© 1999 Macmillan Magazines Ltd

Materials theorists are making hugeprogress in relating macroscopic propertiesof materials to the fundamental atomicconstituents of matter. Both the complexityand the understanding revealed by studiessuch as that by Jhi et al.1 illustrate that,along the road to ‘materials by design’, wecan anticipate traversing some interestingterrain.Barry M. Klein is in the Department of Physics, University of California, One Shields Avenue, Davis, California 95616, USA.e-mail: bmklein@ucdavis.edu

1. Jhi, S.-H., Ihm, J., Louie, S. G. & Cohen, M. L. Nature 399,

132–134 (1999).

2. Toth, L. E. Transition Metal Carbides and Nitrides (Academic,

New York, 1971).

3. Klein, B. M. & Papaconstantopoulos, D. A. Phys. Rev. Lett. 32,

1193–1196 (1974).

4. Papaconstantopoulos, D. A., Pickett, W. E., Klein, B. M. &

Boyer, L. L. Nature 308, 494–495 (1984).

5. Kohn, W. & Vashishta, P. in Theory of the Inhomogeneous

Electron Gas (eds Lundqvist, S. & Marsh, N. H.) 79–147

(Plenum, New York, 1983).

6. Singh, D. J. Planewaves, Pseudopotentials, and the LAPW Method

(Kluwer, Boston, 1994).

7. Zunger, A. in Statics and Dynamics of Alloy Phase Transitions

NATO ASI Series B Vol. 319 (eds Turchi, P. E. A. & Gonis, A.)

361–419 (Plenum, New York, 1994).

news and views

NATURE | VOL 399 | 13 MAY 1999 | www.nature.com 109

— such as structural features (bondlengths), elastic behaviour and lattice vi-bration frequencies — has in many casesagreed with experiment to within a few percent. The only experimental input intothese calculations is an assumed crystalstructure, but even the ground-state crystalstructure can be determined by these meth-ods using a total-energy-minimum searchprocedure7.

The observed maximum in hardnessof carbonitrides occurs for a non-integralvalue of the valence electron concentration,corresponding to major deviations fromthe properly occupied NaCl lattice struc-ture. In general, such a loss of lattice peri-odicity would render accurate theoreticalcalculations impossible. However, one ofthe remarkable properties of the carbo-nitrides is their ‘rigid band’ behaviour,whereby the varying valence electron con-centration (alloying) can, to a very goodapproximation, be modelled by simplechanges to the perfect lattice calculations.This rigid-band-like behaviour is one ofthe reasons that the work by Jhi et al.1 waspossible.

What Jhi et al. have shown is that thereis a correlation between the behaviour ofthe elastic constants and, by implication,the hardness, with the filling of electron-bonding states formed by the carbon andnitrogen p-electrons and the transition-metal d-electrons. In greater detail, there iscompetition between the filling of a set ofstates favourable for elastic strength and aset of states having the opposite effect. Thiscompetition results in a maximum in elas-tic constants such as the shear modulus —which measures the ability of a material torecover from transverse stress — as a func-tion of electron filling, in direct correlationwith a similar maximum for the hardnessthat has been observed in experiments. Inparticular, Jhi et al. find that the maximumin the shear modulus occurs for a valenceelectron concentration corresponding to acarbon or nitrogen vacancy concentrationof 12%, in agreement with material testing.

Particularly interesting is that the bulkmodulus of the carbonitrides, which isdetermined by a homogeneous volumedeformation of the crystal, does not showany maximum as a function of valence elec-tron concentration. The authors point outthat as much as we like to think of the car-bonitrides, with respect to their covalent-like charge density and bonding, as beinganalogous to covalent semiconductors,there are differences. In semiconductor orinsulator alloys the hardness and the bulkand shear moduli vary in the same waywith electron concentration. But the exis-tence of two different types of energy bandsfor the least tightly held electrons in car-bonitrides leads to different responses toshear and volume deformations.

Assemblages of plants present ecolo-gists with some interesting conceptu-al problems. How can a number of

species coexist when they are tapping thesame set of environmental resources? Allrequire light energy, plus a source of atmos-pheric (or dissolved) carbon, together withwater and a fairly standard range of elementsfor their growth and development. So,cohabitation must depend on the plants’capacity to tap these resources in differentways, coupled with their ability to cope withthe stresses that any particular habitat maypresent. Although such issues have beenstudied in terrestrial habitats, aquatic eco-systems — particularly ephemeral (or tran-sient) aquatic habitats — have been neglect-ed. But Jon Keeley of the United States Geo-logical Survey Biological Resources Divisionin California has begun to redress the bal-ance. Reporting in Functional Ecology1, hedescribes his latest findings on the variousphotosynthetic processes found in thevegetation of temporary pools.

Keeley has looked at the photosyntheticecology of shallow pools for many years.In 1981, he discovered a quillwort, Isoeteshowellii, a small aquatic pteridophyte which,unexpectedly, showed the characteristics ofcrassulacean acid metabolism (CAM)2. Thisphotosynthetic system involves the accumu-lation of carbon using the enzyme phospho-enolpyruvate carboxylase as the initial stagein the fixation process, rather than the moreusual ribulose 1,5-bisphosphate carboxy-lase–oxygenase (which results in the forma-tion of 3-carbon products, leading to theterm C3 plants for those that use this system;Fig. 1). Crassulacean acid metabolism ismost frequently associated with arid, terres-trial conditions — the short-term storageof fixed carbon in the form of organic acidsallows plants to accumulate carbon duringthe night, enabling them to conserve waterby keeping their stomata (the pores throughwhich they take in CO2 and, consequently,lose water) closed during the day.

The potential advantage of CAM for an

Photosynthesis

Mixed metabolism in plant poolsPeter D. Moore

Figure 1 Photosynthetic strategies. In his studies of atransient shallow pool, Jon Keeley1 has found that aquatic plants use a variety of photosyntheticstrategies, allowing them to coexist and make best use of the available resources.

C3 plants

CO2 CO2 CO2

C4 plants CAM plants

Phosphoenolpyruvate (C3)

Phosphoenolpyruvate (C3)

Ribulose 1,5—bisphosphate (C5)

2x3—phosphoglycerate(C3)

Oxaloacetate (C4)

Transfer tovascular

bundlesheath

cells

Calvin cycle

In situ

Oxaloacetate (C4)

© 1999 Macmillan Magazines Ltd

aquatic plant lies in the high affinity of phos-phoenolpyruvate carboxylase for CO2. Thisallows the plant to compete for dissolvedCO2 more efficiently than C3 plants whenconcentrations fall to low levels. It also per-mits the plant to continue fixing CO2 duringthe night, when the concentration of carbonin the water rises. Low concentrations of CO2

during the day are relatively common inshallow water — photosynthesis by macro-phytes (larger aquatic plants) and algaequickly exhausts the supply of carbon, asCO2 diffuses much more slowly throughwater than in the atmosphere.

Keeley has now1 examined componentsof the vegetation assemblage from a pool inCalifornia. The pool has been artificiallymaintained over the past ten years by allow-ing it to fill with rainwater in January andthen to dry out in mid-April (this is termed avernal pond). Twenty plant species, whichwere originally introduced from a naturalvernal pond, have become established underthese conditions, and Keeley has examinedthe 15 most common of them, both struc-turally and for their photosynthetic proper-ties.

Keeley found that plants with floatingleaves are normally C3. A CAM system wouldbe of no advantage to these plants, becausetheir stomata are in contact with the atmos-phere, yet they are not short of water. Oneexception is the grass Orcuttia californica,which has aquatic, floating and terrestrialleaves, but uses a C4 system of photosynthesis(which also employs phosphoenolpyruvatecarboxylase as the initial carbon-fixingenzyme). The submerged leaves have all thecarbon-scavenging properties of the CAMsystem, whereas aerial C4 leaves are particu-larly efficient at high temperatures and highlight intensity. The two quillworts studied,I. howellii and I. orcuttii, use CAM photosyn-thesis, but they increase their capacity foraccumulating carbon by using their roots totake up CO2 from the sediment, where respi-ration by invertebrates and microbes enrich-es the supply. In this way, these plants tap asupply of carbon that is not immediatelyavailable to the rest of the photosyntheticfoliage.

Keeley also discovered that other aquaticplant species, such as Eleocharis acicularis,can act as either C3 or C4 plants. The C4 mech-anism has the advantage over C3 in that theplants can avoid photorespiration (whichleads to inefficient carbon fixation in thepresence of oxygen). If the concentration ofdissolved carbon is low and that of dissolvedoxygen is high, it could be valuable to avoidphotorespiration. As the pool dries out,however, the diurnal depletion of the carbonsupply becomes far less extreme, so the valueof the C4 mechanism declines and E. acicu-laris switches from C4 to C3 photosynthesis.But if conditions become very dry, especiallyin the heat of a California summer, the C4

mechanism may again be useful as a means ofwater economy.

Several of the amphibious plant speciesshowed marked changes in morphology,such as tougher stems and thicker cuticles, asthe pool dried out and conditions changedfrom aquatic to terrestrial. This findingshows that these plants can both tolerate andcompensate for the new stresses of a dry life.In general, the species that showed such mor-phological changes on drying were thosethat had evolved from terrestrial ancestorsrelatively recently. Species with a long aquaticpedigree were less morphologically plastic.So, the diversity of responses to environmen-tal fluctuations involves both physiologicaland morphological modifications.

Long-term studies of this aquatic systemhave shown that there is a considerable inter-annual variation in the composition of thepool’s plant community, probably associat-ed with year-to-year variation in the climaticconditions. Keeley believes that this variabil-ity itself may help to promote coexistence of

the various plants present. Here is an eco-system in which inherent instability, and adegree of unpredictability, generates diversi-ty among its components. It is a system inwhich niche diversification among plants,although subtle and somewhat occult, canbe discerned in the way that the availableresources are apportioned3. Given the easewith which this relatively simple ecosystemcan be manipulated experimentally, it is sur-prising that it has not previously been usedas a model to study the assemblage rules ofplant communities — for the variety andflexibility of photosynthetic strategies hereseems to be the key to their continuedexistence.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.uk1. Keeley, J. E. Funct. Ecol. 13, 106–118 (1999).2. Keeley, J. E. Am. J. Bot. 68, 420–424 (1983).3. Crawford, R. M. M. (ed.) Plant Life in Aquatic

and Amphibious Habitats (Blackwell, Oxford, 1987).

news and views

NATURE | VOL 399 | 13 MAY 1999 | www.nature.com 111

Astriking feature of many excitatorysynapses in the central nervoussystem is that the neurotransmitter

glutamate simultaneously activates recep-tors with very different pharmacologicalproperties. Notably, AMPA (a-amino-3-hydroxy-5-methyl-4-isoxazole propionicacid) receptors open and close rapidly andmediate fast signalling, whereas NMDA (N-methyl-D-aspartate) receptors have slow,voltage-dependent kinetics, are permeableto calcium, and are central to the inductionof long-term synaptic plasticity. On page 151of this issue, a report by Mainen et al.1 chal-lenges a widespread assumption about howglutamate activates NMDA receptors.

For AMPA and NMDA receptors todetect the release of glutamate from pre-synaptic terminals, the concentration of thisneurotransmitter in the synaptic cleft mustbe enough to activate them. Accurate esti-mates of the glutamate concentration afterexocytosis have been obtained by measuringthe sensitivity of AMPA or NMDA-mediatedsynaptic signals to rapidly dissociatingantagonists — the higher the concentrationof glutamate, the more easily the antagonistis displaced from the receptors2. These (andother) experimental approaches have con-verged on a picture where AMPA receptorsare activated by the release of a single vesicle,although the receptors may not be saturat-ed3–5. But NMDA receptors are predicted tobe almost completely occupied, the differ-ence being explained by their 100-fold high-er affinity for glutamate6. This conclusion

agrees with the results of experiments thatsimulated diffusion of a few thousand gluta-mate molecules in the sticky and tortuousenvironment of the synaptic cleft7.

The report by Mainen et al.1 comes as asurprise, because it concludes that NMDAreceptors are, in fact, far from saturatedby a single release event. The authors usedtwo-photon laser scanning microscopy tomeasure changes in the intracellular concen-tration of Ca2& in pyramidal neurons fromthe hippocampus8. They filled individualneurons with a high-affinity Ca2& indicator,then held them at a positive membranepotential using a patch-clamp pipette. Stim-ulation of the presynaptic axons triggeredCa2& signals that could be resolved at thelevel of individual dendritic spines. TheseCa2& signals were blocked by antagonists toNMDA receptors, but were not affectedwhen the intracellular Ca2& stores weremanipulated. Importantly, in a proportionof trials the presynaptic stimuli producedno postsynaptic response (a ‘failure’). Thisimplies that the Ca2& signals reflected theactivation of the NMDA receptors by theprobabilistic release of individual quanta ofglutamate.

The authors used a simple strategy to testwhether the NMDA receptors were saturatedwith glutamate. Given that glutamate dis-sociates extremely slowly9, the receptorsshould not respond to a second pulse ofthe neurotransmitter, delivered after a briefdelay, if the first pulse is enough to saturatethem. So, the Ca2& signal after two stimuli

Excitatory synapses

Neither too loud nor too quietDimitri M. Kullmann