that their rapid removal must be possible.The availability of site-specific antibodiesagainst citrullinated histones will allow atleast some of these issues to be addressed.And one final question: is the methylation of lysine residues in histone proteins alsoreversible? ■

Yi Zhang is in the Department of Biochemistry andBiophysics, Lineberger Comprehensive CancerCenter, University of North Carolina at Chapel Hill,Chapel Hill, North Carolina 27599-7295, USA.e-mail: yi_zhang@med.unc.edu1. Kornberg, R. D. & Lorch, Y. Cell 98, 285–294 (1999).

2. Jenuwein, T. & Allis, C. D. Science 293, 1074–1080

(2001).

3. Bannister, A. J., Schneider, R. & Kouzarides, T. Cell 109,

801–806 (2002).

news and views

population made a last stand in the BritishIsles before dying out 10,500 years ago, but the Siberian population persisted for another3,000 years.

What caused the extinction of so manylarge mammals 10,000 or so years ago?Human hunting3, changes in climate or veg-etation, or both4, are often proposed to becausal factors. But the “‘ragged’ nature” (touse Stuart and colleagues’ phrase) of theseLate Pleistocene extinctions, with isolatedpockets of populations surviving for longer,suggests that the extinctions have a complexecology, with no single mechanism respon-sible for the demise of every species in every location.

Theories for both the expansion and theextinction of Irish elk populations, forinstance, often focus on the animals’ hugeantlers, which weighed 40 kilograms andspanned 3.5 metres,making them 30% largerthan those of modern moose. It has been suggested5 that female Irish elk selected maleswith large antlers, as this might have signifiedan ability to find sufficient food to supportbuilding and shedding a rack each year.This ability would then be passed on to theirmale progeny.

But the large antlers, which contained asmuch as 8 kilograms of calcium and 4 kilo-grams of phosphate, would have posed alarge annual nutritional burden on bulls6.The antlers would also be physically unwieldyin dense forests7. So both physical and nutri-tional constraints probably restricted the

Palaeontology

Ecology of ice-age extinctionsJohn Pastor and Ron A. Moen

The last ice age saw the extinction of numerous large mammals — butperhaps not as many as was thought. The woolly mammoth survived to much more recent times, and so, it now seems, did the Irish elk.

Sabre-toothed tigers, mastodons,woolly mammoths — these and manyother spectacular large mammals are

generally thought to have become extinctabout 10,000 years ago, at the end of thePleistocene epoch, otherwise known as thelast ice age. But it’s becoming clear that someof these species clung on tantalizingly closeto the present day.Thomas Jefferson’s instruc-tion to Meriwether Lewis and William Clarkto search for live woolly mammoths in theAmerican West in 1804 was perhaps a littleoptimistic. But the species survived onWrangel Island in the northeastern SiberianArctic until some 4,000 years ago1, making itcontemporaneous with the Bronze Age XiaDynasty in China. On page 684 of this issue,Stuart et al.2 report that another charismaticice-age mammal that was thought to havebecome extinct 10,000 years ago — the giantdeer or Irish elk (Megaloceros giganteus) —survived in western Siberia to the dawn ofhistoric times. The finding lends weight tothe idea that there is no one explanation forthe so-called Pleistocene extinctions.

The Irish elk (Fig. 1) must have cut animpressive figure, standing more than twometres high at the shoulder — about the sameas a bull moose, the largest living member ofthe deer family. But when and why did itbecome extinct? In their investigation, Stuartet al.2 began by carrying out radiocarbon dating of five skeletal specimens, including acomplete skeleton of an antler-bearing male.By combining this information with maps ofthe specimens’ locations, they show that Irishelk were widespread in Europe — from Ire-land to Russia, and from Scandinavia to theMediterranean — before 20,000 years ago.

NATURE | VOL 431 | 7 OCTOBER 2004 | www.nature.com/nature 639

But by the last glacial maximum 15,000 yearsago, they may have been restricted to refugesin the shrub steppes of central Asia. Fromthere, Irish elk apparently recolonized north-western Europe following the retreat of theAlpine and Scandinavian ice sheets during aperiod of climatic warming. The European

Figure 1 Prehistoric giant — artist’s impression of the Irish elk.

NA

TU

RA

L H

IST

OR

Y M

USE

UM

,LO

ND

ON

4. Wang, Y. et al. Science doi:10.1126/science.1101400

(2004).

5. Cuthbert, G. L. et al. Cell 118, 545–553 (2004).

6. Zhang, Y. & Reinberg, D. Genes Dev. 15, 2343–2360

(2001).

7. Grewal, S. I. & Moazed, D. Science 301, 798–802 (2003).

8. Byvoet, P. Arch. Biochem. Biophys. 152, 887–888 (1972).

9. Henikoff, S., Furuyama, T. & Ahmad, K. Trends Genet. 20,

320–326 (2004).

10.Vossenaar, E. R., Zendman, A. J., van Venrooij, W. J. &

Pruijn, G. J. BioEssays 25, 1106–1118 (2003).

11.Nakashima, K., Hagiwara, T. & Yamada, M. J. Biol. Chem. 277,

49562–49568 (2002).

12.Bauer, U. M., Daujat, S., Nielsen, S. J., Nightingale, K. &

Kouzarides, T. EMBO Rep. 3, 39–44 (2002).

13.Chen, D. et al. Science 284, 2174–2177 (1999).

14.Strahl, B. D. et al. Curr. Biol. 11, 996–1000 (2001).

15.Wang, H. et al. Science 293, 853–857 (2001).

16.Schurter, B. T. et al. Biochemistry 40, 5747–5756

(2001).

7.10 n&v 635 AM 1/10/04 5:36 pm Page 639

© 2004 Nature Publishing Group

when transistor size reaches a few tens ofnanometres.

This intrinsic limitation of silicon devices has provoked much speculationabout the ‘next big thing’in information pro-cessing — might it be ‘spintronics’, whichuses an electron’s direction of magnetiza-tion, or spin, instead of its charge as the information carrier? Each spin could carry a bit of information — in effect, a single-spin transistor — which could lead to faster computers that consume less electricity,and alleviating the problem of heat dissipa-tion. In designing spintronic devices, we canexplore the exciting opportunities offered

news and views

Irish elk to productive open environments,with relatively tall willow and birch shrubsthat could be navigated but still supply sufficient calcium and phosphate for antlergrowth6,7. An inability to balance sexualselection for large antlers with nutritionalselection pressures for smaller antlers mayhave led to the Irish elk’s demise in the BritishIsles, particularly as the climate cooledrapidly and caused the vegetation to changeto short-statured and unproductive tundra6.

And what about conditions in Siberia,where the Irish elk evidently survived forsome while longer before its eventualdemise? Stuart et al. suggest that the neces-sary productive and open grass–shrub plantcommunity persisted in the eastern foothillsof the Urals, sandwiched between denseforests on the upper mountain slopes anddry grass steppes in the plains to the east (neither of which could have supported thenutritional demands of the large antlers).The authors also propose that a generalspread of the dry grass steppes over westernSiberia, and the persistence of closed forestsin the mountains, squeezed the grass–shrubhabitat to a sliver and contributed to thedemise of perhaps the last population ofIrish elk about 7,000 years ago. In contrast,woolly mammoths survived on the steppe–tundra of Wrangel Island well beyond theextinction of the Irish elk. Stuart et al. effec-tively link the appearance and disappearanceof Irish elk and woolly mammoths in the fossil records to palaeobotanical data show-ing climate-driven changes in the vegetationcommunities that are thought to have sup-ported these large herbivores.

In recent years, ecologists have shownthat large mammalian herbivores not onlyrespond to changes in the composition ofplant communities,but also themselves con-trol this composition and plant productivity,through nutrient redistribution and selec-tive foraging on preferred plant species8,9.Intriguingly, by trampling tundra mossesand stimulating grass growth through grazing, the woolly mammoth might havehelped to maintain and expand its own preferred dry steppe habitat, even when the climate was cool enough to support a mossytundra10.The shift to mossy tundra in Siberiamight have happened only after humanhunters extirpated woolly mammoths, thusrelieving both trampling pressure on mossesand competition with grasses that were pre-viously stimulated by grazing10. The Irish elkmight also have affected its own habitatthrough browsing and grazing, but thisremains an open question.

So the causes of the Late Pleistoceneextinctions of large mammals are complex.Climate warming or cooling may have set the stage through broad-scale vegetationchanges, and human hunting may have delivered the coup de grâce. But the animalsthemselves contended with pressures from

Materials physics

Doping control for nanotubesReshef Tenne

Adding guest atoms to inorganic nanotubes, known as ‘doping’,influences their room-temperature magnetic properties — propertiesthat could be exploited in ‘spintronic’ devices and computer memory.

Much of the technology underlyingour computers and electronicdevices is based on the transport of

electronic charge across sub-micrometrestructures made from perfectly crystallinesilicon. The dimensions of silicon-baseddevices, transistors in particular, are con-stantly shrinking (90-nm transistor gates arenow in production) and their integrationdensities are increasing — in line withMoore’s law, which predicts at least a dou-bling of the transistor density on a chip everytwo years. Although this trend has been fol-lowed for almost 40 years, the rate of increaseis expected to level off in the near future,

640 NATURE | VOL 431 | 7 OCTOBER 2004 | www.nature.com/nature

nutritional and sexual selection, exacerbated perhaps by how they and their competitorsaffected their own environments. Stuart andcolleagues’ work exemplifies the wealth ofdetail that it is possible to obtain on the spatialand temporal patterns of fossil distribution.This,together with the finely resolved recordsof vegetation change generated by fossilpollen studies,and the development of realis-tic simulations of the physiology and ecologyof extant and extinct large mammalian herbi-vores6,10, suggests that the time is ripe for amore comprehensive analysis of the ecologyof the Late Pleistocene extinctions. ■

John Pastor and Ron A. Moen are at the University

of Minnesota, Duluth, Minnesota 55811, USA.e-mail: jpastor@nrri.umn.edu1. Vartanyan, S. L., Garrut, V. E. & Sher, A. V. Nature 362, 337–340

(1993).

2. Stuart, A. J., Kosintsev, P. A., Higham, T. F. G. & Lister, A. M.

Nature 431, 684–689 (2004).

3. Martin, P. S. in Quaternary Extinctions: A Prehistoric Revolution

(eds Martin, P. S. & Klein, R. G.) 364–403 (Univ. Arizona Press,

Tucson, 1984).

4. Stuart, A. J. Biol. Rev. 66, 453–562 (1991).

5. Geist, V. Nat. Hist. 95, 54–65 (1986).

6. Moen, R. A., Pastor, J. & Cohen, Y. Evol. Ecol. Res. 1, 235–249

(1999).

7. Gould, S. J. Evolution 28, 191–220 (1974).

8. Hobbs, N. T. J. Wildl. Mgmt 60, 695–713 (1996).

9. Jefferies, R. L., Klein, D. R. & Shaver, G. R. Oikos 71, 193–206

(1994).

10.Zimov, S. A. et al. Am. Nat. 146, 765–794 (1995).

Figure 1 Vanadium oxide (VOx) nanotubes, seen in false-colour by scanning electron microscope.Krusin-Elbaum et al.1 show how the magnetic properties of such nanotubes can be engineeredthrough doping with guest atoms.

L.K

RU

SIN

-ELB

AU

M E

T A

L.

7.10 n&v 635 AM 1/10/04 5:36 pm Page 640

© 2004 Nature Publishing Group