) stacked one inside the other was found in Room 12.
An unusual feature of the infant burials was the presence of the skeletons of at least twelve puppies less than six months old and at least one immature dog about one year old, all buried near the premature or neonate babies (Fig. 8, Fig. 9). The dogs were studied by Michael Mackinnon and Stanley Olsen. The skeletons of four were nearly intact, having most or all of their skull, mandibles and body elements. However, one puppy had only its head while four others had their upper skull and body pieces but lacked mandibles. Three puppies had mandibles and body elements but no skull. Finally, the immature dog had most of its body but lacked the entire head. One dog which lacked a skull had been intentionally severed above the pelvis, and one of its two mandibles had been buried with each half (Fig. 14). Also found was a single incisor from a mature dog. None of the dogs was found inside an infant tomb or near older infants.
Dogs were found at virtually all levels of the infant cemetery from the top to bottom. None of the puppy bones shows signs of exposure. This suggests that materials within the cemetery were covered fairly quickly and protected from surface destructive processes. This supports our hypothesis that the cemetery was deposited over a relatively short period of time.
The list of complete dogs arranged by room includes:
ROOM 10
1) Loci 1003, 1005, 1007, 1107. Basket Numbers 4629, 4644, 5162, 5174, 5175, 5379, 5623. L mandible, L temporal 5 cranial elements indeterminate but likely part of parietal and frontal, L, R humerus, R radius, L, R scapula, L ulna, L ilium, R femur, L, R tibia (R distal end only), 7 ribs, 1 vertebra fragment. Has cranium and mandible and post-cranial. Age: circa 4 months. Fairly complete.
2) Locus 1008. Basket 5196. R maxilla (no teeth), L, R frontal, L, R parietal, R scapula, R humerus, R radius, R ulna, L, R femur, L, R tibia, 13 ribs. Has cranium and post-cranial, missing mandibles, Ages: 6 weeks-2 months. Fairly complete.
3) Locus 1000. Baskets 4672, 6544. R frontal, R humerus, R femur. Has cranium and post-cranial, missing mandibles and many post-cranial and cranial pieces. Age: 4 months. Probably not a full interment, possibly scattered pieces, perhaps from natural causes.
ROOM 11 and 12
4) Locus 1400. Basket 67342, 7332, 7350. R, L mandibles (R with di2, dc, dpc, dp3, L with dp4), various cranial pieces including R temporal, R zygomatic, L premaxilla with no teeth, L petrasal, R audintory bulla, frontal fragments, parietal fragments (size indeterminate), R, L radius, R L ulna, L ilium, L R femur (proximal ends only), L calcaneus, R metacarpal 2, scrum fragment, 7 rib fragments, 1 phalanx isolated teeth: dc (max.) dp2 (L), dp3 (L). Age: 4 months. Fairly complete.
5) Locus 1405. Basket 7380. R, L maxillae (R has dp2, dp4, dc, L has di2, di3, dc, dp4), L, R frontal, L parietal, R temporal and auditory bulla, 3 miscellaneous skull fragments (side indeterminate), L, R mandibles (L with di2, dp2, dp3, dp4, R with di2, di3, dc, dp4), R radius fragment, R calcaneus Radius and calcaneus too small for this puppy. Age: 4-5 months. Skull only.
6) Locus 1350. Basket 6486, 6951, 7206, L, R mandible with no teeth, R frontal, R cranial fragment (likely frontal), R ulna, R femur, L, R tibia (distal fragment of R tibia). Age: 4 months. Fairly complete but much post cranial missing, not complete puppy.
7) Loci 1350, 1355. Basket 6486, 6954, 6964, 6989, 7206, 7724, 7757. R mandible with dp2, dp3, L scapula, L, R humerus, R radius, L, R ulna, L, R femur, R tibia fragment, 3 vertebrae, R stragalus. Age 4 months. Fairly complete, lacks cranium, has mandible (only right side).
8) Locus 1400. Basket 7325. R mandible, L ulna, L, femur. Age: 4 months. Very fragmentary, possibly scattered disturbed pieces.
9) Locus 1403. Basket 73662. L premaxilla, R fragmentary premaxilla (?), several indeterminate fragmentary cranial pieces, 14 ribs, 20 vertebrae fragments, R, L ischium, L ilium. Age: 1 month. Fairly complete but lacks mandible.
10) Loci 1403, 1406. Baskets 7403, 7362, 7372. R, L mandible (M3 erupted but only permanent canine, M2 both R, L in mandible), R, L humerus, R, L ulna, R radius, 11 ribs, 1 thoracic vertebra with dorsal spine, L metacarpal, L pelvis (unfused), L femur, R, L tibia, R 3 metatasal. Age: 6-7 months. Intentionally severed puppy with front part in 1403 including ribs, upper limbs, right mandible half, then lower part in 1406 with pelvis left mandible.
11) Locus 1304. Basket7675, 7683. R, L parietals, L, frontal, L temporal, R maxilla, with dp3, dp4, 2 miscellaneous skull fragments, L scapula, r, L humerus, R ulna, 25 ribs, 8 vertebrae fragments, R femur, R tibia fragment, scrum, 2 metapodials, Age 4 months. Fairly complete but lacks mandibles.
12) Loci 1300, 1304, 1305. Baskets 6462, 6807, 7651,7661. R maxilla with dp3, dp4, L mandibles with no teeth, R humerus, L ulna, L radius, R femur, L tibia. Age 4 months. Very Fragmentary.
13) B Loci 1304, 1309. Baskets 7619, 7634, 7640, 7651. L humerus, L, R radius, R ulna, R, L pelvis, R femur, R, L calcaneus, L astragalus, 7 vertebrae, 21 ribs, L metacarpals 2 to 5, R metacarpals 2 to 4, L R metatarsals 3 to 5, 5 phalanges. Age: 1 to 1 ¼ years. Post cranial only, lacks cranium mandibles.
Traces of additional fragmentary dog remains were also uncovered in the cemetery. These include the following:
ROOM 10
1) Locus 1005. Baskets 4544, 5102, 5143. Axis (distal epiphysis is unfused), 2 vertebrae (epiphyses unfused), 1 permanent canine tooth (not deciduous). Age: circa 1- 1 ¼ years. Probably disturbed pieces of same individual.
2)Locus 707. Basket 3080. R tibia, R ilium. Age: 4 months.
ROOM 11/12
3) Loci 851, 852. Baskets 4912, 4160. R, L ulna, R femur, L scapula. Age: 4 months.
4) Locuse 954 Baskets 4395, 6515. Cranial fragments, L humerus, metapodial fragment. Age: 4 months.
5) Locus 1403. Basket 7362. R humerus. Age: 5 months.
6) Locus 1410. Basket 7413, 7430. 3 vertebrae fragments, R scapula fragment. Age: 5-6 months.
7) Locus 464. Basket 4922. Permanent canine tooth. Age: 6 months.
8) Locus 465. Basket 2147. L 3 metatarsals, proximal fragment. Age: 6 months.
9) Locus 1352. Basket 6912, 7708. Thoracic vertebra fragment (unfused epiphyses), phalanx 2, (fused). Age: 6 months to under 2 years.
10) Locus 1350. Basket 6951. Permanent canine tooth. Age: 6 months.
11) Locus 1051. Basket 5263. Phalanx 2 (fused epiphyses). Age: 6 months.
12) Locus 1350. Basket 6903. R ulna, proximal end. Age: 4-5 months.
ROOM 15
13) Locus 1303. Basket 6852. Permanent incisor tooth. Probably from adult dog.
14) Locus 1308. Basket 7627. L femur. Age: 4 months.
ROOM 17
15) Locus 1608. Basket 8277. Proximal end of R femur, epiphysis fused. Age: at least 1 ½ years.
16)
Locus 1608. Basket
8255. Proximal end of R metacarpal
5, epiphysis unfused. Age: 6
months.
The Romans often used puppies as a folk remedy. They were killed and buried in the earth as a sacrifice to malevolent infernal deities (Pliney, XXX. 42 and 64). These puppies may have been killed in an effort to bring health to those still living or to purify the area in a magic ritual, perhaps conducted by local magi or sorcerers (Scholtz, 1937; Collins, 1990 and 1992; Pausanias, III. 14, 8-9; Levy, 1991; Green, 1992). The Romans feared aborted, stillborn or short lived infants, and believed their souls could be used by magi to bring evil to the living (Cumont, 1959; Tertullian, De anima 57; Johnston, 1990). The mothers of these infants were also considered polluted (Plutarch, Superstitione 170B; Boswell, 1988).
To what divinities might such magic offerings be made? Certainly not to the traditional Roman goes such as Jupiter, Juno or Minerva, because such practices are unknown in their cults (Scullard, 1981). One possibility is the goddess Hekate, mistress of darkness, whose consorts were puppies and dogs (the hounds of hell). She was worshipped as one who watched over the souls of infants and aided their passage to the afterworld (Johnston, 1990). Other similar demon-dogs whose names are unknown or unfamiliar to us (such as Geneta Mana) may have been invoked in place of Hekate (Plutarch, Quaestiones Romanae LII).
The discoveries at Lugnano suggest that an epidemic, progressing slowly at first and then proceeding with deadly force, drove survivors to witchcraft even at a time when the Roman empire was officially Christian. But what was the nature fo the epidemic? Consultation with Eskild Petersen, Head of Infectious Diseases at the University of Arizona Medical Center, revealed several possible diseases which may cause pregnant women to abort.
A bacterium called listeria, ingested through infected milk, is one possibility. Another is the bacterium brucella, present in contaminated milk or dairy products from infected cows, sheep or goats. Toxoplasma gondii, a parasite sometimes found in undercooked red meat might also cause abortion. But each of these diseases, while they may cause mass deaths and aborted fetuses, tend to cause many deaths at one time. They probably would not cause the pattern of increasing numbers of deaths over time as is found in this cemetery.
Before suggesting a possible cause of the epidemic, it is useful to discuss the burial in detail.
Forty of the individuals (85.11%) are fetal or neonatal. The age-at-death distribution of the recovered skeletons is
shown in Table 1.
|
|
| Table 1 |
The condition of the bone of the infant skeletons ranges from poor to very good. The poor and friable condition of some of the bones is due to the clay matrix in which the burials were found. The clay retains moisture, creating an environment which encourages the rapid deterioration of organic materials. In addition, some of the amphora burials were in poor condition because ground water had pooled in the vessel and the bones were immersed at some point after they had been interred. The good condition of other bones is attributed to the sandy soil in which those burials were found as sand lets ground water percolate, allowing the bones to remain dry.
Because the bones of these individuals are tiny, they are especially
sensitive to deterioration from moisture. As
a result, a few of the skeletons are complete, but most are incomplete.
The bones of the cranial vault, in particular, are rarely well preserved.
Rodent burrowing, nesting and feeding are other agents that affected the
completeness of several of the skeletons.
Due to the young ages of these individuals at death, few of the forty-seven skeletons exhibit any pathological conditions. With the exception of IB 2, 6, 19, 36, 39, and 40A, no pathologies are present because they have not lived long enough to develop skeletal involvement or any systemic diseases.
Only 14.9% (7 of 47) of the individuals are older than neonatal age, and six of these (12.8%) exhibit cortical bone changes indicating the presence of blood discrasias such as an inherited or acquired anemia. The remains from IB 2, 6, 19, 36, 39 and 40A exhibit a pathological condition of cortical bone known as porotic hyperostosis. This manifests itself as newly reactive subperiosteal bone (Fig. 20) or as a “bone-within-a-bone” morphology in the longbones. Cranial involvement of this same porotic condition is seen in IB 2, 6, 19, and 36 (Fig. 21, Fig. 22). The longbones display this condition in IB 2, 6, 19, 39, and 40A (Fig. 23, Fig. 24, Fig. 25).
There are only two other instances of pathology observed in the excavated skeletons. First, the individual from IB 5 has a healed fracture. Second, one of the skeletons with porotic hyperostosis (IB 36) exhibits enamel hypoplasia on an encrypted adult incisor. The resulting frequencies of both these pathologies are very low (1 of 47 = 2.1%).
It is important to point out that the six skeletons exhibiting porotic
hyperostosis are from individuals who are among the oldest recovered at the
site. These six are old enough to
manifest skeletal signs of a systemic disease process.
The skeleton from IB 4 is the only individual of a relatively advanced
age which does not display this pathology.
This 4 to 6 month old infant exhibits pathological conditions in the cranial and postcranial system. Cranially, the parietals and occipital show a marked degree of porotic hyperostosis , seen as areas of porous, sponge-like bone on the external surface of the cranium (Fig. 21). Areas of both spongy hyperostosis and osteoporotic pitting are present (Walker, 1085).
Postcranially, macroscopic analysis reveals only a small lesion on the
internal surface of a left rib. No
pathologices are observed macroscopically on the longbones but their subsequent
radiographic analysis showed a definite expansion of the cortex on the femora,
tibiae and right fibula. This
appears as a “double-walled” or “bone within a bone” morphology (Angel,
1966) where there are two cortical tables separeated by spidery trabeculae
(Angel, 1967).
These remains, aged birth to two months, are unremarkable except for the
presence of a callous formation on the right clavicle, approximately 17mm from
the acromial end, indicative of a healing fracture which was not probably
sustained during birth.
This 5 to 6 month old infant exhibits the pathological conditions porotic hyperostosis. The parietals and occipital appear porotic with diploic expansion, and both spongy hyperostosis and osteoporotic pitting are evident. The diploe is especially swollen in the area of the spheno frontal articulation. The basiocciput also is swollen or expanded.
The longbones were recovered in a fragmentary state and reveal an expanded cortical region and “double-walled” or “bone within a bone” morphology in the diaphysis. Further radiographic analysis confirms the presence of this pathological condition which is also present in the humeri, radii, ulnae, tibiae, fibulae, and the left femur (Figs. 23, Fig. 24, Fig. 25).
Three of the deciduous canines (both maxillary and one mandibular)
exhibit enamel hypoplastic pits. These
pits were formed approximately one month
prior to death.
The cranial and postcranial skeleton of this 4 to 5 month old infant also exhibits evidence of porotic hyperostosis. Cranially, this pathology can be seen on the inner table of the frontal, the left parietal, the left temporal and the occipital. The outer table of the left parietal and left temporal also shows this condition. Expansion of the diploic space is apparent on the frontal.
Initial macroscopic analysis identified endosteal porosity of both tibiae
and the left radius. The ulnae and
fibulae are fragmented, revealing pathological expansion of the diaphyseal
cortical surfaces and the “double-walled” or “bone with a bone”
morphology. Subsequent radiographic
analyses provide further evidence of cortical expansion in the longbones.
The bones of this 2 to 3 year old child exhibit cranial and dental pathologies. The cranial pathology is known as cribra orbitalia, a form of porotic hyperostosis found on the orbital roof (Fig. 26). The condition is strongly expressed bilaterally in this individual.
The dental pathology is enamel hypoplasia.
A pit, observed on the labial surface of the adult mandibular right
lateral incisor, is located 4.5 mm
from the cemento-enamel junction. This
defect occurred at round 2.0 years of age (Goodman & Rose 1990).
The enamel defect may have resulted from physiological perturbation or
“stress” suring the secretory phase of amelogenesis (Goodmand & Rose,
1990). There are many hypotheses
for the etiology of this enamel defect. The
most common causes are thought to be constitutional disturbances brought on by
nutritional deficiencies or disease. The
presence of cribra orbitalia and enamel hypoplasia in this individual is
consistent with either type of stress.
These 4 to 6 month old remains exhibit porotic hyperostosis in the
longbones. Both radii and both
ulnae are noticeably misshapen and have irregular periosteal bone on the
diaphysis. The left humerus is also
mildly affected.
The longbones of this 4 to 6 month old exhibit porotic hyperostotic
lesions. Initial macroscopic
analysis identified endosteal inflammation on the radii and tibiae. Radiographic analyses revealed that the radii, tibiae and the
femora had expansion of the cortex and the “double-walled” or “bone within
a bone” morphology.
Porotic hyperostosis
is a term that was initially introduced by Angel (1966) although the
pathological condition was first described by Welcker (1888).
In infants and children this pathology can affect the skull and the
longbones. It affects the outer
table of the cranium (symmetrical osteoporosis) or the orbit roofs (cribra
orbitalia), and sometimes the inner table of the cranium.
On the outer table, Saul (1997) has differentiated the lesions into an
“active” variety and a “healed” variety.
The “active” is characterized by a spongy appearance, and the
“healed” by smaller or condensed porosities.
The longbones exhibit swelling of the periosteal layer, or a “bone
within a bone” morphology.
It is commonly suggested that the etiology of the porous cortical bone of this condition is anemia (Angel, 1966, 1967, 1977; Stuart-Macadam, 1987):
“These lesions develop as a result of marrow expansion within the bones of the skull that occurs as part of the body’s response to the inadequate oxygen carrying capacity of the blood. This marrow hyperplasia appears to reflect increased red blood cell production as well as an increase in the number of larger, immature red cell precursors” (Walker, 1985).
 |
|
| Table 2 |
It is believed that the six skeletons from the Poggio Gramignano cemetery
which display porotic hyperostosis suffered from just such an anemic condition.
This anemia may have been either genetically inherited or acquired during
life. Differential diagnoses of
paleophological etiologies in skeletal material is difficult, since osteologists
do not have the luxury of diagnosing the disease process during the patient’s
life and then observing their bones after they die.
As a result, care must be taken in assigning specific diagnoses and
etiologies to skeletal pathologies. It
must be remembered that bones responds to diseases in only a limited number of
ways.
Stuart-Macadam & Kent (1992) have discussed in great detail the traditional views on acquired anemia and introduced new perspectives. The have outlined some of the ways in which an individual can become anemic through an acquired iron deficiency. Struart-Macadam (1992) believes that in past populations “porotic hyperostosis is most probably representative of an acquired iron-deficiency.”
“Nutritional stress” is traditionally viewed as a major factor in causing anemia. This is refereed to as dietary iron deficiency anemia caused by a diet that remains deficient in iron over an extended period of time. As a result there is inadequate absorption of iron by the body.
Anemia of chronic disease is thought to be a defense mechanism by the body that produces low iron levels to combat invading microorganisms. Essentially, the body withholds iron from these microorganisms.
A third cause of acquired anemia is thought to be an increase in intestinal microparasitism which is determined by living conditions: “poor sanitation combined with concentrated populations would increase infection levels” (Reinhard, 1992). Reinhard’s work on coprolites correlates the prevalence of pin worms with the amount of microparasitism in a society. Poor sanitation and hygiene can increase microparasitic levels. Microparasites such as hookworm are the most dangerous, and it is believed that prolonged intestinal bleeding is a likely pathogenic cause of anemia.
Finally, during the infant stage of life, the above factors can also be
interrelated in causing an acquired anemia.
A poor diet can also indirectly produce anemia through an increase in
immunological vulnerability, thereby making the infant susceptible to various
pathogens and chronic disease.
Porotic hyperostosis can also result from inherited anemias such as thalassemia (Cooley’s anemia) and sickle cell anemia. These are hemolytic anemias caused by genetic defects in the hemoglobin structure. Ortner and Putschar (1981) discuss how porotic hyperostosis is a result of an increased demand on the red blood cell (RBC) producing marrow of the long bones and the diploe of the cranium. This increased demand for RBCs caused a rapid increase in the dimensions of the medullary cavity with a concomitant “bone-within-a-bone” appearance in cross-section of the long-bones or the “hair-on-end” morphology observed in cranial radiographs. The occurrence of these blood discrasias or anemias is generally believed to be geographically localized, with thalassemia being more prevalent among Mediterranean people and sickle cell anemia more prevalent in African populations.
In the 1960s and 1970s Angle proposed that an inherited anemia (such as thalassemia or sickle cell anemia) was the main factor for the presence of porotic hyperostosis in Mediterranean antiquity. In particular, he suggest a link between porotic hyperostosis, sickle cell anemia, thalassemia and malaria (Angel 1966).
Angel believed that the more severe forms of these anemias, such as thalassemia major, produces the skeletal pathology porotic hyperostosis seen cranially and postcranially. He believed that the milder forms of the anemias, such as thalassemia minor, presumably imparts a partial “protection” from malaria for the individual.
In the preceding sections the evidence from the cemetery has been
presented. The extraordinary nature
of these finds requires that specialists from a broad range of disciplines (Osteology, Archaeology, Classics, Ancient
History, Palynology, Malariology, Medical Biology, etc.) work together to
discover the cause of death of the infants of Poggio Gramignano and the
significance of this unique cemetery. The
following sections will attempt to deal with these problems systematically and
construct an hypothesis for scholars to consider.
One goal of this paper is the differential diagnosis of the porotic hyperostosis in the infants from the Poggio Gramignano cemetery. It is not scientifically possible at this time to state unequivocally which form of anemia caused the porotic hyperostosis observed in the infants. The various causes of anemia can produce similar pathological responses in bone. For example, Moseley (1965) shows that iron-deficiency anemia in late suckling children can duplicate the porotic hyperostosis of sicklemia. But differential diagnosis can be attempted through a joint venture by the physical anthropologist and the archaeologist: “In other words, consideration of the biotic, physical and cultural ecology of the area must be analyzed” (Martin &t al, 1985).
Thus, if the anemia is a product of dietary iron deficiency, then “the interpretation of porotic hyperostosis requires an archaeological reconstruction of diet in order to determine the potential, and actual uses of resources” (Martin & al, 1985). It is critical to note here that dietary iron deficiency can be ruled out for two main reasons. First, forty-six of the forty-seven individuals were either in the womb or being breastfed, so they were still getting an optimum diet from the mother. Second, the amount and variety of nonhuman animal bone (i.e. pigs, goat/sheep and cattle) found in the fill associated with the burials suggests that nutritional stress would not have been a problem for the 2 to 3 year old (IB 36).
 |
 |
| Table 3 | Table 4 |
Additionally, it is believed that skeletal pathologies related to iron-deficiency anemia are uncommon, and usually limited to the skull vault (Lanzkowsky, 1977: Ortner & Putschar, 1981). Mosely (1963; 1966) reports that longbones are not affected in iron deficiency anemia. At the Poggio Gramignano cemetery only the individual in IB 36 could possibly fit this pattern. That individual exhibits cribra orbitalia without any postcranial pathological involvement. The other five individuals have postcranial porotic hyperostosis, indicating an etiology other than a dietary deficiency.
Therefore, the position of the authors is that the porotic hyperostosis present in five of the individuals at the Poggio Gramignano cemetery is caused by an inherited anemia. Radiographic and macroscopic examinations of the bones form the five affected infants strongly suggests that the observed pathologies are more consistent with anemias such as thalassemia or sickle cell. Baker (1964) notes that in individuals with thalassemia, bony changes are visible in the first six months of life.
The study of the skeletons produces a range of possible causes of death,
but does not explain why the infants died.
In order to formulate an hypothesis about the problem, it is first
necessary to consider the character of the cemetery, the literary evidence for
the history of the area, and the nature of the offerings found in the burials
along with the information derived from the skeletal pathologies.
Evidence suggests that skeletons found in the Poggio Gramignano cemetery were interred over a brief period. In the heart of the cemetery, in Room 11 and 12, a single burial was found in the lowest strata while single and paired burials appeared in the intermediate strata and burials of up to seven infants together were found in the uppermost strata. This pattern suggests that a rapidly growing epidemic, perhaps spreading over some days, weeks or a month caused these deaths. A malarial epidemic, which could create such a scenario, would be appropriate to this geographical area, which is characterized by swampy, marshy conditions along the nearby Tiber River and around the banks of the several lakes in the region. Malaria here was eradicated only in 1962, and even today mosquitoes remain a major problem (Kitron & Spielman, 1989).
Two contemporary ancient sources discuss disease in this region of Italy during the mid-fifth century A.D., the date of the infant cemetery as suggested by associated finds. One source is specific with regard to the area affected, the other more general.
In the summer of A.D. 467 a wealthy Gallic politician named Sidonius
Apollinaris traveled down from Ravenna in northern Italy by the Via Flaminia to
meet with the emperor Anthemius in Rome (Sidonius, Ep. 1.5.6-9).
Although the initial part of the trip went well, Sidonius became ill when
be got between Tuscany and southern Umbria, not far from Lugnano.
He later wrote:
”Here
either the wind Atabulus (the sirocco) from Calabria or the pestilential region
of Etruria infected my body which had been suffocated by breathing the air which
is imbibed in poisoned gasps and which alternates sweats and chills.
Meanwhile,
fever and thirst ravaged the deepest recesses of my heart and bones”.
(translation by Frank Romer)
The symptoms described by Sidonius are consistent with malaria. In summer, mosquitoes would have been plentiful all along the Tiber. The river was not well maintained in this period, and its mouth at Ostia was a particular problem with regard to mosquitoes and malaria (Cellia, 1933; Augustine, Confessions V. 103, IX 8; Bruce-Chwatt, 1980). Sidonius’ presence in this part of Umbria within seventeen years of the estimated date of our infant cemetery may be significant because he is already aware of the reputation of Etruria as a region of pestilence.
A second source, the Novellae divi Valentiniani, is a collection of fifth century A.D. laws which includes a discussion of a visit by Pope Leo I to Attila the Hun in northern Italy near Verona in A.D. 452, about the time of our cemetery. Attila was allegedly dissuaded him by citing that new Roman troops were ready to oppose him and that, since the previous year, famine and pestilence had been a grave problem between Verona and Rome (N. Val. 33). Regrettably, the pestilence is not identified, but when considered with Sidonius’ comment it is not unreasonable to theorize that it may have been an epidemic of malaria.
Associated in abundance with the most recent burials in the cemetery are charred seeds believed to be from the shrub periclymenon (honeysuckle) described by Pliny the Elder in the later 70s A.D. as a cure for diseases of the spleen; this plant was also used as a diuretic, an aid for asthmatics and a medicine for women suffering during childbirth or needing help in expelling the afterbirth (Plinny, XXVII.XCIV. 120 Loeb, with commentary). This associated with birth and its difficulties would make the offering of honeysuckle appropriate in the Lugnano cemetery. In addition, the presence of a cure for diseases of the spleen associated with these burials may be significant because splenomegaly (enlargement of the spleen) is one of the most pronounced symptoms of two of the three types of malaria present in Italy in late Rome antiquity (Plasmodium vivax and Plasmodium falciparum)(Markell & Vogue, 1976; Bruce-Chwatt, 1985).
The inclusion of a toad in the burial of a child might indicate that it was a pet, but in light of the other evidence another interpretation is possible (Fig. 18). Pliny notes that toads were used commonly to ward off storms, eye inflections, earache, toothache, or coughs but they were considered by magi the most useful remedy for relieving tertian or quartan fevers (commonly associated with malaria)(Pliny, XXII.49).
It is possible to suggest that a malaria epidemic would best suit the available evidence as a cause for the deaths of the infants from the Lugnano cemetery. No one piece of evidence points definitely to this conclusion, but when considered together, each piece helps towards the construction of a “most likely” scenario.
Of the 3 types of malaria common to Italy in antiquity (Plasmodium malariae, plasmodium vivax and plasmodium falciparum), Plasmodium falciparum is the type which might cause the scenario witnessed by Lugnano. This type is characterized by a few deaths followed quickly by many, with accompanying aborted fetuses and stillbirths common. Lugnano is situated at the eastern edge of the infamous Tuscan material zone, the borders of which up to World War II differed little from those of the zone in Roman times (Fig. 27; Bruce-Chwatt & de Zuleta, 1980). Individuals living within that zone might hope to have developed some natural resistance to the disease. However, Plasmodium falciparum could occasionally or intermittently spread beyond its endemic limits, behaving as an epidemic and devastating unwary individuals in southern Umbria. A warm, short winter and long summer might trigger such an epidemic.
Dr. Mario Coluzzi, Head of the Istituto di Parassitologia at the Univeristy of Rome believes that Plasmodium falciparum reached its full strength in the later Roman empire, particularly in the fifth century across much of Italy, Sicily and Sardinia. This view has been supported by several scholars (Bruce-Chwatt & de Zulueta, 1980):
“Galen, writing in Rome during the second century A.D., described
the various forms of the disease in terms very close to those found in the
Hippocratic texts. Celsus, also
writing in the second century described unmistakably quartan and benign tertian
malaria and …his description of the malignant or pernicious tertian is
extremely suggestive of falciparum malaria.
It thus seems as if the three species of malaria parasites were present in Italy by the second century A.D.; from then onwards the prevalence of the disease seems to have increased and Jones (1907) and later Celli (1925, translation 1933) attributed in great part the decadence of the Roman Empire to the ravages of malaria. Some new arguments (Zulueta, 1973)…tend to support this view”.
Transmission of the disease along the territory touched by the marshy banks of the Tiber must have been common. If the inhabitants of the Lugnano cemetery lived and worked along the Tiber’s nearby banks as is likely, they would have experienced that full force of such an epidemic. The month of July is the time when honeysuckle goes to seed providing the opportunity for the offering of the seeds found in our cemetery. It is also the month when malaria becomes widespread and severe in Italy (July to September was considered the peak season). Moreover, malarial fevers corresponded to the rising and setting of Sirius, the Dog Star (July 3-August 11), suggesting another reason why the dog sacrifices found at the cemetery may have been related to a malaria epidemic. Malaria cannot be said to be responsible for the fall of the Roman Empire. Economics, movements of peoples in Europe, the introduction of plague and other diseases and internal political unrest all had their roles to play. But this new physical evidence from the Lugnano infant cemetery provides an important corolary for the previously known literary/historical information and suggests that malaria may indeed have been a major player in crippling late antique Rome.
As expected our conclusions have caused a firestorm of criticism. Much of this has occurred before we have been able to offer our conclusions formally. So far as we are aware, there have been three major criticisms of the hypotheses presented in this paper. The first, espoused by Dr. Marshall Becker of West Chester State University in a recent television program in the series Archaeology produced by the Learning Channel, argues that the cemetery was not installed in one short period of time but rather over a period of years. The second argument, also espoused by Dr. Becker, is that there is no specific evidence of the presence of malaria in the infant cemetery. The third argument which usually is offered by medical biologists attending our public lectures is that the evidence of porotic hyperostosis found on some of the infants bones does not indicate that malaria was present, or even if it does, there was not time for the lesions to develop on the bones during a short epidemic.
Regarding the first objection, the presence of puppy bones which have joins of individuals from the top to the bottom of the cemetery suggests strongly that the cemetery fill was deposited over a very short period. If the bones had been left lying around on the surface and buried years later, they would have shown traces of exposure and wear. These bones were put under the ground quickly. The pottery dating evidence, the pottery joins from top to bottom in the cemetery, the uniformity of cultic offering all argue for a single, short phase of deposition, and all independent specialists brought in to examine the cemetery agree with this conclusion. Dr. Becker’s argument for a long duration of cemetery use is based on the fact that other infant cemeteries have been interpreted as long duration gravesites. But there is no requirement that every gravesite must be the same and indeed the unique offerings and infant burial clustering at the Lugnano gravesite that something odd is occurring. In our opinion the long duration argument is absolutely untenable when the evidence is examined.
As to the second objection, that malaria cannot be shown to be present, it must be remembered that archaeologists cannot excavate malaria. That does not mean that as archaeologists we must stop trying to explain what has caused the mass burials we uncovered. The circumstantial evidence presented here is designed to develop an hypothesis which explains a series of events otherwise unexplainable, particularly the extraordinary number of aborted fetuses, the increasing numbers of burials, the short period of time for the installation of the cemetery and the uniqueness of its offerings.
Other interpretations are possible and welcomed but those we have seen such as listeria and toxoplasma gondii fail to explain all the data recovered. Circumstantial evidence is, at least in America, legally admissible and the malaria hypothesis is our most educated guess. To dismiss it without offering an alternative seems to us unfair.
The last objection is the most difficult to deal with. The porotic hyperostosis found on the bones of most of the older infants may or may not have anything to do with malaria. It is possible that its presence did result as a response to malaria which was common in the region over a period of time and perhaps a sickly populace was attempting to resist it through defenses produced by their bodies. However, its presence could just as easily have been triggered by a variety of other causes.
We do not know how the Plasmodium falciparum malaria was spreading during this period through Italy or indeed if it may have been a new strain of Plasmodium falciparum mutating or somehow changing as it adapted to its relatively new vector. It may have been that the people of the area who had been exposed to malaria previously were now seeing something new, in epidemic form, to which they could not adapt. All this however is speculative and for the moment fantastic. Until we know more about the validity of Angel’s conclusions, the precise relationship of porotic hyperostosis and malaria, and how Plasmodium falciparum malaria spread through Italy in late antiquity, we will continue to debate what the strange discoveries at Lugnano really mean.