3: Nutrient Acquisition

Alastair Fitter

3.1 Availability of nutrientS

Plants require about IS essential elements, and with a few importantcltceptions these arc obtained as ions dissolved in soil water. In physio­logical experiments, it is easy to induce a deficiency of any of theseelements, even of those:: such as Zo, Cll or Mo that arc only required inminute quantities, and on some soils these micronuuicnts may benaturally limiting. For example. ~t soils arc wholly organic and lack amineral reservoir; they often induce copper deficiency. To plants grow­ing in most field situations, howcver, the twO clements that mostcommonly determine plant performance are nitrogc:n and phosphorus.The ways in which these two elements becomc:s available to plants differstrikingly. Nitrogen is an abundant clement: air contains around 79%dinitrogen gas (N2), but this is unusable by eukaryotes, which must haveit 'fixed' into ionic form, tither as ammonium (NH4 ·) or nitrate(NO.!-) ions. Plants then conven these to organic forms, and theseorg3nic N compounds are eventually returned to the soil as litter, wherethey are acted upon by microbes (Fig. 3.1). Decomposition results inthe reappearance ofNH4 • ions in the soil and these may experience oneof four fatcs:1 some may be adsorbed onto clay minerals in soil;2 much is oftcn leached out of the rOOt zone into ground WOlter, butthis is more likely if;3 it is converted to nitrate by nitritying bacteria, which may be a majorpathway;4 finally it may be taken up by plant roots or microbes, depending ontheir activity.Importantly, there is no large mineral reserve ofN in soil, both becausethe principal ions of nitrogen (NO.!- and NH4 +) arc soluble in soilsolutions and because they arc rapidly converted by plants and micro­organisms to organic forms. Virtually all the N in soil, therefore, isprescnt as organic N in soil organic matter. Its releasc as inorganic ions(the process known as mineralization) depends on the activity ofdecomposers, and that in turn depends on a range of soil characteristics,such as pH, temperature, oxygen concentration, moisture concentra·tion, and so on. The speed at which mineralization occurs mainlydetermines the fertility of the soil in terms of N. If organic N is

[51]Plant Ecology, Second Edition Edited by Michael J. Crawley

© 1997 Blackwell Science Ltd. ISBN: 978-0-632-03639-4

: ..N01~ •

Nitrification

:~ SoilL7C--'-~c----,-----c--,-~-,---,----'Oe"alh' or~nie N

t· ,

. [JpbQ'1

. . .: :NH," I

I,:," I .,.: ,

.8~;e.,.; ;,__ .. ,. ''''''io. Uptake .' I)

...SOfL:. :.

Fig.3.1 A schematic representation ofrhc nitrogen cycle, showing the majortr.lnsformarions IMI dncrminc pbnt ;av;aibbility. 1be ntc-dctamining stq> is rypicallymincnliution of organic N. The cycle is dominated by organic compounds.

mineraliud, little of the resulting inorganic N remains in solution asions. NH•• ions follow the pathways listed above; ifconvened [0 NO j ­

by nitrifying bacteria then leaching is more likdy, since NO]- ions arcvery \\-~kly adsorbed by soil panicles, and hence easily moved in soil.Under anaerobic conditions, NOJ - ions may be converted [0 N 2 gas bybacteria such as some pseudomonads, which gain energy by using NO]­ions as an alternative electron acceptor.

Most plants can take up N as either NH." or N03 - ions, thoughsome species such as heathers, characteristic of acid peats where nitrify­ing bacteria a~ scarce, cannot cope with N03- because they lack theenzyme system (nitrare and nitrire reducnse) that converts NO)- toNH... (Dirret RI. 1973). N assimilation in these plants requires that theinorganic N be either taken up as NH... ions or convened to them.

In contrast to N, phosphorus is a rare: clement. As with mostnutrient clements (N is the important exception), there is no ,Hmo­spheric reservoir. Rocks typically contain much less than 1% P, and alltypes of phosphate ions are extremely insoluble in combination with thedominant cations in soils, such as aluminium, iron and calcium. Thephosphorus cycle is therefore dominated by a large pool of insolubleinorganic P (Fig. 3.2), aJthough as soils age an increasing amount isfound in the soil organic matter (Fig. 3.3). The availability of P istherefore a function of soil chemistry, whereas that ofN depends moreon soil biology. Other nutrients fall somewhere between these twoextremes, although potassium is different again since it is nOt meta-

Inorganic Organic

..... '. .",-. :1"'" ••• , '-:.~

[53JChapter 3

NutrientAr~ui.ritWn

OrganicPmsOIIOM.

Plants

H2PO.­absorbedinlO clays..,.P in mineral'";:::;:~

precipitates. I"etc.

,,,,::-:-:--'--'--':-;:-~..-:~~.~. --~~g;n~;--

•. compounds

: t .1- _.__ ~ "r~p~a~_,>/'. H p~. _ , 1)/Mlrlera.hzallon

I • I·'in soi solUlion I' .;

" .. .,., .

Fig. 3.2 A schematic rcpre5Cnbtw>n of the phosphorus eyt:1c, showing the majortransfonnations that deurmine plant availability. The concentration of inorganic Pinsoil solution is largely detennincd by lhe equilibrium between adsorbed and dissoh'ai P,whkh is snoogly in &\'our of the: adsorbed P. Inorganic componenl$l)'Pial1y dominateIhe eyt:1c.

Fig. 3.3 Total phosphorusdeclines and phosphorusfractions in the soil changealong a chronoscquence ofsoils on Ihe Franz Josephglacier in New Zc:aland. (FromWalk.er&Sycrs 1976.)

'00

~ 80 &..•,'2 60,X· ..~•(§ 20

o

• ,. Organic phosphorus- AI and Fe phosph8tes- C.lclum phosphates

......," ..

",....",.'-" ",

22000Age of soillyears)

bolized by organisms, never forms organic compounds and is thereforefound only in the inorganic phase of the soil.

3.2 Nutrient uptake by root systems

3.2.1 Transport through the soil

AU the resources used by plants reach the plant by physical transportprocesses: radiation in the case of photosynthetically active radiation

[54}ChRpur3NutrientAClJuisition

(PAR); diffusion in the case or COl and many ions in the soil (notablyH 2PO.-); and convection in the: case ofwartr and other ions in the soil.panicularly those pCCSl:nt at high concentrations. With the exception ofPAR, which is simply intcccl,"ptc:d by plant tissues, fhl:: supply of all otherresources is creatc:d by absorption. It is the: removal OfwaU:T from soil, aprocess ultimately driven by evaporation from the: Icaves, that bringsabout convective: transport of water and its dissolved ions through thesoil. Where: this convective: supply of ions is less than the rate: ofabsorption at the root surface:, depiction there: will create concentrationgradients in the soil and initiate diffusion down those gradients (Finer& Hay 1987). For all resources other than PAR.. it is we of a resourcethat c~ates supply.

Since mechanisms of interception of PAR and absorption CO2 aredtah with in Chapter 1, only absorption of water and minerals arediscussed here. Uptake of water is entirely passive: and depends on theexistence of a water potential gradient between root cdls and soil,generated by evaporation from leaves (see Chapter 2). Since the root cellwalls are freely permeable to water as far as the suberized endoderm is(and as far as the xylem in very young, unsuberized roots), the surfaceare:a for uptake is equivalent to that of all the root cortical cells. Actualtransport into the symplasm may occur anywhere in the cortex, al­though hydraulic considerations make transport through the cell walls(apoplastic transport) the more likely (Newman 1974). This watermovtment produces coO\'ecti\'e transport of dissolved ions to the rOOtsurface, where the cations may either be absorbed by negatively chargedcell wall materials or move passively into the symplasm down theelectrochemical gradient (root cortical cells have a standing potential of- 60 to - ISO mV); anions arc actively absorbed against that gradient.

Since soil water permeates the whole cortex and can therefore be

taken up into the symplasm by any of the cortical cells, whereas manyions tend to move into the symplasm only at or near the epidermis, theeffeeti\'t surface area for absorption of ions is much smaller than that forwater. The convective ion flux F is given simply by the water flux at therOOt surface, V, and the concentration of ions in tht soil solution C.:

(3.1)

(3.2)

When, for a particular ion, F is less than the rate of absorption by rootcells, depletion at the root surface will lead to diffusion of the ion fromthe bulk. soil. The diffusion coefficient for this depends on the moisturecontent of the soil (9), the tortuosity of the diffusion pathway if) andthe reactivity of the ion with the soil (soil buffer power, dC/de .):

dC,D- -D·9I· dC

where D* and D arc the diffusion coefficients for tht ion in soil solutionand ffte solution respectively (Nyc & Tinker 1977). Note that buffer

power enters the equation as its reciprocal. The effect of these relation­ships is that the availability of a particular ion in soil is determined bythe following factors.1 Soil buffer power (dC/dC /): this controls the soil solution COOl,:clltra­tion (Cd in relation to total soil concentration (C) and so the level ofconvective supply. In consequence ions that are strongly buffered in anysoil will tend to have low values of C, (unless the total concentration,C, is very high) and hence low rates of both convective and diffusivesupply. The extreme example is phosphate, where uptake is nearlyalways diffusion-limited. If butT!::1 pVWl:1 i:) very weak., <I:) i~ lhe l:ase fOInitrate (where C _ C I), convection may be inadequate if C, is low. Inthis event, rapid diffusion may lead to exhaustion of soil nitrate stocks.2 Soil water content: the demand for water by leaves, largely to main­tain energy balances, interacts with soil water content to control the rateof water uptake and hence convective supply of ions (Equation 3.1).Where soil water content is sufficiently low to limit convective supply,rates of diffusion are also depressed, because of the effect on diffusioncoefficients (Equation 3.2).3 Soil strncture: the compaction of the soil and the nature of the soilaggregates determine both the size and distribution of soil pores,affecting both the hydraulic conductivity of the soil and the length ofthe diffusion pathway ifin Equation 3.2).

These soil factors control water and ion availability per unit root areaor volume. Thus the actual levels of resources captured are stronglyinfluenced by the total amount of root and its three-dimensional distri­bution within the soil.

~:2.2 Transport across the root

The membranes of plant root cells contain transport proteins that canmove ions across the membrane, with great precision and at very highrates. The details of the transport process are described in several reviews(sec, for example, Clarkson & LOttge 1991). From an ecological stand­point, some of the imponant features of such mechanisms arc:1 they frequently involve movement against concentration gradients,so that the concentration of an ion inside [he root cells can be manytimes that outside, and yet uptake may continue;2 uptake frequently results in the loss of protons (H') to the externalmedium, with consequent changes in the acidity ofthe soil around roots;3 uptake requires the expenditure of energy, and there is therefore anenergetic COSt to ion uptake, which is responsible for part of therespiration in roots.The potential rate orion uptake by roots (especially excised roots, whicharc often used in physiological experiments) under ideal conditions isusually very high, and typically much higher than is observed in intactplants, and even more so than is measurable by plants growing in the

[55]Chapter 3

NutrientAClJuisition

[56]Chapter 3NutrientAcquisition

field. When lettuce plants were grown in carefully controlled conditionsand the nitrate concentration in solution was varied over a range from 0.5to 10 mmoll- I

, their nitrate uptake rate was constant (BJorn-Zandsua &Jupijn 1987); if excised roots arc given such a range of concentrations,the uptake rate would show a classic asymptotic response, increasinginitially as external concentration increased and then saturating. Thistells us that the rate of nitrate uptake is not controlled by the uptakernc=chanism, but by the plant itself: in other words, there is regulation.

This is an immensely important discovery. In effcct, we can regard theuptake syslefll ;I.:> l.-<l.pablc uf a..:yuirillg any iuns thal tIn: plant needs andthat are available at the surface of the root cell membranes. Plant <need' isdetermined by, for example, growth rate. When the plant is growingrapidly, it has a greater demand for nutrients, because metabolic pro­cesses consume nutrients already available. Consequently, the concentra­tions ofeither the ions themselves (as appears to be the case for phosphateand potassium) or a product of their metabolism (possibly amino acids inthe case of N) change in the phloem and hence in the root cells. Thisseems to send a signal to the ion transporters and down-regulates the rateof uptake. There is currently great research activity aimed at discoveringmore about the molecular nature of these transporters, and this shouldhelp to reveal more about the control mechanism.

However, much of this work is done on plants that arc capable ofhigh rates of growth, typically crop plants (and increasingly nowadaysthe tiny annual crucifer Arabidopsis thaliana, Brassicaceae). Many wildplant species are characterized by very low rates of growth, even underoptimal conditions. When presented with high concentrations of an ionsuch as N03 - at the root surface, these species do not down-regulateuptake to the same extent, and may consequently accumulate higherconcentrations of N in tissues than they can use. They thus appearsuperficially to be very inefficient: one measure of this is photosyntheticnitrogen-use efficiency (PNUE), i.e. photosynthetic rate expressed perunit N, rather than leaf area or biomass. Pons et al. (1992) grew fourgraminoids (three grasses and a sedge) at both low and high N supplyrates. Two species (Carexflacca, Cyperareae and Briza media, Poaceae)were inherently slow growing; two (Brachypodium pinnatum, Poaceaeand, especially, Dactylisglomerata, Poaceae) were capable offast growthunder good conditions. At the low rate of N supply, all species grewequally slowly, with a relative growth rate of around 0.05 day-I. WhenN was supplied at an optimum rate, all species grew faster: Carex andBriza approximately doubled their growth rate, but Brachypodium tri­pled it and Dactylis increased it fourfold. PNUE was higher for allspecies under N-poor compared with N-rich growth conditions. How­ever, when N supply increased, Dactylis showed the smallest reductionin PNUE because it had the smallest increase in leafN (Fig. 3.4). Theother species, to varying extents appeared to accumulate N in the leaves,but not to increase photosynthetic rate proportionately.

no~~ 200

<5 160§8" 160

"0 140

.~- 120woz 100~

•+

[57]Chapter 3

NutrientAcquisition

Fig.3.4 Photosynthetic nitrogen-use dfidcncy (PNUE, Ilmol CO2 (mol Nt' S-l) is lhe~tc at which a plant fixes COa as a function of its N content. PNUE dedines :IS learNconccntr:ltion increases, when plants arc fcniliz,cd: solid symbols on (he graph representplants grown at high N. The clTeet is much greater, however, for slow-growing speciessuch as Ihc grasses Brarhypodium pinnlltum (i'>.) and Briul mrdiIJ (';7) and the sedge Cart)(flat'" (0), than for the fast-growing grass Daayfisglomeraca (0) which shows thesmallest increase in leaf N conccntr.nion and linle or no decline in PNUE. (From Ponsrtlll.1994.)

It would be a mistake to equate high PNUE with fitness. In fast­growing plants, the bulk ofN in the leaf is present in the photosyntheticenzymes (ribulose bisphosphate carboxylase oxygenase, or Rubisco, inC3 plants; see Chapter 1); they are fast-growing because they have such alarge photosynthetic capacity. Slow-growing plants may use N in dif­ferent ways: for survival in saline habitats, for defence against herbi­vores, or as a store. The reasons why some plants have the potential togrow rapidly and some do not are poorly understood but there may wellbe trade-offs between maximum growth rate and ability to survive at lowrates of resource supply (see Chapter 4).

3.3 Responses to nutrient deficiency

3.3.1 Modifying the rhizosphere

If plants are deficient in nutrients, important changes in their physiol­ogy occur and these alter the interaction that roots have with thesurrounding soil. All roots are surrounded by a zone of soil called therhizosphere that has been chemically altered by root activity. Many rootactivities affect concentrations within the rhizosphere: for example,evolution of carbon dioxide from rOOt respiration will increase localconcentrations, whereas removal of water and ions due to uptake willreduce them (Dunham & Nye 1974; Hainsworth & Aylmore 1986).Most importantly, roots continually lose organic compounds, by activesecretion, passive exudation from epidermal cells or by the death ofthese and other cells. This supply of organic carbon fuels an activemicrobial community, which itself transforms other compounds and

[58]ChRpter3NutrinltAClJuisitW'I

furthc=r altc=rs the soil environment. Therefore, the rhirosphere is phy­sically, chemically and biologically distinct from the bulk soil. It isimportant to stick to this definition of the tcrm, which is increasinglyand incorrectly used as a synonym for the rooted zone in a soil horizon.

Nuuiem~c:6cient plants have quantitatively and qualitatively dif·ferent patterns of exudation and secretion, and this alters the mizo.sphere microflor.a. In some cues th~ changes in exudation arc dirc:ctresponses to the deficiency. lron-deficic:nt plants may increase the .s«~­

tion of siderophores, chemicals that can chelate Fe ions and prevc:ntthem being immobilized physicochemically by soil (Marschner er aI.1989). In consequence the Fe ions arc more available: for uptake by theroot. However, various microbes in the rhizosphere can bmh produceth~ siderophores themselves (Crowley er Ill. 1991) and degnde them(von Wiren et Ill. 1993), so thu there is uncertainty about the roleplayed by siderophores in soil-grown plants. Similarly, phosphate­deficient plants may increase the activity of their extt3cellular acidphosphatase enzymes (Boutin er al. 1981). It has been suggested thatthis is an adaptive response that enables the plant to decompose organicphosphates to release inorganic ions that can be: taken up by the root,bur there is little evidence to support this.

The most dt3matic impact of the root on the rhiwsphere is relaredto N nutrition. Since N can be: taken up as either NO)- or NH... , thebalance between these two ions has large implications for the electro­chemical balance of the root cells. If the anion is the main form, theplant excretes OH- ions to balance this; if the cation, then it is H' ionsthat are lost. Where NH.. • is the main source ofN, therefore, there canbe: marked acidification of the rhiwsphere, by .sevet31 pH units, aphenomenon that is easily visualiud using pH indicators (Plate I,facing p.366). This acidification may have profound effects on themicrobiology and chemistry of the rhizosphere and hence on plantcommunity structure (.see Chapter 14).

3.3.2 Resource allocation

The best chat3ctcrizcd response to deficiency of soil-based resources is achange in the pattern of growth, favouring root growth over shootgrowth. Most plants grown in ample light but with inadequate nutrientsupply increase root growth relative to shoot growth, producing a highroot weight ft3ction (RWF):

RWF .. root dry weighttotal plant dry weight

conversely plants grown in shade ha\'e low RWF. This is an intuitivelyadaptive response and can be: shown to represent optimal behaviourunder idealiud conditions (lwasa & Roughgarden 1984).

However, incrcasro root growth will not necessarily result in in­creased nutrient upt2ke, and hence alleviation of the deficiency. Since

nutrient uptake depends to a great extent on the geometry of the rootsystem, the greatest return on this investment will be achieved if rootlength is maximized. This implies that the production of fine roots willbe favoured, since they achieve the greatest root length for a givenweight. This can be measured as the ratio of length to weight, thespecific root length (SRL). Generally, plants grown in low nutrientenvironments have higher SRI.. than those well supplied with nutrients(Fig. 3.5), and this is apparently due to the production of thinner roots(Christie & Moorby 1975).

[59]Chspter 3Nutrient

Acquisition

soAnlhoxanthum

odOfSfUm

'.EEE

%0•! so

Holcus IsnafUS0~

1m 40 -

20 30 50

lime (days)

Fig.3.5 Roots generally become thil:ker ovc:rlll as soil fertility increases as shown bychanges in specific root length (SRL: length per unit weight) In a set of eight grassspedes. Time-trends of SRL were vc:ry variable, roots b«oming thicker with age in somespedes and thinner in others, but in almOSt all cases roots of plants al low fertilizeraddition rates were finer than those grown at high rates. Open symbols represent lowfertility, solid symbols, high fenility (from Finer 1985).

Open symbols. cootrol p1.na Solid -vrnbols. split root systems

0714212835

- ..•••••••••••I!!l.~••Fig. 3.6 Specific root length increases in split-root experimenu in those parts of rootsy5lems supplied with N in each of three gnu specks, but the behaviour of the pan ofthe root system deprived ofN differs between species. Circles desctibc: rootS in low Nafter day 14, squara those in high N. (From Robinson & RoNon 1983.)

Some roots arc shan-lived, but others survive for longer and give rise:to new branches. In most species this development is accompanied byradial growth, which means that newly produced roots arc finer onaverage than the existing roots on a plant. Therefore, an increase in rootgrowth following the reallocation of resources caused by nutrient defi­ciency will result in a fall in SRI. simply because there :are mon= youngroots in the root S)'5tem. Evidence from split-root experiments showsthat there is a general decline in mean root diameter in nutrient-poorsoils (Fig. 3.6).

If root systems in infertile soils have different proportions of youngand old fOOts compared with those in richer soils, this suggests that theoverall form or architecture of the root system may also change.Describing complex three-dimensional patterns is notoriously difficuJtbut a set of techniques derived from the m:uhematics of graph theorycan be used (see Box 3.1). This approach can be used to pn=diet thatcertain types of branching pattern are more expensive to construct thanothers, but are also more efficient at exploring soil (Fig. 3.7). Theincreased cost arises because herringbone patterns, which consist of asingle main axis with laterals arising from it, have a high proportion ofinterior links of large diameter, and the increased efficiency arisesbecause, despite their greater cost, they tend not to have links very doseto each other that will simply compete with each other for nutrientions. These: predicted patterns have been confirmed by experiments(F;g. 3.7b).

Adaptive responses to nutrient deficiency based on changes in re­source allocation arc, then=fore, more complex than might initially beimagined, since they involve not only increased root growth in terms ofmass. but also changes in the overall structure of the root system. A

Box 3.1 The architecture of root systems

In an architectural analysis, root systems are treated as sets of links oredges, which arc segments between branching points, nodes or vertices.Links can be exterior (ending in a meristem) or interior. A system has amagnitude, which is the number of exterior links or root tips, and itsbranching pattern or topology is described by the relationship betweenmagnitude and altitude, which is the greatest number of links that can befound in a single path from any root tip to the base of the root system.

[61]Chapter 3Nutrient

kquisition

1.1

8

2,Ibl

8

•

The two root sysrcms (a) and (b) represent extreme topologies. Each hasmagnitude 8: that is to say, each has eight exterior links. The magnitudeofan interior link is the number ofexterior links it feeds.

101

••

3

•

••

•

..~ ... ,.62

[62]Ch.pter3NutrientAClJuisitum

Box 3.1 Continued.

The branching patt~m (topology) an lx: chancrerized by calculatingthe path length of each link., which is the number uf links in [he paththat connects it to the base. In (e) the longest path length is 8; in (d) itis 4. This is caJ.lc:d the altitude of the system. If altitude (a) is plottedagainst magnitude (Il), a simple relationship emerges (e); systems witha single main axis and side branches (herringbone) have the largestpossible altitud~ for a given magnitude, while (d) represents theminimum possible altitude, found in a dichotomous system.

Most root: systems have intermediate root bnnching patterns, rcpre­u:nted by the middle line on the graph. The slope of the line d~ribc.s

the br;rnching pattern and reflects the rules by which the root systemdevdops. Finer (1995) gives a fuJlcr account of this approach.

100r---\"O'~"'II"""--;/--"lI

,., 10

"""100

more radical alteration to the geometry of the plant's nutrient gatheringsystem occurs in mycorrhizal plants.

3.3.3 Symbioses

3.3.3. J Mycorrhizas

Root diameter varies widely. The finest roots can be as great as 1.5 mm(e.g. in Drimys winrm' (Winteraceae); Baylis 1975) or less than 100 f.lmin the common weed CJ~pseJla bursa-paIrons (Brassicaceae) (Fitter &Peat 1994). This must represcnt one of the least well explained patternsof morphological variation in plants, though some at least is probablyrelated to mycorrhizal colonization (sec Fig. 3.8). Since the cost ofconstructing unit length of fine root is proportional to volume (assum­ing constant density), it is also proportional to the square of its

-" [63]"E • Chapter 3E "" • NutrientE0 .' • Acquisitwn;:. 10 ,0 •c • •• •0 8 • .~•• : •c • • ..~ • , , ... .- • • •. •.,

• • •• 0" i . •.. -. •< •• • • •w • •2

Fig.3.7 (a) The efficiency 0.3 0.' 0.5 0.' 0.7 0.8 0.9loJ Topological index

with which a root systemexploits soil in a simulationanalysis increasa as the root 0.9becomes mOll: herringbone in 0.8architecrure (high V;l.lucs of

< 0.7topological index (sec Box •• 0.'3.1». (b) The topological ,index of plant root systems arc

.. 0.5.,highest in plants of low I 0.' • •relative growth rate:, which are 0.3 •typically adapted to growth on ~ •0.2nutrient-poor soils and would

0.7benefit most from highexploitation efficiency. (From 0

0.00 0.08 0.10 0.12 0.14 0.16Fitter et aJ. 1991.) Ibl Mean relative growth rate ld-'l

diameter. A5 we have seen, roots vary by nearly two orders of magnitudein diameter, and so by nearly fOUf orders of magnitude in specific cost(tissue mass per unit length).

An economic approach to understanding this variation (Fitter 1991)must offset the obvious benefit of increased. SRL in fine roots by somedisbenefit. Coarse roots may be favoured because they have greaterdisease or herbivore resistance, because they are longer-lived or becausethey have greater growth potential. In addition, coarse-rooted speciesmay increase their nutrient acquisition ability by being mycorrhizal.Mycorrhizal fungi are symbionts that live in or on plant rootS andextend their mycelium into the soil. There are several different types ofmycorrhiza, but the most widespread is the arbuscular (sometimesknown as the vesicular-arbuscular or VA) mycorrhiza (see Box 3.2 andPlate 3, facing p. 366). In this association the fungus is able to obtainphosphate ions, the most immobile of all plant nutrient ions, from soiloutside the depletion zone in which the root is itself able to forage, andtransport it back to the root. The fungus obtains its carbon from theroot, and is obligately dependent on the plant for its energy require­ments.

The economic explanation for this symbiosis lies in the dimensionsof the fungal hyphae. They arc typically less than SlJ.rn in diameter, atleast an order of magnitude finer than the finest plant roots. In conse-

[64JChapter 3NutnmtAClJuisition

BOl: 3.2 Myconhizas

The term myurrrhjrA comes from twO Greek words meaning fungus­root. Mycorrhiz.as are intimate symbiotic asSOI.:iations offungi and rootS,

and mOSt plants form mycorrhizas under n3.tural conditions. There arcnumerous typeS of mycorrhizal associations, each involving distinctgroups of plants and fungi. All have evolved ~paratcly. The mOSt

widcsprcad t}1X of mycorrhiza is the muscular or vesicular-arouscularmycorrhiza. This is formed between mOSt plant species (probably:around rwo-thirds ofall spc=cies) and a InrncH:d group ot tungJ that arcmembers ofthc: Zygomycorina (order Gloma1<:s); about 120-150 spe­cies of the$(: fungi have been described, but there may be many morc.The fungus is an obligate symbiont that can only survive when linked toa foot from which it obtains all its organic food sources. Inside the root(Plate 3), facing p.366 the fungus forms characteristic arbuscules,branching structures that penetrate the root cortical cell walls andinvaginate the plasmalemma, creating a huge surface area of contact.Many fungi also form storage bodies called vesicles. The hyphae of thefungus either ramify between the cells or pass directly from cell to cell;outSide the rOOt they ramify in the soil, from where they transportphosphate ions back to the root. These fungi are also known to increaseplant uptake of some micronutrienu, provide protection against someroxic ions, fungal pathogens and grazing insects. and to affect plantwater relations. The exact balance of .all thCSe various benefitS ro .aparticular plant growing in the field remains unclear.

The best known rype of mycorrhiu is the eetomycorrhiu, formedbetween some forest trees, especially in bore.al .and cool·temperate for­ests, .and luge fungi ('toadstools') that are mainly members of the8asidiomycotina. Here the fungus forms.a sheath .around the outside ofthe root and spreads between the cells of the cortex, but does notpenetrate cell walls. Most obviously, mycorrhizal roots usually have adifferem growth form, often stubby and dichoromous, Plate 2, facingp. 366 and this is the only type of mycorrhiz.a that can be recognized bythe unaided eye. EClomycorrhizas probably involve less than 10% ofplant species, but more fungal species are k.nown than in the case ofarbuseular mycorrhizas.

The other main mycorrhizal types are taxonomically restricted, inone case to heathers and allies (Ericales: cricoid mycorrhizas) and in theother to orchids. Ericoid mycorrhizas involve a few fungi in the Ascomy­cotina and form dense infections in very fine 'hair-roots', Orchid mycor­rhizas are found only in the orchids, The fungi are mainly members ofthe Basidiomycotina. including some species th.at .are parasitic on trees,but m.any of the fungi .are permanently sterile and cannOl be reliablydassified. The fungi form coils in root cells, but what physiologic.alinteractions occur there is almost completely unknown. Intriguingly,the s.ame coils are found in some .arbuscul.ar mycorrhizal .associations,

quence, at least 100 times the length of hyphae can be constructed for agiven investment of resources, as compared to roots. The hyphae extendthe zone that can be exploited for immobile nutrients (especiallyphosphate) by several millimetrcs. The benefit to the plant is directlyproponional to the diameter of its roots, since this determines thedifferential in cost between growing more roots to obtain P andsubcontracting the task to fungal hyphae. Thick-rooted plants arc there­fore more likely to be mycorrhizal than fine-rooted species (Fig. 3.8).Arbuscular mycorrhizas (AM) are the most widespread type, found inaround two-thirds of all vascular plants. The other, perhaps benerknown type is the ectomycorrhiza, an association between trees and aquite different set of fungi, mostly members of the Basidiomycotina,many ofwhich are familiar woodland toadstools and often characteristi­cally found under a panicular species of tree. Boletus elegam, forexample, is typically found under larches Larix SDp., with which it formsa specific symbiosis. Not all fungi are specific, however; many, like theAM fungi, have wide host ranges. Ectomycorrhizas have marked effectson the root systems of the plant: often the colonized roots have muchreduced growth and they may fork dichotomously. The fungus does notpenetrate the root cells, forming instead a sheath around the outsideand then penetrating between the cells of the cortex, forming a struc­ture known as a Hartig net.

o Never or rarely mycorrhizal

o Occasionally mycorrhizal

• Normally mycorrhizal

[65]Chapter 3Nutrient

&lJuisition

"

.~ 15

~1;

ZEi

5

oLl.-'-_<100 100-200 20<1-300 ,300

Root diameter (I'm)

Fig.3.8 Plants that arc never or ~rely mycorrhizal have panicularly fine roots,confinning theoretical predictions. Data are for British plant species and come from theEcological Aor:a Database. (From Peat & Fitter 1993.)

[66]Chllpttr 3Nutrie"rkfJuuiricn

Although ectomycorrhizal associations arc certainly involved in Ptf2nsfer from soil via fungus to the: roots, their nutritional role isceminly wider. Some fung;al species can decompose organic maner andtransport the N back to the ute. effectively dosing the nutrient cycle.since the N is never free in soil and therefore much less Likely [0 be: lostby leaching or denitrification. Eccornycorrhiu5 arc found in rather fc:wecosystems, most notably the boreal forest. and it seems that they arcfavoured in those ccosystems where N deficiency is a major controllingfactor (Rt2d 1991).

A mird type uf mycorrhiza is different again. Eriwid mycorrhizas arefound in thc' roms of members of the: order Eriales (the heathers andtheir allies). These: fungi belong to the Ascomycotina and thc:rc is noc:vidence: for P transport to the plant. Decomposition of organic N isprobably important, and these fungi appear to be involved in protectingthe plant against toxic metal ions in soil- heathers typically live inextreme, often acid soils.

The fourth mycorrhizal type emphasizes the point that the conceptof mycorrhiza is a functional not a taxonomic one. AU orchids formmycorrhizal associations, but in this case the orchid appears to beparasitic on the fungus, since even in green orchids it is not possible todetect any carbon movement from plant [0 fungus. Some orchids haveno chlorophyll at all, so they clearly cannot supply fixed carbon to thefungus. Since they have no root system either, it is not obvious thatthese orchids are capable of acquiring any resources independently.These non-green orchids are usually referred to as saprophytes, but it isbetter to classify them as parasites. When it is discovered that the fungithey parasitiz.c: can themselves be parasites on other plants, it becomesapparent that the life-style of some orchids can be very complex.

3.3.3.2 Root nodule!

The other major root-microbe symbiosis involves N-fixing bacteria; it ismuch better understood than mycorrhizas, but much more restricted.Most members of the angiosperm family Fabaceae (the legumes), formnodules on their roots that harbour a group of bacteria known asrhizobia. In addition, a taxonomically diverse group of species in aoom17 genera and eight families form rather different nodules with anotherbacterial genus, Frankia (Table 3.1). Rhizobia were once thought to bea rather simple group taxonomically, but molecular techniques haverevealed that they do not form a single taxonomic unit (there are at leastfour genera).

Wild plants frequently appear to be limited by N and so, at fimsight, the selective advantage of possessing bacteria in rOOtS capable offixing atmospheric N1, which is extremely abundant, and converting itto NH.... , which is not, would seem very large. It is surprising thereforethat so few plant species have acquired this trait, and those that have are

Table 3.1 The genera of flowering pbnu thai can form nitrogen· filing :lS5OCi;uions withthe: acrionmycete F,."dia. The dUlrioolion is haphnard monomiully, lXcuring in adi,U$C group of familia., in a few genc:r.a wilhin lOOse familic:$ and only regularly insome of tho5c:.

Family

BetulaceaeCasuarinaceaeCoriariaceacDui!lCollcueElacagnaceac:

Ericxc~

Myric~x

RhamnKcxRosacuc:

GcneD in which mOA specksnonnally form nitrogen.filingsymbioses

AlnUI

CfrmriAiJflWtllS

Hippt.rphlU, !iJNplNn/u.

AI,",II, CAMprfmU.e;,.-'"uPmhM

Genera in whM:h wme spccic:$may form nitrogen·filingsymbioses

Tl'f1I'N, Diull'"RMItus, Drytu, Ctn.llrplU

[67JChtl.pttr3

NutrientAflJuuitio"

not more successful than they are. Although legumes arc extremdywidespread in agriculture, natural communities dominated by legumesor by Frtl."kia-associated species arc much less common. Species of alderAl"us may form almost pure woods and some tropical leguminous trees(e.g. TttTtI.-berJim·a) can be important canopy dominants, but manyN-fixers are rare or uncommon in the wild.

Part of the explanation for this may lie in the cost of the symbiosis.The fixation of N is an energy-demanding process and typically around10% of the carbon fixed in photosynthesis is diverted to the nodule andto the fixation process in an actively fixing, nodulated plant. Howserious this transfer of carbon is to the plant must depend on the extentto which photosynthesis is limited by sink activity (which seems oftell to

be the case) or by the direct impact of environmental factors, notablylack of light (sec Chapter 1). Very few N-fixing plants grow in shade.Plants growing in sun, however, are usually photosynthesizing belowthe maximum ratc that they could achieve, because photosynthetic rateis then dctermined by the metabolic activity of growing points andother plant processes, much in the way that ion uptake is controlled (secabove).

Another reason why N fixation may not be more widespread inplants is that the benefit can only be large in N-deficient soils; yet, by itsvery ellistencc, the N-fixing plant must enrich the soil with fixed N as itstissues contribute to soil organic maner, and so it will reduce the benefitit receives. Nitrogen fertilization of a community rich in legumesquickly reduces their abundance, as shown dearly by the famous ParkGrass experiment at Rothamsted Experimental Station (Fig. 3.9,Plate 12, facing p. 366 and sec Chap[cr 14). N-fixing plants themselvesachieve the same effect over a longer period of time by adding N [Q soil.This is shown in the successional sequence at Glacier Bay, Alaska, wherethe community is dominated by the N-fixing alder Alnus mspa from

,N~.N..P..K(SOI N..P..K(l001 N..P..K(1501P.K

•.,

•.,:: 0.25

j£

f 0.15

;..<£t: 0.05

1

[68JChRpter 3NutrientAt~n

Fig. 3.9 The rdarive abundance: ofkguma. meuural as lhe proportion of shootbtonuss in the June hay crop made up by TnltJIiNIft pr.lnW, LAtW,nu J-tmsis and lAtwtl1nfinJ.atJol, dc:dina with inc:reasing nitrogen, N, addition on those plou wh.ich recti\'(:both polusium, K. and ph05phorus., P, fertilizers (the 6gu«, in pa«nthais islhc Napplication rate in kg/ha/yr). The control plots. whkh m:c:ivc no nuttient fertilizers at

all, have more legumes than the plou receiving Ihc hishcsl rates of nitrogen, bUI lessthan other fertilized pIOIS; kgumcs on the control plot an: P limited. Data trom Crawley& Silvcrtown, unpublished results.

about 30 to 80 years after glacial retreat. During this period, soil Nconcentrations increase rapidly, greatly improving the competitive abil.ity of spruce Piua sirciunsU, which then replaces the aJder (Fig. 3.10).

3.4 Heterogeneity

3.4.1 Patchiness

Because soil is a solid medium, trampon processes within it are slow.Nutrients are therefore rarely evenly distributed, either in time or inspace. To understand how plants acquire nutrients from soil, it isimportant to consider the distribution of nutrient availability and howplants respond to that.

Orgenie Bnd mlnerelhorizon.

Mineral soil only------- -- --//-.L_--"",--,,,M,,---,,,,,,,--,,,,,

Soil age !yeersl

Fig. 3.10 Nitrogen builds up~pidly in the developing soilson newly aposcd glacialmonines at Glacier Bay,A1ub, especially during Iheperiod from aboUl 30 10 80yars after (Olonizalioncorresponding to the rimewhen N-lixing alder (AlJIIIS

crisptt) is dominant. (FromCrodcr& Major 1955.)

Heterogeneity in time arises because of the seasonal nature of manyenvironments. This imposes a regular pattern on many processes, in­cluding decomposition. In temperate ecosystems there is a pulse oforganic matter deposition in autullln and often a peak ofdecompositionin spring or summer. This leads to a clear pattern in the production ofinorganic N (Fig. 3.11), irregularity is imposed by weather variation.

Heterogeneity in space is caused by the pattern in which organicmatter is deposited: leaves and shoot material arrive on the surface andtend to produce a regular gradient in nutrient availability from thesurface downwards. Roots, in contrast, though they may be concen­trated at the surface, occur as discrete patches throughout the soilprofile, providing local hotspots of organic matter and decompositionactivity. The same effect is achieved when earthworms drag leaves downinto soil and when individual soil animals die. The net result of theseprocesses, and of inherent variation in the distribution of soil minerals,is often pronounced spatial pattern, which can be revealed by carefuland stratified sampling procedures (Fig. 3.12).

[69JChapter 3

NutrientAcquisition

Fig. 3.11 NitT:l.leconcentration in the: soilsolution in a de:ciduouswoodland ne:ar York, UK,fluctuates seasonally, withpeaks in e:arly summer and late:autumn corresponding toperiods of maximummicrobia! activity (summe:r)and resource availability (Ie:affall). (R. A. Farley,unpublishe:d data.)

• 1993-94

• 1994-95

It is possible to classify all this variation in terms of the attributes ofthe patches (Fitter 1994). Patch attributes are distribution, extent andnumber. Patches may be distributed randomly, regularly or be clumped,in both space and time; they may be small or large, short- or long-lived;they may be abundant or rare in space or occur frequently or rarely intime. When these attributes arc combined it is obvious that each soilmay have its own unique patchiness, and plants need to be able torespond to these patches effectively.

3.4.2 Response to patches

When roots encounter nutrienr.rich patches they may display a prolifer­ative response that seems intuitively adaptive (Robinson 1994). This

0-<365.<.(0

.<515,.OO!\:;,-----,J,,--==--'"

2.00 2.25 2.50

Organic maner 1%)>.5.,,---===,===---

o7.25

'2'

,.oo~\;;;__----;~--~2.00 2.25 2.50

_ 1.00~= ~_=_""""",

.§. 2.00 2.25 2.50

~~ location of tussock~ 7·50'r---'''-''''-i-,"'''-'''------,

[70]Chapter 3NutrientALquirition

length lml

Fig. 3.12 Soil d\J1noao:riKiQ.show distinct spatial pattern, in this casc: around a singklUssockofthc grusp~ spi€lfr". Soil P, K and OIpnic maner are all grateraround the tussock.. (From Jackson & ~ld\\."(:11 1993.)

rcsponSl= is easy to demonstrate with species such as cc:reaIs (barky,wheat. maize) and with numerous other species, In a classic ~riC$ ofcXJXrimc:ncs by Drew and coworkers (e.g. Drew 1975), onc section of abarley root axis was exposed to IOmmol L- l of N03-, NH~" orphosphate ions, while the rest experienced only 0.1 mmol L- 1

, Thesegment receiving the high concentration showed greatly enhancedgrowth of primary laterals and consequently much denser developmentof secondary laterals. Intriguingly, the roms failed to respond to potas­sium ions, perhaps because potassium ions are not metabolized in theroot; in the case of the other ions, increased uptake would lead to ademand for more melabolites for their utilization and this increasedsupply of fixed carbon may have stimulated growth.

This proliferntive response is so clear and apparently so adaptive thatit has often been assumed that it is universal, but this is not so. In a neatexperiment, Campbell et 4/. (1991) allowed nutrient patches to developin sand culture by dripping solutions onto the sumce ofa hemisphericalbowl. This created four quadrants with no barriers between but withvery well-maintained differences in nutrient concentration. They then

allowed plants of a range of uncultivated species to explore this hetero­geneous environment. Strikingly, whereas some species did, as expected,concentrate root growth in the nutrient-rich areas, others grew theirroots almost randomly through the bowls. The species that were mostprecise in their root placement, developing roots predominantly in thenutrient-rich patches, were the smallest and slowest growing. Those thatignored the patches and grew roots randomly were the largest, fastestgrowing and most competitive (Fig. 3.13). The authors pointed outthat this suggested that a contrast should be drawn between the scaleand precision of foraging.

[7l]Chapter 3

NutrientAcquisition

••

~. .•

•

•<.i'fiQo

" "~ E 0.9g.~

i ~ 0.80<

'0 :E 0.1g2'f'l::l8. ~ Q.6, ~

c.. ';ij 0.5,':-----::-----::c;----;:-:0.1 0.2 0.3 0.4

0.174 logR .. O. 100 logWsd

Fig. 3.13 The precision with which plants allocated new root growth to nunicnHichpatches in the experiments ofCampbdl et at. (1991) was greatest in small plants andleast in large ones. The x-uis in this graph is a composite variable derived from multipleregression of precision of allocation on relative growth rate (R) and seed weight (IV1<1);high values represent large and/or fast'growing plants. (From Fitter 1994.)

3.4.3 Turnover

The fact that plants do not always exploit nutrient-rich patches byproliferation highlights the complex economics of patch exploitation.For proliferation to be an advantageous response, the benefit from theincreased. root density must exceed the costs of the new roots. If a patchis short-lived, the gain may in fact be small, but the cost may be largeand the need for maintaining the roots will continue for as long as theroots survive. The coarser the roots, the less likely it will be that theconstruction costs will be recouped. Coarse-rooted species might there­fore be expected to be less likely to exploit patches, and this appears tobe what occurs.

Another important distinction between coarse and fine roots seemsto lie in their longevity. Fine roots are often remarkably short-lived: inrhizotron studies, where roots are observed against glass walls or thesides of tubes inserted into soil, the half-life ofa population of fine rootscan be less than 10 days (Fig. 3.14). In other words, half of a set of rootsinitiated at a particular time may be dead 10 days later. This rapid

• Great Dun Fell, .pring• Great Dun Fell, summero Newton Rigg. spring

.. Newton Rigg, summer

o

'- e •• _." •......-" _.....0 ••

' ..p •..

'.'.0

'•.....'':---'''''''''''''''''''O-:----=--''>'"--::---'-''"":cM.y 93 Jul93 Sep 93 Nov 93 Jln 9.

,

[72]Chapter 3NutrientAClfuiI;tiqn

fig. 3.1. Survival of cohons of mou under ill gruslands~ growing either U lhe

wmmir ofGf'C;<It Dun Fell in nonhcm England (air. 845 m) or at Newton Rig (all.170 m). Mean survival time is always lower for spring cohorts comp;al"td with summercohorts, and for low compared with high alrirudc cohoru. (G. K. Self & T. K. Brown,unpublished data.)

turnover of foots has enormous implications for studies of carboncycling, since it means that estimates of below-ground productivitybased on biomass measurements are almost always too low. For studiesof resource acquisition, it suggests the possibility of responding tospatial and temporal hcrc:rogeneity in nutrient supply by the mainte­nance of a labile root system, in which roots arc continuously develop­ing in an=as of high nutrient supply and then dying.

What appears to be unknown is whether these roots scnesce in thesense: that leaves do, with the nutrients being withdrawn and re-used, orwhether they JUSt die or are killed by soil organisms. If there is nonutrient recycling the adaptiveness of rapid turnover is less obvious.This is an area when= new research is urgently needed. Nevertheless,heterogeneity of nutrient supply and the nature of plant response to itare too often neglected in studies of mineral nutrition that have theirintellectual springs in the controlled environment chamber and thehydroponic culture. An ecological approach must be more aware of thenature of the soil environment.

3.5 Summary

The acquisition of nutrients from soil by plant roots cannot be under·stood by reference: to the physiology of ion uptake alone. The move·ment of ions from soil to foots represents two processes in series and thelimiting step is frequently that through soil. Plant responses to thislimitation include direct modifications of the soil around the roots, andcomplex changes in the architecture of root systems. In addition, mOStplants an= symbiotic, typically with mycorrhizal fungi; N-fixing bacteriaalso promote nutrieO[ acquisition in some species. The responses thatplants make to nutrient limitation arc complicated by the variability ofnutrient supply in time and space.