SCIENTIFIC CORRESPONDENCE western Europe during most of the present -5 °C for approximately 1,500 yr after its throughout the Holocene. We maintain warm period, the Holocene, and only fall at a relative age of 7,600 yr, with only that the reconstruction of colder winter arrived there long after conditions one short colder episode during this conditions at Bispingen during the latter favourable to it were established, perhaps interval. This is consistent with the palyno­ two-thirds of the Eemian concurs with the promoted by the introduction of agri­ logical evidence of H helix and J. aquifoli­ palynological record and is the most parsi­ culture and associated land-use changes. um. Pollen of both occurs consistently in monious explanation of observed differ­ Nonclimatic shifts in the vegetation consti­ samples during and before this 1,500-yr ences in forest composition between the tute a definite source of error in palaeo­ interval, but is represented extremely latter parts of the Holocene and Eemian in climate reconstructions based on modem sparsely thereafter. The isolated occur­ this region. The principal features of the vegetational analogues. Control by indica­ rence of both in a sample with relative age reconstruction are consistent with the tor plants is therefore important. Field et about 5,000 yr coincides with a peak in GRIP ice core data15, and with data from al. also attempted a climate reconstruction reconstructed MTCO, the upper 95% con­ ocean and shelf cores11•12. The Bispingen for the Grande Pile location, about 640 km fidence interval of which reaches >-5 ° C. region is likely to experience climate SSW of Bispingen and 330 m above sea Given that these 'indicator' tax.a were not changes associated with such changes in level, which today has a MTCO of +0.6 °C. used in our reconstruction, their pattern of circulation and/or surface temperature in In this reconstruction the interval with occurrence provides independent support the Norwegian and North seas, although extreme winter coldness is absent. This is rather than contradicting our reconstruc­ there is evidence of Eemian climate fluctu­ also the case in the palaeoclimate recon­ tion. As to their wider occurrences in ations elsewhere in Europe16• Further stud­ structions based on Eemian pollen dia­ Eemian records, these can be systematical­ ies are needed to establish the extent, grams from Grande Pile by Guiot8 and ly evaluated only by considering their rela­ timing and duration of such fluctuations in Guiot et alY0, in which both MTCO and tive spatial and temporal contexts. We different parts of the continent. annual mean temperature during the make no claim that the palaeoclimate his­ Brian Huntley whole Eemian interglacial are similar tory reconstructed at Bispingen is univer­ Department of Biological Sciences, to present-day conditions. We therefore sal; indeed, we draw attention to contrasts University of Durham, suggest that the severe winter spell during with sites further south and/or west. Durham DH1 3LE, UK the Eemian at Bispingen proposed by Field Aaby and Tauber suggest that "the dis­ Michael H. Field et al. is an artefact caused by limitations in crepancies" they identify may result from Laboratoire de Botanique Historique et their climate reconstruction method. difficulties in finding appropriate ana­ Palynologie, Bent Aaby logues. Although, as we indicated1, this Universite de Droit d'Economie et des Henrik Tauber may apply to the more extreme cold oscil­ Sciences d'Aix-Marseille, Boite 451, The Natural Science Research Unit, lations reconstructed, the degree of analo­ 13397 Marseille Cedex 13, France The National Museum, Ny Vestergade 11, gy generally was comparable to Holocene DK-1471 Copenhagen K, Denmark samples. They further suggest that the absence of beech (Fagus sylvatica L.) from Lie detection HUNTLEY AND FIELD REPLY - Aaby and northern Europe during the Eemian result­ Tauber assert that the North Atlantic ed from unknown but nonclimatic factors SIR - The results obtained by R. Wiseman Ocean and Norwegian Sea were "at least and that the resulting difference in forest on lie detection, published in Nature 373, as warm as today" during the period of composition underlies the reconstruction 391; 1995, but which I read about in the "supposed severe winter". Recent high­ of colder winters than today. While we April issue of Monthly Nature, must be resolution records of sea-surface tempera­ agree with the latter conclusion, the source regarded as dubious, because they were ture 11, however, show marked cooling in cited by Aaby and Tauber in support of obtained in uncontrolled conditions via both areas no more than 5,000 yr after the their assertions about F. sylvatica 7 is contra­ polls in the mass media. It is well known stage 6/5e termination. In the Norwegian dicted by recent work13 showing F. sylvatica that the audience of every newspaper, Sea, temperatures were similar to those of occupies a very similar climatic range to radio or television station is different. And modern times during only the initial short that occupied in eastern North America by differences in frequency of lie detection warm period of Se, subsequently falling to Fagus grandifolia Ehrh. It has never, to our may be the result, not of differences of cues values comparable to those during glacial knowledge, been suggested that human used by observers (visual, verbal, vocal), stages 11. In the North Atlantic, early warm interference has influenced the distribution but of differences in the proportions of period temperatures around 1.7 °C higher of the latter. Furthermore, Prentice et al. 14 skills in the different groups of readers than present fell by around 2 ° C at the demonstrate that F. grandifolia maintained (intelligence, perspicacity and so on). same time11 • Such a sea-surface tempera­ its range in equilibrium with climate Another possible source of large bias is ture scenario is consistent with stronger the self-selection of people who respond­ mid-Eemian cooling in northern and 1. Field, H. M., Huntley, B. & Muller, H. Nature 371, ed to the request for participation. northeastern compared with western and 779-783 (1994). Numerous television viewers could not 2. Dansgaard, W. et al. Nature 364, 218-220 (1993). southern Europe and generally concurs 3. Kellogg, T. B. Boreas 9, 115-137 (1980). detect the lies, fewer Daily Telegraph with the mechanism we hypothesized1• 4. Iversen, J. Geo/. For. Stockh. F6rh. 66, 463-483 (1994). readers could, but many radio listeners Our hypothesis is also supported by 5. Muller, H. Geo/. Jb. A21, 149-169 (1974). were able to detect the lies. This outcome 6. Menke, B. & Tynni, R. Geo/. Jb. A76, 3-120 (1984). evidence of sea-surface temperature 7. Huntley, B. & Birks, H.J. B. An At/as of Past and Present is probably the result of selection, as dif­ instability in the North Sea during the Pollen Maps for Europe: 0-13000 Years Ago (Cambridge ferent proportions of people in the three Univ. Press, UK, 1983). Eemian; two marked coolin§ episodes are 8. Guiot. J. Palaeogeogr. Palaeolimnol. Palaeoecol. 80, groups may have failed to respond out of reported north of Denmark1 . 49-69 (1990). lack of enthusiasm, being too busy and so 9. Guiot, J., Pons, A., Beaulieu, J. L. & Reille, H. Nature Aaby and Tauber suggest that our 338, 309-313 (1989). on. The results obtained by Wiseman can reconstruction "is directly contradicted" 10. Guiot, J. et al. Pa/aeogeogr. Palaeolim. Palaeoecol. 103, be explained on this basis. by occurrences of ivy (Hedera helix L.) and 73-93 (1992). These are only two of the many possi­ 11. Cortijo, E. et al. Nature 372, 446-449 (1994). holly (Ilex aquifolium L.) pollen at Bispin­ 12. Seidenkrantz, M.-S., Kristensen, P. & Knudsen, K.L. J. ble sources of bias in an uncontrolled gen and in other Eemian records, as these Quat. Sci. 10, 77--82 (1995). study of this type. 13. Huntley, B., Bartlein, P.J. & Prentice, I.C. J. Biogeogr. 16, tax.a are today found only where mean 551-560 (1989). Vasily Vlassov temperature of the coldest month 14. Prentice, I. C., Bartlein, P. J. & Webb, T. Ill Ecology 72, Department of Aerospace Medicine, 2038-2056 (1991). (MTCO) is :2:: - 5 ° C. Careful examination 15. GRIP Members Nature 364, 203-207 (1993). Saratov Medical University, of our reconstruction shows MTCO is 16. Thouveny, N. et al. Nature 371, 503-506 (1994). PO Box 1528, Saratov 410601, Russia 28 NATURE · VOL 376 · 6 JULY 1995