Derek Hamilton
University of Glasgow, Scottish Universities Environmental Research Centre, Professor of Archaeological Science
My research interests lie in the use of Bayesian statistics in the modelling of absolute and relative dating evidence, thus providing a framework within which archaeologically relevant questions of chronology and temporality may be addressed. My PhD research focused on building absolute chronologies for later prehistoric settlement sequences in central Britain and using these to test models of cultural change. Prior to my current position, I was involved in the use of both geophysical and aerial/spaceborne remote sensing techniques for the identification, mapping, and spatial analysis of archaeological sites. As a result, my future research plans include building on these foundations of space and time, to ultimately develop a quantitative methodology whereby integrated spatio-temporal analyses can be undertaken on a regional level.
Address: Scottish Universities Environmental Research Centre, Scottish Enterprise Technology Park, Rankine Avenue, East Kilbride, G74 0QF, UNITED KINGDOM
Address: Scottish Universities Environmental Research Centre, Scottish Enterprise Technology Park, Rankine Avenue, East Kilbride, G74 0QF, UNITED KINGDOM
less
InterestsView All (24)
Uploads
Papers by Derek Hamilton
revealing the very well-preserved remains of an Iron Age
settlement, the wetland context ensuring that the timber
structures have remained intact and that the detritus of daily
occupation survives for us to pick apart and understand. One of
the structures in this settlement is an exceptionally well-preserved
roundhouse, the material remains of which have been subjected
to a barrage of analyses encompassing the insect, macroplant,
bone and wood assemblages, soil micromorphology, faecal
steroids, radiocarbon-dating and dendrochronology. These will
enable us to address some of the key issues regarding the life
cycles of Iron Age roundhouses, from conception and
construction, use of internal space, nature of occupation and likely
function, through to abandonment. Critically, we are now able to
view that life cycle through the lens of a tightly-defined
chronology bringing us close to the ‘ …short-term timescales of
lived reality’ [Foxhall, L. 2000. “The Running Sands of Time:
Archaeology and the Short-Term.” World Archaeology 31 (3):
484–498].
proportion of their diet must have contained some non-terrestrial protein that produced a reservoir effect. As freshwater fish from Lake Mývatn and marine fish have similar bone collagen δ13C values, using only one isotope to differentiate between the two food sources was not feasible. Therefore, δ13C, δ15N and δ34S isotope values for various terrestrial animals, and marine and freshwater fish were incorporated into the Bayesian mixing modelling program FRUITS (Food Reconstruction Using Isotopic Transferred Signals) to more accurately determine the diet of each individual. Their 14C ages were then corrected for both a freshwater and marine reservoir effect before Bayesian chronological modelling was employed to provide a more robust time frame for when the cemetery was in use. Previous findings suggested activity began
shortly after landnám, however this study indicates that the cemetery was in use between the 11th and 13th centuries.
A previous multi-isotope study of archaeological faunal samples from Skutustaðir, an early Viking age settlement on the southern shores of Lake Myvatn in north-east Iceland, demonstrated that there are clear differences in d34S stable isotope values between animals deriving their dietary protein from terrestrial, freshwater, and marine reservoirs. The aim of this study was to use this information to more accurately determine the diet of humans excavated from a nearby late Viking age churchyard.
Materials and Methods:
d13C, d15N, and d34S analyses were undertaken on terrestrial animal (n=39) and human (n=46) bone collagen from Hofstaðir, a high-status Viking-period farmstead 10 km north-west of
Skutustaðir.
Results:
d34S values for Hofstaðir herbivores were 6‰ higher relative to those from Skutustaðir (d34S: 11.4 ± 2.3‰ versus 5.6 ± 2.8‰), while human d13C, d15N, and d34S values were broad ranging (-20.2‰ to -17.3‰, 7.4‰ to 12.3‰, and 5.5‰ to 14.9‰, respectively).
Discussion:
Results suggest that the baseline d34S value for the Myvatn region is higher than previously predicted due to a possible sea-spray effect, but the massive deposition of Tanytarsus gracilentus (midges) (d34S: -3.9‰) in the soil in the immediate vicinity of the lake is potentially lowering this value. Several terrestrial herbivores displayed higher bone collagen d34S values than their contemporaries, suggesting trade and/or movement of animals to the region from coastal areas. Broad ranging d13C, d15N, and d34S values for humans suggest the population were consuming varied diets, while outliers within the dataset could conceivably have been migrants to the
area.
herbivores and freshwater fish, while freshwater biota displayed δ13C values that were comparable with marine resources. Therefore, within this specific ecosystem, the separation of terrestrial herbivores, freshwater fish, and marine fish as components of human diet is complicated when only δ13C and δ15N are measured. δ34S measurements carried out within a previous study on animal bones from Skútustaðir, an early Viking age settlement on the south side of Lake Mývatn, showed that a clear offset existed between animals deriving their dietary resources from terrestrial, freshwater, and marine reservoirs. The present study focuses on δ13C, δ15N, and δ34S analyses and radiocarbon dating of human bone collagen from remains excavated from a churchyard at Hofstaðir, 5 km west of Lake Mývatn. The results demonstrate that a wide range of δ34S values exist within individuals, a pattern that must be the result of consumption of varying proportions of terrestrial-, freshwater-, and marine-based resources. For that proportion of the population with 14C ages that apparently predate the well-established first human settlement of Iceland (landnám) circa AD 871 ± 2, this has enabled us to explain the reason for these anomalously old ages in terms of marine and/or freshwater 14C reservoir effects.
in Mainland and Island Southeast Asia
Radiocarbon results from houses, pits, and burials at the SunWatch site, Dayton, Ohio, are presented within an interpretative Bayesian statistical framework. The primary model incorporates dates from archaeological features in an unordered phase and uses charcoal outlier modeling (Bronk Ramsey 2009b) to account for issues of wood charcoal 14C dates predating their context. The results of the primary model estimate occupation lasted for 1–245 yr (95% probability), starting in cal AD 1175–1385 (95% probability) and ending in cal AD 1330–1470 (95% probability). An alternative model was created by placing the 14C dates into two unordered phases corresponding with horizontal stratigraphic relationships or distinct groups of artifacts thought to be temporally diagnostic. The results of the alternative model further suggest that there is some temporal separation between Group 1 and Group 2, which seems more likely in the event of a multicomponent occupation. Overall, the modeling results provide chronology estimates for SunWatch that are more accurate and precise than that provided in earlier studies. While it is difficult to determine with certainty if SunWatch had a single-component or multicomponent occupation, it is clear that SunWatch’s occupation lasted until the second half of the AD 1300s.
revealing the very well-preserved remains of an Iron Age
settlement, the wetland context ensuring that the timber
structures have remained intact and that the detritus of daily
occupation survives for us to pick apart and understand. One of
the structures in this settlement is an exceptionally well-preserved
roundhouse, the material remains of which have been subjected
to a barrage of analyses encompassing the insect, macroplant,
bone and wood assemblages, soil micromorphology, faecal
steroids, radiocarbon-dating and dendrochronology. These will
enable us to address some of the key issues regarding the life
cycles of Iron Age roundhouses, from conception and
construction, use of internal space, nature of occupation and likely
function, through to abandonment. Critically, we are now able to
view that life cycle through the lens of a tightly-defined
chronology bringing us close to the ‘ …short-term timescales of
lived reality’ [Foxhall, L. 2000. “The Running Sands of Time:
Archaeology and the Short-Term.” World Archaeology 31 (3):
484–498].
proportion of their diet must have contained some non-terrestrial protein that produced a reservoir effect. As freshwater fish from Lake Mývatn and marine fish have similar bone collagen δ13C values, using only one isotope to differentiate between the two food sources was not feasible. Therefore, δ13C, δ15N and δ34S isotope values for various terrestrial animals, and marine and freshwater fish were incorporated into the Bayesian mixing modelling program FRUITS (Food Reconstruction Using Isotopic Transferred Signals) to more accurately determine the diet of each individual. Their 14C ages were then corrected for both a freshwater and marine reservoir effect before Bayesian chronological modelling was employed to provide a more robust time frame for when the cemetery was in use. Previous findings suggested activity began
shortly after landnám, however this study indicates that the cemetery was in use between the 11th and 13th centuries.
A previous multi-isotope study of archaeological faunal samples from Skutustaðir, an early Viking age settlement on the southern shores of Lake Myvatn in north-east Iceland, demonstrated that there are clear differences in d34S stable isotope values between animals deriving their dietary protein from terrestrial, freshwater, and marine reservoirs. The aim of this study was to use this information to more accurately determine the diet of humans excavated from a nearby late Viking age churchyard.
Materials and Methods:
d13C, d15N, and d34S analyses were undertaken on terrestrial animal (n=39) and human (n=46) bone collagen from Hofstaðir, a high-status Viking-period farmstead 10 km north-west of
Skutustaðir.
Results:
d34S values for Hofstaðir herbivores were 6‰ higher relative to those from Skutustaðir (d34S: 11.4 ± 2.3‰ versus 5.6 ± 2.8‰), while human d13C, d15N, and d34S values were broad ranging (-20.2‰ to -17.3‰, 7.4‰ to 12.3‰, and 5.5‰ to 14.9‰, respectively).
Discussion:
Results suggest that the baseline d34S value for the Myvatn region is higher than previously predicted due to a possible sea-spray effect, but the massive deposition of Tanytarsus gracilentus (midges) (d34S: -3.9‰) in the soil in the immediate vicinity of the lake is potentially lowering this value. Several terrestrial herbivores displayed higher bone collagen d34S values than their contemporaries, suggesting trade and/or movement of animals to the region from coastal areas. Broad ranging d13C, d15N, and d34S values for humans suggest the population were consuming varied diets, while outliers within the dataset could conceivably have been migrants to the
area.
herbivores and freshwater fish, while freshwater biota displayed δ13C values that were comparable with marine resources. Therefore, within this specific ecosystem, the separation of terrestrial herbivores, freshwater fish, and marine fish as components of human diet is complicated when only δ13C and δ15N are measured. δ34S measurements carried out within a previous study on animal bones from Skútustaðir, an early Viking age settlement on the south side of Lake Mývatn, showed that a clear offset existed between animals deriving their dietary resources from terrestrial, freshwater, and marine reservoirs. The present study focuses on δ13C, δ15N, and δ34S analyses and radiocarbon dating of human bone collagen from remains excavated from a churchyard at Hofstaðir, 5 km west of Lake Mývatn. The results demonstrate that a wide range of δ34S values exist within individuals, a pattern that must be the result of consumption of varying proportions of terrestrial-, freshwater-, and marine-based resources. For that proportion of the population with 14C ages that apparently predate the well-established first human settlement of Iceland (landnám) circa AD 871 ± 2, this has enabled us to explain the reason for these anomalously old ages in terms of marine and/or freshwater 14C reservoir effects.
in Mainland and Island Southeast Asia
Radiocarbon results from houses, pits, and burials at the SunWatch site, Dayton, Ohio, are presented within an interpretative Bayesian statistical framework. The primary model incorporates dates from archaeological features in an unordered phase and uses charcoal outlier modeling (Bronk Ramsey 2009b) to account for issues of wood charcoal 14C dates predating their context. The results of the primary model estimate occupation lasted for 1–245 yr (95% probability), starting in cal AD 1175–1385 (95% probability) and ending in cal AD 1330–1470 (95% probability). An alternative model was created by placing the 14C dates into two unordered phases corresponding with horizontal stratigraphic relationships or distinct groups of artifacts thought to be temporally diagnostic. The results of the alternative model further suggest that there is some temporal separation between Group 1 and Group 2, which seems more likely in the event of a multicomponent occupation. Overall, the modeling results provide chronology estimates for SunWatch that are more accurate and precise than that provided in earlier studies. While it is difficult to determine with certainty if SunWatch had a single-component or multicomponent occupation, it is clear that SunWatch’s occupation lasted until the second half of the AD 1300s.