Exploring a Neolithic neighbourhood at Llanfaethlu
Since 2014, archaeological work at Llanfaethlu, on the north coast of Anglesey, has been revealing the remains of the first early Neolithic multi-house settlement to be found in north Wales. Catherine Rees and Matthew Jones explain further.
When our work began at Llanfaethlu, the site had already been identified as being in an archaeologically significant area. A desk-based assessment carried out ahead of the work (which was undertaken on the site of a new community school) had highlighted the presence of several early medieval cemeteries and a number of prehistoric sites in the immediate vicinity, so we knew that this was an area with great potential. Even so, the discovery of a cluster of four early Neolithic houses dating back almost 6,000 years came as a huge, but very important, surprise. The find is unprecedented in north Wales, where only three single houses of this period were previously known.
Since then, we have been working closely with the developer, Isle of Anglesey County Council, and Gwynedd Archaeological Planning Services, to excavate and record these rare remains (see CA 315). In addition to the structures, we have uncovered a sizeable assemblage of Mesolithic flint; several thousand Neolithic artefacts, including pottery and stone tools; two enigmatic burials; late Neolithic/early Bronze Age pit groups; a burnt mound, and much more. As we write, environmental samples from the site are being wet-sieved and processed, and to-date this work has yielded a wealth of evidence to help build up a detailed picture of what sort of landscape the Neolithic houses had stood in – including seeds, burnt wood and twigs, acorns, and thousands of hazelnut shells. We have also uncovered a number of carbonised posts and planks belonging to the buildings themselves, allowing us to get a much clearer idea of what they looked like. So let us take a tour through this Neolithic neighbourhood.
The structure that we dubbed ‘House 1’ was the largest of the four, an impressive longhouse measuring 21.6m in length by 7m wide. Interestingly, as we examined its footprint more closely, we could see that the building’s outline was curiously non-uniform, with a number of methods used in its construction. The north-eastern wall was formed from a row of five stone-lined post-holes, at least 40cm in diameter and 20-25cm deep, while its south-eastern partner was not only very different in make-up, but showed considerable variation in technique along its length. Part was made up of beam-slots, part of post-holes, and at the north-eastern end we found a series of ten stake-holes.
It is not yet clear whether these posts had replaced an earlier beam or vice versa, but this area also held tantalising clues to a dramatic episode in the dwelling’s past. As we excavated this portion of the wall, we discovered reddened, heat-affected soil and, within this, burnt planks and the tip of a charred oak post. That traces of burning can be seen within the lowest level of the original wall slots hints that fire may have been involved in some kind of significant act during the closing down of one phase of the house – perhaps it was put to the torch before being renewed in the same spot (the wall slot was later re-cut in exactly the same location) – although, of course, these charred remains could just as easily echo an unfortunate accident.
It was this wall that contained the dwelling’s doorway, defined by two post-holes, between which we found two large, flat stones creating a step or threshold into the building. Like the burnt wood found nearby, these remains also hint at closing rituals, perhaps marking the end of the longhouse’s life or a particular phase of use: directly behind the threshold was a stone-filled pit, which blocked the entrance and may have contained a pair of posts sealing off the old way in.
Moving around the structure, the south-western wall was the least substantially built of the four, formed from pairs of stone-lined post-holes that seem to have been sporadically replaced over time, between which may have stood wattle-and-daub screens. The north-western wall was more robustly built, using both beam-slots and postpits – though none of the walls were as well built as the south-eastern side, the front, which faced the other houses. It is interesting to note the marked differences in the investment of resources between this part of the structure and its much less substantial rear. Was this disparity due to practical concerns, or might we see a more house-proud motivation, creating an impressive frontage purely for display?
As for how the house’s interior had been arranged, rows of post-holes and wall slots running at 90 degrees to the outer walls created separate compartments – something seen within all the structures. In House 1, these lay on either side of a central row of post-holes running the length of the building, splitting the interior in three. Two of these segments were busy with internal features, but the backmost portion was apparently empty – it could be that this area was used to house livestock, brought inside for shelter and to provide warmth for the building’s human occupants.
Of the myriad pits and post-holes that had been dug in the other parts of the longhouse, one in particular leaped out at us as something unusual. This steep-sided post-hole, half a metre deep and over 70cm across, had clearly received special attention from its creators. At the bottom of the hole, we found the cremated remains of a sheep or possibly a deer leg-joint (these species are suggested by the survival of an intact astragalus bone, part of the ankle), which had apparently been burnt within the pit itself, or at least placed in the ground while it was still very hot, as high temperatures had reddened the surrounding clay.
It is possible that this special – perhaps ritual – ‘structured deposit’ had marked an early stage in the house’s construction: the remains were covered by a burnt quern stone that also formed the base of a post-hole. Here, too, we can see what happened at the end of the life of the hole (and possibly the house): the timber was removed from its slot and the hole backfilled with stone – during which a beautifully polished rubbing stone was carefully pressed into the side of the pit – closing the earth back over the buried items. It is fascinating to think of how one feature could represent the entire sequence of establishing and then decommissioning the house – we hope further dating analysis will help to clarify the timeframe over which this activity took place.
Sign of a shrine?
A short distance away, ‘House 2’ was another fairly large building measuring around 16.6m by 9.6m, though its precise design is still being unravelled – at first we thought we were looking at a square structure with a porch-like entrance, but further digging revealed that the ‘porch’ was in fact a continuation of the main exterior, and we are yet to locate the actual entrance.
Inside the house we found another grid of internal postholes, some linked by hollows filled in rather haphazard fashion with packing stones. This apparent randomness leads us to imagine that, rather than being beam-slots, the shallow trenches could have supported a wattle-and-daub structure with stones packed around the hazel uprights. As for how these flimsy screens may have been used, they mark a clearly defined square within the main building that measured around 1.6m2. There is no obvious way inside this space – it is possible that there was a higher opening in the partition, through which one stepped to gain entry – and we can see that its walls had been replaced or renewed in the same position at least once. Such small internal chambers are very unusual, if not completely unknown – Corbally 5, at Kildare, and Stretton-on-Fosse, Warwickshire, have similar carefully delineated partitions – but for now their purpose remains obscure. They could have been as mundane as a separate storage area, or as mystical as a shrine.
Another enigmatic aspect of the house lay outside its walls: just to the north we found a particularly large post-hole, around 65cm in diameter, containing stone packing that could have supported a very large post indeed. This is, of course, speculation, but we wonder if this single, substantial timber could have been carved or otherwise adorned to serve as some kind of marker or totem for the house.
Our third structure was harder to distinguish, being partly overlain by a cluster of Middle Neolithic (c.3300- 3000 BC) pits. We have attempted to unpick which holes belong to the house, and which to this later group, through pottery analysis. The later features contain fragments of Mortlake pottery, while those thought to be earlier do not – some being empty, others containing Irish Sea ware. This is a rather blunt method, however, and we look forward to finding out what post-excavation analysis reveals.
As far as we can see, the building was formed from a mix of post-holes, stake-holes, and beam-slots/hollows, creating a structure measuring around 6m by 11.5m. Inside, like its neighbours, it was divided into clear zones, while the centre housed an area of heat-affected clay, probably a hearth. Discoloration of the natural clay speaks of a large fire being repeatedly re-lit in the same spot, and its central location would have provided light for the whole main area of the house. In a compartment to the rear, we found a large post-hole filled with packing stones and struck flint, Irish Sea ware pottery, and hazelnut shells. Lying opposite the hearth in a gap between two beam-slots, the hole may have had some kind of special function; the same area of the house also contained a post-hole (one of a pair at its entrance) into which had been placed a leaf-shaped arrowhead and a fragment of a distinctive polished stone axe.
This latter find came from the Graig Lwyd axe factory in Penmaenmawr, the largest Neolithic axe-producing site known in Wales. Its products not only dominated the local market, but are seen as far afield as Yorkshire, the Peak District, and the Midlands. The presence of such an axe at Llanfaethlu places our site in a much wider context of commercial and cultural contacts. The arrowhead, meanwhile, seems to have been burned after manufacture. We suggest that these ‘special’ items had been specifically selected to be buried as an act of ceremonial ‘closing’ during the decommissioning of the house and the removal of the posts that once occupied the pits.
The fourth and final house was discovered more recently (see CA 323), and we were not able to excavate its entire footprint as it lay partially outside the area of our investigation, and under some inconveniently placed overhead cables. Outside the area proposed for development, moreover, a later trackway has been cut through where the house would once have extended. Even with only half the structure exposed, though, we can already tell that this was another imposing building, possibly the largest of the four – it measured some 11.5m in width, and while we have currently uncovered about 12m of its length, it continues well outside the trench, and we suspect its full extent could have been closer to 20-25m.
In make-up, House 4 is not unlike its neighbours, with rows of inner and outer post-holes, a hearth area, and a series of narrow slots for wattle-and-daub partitions – but it also yielded something much more unexpected: human remains. This was a significant find: bone preservation at Llanfaethlu, and in north Wales more generally, is very poor due to the high acidity of the local soil, which eats away at organic materials. Neolithic remains are rarer still, and when they do survive they are usually found in megalithic tombs, so to find a burial of this period in less monumental surroundings was truly exciting.
The grave lay just outside House 4 in a shallow pit only large enough to have contained a body laid in the foetal position. Very little of its occupant remained, but we did recover a small number of teeth belonging to a teenager or young adult. There were no obvious grave goods, but as the fill contained two worked flints, and the burial lay right beside a pit holding over 200 early Neolithic pottery sherds, we suspect that the burial is contemporary with House 4. This will have to be confirmed through radiocarbon dating, but if it does prove to be the case, such an association would be very interesting: the burial of individuals near or under houses is not uncommon in continental Europe, and such practices hint at a culture where ritual and secular lives were far more integrated than ours are today.
For all that such discoveries are very rare in the region, these were not the only human remains found at Llanfaethlu. A second set was uncovered on the site, but in a very different context. While investigating an area that initially appeared to be a scatter of well over 600 flints, we found a series of tree-root holes into which pieces of human bone had been tucked. Unlike the first set, these remains – comprising teeth, fragments of jawbone, and possibly part of a long bone – are unlikely to have been articulated when they were consigned to the ground. There was no associated pottery to hint at the date of this burial, but interestingly the overlying flint scatter appears to use exclusively beach flint, something linked to earlier Neolithic/Mesolithic contexts elsewhere on the site.
The discovery of these two sets of human remains is fascinating, not only for their rarity, but because of the insights they grant us into how varied Neolithic mortuary practices may have been. At Llanfaethlu, we have a more formal grave associated with a house, and bones being placed in natural features – tantalising glimpses of burial customs beyond better-known activities observed in tombs. We can only speculate on what other practices might yet be discovered, or what traditions may have once been commonplace, but have not survived for us to record today.
We hope that further analysis, particularly radiocarbon dating, will help to refine our understanding of how long Llanfaethlu’s early Neolithic occupation lasted, and how the four houses relate to each other. Dating evidence from other comparable Neolithic sites, together with the fact that all four of our structures are associated with very similar Irish Sea ware ceramics, suggests that the settlement lasted for around two centuries between 3800 and 3600 BC; but we may yet gain a more detailed picture.
Dating sites through material culture alone is problematic, as changes in artefacts do not necessarily mean changes in population, nor do they always signify that a site has been abandoned and returned to; this is the difficulty with interpreting the later Neolithic occupation of Llanfaethlu.
As we mentioned above, House 3 was cut by a series of later features containing characteristically middle Neolithic material. This included four hearths, a scatter of post- and stake-holes, and 19 bowl-shaped pits, some as wide as 1.2m in diameter. Some of the hearths contained stones that may have been involved in cooking, while one was surrounded by a complex arrangement of small holes that could once have supported some kind of frame used to suspend items over the fire for cooking or drying.
Despite their diverse functions, most of the cut features had one common element, in that they contained pieces of Mortlake ware pottery. This highly decorated style of bowl, with heavy rims and high shoulders, was originally thought to be characteristic of central and southern England, but it is now being found in increasing quantities in north Wales. The pieces we are finding at Llanfaethlu show particularly strong similarities to examples from Parc Bryn Cegin, Llandygai; Borras Quarry, Wrexham; and Brookhouse, Denbigh, extending the site’s cultural connections once more. Some of the quality of potting and decoration that we have seen at Llanfaethlu, moreover, is quite exceptional.
It was not only ceramics that came out of the pits – these holes also yielded quantities of worked stone, including local and imported flint, chert, and fragments of at least two more Graig Lwyd axes. The concentration of features within this pit group, some of which intercut, together with the number of hearths, strongly suggests repeated use of the same site over a period of time, reinforcing the strong sense of place that we have observed at Llanfaethlu. This, however, is slightly contradicted by the widespread distribution of sherds from individual pots in multiple pits, which would suggest that all the pits were being filled at approximately the same time, and that the period of occupation might be relatively short.
Further mysteries come from the pits themselves: in at least two, we have found both Mortlake ware and Grooved ware – something that raises a bit of a problem, as Grooved ware is generally thought to have been popular some 500 years after Mortlake ware. True, both styles of pottery were recently seen in similar pit clusters at Parc Bryn Cegin, Llandygai, but they were never found together in the same hole – and there radiocarbon dating gave a currency for Mortlake ware in the 33rd-30th centuries BC, and for Grooved ware in the 29th-26th centuries BC. What the mixed material at Llanfaethlu means is something that post-excavation analysis may help unravel: could we be looking at the selection and curation of ancestral pieces?
Grooved ware was not confined to the few sherds found in these pits, though: we saw it in much greater quantities in five groups of late Neolithic pits that lay to the south of the houses. These groups generally contained 15-20 pits which were, on average, much larger and deeper than those containing the Mortlake pottery, and many intercut, but radiocarbon dating is needed to see if we are looking at intensive use over a short period of time, or something smaller scale over a longer duration. Interestingly, although sherds of pottery from the same vessel were found in multiple pits, to-date this has only been observed within groups: there have yet to be pieces of the same vessel found in more than one group.
These features also tell us that by now the site’s trade networks had widened: the flint in the pits was no longer almost exclusively from local beach pebbles, but a high-quality imported material. Graig Lwyd axe fragments were also found in far greater numbers, and were larger pieces and complete artefacts. The flint tools present had also changed: in this period, we see the emergence of serrated cutting tools. As with the pottery, axe fragments were also found distributed between multiple features but not between different pit groups – and, fascinatingly, within these later pit clusters we also saw materials associated with the manufacture and modification of stone axes. This included two stone axe polishers, as well as a large piece of unworked Graig Lwyd stone, while fragments of stone axes with polished elements were also found reworked into other tools, such as a large Graig Lwyd scraper.
There was one last mystery to come: one of the pit groups was associated with the partial remains of an enigmatic curved feature, around 11m in diameter, that contained stone-lined post-holes. These remains are still under investigation and their purpose remains obscure. Might they be part of a house, or a more monumental construction like a henge or a barrow?
Burnt mounds and belonging
The final feature that we uncovered at Llanfaethlu was a small burnt mound lying about 20m south of the later Neolithic pits. Burnt mounds are fairly ubiquitous in Ireland (where they are called fulachta fiadh – see CA 256) and parts of Britain, including north-west Wales, and although their use spans the late Neolithic to the early medieval periods, they most often date from the Bronze Age (we are yet to get radiocarbon dates back for our example). They have a distinctive distribution pattern, lying on the margins of wetland or near a stream or other source of water, and while there is no stream at Llanfaethlu, the land around the mound was the only part of the site where we found post-medieval land drains, suggesting it had once been a wet area, possibly home to a spring.
Characterised by a heap or spread of fire-cracked stones and charcoal, these mounds are generally associated with a pit or trough lined with clay or timber. It is thought that this was filled with water, into which heated stones were placed to bring the liquid to temperature. The purpose of this act has long been debated, but a variety of functions from saunas, brewing beer, and other industrial processes like dyeing, fulling, or bending timber, have been suggested. We can only wonder what activities might have been carried out at Llanfaethlu.
Although aspects of the site’s archaeology remain enigmatic, what is clear is that Llanfaethlu is an important addition to our understanding of the Neolithic period; to-date it remains the only multi-house settlement of early Neolithic date to be found in north-west Wales, and will undoubtedly become a ‘type site’ in the study of British prehistory. While it shows some striking similarities with the houses at Llandygai, near Bangor, and Parc Cybi at Holyhead, Llanfaethlu stands apart in terms of the artefacts excavated there, the level of preservation, and the size of its houses. There is a strong resemblance to Irish sites, though, where we sometimes find a recurring pattern of two or three buildings clustered together (see CA 296).
At Llanfaethlu, the location seems to have been chosen as a long-lived focal point, potentially used over a period of at least 1,000 years. It seems to be a place whose physical remains endured for a long time, as did the respect in which it was held by its prehistoric occupants – the later pits nestling between two structures suggest that at least some of the houses must have remained visible above ground for centuries after they were abandoned, either as an empty, decaying ruin, or as earthworks.
These house-builders were not the first group to make use of the site, though: there was also a strong late Mesolithic presence. It is relatively common to find a small residual scatter of Mesolithic flint on later sites, but at Llanfaethlu we found over 100 pieces – a number that is likely to increase as Dr Ian Brooks continues his examination of the material. The flints pre-date the Neolithic houses, possibly by centuries, which raises interesting questions about the relationship between these Mesolithic people and the later settlement. Are we looking at an incoming population bringing new ‘ways of living’, a local adoption of new technologies, or something in between? Either way, this was a location that was evidently returned to again and again; we are only the latest visitors to explore the land around Llanfaethlu.
This feature was published in CA 332.