Few topics in ecology are as intriguing or bedeviling as comparative studies of different types of mycorrhizas formed in the same plant species. Attempts to determine the relative benefits from each fungal type to the host plant are fraught with difficulties (Jones et al., 1998), and for this reason plants that form tripartite associations with arbuscular and ectomycorrhizal fungi make ideal experimental systems. Just such a tripartite system is explored in a report in this issue by Chen et al. (pp. 545–556).

Why compare mycorrhizas? Although both ecto- and arbuscular mycorrhizas are generally known to increase the uptake of nutrients such as phosphorus and nitrogen in infertile soils, their functions and the benefits they provide to their host plants may not be equivalent (Jones et al., 1998). Hence the ecological benefits of the two types of symbiosis are of great interest to plant community ecologists and researchers studying nutrient cycling. Arbuscular mycorrhizas are formed as a symbiosis of plant roots with primitive fungi of the order Glomales (Zygomycetes), while ectomycorrhizas are formed with higher basidiomycetes and a few ascomycetes. Only a few plant genera and species are capable of forming both types of association, sometimes on the same root tip (Fig. 1), but it is these that are so useful in comparative studies.

Several generations of soil microbiologists and plant pathologists have attempted to unravel the microbial composition, population dynamics and various life-sustaining processes in the soil. However, until recently, this basic resource for plant growth and nutrient recycling has to a large extent remained a black box. Studies have been made of the behavior of individual microorganisms or their population dynamics, but the relationship between the growth of individual organisms or parts of organisms (such as fungal hyphae) and that of populations and communities has been largely unknown. In particular, the spread of fungal colonies in soil has been an enigma, but now Bailey et al. (pp. 000–000 in this issue) have clearly demonstrated that spatial behavior of Rhizoctonia colonies can switch from finite to invasive expansion dependent on the non-linear relationship between the distance among substrate particles and the probability of colonization.

Cadmium accumulation and tolerance are discussed in a New Phytologist article by Krämer (2000), which comments on a paper by Lombi et al. (2000). In this context, a number of additional points should be made, putting the role of humans in the evolution of metal resistance into context and emphasizing what is the ‘hype’ of phytoremediation.

It is important that sites created by humans should not be overemphasized in considering the evolution of metal resistance. Plants resistant to heavy metals have their primary sites not on these, but on soils where ores are outcropping, the so-called metalliferous or orogenic soils (Ernst, 1974). Over thousands of years, natural exposure to a surplus of various metals, depending on the mineralization process, has driven the evolution of metal resistance in many plant species under the local environmental conditions. Many publications have shown that Thlaspi caerulescens can hyperaccumulate Zn (e.g. Vázquez et al., 1992), and accumulate other heavy metals such as Cu and Pb depending on soil chemistry (e.g. Baker et al., 1994). It has been known for more than 30 years that T. caerulescens gives a good response to experimentally supplied high Zn levels (Ernst, 1968). One of the ecological effects of hyperaccumulation of heavy metals is a defence against herbivorous insects (Boyd & Martens, 1994). This effect is enhanced by a preferential accumulation of heavy metals in the epidermal leaf layer (Heath et al., 1997).

I. INTRODUCTION 360

II. PHOTOINHIBITION AND ACTIVE OXYGEN 360

III. OXYGEN AS AN ELECTRON ACCEPTOR 362

1. Oxygen ‘poises’ electron transport and carbon assimilation 362

2. The role of oxygen in ATP synthesis 364

3. How fast is O2reduction at PSI? 364

4. Chloroplastic processing of H2O2 366

IV. REDOX REGULATION OF PHOTOSYNTHETIC METABOLISM 368

1. The thioredoxin system 368

2. Manipulating the expression of thiol-regulated enzymes 369

3. Modifying sensitivity to thiol regulation 369

V. PHOTORESPIRATION 369

1. The pathway and its genetic manipulation 369

2. Engineering plants that photorespire less? 371

3. Is photorespiration important in energy dissipation? 372

4. Production and processing of photorespiratory H2O2 373

5. Catalase and foliar H2O2levels 374

6. Catalase and non-photorespiratory H2O2generation 375

VI. RESPIRATION 376

1. ‘Photosynthetic’ respiration 376

2. AOS in the mitochondrion 376

3. AOX: regulation and significance to photosynthesis 377

VII. PHOTOSYNTHESIS AND REDOX SIGNAL TRANSDUCTION 378

1. The need for sensors, signals and transducers 378

2. Signal transduction at the local level 378

3. Remote signalling and responses leading to acclimation of photosynthesis? 379

4. Interactions between AOS, NO., and antioxidants 380

VIII. CONCLUSIONS 380

Acknowledgements 381

References 381

The gradual but huge increase in atmospheric O2 concentration that followed the evolution of oxygenic photosynthesis is one consequence that marks this event as one of the most significant in the earth's history. The high redox potential of the O2/water couple makes it an extremely powerful electron sink that enables energy to be transduced in respiration. In addition to the tetravalent interconversion of O2 and water, there exist a plethora of reactions that involve the partial reduction of O2 or photodynamic energy transfer to produce active oxygen species (AOS). All these redox reactions have become integrated during evolution into the aerobic photosynthetic cell. This review considers photosynthesis as a whole-cell process, in which O2 and AOS are involved in reactions at both photosystems, enzyme regulation in the chloroplast stroma, photorespiration, and mitochondrial electron transport in the light. In addition, oxidants and antioxidants are discussed as metabolic indicators of redox status, acting as sensors and signal molecules leading to acclimatory responses. Our aim throughout is to assess the insights gained from the application of mutagenesis and transformation techniques to studies of the role of O2 and related redox components in the integrated regulation of photosynthesis.

Physcomitrella patens, recently renamed Aphanoregma patens, has been transformed with the plasmid, pBI426. On selective medium approx. 30% of regenerants expressed the transformed phenotype transiently (transients). The remaining 70% (transformants) retained their transformed phenotype (GUS-positive and resistant to G418) indefinitely when subcultured repeatedly on selective medium. However, most lost this phenotype after one or two passages through nonselective medium (unstable transformants). Approximately 0.2% of transformants retained their transformed phenotype after numerous passages through nonselective medium (stable transformants). Using PCR methodology, it has been shown that loss of the transformed phenotype by unstable transformants is invariably accompanied by disappearance of the transgenic DNA. Southern blot analysis data argue strongly that unstable transformants cultured under selective conditions contain unintegrated pBI426 as circular concatenates consisting of 3–40 copies of the plasmid. Under selective conditions, it appears that replication and/or partitioning of these extrachromosomal concatemers might be growth rate-limiting. This is the first report of a transgenic, autonomously replicating extrachromosomal element in a photosynthetic plant. A single copy of pBI426 has been inserted into the moss genome in each of three stable transformants analysed.

The three youngest age-classes of needles of Norway spruce (Picea abies) were collected from four sites in the Krusne Hory Mountains (Czech Republic) characterized by different levels of damage caused by environmental pollution. Histochemical methods did not reveal any differences in localization of phenolics among the needles. Mesophyll cells close to the epidermis of needles and cells around resin ducts and substomatal cavities often accumulated higher amounts of phenolics than the rest of the mesophyll cells, but this was independent of age and damage. Needles of different age- and damage-class did not show any marked changes in general lignification pattern. However, a lower intensity of histochemical detection of lignin was observed in needles from the most damaged site. This finding was confirmed by chemical analysis using thioglycolic acid. Generally, the amount of lignin in mesophyll cells was lower in damaged trees than in healthy ones. Using the Folin–Ciocalteau method, no significant differences in the total content of phenolics were observed in the needles, although HPLC revealed marked alterations in the forms of seven phenolic acids. Concentrations of conjugated forms of phenolic acids (esters and glycosides) were higher in damaged needles (255.9 μg g−1 f. wt) than in healthy needles (189.8 μg g−1 f. wt). By contrast, content of esterified phenolic acids incorporated into cell walls was higher in needles from healthy trees (101.1 μg g−1 f. wt) than in damaged needles (78.3 μg g−1 f. wt). Marked differences were also observed in the activity of soluble peroxidases, although the activity of ionically bound forms was approximately the same in healthy and damaged needles. The total amounts of chlorophylls and carotenoids decreased as environmental damage increased.

Growth responses of two provenances of European beech (Fagus sylvatica) were studied. The seedlings were grown in closed-top chambers at four temperature regimes (−2.9 °C below ambient, ambient, +2.3 °C and +4.8 °C above ambient) in combination with two CO2 partial pressures (40 Pa and 74 Pa). Growth was followed by making destructive harvests c. every 25 d from germination in early June to senescence in late September. Allocation patterns were significantly affected by the temperature regimes. However, changes in dry matter allocation and morphology associated with the different treatments at a given time were mostly a result of differences in tree size. Temperature regimes only had a significant effect on the relative growth rate, RGR, at the beginning of the experiment. In contrast to temperature, high [CO2] increased RGR throughout the experiment when compared with plants of equal size. As the trees increased in size net assimilation rate, NAR, decreased but the effect of [CO2] on both NAR and RGR had a tendency to increase. Increases in NAR caused by elevated [CO2] were partly counteracted by reductions in the leaf area ratio, LAR. Reductions in LAR were caused by concomitant reductions in specific leaf area, SLA, whereas the level of [CO2] did not significantly affect leaf weight ratio, LWR, nor other dry weight ratios. The interactions between temperature and [CO2] are highly dependent on whether they are expressed as instantaneous values for plants at a common age or instantaneous values at a common size (and thereby extracting the effects of ontogenetic drift). When comparing instantaneous values at common sizes, the positive effect of [CO2] on RGR increased with plant size in every temperature regime. This also occurred in every temperature regime when comparing plants of equal age but the response to [CO2] was less. The effect of [CO2] on RGR was dependent on growth temperature. The positive effects of elevated [CO2] on RGR were less than the positive effect on photosynthesis. The two provenances did not differ significantly in the response of RGR to [CO2] which is in agreement with measurements of photosynthesis.

The effects of elevated concentrations of atmospheric carbon dioxide and ozone on diurnal patterns of photosynthesis have been investigated in field-grown spring wheat (Triticum aestivum). Plants cultivated under realistic agronomic conditions, in open-top chambers, were exposed from emergence to harvest to reciprocal combinations of two carbon dioxide and two ozone treatments: [CO2] at ambient (380 μmol mol−1, seasonal mean) or elevated (692 μmol mol−1) levels, [O3] at ambient (27 nmol mol−1, 7 hr seasonal mean) or elevated (61 nmol mol−1) levels. After anthesis, diurnal measurements were made of flag-leaf gas-exchange and in vitro Rubisco activity and content. Elevated [CO2] resulted in an increase in photoassimilation rate and a loss of excess Rubisco activity. Elevated [O3] caused a loss of Rubisco and a decline in photoassimilation rate late in flag-leaf development. Elevated [CO2] ameliorated O3 damage. The mechanisms of amelioration included a protective stomatal restriction of O3 flux to the mesophyll, and a compensatory effect of increased substrate on photoassimilation and photosynthetic control. However, the degree of protection and compensation appeared to be affected by the natural seasonal and diurnal variations in light, temperature and water status.

Seedlings of Quercus petraea (oak), Fraxinus excelsior (ash) and Pinus sylvestris (Scots pine) were grown at two CO2 concentrations with two O3 and two water supply treatments for 3 yr in a factorial experiment. Oak was the most responsive species to all three treatments; elevated CO2 and irrigation increased biomass by an average of 79% and 41%, respectively, whereas the ozone treatment resulted in a 30% reduction in growth. Significant treatment interactions in this species demonstrated that CO2 ameliorated and irrigation exacerbated the effects of ozone. For Scots pine and ash, irrigation and elevated CO2 increased growth by approx. 60% and 20%, respectively, whereas ozone had no detectable effect on ash and resulted in a 15% reduction in growth of Scots pine. Carbon partitioning to the shoot was enhanced by both the CO2 and H2O treatments in oak, while branching was also increased in this species in response to elevated O3, resulting in changes to the allometric relationships. CO2 enhanced leaf production in oak and Scots pine, and together with the promotion of shoot allocation, this indicates an increased susceptibility to windthrow. Biomass accumulation expressed as relative growth rate, suggested three different time-dependent growth responses to elevated CO2; the CO2 fertilization effect was maintained through the third year of growth in oak, had disappeared in Scots pine and a negative effect was evident in ash. Foliar nitrogen and chlorophyll concentrations indicated a CO2-induced nitrogen deficiency in oak and ash, but not in Scots pine. Chlorophyll degradation in response to ozone was observed in oak, an effect that was enhanced by irrigation and reduced by CO2, presumably through stomatal mediated changes in effective ozone dose. These results therefore suggest that elevated atmospheric CO2 concentrations will enhance growth of some UK forest tree species, although this might only be apparent during the juvenile phase. However, nitrogen deficiencies might limit this enhancement on some sites while changes in allocation and leaf area might promote susceptibility to windthrow. Elevated CO2 also provides some protection against ozone pollution, especially in combination with limited soil moisture availability. These interactions between CO2, ozone and water supply should be taken into account when predicting the effects of environmental change on tree growth and forest productivity.

We assessed the ability of several populations of the metal-hyperaccumulator species, Thlaspi caerulescens, to mobilize non-labile cadmium in soils historically contaminated by Pb/Zn mine spoil or sewage sludge. Radio- labile Cd was determined chemically as an ‘E-value’, [CdE], and biologically as an ‘L-value’, [CdL]. For comparison, chloride-extractable Cd, [Cdchlor], was also determined using 1 M CaCl2 as a single-step soil extractant. Values of [CdL] were measured for six populations of T. caerulescens that varied substantially in their ability to assimilate soil Cd, and a non-accumulator species with a similar growth habit, Lepidium heterophyllum. Seeds were sown in soil spiked with 109Cd and grown for 9–12 wk in a controlled environment room. Values of [CdL] were determined from the specific activity of 109Cd and concentration of Cd in the plant leaves. For the six soils studied, [CdE] ranged from 4.9 to 49% of total soil Cd [CdT]. Values of [CdL] were, in general, in close agreement with both [CdE] and [Cdchlor] and substantially less than [CdT]. However, [CdL] showed no correlation with the concentration of Cd in plant tissue, [Cdshoot]. This suggests that, in the soils studied, T. caerulescens did not mobilize non-labile soil Cd by producing root exudates or altering rhizosphere pH. The results imply that there may be significant restrictions to metal bioavailability, even to hyperaccumulator species, in heavily contaminated soils in which a large proportion of the metal may be present in ‘non-labile’ forms.

Two-yr-old Scots pine (Pinus sylvestris) seedlings were grown in sand culture for 1 yr with a generous supply of a balanced nutrient solution. Trees were repotted into clean sand in February 1998 and given either a reduced or adequate nutrient supply containing enriched 15N, 41K and 26Mg to label nutrient uptake during spring 1998. Trees doubled their biomass during the experiment. Whole-tree net photosynthesis was reduced by 43% after 95 d in trees that received the lower nutrient supply (P < 0.001), although differences in biomass between the two treatments were less pronounced. Remobilization contributed 83, 82 and 52% of the N, K and Mg, respectively, used to support growth of new tissues in trees that received reduced nutrient supply. Those receiving the higher nutrient supply still obtained 44–59% of nutrients used for spring growth of new tissues from remobilization. Current nutrient supply had no significant effect on the amount of N or Mg remobilized to new tissues but K remobilization was less in trees that received the lower nutrient supply (P = 0.025). The importance of remobilization in young trees and problems associated with quantifying internal cycling of nutrients are discussed.

Hordeum spontaneum shows a large genetic variation and occupies a wide range of different habitats. The aim of this study was to quantify variation in growth characteristics of H. spontaneum from different sites in Israel and to relate this variation to different environmental conditions. To this end, 84 accessions of 21 populations were grown in a growth chamber in near-optimal conditions and a range of physiological, morphological, allocation- related and chemical characteristics were measured. These parameters included rates of photosynthesis, shoot and root respiration, specific leaf area, biomass allocation and seed mass. Averaged over all traits variation explained by differences between populations was 26%, between accessions 21%, whereas that within accessions was 53%. By contrast with most genetic studies, we found variation between populations larger than between accessions. The largest between-population variation (46%) was for morphological traits. In particular, seed mass, leaf thickness and leaf width differed strongly between populations. Variation in growth characteristics between populations was poorly related to mean annual rainfall, mean humidity or January temperature at the sites of origin. We expect that differences between populations to be larger and correlation with environmental parameters stronger in plants grown in stressful conditions. According to our study, seed mass is more important than relative growth rate in determining variation in early plant biomass in H. spontaneum.

Following a 3-wk pretreatment with 150 mM NaCl, Sorghum plants were able to survive exposure to 300 mM NaCl, a lethal concentration for nonpretreated plants. This response is termed salt adaptation. Although the population was initially homogeneous, Na-includer and Na-excluder individuals coexisted after the achievement of the salt-adaptation response, and a large diversity in size of the seeds produced was observed at the end of the life cycle. Offspring of salt-adapted plants were exposed to a new salt-adaptation treatment. Even in the absence of selection, the proportion of Na-excluder individuals in the progeny was significantly increased. Offspring germinated from small seeds differed significantly from plants first-exposed to a salt-adaptation treatment, whereas offspring from large seeds displayed intermediate characteristics. This suggests that some of the adaptive changes were transmitted through embryo imprinting. The importance of this phenomenon for emergence of a newly adapted ecotype is discussed.

In normal populations of the common grass Holcus lanatus there is a polymorphism for arsenate resistance, manifested as suppressed phosphate uptake (SPU), and controlled by a major gene with dominant expression. A natural population of SPU plants had greater arbuscular-mycorrhizal colonization than wild type, nonSPU plants. It was hypothesized that, in order to survive alongside plants with a normal rate of phosphate (P) uptake, SPU plants would be more dependent on mycorrhizal associations. We performed an experiment using plants with SPU phenotypes from both arsenate mine spoils and uncontaminated soils, as well as plants with a nonSPU phenotype. They were grown with and without a mycorrhizal inoculum and added N, which altered plant P requirements. We showed that grasses with SPU phenotypes accumulated more shoot P than nonSPU plants, the opposite of the expected result. SPU plants also produced considerably more flower panicles, and had greater shoot and root biomass. The persistence of SPU phenotypes in normal populations is not necessarily related to mycorrhizal colonization as there were no differences in percentage AM colonization between the phenotypes. Being mycorrhizal reduced flower biomass production, as mycorrhizal SPU plants had lower shoot P concentrations and produced fewer flower panicles than non-mycorrhizal, nonSPU plants. We now hypothesize that the SPU phenotype is brought about by a genotype that results in increased accumulation of P in shoots, and that suppression of the rate of uptake is a consequence of this high shoot P concentration, operating by means of a homeostatic feedback mechanism. We also postulate that increased flower production is linked to a high shoot P concentration. SPU plants thus allocate more resources into seed production, leading to a higher frequency of SPU genes. Increased reproductive allocation reduces vegetative allocation and may affect competitive ability and hence survival, explaining the maintenance of the polymorphism. As mycorrhizal SPU plants behave more like nonSPU plants, AM colonization itself could play a major part in the maintenance of the SPU polymorphism.

This study examines the microtubular cytoskeleton during megasporogenesis in the Nun orchid, Phaius tankervilliae. The subepidermal cell located at the terminal end of the nucellar filament differentiates first into an archesporial cell and then enlarges to become the megasporocyte. The megasporocyte undergoes the first meiotic division, giving rise to two dyad cells of unequal size. Immunostaining reveals that microtubules become more abundant as the megasporocyte increases in size. Microtubules congregate around the nucleus forming a distinct perinuclear array and many microtubules radiate directly from the nuclear envelope. In the megasporocyte, prominent microtubules are readily detected at the chalazal end of the cell cytoplasm. After meiosis I, the chalazal dyad cell expands in size at the expense of the micropylar dyad cell. At this stage, new microtubule organizing centres can be found at the corners of the cells. The appearance of these structures is stage-specific and they are not found at any other stages of megasporogenesis. The functional dyad cell undergoes the second meiotic division, resulting in the formation of two megaspores of unequal size. The chalazal megaspore enlarges and eventually gives rise to the embryo sac. As the functional megaspore expands, the microtubules again form a distinct perinuclear array with many microtubules radiating from the nuclear envelope. A defined cortical array of microtubules has not been found in P. tankervilliae during the course of megasporogenesis.

Results presented in this paper demonstrate the relationship between two data sets from Mubwindi Swamp, south-west Uganda. The first data set is from a survey of dryland montane forest within the immediate catchment and the second is in the form of pollen spectra recovered from surface samples of sediment. Numerical analysis grouped pollen types encountered into five categories, according to their level of representivity when compared with abundances of the same taxa in the sampled vegetation: under representative, moderately under representative, representative, moderately over representative, and over representative. The northern part of Mubwindi Swamp also yielded a 5-m-long sequence of organic-rich sediments (core MB6) from which sub-fossil pollen and radiocarbon data have been previously obtained. This provides an opportunity to apply calibration factors derived from modern pollen–vegetation relationships to sub-fossil pollen data that were obtained from the same site. These calibration factors effectively down-weighted the influence of over representative pollen types and permitted the discrimination of a phase of forest disturbance that had not previously been so clearly visible in sub- fossil pollen data.

The development of cereal cyst nematode (CCN; Heterodera avenae) induced syncytia in the host roots of infected resistant bread wheat (Triticum aestivum cv. AUS10894), diploid wheat (Aegilops tauschii), barley (Hordeum vulgare cv. Chebec and cv. Galleon) and in the susceptible wheat cv. Meering and barley cv. Clipper were studied over a period of 13 d. The resistance to CCN in these cereal plants is conferred by the resistance genes Cre1 in the wheat cv. AUS10894, Cre3 in A. tauschii, Ha2 in barley cv. Chebec and Ha4 in barley cv. Galleon. Anatomical observations were made on the development of the syncytia in CCN-infected wheat and barley roots, which carry each of these four sources of resistance genes. Accelerated development of the syncytia in resistant plants, especially in the barley cultivars, was observed. The sites of syncytia development in susceptible wheat and barley were also closely associated with the vascular tissues in the stele, but less so in the resistant plants. The syncytia in the infected susceptible wheat and barley were also metabolically active at day 13. By contrast, the syncytia of resistant wheat plants carrying the Cre1 or Cre3 genes remained extensively vacuolated and less metabolically active. In barley plants with the Ha2 or Ha4 genes, the syncytia appeared non-functional and in early stages of degeneration by day 13 after inoculation.

In this paper we distinguish between invasive and noninvasive (finite) saprotrophic spread of the soil-borne fungal plant pathogen, Rhizoctonia solani amongst discrete sites of nutrient resource. Using simple concepts of percolation theory, we predict the critical threshold distance, associated with a threshold probability, between donor (colonized) and recipient (uncolonized) nutrient sites at which R. solani can spread invasively by mycelial growth through a population of nutrient sites on a lattice. The critical distance for invasive spread is estimated from colonization profiles derived from placement experiments that summarize the probability of colonization with distance between replicated pairs of colonized and uncolonized sites. Colonization profiles were highly nonlinear, decaying sigmoidally with distance. Thresholds for invasive spread were predicted at inter-site distances of 8.1 mm and 11.8 mm for sites of low and high nutrient agar, respectively. In population experiments with inter-site distances below the predicted thresholds, the spread of the fungus was invasive in all replicates. At large distances (>10 mm for low, and >14 mm for high nutrient sites) the spread of the fungus was always finite, with the proportion of finite replicates decreasing sharply close to the percolation threshold. Invasive spread did not depend on the furthest extent of growth of the fungus but on distances predicted by the percolation thresholds. Invasive spread of the fungus is also examined in a more natural and variable, nonsterile system involving the growth and colonization of a lattice of poppy seeds over sand. The system is characterized by a decay in the probability of colonization between older poppy seeds, which effectively ‘quenches’ saprotrophic spread. Hence in the population experiments with poppy seeds all growth was ultimately finite. The threshold distance, corresponding to the critical percolation probability for invasive growth changed from 18 mm to 4 mm over 21d leading to a switch from invasive to finite growth. We conclude that percolation theory can be used to link the growth of individual mycelial colonies to the formation of patches that result from the colonization of particulate organic matter. The nonlinearity of the colonization profiles combined with the presence of a percolation threshold means that small changes in the distance between nutrient sites can result in large differences in final patch size. The rapid decay of particulate organic matter in a more natural system can have a profound effect on the dynamics of colonization, restricting saprotrophic invasion of the soil. The consequences of invasion thresholds for colony growth of saprotrophic and parasitic fungi in dynamical systems are briefly discussed.

Eucalyptus species are considered to have ectomycorrhizas (ECM), but many also have vesicular–arbuscular mycorrhizas (VAM) and their relative importance is unclear. Interactions between ECM and VAM fungi colonizing roots of Eucalyptus species were examined in a glasshouse experiment. This experiment investigated competition between these two types of fungi and compared benefits provided to Eucalyptus globulus and E. urophylla. Eucalyptus seedlings were inoculated with spores of the ECM fungus Laccaria lateritia and/or pot- culture soil for a VAM fungus (species of Glomus, Acaulospora or Scutellospora). Initial inoculum levels were important, as VAM fungi became established much more rapidly than Laccaria. In plants with both types of mycorrhizas, Laccaria mycorrhizas substantially increased after 2 or 3 months and the proportion of roots with VAM declined. However, the proportion of roots with VAM also decreased significantly in plants without ECM after 2 months. Thus, the relative susceptibility of eucalypt roots to these mycorrhizal associations changed. Substantial growth responses to mycorrhizal inoculation occurred when a low concentration of phosphorus fertilization was used (5 mg kg−1), but not at a higher concentration (10 mg kg−1). Treatments where Laccaria was applied, alone or in combination with a VAM fungus, resulted in the largest growth increases relative to nonmycorrhizal plants. An Acaulospora isolate was the most effective VAM fungus for E. urophylla although other VAM fungi also increased growth relative to the control. VAM fungi alone had little effect on E. globulus growth, but plants inoculated with both ECM and VAM fungi were larger than plants only inoculated with Laccaria. ECM fungi had a major impact on root system form by reducing the proportion of fine roots (specific root length) relative to nonmycorrhizal plants or those with VAM. Changes in root colonization patterns over time must be considered in studies of ECM/VAM interactions, as there can be substantial changes in the relative importance of the two associations.