Dating a Crannóg: Stratigraphy, Radiocarbon and a Drainage Ditch

The most defensible sequence is Layer 4, then Layer 3, then Layer 2, then Layer 1, but the team must state that the ditch makes parts of it insecure. Layer 4 is lowest and undisturbed, and the hazel stake gives a calibrated radiocarbon range of AD 1010–1160, so it is the earliest dated context. Layer 3 lies above it and contains an oak plank felled in AD 1128: the timber cannot have been used before it was felled, so Layer 3 was formed after 1128, which is consistent with the C14 range for Layer 4. Layer 2 lies above Layer 3 and contains a coin minted 1272–1307, giving a terminus post quem of 1272 for that layer. Layer 1 is topsoil and contains modern material, so it is latest. The problem is the nineteenth-century drainage ditch, which cut through Layers 2 and 3. That means objects can have moved between them. The corroded coin found in Layer 3 may be intrusive from Layer 2 or from the ditch fill. The team should present the sequence as a relative stratigraphy with disturbance, not as four sealed and perfectly dated episodes.

The team should resist a simple yes or no. The evidence supports activity in the twelfth century: the oak plank was felled in 1128, and the C14 range for Layer 4 overlaps that period. That is before the first Norman arrivals in 1169, but it does not show that the site was abandoned then. Layer 2 contains a coin minted between 1272 and 1307, which indicates activity after the Normans had arrived. However, the drainage ditch cut Layers 2 and 3, so the coin's position may be disturbed, and the corroded coin in Layer 3 cannot be securely tied to twelfth-century occupation. A defensible answer would be: 'The crannóg was certainly in use in the twelfth century. There is evidence of later, thirteenth-century activity in Layer 2, but because of the ditch we cannot yet say whether this was continuous occupation, reuse, or redeposition. The site was therefore not necessarily abandoned before the Normans.' The team should ask for more dating, careful plotting of finds, and perhaps excavation of undisturbed deposits.

Radiocarbon and dendrochronology answer different questions. The C14 date comes from a hazel stake in Layer 4 and dates the death of the wood; after calibration it is a probability range (AD 1010–1160), not a single year. Dendrochronology on the oak plank from Layer 3 gives a felling date of AD 1128, which is much more precise, but it only dates when that tree was cut. The plank could have been reused or stored before it entered the layer. The two dates also come from different layers and materials. Averaging them would produce a meaningless number because it would mix a range from one context with a precise date from another and would ignore the stratigraphy. Instead the team should use each as a terminus post quem or a cross-check: the C14 range and the dendro date are compatible, which strengthens the case for twelfth-century activity, but neither dates the whole site. The ditch disturbance is a further reason to avoid over-precision.