Research reviews Biotic and abiotic consequences of differences in leaf structure

16 Pages·1999·0.21 MB·English

NewPhytol.(1999),143,3–18
Research reviews
Biotic and abiotic consequences of
differences in leaf structure
VINCENT P. GUTSCHICK
Department of Biology, MSC 3AF, New Mexico State University, Las Cruces, NM
88003, USA (tel ›1 505 646 5661; fax ›1 505 646 5665; e-mail vince!nmsu.edu)
Received 5 October 1998; accepted 9 April 1999
summary
Both within and between species, leaves of plants display wide ranges in structural features. These features
include: gross investments of carbon and nitrogen substrates (e.g. leaf mass per unit area); stomatal density,
distributionbetweenadaxialandabaxialsurfaces,andaperture;internalandexternalopticalscatteringstructures;
defensivestructures,suchastrichomesandspines;anddefensivecompounds,includingUVscreens,antifeedants,
toxins, and silica abrasives. I offer a synthesis of selected publications, including some of my own. A unifying
themeistheadaptivevalueofexpressingcertainstructuralfeatures,posedasmetaboliccostsandbenefits,for(1)
competitiveacquisitionanduseofabioticresources(suchaswater,lightandnitrogen)and(2)regulationofbiotic
interactions, particularly fungal attack and herbivory. Both acclimatory responses in one plant and adaptations
over evolutionary time scales are covered where possible. The ubiquity of trade-offs in function is a recurrent
theme;thishelpstoexplaindiversityinsolutionstothesameenvironmentalchallengesbutposesproblemsfor
investigatorstouncovernumerousimportanttrade-offs.Ioffersomesuggestionsforresearch,suchasontheneed
formodelsthatintegratebioticandabioticeffects(thesemustbehighlyfocused),andsomespeculations,suchas
on the intensity of selectionpressures for thesestructures.
Key words: leaf anatomy, leaf morphology, gas exchange, herbivory, leaf nutrients, chemical defence,
optimization, trade-off.
introduction (Ehleringer & Hammond, 1987); removing tri-
chomes can demonstrate their effectiveness against
Both between and within plants (individuals or herbivory (Kanno, 1996). Many other quantitative
species), leaves are diverse in structure. Variations benefits (and costs) have been investigated. In this
are prominent in, for example, linear dimensions, review, I concentrate on the costs, benefits and
dissection of the margins, dry mass per unit area, associated trade-offs for distinct structural features,
nutrient content, cell size, optical scattering and above the level of genes and biochemistry. Many
absorption,stomataldensityandapertures,presence features,suchastrichomesorstomata,showadaptive
or absence of trichomes, and cuticle composition. value for multiple purposes (for an example, see
Thisdiversityisdemonstrablyundergeneticcontrol, Press, 1999). For example, trichomes can function
including the plastic responses such as leaf size and both in energy balance and in defence, and in
mass per unit area in response to the light en- defence against both microbes and herbivores.
vironment. We presume that the differences are Trade-offs between costs and benefits, particularly
largely adaptive, either for acquiring resources as marginal costs and benefits, are expected to be
(mostly in photosynthesis) or in biotic interactions, close if function is well optimized by natural
such as retarding leaf herbivory. In some cases, the selection.Indeed,closebalancesareoftenobserved.
adaptive value of a structural feature can be dem- Thisclosenesssupportsthevariabilityofstructures,
onstrated directly, by using some innovative and it explains partly how very divergent structures
methods. Constraining leaf angle displays can dem- can provide adaptation to the same selection
onstrate the value of leaf angles for photosynthesis pressures in different plant species.
Printed from the CJO service for personal use only by...
4 V.P. Gutschick
1997). Aperture responds to at least three major
leaf structure affects resource
physiologicalstatevariables(Tardieu,1994;Tardieu
acquisition and use
& Simmoneau, 1998): (1) photosynthetic metab-
olites, so that conductance keeps pace with need for
The cuticle
CO substrate (in the leaf interior; Mott, 1988); (2)
#
The cuticle, present in almost every land plant, is hormonesorregulators,particularlyabscisicacidor
foremost a barrier against water loss as well as ABA (as a water-stress signal), primarily from the
against pathogen invasion. The cuticle also offers roots (Blackman & Davies, 1985; implicating cyto-
much protection against a loss of solutes to rain kinins;Zhang&Davies,1990;Tardieuetal.,1993,
(leachingbythroughfall),althoughsomedoesoccur, 1996);and(3)hydrauliclinkages.Hydrauliclinkage
especiallyinacidconditions(Pearcy&Baker,1991). is overall to the bulk water status of the plant but
Because leaves flex in the wind and other stresses, local linkages to the epidermal cells are responsible
the cuticle must be flexible or layered. Abrasion of for stomatal responsiveness to humidity (Haefner et
the cuticle in high wind can increase water loss by al., 1997). Bunce (1997) describes how the three
transpiration(Grace,1974;Pitcairn&Grace,1985). physiological variables can be linked and how the
The layer of wax or cutin is commonly rather thin linkagescanbededucedexperimentally.Itisnotable
(several lm), except in many xeromorphic plants, that the upper and lower (adaxial and abaxial) leaf
where it can reach 60lm (Ihlenfeldt & Hartmann, surfaces can differ markedly in stomatal density
1982). Consequently, the metabolic cost of con- and in physiological responsiveness (Pospı!silova! &
structingthecuticleistypicallyafewpercentoftotal Sola!rova!, 1980).
leaf construction cost. (This accounting excludes Jones (1998) distinguishes between three major
surfaceresins,whichcanmakeuphalftheleafmass adaptivefunctionsofstomata:optimizingthetrade-
in species such as Larrea tridentata.) The thickness off in taking up CO while losing water; controlling
#
does vary, and so does the corresponding water the risk of dehydration, particularly poising the leaf
permeability, by about one hundredfold (Schreiber waterpotentialabovethepointofcatastrophicxylem
& Riederer, 1996). This variation is often in cavitation(Tyree&Sperry,1988);andregulatingof
acclimation to water regimes (Turner, 1994). temperaturebytranspirationalcooling.Thesefunc-
tionswillbediscussed individuallyhere.A uniform
framework to explain all these functions simul-
Stomata
taneouslyis not yet available, either mechanistically
These dynamic pores are present at densities of or evolutionarily (that is, demonstrating the adap-
severalhundredpersquaremillimetre.Notallplants tiveness, or cruder optimality, of observed behav-
have them; indeed, Woodward (1998) poses the iour).
question,‘Arestomatanecessary?’Allmajortaxado
possess stomata now, having evolved increasing Optimization of assimilation rate}transpiration rate
densities of them though the Upper Carboniferous (A}E) by stomata. Stomata cost almost nothing to
period.Attheleast,stomataarerequiredforcontrol develop. Similarly, they cost little metabolic energy
of the exchange of CO for water vapour, which is tooperate(Assmann&Zeiger,1987).Consequently,
#
also inherently related to transpirational cooling. the costs and benefits in their operation are almost
Woodward (1998) begins with cooling as a need in whollythoseofresourceuse(water,CO ,nutrients)
#
full sunlight and notes that cooling is effective even thattheymodulate.Themostimportanttrade-offis
for plants of short stature, poorly coupled to the thatofphotosyntheticCO gainagainstwaterloss,or
#
atmosphere (as defined by Jarvis & McNaughton, AagainstE.Cowan&Farquhar(1977)proposedan
1986). This occurs because the stirred part of the optimizationprinciple,thatstomatashouldmaintain
troposphere (the convective boundary layer) typi- aconstantratioofthemarginalincreaseinCO gain
#
cally maintains humidities well below saturation, to marginal increase in transpiration:
even in the presence of much evapotranspiration.
ƒA}ƒEflconstant. Eqn1
Muchmoreattentionhasbeengivento CO –water-
#
vapour exchange, as will be discussed shortly. The basis of this principle is that there is a
Asapopulation,stomatacanbedescribedbytheir metaboliccost(proportionaltoE)ofmaintainingthe
arealdensityorbythefractionofepidermalcellsthat magnitude of E required for a given A: the cost of
they represent (stomatal index). They are further constructingandmaintainingrootsandothertissue.
described by their distribution of apertures. Inanexceptionalcombinationoftheoryandexperi-
Althoughthe area-averaged effect ofstomatal open- ment, Givnish (1986) showed how the constant
ing is to confer a conductance that is controlled could be evaluated for particular plants and growth
physiologically, not all stomata are open equally. A conditions. He included the effects of water stress,
typical histogram of apertures is unimodal, but not just water use, and extended the theory to
transients in light or humidity can induce broader allocation of root and shoot.
and even multimodal distributions (Buckley et al., How do stomata (or conductance per leaf area,
Consequences of differences in leaf structure 5
g) regulate both A and E? We must specify We thus obtain an expression for instantaneous
s
themicrometeorologicalenvironment:fluxdensities water-use efficiency of the leaf:
of photosynthetically active radiation PAR, near-
A
infrared (NIR) and thermal infrared (TIR); air WUEfl
temperature, humidity, CO concentration, and E
#
windspeed; and the resistance of any canopy C g! e fie
fl a bs leaf air
boundary layer between the leaf and our point of 1›g! }g g
measuringthemicrometeorologicalvariables.Three bs m bs
majorprocessesthenmustbemodelled:leafassimi- C g! 1
fl a bs Eqn2d
lation, leaf energy balance, and stomatal response e fie g 1›g! }g
leaf air bs bs m
to the leaf environment. The equations can be
In the final formula, the first factor is dependent
formulated to a very good level of accuracy,
mostly on the external environment: C }(e fie ).
althoughthesimultaneoussolutionismathematically a leaf air
(Ofcourse,leaf-interiorvapourpressureisafunction
challenging (Collatz et al., 1991; V.P. Gutschick,
of leaf temperature, which is affected by the water-
unpublished). A simpler, partly qualitative view-
vapourconductance,acomplicationthatwedismiss
point can be taken, to show that increasing g
s for now and that has an equally complex resolution
increases A but decreases water-use efficiency
(WUE): WUEflA}E. Consider first the transport in theory and experiment.) The second factor is
almost a constant, having a narrow range from 0.62
of water vapour, through the conductance of the
to 0.72. The final factor expresses the physiological
stomata (g) and the conductance of the boundary
s control of WUE by the factor g! }g . Physical
layers (leaf and canopy taken together, with total bs m
conductance g ). The total conductance is g fl conductance and biochemical capacity act in com-
1}(1}g›1}g ),busing the addition of series rebssis- pletelycomplementaryfashions.Inanyenvironment
tancess(1}cobnductance). Clearly, stomata control with boundary layers fixed by leaf dimensions and
windspeed, a higher stomatal conductance increases
only part of this conductance, a point to which we
g! andconfersbothhighA(Eqn2c)andlowWUE
must return. The transpiration rate is simply this bs
(Eqn 2d). The trade-offis so sharpthat most plants
total conductance multiplied by the difference in
controlg soastokeepC}C inaverynarrowrange,
water vapour ‘concentration’ from leaf interior s s a
to stirred air outside the boundary layers: Efl c. 0.7 for plants with the C$ pathway (Bell, 1982;
g (e fie )}P.Thisformulaincorporatesthetotal Wong et al., 1985).
bs leaf air StomatahavesomewhatdifferentleverageoverA
air pressure, P, so that E is formulated in terms of
and WUE than in this simple model, because there
thewater-vapourmolefraction.Conductanceisused
in the familiar molar units (mol m−# s−"), which are arebiophysicalfeedbacks:(1)anincreaseingsleads
todecreasedleaftemperature,whichflattenstherate
lessdependentontemperatureandpressurethanthe
old velocity units (m s−"; see Jones (1992), pp. of decrease of WUE; (2) within the leaf and canopy
boundaryresistances,increasedg humidifiestheair,
54–55).Next,considerthetransportofCO#,through s
similarlyflatteningthedropofWUE(anequivalent
similar physical paths, followed by its biochemical
statement is that stomatal control is diluted by the
reaction in mesophyll cells. The physical-path
boundary-layer resistance, as is apparent in the
conductance of CO from its partial pressure
#
formulaforg alreadydescribed);(3)lossesofwater
in free air (C ) to that in the substomatal cavity bs
a to soil evaporation or to competitors decrease as g
(C) is very similar to that for water vapour: s
g! sfl1}(1.6}g›1.37}g ); the factors 1.6 and 1.37 increases.Thislastfeedbackarisesfromtwoeffects;
bs s b (a) a more humid canopy decreases the gradient in
account for the lower diffusibility of CO than of
#
water-vapour pressure from soil to canopy air, and
watervapour.Thereactionrate,whichissimplyA,
(b) in the long term, high g confers higher A and
can be approximated (from full enzyme kinetics, as s
faster growth. The canopy closes earlier and sup-
in Farquhar et al. (1980)) as being proportional to
presses soil evaporation (analogous results were
the CO mole fraction in the substomatal cavity,
#
C}P, yielding Aflg C}P in units of mol m−# s−" reported as a function of crop planting density by
s m s Richards(1991)).Incertainsetsofconditions,WUE
(commonly quoted in micromoles, not moles).
canevenrisewithmodestincreasesing ;acaseina
Empiricallyandenzyme-kinetically,g issimplythe s
slopeofAagainstC}P,whichisfairlymconstantover fieldexperimentisreportedbyMeinzeretal.(1997).
s Ingeneral,thesefeedbacksdilutethecontrolofboth
modest ranges of C.
s A and E or WUE by stomata, for the canopy as a
With these definitions, we obtain
whole. For individual leaves, and especially for
Aflg C flg! (C fiC) Eqn2a
m s bs a s individual plants competing with the group, many
SolvingforC andexpressing Ainterms offree-air feedbacks such as changes in canopy humidification
s
CO# level and the conductances, we have: do not apply, and stomatal control retains much
C flC g! }(g! ›g ) Eqn2b value.
s a bs bs m
AflC g g! }(g! ›g )flC g! }(1›g! }g ) Eqn2c Discussions of the value of stomatal control in
a m bs bs m a bs bs m balancingtheinstantaneousratesofcarbon(C)gain
6 V.P. Gutschick
and water loss are incomplete. Much more general (h andC are,respectively,therelativehumidityand
s s
models can be constructed. If we consider only the CO concentration at the leaf surface, beneath
#
water:(1)watercostscanvarywithsoilwaterstatus anyleafboundarylayer;theslopemandtheintercept
(morerootmassisneededasthisdeclines);(2)water b are both measures of commitments to use water
availability rates can be constrained by the decrease andtofavourassimilationoverwater-useefficiency).
in water potential that threatens to cause xylem Stomata that respond according to this form do
cavitation;(3)therateofwateruse,andthusWUE, decrease g as evaporative demand (closely pro-
s
can be less relevant than constraints on total water portional to 1fih) rises. The apparent response to
s
availability (water volume in the potential rooting humidityactuallyderivesfromadirect,mechanistic
volume). Consider the change from conditions of response of g to transpiration rate (Mott & Park-
s
unlimited water bearing a cost of acquisition (root hurst, 1991) and, more specifically, to epidermal
function) to a limitation in volume. The value or transpirationrate(Saliendraetal.,1995;Haefneret
weighting of WUE then increases relative to A. al., 1997). However, as leaf temperature rises, the
Models with varied degrees of inclusiveness and resupplyofwaterisalsoactivated,suchthatthenet
complexity exist, predicting varied optimal pro- response of g is close to a response in (e fie )}(a
s leaf air
grammes for stomatal control (Berninger et al., temperaturefunctioncloselyparallelinge );thus,a
leaf
1996; Haxeltine & Prentice, 1996; Santrucek & responsetoe }e ,orrelativehumidity.Haefneret
air leaf
Sage,1996).Manypredicted(andobserved)changes al. (1997) demonstrated the realism of the full
instomatalconductance,frombothaperturecontrol hydraulicmodel,notonlyforbulk leafconductance
and stomatal density development, apply over the butalsoforitspatchybehaviourandforitstransient
spanofleafdevelopmenttime,notjustasaresponse behaviour,oppositeindirectiontothefinalresponse.
to the immediate environment. In a viewpoint
covering the longest time spans, we must consider Regulationofleafwaterpotentialbystomata.Stomatal
theevolutionofdevelopmentalcontrolsoverstoma- conductanceandtranspirationincreasetogether,and
tal density and physiological controls over aperture. thewaterpotentialdecreasefromsoiltoleaf,wfiw ,
s L
Robinson (1994) argues that plant families or taxa increaseswithE.Modelsofvaryingcomplexityand
that evolved earlier were constrained, and therefore inclusiveness show how w responds to g (Jones,
L s
did not develop as ‘efficient’ stomatal control. A 1992,especiallyp.158etsqq.).Manyofthesemodels
precisedefinition of efficiencymust bedevelopedto are used to argue that g can be set to maintain w
s L
clarify Robinson’s point fully. One must also ask in above the point of catastrophic xylem cavitation
what trade-offs these older taxa excel, so that they (Tyree & Sperry, 1988), which is very expensive in
are not extinguished by modern taxa. lost function in leaf and stem. (Partial cavitation, as
We should also consider other resources, such as is often observed in field conditions (Meinzer et al.,
N,aschangingthecost–benefitstructureforstomatal 1997) might nevertheless be within the optimal
control. The simple arguments already discussed behaviour.) However, stomata do not respond di-
assumedthatmesophyllconductance,orsomeother rectly to bulk w but to particular combinations of
L
measure of photosynthetic capacity or investment, epidermal and guard-cell water potentials (Haefner
was given a priori. The relative values and avail- etal.,1997).Aspecificstructural(hydraulic)linkage
abilitiesofwaterandNmustactuallybebalanced,so enforces this form of response. Leaves also respond
that g and g (or leaf N content) are optimized toABA(Tardieu&Simmoneau,1998)asasignalof
s m
together. This will be discussed further in a later root or soil water status; perhaps the ultimate
section. responseismoredirectlytosoilmechanicalstrength
For stomata to (nearly) optimize A}E, they must (Tardieu,1994;Masle,1998)thantowaterpotential
respond appropriately to environmental signals. To alone,withtheformerasabetterindicatoroffuture
respondtoA,theymustrespondtoaphotosynthetic prospects of water extraction. The resultant com-
metabolite. This metabolite must be near or in the binationofresponsetohydraulicsignalsandtoABA
guard cells (Jarvis & Davies, 1998). No metabolite can result in w that is stable, or at least kept above
L
has yet been identified, although it has been a ‘floor’ value, in many plants called isohydric
demonstrated that stomata respond to internal CO (Saliendra et al., 1995; Tardieu & Simmoneau,
#
partial pressure in the leaf (Mott, 1988). Stomata 1998).
must also respond to E or to the atmospheric Stomatal regulation in response to both A}E and
humidity (absolute or relative) that helps to de- waterstatuscanbejoinedviamechanisticresponses
termineE.Empirically,theresponseofg tothefull of g to photosynthetic metabolite(s), transpiration
s s
setofenvironmentalconditions,includinghumidity, and ABA, as already noted.
is often well approximated by the Ball–Berry model
(Ball et al., 1987), but not always (Jarvis & Davies, Regulation of leaf temperature by stomata. Leaf
1998; Dewar, 1995): temperature affects all manner of resource use.
Increasing temperature (T), up to an optimum, is
g flmAh}C›b Eqn3 desirable for activating CO assimilation and in-
s s s #
Consequences of differences in leaf structure 7
creasing the photosynthetic N-use efficiency. It can E to support a given A decreases. This devel-
also aid photosynthate transport and organ de- opmental responseis seen overevolutionary time as
velopment. Extremes are to be avoided. At high T, wellasinthelifetimeofsingleplants(Wagneretal.,
thermaldamage can occur;stomatado open athigh 1996). Of course, it is variable with leaf position in
T,perhapsadaptivelytolimitT.LowTcanleadto thecanopy;Pooleetal.(1996)cautionthatthismust
chillingandfreezinginjuries;insomeplants,chilling be accounted for in interpreting data, especially
aloneisnotdamaging,butitiswhencombinedwith palaeontological. The response of stomatal density
high light levels (Ball et al., 1991, 1997). seems to be independent of life form (e.g., herb or
Both stomatal conductance and leaf geometry tree) but dependent on exposure and on initial
affectleafT.Atsteadystate,leafTadjuststobalance stomatal density (Beerling & Kelly, 1997).
the energy fluxes per unit leaf area: Changes in other environmental variables such as
humidityorPARfluxdensityoverthedurationofa
0flQ+SW›Q+TIRfiQ−TIRfiQEfiQcc Eqn4 leaf’s growth also affect the stomatal density as well
as the stomatal control programme (the short-term
(Q+ is the net rate of absorption of shortwave
SW responsivenessofg toenvironmentalvariablessuch
radiant energy, which depends on spectral absorp- s
as radiative fluxes and humidity, as is expressed in
tivity and leaf display angle). The influx of thermal
Eqn 3) (Bunce, 1998). The control programme
infraredradiantenergy(Q+ )isalmostindependent
TIR (expressed, e.g., as magnitudes of slope m and of
of leafstructure orofdisplay angle, depending only
interceptbinEqn3)acclimatesatelevated[CO ].A
on surrounding temperatures and the nearly in- #
significantpartoftheacclimationmightbetoaltered
variant leaf thermal absorptivity. The next three
water status (Morison, 1998). Typically, the ac-
terms for energy losses all depend on leaf T,
climation preserves the ratio of internal to ambient
increasing (mostly nonlinearly) with T. Thermal
CO partial pressures, C}C (Morison, 1998). We
infraredlosses(Q− )increaseasthefourthpowerof # s a
TIR canrewriteEqn2d,forwater-useefficiency,usingA
T but are essentially independent of leaf structure,
as in the far right-hand side of Eqn 2a, to obtain
display angle or physiology. Evaporative cooling
(hQeaEt)iosfsivmapployritzhaetimonolaorftrwaantsepri.raTtiohnisractoeotliimngesrtahtee WUEflg!bgsC(ae(1fifiCes}C)a) Eqn5
responds to stomatal conductance (a component of bs leaf air
total conductance g! ; see the discussion around Giventhatg! }g varieslittle(seepreviously),we
bs bs bs
Eqns 2a–2d) and to T (because internal vapour seethatasambientCO pressureC risesatconstant
# a
pressure in the leaf rises exponentially with T). C}C , there is an increase in WUE (and in water
s a
Convective and conductive cooling to air (Q ) status).
cc
increases linearly with T. As stomatal conductance
increases, Q increases. This affords stomata a
E
modest control of T, over a range in the order of Trichomes
10(cid:176)C, depending on, for example, windspeed. I use
These leaf hairs function in defence (see later), but
theterm‘modest’becausetranspirationchangesless
also affect gas exchange and temperature. In many
than proportionally with g, because stomatal re-
s plants, trichomes decrease the absorption of short-
sistance is diluted in boundary-layer resistance and
wave radiation by leaves and keep them cooler
there is a negative feedback (increasing g also
s (Ehleringer, 1981; Baldocchi et al., 1983). The
decreases the vapour-pressure deficit, e fie in
leaf air silversword plant on Mt Haleakala, Hawaii, is an
Eqn 2a). Empirically, there is little evidence of a
exception:itusespartlyfocusedlightreflectedfrom
direct response to T under normal environmental
trichomes to keep its apical meristem very much
conditions,asexpressedinthegeneralsuccessofthe
warmer than ambient air, to aid its development
Ball–Berry and related models;inthese models,the
(Melcher et al., 1994). The principal cost of
onlyeffectsofTareonA.Theadaptivereasons are
pubescence to alter leaf T is decreased light in-
not clear.
terception. This might be a negligible cost in high-
light (light-saturated) environments, or even a
Acclimationofstomataldensityandcontrolprogramme benefit, from the avoidance of photoinhibition (see
to long-term environmental conditions. Stomatal den- Press (1999) for a further discussion of leaf pu-
sity decreases as atmospheric CO concentration bescence).
#
rises; the stomatal index decreases even more Trichomes also keep water droplets off the leaf
regularly(Morison,1998).Thisdecreaseisarguably surface and the stomata (Brewer & Smith, 1997),
veryadaptive(Kurschneretal.,1998),givenalinkto whichhelpstomaintainleafgas exchange(Smith &
stomatal conductance (note that the conductance McClean,1989;Brewer&Smith,1995).Incalcicole
contribution of individual stomata can also change, species,trichomesactassinksforexcesscalciumthat
e.g., if the size or maximal aperture changes). wouldotherwisecausestresstotheplantsandresult
Simply,assimilationisenhancedwithoutcost,orthe in stomatal closure (De Silva et al., 1996).
8 V.P. Gutschick
usingwater.Anextensivediscussion,pointingtothe
Overall leaf size, shape and display
abundant detail in the literature, has been given
Leaves range widely in linear dimensions, from elsewhere (Gutschick, 1997). Several interesting
millimetres to nearly 1 m; they also vary in shape, phenomenacanbesummarizedhere.Oneisthatthe
from nearly circular with entire margins to deeply optimal leaf angle varies with depth in the canopy
lobed or serrated margins. A highly dissected leaf (Loomis & Williams, 1969; Duncan, 1971;
margindecreasestheeffectivesizeoftheleaf. Small Niinemets, 1998b) and with the relative importance
size or dissection thus increases the boundary-layer of WUE over assimilation and growth rate (plants
conductance,whichisproportionaltoo[windspeed} can change solar tracking modes with changes in
(lineardimension)].ThefractionalcontrolofAand water status (Forseth & Ehleringer, 1983; Reed &
E by stomata is kept higher than in large or entire Travis, 1987)). Superior light-use efficiency is
leaves.Also,heattransportisfacilitated,sothatleaf achievedwithleavesthataremoreerect.Cropshave
temperaturesareheldclosetoairtemperature.This been bred for this trait (Trenbath & Angus, 1975),
can bear a cost, in that it decreases leaf cooling and which apparently gives up the stronger shading of
WUEgainsathightranspirationrates.Asabenefit, competitors afforded by more planophile leaves.
leaves suffer less extremes of temperatures, neither Canopiesasawholeshowdifferent(higher)light-use
highTinhighsunandlowtranspiration,norlowT efficiencies than individual leaves (Monteith, 1994).
atnight under radiativecooling (radiation frosts are Finally, it is inadequate to characterize a leaf’s light
a hazard; Leuning, 1988). All these effects are interception and light-use (or N-use) efficiency on
modulated by leaf position within the canopy, of the basis of the total of direct and diffuse in-
course. Deeper in the canopy, wind penetration is terception. The distribution between direct and
decreased, as is radiative input both in shortwave diffuselightisimportant,bothforsingleleaves(fine
andthermalradiation.Asaresult,thetrade-offsvary leaves that give diffuse solar shadows or penumbras
with position and so can the leaf shape. are beneficial; Gutschick, 1991) and for whole
Size as linear dimension and thickness affects the canopies (Leuning et al., 1998).
efficiency of resource use. Larger, thicker petioles
are demanded for broader and thicker leaves. ‘Sun’
leaves are thicker than shade leaves, for example.
Mesophyll structure, particularly total mass and
Costs of petioles actually make sun plants less
nitrogen investments
effective than shade plants of the same size, for
interceptingPAR(Sims&Pearcy,1994).Why,then, The development ofthe palisade layer(s) of cells, as
does sun architecture occur? Vallardes & Pearcy number and length, is most responsive to light
(1998) propose that shade plants would be more levels.Sunleavesaremarkedlythickerandcanhave
damaged by photoinhibition, from high PAR in- additional palisade layers compared with shade
terception on leaves of limited electron-transport leaves (Nobel & Hartsock, 1981; Thompson et al.,
capacity. Niklas (1992) argues that in one species of 1988). Interms ofresourceuse, developmentof the
plant, petiole investment is excessive for light mesophyll (palisade plus spongy mesophyll) repre-
interception, summed over the day. I suggest that sents foremost the investment in N and in total
the architecture might be closer to optimal if one construction costs. Therefore, the remaining dis-
weretoaccountforsunlightinterceptionbeingmore cussionwillfocusonNandontotaldrymassperleaf
valuable early and late in the day, when vapour- area.
pressure differences are smaller and WUE is larger. Nitrogen is allocated to leaves increasingly as the
Compound leaves require a higher investment in opportunity for photosynthesis increases, especially
support (rachis plus petioles) than do simple leaves with increasing PAR flux density. The declining
of the same area. Givnish (1978) argues that investmentinoldleavescanbereversed(Johnstonet
compoundleavesareneverthelessacheapdisposable al.,1969).Thisisclearlyadaptive,atleastforplants
structure in seasonally dry tropics. They decrease in which fast growth is valuable, and the patterns
waterlossfrom(absent)branchesinthedry season. alongthegradientofmicroenvironmentsinacanopy
Niinemets (1998a) also notes that petioles in com- havebeensoanalysed(Hirose&Werger,1987);for
poundleavesarelowinNcontentandaretherefore a first-principles derivation of the optimum in-
cheaper to construct in terms of the most limiting vestment in overall mass, see Gutschick & Wiegel
resource. (1988).AccompanyingtheincreasingmassofNper
Leafdisplayanglepresentsarichnessoftrade-offs leafareainhighsunlightisanincreasingpartitioning
in leaf function. Angles that favour high light to carboxylation enzymes at the expense of light-
interception(normaltothesun,andperhapsactively capturing chlorophyll complexes (Cowan, 1986;
tracking the sun) favour high efficiencies in using Evans, 1989), another clearly adaptive pattern. In
nutrients and N, but low efficiencies in using light general, the patterns tend to maximize canopy
(much light is intercepted at irradiances far ex- photosynthetic rate per total mass or per total mass
ceedingthelight-saturation pointofthe leaf)and in of N. These rates per mass are much more directly
Consequences of differences in leaf structure 9
relatedtorelativegrowthratethanareratesperleaf PAR absorptivity than is optimal) to deprive com-
area (Gutschick, 1987). peting plants of light. Conversely, plants with
TomaintainanoptimaldistributionofNbetween decreasedchlorophylllevelswereproposedasbeing
leaves, N in old leaves that have been overtopped superior in photosynthesis (Gutschick, 1984); field
must be moved to new leaves. Such resorption and testsborethisoutinsoybean(Pettigrewetal.,1989).
remobilizationofNiswidelyobserved,whereasthe A second determinant is niche differentiation
fraction reabsorbed is moderate, as it is for most between life forms, such as trees versus grasses.
nutrients(one-halforless,onamassbasis;Vitousek, Trees commonly have low A per leaf area
1982), notably less than the optimality models (Wullschleger,1993)(theexceptionscitedbyNelson
predict. A fundamental limitation on the remobil- (1984) remain exceptions). Their strategy of de-
ization of all nutrients is that enough of the velopment and of N use in particular differs from
biochemical and phloem-transport systems must be that ofgrasses. Inregions with seasonal leafing-out,
maintained to catabolize cell contents and to export established trees use N that was reabsorbed to the
nutrients in low-molecular-mass compounds. trunk(Ryan&Bormann,1982)foranearlyleafflush
The distribution of N investment, or total mass thatestablishesasuperiorclaimtolightinterception.
per leaf area, between different leaves on a plant is However, young trees are often outcompeted by
only part of the pattern. What sets the absolute grasses.Itremainspuzzlingtomewhyjuveniletree
magnitude of N per leaf area, N , or the related foliageisnotprogrammeddevelopmentallytoattain
a
fractional content of N, f ? There are substantial highN andA.Certainly,suchplasticityispossible,
N a
differences in N or f between plants in similar at least in shape, and might extend to N . De-
a N a
microenvironments (say, trees compared with velopmental plasticity can have little or no cost (J.
grasses in the same geographic location). One Schmitt,pers.comm.).Onepossibilityisthatgrasses
determinant is certainly functional balance between are always superior in A per unit mass, given their
rootandshoot.StrongphotosyntheticfunctionofN very low investment in supporting structures. (A
in the shoot, as in high light, dilutes N, whereas similar argument can be made for vines.) Thus, the
strong root function in acquiring N increases f . A competition for A per unit mass is lost, and trees,
N
‘passive’ balance sets f in this accounting committed to a woody structural base, however
N
(Gutschick, 1993; Gutschick & Kay, 1995). Rate modest at first, must express superiority in other
limitations on the uptake of N (or P) lead to low f . resource use or later in time.
N
High [CO ] increases shoot function and leads A third determinant is the trade-off between the
#
similarly to low f , as is widely observed, both in instantaneous photosynthetic N-use efficiency
N
current experiments and in comparisons of leaves (PNUE) and WUE. High N content, absolute or as
from earlier centuries with present-day leaves a mass fraction, can confer high WUE. It increases
(Pen4uelas & Matamala, 1990). Low-N soils often the mesophyll conductance and decreases g}g ,
s m
leadtothickorsclerophyllousleaves,whichnotonly which increases WUE (see Eqn 2d). However,
havelowf butgreatthickness,ordrymassperunit assimilation per mass of leaf is a modestly declining
N
area, and long leaf lifetimes. Givnish (1979) ex- functionofmassperunitarea(Gutschick&Wiegel,
plainedthecombinedpattern.Hedemonstrated,for 1988),atallmagnitudesofmassperunitarea.Thus,
example, that greater rates of increase of lifetime A assimilation per unit mass of N declines similarly.
occur at greater leaf thickness; only with great Furthermore, if high g decreases C, the carboxyl-
m s
thickness does the marginal benefit equal the mar- ationratepermassofRubiscoenzymedeclines,also
ginalcostofconstructingtheleafandsupportingits decreasing A per mass of N in the whole leaf. The
function. This pattern should nevertheless be re- magnitude of N that is optimal depends on the
a
examinedinthelightofrecentfindingsshowingthat relative costs and benefits of water and N. High N
construction costs do not vary systematically with availabilityfavourshighWUEandlowPNUE.Such
leaf structure (Poorter & Villar, 1997). trade-offsareseenindifferentshrubspecies(Fieldet
Theoptimumforaveragemass(orN)perleafarea al., 1983), but their generality is questionable
in the whole canopy is rather broad and flat. (Meinzer et al., 1992, Poorter & Farquhar, 1994).
Consequently, even substantial deviations from the
optimum bear little cost in photosynthetic per-
Kranz anatomy
formance.Gutschick&Wiegel(1988)proposedthat
canopies develop with a mass per unit area that is The specialization of leaf cells into mesophyll and
well below the optimum. The extra leaf area bundle sheath is of most marked value in allowing
developable per unit mass can decrease light avail- the C path of photosynthesis. Numerous articles
%
ability to competing plants. The hypothesis has not have reviewed the advantages of C photosynthesis
%
beentestedyet.Theoptimummassperleafunitarea in the efficiencies of using water, N, light, and even
increases with leaf area index; indeed, canopies do C substrates themselves (see, e.g., Ehleringer &
follow this trend. A broadly related hypothesis is Monson(1993),whoalsopresenttheargumentsthat
that leaves absorb excess light (by having higher thepathwayevolvedinresponsetolow[CO ]inthe
#
10 V.P. Gutschick
Mioceneepoch).The assimilationratesofC plants been investigated several times (Gutschick, 1984;
%
are far less sensitive to ambient [CO ] than those of Parkhurst, 1994). Modelling demonstrates that the
#
C plants, because C plants have a CO -concen- exactshapeoftheseprofilesisnotcritical,buthaving
$ % #
trating biochemical pump. The relative perform- anappreciablegradientallowsahighactualquantum
ancesofthetwopathwaysinthecurrentlyincreasing yieldinhighlight.Interestingly,thewhole-leafrate
CO levels are of intense interest. I still regard it as of photosynthesis is not strongly altered by the
#
difficult to explain why C plants have not replaced distribution of stomatal conductance between top
%
C plants even more extensively. A common ar- andbottomsurfaces,eventhoughCO isconstrained
$ #
gumentisthatC plantsaremoredown-regulatedby to enter the leaf at the surface with the lowest light
%
lowtemperatures.However,someC plantsfunction level (Gutschick, 1984).
%
wellatlowT;ifthiscanevolveinafewfamilies,why More details of leaf optical structure and the
hasitnotdone soin allthefamiliesinwhichthe C effectsonleafperformancearegivenbyEvans(1999)
%
path evolved independently? (This topic is also and Han et al. (1999).
discussed by Press, 1999.)
Leaf optical structure for UV protection
Joint ‘optimization’ of nitrogen content and leaf
Ultraviolet B and C flux is damaging to DNA in all
lifetime
cells, including plant cells. Leaves deploy UV-
Avarietyofargumentshavebeendevelopedtorelate absorbing compounds (especially flavonoids) in the
these traits. One early argument is that leaves with epidermis and throughout the mesophyll. As one
high N content per unit mass have high costs of might expect, the degree of protection is greatest in
construction and should require longer lifetimes to sunleavesmostexposedtoUV,andinlongest-lived
pay these costs back. A more careful analysis of leaveshavingthegreatesttime-integrateddose(Day
payback rates, and field measurement, shows the et al., 1993). The cost of UV screens is currently
reverse: high-N leaves are short-lived (Williams et being assessed. The demand for screening varies
al., 1989; Reich et al., 1997). At the other extreme, with depth in the leaf and also laterally, because
evergreens have long lifetimes that are commonly plant cells both scatter and focus radiation. Conse-
correlatedwithlowN .However,thistraitneednot quently, screening is being investigated as a fully
a
simplyoptimizePNUEevaluatedoverthewholelife three-dimensional phenomenon (Alenius et al.,
cycle. Jonasson (1989) found that, in five shrub 1995).
speciesoverwidelydifferentlocations,leaflifetimes
were not markedly long, nor was the efficiency of
leaf structure affects interactions
reabsorbing N from senescing leaves high. He
with herbivores and pathogens
proposed that evergreenness was an adaptation to
low rates of N supply from soil.
Mechanical defences
ExtensivereviewshavebeenpresentedbyGarnier
& Aronson (1998) and by Grime et al. (1997); and Waxy cuticle. This is the primary defence against
foradiscussionofnutrientutilityasaffectedbyboth microbial and viral invasion, as well as a barrier to
leaf lifetime and nutrient resorption see Eckstein et water and solute loss (Hadley, 1980). Thickness is
al. (1999). quite variable, as already noted; chemical com-
position and fine mechanical structure are perhaps
more important in functions such as providing a
Leaf optical structure, relevant to PAR absorption
barrier against water loss (Kerstiens, 1996). Fungi
Manystructureswithintheleaf,particularlythecell had a key role in the evolution of cuticle properties
walls, scatter light (Fukshansky, 1991). The com- (Taylor & Osborn, 1996). Penetrating the cuticle
bined scattering and absorption of light lead to a remains a key step in the fungal invasion of leaves
steepgradientoftotalfluxdensitywithdepthinthe (Mendgen et al., 1996). Fungi can degrade the
leaf.Thus,thephotosyntheticratealsovariessteeply cuticle (Commenil et al., 1998; Sugui et al., 1998),
with depth. However, novel leaf-sectioning exper- butleaves candetect the products(Schweizeretal.,
iments reveal that photosynthesis does not fall off 1996) and initiate other defences such as the
with the same profile as light (Nishio et al., 1993). hypersensitive reaction.
Thisistrueinparticularathighlightfluxdensities.
At low light, photosynthesis does follow the light Trichomes. These leaf ‘hairs’ may be extensions of
absorptionprofile,sothatnolightiswastedinexcess single epidermal cells. They may also themselves
absorptionandthequantumyieldreachesitslimiting be multicellular (Esau, 1965). Trichomes occur in
value near 0.05 mol CO mol−" photons. The almost every plant family (Johnson, 1975) and
#
adaptive value of the manner in which light ab- commonly have a defensive value (Levin, 1973)
sorption (pigment concentration) and enzyme ca- because they impede herbivores mechanically (in-
pacity are distributed across the leaf thickness has cludingmakingattachmentdifficult)orirritatethem.
Consequences of differences in leaf structure 11
Trichomescanalsosecretedefensivecompounds,as (Baur et al., 1991). Trichomes can also alter the
discussed later. Other functions of trichomes are susceptibility of plants to fungal infection. For
known: (1) acquiring resources (leaf trichomes example, Wilson & Hanna (1998) found a positive
hold water and absorb both water and nutrients in correlation of pubescence with severity of smut
some bromeliads (Raven et al., 1992), while root (Moesziomyces penicillariae) in pearl millet. En-
trichomes are commonly known as root hairs, hancedtrappingoffungaloroomycotalsporesmight
performing the same function); (2) limiting the be the cause.
interceptionofUVradiationoroftotalradiationand Estimating the total cost of trichome presence is
hence limiting leaf temperatures (Ehleringer, 1981; thereforesomewhatinvolved.Thecombinedcostof
Baldocchietal.,1983);and(3)excretingexcesssalt, trichomes and related chemical defences can be
as in mangroves and many other species in diverse moderately significant and therefore subject to
biomes. measurable selection pressure. Mauricio & Rausher
Trichomes can deter feeding by small inverte- (1997)comparedstandsofArabidopsisthalianawith
brates(Letourneau,1997).Kanno(1996)established andwithoutnaturalinsectenemies.Thepresenceof
thisforsoybeansattackedbyfalsemelonbeetles.In insects selected for greater production of trichomes
addition to correlating the extent of herbivory with anddefensivechemicals,theglucosinolates.Boththe
hairiness, he manipulated hairiness by shaving benefits of resistance and the costs in growth were
leaves. Finally, he showed that trichomes were resolvable. On the scale of a single plant generation
responsible,ratherthansolvent-extractablechemical (that is, showing plastic response, not natural
defences, by applying the latter in reciprocal treat- selection), Wilkens et al. (1996) related trichome
ments.Inotherplant–herbivoresystems,theprotec- density of the tomato to resource availability and to
tion afforded by trichomes is less marked. Gannon caterpillarforagingonleaves.Theytestedamodelof
& Bach (1996) found that the development of bean trade-offs of defence costs against benefits. This
beetle larvae on soybeans was variously retarded or model focuses on net growth potential, and it
accelerated by trichome density, according to the predicts the highest levels of defence at an in-
specific larval stage. Nevertheless, the hairiest termediateavailabilityofresources.Themodelheld
leaves induced a much higher mortality in larvae. forvariations inavailability oflight, but less closely
Interestingly,stingingtrichomesontwoherbaceous for water. Overall, trichomes seem to have modest
plantspecies(UrticadioicaandLaporteacanadensis) defensive value (quantitative, not absolute) and
didnotdeterfourspeciesofinvertebrateherbivores modest cost. They are clearly not required, given
(Tuberville et al., 1996). Such stinging trichomes that many plants have few or no leaf trichomes, yet
are known to be more effective against large they survive. Plants have other methods of defence
(mammalian) herbivores. or avoidance or leaf herbivory to use in various
Trichomesbearcosts.Substratesarerequiredfor combinations,asweshallseelater.Thedefencesare
constructing them, perhaps several percent of leaf effective, in that leaf herbivory is estimated at only
construction costs. Another cost is a ‘lost-oppor- 5–20% in some representative ecosystems (Golley,
tunity’ cost, the decrease in light interception. 1977;Schowalteretal.,1981).Thecostsoverallcan
Typically, the effect on photosynthesis of directly be substantial, particularly for perennials. Tropical
sunlit leaves is very small. These leaves are light- forestsseemtosufferhigherherbivoryratesthando
saturated, and the principal effect might be a temperate forests, especially in young leaves and
lowering of leaf temperature. A 1(cid:176)C decrease might understoryleaves,andmoreinseasonallydryforests
be caused by a 10% decrease in PAR absorptivity than in wet forests (Coley & Barone, 1996).
(depending on boundary-layer and stomatal con-
ductances). This T decrease can decrease light- Spines. Plants have a variety of sharp structures for
limited photosynthesis by approx. 6%. However, a defence: spines, which are modified leaves or
water-limited sun plant also improves its water-use stipules; thorns, which are modified branches; and
efficiency,toadegreedependentonairhumidityand prickles, such as sand burrs, which are epidermal
stomatal responses, for example. Returning to lost growths. Here discussion will be limited to spines,
photosynthesisasacost,thiswillbemostsignificant but the cost–benefit analyses apply quite directly to
for shade-lit leaves. A 10% decrease in PAR thorns and prickles. Spines deter large herbivores,
absorptivity for these leaves translates to a 10% primarily mammals. Mammals have two basic
decrease in photosynthesis. However, shade-lit methods of eating leaves: wholesale, with branches
leavesdonotcontributeheavilytophotosynthesisin (pruning),andselectively,forleavesalone(picking).
a whole plant canopy (as in crop monocultures). The formidable spines of Acacia tortilis in East
Even in a dense canopy, only approx. 20% of Africaseemtoprotectneighbouringtrueleavesfrom
photosynthesisisbyshade-litleaves,asindicatedin pruning by goats, but not from picking (Gowda,
many model studies. 1996). They preserve not the leaves themselves but
Because trichomes bear costs,their density seems the potential to regenerate leaves from meristems.
toberegulated(induciblebythreats)insomespecies Also preserved are carbohydrate and nutrient re-
12 V.P. Gutschick
serves in the branch tissues. These are smaller but (Richards & Caldwell, 1985). Clonal growth is
critical benefits. The cost of the spines was more another way of protecting a sufficient number of
thanrepaidbytheshootbiomasssaved.Anattractive meristems.
prospect is constructing and testing a model for the
wholelifecycle,toassessdeferredbenefitsandcosts
Chemical defences
such as these.
Classes of compound. Chemical classifications of
Tough tissues. Within leaves, the tissues with the defensive compounds are readily found in both the
highest mechanical resistance or toughness include introductoryliterature(Salisbury&Ross,1992)and
thelargeveins.Toughnessisconferredbythickcell the specialist literature (Rosenthal & Berenbaum,
walls and lignification in xylem vessels and in non- 1991). In addition to chemical structure, an im-
conducting fibre cells. Choong (1996) found quan- portant distinction (Feeny, 1975) is between the
titative relations of cell-wall volumetric content and quantitative defences present in large quantities and
fibre content to toughness in Castanopsis fissa. thehigh-potencyqualitativedefences.Intheformer
Toughness protected older leaves, whereas younger group,tanninsandresinshavehighmetaboliccosts,
leaveshadmoreprotectionbyphenoliccompounds. bothfortheproducingleafandfortheherbivorethat
(The protective value of lower N content in old detoxifiesthecompounds.Thelattergroupincludes
leaves was not evaluated.) cardiac glycosides (Boppre, 1978) and insect hor-
mones or analogues (Slama, 1980, 1987; Bowers,
Abrasives.Themostwidespreadabrasivecompound
1991).Thesetoxinsarepresentatlevels!1%indry
issilica,effectivelyrestrictedtothegrassesandsome
mass but are highly effective at deterring herbivory
minortaxasuchasEquisetum.Thesilicaingrassesis
(digitoxin) or at killing herbivores (insect juvenile
hydrated,thatis,opal(Baker,1960).Itsabrasiveness
hormones such as juvabione, which prevent normal
(Esau, 1965) limits herbivory on grasses, which
metamorphosis). The potent chemicals trade off
otherwise are highly attractive for their absence of
their high effectiveness against (1) limited range
woody tissue and of toxins. One of the most
(e.g.,cardiacglycosides,unliketannins,donotaffect
remarkable hypotheses, being examined currently
invertebrates) and (2) a negligible cost to the
(Kaiser,1998)isthatthewidespreadreplacementof
herbivore once a biochemical adaptation evolves to
othervegetationbygrassesintheMioceneeradrove
avertthedefence.Itiswellknownthatsomeinsects
theextinctionofNorthAmericanhorses,whichhad
can even turn the glycosides into protectants for
lower rates of tooth growth than horses on other
themselves against their own predators (Boppre,
continents.
1978;Holzinger&Wink,1996).Tanninsactbroadly
Stomataldesign.Doesthedistributionofstomata,or against both invertebrates and vertebrates but re-
theirindividualstructure,helptodecreasetheriskof quire large investments of C.
fungal entry into leaves? I once speculated that the Moretraditionalcategorizationsbymodeofaction
distributionofstomatafavouringabaxialoveradaxial resolve the antifeedants, the toxins, the anti-
surfaces might be so protective (Gutschick, 1984). digestants and the phytohormones. One antifeedant
However,themacroevolutionofplantsseemsnotto shared by several plant species is 2,4-dihydroxy-7-
favouraparticularpatternofadaxial:abaxialratioin methoxy-1,4-benzoaxazin-3,1 (DIMBOA) (Barry et
stomatal density (Beerling & Kelly, 1996). It seems al., 1994); resistance to corn borers, for example, is
that the structure of individual stomata is more quantitatively related to DIMBOA concentration.
important in deterring fungal invasion. The top- Potent toxins that are effective against almost all
ographyofstomatacaninducefungitoform(ornot herbivores (Rosenthal & Janzen, 1979) include
toform)appressoria,thehyphaethatarespecialized alkaloids, glucosinolates and cardiac glycosides.
topenetrateothercells(Readetal.,1997).Protection These first two classes of chemical protectants are
canbeofferedbyathickwaxlayer(Rubiales&Niks, most common in crop plants and their relatives
1996)orbywaxyplugsthatbearacostofdecreased (Letourneau,1997).Aswithalldefences,theycanbe
overcome by some herbivore species or races.
photosynthesis (Brodribb & Hill, 1997).
Herbivores must expend both energy and nutrients
Other structures, and overall architecture. A number indetoxifyingplantprotectants(Foley,1992),which
ofplantspecieshavesmallstructures(domatida)that gives the chemicals a remanent value as defences
harbour mites, which in turn can clean away fungal even against capable herbivores.
spores (O’Dowd & Willson, 1989). More generally, TheenergeticandNcostsofN-baseddefencesare
grasses protect their meristems from most grazing significant.However,asNavailabilitytoplantsrises,
animals by their position at or below soil level. these costs are diluted in a larger flux of metabolic
That is, aboveground, grasses are almost all leaf. In energy. One might hypothesize that defences are
addition, the bunchgrass growth habit allows the most supportable at high N, given this dilution of
rapid recovery of tissue growth after grazing; costs and the greater benefit of defence (plants are
carbohydrates are mobilized rapidly from roots more at risk at high N content because they are