document

Nature © Macmillan Publishers Ltd 1998

8

Other traffic may have run in the oppositedirection. For example, it is thermodynami-cally difficult to form peptides out of individ-ual amino acids in the open ocean, but amino acids have been converted to peptidesin experiments that are claimed to modelhydrothermal settings11. We don’t knowwhether amino acids themselves can form atvents, but they might have been producedcopiously at the Earth’s surface12, and wouldalso have arrived on meteorites13. So perhapsthey were delivered from the surface to thevents, and only there linked into peptides.

But none of these possibilities should beoverplayed; there is still too much we do notknow. For example, nitrites and nitrateswould have been created by a number ofsources on the early Earth, including light-ning in clouds of volcanic ash14, and thenreduced to ammonia by Fe2+ in the ocean15.These mechanisms may have produced asmuch ammonia as the mineral-catalysedreduction identified by Brandes et al.1.

It is intriguing to ask how early terrestrialhydrothermal environments differed fromthe wet, organic-molecule-rich environ-ments in the parent bodies of carbonaceouschondrite meteorites early in the Solar System’s history3. Evidence from mineralogyand organic chemistry makes it clear that theMurchison meteorite, for example, experi-enced liquid water for the first 10,000 years orso of its history, during which time its aminoacids were probably synthesized. Yet there is

no evidence for peptides in Murchison16, andin general it appears that prebiotic evolutionin that object did not proceed beyond simplemonomers. Why not? Does this cast doubt onthe hydrothermal-origins hypothesis, or wassomething missing, or the time too short?Within the deep interiors of large asteroids,where liquid water may have persisted for a hundred million years, could prebioticchemistry have proceeded much further3?This is, after all, comparable to the time available8 for the origin of life on Earth.Christopher Chyba is at the SETI Institute, 2035Landings Drive, Mountain View, California 94404,and in the Department of Geological andEnvironmental Sciences, Stanford University,Stanford, California 94305, USA.e-mail: chyba@seti.org1. Brandes, J. A. et al. Nature 395, 365–367 (1998).

2. Gold, T. Proc. Natl Acad. Sci. USA 89, 6045–6049 (1992).

3. Chyba, C. F. & McDonald, G. D. Annu. Rev. Earth Planet. Sci.

23, 215–249 (1995).

4. Carr, M. H. et al. Nature 391, 363–365 (1998).

5. www.jpl.nasa.gov/ice_fire//AO_002.htm

6. Stevens, T. O. & McKinley, J. P. Science 270, 450–453 (1995).

7. Anderson, R. T., Chapell, F. H. & Lovley, D. R. Science 281,

976–977 (1998).

8. Sleep, N. H. et al. Nature 342, 139–142 (1989).

9. Holm, N. G. Orig. Life Evol. Biosph. 22, 5–14 (1992).

10.Sagan, C. & Chyba, C. Science 276, 1217–1221 (1997).

11.Huber, C. & Wachtershauser, G. Science 281, 670–672 (1998).

12.Schlesinger, G. & Miller, S. L. J. Mol. Evol. 19, 376–382

(1983).

13.Chyba, C. & Sagan, C. Nature 355, 125–132 (1992).

14.Navarro-Gonzalez, R., Molina, M. J. & Molina, L. T. Geophys.

Res. Lett. 25, 3123–3126 (1998).

15.Summers, D. P. & Chang, S. Nature 365, 630–633 (1993).

16.Cronin, J. R. Orig. Life Evol. Biosph. 7, 343–348 (1976).

news and views

330 NATURE | VOL 395 | 24 SEPTEMBER 1998

Starch, derived from the roots of tuber-ous plants, is the staple dietary energyresource for the peoples of many trop-

ical countries. These plants, such as manioc(or cassava, Manihot esculenta; Fig. 1) andsweet potato (Ipomoea batatas), are widelyused, yet little is known about their origins.Undoubtedly they evolved in Central andSouth America, but beyond that there is littleevidence of where or when they were firstbrought into cultivation. The main problemfor archaeologists is the lack of fossils —unlike the grasses, tuberous plants have nopersistent parts that survive in sites of formercultivation or food preparation. But anapproach that may open up new opportuni-ties for tracing the roots of Neotropical agri-culture is now reported in the Journal ofArchaeology. Dolores Piperno and IreneHolst1 show that starch grains belonging tosome of the plants in question have survivedon the surfaces of prehistoric stone tools.

Old World archaeologists tracing the his-tory of the cereals have many advantages overtheir New World colleagues. Grasses storetheir starch in tough seeds that are easily

carbonized and preserved in an intact state,allowing them to be identified with preci-sion to species level. The grasses also containopal phytoliths — silica bodies with distinc-tive, often angular, shapes — that are rela-tively inert and survive after organic tissueshave decomposed. Moreover, the domesti-cated cereals produce pollen grains with features that allow them to be separatedfrom other grasses, providing opportunitiesfor the history of cereal-based agriculture to be traced. By contrast, the Neotropicalstarch-producing root crops, such as man-ioc, have no identifiable phytoliths, producefew pollen grains and do not carbonize, so ithas been difficult to work out their history.

Starch grains vary in size, shape andstructure, and it has been claimed that theyare distinctive enough to allow species deter-mination2. So if starch grains could be foundin a datable situation, they could provideevidence for the use of certain plants in thepast — including the Neotropical root cropswhose history has proved so elusive. Themost obvious place to look for such starchgrains is on the tools used for grinding and

Plant domestication

Getting to the roots of tubersPeter D. Moore

100 YEARS AGOSuppose I toss a penny, and let it fall onthe table. You will agree that the face ofthe penny which looks upwards isdetermined by chance, and that with asymmetrical penny it is an even chancewhether the “head” face or the “tail” facelies uppermost. For the moment, that isall one can say about the result. Nowcompare this with the statements we canmake about other moving bodies. Youwill find it stated, in any almanac, thatthere will be a total eclipse of the moonon December 27, and that the eclipse willbecome total at Greenwich at 10.57 p.m.;and I imagine you will all feel sure onreading that statement, that whenDecember 27 comes the eclipse willoccur; and it will become total at 10.57p.m. It will not become total at 10.50p.m., and it will not wait until 11.0 p.m.You will say, therefore, that eclipses ofthe moon do not occur by chance. Whatis the difference between these twoevents, of which we say that onehappens by chance, and the other doesnot? The difference is simply a differenceof degree in our knowledge of theconditions. The laws of motion are astrue of moving pence as they are ofmoving planets...From Nature 22 September 1898.

50 YEARS AGOThat Newton had a just appreciation ofthe work of Huygens and fullyunderstood it is significant, becauseHuygens signally failed to comprehendNewton’s full achievement, although herealized Newton’s greatness as amathematician and as an experimenter.He criticized Newton’s fundamental workon colour because it did not explain theultimate nature of colour — “Besides, if itshould be true that the rays of light, intheir original state, were some red,others blue, etc., there would still remainthe great difficulty of explaining, bymechanical principles, in what consiststhis diversity of colours”. He did notunderstand Newton’s “But to examinehow Colors may be explainedhypothetically is beyond my purpose”.Huygens himself wrote little about colour,since the problem as he conceived it, tofind a mechanical explanation, seemedto him intractable. … To say this is not todisparage Huygens, whose fundamentalachievements make a formidable list. From Nature 25 September 1948.

Nature © Macmillan Publishers Ltd 1998

8

news and views

NATURE | VOL 395 | 24 SEPTEMBER 1998 331

Daedalus

A magnetic speaker The crudest component of a sound systemis its loudspeaker. The most flawlessrecording, perfectly digitized, recoveredand ultra-linearly amplified, must still beheard via a clumsy cone of plastic or paperflapping about in time with it. Daedalusnow has a better idea. His new‘Paraspeaker’ has no moving parts at all.

The Paraspeaker exploits the fact thatthe oxygen of the air is paramagnetic. So itis attracted into, and compressed by, amagnetic field. Vary that field at an audiofrequency, and the pressure must vary insympathy. This, says Daedalus, is whymany transformers produce a steady hum,though they have no moving parts. The airaround the transformer is alternatelycompressed and released by its straymagnetic field, and speaks up at its linefrequency. Well — twice that frequency,actually. For, annoyingly, the inducedpressure varies not with the field, but thesquare of the field. This gross nonlinearitydoubles fundamental frequencies, andadds other frightful distortions as well.

Daedalus is undismayed. Electronicsignal-processing is so fast and precisethese days that this and other problems canbe sorted out electronically in real time.The Paraspeaker will simply be a large coilof wire set in a hole in a baffle-board; butthe electronics driving it will be subtle andcomplex. The audio signal will go througha square-rooter and programmable signalconditioner while it is still digital. Afteranalog conversion, a dedicated amplifierwill feed it to the Paraspeaker.

The Paraspeaker will be a ferociouslyinductive load; even the best analogamplifier may falter in driving it. Thesignal conditioner has the job of tweakingthe digital signal so as to cancel theaberrations that it will meet downstream.Once programmed to counter the flaws ofthe following amplifier, it will ensure that aperfect signal reaches the Paraspeaker. Soperfect sound will emerge. Its vacantcentral hole will be an open window ontransparent, flawless sound.

The hi-fi community will be entranced.Yet the major benefactors will be theboom-box community, and its victims. Forthe boom box has but one advantage over asystem using headphones: it avoids theirdiscomfort and blanking out of ambientsound. But Paraspeaker earphones, beingsimply hollow coils, do not obstruct theears. They can be worn 24 hours a day. Themusic addict will be able to enjoy hisnarcotic endlessly, without deafening therest of us with his damnable loudspeakers.David Jones

preparing food. It has been shown that thestone hunting weapons of prehistoric soci-eties may carry traces of the victim’s blood,and that the haemoglobin can be identifiedto species level3. It is possible that the roughsurfaces of grinding tools could, similarly,carry a residue of informative starch grains.

This approach was used by Thomas Loy4

in his research into prehistoric remains fromthe Solomon Islands. He examined stonetools, dating from 27,000 years ago, andfound starch grains belonging to Colocasiaesculenta (taro). This plant has a starch-richcorm (an underground, swollen stem base),and is now widely cultivated in the tropics.Loy’s finding provided circumstantial evidence for use of this plant as food in thelate Pleistocene and, possibly, even its earlydomestication. The work also confirmed thepotential of starch-grain analysis as a meansof tracing the history of early food plants.

Piperno and Holst1 have now appliedsimilar techniques to investigate manioc andarrowroot (Maranta arundinacea) in thehumid tropical lowlands of Panama. Theyexamined a number of archaeological sitesfrom which milling stones were available,and extracted starch grains from the crevicesin the rough surfaces of these stones using a binocular dissecting microscope and a needle. Not only did they find starch grains,but they also had considerable success inidentifying these on the basis of an extensivetype collection that they had established.The oldest site was a rock shelter wheregrinding cobbles dating from the pre-ceram-ic stage (8,000 to 5,000 years ago) were re-covered from sediments in which phytolithsof bottle gourd, squash and maize had been

found. Starch grains that compared well withthose of manioc were found on the cobbles,confirming the early use of this plant in the lowland tropics. The authors also foundstarch grains of maize and arrowroot onthese implements.

Much more work needs to be done on the taxonomy of starch grains, particularlythose from potential crop species. But thesepreliminary results seem to establish that avariety of starch-producing plants, includ-ing manioc, were being exploited in centralPanama about 7,000 years ago. Although wedo not know whether these plants were gathered from the wild, or whether they werecultivated, palaeolimnological work in thearea indicates some slash-and-burn clear-ance of the forest at that time. This strength-ens the case for the development of agricul-ture as an explanation for the presence ofthese plants. The occurrence of other knowncrop species in association with the maniocand arrowroot gives weight to this argument.

Most important of all is the confirmationthat recognizable starch grains can persistover many millennia in the Neotropics —this opens up new opportunities for deter-mining, with precision, how these cropspecies with an invisible past were domesti-cated.Peter D. Moore is in the Division of Life Sciences,King’s College, Campden Hill Road, London W8 7AH, UK. e-mail: peter.moore@kcl.ac.uk 1. Piperno, D. & Holst, I. J. Archaeol. Sci. 25, 765–776 (1998).

2. Banks, W. & Greenwood, C. T. Starch and its Components

(Edinburgh Univ. Press, 1975).

3. Loy, T. H. Science 220, 1269–1271 (1983).

4. Loy, T. H. in Tropical Archaeobotany (ed. Hather, J. G.)

(Routledge, London, 1994).

P. K

AR

UN

AK

AR

AN

/HO

LT S

TU

DIO

S

Figure 1 Ancient roots — heaps of harvested manioc (Manihot esculenta) for sale at a market.Piperno and Holst1 show that starch grains from manioc, and other Neotropical root crops, havesurvived on the surfaces of ancient stone tools.