Blue Bridge Lane
&
Fishergate House
Printed from the APC web site: navigation and non-essential images removed.
Please view on-line for full content (URL at end of document).
Excavations at Blue Bridge Lane and Fishergate House encountered deposits dated from the Roman period to the 20th century and recovered a large faunal assemblage. Analysis was conducted on a total of 52,715 fragments: 9937 hand-collected bones, 18,356 fragments recovered by coarse-sieving, 9593 bones from fine mesh sieving and 14,824 fragments from flotation residues. The majority of faunal remains were dated to the Anglian period and analysis of this material showed it to be very similar in character to that from 46-54 Fishergate but also more generally to assemblages derived from sites identified as wics. As such, evidence for external supply within a socially embedded system of food rents was found in the form of a very narrow resource base consisting largely of cattle, sheep and pigs, the meat products from which were parsimoniously extracted. Subtle differences between the assemblages from 46-54 Fishergate and from Blue Bridge Lane may suggest that the latter site was more peripheral, potentially less intensively occupied and could well have been involved in production of woollen goods and more possibly, limited stock rearing, particularly of birds. A large assemblage of fish bones was recovered and found to be very similar in character to that from 46-54 Fishergate, and again could plausibly derive from reallocation of specific food renders coupled with a small amount of opportunistic procurement by the inhabitants themselves. Detailed analysis of the taphonomy of the Anglian pits indicated that bone was often redeposited into a cessy environment after accumulating on middens in an already highly fragmented state.
Analysis of the smaller post-Conquest and early monastic assemblages indicated a greater diversity of bird and fish resources, the former, along with rabbits and hares, likely to imply the presence of high status diners throughout the medieval phases, while the fish bone indicated the development of marine fisheries and a widening market. Provisioning of the Gilbertine priory in Periods 5 and 6 may have been via the institution's rural estates, but from Period 7 there is greater evidence for market produce.
This document reports on the zooarchaeological analysis of approximately 1600 litres of animal bone recovered by hand-collection, coarse-sieving (to 10mm), fine-mesh sieving (to 2mm) and flotation (residue to 1mm) from an archaeological evaluation and excavation undertaken by Field Archaeology Specialists Ltd (FAS) in association with Mike Griffiths Associates on behalf of Rank Leisure. Fieldwork took place between July 2001 and July 2002 following a programme of evaluation between December 2000 and January 2001 on the site of the Mecca Bingo Hall, Blue Bridge Lane, York.
Excavation encountered 574 features and 1285 contexts; zooarchaeological analysis targeted bone from feature fills because these were more likely to contain primary refuse, with the result that material from 328 contexts was recorded. This represented the majority (approximately 70%) of the excavated faunal assemblage, the remainder of which was distributed among c.300 bone-bearing floating contexts, presumed to be secondary in nature. Based upon pottery spot-dates, recorded features and contexts containing faunal material originated from deposits dated to the 2nd to 20th century. To simplify analysis and to produce more statistically significant samples, contexts were collated into period groups as follows: Period 2: Roman (1st to 3rd century), Period 3: Anglian (late 7th to mid-9th century), Period 4: Anglo-Scandinavian (10th to mid-11th century), Periods 5 & 6: Post-conquest/early priory (mid-11th to 13th century), Period 7: monastic industry (14th century), Period 8: late monastic industry- (late-14th to mid-16th century) and Period 9: post-Dissolution/medieval (late 16th to 18th century).
Following the recommendations of the assessment report, the aim was to record bones from well-dated features. Those of Anglian, Anglo-Scandinavian and medieval date were to be subjected to some degree of analysis, in terms of domesticate proportions, age and size dependant upon the availability of statistically significant data for each category within each period, while those of other periods from well-dated features were to be archived. Of particular interest was the comparison of the Blue Bridge Lane assemblage with those from nearby contemporary sites such as 46-54 Fishergate and from Fishergate House as well as the opportunity to use the excavated evidence to test the hypothesis that Anglian deposits relate to an organised trading settlement under royal patronage analogous to the wics identified at such sites as Southampton, Ipswich, and London. While bones of medieval date are certainly not rare, the Blue Bridge Lane assemblage provides an opportunity to examine material from an urban religious institution and also to compare assemblages from different parts of the Gilbertine Priory encountered during the Fishergate excavations.
It is important to note that the recovery strategy at Blue Bridge Lane was as comprehensive as possible. Large volumes of sediment were sampled from targeted features on the basis of feature type, period, context status (whether primary or secondary), organic preservation and recovery of dating evidence. Zooarchaeological remains were systematically collected through the use of four recovery techniques of differing intensity.
Soil samples ranging in volume from 6 to 50 litres (normally 30 litres) were taken from 220 contexts selected on the basis of possibility of organic preservation, the presence of small artefacts or ecofacts (such as beads, cremated bone or fish bone) or the importance of the deposit in terms of primacy or date. This material was washed down in a water recycling flotation (Sir~f) tank to remove the fine silt and clay particles, with the light fraction being washed into a 250 micron sieve and the more dense residue trapped in a 1mm mesh within the flotation tank. Where organic preservation was suspected, between one and ten litres of soil were retained for further analysis. Once dried, the residue was screened into three fractions (more than 5mm, 2 to 5mm and less than 2mm) of which all but the smallest were selectively sorted for both environmental and cultural material. Bone from 163 of the total 220 sampled contexts was subjected to analysis.
A second method was fine-mesh sieving (fms) to 2mm using a pressurised water jet. Samples up to 100L in volume were taken from 84 selected contexts with the residue treated in a similar manner to that from flotation. Bone deriving from the fine-mesh sieving of all 84 contexts was subjected to analysis.
A third method of recovery was coarse-sieving, which was analogous to the dry sieving carried out at 46-54 Fishergate and was applied to selected contexts in order to maximise and to standardise the recovery of artefacts and faunal material; a minimum of 100L of sediment (where available) was washed over a 10mm mesh using a pressurised water jet. Where this 100L represented less than half of a context, additional material was sampled until 50% or, in the case of specific features (such as Anglian pits), 100% of the deposit had been processed. Bone from 167 out of 345 contexts sampled in this manner was subjected to analysis. Any remaining sediment and bones from non-targeted contexts were subjected to hand-collection, whereby all visible bones were gathered and retained. Of 393 contexts yielding hand-collected bone, 207 were analysed.
The extensive programme of recovery, particularly flotation (flot) and fine-mesh sieving, produced an assemblage of c.100,000 fragments of mammal, bird and fish bone, of which only partial analysis was possible. Contexts that produced small amounts of unidentified material were ignored for the purpose of analysis, while only identifiable material from flotation and fine-mesh sieved residues was recorded, with more concise notes made on the remaining material. In addition, where several samples were available from flot or fms, only one was fully recorded with the rest rapidly scanned for taxa not identified in the original sample.
Zooarchaeological remains were recorded using a modified version of the York System computerised database (Barrett et al 2003), with subjective and semi-quantitative notes made on the state of preservation ('excellent', 'good', 'fair' or 'poor'), angularity ('spiky', 'rounded' or 'battered') and colour as well as the degree of fragmentation and the proportions of butchery, burning, gnawing and fresh breakages as expressed in percentage ranges. Percentage ranges for butchery, burning, gnawing and fresh breakages are expressed in words as follows: low (0-10%), moderate (10-20%), high (20-50%) and very high (50%+). Fragmentation is referred to as low (<20% bones measuring 0-5cm, >50% measuring 5-20 cm, some bones >20cm across), moderate (20-50% of fragments measuring 0-5cm, >50% measuring 5-20cm), high (>50% fragments measuring 0-5cm, 20-50% 5-20cm) and very high (>50% fragments measuring 0-5cm across and <20% fragments 5-20cm). Data was manipulated and figures and tables were prepared using Microsoft Excel 2002.
Identifications were made using the FAS and University of York Palaeoecology reference collections and recording followed the Environmental Archaeology Unit (EAU) protocol for recording animal bones (Dobney, Jaques and Johnstone 1999) which, to increase speed of analysis and to maximise the potential of the most informative elements, advocates the recording of a specific suite of 'A bones' using the bone zones of Dobney and Reilly (1988). In addition, to aid determination of the final epiphyseal fusion stage, cervical, thoracic and lumbar vertebrae were also recorded to species if more than 50% of the vertebral body (Zone 1) was present. The remaining elements were not identified to taxon, regardless of completeness. Instead, along with less complete elements, these were identified to anatomic element where possible, and recorded generally as bird, medium mammal 2 (dog, cat or rabbit sized), medium mammal 1 (caprovid, pig and small deer sized), large mammal (cow, horse and large deer sized) or unidentified.
Sheep and goats were distinguished on the basis of the deciduous fourth premolar, distal humerus and tibia, proximal and distal radius, astragalus, calcaneus and the third phalanx according to the criteria of Boessneck (1969), Payne (1985) and Prummel and Frisch (1986).
Tooth-wear patterns for caprovids were recorded and aged using the schemes of Payne (1973, 1987), while those for cattle and pigs used Grant (1982) for tooth-wear but were aged using the system of O'Connor (1988). Mandibles with less complete tooth rows were assigned to age groups on the basis of comparison with more complete ageable mandibles from the assemblage. The same was true of loose third molars (M3) and deciduous fourth premolars (dp4). It was more difficult to assign an age to loose first and second molars, since these are virtually indistinguishable. In an attempt to categorise them, these were measured and compared to teeth in more complete mandibles. Where it was still not possible to assign these teeth to an age group, or where sample size was already sufficient, they were ignored. Horses were aged using the schemes of Levine (1982).
Mammal bones were recorded as 'juvenile' if the epiphysis was unfused and if the epiphysis or metaphysis was spongy with billowing growth surfaces. If the bone was particularly small, then it was described as 'neonatal', although bones described thus could derive from animals several months old.
Epiphyseal fusion followed Silver (1969). In an effort to reduce a bias towards immature bones when both the diaphysis and epiphysis were recorded, in the case of both fused and unfused elements, only the most abundant bone zone was used for fusion figures in each case. In several cases a vertebra might be fusing at one end of the centrum but could still be unfused at the other. In such cases, these vertebrae were recorded as 'fused'
Measurements followed those outlined in the EAU protocol, utilising those illustrated by Von Den Driesch (1976) and Dobney at al (1999).
Butchery was recorded according to the aspect of the affected zone, and the number, direction and type of marks, defined as cuts (shallow and comparatively long), chops (wide, deep and comparatively short) and splits, where the bone had actually been cloven. In addition, a record was made of fragments which while not bearing any visible blade damage, had smooth, bevelled fracture surfaces and were highly likely to have been cloven or broken up while the bone was still fresh (for the removal of marrow for instance). Such bones were categorised as either having wet bone spiral fracture or longitudinal wet bone split (sagittal, coronal or undetermined).
Pathology was recorded according to the aspect of the affected zone, with the extent and nature of the pathology being described.
Several quantification techniques were used in an attempt to tackle biases associated with sample size, recovery strategy and differential preservation of more robust bones and taxa. Raw counts of every identifiable fragment within the confines of the recording protocol, the Number of Individual Specific Parts (NISP), were used for the bones from each period. These figures have the advantage that they produce a large sample, but create problems in that with such heavy fragmentation, particularly in the case of larger bones and easily identified mandible and maxillae constituents, several parts of the same bone can be recorded separately. When this technique is combined with poor preservation and a system of hand-collection, a bias towards larger animals, particularly cattle molars, is inevitable.
A second technique is MNE, the Minimum Number of Elements. This method attempts to cancel out the bias created by heavy fragmentation by counting up the number of diagnostic zones recorded for each element, and using only the most common individual zone, ignoring those to which no zones are allocated (such as maxillae, incisors, premolars and maxillary molars). In effect each bone, regardless of fragmentation and dispersal, can only be counted once and is thus considered 'non-repeatable'. For the purposes of analysis, each counted bone is a composite of a number of fragments from that period, despite the real possibility that every one of those fragments could have derived from a separate animal.
A third quantification technique, Minimum Number of Individuals (MNI), seeks to extrapolate the minimum number of animals present on the basis of the most abundant zone from the most common element modified by body side. This technique attempts to overcome the biases produced by heavily fragmented larger mammal bones but in so doing, greatly reduces the statistical significance of the dataset. In addition, where selected joints of some animals are imported onto sites, the importance of such animals can be over-estimated when compared to those reaching the site intact. Since the data for this method used MNE data, those figures for MNI are actually a count of the most common zone of the most frequent element, regardless of side.
A final technique is that employed by O'Connor on the Anglian remains from 46-54 Fishergate (1991), whereby eight selected elements (mandible, axis, scapula glenoid, distal humerus, proximal metacarpal, pelvic acetabulum, distal tibia and proximal metacarpal) were used for quantification. When used on material from Fishergate House, counts were based on the most abundant zone for each of these elements. This technique has the advantage of concentrating on non-repeatable and robust elements that are large enough to stand a high chance of recovery in a good state of preservation regardless of taxon. The disadvantage is that sample size is reduced below that of MNE.
Bone from a single context dated to the Roman period, being the backfill of flanking ditch F43 was analysed. The condition of bone was not well-preserved and had suffered some degree of surface abrasion.
Bone from a total of 126 contexts deriving from hearth F246 and from 22 pits was analysed. Bone preservation was generally recorded as good, and angularity mostly spiky but many fragments were at the lower range of these categories, verging on fair and rounded respectively and with some degree of surface abrasion, suggesting that many bones had been redeposited from their primary contexts. This was most apparent when the basal deposits and later backfills from most features were compared, the latter being slightly more likely to be rounded or abraded. Pits F184, F225, F353, F402, F408, F413 and F521 were significant in containing neither chewed fish bone nor faecal concreted fragments. Where present in other features, faecally concreted fragments and chewed fish bones were somewhat more common in the lower fills, although not exclusively so. The degree of fragmentation was not consistent between contexts within features but overall was high, with well over 50% of fragments recorded as under 5cm across and very few more than 20cm across. Only within pits F442 and F546 was fragmentation generally more moderate. Carnivore gnawing was mostly very low, affecting less than 1% of fragments in most contexts and never exceeding 5%. There appeared to be no real correlation between the backfill sequence and the intensity of carnivore gnawing. The proportion of burnt bones and the nature of burning (whether calcined, charred or scorched), was similarly variable within features but ranged from 0% to 40%. On average, butchery marks were recorded on 13% of Anglian bones, but there was a great deal of variation within features, varying from 0% to 27%. A highly detailed analysis of the taphonomic and formation processes of the Anglian pits can be found in section 2.8.
Bone was analysed from six contexts from three features, comprising well F46, posthole/pit F393 and pit F460. With the exception of the bones from well F46, the rounded state of which indicated redeposition (along with a number of fragments of Anglian pottery), most bones appeared to be well-preserved and spiky enough to have suffered little post-depositional disturbance. Fragmentation, particularly in F393, tended to be high, while burning and gnawing was quite low, with limited evidence in each feature. Evidence for human faeces was absent, although a bit of dog coprolite had found its way into F393. Butchery was again variable, averaging 12%.
Bone from a total of 51 contexts from 30 features was analysed, including floor F547, scoop F498, posthole F424 and 27 pits, most of which, with the exception of F4, had limited fill sequences compared to those of the Anglian period. Many of these features, apart from F4, F252 and F256 contained relatively small numbers of fragments when compared to earlier periods. While angularity was frequently recorded as spiky, bone preservation varied between fair and good and was not always consistent between contexts within the same feature, while variably preserved bones implying a more heterogeneous origin were recorded in F239 and F357. Surface damage, while present in some features, notably F4, F230, F255, F256 and F457 was not as common as in contexts dated to the Anglian period. Fragmentation, although still high among many contexts containing less well-preserved bones, was overall less severe than the Anglian material, although bones exceeding 20cm in length were still a rarity. The amount of carnivore gnawing was low, and was absent from almost half of the contexts and from 15 entire features (although considering that many of these contained only one or two fills such results may not be of great significance). The proportion of burnt bone was again variable between features, although not inconsistent between contexts within the same feature. Again, this may relate to the small number of contexts within each feature. The evidence for human faeces was more limited, being restricted to a total of eleven features. Chewed fish bones were recovered from primary deposits in F457, F503 and cess pit F396 which also contained faecal concreted bones. However, with this exception and that of F256, faecal concreted bones recovered from secondary deposits in F4, F255 and F395 may represent reworking of earlier material. Overall, butchery marks were observed on about 9% of fragments, but between individual contexts the proportion was still highly variable, ranging between 0% (in almost half of all contexts) and 29%.
Bone was analysed from a total of 61 contexts from 27 features, including cess pit F150, scoop F178, kiln F218, ditches F219 and F474, quarry pit F223, hearth F334 and 20 pits. Of these, only F150, F235 and possibly F219 contained significant amounts of faunal material. Preservation was generally good, occasionally fair, angularity mostly spiky, and surface abrasion was restricted to ten contexts, mostly from ditch F19 or from pit secondary backfills. Bone from secondary backfills in pits F259 C1585 and F397 C1884 was variable in preservation, angularity and colour, implying an extremely heterogeneous origin for much of this material. Fragmentation was moderate or high, but there was no consistency between contexts within the same feature; this, and the variable butchery, burning and gnawing may in part relate to the small numbers of fragments from many Period 7 deposits. Butchered bones accounted for 6% of the recorded sample, but within contexts this figure fluctuated from 0% to c.30%. Burning was less common than in previous periods occurring in only 21 contexts and 13 features, rarely exceeding 10%. Dog gnawing was even less frequent, recorded in small amounts in 17 contexts from ten features. Evidence of the human excrement was again limited, chewed fish bones being recorded from pits F333 and F397 along with cess pit F150. The latter two features contained faecal concreted bone the presence of which was also recorded in pits F218, F259, F401 and F425. A single concreted bone from ditch F219 is likely to have been reworked and the same may be true of some of the bone from F235 which also contained a rather rounded off-cut from antler working.
Bone deriving from 72 contexts and 21 features was subjected to analysis including kiln F58, ditch F208, hearth F329, quarry pits F162 and F215 as well as from 15 pits, a number of which, notably pits F77, F198, F242, F245 and F359 and quarry pit F162 produced significant amounts of bone. Preservation was more frequently good and spiky than fair and rounded, with greater consistency between contexts within many features. Variable material suggesting a more disparate origin was noted in the backfills of kiln F58 C1368 C1947 and pits F77 C1744, F78 C1206, F198 C1384 and F254 C1680. A number of secondary backfills, particularly ditch F208 C1400, and pits F78 C1552, F198 C1443, F245 C1546 and C1548, F252 C1572, F269 C1593, F77 C1738 and C1742 as well as quarry pit F215 C1412, bore some degree of surface abrasion, indicative of movement from a primary context of deposition. The same was true of primary refuse deposit C1569 in F253. Overall fragmentation tended towards moderate and in certain features there was greater consistency between contexts, notably pit F77, quarry pit F162 and pit F220 where fragmentation was low to moderate and in pit quarry pit F215 and pit F254 where fragmentation was high. Dog gnawing was noted in 28 contexts from 15 features but generally in low amounts, rarely exceeding 10%. Of particular interest was pit F242 C1528 where 10% of bones, often those of birds, had been gnawed by cats.
Possible evidence for human faeces was provided by chewed fish bones from the primary fills of F77, F126, F242, and F253 while faecal concreted bones were recorded only from secondary backfills in kiln F58 and pits F77, F211 and C359. Butchery, with a mean of 11% was notably higher than the preceding period but, as usual, was extremely variable between contexts even within the same feature, ranging from 0% to 32%. Burnt bones were recorded in 35 contexts from 17 features ranging from 0% to 20%, but rarely exceeding 10%. Within quarry pit F162 most contexts contained a similar proportion of burnt bones, ranging from 2% to 4% and the backfills of pit F245 were again similar in character.
Bone was analysed from four contexts from pit F88 and three from ditch F200, neither of which contained significant amounts of material. Preservation and angularity were generally recorded as good and spiky, and fragmentation was moderate. Burning and carnivore gnawing were not particularly common and an average of 9% of bones had been butchered. A single bone from ditch F200 C1387 had faecal concretions, but since this was a bone-working off-cut, it is likely to be redeposited from earlier material.
A total of 242 bones were recovered by hand-collection (109 fragments), coarse-sieving (10 frags), fms (121 frags) and flotation (2 frags) of which only 28 could be identified to taxon (Tables 1, 2 and 3). Cattle was predominant, followed by equal numbers of horse and caprovid along with single pig and human bones. Fish were represented by five herring bones.
Almost 35,000 bones from all recovery techniques were dated to the Anglian period of which over 8700 were identified to taxon (about half of which were fish). A total of 3807 fragments were identified as belonging to the main domesticates, which overall suggested proportions of 51% cattle, 37% caprovid and 12% pig. When recovery technique is considered (Figure 1, Table 1.1), the proportion of pig remains fairly stable at around 10% to 12%, but cattle varies from 63% (hand-collected) to 42% (fms) with a concomitant increase in the proportion of caprovid. Use of MNE and O'Connor quantification techniques show an increase in caprovid at the expense of cattle as quantification refinement is increased, with the proportion of pig bone remaining fairly stable at around 11-13% (Figure 2). Using O'Connor counts caprovid bones account for 50% of the main domesticates, cattle 39%, while MNE suggests slightly more even proportions. MNI figures suggest that caprovid bones are as common as those of cattle, but the pig figures, based on 31 calcanei, increase to 23%. There was not a great variety of other taxa, of which horse, with 57 fragments was most important. A total of 36 cat bones were present, including a partial skeleton of a newborn individual from pit F241 C1531. There was a total of 27 fragments of deer, all antler with the exception of the roe deer metacarpal from pit F442 C2011, and with the majority (20 worked cortex fragments) coming from F442 C1951. The remaining five fragments were recovered from separate contexts in pits F273, F351 and F381. While the majority of distinguishable caprovid remains were identified as sheep, 15 bones from nine features were recorded as goat. These included a single horncore, a calcaneus, a metacarpal and two phalanges, but otherwise comprised major limb bones. Seven human bones were recovered from pits F13, F351, F381 and F520 and with the possible exception of the deciduous molar from F13 C1144, and are most likely to be reworked Roman material. Six dog bones included a short and slightly recurved humerus from F353 C1449.
There were a total of 255 bird bones identified to taxon, of which 177 (69%) were domestic fowl (Figure 3). A total of 63 grey goose elements accounted for 25% of the bird assemblage. Less common were wild bird taxa including three bones of barnacle goose, a possible brent goose and two ducks, comprising mallard and shoveler. The single wood pigeon bone could derive from food remains, but black-headed gull, raven and rook/crow are more likely to have been the bodies of scavengers. Two bones each of wren and passerine are again unlikely (but not definitively) to come from food refuse, and could represent birds caught by local cats. Their presence would indicate that at least some trees, shrubs or hedges were growing in the local area. Small vertebrate remains, particularly shrews, but also voles, wood/yellow necked mice and hedgehog suggest the presence of extensive or unkempt vegetation (Figure 6). House mice were also present, likely to be attracted to stored food. Most amphibian remains are likely to be accidentally incorporated into deposits due to pits acting as traps and are not particularly indicative of local environmental conditions..
The fish assemblage was dominated by the bones of eel, 3112 elements accounting for nearly 78% of all identified piscine fragments (Figure 4). Over 500 herring bones represented another 13%, while the remaining 9% consisted of the bones of salmonids (comprising salmon/trout, salmon, pike and smelt), cyprinids (including carp, barbel, bream, roach, rudd, dace and gudgeon), flatfish, and very small numbers of gadidae (cod, whiting, haddock), cartilaginous fish and shads (Figure 5). In each case, bones of the head as well as vertebrae were represented, although in the instance of both eel and herring there seemed to a disproportionate amount of vertebrae to head bones. In the case of eel, the most abundant skull element was the articular, of which 22, a mere 0.7% of the assemblage, were recorded, as opposed to 3001 vertebrae. The most common non-vertebral element was the cleithrum, with 37 examples. Although the figures for this element are not great, they are significantly higher than those for the more ventral skull bones, of the basioccipital and also of the first abdominal vertebrae. It is possible that at least some eels reached the site in a prepared, and possibly preserved state, a process involving head removal, often ventral to the cleithrum. A similar pattern may be true of herrings. The robust and easily recognisable otic bullae accounted for about 4% of all herring bones from the entire Blue Bridge Lane assemblage, but in the case of Period 3, this proportion was only 1%. It seems likely that at least some herrings were provided in a preserved state.
Bones of eel were quite variable in size, representing both large and small specimens. Conversely, bones of pike seemed to indicate small specimens, as did those of the cyprinids, many of which derived from individuals between 10cm and 15cm long. However, pike, salmon/trout, smelt and eel bone frequently showed evidence of having been chewed, while such marks on cyprinids were much less common. In addition, the robust cyprinid pharyngeal bone was over-represented relative to vertebrae, which is likely to relate to taphonomy and to ease of recognition during sample sorting.
A total of 1069 bones were recorded from Anglo-Scandinavian deposits, of which 184 fragments were identified to taxon. The majority of these were of the main domesticates, but data was sufficient only for NISP and MNE quantification (Figure 7). These indicated that the ratios of cattle to caprovid were similar to those of Period 3, with MNE indicating more even numbers. Regardless of quantification technique, there appeared to be an increase in the proportion of pig from around 11% to 16% between Periods 3 and 4. The range of other taxa was extremely limited, confined to two horse bones, a fragment of red deer antler from F393 C1062, a possible fallow deer medial phalanx from F46 C1167, seven chicken bones, and single examples of barnacle goose and herring.
Of 4868 bones recorded from Period 5 and 6, 1245 were identified to taxon, including 498 bones of the main domesticates. Each of the quantification techniques indicate a rise in the importance of cattle at the expense of caprovid (Figure 8), while the proportion of pig ranging from 13% to 17% is very similar to that from Period 4, but still higher than Period 3. There is a much more diverse range of taxa in this period, including game animals such as rabbit, hare, and roe deer. Small numbers of cat, dog and horse are also present. Proportionately, chicken had become less common than in Period 3, although the figures for goose are not dissimilar (see Figure 3). Instead, an array of wild birds is present, comprising a number of duck species, partridge, golden plover, woodcock, as well as a significant number of turdids, including blackbird and thrush as well as passerines. The latter taxa could have been eaten, but it is also possible that they represent birds living locally in trees, hedges and bushes. Small mammal elements identified to taxon comprised individual bones of shrew, wood/yellow necked mouse, house mouse and black rat.
Among the fish bones, of which 654 were identified to taxon, there is a marked increase in the numbers of herring relative to those of eel, and the contribution of other fish, which in Period 3 barely reach 9%, are 34% in Period 4 (Figure 5). The most important of these other fish are the gadidae, rising from 0.3% in Period 3 to 20% in Period 5, and include larger numbers of cod/saithe/pollack (of which only cod was positively identified) and whiting as well as lower numbers of deeper water taxa such as haddock and ling. Flatfish and cartilaginous fish also become more important, while the proportion of salmonids remains fairly similar, but within this family, smelt is only of tertiary importance relative to pike and trout/salmon when in Period 3 it is the most common taxa. The decrease in cyprinid diversity is most likely to relate to the smaller sample size in Period 5, with their bones still accounting for 2% of the fish assemblage. Element distribution suggests that fish were likely to have reached the site intact, with both gadids and herrings supplied with their heads still attached.
A total of 3338 bones were recorded from Period 6 deposits, of which 1560 were identified to taxon. The 361 domesticate bones indicate that the ratios of caprovid and cattle were similar to the preceding periods, with cattle remaining predominant. Regardless of quantification technique there is a marked decrease in the proportion of pig to between 8% and 10%, with even MNI illustrating a drop from 17% to 12% (Figure 9). Other taxa include rabbit and horse with a particular concentration in F219, several cat bones possibly from the same individual distributed among five contexts in ditch F219, and an almost complete skeleton of an adult medium sized dog from C1946, again in ditch F219. The same feature also produced a near complete skeleton of an adult sheep. The profile of bird proportions is very similar to that of Period 5, particularly in the case of passeriformes and turdidae. The main differences are the absence of ducks, replaced instead by wader and snipe and that scavengers, represented by a single corvid bone, make a reappearance. Small mammal bones identified to taxon indicate a slight predominance of house mouse, but voles (including field and bank voles) are also well-represented, while black rat, mole, watervole and apodemus were also identified. The pattern of fish exploitation however is markedly different from any of the preceding periods. A total of 923 herring bones make up 92% of the identified fish assemblage, with eel accounting for another 5%. Of the remaining 3%, gadidae, including, cod, whiting, ling and a number of haddock are the most significant, followed by flatfish, cartilaginous fish, cyprinid, shad and salmoniforme. The distribution of elements implies that herrings were brought to the site whole, and that at least some gadids were also supplied with their heads attached.
Of 7914 fragments, 4888 were identified to taxon. Within the total of 669 main domesticate bones, cattle are slightly more common than caprovid, but compared to previous periods, there is an increase in the proportion of caprovid relative to cattle. In the case of NISP and MNE the two taxa are close to equality, while O'Connor counts suggest an increase in caprovid to 51% and a decline in cattle to 42% (Figure 10). Also apparent is an increase in the proportion of pig, rising to between 11% (O'Connor) and 13% (MNE) and actually doubling to 18% in the case of NISP. Other taxa comprise rabbit, hare, cat, including one part skeleton and the remains of at least one other individual from F253 C1569, an adult dog skull and cervical vertebrae from F162 C1353, horse, roe deer, and a fallow deer humerus from F77 C1743. Of particular interest is a single butchered cetacean vertebra from F360C1810 which is larger than the common porpoise in the palaeoecology reference collection.
Bird bone proportions are not so different to preceding periods, although there is a slight increase in the overall proportion of chicken bones largely at the expense of goose. The proportions of turdids and passerines also decline, a feature that cannot be attributed to less intense recovery in Period 8 (in fact, a greater proportion of Period 8 contexts were subjected to fms and flot when compared to Period 7). A possible explanation might be that their decline may relate to a loss of habitat. Scavengers also seem to increase, with three records of raven and one of red kite. These might again suggest that actual occupation was not particularly intense, perhaps associated with intermittent rubbish dumping. Their incorporation into pits would either suggest deliberate culling by humans, predation by brave carnivores, or natural death which at some point was tidied up. After chicken and goose, the most abundant taxon was rock/stock dove. While these might again represent opportunistic scavengers much as they do in modern cities, it is more likely that they represent animals specifically kept for food. Bones from woodcock, plover, partridge and pheasant continue to indicate that waste from high status tables continued to be deposited.
Period 8 deposits contained fairly even numbers of house mouse, vole and shrew, with black rat, mole, watervole and apodemus also represented in smaller numbers.
The Period 8 fish assemblage is again dominated by herring (3699 fragments = 93%), but this disguises a number of diversions from Period 7 deposits among the less well-represented taxa. The proportion of eel falls to 1.3%, and other freshwater fish such as cyprinids and salmonids are only a tenth as frequent as in Period 7; euryhaline taxa including flatfish and rays are only half as abundant. These fish are replaced by gadidae, the proportion of which more than quadruples, with haddock the predominant taxon. The element distribution certainly suggests that haddock, herrings and eels reached the site with heads attached, as did a number of cod. The lack of head elements from ling and whiting may relate to the relatively small numbers of bones from these taxa. Of particular interest were C1580 and C1590 from pit F126 which contained very little material apart from thousands of herring bones likely to derive from hundreds of individuals. All elements appear to have been represented, there was no crushing associated with having been chewed and defecated and observations made during excavation indicated that many of the bones articulated. It would instead appear that a large number of fish had been dumped intact.
Of 239 fragments, 125 were identified to taxon of which 36 were main domesticate bones. These suggested rather even numbers of cattle and caprovids along with three pig bones. Bird bones were restricted to five chicken and a single goose, fish to 75 herring and one eel bone.
There are a total of 86 mandibular elements which could be categorised into O'Connor age groups, either using his specific criteria, or through means of comparison with more complete tooth rows (Tables 5 and 6). In addition, there were 86 loose first or second mandibular molars which could potentially be fitted into age categories on the basis of their size and the presence and shape of interstitial wear facets, but were used only to support the more reliable data. There are only two ageable elements from Period 2 indicating animals in O'Connor's adult1 and adult2 age groups.
In Period 3 there are 52 ageable mandibular elements indicating animals in each of O'Connor's juvenile to elderly age groups (Figure 11), but within this distribution there are a number of distinct peaks, notably in the elderly category, suggesting that 38% of animals were 8 years old or more when slaughtered. Another fifth of cattle were killed as older adults (5 to 8 years old) while a quarter were killed as subadults (18-36 months) with a more or less even distribution between the first and second subdivisions of this age group. Overall, about a third of animals were killed before reaching adulthood. A brief examination of the 43 loose first or second molars showed a pattern consistent with that of the more easily aged elements with a majority of heavily worn teeth likely to have come from elderly or older adult individuals. Along with a number of teeth at wear stages A to F, plus one unworn, there were seven deciduous fourth premolars (dp4) with wear stages of J and K, likely to have come from the younger animals (immature to subadult2) already identified.
There are a total of 681 elements available for epiphyseal fusion analysis from Period 3 (Figure 12, Table 7). These suggested that about 1% of cattle on the site died at, or shortly after birth, but thereafter a fairly even pattern of culling was indicated, with between 14% and 17% of animals killed off in each of the early, intermediate and late fusing groups. Overall, there is broad agreement between the fusion and mandible data. The figure of 51% of animals surviving beyond the late fusion stage (42-48 months) is not unlike the 63% of mandibles categorised beyond the adult1 age group (over four years) and the smooth curve of the fusion graph is similar to that of the mandibles up to the adult2 stage. Mandibles within adult3 and elderly age groups can be seen as analogous to the final epiphyseal fusion stage (based on fusion of the vertebral centra) which represents individuals between 5 and 7 years old. As such, there is a slight discrepancy as the fusion data suggests that 42% of animals survived beyond about five or six years old whereas in the case of the mandibles, this is closer to 60%. Examination of the wear stages of adult3 third molars suggested animals closer to seven years old than five, but it is possible that cattle in Period 3 were slow to reach skeletal maturity, particularly as the proportion of fused vertebrae is very similar to that of elderly individuals. An alternative might be that these animals had a very coarse diet, leading to more rapid toothwear and animals appearing older than they actually were. Taphonomic factors seem less likely, as fused vertebrae are more likely to survive than unfused, the reverse of the situation here.
Period 4 produced an adult2 and an elderly mandible along with a loose dp4 which at stage J could fit into either immature or subadult1 groups, although a median age of about 18 months might be appropriate. With only 38 elements available for analysis from Period4, there was insufficient data on which to draw any conclusions. The overall pattern seemed broadly similar to that from Period3 however, but there was an absence of animals killed before about 18 months old.
There were 11 ageable mandibles attributed to Periods 5 and 6, which although far from an exhaustive data set, suggest that again about a third of the cattle were killed before reaching adulthood. There is a single juvenile specimen and three more animals were killed as subadults. Slaughter of elderly beasts was still common, but the contribution of adult2 and 3 individuals is proportionally more important than in Period 3. Wear recorded on the ten isolated first or second molars was consistent with the pattern indicated by the more complete elements. Period 5 produced 114 elements suitable for epiphyseal fusion analysis the results of which broadly agreed with those of the mandible data. The figure of 71% survival beyond the intermediate fusion stage (24-36 months) is not dissimilar to survival into adulthood as indicated by the mandibular data (64%). Of particular interest is the complete agreement between fusion and mandibular data regarding an absence of individuals slaughtered between about 3 and 4 years old, i.e., those within the adult1 mandible age group and those between the intermediate and late fusion groups. There is less good correlation between the final fusion stage (67% of animals surviving to skeletal maturity) and of adult3 and elderly mandibles (45%). If the adult2 mandibles are considered, then the figure (64%) is actually very similar, so it is possible that vertebral fusion among Period 5 and 6 cattle started quite early. However, the discrepancy is as likely to arise from the rather small dataset available for this period.
Period 7 deposits produced five mandibles ranging from juvenile to elderly but of nine loose first or second molars, six were at stages H and J, indicating a predominance of older adult animals, while three more, either unworn or lightly worn, are likely to have come from juvenile, immature or younger subadult animals. A total of 70 elements from Period 7 was an inadequate amount of data for reliable analysis, with figures suggesting that roughly 60% of animals reached adulthood while a third survived to be skeletally mature; certainly nothing that would disagree with the equally limited amount of mandible data.
Thirteen mandibles from Period 8 all represent adult individuals and demonstrate a marked predominance of elderly cattle (9 mandibles =69%). Of 19 loose first or second molars, thirteen were heavily worn between stages H and M, consistent with a preponderance of older beasts, but three at stages B or C and two at stage F would represent immature and subadult individuals. The 120 elements from Period 8 available for epiphyseal fusion analysis demonstrate the presence of a small proportion of immature (c.7%) and subadult (c.17%) animals indicated by loose teeth and also showed that a high proportion (71%) reached adulthood. A marked departure from the mandibles was the very low proportion of fused vertebrae, suggesting that only 6% of animals reached skeletal maturity. Considering the high numbers of elderly mandibles and heavily worn loose teeth, this seems an unlikely situation.
There are a total of 98 and 104 mandibular elements attributable to Payne and O'Connor age groups respectively (the discrepancy arising from the need for both M2 and M3 for the assignation of certain Payne ages) (Tables 8, 9 and 10). The majority of these (63 and 65) belonged to Period 3 deposits (Figures 13 and 14). These indicated that caprovids were slaughtered at every stage between the neonatal and elderly age groups but over half were killed before reaching adulthood (about two years old), many between about 12 and 18 months old (O'Connor's subadult1). Another 15% of animals fall into Payne's two to three year old age group, but from the fact that there are two distinct wear stage groups that fit quite neatly into O'Connor's adult1 and adult2 groups it is probable about 5% were killed at the start of their third year and 10% at the end. Another 30% of mandibles fall into O'Connor's adult3 age group representing animals between about four and seven years old. Of these, the Payne age groups would suggest that 17% were killed off towards the end of their fourth year (thus just missing O'Connor's adult2 group) whereupon smaller numbers were slaughtered throughout the remaining age categories with only about 7% surviving beyond eight years old.
There are 633 elements available for epiphyseal fusion analysis from Period 3 (Table 11, Figure 15), the results of which display broad agreement with the tooth-wear analysis. There is a slight bias towards older animals, the epiphyseal data indicating that 57% of caprovids survived into adulthood (the end of the second intermediate fusion stage at c.18 to 30 months) as opposed to the 47% suggested by the mandibles, but this could relate to the greater durability of fused elements. There is an equally small proportion of neonatal elements but the number of animals killed off during their first year (early fusion stage c.10 months) is under-represented compared to the mandibles. Similarly, a quarter of mandibles were of subadult individuals between about one and two years old which, in age terms, straddle both the first intermediate and the earlier fusing elements (distal tibia and metatarsal) of the second intermediate fusion groups. Considering just these elements provides very similar figures to the mandibular data, and again suggests that more of these subadult animals were killed between 12 and 18 months old than those between 18 and 24 months old. There is good agreement between the proportion of mandibles falling into O'Connor's adult3 group (four to seven years old) and of fused vertebrae in the final epiphyseal fusion group (c.four to five years old). The high number of vertebral epiphyses recorded as 'fusing' compared to those actually fused combined with the large proportion of mandibles falling into Payne's three to four year old age group would lend support to the idea that many animals were killed at about four years of age. Epiphyseal fusion also corroborates that few animals were killed between about 36 and 42 months (late fusion stage), roughly equivalent to O'Connor's adult2 age group.
There are seven ageable mandibles from Period 4, representing individuals slaughtered between about 6 months and four years old while a loose first or second molar could have come from a slightly older animal. Three of these mandibles would appear to indicate animals culled towards the end of their fourth year. There are only 26 elements suitable for epiphyseal fusion analysis, but this limited evidence could be used to support the mandibular data, with the 'fusing' rather than 'fused' state of the majority of vertebrae indicating animals likely to have been around four years old.
Periods 5 and 6 also produced seven ageable mandibles ranging from about two months to four years old, with four specimens likely to have been killed towards the end of their fourth year. Four loose first or second molars fit into already identified age categories. The 92 elements available for epiphyseal fusion analysis are hardly an exhaustive data set, the sample size being reflected in the somewhat erratic nature of the fusion graph (Figure 16). This 'blip' in the first intermediate fusion group appears in every other than Period 3 and 4, and is likely to be a combination of sample size, recovery and taphonomy. Overall, the fusion data indicates that a large proportion (almost 80%) of animals survived to adulthood with about a third of animals killed as young adults. Older adults indicated by the presence of fused vertebrae, suggest that over 40% of animals survived to at least four or five years old.
Four ageable mandibles from Period 7 indicate two animals killed at about two years old and two more between four and eight years old. A loose dp4 also indicates the presence of a sheep about one year old while four loose first or second molars were at wear stages consistent with already identified groups. No firm conclusions can be drawn from the total of 46 elements used for epiphyseal fusion analysis, beyond the fact that subadult animals not represented by the mandibles were likely to have been present. The contrast between the figures for the second intermediate and the late fusing groups along with the proportion of 'fusing' compared to 'fused' vertebrae might suggest that many of the animals were slaughtered as young adults, with few exceeding about five years of age.
The 20 mandibles from Period 8 display a range of ages from juvenile (c.two months) to about eight years old, and suggest that a fairly small proportion, about 12%, all juvenile, were culled before reaching adulthood (Figure 15). Of the remainder, even numbers (c.19%) seem to have been killed off towards the end of their third and fourth years while the similarity of wear among those killed in Payne's group G (4-6 years old) might imply a specific cull of 30% of the caprovids at about five years old. The remaining 19% fell into Payne's 6-8 year old age group, and from the similarity of the tooth wear forms, it is tempting to interpret a specific cull of older animals before they became too elderly. Results of epiphyseal fusion analysis based on a total of 103 elements are variably supportive of the toothwear. The juvenile individuals indicated by the mandibles are not reflected in the fusion data, with an absence of unfused bones from the early fusion group, while the admittedly questionable data (based on only 6 elements) from the first intermediate fusion stage would suggest the presence of at least some subadult animals that are otherwise absence from the mandibular data. There is greater agreement concerning the figures for adult animals suggesting that over three quarters of caprovids reached adulthood. Of these, there is good agreement with the mandible data for the proportion of animals aged four years old or more, particularly given that the majority of vertebrae were recorded as 'fusing' rather than 'fused'.
There are a total of 34 pig mandibular elements to which ages can be ascribed, of which 14 are from Period 3 (Tables 12 and 13, Figure 17). These indicated a broad range of ages from neonatal to elderly with very little concentration in any particular age group. Over half the pigs were killed before reaching adulthood, most under the optimal size likely to be indicated by the subadult age groups. The proportion of adults, particularly of older individuals is high. The 196 elements used for epiphyseal fusion analysis tell a rather different story (Table 14, Figure 18). There is initial agreement, with both mandibles and epiphyses suggesting that about 20% of pigs were killed below the age of a year (although the limited number of new born bones compared to mandibles is likely to relate to taphonomic destruction), while about another 40% are killed off by the end of their second year. However, the fusion data suggests that barely 9% of pigs survived beyond about two and a half years (and continued to do so up until at least the final fusion stage of about five years old), the mandibles, more than double this. By implication, about 30% of mandibles should fall into the adult2 age group, but less than half of this proportion were recovered. Given the limited sample size, it is difficult to attribute much significance to these observations, but they are none the less intriguing. There was insufficient material from any other period on which to conduct epiphyseal fusion analysis.
Two mandibular elements from Period 4 indicated adult2 and elderly individuals while the six mandibles from Periods 5 and 6 deposits and three from Period 7 displayed a more typical age distribution, with only one mandible in each period belonging to an adult. The nine mandibles from Period 8 ranged from neonatal to adult, with four classified as adult2, three of which had similar wear patterns and were likely to have been about two years old.
There are a moderate number of horse mandibular elements, but the majority of these comprise isolated teeth meaning that it is rarely possible to identify the exact tooth, let alone ascribe an age to it. Although there were insufficient bones for epiphyseal fusion analysis, with the exception of one or two vertebrae, the majority of epiphyses appeared to be fused. In Period 2 there is an unworn canine from Period 2 F43 C1163, likely to have belonged to an animal about four or five years old. There are a number of ageable elements from Period 3 deposits, many of which suggested the presence of animals between three and nine years old, along with some much older individuals, including a probable three year old from F351 C1763, a seven year old (with the characteristic seven-year hook on the corner incisor) from F13 C1065 and a fusing thoracic vertebra from F353 C1449 likely to belong to a horse of about five or six years old. The same context contained a very heavily worn pair of maxillary molars originating from a much older individual while the unfused thoracic vertebra from F381 C1846 must have come from an animal under five years old. Finally, an incisor from F246 C1543 is likely to have come from a horse about eight years old.
In Period 4 there are three consecutive molars from F46, the slight wear on the third molar suggesting a beast of between four and a half to six years old. The available evidence from Periods 5 and 6 also seems to indicate fairly young individuals, with teeth from F396 C1864 and F4 C1006 and a fusing thoracic vertebra from F457 C1976 suggesting three animals between about three and five years old. In Period 7 there was evidence for a similarly aged animal from F219 C1485 and from F397 C1884 another aged at least nine years old. F200 C1387 allocated to Period 9 contained an interesting mandible with distinctive wear on the anterior surface of the second premolar likely to be the result of bit wear, while the heavily and irregularly worn molars indicated an old animal.
In order to base element analysis on the number of non-repeatable elements, the figures from MNE were used. In addition, because MNE figures take no account of body side and in order to account for the over-representation of multiple elements (such as phalanges and vertebrae) and the under-representation of single elements (atlas and axis) certain elements were subjected to the following modifiers: all axis and atlas counts were doubled, all phalanges were divided by four, all cervical vertebrae divided by 2.5, thoracics divided by 6.5 (7 in the case of pig) and lumbar vertebrae divided by three (3.5 in the case of caprovids).
While there is some degree of fluctuation in the Period 3 assemblage, there is a fairly even distribution of cattle and caprovid elements (Figures 19 and 24, Table 15). Higher proportions of elements such as mandible, scapula, calcaneus, astragalus and the metapodialsls are likely to relate to their greater robustity, particularly when compared to more fragile horncores, femora and vertebrae. This taphonomic influence is also indicated by the slight predominance of the distal tibia and humerus as well as the proximal radius and femur when compared to their less robust opposite ends (most apparent in the case of caprovid bones). The small numbers of caprovid phalanges could relate to the fact that their small size makes them hard to recover by hand-collection, but considering that this method of recovery accounted for less than half of the Anglian material and that the astragalus, an element equally likely to avoid detection, is present in reasonable numbers, it is possible that many phalanges may have been removed with hides. The slightly low count for cattle humeri might be considered somewhat anomalous as it is a robust element and not so small as to be easily missed by either collection strategy, particularly when it is observed that the adjoining elements, the scapula, radius and ulna are all well-represented. Consumption and disposal of whole or near-complete animals is particularly well illustrated by the example of pit F388 C1856 where the presence of an animal burial was recorded in the field, due to the articulated state of much of its contents. Closer examination revealed that while there was evidence for only one skull, at least three cattle were represented, two of similar size (including almost complete vertebral columns and some paired elements) and a third with fused vertebral epiphyses. The distribution of pig elements (Figure 29) shows one or two interesting trends, although as with cattle and caprovids, there is a propensity for greater representation of the more robust and easily recovered elements, the calcaneum in particular. Of interest, especially when compared to the high counts from cattle and caprovid as well as the large numbers normally recovered from archaeological sites, are the low numbers of pig mandibles. Compared to caprovids and cattle, the proportions of vertebrae and pelvic acetabulae are also low. In the case of the vertebrae, this could in part relate to the higher susceptibility to taphonomic destruction of pig vertebrae, greater numbers of which are likely to be unfused but, any post depositional destruction of unfused vertebrae was insufficient to skew the epiphyseal fusion data in the case of any of the main domesticates. In addition, the pelvis is a robust element which should survive in reasonable numbers. When combined with the low metatarsal and phalanx counts, this may illustrate further evidence for the supply of a certain proportion of dressed pig carcasses. The element distribution of the small number of horse bones (Figure 34) was not so different from that of cattle, and was sufficient to suggest that bones from whole animals had been deposited on the site.
There are no element concentrations that might indicate that Period 3 deposits were receiving craft production waste (such as horncores and distal limb elements), although if it were not for the high numbers of scapulae it would be tempting to identify the presence of a high proportion of primary butchery refuse. Overall, it would appear that in the case of cattle and caprovids, bones from the primary and secondary butchery of complete animals and waste from the table are all represented while in the case of pigs, bones from both whole animals and dressed carcasses are likely.
The amount of material available for analysis from Period 4 was too limited to be truly informative (Table 16, Figures 20, 25 and 30), in the case of cattle throwing up such oddities as five times the number of calcanei as astragali and an absence of metacarpals. The relatively large proportions of cattle mandibles, atlas, calcanei and even the normally easily destructible horncores might imply a certain concentration of primary butchery waste, but the counts for sheep (with the exception of the mandible) and pig implied the presence of more complete animals.
As in Period 3, the Period 5 and 6 cattle element distribution is more likely to be informative of fairly harsh taphonomic processes with a predominance of the more robust portions of bones (Table 17, figures 21, 26, and 31). Again it would appear that material from both butchery and kitchen waste is represented. In the case of sheep, there would appear to be a slight bias towards the limbs, particularly when compared to the mandible the counts for which are low compared to other periods, but also to the vertebrae, distal limb bones and the pelvis. It is possible that dressed caprovid carcasses and possibly even selected limbs were brought to the site during this period. However, the sample size is not vast and the high proportion of atlas vertebrae might suggest that even if sheep were provided in a butchered state, elements of the vertebral column were likely to be as important as the limbs. The amount of pig bone is really too limited for reliable analysis, but it is tempting to see the preponderance of ulna and tibia as possible evidence for importation of limb joints.
While in Period 7 there is still disproportionate preservation of the more robust cattle elements, compared to the previous periods there are more significant numbers of upper forelimb bones in relation to many of the distal limb bones, and the pelvis is particularly well-represented (Table 18, Figures 22, 27, 32). Discrepancies between more and less robust proximal and distal portions of bones indicate that taphonomic factors are as intense as in previous periods, so it is tempting to interpret specific selection of larger numbers of upper forelimb bones, although the number of fragments available for analysis is small. In the case of sheep, the bias towards the upper limbs, particularly the front leg, is more noticeable when compared to figures for the mandible and metacarpal, but again the numbers are too limited to be reliable. This fact is illustrated by the absence of the ulna and the presence of five astragali and only a single calcaneus. The Period 8 cattle distribution is more similar to that of Periods 3 and 5 and 6, although the numbers of phalanges are quite low, while the element distribution for pig may indicate the butchery and deposition of complete animals rather than selected elements. Caprovids however display a distinct lack of metapodialsls and phalanges, suggesting that these may have been removed with fleeces while the heads were left on (Table 19, Figures 23, 28 and 33).
An overall trend observable in Periods 5 to 8, to a lesser extent in Period 3 and only affecting caprovids in Period 4, is the under-representation of the horncore. These are delicate elements, but the fact that less robust cattle horncores were better represented than those of caprovids, and the low frequency of horncore fragments identified as large or medium mammal, may suggest that these elements were not regularly deposited at Blue Bridge Lane and were instead removed for craft-working at another location.
Due to the fragmented state of much of the material only limited metrical analysis was feasible, being largely restricted to breadth and depth measurements of the more robust bones. It was possible to calculate withers heights for 26 cattle bones, with three from Periods 5 and 6, one each from Periods 7 and 8 and the rest from Period 3. The 21 reconstructed shoulder heights from Period 3 had an average height of 1147mm and ranged from 1053mm to 1275mm falling into two main groups (Figure 35). One group comprised eight more diminutive individuals ranged between 1053mm and 1090mm while another five varied from 1215mm to 1248mm. Three more fell between these groups while a fourth was the largest. All the figures for Periods 3 and 7 fell around those of the Period 3 diminutive group, while that from Period 8 was a larger individual 1236mm at the shoulder.
Measurements from 21 cattle horncores show a number of trends, with division into two definite groups and a more disparate group when the basal circumference (BC) is plotted against the index of minimum basal diameter divided by maximum basal diameter (46/45x100) (Figure 36). Plotting these figures with those from Armitage and Clutton-Brock (1976) would suggest that the division of horncores into cows (high index, low circumference), bulls (low index and circumference) and oxen (low index and high circumference) is complicated by the fact that the genders do not always separate out particularly well, but also by the possible presence either of medium- and short- horn types or of an intermediate form. In Period 3 it is tempting to interpret the following: two medium horned oxen, two medium horned cows, three short horned cows and one short-horned bull. In addition, there are four rather indeterminate specimens perhaps more likely to be male than female. The single horncore from Period 4 is likely to be of a cow of short- or medium-horned type; that from Periods 5 and 6 is likely to be of a medium-horned type, but is intermediate between ox and cow. The situation is similar in Period 7 where the animals are of medium-horned type but of rather indeterminate gender. In the case of the four examples from Period 8 it is tempting to identify two medium-horned oxen, a cow and a bull. The fact that element analysis indicates that many horncores were deposited away from the site along with the greater robustness of short bull horncores, compared with those of cows means that the overall results are not particularly useful in reconstructing gender proportions, although they do provide a rough guide when interpreting the results of other metrical analyses.
Plotting measurements of the metacarpal, metatarsal and the astragalus provide the clearest indication of the probable gender structure of cattle brought to the site during Period 3. Theoretically, metapodialsls are ideal for determining gender proportions for several reasons: the distal epiphysis fuses by about two years of age, long after castration is likely to have taken place, and before the majority of animals are likely to have been culled; plotting length against breadth measurements should produce graphs in which smaller cows, taller but more gracile oxen and larger, more robust bulls are easily distinguishable. Unfortunately, the number of complete metapodialsls was insufficient for analysis, and instead reliance was placed upon plotting length and depth measurement, which are more an indication of robustness than of size. Despite this, three distinct groups, likely to relate to cows, oxen and bulls can be clearly distinguished. Plotting metatarsal distal medial trochleal depth (Dvm) against distal articular breadth (BFd) produces a pattern with nine cow-sized beasts, nine castrates and a couple of bulls (Figure 37). Measurement of the distal medial trochleal depth (Dvm) against lateral depth of the medial distal condyle (Dem) of the metacarpal indicated 14 cows, six castrates and two bulls (Figure 38). The astragalus, a robust element without epiphyses, was the only bone to provide a large number of length measurements. Plotting the lateral depth (di) against greatest lateral length (Gli) indicated a gender ratio of 23 cows to eight castrates to one bull (Figure 39). While the sample size is not vast it is interesting that the distal metapodials produced more even gender ratios than the astragalus, the opposite of what one might expect in a closed system, since it is far more likely that male cattle will be slaughtered before reaching skeletal maturity (and thus will not be measured). It would appear then that many of the animals slaughtered before about two years of age (roughly a quarter according to epiphyseal fusion and tooth wear) are likely to be female.
Data is insufficient to say much about gender ratios in each of the succeeding periods, other than the observation that there were a majority of females in Periods 5 and 6 along with a single castrate, but more even numbers of cows, oxen and bulls in Period 8.
Although the amount of caprovid metrical data is greater than that of cattle, it is hard to interpret the graphs once these figures are plotted as there is a tendency for measurements to either fall into up to five subgroups or for rather amorphous groupings to occur, a characteristic observed by Crabtree (1994 49) concerning non-sexual dimorphic size variation among Anglo-Saxon sheep from East Anglia. Caprovid withers heights were calculated on the basis of 45 elements including three from Period 8, four from Period 7, seven from Periods 5 and 6, and the rest from Period 3 (Figure 40). Those from Period 3 ranged from 474mm to 641mm with an average of 558mm. Within this, there was a particular group of 13 heights between 515mm and 550mm. Thirteen measurements lay between 550mm and 600mm with a certain degree of clustering at the lower end of this limit while only four exceeded 600mm. There were also two very small examples between 450 and 500mm. It is tempting to interpret all those below 550mm as ewes, but beyond this the measurements are likely to represent ewes, rams and wethers. As with cattle, the tallest individuals are likely to be castrates, since this process extends the growth period by delaying epiphyseal fusion. Period 5 and 6 measurements ranged from 478mm to 616mm with a mean of 539mm, the four Period 7 measurements from 512mm to 577mm and the three Period 8 heights from 541mm to 563mm.
Measurement of the proximal metatarsal epiphyseal depth against epiphyseal breadth (Dfp/BFp, Figure 44) produce a rough bimodal pattern with nine smaller and eight larger individuals plus a much larger outlier, while plotting calcaneus depth (ds) against articular length (C+D) produces a possible trimodal pattern with ten smaller individuals, seven larger and three large outliers (Figure 43). It is tempting to interpret the smaller examples as ewes, those in the middle as wethers and the most robust as rams. There are hints of this trimodal pattern when measurements of the distal tibia (depth Dd/ breadth Bd) (Figure 41), distal humerus (mid-trochleal notch HTC/trochleal breadth BT) (Figure 42) and astragalus (lateral depth Di/ lateral length Gli) (Figure 45), but gender groups are hard to define with any clarity.
The typically unfused state of the majority of the pig bone assemblage combined with the limited number of teeth meant that there were very few measurements that could be plotted. There was a total of 17 measurable astragali, two from Period 8, the rest came from Period 3. Plotting two length measurements, greatest medial length (Glm) against greatest lateral length (Gli) (Figure 46) is less likely to display metrical separation of gender than would plotting figures from two different dimensions, but it is possible to distinguish one or two subgroups, including one cluster of smaller females, another of slightly larger probable females and a third group likely to be male, both entire and castrated with one or two really tall beasts.
A pair of horse metacarpals from Period 3 deposits provided withers heights of 1429mm (14 hands) and 1359mm (13 hands 1") while a metatarsal from Period 9 indicated an individual 1450mm (14 hands 1") at the withers.
As observed in Section 2.1, there is a steady decline in the proportion of observed butchery marks between Period 3 (14%) to Period 7 (6%), with an increase in Period 8 (12%). There is also some degree of fluctuation in the type of butchery marks observed, although given the small sample sizes from Periods 4, 5 and 6, and 7, the significance of these trends is hard to assess (Table 20, Figure 47). In all periods the majority of butchery marks appear to have been made by splitting bones into fragments (about 65% in Periods 3, 5 and 6, 7 and 8 and over 85% in Period 4), either cleaving through them longitudinally or transversely using a heavy chopping tool which left a characteristic large straight-edged linear mark often succeeded by a spiral or undulating wet-bone fracture along the axis of the blow. Either due to post-butchery fragmentation or to the employment of a heavy blunt instrument, these typical wet-bone fractures are often present without the chop mark. The relative proportion of splits and wet-bone fractures varies by period with more or less even numbers in Periods 3 and 5 and 6, a marked tendency towards wet-bone fractures in Period 4 and a predominance of chop-marked splits in Periods 7 and 8. It may be no coincidence that the greatest proportions of wet-bone fracture relative to splits are in periods where preservation was less good and fragmentation was higher.
Chop and knife marks made with lighter tools account for a much smaller proportion of the assemblage, and while their relative proportions vary, that of chop marks could be considered to be fairly constant between 10% and 15%. Knife marks are likely to be under-represented, as it is highly likely that much of the material recorded as unbutchered (i.e., the vast majority of the assemblage), would have been skilfully disarticulated or filleted by slicing through tendons, leaving no trace.
Within each period there are patterns in the intensity and type of butchery applied to different taxa, but it is only in the large amount of Period 3 material that these trends are particularly clear (Figure 48). In Period 3, with an overall average of 14% butchered bones, 16% of caprovid, cattle and large mammal bones, are affected, but only 7% of pig. Surprisingly, at 21%, the proportion of butchered horse bone is greater than that of cattle, while the 21% of goose bones compared to 8% of chickens is likely to reflect the greater size of the former. The rough proportions of the individual butchery techniques recorded on cattle, horse and large mammal bone are very similar, with high numbers of split elements relative to knife cuts in particular. Knives were more important in the butchery of the medium mammals, especially in the case of pig where more than half of the observed marks were knife-cuts; knives were almost the only tool used in the case of birds. Large numbers of wet-bone fractures recorded on large and medium mammal bone relate to the splitting of long bone shafts. In the later periods it is interesting to note that with the exception of Period 4 where the number of bones is small, the proportion of butchered horse bone is greater than that of other animals, while that of pigs remains consistently low (Figure 49). From Period 7 (Figure 51) onwards there is also the impression that pig bones are treated differently from those of other animals, receiving more chops than knife-cuts and in Period 8 (Figure 52) in particular, their bones are far less likely to be split.
As might be expected, there are clear patterns in the type, intensity and location of butchery marks on different elements. Among Period 3 cattle bones the horncore, axis, humerus, femur and metapodialsls are the most heavily butchered elements, while the phalanges, major tarsals, ulna and mandible received very little attention (Table 21, Figure 53). Almost a quarter of all recovered horncores are split transversely just above the base, while another 15% bear chop marks in a similar location, indicating that the horn sheath was chopped through and tugged off, possibly after the underlying connective tissues had decomposed. The unmarked horncores which form the majority of the assemblage are likely to have been treated in a similar manner. The extant butchery marks appear to have been delivered from a variety of angles, suggesting that the head was already removed. Butchery of the axis is largely confined to splits, mostly associated with separation from other vertebrae, but sagittal splitting was also observed. The atlas also received a fair amount of attention, and was often split sagittally, and although such marks were seen on a few other vertebrae, they were not particularly common. Instead, a moderate proportion of vertebrae were longitudinally split part-way between vertebral spine and transverse process. Transverse splits dividing up individual or groups of vertebrae are most common, followed by splits consistent with the removal of ribs and transverse processes. Chops and more rarely knife marks likely to be associated with this process were also occasionally recorded.
All of the major upper and distal limb bones were affected by a certain amount of wet-bone fracture likely to relate to the deliberate splitting or smashing of these bones for the removal of marrow. This process of secondary butchery is likely to have been common on many elements, but is clearly indicated by the heavily butchered metapodials. There are a reasonable proportion of both the metatarsal and metacarpal with cut marks around the distal articulation, notably transverse knife marks across the dorsal ligament attachments relating to the removal of the phalanges, most likely with the hide attached (this process can be seen on several phalanges, where heavier butchery marks on the dorsal surfaces relate to the cutting of ligaments with knife marks on lateral and ventral surfaces probably pertaining to skinning). Once the metapodials had been separated from the phalanges and the podials, they frequently show evidence of having been split sagittaly, either proximally or between the condyles, or in one case, from both directions. Butchery of the hind leg was concentrated around the pelvic acetabulum and the femoral head, the latter frequently being split longitudinally to the axis of the limb, occasionally causing damage to the ischial and iliac margins of the acetabulum. Also common, appears to have been splitting of the femoral head transverse to the axis of the limb, causing splitting across the acetabulum when still articulated. Such processes seem hard to conduct unless the limb was first removed from the torso, and evidence of this is provided by splits across the ilium and ischium of pelvis fragments recorded either as cattle or large mammal. Separation of other hind limb elements is indicated by cuts and chops often around the proximal and distal tibia, which then seems to have been split sagittally for marrow extraction, and on the ventral surface of the astragalus.
The forelimb seems to have been treated in a similar manner, with chops occasionally concentrated round the proximal humerus but more frequently around, or just above, the scapula glenoid. Cuts in a similar location support the idea that much disarticulation was conducted with a sharp knife and this was likely to be the case for the elbow joint, where cuts and chops around the distal humerus, proximal radius and ulna were indicative of ligament slicing, rather than crude splitting, which was only rarely observed on the ulna. Splits on the proximal radius and distal humerus are longitudinal and are much more likely to relate to marrow collection than to disarticulation.
Although the amount of material available for butchery analysis from other periods was not great, the assemblages from Periods 5 and 6 (Figure 54) and particularly 8 (Figure 55) did show one or two differences to the Anglian material. Chief among these was the fact that the metapodials showed no sign of knife marks associated with skinning, and while they might occasionally suffer wet-bone fracture, concrete evidence for sagittal splitting was absent with the exception of a few examples from Period 5 F4. A slightly higher proportion of knife marks were observed on the phalanges, suggesting hide removal further down the limb. Horncores were much less commonly butchered, as was the scapula, with separation, particularly in Periods 5 and 6, concentrated around the proximal humerus. Treatment of the vertebrae in Periods 5 and 6 is more akin to that recorded from Period 3 but the majority in Period 8 have been sagittally split very close to the spine, indicating division of carcases into sides of beef. Splitting for rib removal was also observed in a number of cases. Disarticulation of the lower leg by splitting through the astragalus and calcaneus, particularly just below the tuberosity in the case of the latter, was a feature of Period 8 butchery and this was accompanied by various chops and cuts along the shaft of the tibia, possibly from filleting as well as one or two cases of sagittal splitting. Division of the elbow constituents seems to have been achieved by splitting through the ulna at various points around the articular surface.
Butchery patterns observed on Period 3 sheep (Figure 56) share several similarities with those of cattle, including a high proportion of horncores with transverse chops, along with heavy butchery of the axis, atlas and metapodials the latter of which occasionally have knife and chop marks around the dorsal and lateral margins of the proximal and dorsal ligament attachments and frequently have been sagittally split between the condyles. The pelvis was often split around the ischium and the pubis, but less commonly across the acetabulum; cuts and chops in the same location indicated that more delicate techniques were often employed. Butchery of the major limb bones is limited compared to those of cattle, likely to be a reflection of relative size. Compared to the appendicular skeleton, caprovid vertebrae were heavily butchered, mostly by transverse splitting to separate individual or groups of elements, but splits across the transverse process close to the vertebral body and others between the neural spine and transverse process were reasonably well-represented; cuts and chops in similar locations indicate filleting of meat.
There are too few caprovid elements to say much about butchery in later periods (Figures 57 and 58) although in Period 8 the metapodials were not butchered, but disarticulation of the hind leg could include splitting through the major tarsals, much as was the case with cattle. Butchery of the pelvis in Period 8 followed a similar pattern to that of Period 3, while the large amount of butchery meted out to vertebrae almost exclusively comprised sagittal splitting close to the neural spine, similar to that of contemporary cattle bones, with a few splits arising from rib removal and transverse separation.
The very limited data for pig butchery means little can be deduced beyond the possibility that most butchery was conducted with a sharp knife and that vertebral columns were treated in a similar manner to those of caprovids (Figure 59). Nor can much be said about horse butchery beyond the fact that in Period 3, horse bones (Figure 60), including the vertebrae, were treated in a very similar way to those of cattle. A metatarsal from F388 C1843 bore a concentration of about 20 deep knife cuts around the entire circumference of the bone, although such a pattern does not seem functional, either in terms of butchery or bone object manufacture. That horse metapodials were used as a raw material is indicated by a faecal concreted metatarsal fragment from which longitudinal slivers, possibly for blanks, had been sliced from the ventral surface from F200 C1387 (identified as large mammal) from Period 9, but likely to have been reworked from an earlier feature.
There are almost 150 cases of recorded pathology, the most common of which, with 44 incidences are arthropathies. These include examples of eburnation, ossification of ligaments, development of exostoses, fusing together of elements and splaying of joint surfaces, and are largely restricted to cattle carpals, tarsals and phalanges along with a few vertebrae, particularly in the case of caprovids, several horse calcanei and a pair of chicken coracoids. Splaying of the medial distal condyle was observed on four cattle metatarsals (and associated with eburnation and grooving in two cases) from Period 3 while an eburnated patella and a phalanx with ossified ligaments from Period 8 F77 C1744 is likely to have come from the same horse. Also very common are traumatic injuries, including fractures (often healed) and more frequently osteochondritis. These possible osteochondritic lesions of the medial proximal articular surface comprise shallow pitted sub-round depressions up to 2mm deep and 12mm across, often with a slightly raised reactive margin. These affected seven out of 43 metacarpals in Period 3, three out of six from Period 5 and 6, one out of four from Period 7 and seven out of 11 in Period 8. Similar injuries were observed on the articular surfaces of several phalanges, as well as a mandible, a scapula and a horse patella from Period 3 and a pig scapula from Period 8. Fractures are largely restricted to large mammal ribs and a few vertebrae with five exhibiting fracture callus formation to varying degrees, while three large mammal and three medium mammal ribs had healed with varying amounts of callus remodelling. Fractured bird and amphibian bones were also observed. Extensive ossified haematomas were noted on sheep metapodials from Periods 5 and 6, and 7, the former of which appeared smooth and well healed, while there are two cases of 'Selby' pathology. This buttressing of the dorsal margins of caprovid proximal metapodialsls, includes a mild case from F381 C1836 and more severe from F546 C2194, both from Period 3. A cattle metatarsal from Period 7 F219 C1428 has slightly backward angled condyles, a condition that may arise from the application of undue stress to the joint (such as from traction) before the distal epiphysis had fully fused. Of interest is a humerus from Period 3 F164 C1337 belonging to a goose that appears to have survived an attack from a carnivore. The bite had left a sub-circular depressed fracture about 6mm across on the dorsal surface of the shaft, just above the distal epiphysis. A patch of raised, pitted, greyish additional bone extended for 15mm around and above fracture indicated that the bird survived the experience at least long enough for the injury to become severely infected.
There are 14 possible cases of infection evidenced by plaques of pitted or striated greyish additional bone, often located on the ribs but also on the skull. Considering the large amount of material from Period 3, the incidence of infection seems greater in Periods 7 and 8, but the figures are too small to be significant. The seven cases of periodontal disease affected caprovids and cattle, and included inflammation of the tooth socket, alveolar recession and ante-mortem tooth loss, the latter indicated by irregular wear on one cattle mandibular, and three maxillary teeth. Other oral diseases are restricted to 12 cases of dental calculus; the incidence was not severe and was particularly common in caprovid jaws.
Other conditions were congenital in nature, such as reduction of the third molar hypoconulid which affected three cattle in Periods 5 and 6, and 7, while a caprovid mandible from Period 3 showed evidence either of congenital absence of the second premolar or the loss of this tooth a long time before death. From Period 3 F381 C2063 there was a caprovid maxilla fragment with the third premolar erupting at a decidedly backward angle, from Period 7 F219 a dog with a rotated second mandibular premolar and from Period 8 F162 C1352 a cow with the same condition. Also likely to be congenital from Period 3 F381 C2049 was a large mammal (probably cattle) thoracic spine bifurcated at the top, sadly no longer believed to have belonged to an imported Zebu cow. Very large foramina running through the vertebral body may also be congenital, affecting two cattle bones in Period 5 and another in Period 3.
The 22 identified Anglian pits form an important series of features often displaying a complex sequence of fills and backfills. An attempt was made to characterise the constitution and origins of the contexts within these pits so that the results of such analysis might elucidate aspects of the nature of Anglian occupation which is otherwise hard to define on the basis of scant structural evidence. These pits can be divided into a series of groups based on spatial arrangement as follows: Pit Group (PG) 1: F273, F351, F359, F458 F521 and F557; PG2: F13, F164, F353, F381, F388, F402, F408 and F246 and F413; Double Pit Group (DPG) 1: F442 and F508; DPG2: F143, F241 and F225; DPG3: F427 and F437. All except DPG3, located in Int22, were recorded in Int15. In addition, individual pits F520 and F546 were located in Int16 and Int24 respectively. PG2 is particularly significant since it lies within the main focus of Anglian occupation indicated by a concentration of personal items and building debris. Analysis was conducted stratigraphically by context on a number of taphonomic, cultural and economic factors including the proportion of carnivore gnawed bones, the presence of faecal concreted material, chewed fish bone and of small vertebrates (likely to have fallen into the pits, although the exact numbers of these, particularly in the case of amphibians, could be influenced by seasonal behaviour) combined with the state of preservation, angularity and fragmentation, lending some clue to the origin of deposits and the rapidity of burial; the proportion and type of burning as well as that of butchery, element distribution and species proportions were considered.
Analysis of species proportions was based upon raw fragment counts (NISP), as it was the only technique capable of providing adequate sample sizes, and while not necessarily the most sensitive of techniques, accuracy is sufficient for intra-site analysis of rough proportions. Element analysis also used raw fragment counts, one method considering all mammal remains (including bones identified as large and medium mammal) and another restricted to bones identified as cattle, caprovid and pig. To simplify the issue, elements were divided into body zones and further categorised into what might be considered primary (low meat bearing) and secondary (high meat bearing) butchery waste, although such matters are likely to be culturally and economically influenced. Elements likely to be removed during primary butchery (and therefore categorised as such) include the skull (cranium and horn core) the jaws (along with loose teeth) and the distal limbs (carpals, tarsals, metapodials and phalanges) while elements bearing greater amounts of meat were categorised as forelimb (scapula, humerus, ulna and radius), hindlimb (pelvis, femur, tibia and fibula), limb (unidentified long bone shaft) and torso (vertebrae and ribs). The greater robustness of mandible and distal limb elements means that they were much more frequently identified to taxon compared to the other elements, and in the case of the domesticate analysis, there was clear over-representation. The reverse is likely to be true when all bones were analysed, with greater fragmentation of limb, skull, vertebral and rib fragments skewing the results towards these elements. As such, it is probable that neither technique yields a particularly accurate representation of the material originally deposited, but they do allow some basis for intra-site comparison. Intensity of bone alterations such as burning was based on a comparison with an average by context (not on an average of all fragments by period). The average proportion of burnt bone within a context was 9.5%, while that of carnivore gnawed bone was 0.8% and butchered bone 11%. Roughly 70% of observed butchery marks were made by cleaving through the bone, either by splitting with a chopping tool or by smashing, with the remainder comprising knife and chop marks. Contexts with a distribution of butchery marks roughly matching these figures are described as having 'mixed' butchery.
Species proportions between F273, F351, F359 and F458 were broadly similar (Figure 61) (F521 contained a single cow bone and has been discounted from further analysis), with a slight predominance of cattle followed by caprovid then pig. Compared to other features, caprovids were marginally more important in F273, pig more so in F351 and F359, the latter feature also having relatively more horse bones, small numbers being present in all PG1 features. Only F273 and F351 contained geese in very limited numbers. Small mammal remains were recovered from F273, F351 and F458. House mouse (12 bones) closely followed by vole (9 bones) were the predominant taxa, but shrew bones were also present (6 bones), particularly from F458. Fish bones were recovered from F273, F351 and from F458 with interfeature comparison displaying a pleasing degree of similarity, with a predominance of eel bones and very few herring bones often outnumbered by those of cyprinid and salmonid (Figure 62).
The absence of bone of any description from the basal fill C2032 may suggest either that the pit was initially dug to receive cess (in which case a few fish bones might be expected), or, that there was sufficient time between cutting the feature and commencement of filling for some silting to occur. Overall, the presence of faecal concreted bones from all of F458's bone-bearing contexts and the presence of chewed fish bones in C2030, C1974, C2031, and C2043 would suggest that large amounts of human faeces were likely to have been deposited into the pit, but these seem to be mixed up with other waste. C2030 has a high proportion of burnt bone, indicating that hearth ash may have been a major component in this particular primary layer while relatively high amounts of carnivore gnawed bone would indicate that for at least some of the material, the pit was not the primary refuse location. C2055 contained a particularly high proportion of carnivore gnawing, while bone becomes more abraded towards the top of the backfilling sequence, often with low amounts of burnt bone. Where butchery is recorded, it is fairly consistent, affecting between 5% and 7% of bones in most contexts and is almost exclusively restricted to splits and wet bone fractures. Small vertebrates were present in most layers but not in large numbers, and it is tempting to see backfilling occurring fairly rapidly, with dumps of redeposited midden material interspersed with human waste. In the semi-liquid environment typical of a feature receiving fresh human waste, it is likely that material will percolate into lower layers, causing faecal concretions to occur in underlying deposits while it is also possible that raising of the water table, more than likely in a riverside location, could cause upward movement of solvent faecal minerals. Thus, even though rapidly deposited dumps of bone may have been interspersed with longer periods of faecal deposition, evidence for the latter material is susceptible to dispersal, making contexts appear more homogenous. Element analysis indicated that there were no particular concentrations of high or low meat bearing elements, although C2030 contained a slightly greater proportion of more meaty elements, as did C1969 and C2041.
The distribution of faecal concreted bone and chewed fish vertebrae indicated that the basal fills, C2104, C2096, C2103 and C2095 received a large amount of faecal material, while chewed fish bones were also present in upper layers C2093 and C1645. Preservation and angularity in lower deposits tended to be better, while the incidence of dog gnawing increases towards the top of the sequence, being notably high in C1645. Surface abrasion was not as common as among other features. The amount of burnt bone is also very high in C1645 and C2093, and is likely to indicate the inclusion of a large proportion of ash, while that from C2104 is especially low. Fragmentation was high throughout (very high in C1645), with butchery, where recorded, mainly between 6% and 10% and mostly comprising splits or wet bone fractures. Large numbers of small vertebrate remains were identified from basal fill C2104, but are fairly limited elsewhere in the sequence. Element analysis indicated a mixture of material with no particular concentrations. Overall, initial deposition seems to have been slow and ordurous, followed by slightly more rapid backfilling with domestic waste, at least some of which was initially deposited elsewhere (albeit only briefly in all likelihood) before redeposition. On the basis of small vertebrate remains, the pace of backfilling seems to increase towards the top of the sequence, which also receive a high proportion of fuel ash, much of which from C2093 is calcined, and may indicate the inclusion of some industrial rather than domestic fuel waste.
The basal fills of F351, including C2108, C2115 and C2114 contained evidence for human waste in the form of feacal concreted bones, or in the case of C2114, a lot of crushed fish vertebrae. No fine material was sorted from C2108, so it was not possible to determine the speed at which this basal deposit had developed, but the large number of small vertebrate bones from C2114 suggested a fairly slow rate of deposition. Dog gnawing and surface abrasion were very low and it is possible that bone material was dumped directly into the pit rather than developing on a midden, or if not, had not been left exposed for a long time. C2111 contained no evidence for faecal material, and may either have developed elsewhere (but was neither abraded or gnawed) or more likely, accumulated after the pit stopped receiving cess. Material in C2108 was much more variable in character, and is likely to have developed in several locations before final deposition. The amount of burnt bone was variable between contexts, but was notably high in C2114, and particularly low in C1763. With the exception of C2114, about 10% of bones had butchery marks where recorded which for the most part were splits and wet bone fractures, although C2115 was more mixed. This context, along with C2113, was exceptional in containing no distal limb elements, but the distribution of high and low meat bearing elements was otherwise mixed throughout the feature, perhaps with the exception of C2114, which contained less distal limb and head bones. A single piece of worked deer antler was recovered from C1763.
All three of the fills of this feature contained faecal concreted bones, although the lack of sampling meant that chewed fish and small vertebrates were not recovered, and no bone of any description was recovered from basal fill C1940. The extremely high proportion of gnawed bones recorded in C1866 and C1855 is somewhat anomalous given a state of preservation verging on excellent. C1866 seems to contain a high proportion of low meat bearing bones, and combined with the absence of burnt bones indicating fuel ash, it is tempting to see this as rapidly redeposited primary butchery refuse. C1863 on the other hand contains a more mixed assemblage, some fuel ash and a more average amount of dog gnawing, suggestive of more primary deposition. This latter deposit also differs in having a relatively high proportion of knife and chop-marked bone, compared to the majority of split and wet bone fractured fragments observed in C1885 and C1866.
There is a greater variety of species proportions in PG2 features compared to those of PG1, with F353, F388, F402 and F408 containing a very high proportion of cattle bones, with between three and five times the number of caprovid bones. Cattle are also predominant in F381 and F413, but the ratio is closer to 3:2 while in F13 and F164, caprovid are slightly more common than cattle. The proportion of pig bone is surprisingly consistent throughout, with the notable exception of F353 which contained only about half as much pig but did have a high number of horse bones which, with the exception of the small proportion from F13, were otherwise absent or present in very small numbers from Pit Group 2 features compared to Pit Group 1. The proportion of bird bones is also fairly constant, although noticeably higher from F408 and lower from F388 and F402 and these two, along with F353, contained no goose bones which are otherwise present in all PG2 features. The broad similarity between F388 and F402 is likely to relate to the fact that F402 cuts F388, and is thus likely to contain a lot of residual material from the earlier feature. The amount of small mammals is low compared to the bones of amphibians, and with the exception of F381 (more house mice), the bones of vole are predominant, while apodemus sp. (wood or yellow necked mouse) was also well-represented. Shrew is much less common, with only a single bone from F381. The relative proportions of fish bones, with the exception of F381 which closely resembles those from PG1, are both different from PG1 features and much more variable. While the actual numbers of fish bones from PG2 features are often less than those from PG1, there appears to be a marked increase in the proportion of herring, (sometimes exceeding eel, as in the case of F164) while that of salmonid is lower. There is also a little more dietary variability, with slightly larger (but never vast) numbers of the more minor taxa such as flatfish.
The backfilling sequence within the funnel shaped cross-section of F13 is the most complex of any of the features excavated at Blue Bridge Lane. The earliest fills, C1913, C1881, C1148 and C1908 all contained faecal concreted bone while a few chewed fish vertebrae are present from C1881. Preservation tends to be good and spiky and fragmentation throughout is high, but surface abrasion is common (but not universal) in contexts above C1881. Dog gnawing is low overall, with no particular correlation with the presence of surface abrasion. Burning, mainly charring, is particularly high in C1142 and C1143 but also C1065, but frequently is absent. Butchery is dominated by splits and wet bone fractures and is particularly high in C1148, C1064 and C1065, but also in C1143 and C1880, otherwise ranging from 8% to 13%. All contexts contained a mixture of more and less meaty elements. Overall, it would appear that initial deposition of cess and fragmented mixed bone (most likely to be deposited directly but with some likely to be the result of cleaning up after local dogs) was fairly slow, allowing a large number of small vertebrates to fall to an unpleasant death (C1913, C1881). These initial deposits were followed by (or potentially accompanied by) several deposits of heavily butchered bone (C1148, C1147) both with surface abrasion and with a high level of carnivore gnawing from C1147, likely to be redeposited midden material dumped into a slowly accumulating faecal matrix which again attracted large numbers of small vertebrate bones. Backfills C1145, C1146, C1904 and C1908 seem to be rapidly deposited dumps of bone refuse with varying amounts of ash, likely to contain at least some proportion of fairly quickly accumulated and redeposited midden material with a minimum of surface abrasion and dog gnawing. C1144 is similar in character, but the amount of small vertebrates might suggest either that accumulation was slower, or was during a part of the year when amphibians are more mobile. The same is true of C1064, which is likely to have contained at least some faecal material, and C1065 which shows very little abrasion, and are interleaved with more rapidly deposited layers of similar material containing varying amounts of what is likely to be domestic hearth ash. The final context, C1069, contains no bone, and is likely to be a final deliberate sealing layer.
The two contexts from F164 show little evidence of faecal material with the exception of a crushed fish vertebra from C1341 and a few faecal concreted bones from C1337. The majority of material from the feature is quite heavily butchered, well-preserved and spiky, with limited gnawing and ash component while the sparse numbers of small vertebrate remains indicate rapid deposition. The bone from C1341 is only moderately fragmented and contains a high proportion of more meaty elements the butchery of which has largely been conducted by knife cuts and by chopping. That from C1337 is more mixed in character, with butchery mostly comprising wet bone fractures. It is tempting to suggest that material may have developed quite quickly on a midden, before being redeposited with some domestic fuel ash, and in the case of C1337, including either a limited faecal component or, bones redeposited from a cess pit.
The single backfill C1449 contained a rather disparate assemblage of material, much of it variable in terms of preservation and angularity and suggestive of heterogeneous origins. It is mixed in terms of meat quality and contains no evidence for faecal material. Butchery is high, and is largely restricted to splits and wet bone fractures while a very high number of small vertebrate remains indicate either a slow build up of material, or, that the lined pit lay empty for a long time before backfilling. The latter seems likely.
The basal fill C1887 contained only fish bone (some chewed), while lower fills C1858, C1859 and C2062, also exhibited good evidence for cess, including faecal concreted bone and chewed fish vertebrae. Rather limited small vertebrate remains suggest either a more rapid process or the presence of some sort of structure around the pit. Further up the sequence there is increasing evidence for the presence of fuel ash and surface abrasion, suggesting that midden material was periodically dumped, although reasonable numbers of small vertebrate bones might suggest that there were reasonable time gaps between each backfilling event. Chewed fish vertebrae in intermediate layers C1851, C1852 and C1853, C2054 and C2063 suggest the incorporation of a more limited amount of faeces. Bone waste seems to be mixed in terms of meat quality, only backfill C1847 containing a slightly greater proportion of higher meat bearing elements. Butchery ranges from 8% to 24% (in C1851) with most around 12% and tend to be more mixed, although bones from C1848, C1849 and C1850 exhibit a high proportion of chop and knife marks while spilt and wet bone fractured elements were more common in C2062 and C2063.
With the exception of a chewed fish vertebra from C1845, there is little evidence for faeces within this feature. Basal deposit C1856 comprises partly articulated cattle bones, mostly complete vertebrae, representing the remains of at least three individuals. A single gnawed vertebra is the only incidence of carnivore activity in the feature, and this, combined with the state of articulation and the lack of surface abrasion indicates that material was deposited directly into the pit, most likely quite quickly considering the few small vertebrate bones recorded. C1845 is more mixed in terms of meat quality and is moderately fragmented, while that from C1857 contained a higher proportion of possible primary butchery waste, but the variable state of preservation, angularity and colouration along with the greater fragmentation within this context might suggest the presence of bone from a number of sources and is likely to indicate a deliberate single backfilling incident.
There was a single backfill, C1883, which contained a slightly higher than normal proportion of low meat bearing elements, particularly of skull fragments, but was otherwise mixed in terms of meat quality. No gnawing was recorded although quite severe surface degradation might indicate redeposited midden material. There is no evidence for faecal material while relatively low numbers of small vertebrate remains might indicate rapid deposition. Burnt bones are quite common.
The four contexts from this feature contained no evidence of faecal material, although only the two lower fills were sampled. The amount of bone was not particularly vast, although the upper fills contained a greater proportion of high meat bearing elements while the lower were more mixed. Fuel ash seems to be an important component particularly in all but upper fill C1892, while butchery marks are more variable. Limited carnivore gnawing and an absence of surface degradation suggest either redeposition of rapidly accumulated midden material or direct deposition into the pit, which on the basis of the limited amount of small vertebrates from the sampled lower fills, was also likely to have been speedy.
There is no evidence for faecal material from either of the two fills which contained fairly small assemblages of mixed bone, heavily butchery marked in basal fill C1903 and with a moderate fuel component likely in C1897. No samples were taken, so rapidity of deposition on the basis of small vertebrate remains is impossible to assess.
No animal bone was recovered from backfill F508 C2073. F442 contained about twice as many cattle bones as caprovid, a fairly typical proportion of pig bones and a few bits of horse. The level of bird bones is also fairly typical, with goose well-represented but the most significant feature is the amount of worked red deer antler offcuts which although present in F273, F351 and F381 are rare from these features. Small mammal remains are restricted to a single wood/yellow necked mouse bone while the distribution of fish taxa is markedly different from any feature explored so far. Although the numbers are small, there are almost as many salmoniforme bones as eel, with herring making a very small contribution similar to PG1 features.
Evidence for a faecal component was limited from this pit, restricted to concretions observed on bones in C2011 and a single chewed fish vertebra C1951, despite the fact that all contexts were sampled. Basal fill C2026 contained little bone of any description while the remainder was fairly mixed in terms of meat quality, but with a sligh
