International Journal oj Osteoarchaeology, Vol. 7: 221—229 (1997) Infant Taphonomy HERVE GUY1, CLAUDE MASSET2 AND CHARLES-ALBERT BAUD3 ^Association pour les Fouilles Archéologiques Nationales, Service Départemental d'Archéologie du Val-d'Oise, France; 2Laboratoire d'Ethnologie Préhistorique, Universitě Paris-1, France; and ^Centre Medical Universitaire, 1211 Geneve 4, Switzerland ABSTRACT In almost all living creatures, in Primates as well as in seventeenth-eighteenth century human populations, a high infant mortality is the rule; therefore, the scarcity of children's bones in cemeteries is suspicious from a demographic point of view. Though possible in some cases, sociological causes appear less important than the peculiar behaviour of infants' bones in the tomb. This paper examines the physico-chemical properties of infants' bones and their consequences for the preservation of archaeological samples; it proposes a new way of approaching distributions at death in the past. © 1997 by John Wiley & Sons, Ltd Int. J. Osteoarchaeol. Vol. 7: 221-229 (1997) No. of Figures: 1 No. of Tables: 6 No. of References: 58 Key words: infant mortality; palaeodemography; bone minerals; taphonomy. The problem We know that cemetery excavation rarely yields numerous remains of very young children. This fact, although ordinary, is none the less surprising in many respects, if only because of the great variety of situations observed. Limiting our study to the remains of children under the age of 1 year, at least in sites where all other age classes are represented, we note that they may oscillate from zero to 25 per cent, and even 33 per cent, of the total buried: Fonyód, in Hungary, is an example of the first case,1 Pines Point, in Arizona, or Lerna, in Argolid, illustrating the second case, respectively 23.82 per cent and 35.9 per cent.3 If we leave aside these extreme cases, Table 1, which shows the data from 10 Hungarian cemeteries dating between the tenth and twelfth centuries, gives an idea of the situations encountered most often. It both expresses the amplitude of current variations and shows that the proportion of children under 1 year of age fluctuates frequently around 5 or 6 per cent. Table 1. Proportion of newborns in some Hungarian cemeteries from the Middle Ages (tenth to twelfth centuries), by decreasing infant mortality. After Acsádi and Nemeskéri4 Total buried Number between Proportion under 1 year (all ages: 0 -co) 0 and 1 year (per cent) Fiad-Kérpuszta 395 72 18.2 Zalavar village 141 19 13.5 Oroszvár 115 9 7.8 Halimba-Cseres 932 62 6.7 Zalavar castle 426 23 5.4 Skékesfehérvár Bikasziget 75 4 5.3 Somogy-Vasas 162 8 4.9 Skékesfehérvár Szárazrét 117 5 4.3 Zsitvabenesenyö 73 3 4.1 Zalavar chapel 177 5 2.8 CCC 1047-482X/97/030221-09$17.50 © 1997 by John Wiley & Sons, Ltd. Received 29 November 1996 Accepted 28 February 1996 222 H. Guy, C. Masset and C.-A. Baud Table 2. Infant mortality in a few 'pre-Jennerian' populations, after different authors (Saint-Laurent-des-Eaux, Auneuil;5 France 1740;6 India;7 Geneva;8 Sweden9). Out of 1000 live births, the proportion deceased beefore the age of 1 (per cent) Saint-Laurent des Eaux (Sologne, France; seventeenth century) France 1740-1749 Auneuil-en-Beauvaisis (France, seventeenth century) India 1901-1911 Geneva 1625-1684 Sweden 1757-1759 32.6 29.6 28.8 28.7 26.4 22.7 With reference to the populations known from historical demography, at least to those who lived prior to the invention of vaccination by Jenner ('pre-Jennerian' populations), such proportions are surprisingly low. Formerly, the recording of infant deaths was not satisfactory, precisely because they were so common place. When we can determine relatively reliable figures, however, these are always high and only exceptionally fall below 25 per cent of live births, as shown by Table 2. After the age of 1, death rates decrease rather sharply, but not enough for half of children to reach adulthood (Table 3). We are no longer used to such high death rates in infancy, which were normal in the past, just as among other living species. We can mention the hundreds of thousands of alevins from certain fishes, but it is more interesting to observe our relatives, the primates, at least those for whom a figure was published. To ensure a valid comparison, we shall partly discard infant mortality sensu stricto and consider mainly death prior to puberty. Considering these figures, which largely integrate old Regime populations, we find it hard to imagine that the scarcity of infant remains in our cemeteries can be a true reflection of a demographic fact. If such were the case, we should have to admit the following: that, in our species, infant and child mortality, as far back as prehistory, had clearly broken away from what can be observed among wild animals,- that this peculiarity had lasted almost unceasingly until the end of the Middle Ages,- that then it had been replaced, between the sixteenth and eighteenth centuries, by a brief episode of the 'wild animals' type, precisely at the moment when parish registers came into use. This latter point is Table 3. Proportions of immature deaths in some species of primates, after different authors (for Macaca;w Pan;u Propithecus (Sifaka);12 pre-Jennerian Homo sapiens: here, France 1740-17496). Deaths per 1000 births During the first Between birth year and puberty Macaca fuscata 600 Pan troglodytes (East 580 Africa) Propithecus verreauxi 390 ? (Madagascar) Macaca mulatta (Caribbee) 300 Pre-Jennerian Homo 300 550 sapiens all the more strange because no author in this otherwise well-documented period seems to have noticed such a demographic seism. Indeed, the change would have been very swift, because the above-mentioned cemetery of Fonyód dates from between the fourteenth and sixteenth centuries.1 In it, children under 5 years old number 2.5 per cent of the total, whereas, a century later, in Geneva, they will represent 44.9 per cent, almost 20 times as many, according to the parish registers.8 A sociological interpretation? Palaeodemographers, confronted by so marked and extended an anomaly, have been divided in opinion. A number, despite the above improbabilities, accepted without argument the scarcity of infants in cemeteries, to the point of calculating on this basis the life expectancy at birth in their cemetery population. Knowing how heavily infant mortality presses on life expectancy at INT. J. OSTEOARCHAEOL, Vol. 7 221-229 (1997) © 1997 by John Wiley & Sons, Ltd. Infant Taphonomy birth, we have to be cautious as regards the quality of the results thus obtained. Others, such as Acsádi and Nemeskéri,4 did not underestimate the difficulty. They refused to admit that children's bones may be worse-preserved in the earth than adults', and they looked for other explanations. Notably, they proposed the shallower depth of children's graves, which would have been more exposed to ploughing. Their demonstration is convincing (4, p. 239), and would have been even more so if applied to a site other than Fiad-Kérpuszta, the least problematic of all their cemeteries (Table 1 above). Besides, when reading their text carefully, we note that a shallow grave entails destruction of remains only in sites damaged by erosion,- consequently, the plough is not involved alone. Their explanation is therefore not applicable for non-eroded sites, or for those sites (the overwhelming majority) where the proportion of infants is far from reaching that of Fiad-Kérpuszta. Infanticide was suggested, and especially the exposure of new-borns. As a high infant mortality was common place, it was also thought that, in the past, a real funeral ceremony would not be organized for infants, at least by non-moneyed people: they no doubt would often get rid of the small remains in secret, as proved by a few findings, for instance from Gailhan13 or from Salleles-d'Aude.14 What is awkward in such theories is their too universal character, suitable for all populations in the world,- it is difficult to believe that all funeral customs would have been so closely related. Moreover, these theories cannot be admitted for medieval populations who punished infanticide and practised the baptism of neonates, because, even though the small bodies could not be interred anywhere else than in hallowed ground, their cemeteries yield as few infant remains as older necropoles. This is the case for the above Hungarian sites. Still more convincing is the Saint-Maclou parish, at Pontoise (France, Val-d'Oise), in the eighteenth century, a site where both a partial excavation of the cemetery and the relevant registers are available.15 In its ground, where bones are well preserved, the rate of infants represents a fifth of what should have been found (see Table 5). Let us not imagine an ad hoc ritual, which would have excluded infants from the communal cemetery or 223 Table 4. Bone mineral and water content (after Vinz19). Age Bone mineral Bone water groups Age content content I 0-2 weeks 64.3 19.8 II 3-5 months 62.9 20.5 III 7-11 months 62.9 19.4 IV 1.5-2.2 years 63.7 16.9 V 4-13 years 66.0 14.2 VI 18-40 years 66.8 12.1 relegated them into some recess of the cemetery. The funeral customs of the Pontoise population living 200 years ago are well known, and we know in particular how they would proceed with the small corpses whose baptism and decease had been duly entered into their parish register. A taphonomic process? All these reflections recall us back unrelentingly to the hypothesis of differential destruction. This conjecture could account for the generality of the phenomenon and for the amplitude of the variation between sites, and even between individuals. More than 40 years ago, Angel believed that infant remains disappeared more readily than adults, he even rectified his palaeodemographic curves according to this conjecture, which he credited, but was unable to demonstrate.316 Others, such as Nemeskéri, as seen above, shrank from the subjective aspect of such a rectification, preferring the supposedly more secure field of raw data from excavation. This is also the case of Moore et at,17 who thought that, in this field, the deficiency of ethnographic data would have made it necessary to take the findings from cemeteries for gospel truth. Lastly, others—who express the current opinion — surprised by the relatively good preservation of infant bones when they are found, cannot imagine that most have been lost without trace: at this juncture, they would expect a majority of small remains to be left, even if in a bad state. This expectation is rather unfounded, but it played an essential role in many anthropologists' reluctance to admit a taphonomic difference to the prejudice of infants. If such a difference does exist, then it almost involves a law of all or © 1997 by John Wiley & Sons, Ltd. INT. J. OSTEOARCHAEOL., Vol. 7 221-229 (1997) 224 H. Guy, C. Masset and C.-A. Baud Ven-110 N = 65 r = 0,32 1,0-, 0,8 0,6 0,4 0,2- 0,0 20 40 60 80 100 Fiad : children alone N = 183 r = 0,65 Ven-l 10 : children alnne N = 31 r = 0,44 1,0- 1,0-, S 0,8 - 1 0,8- 1 °.