AMERICAN JOURNAL OF PHYSICAL ANTHROPOLOGY 87:365-372 (1992)

Reliable Identification of Human Albumin in Ancient Bone Using ELISA and Monoclonal Antibodies C. CATTANEO, K. GELSTHORPE, P. PHILLIPS, AND R.J. SOKOL Regional Blood Transfusion Centre (C.C., K.G., R.J.S.1 and Uniuersity Department of Archaeology and Prehistory (C.C., P.P.), Sheffield, United Kingdom

KEY WORDS

Blood proteins, Archaeology, Immunology, Hae-

matology

ABSTRACT

In order to help reconstruct ancient dietary, domestic, and ritual behaviour, a method was developed to identify the blood protein albumin in ancient skeletal material. This was an inhibition enzyme-linked immunosorbent assay (ELISA) using a monoclonal antibody of IgG class against human albumin. With fresh material, the technique gave consistent and specific results and could detect as little as 10 ng of albumin. Extracts of bone from the English Civil War (A.D. 16441, Mediaeval (A.D. 1100-14001, Early Saxon (A.D. 450-6001, Roman (A.D. 100-2001, Iron Age (ca 400 B.c.) and Bronze Age (2200-1700 B.C. cal.) periods were then tested, samples offresh human and animal bone being included as positive and negative controls, respectively. Albumin was demonstrated in human bone from all periods; there was no evidence of cross-reactivity with animal material. Detection seemed to depend largely on amount of sample and chronological age; other factors, such as physical integrity of the specimen and soil characteristics, appeared t o be less important. Preliminary studies of other ancient skeletal remains showed that animal species could be readily identified and that albumin was probably still detectable in cremated material. It is concluded that our method provides a tool specific and sensitive enough for the reliable identification of the speciesorigin of small fragments of bone (and possibly of blood stains) and will thus allow insight into past behaviour patterns. ELISA may also be suitable for identifying other molecules (such as HLA and ABO) which would help determine racial affiliation and disease predisposition among ancient populations.

Reliable identification of blood proteins in ancient skeletal material and on artifacts such as tools or altars can supply crucial information on past dietary and domestic behaviour, warfare, and ritual activity; it may also shed light on the population genetics and health status of our ancestors. It is of primary importance, though, to prove that any biomolecules surviving in ancient material are unequivocally identified by valid and reliable methods (i.e., specific and sensitive) which include the testing of appropriate fresh and ancient controls. It is also important that ambiguity stemming from interference by external factors is eliminated, especially in the archaeological context where

@

1992 WILEY-LISS,INC

contamination from soil and micro-organisms and the effects of time might complicate assays or interpretation of results. The solution t o these problems and thus the key to such archaeological information can be supplied by current biomedical tests. Blood and, t o a certain degree, other tissues offer a vast range of proteins each capable of giving valuable information. Unfortunately, the more informative proteins present in bone or in blood stains are less abundant

Received December 11,1990; accepted October 15,1991. Address reprint requests to C. Cattaneo, Regional Blood Transfusion Centre, Longley Lane, Sheffield S5 7JN, U.K.

366

C. CATTANEO ET AL

and/or less resistant to post-mortem degradation (Hedges and Wallace, 1978). For example, the identification of biomolecules such as the ABO blood group antigens (Mollison et al., 1987) and the human lymphocyte antigens (HLA) (Bender, 1984) in ancient material could provide an understanding of racial affiliation and predisposition to disease in past human populations. However, survival, extraction, and identification of HLA are problematic (Stastny, 1974; Cattaneo, 1988) (although they have been identified in some instances where preservation conditions were exceptional [Hansen and Gurtler, 198311, whilst ABO antigens are known to suffer alteration by soil microorganisms (Kubo, 1989).Yet there are other proteins, in blood especially, which give less specific information but which are found in larger amounts in animal tissue and apparently survive well in time; examples include haemoglobin (Ascenziet al., 1985; Smith and Wilson, 19901, albumin (Lowenstein, 1981; Loy et al., 1990; Cattaneo et al., 1990) and IgG (Fletcher et al., 1984; Loy and Wood, 1989; Loy et al., 1990). Recently, controversy has arisen over the immunological identification of biomolecules in ancient material. The earliest studies in this field date back to the late 70's and early go's, and although only sporadic and incomplete research has been performed, identification of a variety of substances has now been claimed. The first workers were generally interested in the extraction of protein from ancient human bone; they succeeded in identifying collagen and albumin in fossil bone using radio-immuno assay (RIA) (Lowenstein, 1981) and in detecting haemoglobin in Roman, Iron Age, and Copper Age human bone by an immunoblotting technique (Ascenzi et al., 1985). Enzymelinked immunosorbent assays (ELISA) have recently been used to detect cervid blood on lithic artifacts (Hyland et al., 1990), human haemoglobin in bone (Smith and Wilson, 1990) and human albumin in Australian Palaeolithic rock-art (Loy et al., 1990).However, none of these studies seems to have met all the requirements necessary for this type of work, namely sensitivity, specificity, replicability, safety, and consistent use of adequate control samples. In order to devise a method which would meet these requirements, we chose to work on the extraction and identification of the blood protein albumin from ancient human

bone. Albumin is abundant in blood (40gA circa), and preliminary studies showed that it seems to be resistant to the effects of time and harsh environmental conditions (Cattaneo et al., 1990). The focus on bone, rather than stains on artifacts, resulted from the need to test and refine the method with material which could be safely identified by another means (i.e., skeletal morphology).In this paper, we describe an inhibition ELISA using monoclonal antibodies which gives specific results at a high level of sensitivity: human albumin was detected in small quantities of English Civil War, Mediaeval, Early Saxon, Roman, Iron Age, and Bronze Age bone, the latter dating t o the late third or early second millenium B.C. MATERIALS AND METHODS

Studies were carried out t o identify human albumin in ancient skeletal material from six individuals from the Abingdon Civil A.D.1644 (specimens War cemetery, U.K., ABVR 2054, 2057, 2127,2282, 3070, 3115); two individuals from the Chichester Mediaeval cemetery, U.K., A.D.1100-1400 (specimens CHS 230, 279); twelve individuals from the Berensfield Early Saxon cemetery, U.K., A.D.450-600 (specimens BER 20, 28, 32, 43, 43ii, 108, 109, 126, 127, 141, 152, 164); five individuals from the Casteggio Roman cemetery, Italy, A.D. 100-200 (specimens CAST 11, 12(2), 20, 21(1), 21(2)); two individuals from the Garton Slack Iron Age cemetery, U.K., ca 400 B.C. (specimens GS 1, 4); and four individuals from East Yorkshire Bronze Age burials (Garton Slack and Mortimer collection), U.K., 2200-1700 B.C. cal. (respectively, specimens GS 29-2, 29-5; MORT 23,491. Skeletons from most sites were in fairly good condition (Fig. 1);however, those from the Early Saxon Berensfield cemetery were, in general, poorly preserved.

Preparation of skeletal samples Cores of bone (1.5 cm in diameter) were taken from preferably 3 vertebral bodies of the individual (Fig. 2) using a small power drill. On average circa 10 g of bone dust were obtained, though more was taken if possible. The cores were crushed mechanically, the resulting powder was added to 200 ml of 10% K, EDTA in distilled water (pH lo), and the mixture was subjected to ultrasonic disintegration at 2,000 Hz for 2 min. The mixture was incubated at 4°C for one week, being

367

ALBUMIN IN ANCIENT HUMAN BONE

of biotin-labelled goat anti-mouse-IgG, biotinylated Streptavidin horseradish peroxidase complex, and substrate (5-amino-salycilic acid/H,O,) were added sequentially at 30 min intervals, the plates being washed with phosphate buffered saline (PBS), pH 7.0, between steps. The plates were examined after 60 min. A reduction in colour development a s compared with the fresh animal control extracts represented the presence of human albumin. Controls

Fig. 1. Skull and mandible of specimen GS 1from the Garton Slack Iron Age cemetery (ca 400 B.c.), illustrating the generally good state of skeletal preservation in this study.

shaken a t intervals, and was then centrifuged a t 4,OOOg for 15 min. The supernatant was dialysed through visking tubing against distilledhap water for 24 hr, by which time its pH was between 7 and 8; it was then concentrated with a cellulose triacetate filter (limit m.w. 20,000) to l ml circa. The deposit remaining after centrifugation was incubated with K, EDTA for another week and a second extract was made. Each extract, which could be stored frozen if required, was tested either separately or after pooling and further concentration.

The study was controlled by including material from a variety of species each time testing was carried out. Positive controls were human sera (diluted 1/100) and extracts from fresh human bone. Negative controls were bovine, fetal calf, horse, pig, and sheep sera (diluted 1/100) and extracts from fresh ox, pig, and sheep bone, together with extracts of soil. Individual examples of Mediaeval bone from ox, pig, horse, chicken, and sheep and of Roman ox bone also served as negative controls. Phosphate buffered saline was used a s a system control. In the case of the fresh bone, the extraction procedure was facilitated by freeze-drying the sample before coring and crushing. The sensitivity of the method was 10 ng; it was determined by testing increasing dilutions of a standard human serum containing known amounts of albumin. The ability to give consistent findings was shown by the strong positive results obtained when serum samples from 112 adult humans were tested. Local Ethical Committee approval was given for the use of fresh human material. The samples of bone were obtained during hip or knee replacement surgery; informed consent was obtained from each donor and anonymity preserved. RESULTS

The results of the studies to detect albumin in ancient human bone are given in Table 1, together with those for the experimental control specimens; there was no Inhibition ELISA cross-reactivity between human and animal Five microlitres of doubling dilutions (1/ material. An example of a Terasaki plate, 100 to 1/12,800) of a murine anti-human- demonstrating human albumin in Roman albumin monoclonal antibody of IgG class and Bronze Age bone extracts and control (Serotec 321) were incubated with equal reactions, is shown in Figure 3 in order to amounts of extract in Terasaki plates for 60 illustrate both the technique and the quality min; 5 pl of the mixture were transferred to of the results obtained. The reaction scores separate plates which had been coated previ- (Table 1 and Figure 3) are based on the ously with human serum. After 60 min, 5 ~1 degree of colour inhibition; using the results

368

C. CATTANEO ET AL.

Fig. 2. Two vertebrae showing the small amount of damage caused to the vertebral bodies by sampling. The morphological features likely to be ofmost value to palaeopathologists are untouched; specimen GS 1from the Garton Slack Iron Age cemetery (ca 400 B . c . ) .

of the studies with standard human serum, strong positive (+ +), positive (+), and negative (-) scores correspond to > 80 ng, 10-80 ng, and < 10 ng of albumin, respectively. DISCUSSION

We have succeeded in unequivocably identifying the blood protein albumin in Bronze Age human bone from 2200-1700 B.C. (Fig. 3), as well a s in human skeletons from Civil War (A.D. 1644), Mediaeval A.D. 1100-14001, Early Saxon (A.D. 450-600) Roman (A.D. 100200) and Iron Age (ca 400 B.c.) burials (Table 1). These results were specific and were obtained even though skeletal preservation was poor in several cases; there was no crossreactivity between humans and other species, whether ancient or modern specimens were tested (Table 1). The results suggest that variables such as soil factors (known, for example, to affect the

preservation of ABO blood groups [Kubo, 19891 and haemoglobin [Quarino and Kobilinsky, 19881) and physical integrity of the osseous material do not influence the retrieval and identification of albumin a s strongly as expected. In fact, the sandy gravel of the Berensfield site did not seem to make identification of albumin in the Saxon skeletons any more problematic than in the Roman ones which had been preserved in clay. Even the differences in pH among the various sites did not appear to bear any relation to the survival of albumin. These factors no doubt affect antigen survival and may account for the negative results in some of the Saxon, Roman, and Bronze Age specimens; however, the more crucial variables are thought to be the amount of material used and its chronological age. Only small amounts of the Roman material were available and it is likely that albumin would have

369

ALBUMIN IN ANCIENT HUMAN BONE

TABLE 1. Detection of human albumin in ancient human bone and in control samples using inhibition ELISA and monoclonal antibody Old human bone extracts Specimen reference ABVR 2054 ABVR 2057 ABVR 2127 ABVR 2282 ABVR 3070 ABVR 3115 CHS 230 CHS 279 BER 20 BER 28 BER 32 BER 43 BER 43ii

BER

ioa

BER 109 BER 126 BER 127 BER 141 BER 152 BER 164 CAST 11 CAST 12(2) CAST 20 CAST 21(1) CAST el@) GS 1 GS 4 GS 29-2 GS 29-5 MORT 23 MORT 49

Period Civil War Civil War Civil War Civil W a r Civil War Civil War Mediaeval Mediaeval Early Saxon Early Saxon Early Saxon Early Saxon Early Saxon Early Saxon Early Saxon Early Saxon Early Saxon Early Saxon Early Saxon Early Saxon Roman Roman Roman Roman Roman Iron Age Iron Age Bronze Age Bronze Age Bronze Age Bronze Age

Score'

++ + ++ ++ + + ++ +

+ ++

+ + + ++ -

+++ +

Control sera and extracts' Specimen Human Serum Fresh Human Bone Bovine Serum Fetal Calf Serum Horse Serum Pig Serum Sheep Serum Goat Serum Chicken Serum Fresh Ox Bone Fresh Pig Bone Fresh Sheep Bone Mediaeval Ox Bone Mediaeval Pig Bone Mediaeval Horse Bone Mediaeval Chicken Bone Mediaeval Sheep Bone Soil

Score'

++

++

-

+

lRepresentative findings: as control samples wereincluded every time the assay was carried out, sera, fresh bone, and soil extracts were tested on several occasions, with consistent results. 2Reaction score:++,strongly positive; f,positive; -,negative for the presence of human albumin (correspondingapproximately to >80 ng, 10-80 ng, and < 10 ng, respectively).

been detected in more cases had larger samples been tested. This view is in keeping with the report that the amount of haemoglobin detected in Bronze Age bones was only 1/10 of that found in recent specimens (Ascenzi et al., 1985).It is possible that the mode and depth of burial may also have been relevant, although, as found in another study (Smith and Wilson, 19901, we cannot currently distinguish which are the significant burial conditions affecting antigen survival. In the present work, the one case which seems to illustrate the importance of sample quantity and/or mode of burial is the Bronze Age skeleton (GS 29-2) which gave a positive result on the first extraction (AB.BA, Fig. 3); this particular individual was the only one who had been buried in a coffin, and there was also a larger amount (70 g) of material available for testing. Clearly, many impor-

tant questions are raised and priority must be given to a series of taphonomic experiments designed to elucidate factors affecting antigen survival in buried bone. The method we devised, an inhibition ELISA performed with species-specificmonoclonal antibodies, provided the very high levels of sensitivity and specificity necessary for the present investigation: here, the main objectives were the detection and species identification of blood protein in progressively older material and the exclusion of false results due to cross-reacting contaminating factors. The monoclonal antibody was the key and our goals could be achieved most effectively by using this extremely specific and sensitive reagent and, additionally, by including adequate fresh and ancient control samples whenever the tests were performed; the necessity for these controls cannot be

C. CATTANEO ET AL.

370

Dilutions of Monoclonal Anti-human Albumin

FB.H

Inhibition Results

++

FB.0

FB.S AB.R AB.BA

I

+ +

Fig. 3. Examples of results obtained with bone extracts tested against doubling dilutions ( M O O to 1/12,800)of monoclonal anti-human-albumin. The reduction in colour development ( + +/+) seen with fresh (FB.H), ancient Roman (Al3.R) and Bronze Age (AFLBA) human bone extracts, as compared with fresh ox (FB.0) and sheep ( F B S ) bone extracts (-), denotes the presence of human albumin. The Roman bone (specimen CAST 20) is from the Casteggio Roman cemetery (A.D. 100-200) and the Bronze Age bone (specimen GS 29-2) is from the Garton Slack burial site (2200-1700 B.C. cal.).

overemphasised. We had previously evalu- tages are long shelf-life for the reagents ated the anti-human-albumin t o ensure that and avoidance of radioactive material and it was not recognising an epitope common to complicated measuring equipment. Using other species (Fletcher et al., 1984; Hyland Terasaki plates, as opposed to conventional et al., 1990) or to other serum proteins; we micro-ELISA plates, allowed testing of very had similarly checked that human material small amounts of extract. This was impordid not cross-react with monoclonal anti- tant, both because concentrating the extract bovine-albumin. We are thus confident that could be crucial for the detection of albumin a positive result in this study reliably indi- and also because it allowed several expericated the specimen’s human origin. In order ments to be carried out with comparatively to refine the technique further, we are now small samples (while any remaining could be investigating a method for measuring colour stored frozen for future use with different development so that positive and negative and newer techniques). With concentration, results will be defined quantitatively. In this a doubtful result became positive and a + way, weak positive reactions, where the co- reaction became + +, whereas negative conlour diminution is not sufficient t o be reli- trol samples were unaffected. The advantage ably scored as positive, will not be missed. of using a non-competitive inhibition assay We preferred the use of ELISA with mono- (Clark and Engvall, 1981), rather than the clonal antibodies for this type of archaeolog- more common method whereby the antigen ical investigation. In comparison to methods (in this case ancient albumin) is allowed to using micro-ELISA plates or RIA, our tech- adhere directly onto the plastic plate, is innique offers an easy, cheap, and safe way creased sensitivity. of detecting extremely small amounts (10 ng) Vertebrae were preferred for this study of protein while guaranteeing the maxi- and in an earlier one (Ascenzi et al., 1985)as mum specificity attainable. Other advan- they are highly haemopoietic, even in adults.

ALBUMIN IN ANCIENT HUMAN BONE

Other bones, including femoral and humeral heads, skull, and ribs could have been used. One investigation found more haemoglobin in ancient femora than in vertebrae (Smith and Wilson, 1990) and this was thought t o reflect femora generally being one of the hardest, least porous, and best preserved of remains. Our use of a core of bone from a vertebral body (previously evaluated as showing no interesting pathological features) had the advantage of causing relatively little damage to a valuable archaeological resource (Fig. 21, although it did not yield as much material for testing as crushing the whole bone (Ascenzi et al., 1985). The question arises as to whether albumin is the most suitable blood protein for this work or whether other proteins such as IgG (Fletcher et al., 1984; Loy and Wood, 1989), haemoglobin (Ascenzi et al., 1985; Smith and Wilson, 1990)or even osteocalcin (Ulrichet al., 1987) would be better. Both the present study and other work (Loy et al., 1990) show that species-specificalbumin can be detected in ancient material. Since anti-human-IgG reagents are readily available in our laboratory, a preliminary investigation was carried out during this project to determine the feasibility of detecting IgG; albumin seemed t o be by far the better molecule to look for, which is not surprising if one considers that it is found in blood in much greater quantities (35-50 g/l) compared with IgG (5-15 gA). Haemoglobin may also be a suitable target as large amounts are present in blood (130170 gA). Unfortunately, a suitable monoclonal antibody of IgG class specific for human haemoglobin does not appear to be available at present. Determining the merits of different proteins for archaeological use is clearly an important area for future work. ELISA is an extremely versatile technique and could possibly be used to identify more labile antigens, such as ABO and HLA, with the intention of obtaining information on racial affinity and past disease patterns. Applied to prehistoric and fossil material, it could shed light on domestication and evolution of species; in such studies, measurement of cross-reactivity would be used to construct species relationships and phylogenetic trees (Lowenstein, 1981; Smith and Wilson, 1990); polyclonal antibodies might be better for this purpose. ELISA will also complement the emerging techniques of nucleic acid analysis, particularly the use of polymerase chain reaction (PCR) to amplify DNA sequences in

371

ancient human bone (Hagelberget al., 1989). The recent success of DNA research in ancient material should not cast a shadow over the present work. Although DNA seems to resist harsh agents and time, there are problems concerning the very low yields obtained when cloning ancient DNA (Paabo, 1986), and it has been suggested that the nucleic acids found in ancient mummies may, in fact, be RNA (Venanzi and Rollo, 1990).Also, avoiding contamination during assay is a major concern with the relatively new PCR techniques. These technical problems do not exist with our system. Although the present study was confined to known human material, the method could be extended to other species and be used to determine the origin of unidentifiable fragments of bone and of stains on artifacts. Extracts from unknown specimens would be tested with a range of specific monoclonal antibodies. Our current investigations are progressing along these lines: we have now produced monoclonal antibodies against the albumin of several species and have already detected bovine albumin in ox bone from the Roman period. We are also examining the survival of albumin in blood stained material inhumed under controlled conditions. Of particular interest are the findings of a pilot study testing cremated bone of known species origin. One of two Roman cremations showed the presence of human albumin, as did the positive control, a sample of human serum which had been placed in a bath of boiling water for one hour; there was no cross-reactivity with cremated animal material. It is therefore thought likely that albumin, or a significant epitope, can survive at high temperatures, as has been shown already for osteocalcin (Ulrich et al., 1987). If these preliminary results are confirmed, the possibilities for future archaeological studies are clearly very exciting. In conclusion, the present work offers a reliable and extremely sensitive and specific method for identifying albumin in ancient bone, and shows how this provides a means which will help reconstruct dietary, ritual, and domestic practices of our ancestors. ACKNOWLEDGMENTS

We thank Mr. Tim Allen from the Oxford Archaeological Unit for the Civil War bone, Mr. David Miles of the Oxford Archaeological Unit and Professor Keith Branigan of the University of Sheffield for the Berensfield

372

C. CATTANEO ET AL

human bone, Dr. K. Manchester and Dr. Charlotte Roberts from the University of Bradford for the Chichester mediaeval bone, Dr. L. Castelletti, director of the Museum and Archeobiology Laboratory of Como, and Dr. G. Inzaghi of the Museum of Casteggio gio and the Sovrintendenza of Milan for the Roman material, Mr. A. Foxon, keeper of the Hull East Riding Museum, for the Iron and Bronze Age specimens, Dr. Paul Halstead from the Department of Archaeology and Prehistory, Sheffield, and Mr. David Berg of the Wakefield Museum for the ancient animal control samples, and Mr. C.J.M. Getty and Mr. T. Smith for supplying fresh human bone samples. We also thank Mr. D. Allen of the Department of Medical Illustration, Northern General Hospital, Sheffield, for his expert assistance. LITERATURE CITED Ascenzi A, Brunori M, Citro G, and Zito R (1985) Immunological detection of haemoglobin in bones of ancient Roman times and of Iron and Eneolithic Ages. Proc. Natl. Acad. Sci. U.S.A. 82t7170-7172. Bender K (1984) HLA System, 2nd ed. Frankfurt, Biotest, Biotest Bull. 2 (no. 2). Cattaneo C (1988) Palaeoimmunology: The importance of antigens in archaeology and the feasibility of their retrieval from ancient human skeletal material. MA thesis. University of Sheffield, England. Cattaneo C, Gelsthorpe K, Phillips P, and Sokol RJ (1990)Blood in ancient human bone. Nature 347:339. Clark BR, and Engvall E (1981) Enzyme-linked immunosorbent assay (ELISA): Theoretical and practical aspects. In E.T. Maggio (ed.):Enzyme-Immunoassay. Boca Raton: CRC Press Inc., pp. 168-179. Fletcher SM, Dalton P, and Harris-Smith PW (1984) Species identification of blood and saliva stains by enzyme-linked immunosorbent assay using monoclonal antibody. J. Forensic Sci. 29t67-74.

Hagelberg E, Sykes B, and Hedges R (1989)Ancientbone DNA amplified. Nature 343t485. Hansen HE, and Gurtler H (1983) HLA types ofmummified Eskimo bodies from the 15th century. Am. J . Phys. Anthropol. 61r447-452. Hedges REM, and Wallace CJA (1978) The survival of biochemical information in archaeological bone. J . Archaeol. Sci. 5t377-386. Hyland DC, Tersak JM, Adovasio JM, and Siege1 MI (1990)Identification of the species of origin of residual blood on lithic material. Am. Antiq. 55:104-112. Kubo S (1989) Changes in the specificity of blood groups induced by enzymes from soil fungi. J. Forensic Sci. 34t96-104. Lowenstein JM (1981) Immunological reactions from fossil material. Philos Trans R Soc Lond [Biol]. 292:143-149. Loy TH, Jones R, Nelson DE, Meehan B, Vogel J , Southon J, and Cosgrove R (1990)Accelerator radiocarbon dating of human blood proteins in pigments from Late Pleistocene art sites in Australia. Antiquity 64t110116. Loy TH, and Wood AR (1989) Blood residue analysis a t Cayonu Tepasi, Turkey. J. Field Archaeol. 16r451460. Mollison PL, Engelfriet CP, and Contreras M (1987) Blood Transfusion in Clinical Medicine, 8th ed. Oxford: Blackwell Scientific Publications. Paabo S (1986) Molecular genetic investigation of ancient human remains. Cold Spring Harbor Symp. Quant. Biol. 51t441-446. Quarino L, and Kobilinsky L (1988) Development of a radio-immunoassay technique for the detection of human hemoglobin in dried bloodstains. J. Forensic Sci. 33t1369-1378. Smith PR, and Wilson MT (1990) Detection of haemoglobin in human skeletal remains by ELISA. J. Archaeol. Sci. 17t255-268. Stastny P (1974) HL-A antigens in mummified preColumbian tissues. Science 183:864-866. Ulrich MMW, Perizonius WRK, Spoor CF, Sandberg P, and Vermeer C (1987) Extraction of osteocalcin from fossil bones and teeth. Biochem. Biophys. Res. Commun. 149:713-719. Venanzi FM, and Rollo F (1990) Mummy RNA lasts longer. Nature 343:25-26.