Dec 15, 1990 - the bone in the reworked units has been quantified, using a system of recording the 'zones' ... bones from Runnymede that they would provide.
International Journal of Osteoarckaeology, Vol 7: 73-89 (19 9 1 )
‘Rid Grasse of Bones’: A Taphonomic Study of the Bones from Midden Deposits at the Neolithic and Bronze Age Site of Runnymede, Surrey, England DALE SERJEANTSON* Birkbeck College Centre for Extra-Mural Studies, London ABSTRACT
Runnymede has large samples of Neolithic and Late Bronze Age animal bones, with contrasting preservation conditions in both periods. The bone evidence has been used to interpret the formation of the site deposits. There are few articulated bones, and no joins were found in butchered bone, indicating that the area studied did not contain primary refuse. Various aspects of bone alteration have been analysed: (i) the proportion of bones with very good surface preservation was high in the in situ Neolithic excavation units and the basal Bronze Age midden, but bones in the upper units were mostly eroded. These units are reworked flood deposits. The greater degree of fragmentation of the bone in the reworked units has been quantified, using a system of recording the ‘zones’ present on each bone, which allows calculation of the fraction present. It isalso demonstratedthat the reworked units contain a lower proportion of identified bones and a higher proportion of teeth and jaws than the units with well-preserved bone. (ii) Quantification of canid gnawing shows, unexpectedly, that more was recorded on well-preserved bone. Thus recognition of gnawing depends on bone condition. This also confirms that most of the erosion of the bone surface is a post-depositional phenomenon. The sequence of activities is therefore interpreted as follows: meat was cooked and consumed, and the bones discarded for the dogs. At a later stage, larger bones were picked up and thrown away in the river or midden. Some ethnographic examples of periodic cleaning of farming settlements are cited. Keywords: Animal bones, Taphonomy, Gnawing, Midden deposits, Runnymede, Prehistoric Britain
Introduction For the past 15 years the study of animal bones has focused on methods by which the bones from a settlement can be used to infer subsistence and animal husbandry practices. The problems of interpretation that arise from butchery and bone processing, transport and selective deposition of bones, carnivore gnawing, and differential preservation have been acknowledged, and several illuminating studies ’Present address: Institute of Archaeology, University College London, Department of Human Environment, 3134 Gordon Square, London W C l H OPY. 1047-482x/91/020073-17$08.50 1991 by John Wiley & Sons, Ltd
@
make use of these aspects of the bone assemblage. They have hardly been confronted until now in studies of permanently occupied farming settlements, where the trajectory of discarded bones is commonly more complicated than on sites of hunter-foragers. The very optimistic assertion, made as long ago as 1978, that ‘It is possible to reconstruct the impact such processes . . . have had upon our sample” has not been borne out. At that time notable progress had been made in distinguishing the effects of but we were far from being able to separate the effects of this from post-discard and post depositional damage. Some studies have been Received 15 December 1990 Revised 12 February 1991
D.Serjeantson
74 published that propose objective criteria for interpreting well-preserved and poorly preserved deposits,687but interpretations continue to be put forward of husbandry and on-site activities in which differential preservation is not explicitly taken into account. It was apparent from initial work on the animal bones from Runnymede that they would provide an ideal sample from which to examine various aspects of bone alteration and to look at the effects of differential preservation. Two distinct cultural horizons have been investigated belonging to two main episodes in prehistory: the Middle Neolithic and the Later Bronze Age, and in both periods the bones are abundant, and vary considerably in surface condition. Research excavations were carried out by Dr S. P. Needham for the British Museum between 1984 and 1989.8'9The calcareous silts of the flood plain had sealed the deposits and large quantities of cultural debris, including animal remains, survived. Those from the base of the deposits are in very good condition; in the uppermost layers, bones, and indeed the artefacts, are less well preserved, and geomorphological work has shown that these were flood deposits in which the cultural debris has been reworked. From the 1986 season onwards the author worked on the bones on site each summer while excavation was in progress and at the Birkbeck College Centre for Extra-Mural Studies at other times. To date, ca. 5 3 000 fragments have been examined (ca. 12 OOO identified bones) from the site as a whole, about 35 per cent of the expected total. The study had three main aims. The first was to use the evidence of bone alteration t o interpret processes of site formation; following on from that the information so gained can be used to interpret both on-site activities and broader aspects of animal management at the settlement. This paper reports on results based on detailed analysis of two areas of the site. Needham and Sorensen" have published a study of midden accumulation in a limited area of the Bronze Age deposits using the evidence of pottery breakage and distribution, and the bone study is also intended in part to complement their findings.
Table 1. Runnyrnede:sequence of deposits and dates in the areas studied. Area Unit __
A6 A16 A16 A16 A16 A16 A16 A16 A16
Archaeological deposit
Date (calendar years BC)
B
Refuse in river's edge Late tenth to early ninth century N Flood reworked Eighth to seventh occupation century M Occupation Eighth century Midden (upper) JKL Late ninth to early eighth century I Midden (lower) Late ninth to early eighth century Features H Ninth century G Hiatus CDEF Flood reworked After 3400 occupation 6 Occupation 4000-3400
The study sample The sample studied here is from Area 16 East (A16E) and Area 6 (Figure 1).The archaeological interpretation, based on observations made on site, stratigraphic analysis, and subsequent wnrk on the excavated pottery, suggests that 19 stratigraphic units can be distinguished in A16E (Table I).The lowest (unit B) contained a spread of discarded settlement material of Middle Neolithic date, ca. 4000-3400 calendar years BC, and the layers above (units C-F) contain the same or similar material, which has been reworked t o various degrees. Most material from unit F lies at its base and is thought to be only slightly reworked from B; while C, D and E are more disturbed. The hiatus in occupation continued during a period when floodplain silts covered A16E to a depth of up to 300 mm (unit G). In the Bronze Age, after an episode of standing structures (unit H), the deposits, up to 1 m deep, appear to have formed a midden that contained dumps of bones and pottery and other cultural debris (unit I-L). Above this is more weathered and broken cultural debris deriving from occupation in the vicinity (M), and the uppermost layers (unit N) are disturbed material that has been reworked by flooding. The Bronze Age deposits date from about 900-700 calendar years. BC. For comparison with the dry land deposits of A16E, the bones from one unit of Area 6 were
8
P m
Figure 1. Runnymede Bridge: site plan with excavation areas numbered.
ixcavation and Salvage 1976 -1989
WNNYMEDE BRIDGE, Egham B E R K S H I R E
D. Serjeantson
76 re-examined. The material here was from deposits representing the opening phase of settlement (A6, unit B), a dark charcoal-rich sediment containing well-preserved refuse on the edge of the river, which soon silted up in this 11 area. The complete bone assemblage from A6 was studied as part of the rescue programme and has already been reported.” For the present purpose, bone condition and fragmentation were specifically examined, so that these aspects could be compared directly with material from the dry land assemblage. For this study ca. 7500 bones from A16E and 178 from A6 have been considered in detail. The great majority are from the main domestic food animals: cattle, pig and sheep. The only other species present in the Neolithic are red deer and dog, but in the larger Bronze Age sample, goat, horse, red deer, dog, badger, wild cat, and otter bones are found as well as a small number of bird and fish bones. In this report I have referred to sheep throughout; but a small proportion of distinctive horn cores and other bones of goat have been identified, so a few goats were also present.
Articulated bones and bone joins Conjoins of flint are now regularly used, particularly on sites of hunter-gathers, to demonstrate site activities and links between area. A notable feature of the early phase of the A16 Bronze Age midden was the number of joins that could be made between sherds and the fairly complete pots found.” The significance of articulating bone and conjoins also has been but this work has been successful on sites of hunter-gatherers, where it appeared that occupation was short-lived or seasonal, rather than on sites of settled farmers; though one exception is the interesting butchery and bone joins at the Neolithic cave site at Fontbregoua, F r a n ~ e . ” ” ~ A search has been made for associated bones, that is: articulated bones, pairs, unfused epiphyses found with their shafts, and joins of butchered bones. It can be said at the outset that such finds are in fact very rare.
The potential importance of finds of articulated bones for demonstrating links between contexts and activity areas was recognized from the beginning of the excavation and they were recorded on site and planned. No complete or partial animal skeletons were found in A16E, although some have been found elsewhere at Runnymede in Bronze Age deposits. However, among the other bones in the lowest layers of the Bronze Age midden (unit I, contexts 872/873) (Figure 2), was the partial skeleton of a lamb. Apart from an astragalus and calcaneum found together, it was not in articulation in the ground, but enough of the skeleton was present to suggest that a single animal was represented, a lamb of less than 3 months, based on the fusing distal humeruses and the maxillary dentition, in which M I has begun to erupt through the bone but is unworn. Limb bones were found with their epiphyses and bones have not been gnawed. The find suggests that the complete carcass had been buried elsewhere, and later, after decay, the bones were put in the midden. Other smaller groups of articulated bones from unit I are: three lumbar vertebrae from an older sheep (context 8731, five thoracic sheep vertebrae, the radius and ulna of a calf (894), and a pair of sheep tibias and one astragalus (868).In unit J a pair of sheep metacarpals was present (context 836). Two sets of articulating phalanges, one of sheep and one of pig, were found in the occupation deposits on top of the midden (819, 829, unit M). Only a small numer of unfused bones were found complete with their epiphyses.17 The proportion has been calculated for each of the Table 2 . Numbers of immature long bones of cattle, pig and sheep showing proportion found with epiphyses (EP). Unit
N
M JKL I H
F CDE B
N
N
unfused
with El’
7 108
0
54
1 13 0 1 0
89 4 17 6 27
2
2
Percentage with EP 0.0 1.9 1.9 14.6 0.0
5.9 0.0 7.4
77
'Rid Grasse of Bones'
c
0
-*
a, Q)
Y Lo
c
0 u)
e,
0
n
u)
e,
c
0
13
ma,
2 r'
W -0
E E
W 0
m
w C
e m
D. Serjeantson
78
midden units, and is shown in Table 2. The only unit in which than one or two were found was the base Of the Bronze Age where 13 and 89 unfused shafts were found with epiphyses, of which five are from the partial skeleton described above. Most of the joins are in the same context and the same metre square, but when all the metapodial fragments were examined together (see below), two epiphyses were found from a metatarsal in the same context (873), but adjacent metre squares. From the upper midden (J), the two unfused halves of the metacarpal shaft of a newborn or very young lamb were found in succeeding contexts (836 and 849), a join which helps to confirm a link between these. The percentages for the Neolithic are higher than most of the Bronze Age layers, 7.4 per cent in the undisturbed midden and 5.9 per cent above, but the sample of immature limb bones is small. Much of the evidence discussed below suggests that the bones once diposed of in the midden were then undisturbed, so it is significant for the interpretation of site formation processes that few bones were found complete with epiphyses. Not only were they disarticulated but most had also been lying around long enough for the epiphyses t o have become detached before the bones were buried. This is an important conclusion, because superficial consideration might have suggested that burial was relatively rapid in the base of the midden. Work on other sites has shown that the linking of two parts of a bone broken in butchery or by other means can illuminate use of space within a camp site.13’14This clearly has potential for settled sites, and a pilot study was undertaken with the bones from the Bronze Age midden. The quantity of excavated material was such that examination of the whole assemblage from the midden for joins, even from one trench, was not practicable, so the study was restricted to the metapodials of cattle and sheep, as even small shaft fragments of these are readily recognizable. All fragments from the Bronze Age midden and occupation units were looked at together. Although the study did reveal the parts of immature bones that belonged together, which have been referred to already, no joins were
Table 3. Sheep and cattle metacarpals (MC) and metatarsals (MT) from the Bronze Age midden and occupation (units I, J, K, L and M): numbers and percentages of bones that are complete, chopped, gnawed, probably gnawed. with recent break, with ancient break, and burnt. Sheep MC MT Complete Chopped Gnawed Gnawed? Recent break Ancient break Burnt Total
Cattle MC MT
Total Percent
6 0 8 1 1 3 2
6 2 11 2 0 13 0
6 1 1 0 0 0 0
0 1 2 3 0 2 0
18 16.9 4 2.8 22 31.0 6 8 . 5 1 1 . 4 18 25.4 2 2 . 8
21
34
8
8
71
100
found of broken or chopped metapodials. In Table 3 the different causes of bone damage are set out, and it can be seen that 16.9 per cent of the metapodials of cattle and sheep are complete, (except for the loss of epiphyses), 2.8 per cent are partly burnt, and 31 per cent have unequivocal evidence for gnawing. The damage to a further 8.5 per cent was probably the result of gnawing, and this also may be the cause of the damage t o some of the 25.4 per cent of bones with ancient breaks caused by unknown mechanical factors, of the type that cannot be reconstructed with certainty; they are the sort of breaks that may be made by dogs,18 by crude butchery where the percussion mark is missing, by trampling, or by sediment compaction. Thus we can expect joins only in the 2.8 per cent of bones that were deliberately chopped or smashed, and possibly in the 25.4 per cent with ancient breaks and damage of unknown cause. The relatively sparse findings of bones either with their epiphyses or in articulation argues for a considerable degree of disturbance of the bones before they reached the midden. Given this conclusion, it is perhaps not surprising that few joins of butchered bones were found. This contrasts with the pottery of the lower midden, among which many joins were found. The explanation must lie in the fact that pottery is useless for most purposes after it is broken, so tends to get thrown away quickly. By contrast bones were dispersed during cooking and consumption of the meat, and when the human population had finished with them they were eaten by the dogs. Only after that were they
'Rid Grasse of Bones'
79
Figure3. Examples of boneswith very good surface preservation (unit I,context 876) and eroded bones (unit N, context814). Middle top: radius of sheep or goat with carnivore gnawing; right top: cattle rib with chop at distal end.
D. Serjeantson
80
Surface Condition
RIVER (66)
The condition of the bone surface varies from very good to eroded (Figure 3).19Five classes of surface preservation have been used, and were recorded on all bones identified to anatomical element, other than teeth and horn cores. The results are shown in Figure 4. The preservation categories are:
N (157)
M (617) JKL (591) I(790) H (102)
G (35) CDEF ( 1 1 2 )
n
(168)
I 0
20
,
,
40
60
80
100
Percent Very good
Sample
si7c
Goad
1 7Eroded
~n brackets
Figure 4. Histogram showing proportions of bones in three categories of surface preservation: very good, good, and eroded.
tidied away on to the midden, and that sporadically. It is probable that the failure to find joins is typical of bones from permanent settlements, since it must be rare for such settlements to contain deposits of any size that represent single episodes of discard, which have not been disturbed subsequently by later activity on the site. The problem of looking for joins is clearly very much more complex on large permanently settled sites than on hunting camps.
Bone alteration Three aspects of bone modification have been considered: surface condition, bone fragmentation and bone destruction by gnawing. Butchery also has contributed to bone fragmentation; it is routinely seen on the main cattle bones but less frequently on pig and sheep bones. It is the subject of continuing research and is not discussed further here. Some of the bone is burnt, but the effects are insignificant in the two areas studied here.
Very good preservation. Few macroscopic traces of erosion. Many bones from the river and the Bronze Age midden fall into this category, but only a very few examples of bones from the Neolithic units. Good preservation. The best preserved Neolithic bones fall into this class. Though in many respects the surface is well preserved, and there is little rootlet damage, the bones are more denatured and fossilized than those from the later deposits. In Figure 4 they have been combined with class (iii). Good preservation. Slight erosion of the surface and etching by rootlets. Much of the Bronze Age bone falls into this class. Eroded. The bones are bleached, the surface is cracked and the cortex partly flaked off. Over 90 per cent of the bones from the least disturbed Neolithic deposits (unit B) were in good condition; in the disturbed layers above (units C-F) this figure is 49 per cent. Of the very small number of bones recovered from the floodplain silts deposited during the hiatus (unit G) between the Neolithic and Bronze Age occupation, all but about 5 per cent were eroded. The lowest Bronze Age features contained a small number of bones (unit H), just over half (54.9 per cent) of which are in very good condition. At the bottom of the midden, however, 86.1 per cent of the bones were in very good condition, with the proportion diminishing in the upper midden (units, J, K and L) to just under 50 per cent. Once the midden ceased to accumulate, local activity produced a slower accumulation of background rubbish (unit M), bone condition is mixed with 36.3 per cent eroded, and only 29.7 per cent in the best category. In the disturbed layers above (unit N)
la
b
lb'
lc
fmralhad trocb.nter mjor prox. rd. part o f shmft prox. lat. p a r t of e b f t d t a t . d.part o f shmft diet. lac. part of s h a f t t, iDclclde suprs-coodyloid
8
I
UticUlD~
3 4
5 d i e t . d.put o f . h f t 6 d i s t . ht. p u t of .hft 7 d.mi& of d i s t . uticohtia to L.d* poon for ~ m l p l l ~ 8 lat. aide o f diBt. articohtia t o inclodt poon for u c m @ l ~
for-
ht. facet of pro.. UtLcaIAtioa 3 r i d e o f CrLu tlbt. 4 pror. d r f t . to incllllk
2
1 md. f v c t of prox.
WIA
Q9
uticlrlrtion lateral condyle of d i s t . articlrlation
fasa 7 d.condyle of d i e t .
1 2 3 4 5 6
Figure 5. Examples of bone zones: (a and b) humerus; (c) radius; (d) femur; (e and f) tibia.
articPlrtion 2 lat. s i d e of prox. articlrlrtion 3 prox. d. p u t of . h f t 1 pro. let. put of Mt q include t o p of p m n f o r rsdirs 5 d i e t . d.put of aImft 6 d i e t . ht. p u t of s&ft 7 d.part of d i s m l a r t i c o h t i o n to ilIclde f a c e t f o r a r t i c n l a t i o n with r a d i a l carpal 8 lat. put o f d i a t . a r t i c o h t i o n to i n c l u d e f w t f o r Ir+icPlrtioa with i n t e r r d i a t e urpl
1 d.aide Of prox.
LaDIus
'
I
1 head of h r r r u s 2 lateral t u b e r o s i t y 3 prox. d. p a r t of a h f t L prox. let. p a r t of a h f t t o i n c l u d e del,toid t u b e r o e i t 7 5 d i s t . r d . p a r t of r h f t 6 d i e t . 1st. p a r t o f a h f t t o include 1st. condyloid crest 7 4 . pert of a r t i c u l a t i o n t o include d.condyle 8 1st. p a r t Of a r t i c u l a t i o n t o include lateral condyle
k wup
D. Serjeantson
82
the proportion in very good condition dropped to 5.7 per cent and the proportion of eroded bone was highest, at 38.9 per cent. Of the bones from the river (A6 B) 63.6 per cent were in very good condition, but there was an admixture of 16.6 per cent which are slightly eroded and nearly 20 per cent were eroded. This reflects the mixed origins of the unit B assemblage. The bones with very good condition must have been buried soon after they were discarded, as they lack evidence of weathering. They then remained protected from disturbance and other damage. Those from the river were protected rapidly by accumulating silts and those from A16E by quickly accumulating midden deposits. Those with slightly eroded surfaces also have been protected from trampling and reworking, but have been subject to other damage in the soil, probably because they were nearer the ground surface. As well as being exposed to trampling and other disturbance, the very eroded bones have suffered weathering and damage from agencies such as water percolation, and frost and thaw. The sediments as seen on site and analysed by S. Limbrey have been deposited by overbank flooding, which occurred twice in the sequence, once after the Neolithic material was deposited (units C-F), and again at the end of the Bronze Age occupation. The eroded condition of the bones from these layers fits this interpretation of the sediments.
use variable numbers of zones for each bone, which may provide greater detail of which parts of the bone are present, especially in larger mammals. In this system the use of a standard eight zones for each bone allows more ready comparisons of the proportions of each bone present. The proportion with one-eighth or less ( < I 2 3 per cent), a quarter (25 per cent) and more than a quarter present are set out in Figure 7. The bones from the river are least fragmented. This fits very well with the notion of the river as a ‘toss zone’ into which rubbish was thrown.25Only the larger bones were picked up and thrown into the river (see below), and these often will have been the most complete. Those from the bottom of the midden were also relatively unfragmented, with proportions similar to those from the river. As midden depth decreases they are more fragmented, and in the reworked upper layer the largest number of bones with
