In their reply to our recent paper (Blockley et al. 2000), Housley et al. (2000) make four substantial points. Firstly, they assert that our critique of their two-stage re-colonization model rests solely upon radiocarbon calibration. Secondly, and consequently, they point to problems with Late Glacial calibration curves. Thirdly, they argue that radiocarbon calibration should be advanced only for sound archaeological reasons. Finally, they state that our approach is environmentally deterministic and that we have demonstrated only a weak correlation between human demographic change and rapid climatic amelioration.
Housley et al. (2000) argue against the use of Late Glacial calibration curves, and in particular state that `it is because the calibration data are so heavily smoothed that Blockley et al. dispute our notion of a northward movement of people'. Calibration is not necessary to dispute the proposed northward movement. A moving sum based on point estimates creates a distribution of dates, the first `bin' of which was interpreted by Housley et al. (1997) as a `pioneer phase', and the mode as a `residential phase'. If the moving sum is re-plotted to include the 2[Sigma] error, it can be seen that these distributions substantially overlap (Blockley et al. 2000: figure 1b). Furthermore, if the earliest date in each region is plotted with its 2[Sigma] error, it can be easily shown that all these dates are contemporaneous, unless the Upper Rhine and the British Isles are considered in isolation. Simply applying the 2[Sigma] lab errors to the uncalibrated data demonstrates that there is no evidence for a systematic northerly migration, and that the model proposed is an artefact of the moving sum method.
In our original critique of Housley et al. (1997) we calibrated their dates using InterCal 93 (Stuiver et al. 1993) since it is not constructive to criticize a paper using data unavailable at the time of writing. Solely as a check, we also used the latest curve, InterCal 98 (Stuiver et al. 1998). In their reply, they defended the use of uncalibrated dates on the grounds that the marine-based InterCal curves are unreliable, particularly due to fluctuations in the marine reservoir effect. There are, however, a number of terrestrial curves (e.g. van der Plicht 1999; Wohlfarth 1996) which cover this period. Although there is not complete agreement, they conclusively demonstrate that the uncalibrated radiocarbon timescale in the Late Glacial is incorrect and non-linear. The consequent expansion and compression of calendar time means that the true chronological relationship between uncalibrated dates is not known. Archaeologists are no longer in a position to use uncalibrated dates to examine chronological patterning. This is the most important archaeological reason for promoting radiocarbon calibration.
In applying their two-stage recolonization model for northwestern Europe Housley et al. (1997: 26) assume a priori that the region was abandoned due to climatic deterioration in the Pleniglacial. This in itself is environmentally deterministic in that it does not allow for human ability to adapt to a harsh climate. Having assumed abandonment, however, the only explanation for repopulation is recolonization. We simply suggest that the Palaeolithic population adjusted itself demographically to the changes in resources in the region as a result of climate change. The reduced population in the region, which may have been primarily centred in the North Sea basin and other now-inundated low-lying regions, may have been too small for us to observe archaeologically. Since the submission of Blockley et al. (2000), …