Contrasting responses to mycorrhizal inoculation and phosphorus availability in seedlings of two tropical rainforest tree species


  • Agnès De Grandcourt,

    1. UMR INRA ENGREF CIRAD ‘Ecologie des Forêts de Guyane’, Campus Agronomique, 97387 Kourou cedex, Guyane Française;
    Search for more papers by this author
  • Daniel Epron,

    Corresponding author
    1. Laboratoire de Biologie et Ecophysiologie, Université de Franche-Comté, F-25211 Montbéliard cedex, France;
    2. UMR INRA UHP ‘Ecologie et Ecophysiologie Forestières’, Université Henri Poincaré, F-54506 Vandoeuvre cedex, France;
    Search for more papers by this author
  • Pierre Montpied,

    1. UMR INRA UHP ‘Ecologie et Ecophysiologie Forestières’, Centre INRA de Nancy, F-54280 Champenoux, France;
    Search for more papers by this author
  • Eliane Louisanna,

    1. UMR INRA ENGREF CIRAD ‘Ecologie des Forêts de Guyane’, Campus Agronomique, 97387 Kourou cedex, Guyane Française;
    Search for more papers by this author
  • Moïse Béreau,

    1. UMR INRA ENGREF CIRAD ‘Ecologie des Forêts de Guyane’, Campus Agronomique, 97387 Kourou cedex, Guyane Française;
    Search for more papers by this author
  • Jean Garbaye,

    1. UMR INRA UHP ‘Interactions Arbres–Microorganismes’, Centre INRA de Nancy, F-54280 Champenoux, France
    Search for more papers by this author
  • Jean-Marc Guehl

    1. UMR INRA UHP ‘Ecologie et Ecophysiologie Forestières’, Centre INRA de Nancy, F-54280 Champenoux, France;
    Search for more papers by this author

Author for correspondence: Daniel Epron Tel: +33 383684249Fax: +33 383684240Email:


  • • This work aimed at understanding the role of mycorrhizal status in phosphorus efficiency of tree seedlings in the tropical rainforest of French Guyana.
  • • Mycorrhizal colonization, growth, phosphorus content, net photosynthesis and root respiration were determined on three occasions during a 9-month growth period for seedlings of two co-occurring species (Dicorynia guianensis and Eperua falcata) grown at three soil phosphorus concentrations, with or without inoculation with arbuscular mycorrhizas.
  • • Seedlings of both species were unable to absorb phosphorus in the absence of mycorrhizal association. Mycorrhizal seedlings exhibited coils that are specific of Paris-type mycorrhizae. Both species benefited from the mycorrhizal symbiosis in terms of phosphorus acquisition but the growth of E. falcata seedlings was unresponsive to this mycorrhizal improvement of phosphorus status, probably because of the combination of high seed mass and P reserves, with low growth rate.
  • • The two species belong to two different functional groups regarding phosphorus acquisition, D. guianensis being an obligate mycotrophic species.


Regeneration stages of tropical rainforest trees are critical steps for forest dynamics (Whitmore, 1996). Initial seedling growth is little affected by resource availability as long as it depends on seed reserves, but becomes dependent on external supply of resources as soon as these reserves are depleted (Kitajima, 2002). At the regeneration stage, tropical tree seedlings are thought to compete for light, water and nutrients, and the spatial heterogeneity of these resources may contribute to species richness because of differentiated responses among species. Among these factors, soil phosphorus availability in tropical rainforests has often been recognized as a major limiting factor for primary production and tree growth, especially on old ferrallitic soils (Vitousek & Denslow, 1986; Korning et al., 1994).

Several mechanisms could act to improve phosphorus efficiency in phosphorus-limited plants, such as a reduction of phosphorus requirement for growth (i.e. an increase in phosphorus use efficiency), or changes in root architecture, root growth, root–shoot ratio, carbohydrate partitioning or root respiration cost. In addition, mycorrhizal symbiosis can account for an increase in phosphorus uptake at low soil phosphorus concentration by increasing the volume of soil prospected, thus allowing access to additional nutrient sources. Improved phosphorus nutrition would, in turn, enhance growth by increasing either the rate of photosynthesis per unit leaf area, or leaf area itself (Rodriguez et al., 1998). However, mycorrhizal symbiosis is associated with a carbon cost for the plant because of the construction and energetic requirements of the fungal associate (McArthur & Knowles, 1993; Pfeffer et al., 1999). Indeed, growth depression of mycorrhizal plants has often been reported and attributed to the carbohydrate requirement of the symbiont (Peng et al., 1993). The effects of the mycorrhizal status on seedlings will be positive if benefits for the seedlings, in terms of phosphorus uptake, and ultimately carbon acquisition and growth, are higher than the carbon costs as a result of the fungal association.

The phosphorus acquisition efficiency of seedlings can be defined as the amount of phosphorus taken up per unit of carbon expended. This has rarely been studied in the first regeneration stages of tropical rainforest trees. In the French Guyana tropical rainforest, arbuscular mycorrhizas are the predominant symbiotic association (Béreau et al., 1997), and they are well known to improve phosphorus nutrition and growth in soils with low phosphorus availability (Marschner & Dell, 1994; Smith & Read, 1997). However, their importance for seedling establishment, survival and release in tropical forests is understudied. Differences in mycorrhizal colonization or mycorrhizal efficiency between tree species might account for interspecific differences in the competitive ability of seedlings to grow and survive in specific niches (Janos, 1983). Both seed size and successional status of the species are thought to have important implications for mycorrhizal colonization of tropical seedlings, but contradictory results have been reported (Janos, 1980; Metcalfe et al., 1998; Siqueira et al., 1998; Kiers et al., 2000; Zangaro et al., 2000). The two climax species chosen for this study were two co-occurring, nonN2-fixing, shade hemitolerant, Ceasalpiniaceae: Dicorynia guianensis and Eperua falcata (Guehl et al., 1998). Both belong to the same successional group and exhibit Paris-type mycorrhizal association, but have different seed size. It has previously been shown that initial growth of D. guianensis seedlings clearly depends on endomycorrhizal symbiosis (Béreau et al., 2000). By contrast, seedlings of the large-seeded species E. falcata seemed less responsive to their mycorrhizal status (unpubl. preliminary data; see also Baraloto, 2001). Confirming these results would corroborate the hypothesis that even within a successional group, initial growth of large-seeded species is less-dependent on external resource acquisition by mycorrhizas.

In the present study, we assessed the hypothesis that phosphorus acquisition and seedling growth depend on the mycorrhizal status in phosphorus-limited soil and that two species differing in seed size and rooting pattern exhibit a different behaviour. We also examined whether this dependence on the mycorrhizal status for phosphorus acquisition and growth is related to soil phosphorus availability. Mycorrhizal colonization, growth, phosphorus content, net photosynthesis and root respiration were determined three times during a 9-month course of growth for seedlings of both species grown in an open-sided glasshouse at three soil phosphorus concentrations, with or without mycorrhizal inoculation.

Materials and Methods


Seeds were collected in the Paracou experimental forest of the CIRAD-Forêt near Sinnamary in French Guyana (5°15′ N, 52°55′ W; Schmitt & Bariteau, 1990).

Seeds of D. guianensis (average fresh mass 1.36 g) were collected in April 1998 during the last important fruiting period for this species in this site. The seeds are known to retain their germination capacity over several years. The dormancy of stored seeds was broken by soaking them in pure sulphuric acid for 10 min and rinsing five times with sterile distilled water (Béreau et al., 2000). Seeds of E. falcata (average fresh mass 9.22 g) were collected in April 2000 and used immediately because they germinate as soon as they fall to the ground. Seeds of both species were surface-sterilized with 0.1% mercuric chloride for 5 min and washed four times with sterile water. Seeds were then germinated on sterile paper moistened with distilled water for D. guianensis and on sterile soil for E. falcata.

Substrate preparation, phosphorus supply and mycorrhizal inoculation

The topsoil (15 cm) of a ferralsol (WRB, world reference basis) was collected near the seed collection area and sieved through a 1-cm mesh. Sieved soil was then mixed with 1 : 3 (v : v) white sand and steam disinfected at 90°C three times for 2 h each time with 2-d intervals. Sterilized soil was transferred in 5-l cylindrical containers (23 cm diameter, 16.5 cm high) containing 6 kg dry weight of soil each. A check was made to ensure that sterilization did not alter the chemical composition of the substrate, which is given in Table 1 (Soil Analysis Laboratory, INRA-Arras, France). This soil is representative of the region and presents all the characteristics of a ferralsol: acidic pH, low exchangeable cations (CEC) and a high Al3+ concentration.

Table 1.  Chemical properties of the topsoil used as substrate, before and after sterilization. Values are on an oven-dried soil mass basis
 Before sterilizationAfter sterilization
pH (H2O) 5.0 4.9
pH (KCl 1 n) 4.1 4
CEC (mEq kg−1)3026
C (g kg−1) 9.41 9.38
N, total (g kg−1) 0.7 0.69
Total Al (g kg−1) 1.41 1.17
Total Fe (g kg−1) 2.36 2.15
Total P (g kg−1) 0.033 0.037
Extractable P (g kg−1)a 0.006 0.006

Extractable soil phosphorus was determined using Duchaufour's method, which is more appropriate for acidic forest soils (Duchaufour & Bonneau, 1959). Briefly, a dual extraction was applied, using H2SO4 (0.002 m) and NaOH (0.1 m). Both extracts were further pooled and phosphorus was spectrophotometrically determined at 825 nm from the formation of a phosphomolybdate complex. The alternative Olsen's method (Olsen et al., 1954) yielded some values below the detection threshold. Total phosphorus concentration was only five times higher than extractable phosphorus, highlighting the fact that the phosphorus stock in this soil is limited. Three phosphorus treatments were established: P0, low phosphorus treatment, corresponding to the initial phosphorus concentration in the substrate (no addition); P1, intermediate treatment with a supply of 0.008 g P kg−1 soil as P2O5; P2, high phosphorus treatment with a supply of 0.04 g P kg−1 soil as P2O5. Because the phosphorus fertilizer also contains CaO, this was added to provide the same quantity of Ca for each phosphorus treatment. Fertilizers were added to each container individually and mixed into the substrate. Extractable soil phosphorus in each pot was determined at the time of harvest using the Duchaufour's method.

Within each phosphorus treatment, half of the seedlings (M+ treatments) were inoculated with a mixture of crushed fresh roots of adult D. guianensis and E. falcata trees collected in the Paracou forest. Mycorrhizal colonization of these roots was checked and greater than 75% for both species. The same quantity of steam-sterilized crushed roots was added to the uninoculated treatments (M–). Each pot received filtered soil solution (1 : 1, v : v; Whatman paper, 7 m pore diameter) retaining fungal spores and nematodes but not bacteria. Inoculation was made superficially by digging a 3-cm deep hole at the centre of each pot.

Experimental design

Each pot received one germinated seed, and soil was watered twice a day using an automatic drip-irrigation system. The experiment was conducted for 9 months in an open-sided glasshouse from April 2000 to February 2001. Irradiance in the greenhouse was reduced to about 30% of the outside value by a high-density polyethylene net (Bouillon SA, Caudry, France). Three harvests were done on days 117 (August, 2000), 202 (December, 2000) and 281 (February, 2001). On day 174, each one of the remaining seedlings received 250 ml of a nutrient solution containing 720 mg l−1 NH4NO3 and 300 mg l−1 KNO3. Pots were placed on wooden grids (1 × 1 m) previously washed with bleach to avoid contamination of nonmycorrhizal seedlings. Twelve grids with mycorrhizal seedlings and 12 with nonmycorrhizal seedlings were randomly arranged in the glasshouse. Fifteen seedlings were randomly disposed on each grid. Finally, the factorial combination of the two experimental factors resulted in the following six different experimental treatments within each of the two species: P0M–, P1M–, P2M–, P0M+, P1M+ and P2M+.

Gas exchange measurements

Root respiration was measured at the three harvest times. Entire root systems were extracted and gently washed. Shoots were cut at collar level and the cut end was covered by a piece of Parafilm to minimize the local respiration flux induced by wounding. Whole root systems were individually placed in an appropriated respiration chamber (0.25, 1 or 4 l depending on the size of the root system). Respiration was assessed using a closed system with an infrared gas analyser Li-Cor 6200 (Li-Cor Inc., Lincoln, NE, USA), by performing four successive measurements, each corresponding to a [CO2] increase of 10 mol mol−1 between 340 mol mol−1 and 380 mol mol−1[CO2]. The temperature of the laboratory was 25°C. Root temperature was measured with a contact thermocouple. The respiration of the extraradical mycelium was not included in our measurements because it had been washed away together with the soil when cleaning the roots. Root respiration was expressed on a per dry mass basis.

Photosynthesis at saturating irradiance (700 mol m−2 s−1 photon flux density) was measured in the greenhouse on two leaflets of each plant the week before each harvest with a gas exchange system using an infrared CO2 analyser (CIRAS 1, PP System, UK).

Plant parameters

At each harvest, seedlings were separated in their different components: fine roots (diameter < 2 mm), coarse roots, stem, rachis and leaflets. Before drying, total leaf area of each seedling was measured with a Li-3000 area meter (Li-Cor), and stem length and leaflet number were noted. The different components were then dried at 60°C and weighed. The carbon concentrations of roots, stem and leaflets were determined with an elemental analyser (ThermoQuest, Milano, Italy). Phosphorus concentrations were determined by atomic emission spectrometry (ICP; Jobin Yvon Emission, Lonjumesu, France). At the first and at the second harvest, there was not enough plant material for individual analysis (1.5 mg for elemental analysis and 120 mg for AES). Thus, seedlings were pooled in order to provide three replicates per treatment.

Mycorrhizal colonization of roots

Mycorrhizal colonization was quantified for each root system on a sample of fine roots accounting for c. 1% of total root system, according to Trouvelot et al. (1986). After staining fungal structures with fuchsin, root segments were observed at ×125 magnification in 100 microscope fields. For each field, the dominant fungal structure (mycelium, intracellular coils, vesicles or arbuscules) was noted (Béreau et al., 1997). The results were expressed as the percentage of fields with each type of fungal structure, and as the per cent of fields with any fungal structure.

Phosphorus acquisition efficiency

Focusing on the two last harvests, an attempt was made to estimate the phosphorus acquisition efficiency and its components. Phosphorus acquisition efficiency is the amount of phosphorus acquired per unit carbon spent (i.e. uptake over cost). Net uptake rate (mg P g−1 d−1) is the amount of phosphorus gained per unit of time and per unit of root dry mass, and was calculated as

[(P3 − P2)(ln W3 − ln W2)]/[(W3 − W2)(t3 − t2)]

(P is the seedling phosphorus content; W is the root dry mass; t is the day of harvest, with the subscript number referring to the harvest).

Carbon cost (g C g−1 d−1) includes carbon that is deposited during root growth and carbon lost during root respiration, and it was calculated as

C(ln W3 − ln W2)/(t3 − t2) + R

(C is the carbon content of the root tissue; R is the average rate of respiration at the second and third harvests). Root carbon contents were not significantly different between species, mycorrhizal treatment and phosphorus treatment (not shown) and an average value was used (0.45 g C g−1). Other carbon costs such as respiration and turnover of the extraradical mycelium, or root exudation, were not included in the calculation. Uptake rates, costs and efficiencies were calculated using mean values per treatment.

Statistical analysis

Data were analysed using mixed linear models with mycorrhizal treatments and phosphorus treatments as fixed effects and grids as random effect to take into account possible covariance inside grids. Each species was treated separately. Contrasts were used to test the effect of mycorrhizal inoculation and then phosphorus effect inside each mycorrhizal treatments, after checking that the overall model was significant. Phosphorus effect across mycorrhizal treatments was not considered because of significant interactions. Analyses were performed with SAS MIXED procedure of sas/stat 8.1 package (SAS Institute Inc., Cary, NC, USA).


Soil phosphorus availability and mycorrhizal colonization of roots

Extractable phosphorus concentrations averaged over mycorrhizal treatments and dates were 0.006, 0.010 and 0.021 g P kg−1 soil for P0, P1 and P2, respectively (i.e. 1.5 and 3.3 times higher for P1 and P2 than for P0). The three treatments fell within the range of observed concentrations in 60 soil cores collected in the root zone of saplings of these two species in Paracou (unpubl. data). Mycorrhizal inoculation did not significantly affect soil phosphorus concentrations, which were similar for the three harvests (data not shown).

No mycorrhizal colonization was observed in uninoculated seedlings of both species. The mycorrhizal colonization of inoculated seedlings increased over the experimental period and was higher for D. guianensis (between 60% and 95%) than for E. falcata (between 45% and 75%) at the end of the experimental period. No significant difference in mycorrhizal colonization of roots was observed among phosphorus treatments. The most abundant mycorrhizal structures found in root cells at the third harvest were coils in D. guianensis (93%), and coils and mycelium in E. falcata (59% and 28%, respectively, Table 2).

Table 2.  Relative abundances of mycorrhizal structures in root cells of Dicorynia guianensis and Eperua falcata at the third harvest for the three phosphorus treatments
 Phosphorus treatment
  1. P0, Low phosphorus treatment, corresponding to the initial phosphorus concentration in the substrate (no addition); P1, intermediate treatment with a supply of 0.008 g P kg−1 soil as P2O5; P2, high phosphorus treatment with a supply of 0.04 g P kg−1 soil as P2O5. Differences between phosphorus treatments were not significant (P = 0.05).

D. guianensis
Mycelium 2% 2% 1%
Vesicles 5% 6% 4%
Arbuscules 0% 0% 0%
E. falcata
Arbuscules 0% 0% 0%

Biomass and leaf area

In D. guianensis, dry mass and leaf area of nonmycorrhizal seedlings only slightly increased between the first and the second harvest, and even decreased afterward in the P1 and P2 treatments as they lost leaflets (Fig. 1a,c). By contrast, the dry mass of mycorrhizal seedlings increased with time, especially in the P1 and P2 treatments. At the third harvest, there was a positive effect of mycorrhizal colonization on seedlings biomass (P < 0.001) and a positive effect of phosphorus addition for mycorrhizal seedlings (P < 0.001). There was also a pronounced effect of mycorrhizal colonization on leaf area expansion (Fig. 1b, P < 0.001) that was positively enhanced by phosphorus addition (P < 0.001). At the end of the experiment, the total leaf area was more than four times higher in mycorrhizal seedlings than in nonmycorrhizal seedlings in the P2 treatment. This increase in leaf area was mostly caused by an increase in mean leaflet area (not shown).

Figure 1.

Total dry mass (a,b) and leaf area (c,d) of mycorrhizal (filled symbols) and nonmycorrhizal (open symbols) seedlings at three harvest dates for P0 (diamonds), P1 (squares) and P2 (circles) phosphorus treatments in Dicorynia guianensis (a,c) and Eperua falcata (b,d). Each point is an average value obtained on 5–11 seedlings. M indicates to significant effects of mycorrhizal inoculation, and P/M+ indicates significant effects of phosphorus addition for mycorrhizal seedlings (P < 0.05, contrast analysis).

In E. falcata, seedling dry mass increased almost similarly in all treatments except for the phosphorus-supplemented nonmycorrhizal seedlings between the second and the third harvest (Fig. 1b). The effect of mycorrhizal colonization on seedling biomass was not significant. There was no effect of phosphorus addition on the biomass of mycorrhizal seedlings while a surprising negative effect of phosphorus addition was observed for nonmycorrhizal seedlings (P < 0.01). At the third harvest there was a slightly positive but nonsignificant effect of mycorrhizal colonization on leaf area in this species (Fig. 1d, P = 0.062).

There was a positive effect of mycorrhizal inoculation on the leaf area ratio in D. guianensis (P < 0.05) but not in E. falcata. This positive effect was fully explained by an increase in the leaf mass ratio (P < 0.01), while the specific leaf area remained unchanged (not shown). Mycorrhizal status did not affect root–shoot ratio in either species (not shown).

Phosphorus content

Total phosphorus content did not increase with time in the nonmycorrhizal seedlings of both species, while it did in the mycorrhizal seedlings, with a positive effect of phosphorus addition (Fig. 2a,b). At the third harvest, there was a positive effect of mycorrhizal colonization on phosphorus content (P < 0.001 for both species) and a positive effect of phosphorus addition for mycorrhizal seedlings (P < 0.001 for both species). At the last harvest, phosphorus content was 15 times higher in mycorrhizal seedlings than in nonmycorrhizal ones for D. guianensis and three times higher for E. falcata in the P2 treatment.

Figure 2.

Total phosphorus content (a,b) and whole-plant phosphorus concentration (c,d) of mycorrhizal (filled symbols) and nonmycorrhizal (open symbols) seedlings at three harvest dates for P0 (diamonds), P1 (squares) and P2 (circles) phosphorus treatments in Dicorynia guianensis (a,c) and Eperua falcata (b,d). Each point is an average value obtained on three replicates for the first and the second harvest, and an average value on 5–11 seedlings for the last harvest. M indicates to significant effects of mycorrhizal inoculation, and P/M+ indicates significant effects of phosphorus addition for mycorrhizal seedlings (P < 0.05, contrast analysis).

In the nonmycorrhizal seedlings of both species, whole-plant phosphorus concentration decreased between day 117 and day 202 and remained constant afterwards, while it increased in mycorrhizal seedlings of D. guianensis (Fig. 2c,d). Mycorrhizal seedlings of the two species had higher phosphorus concentrations than nonmycorrhizal seedlings at the third harvest (P < 0.001 for both species), with a positive interaction with phosphorus availability for D. guianensis (P < 0.001) and for E. falcata (P < 0.05). This increase in phosphorus concentration in mycorrhizal seedlings was similar in all plant parts for both species (not shown).

Root respiration and leaf photosynthesis

Specific root respiration rate decreased in both species during the experimental period and it was higher in D. guianensis than in E. falcata. Root respiration was significantly higher for mycorrhizal seedlings than for nonmycorrhizal in both species at the third harvest (P < 0.01 and P < 0.001 for D. guianensis and E. falcata), and the increase was more pronounced in E. falcata (Fig. 3).

Figure 3.

Specific root respiration at three harvest dates for P0 (diamonds), P1 (squares) and P2 (circles) phosphorus treatments in Dicorynia guianensis (a) and Eperua falcata (b). Each point is an average value obtained on 5–11 seedlings. M indicates significant effects of mycorrhizal inoculation (P < 0.05, contrast analysis).

Leaf photosynthesis (Fig. 4) was significantly enhanced in mycorrhizal seedlings as compared with nonmycorrhizal ones for the third harvest in D. guianensis (P < 0.001) and for the three harvests in E. falcata (P < 0.05).

Figure 4.

Leaf photosynthesis at three harvest dates for P0 (diamonds), P1 (squares) and P2 (circles) phosphorus treatments in Dicorynia guianensis (a) and Eperua falcata (b). Each point is an average value obtained on two leaves of 5–11 seedlings. M indicates significant effects of mycorrhizal inoculation (P < 0.05, contrast analysis).

Phosphorus acquisition efficiency

Because of lower root respiration rates and lower root growth rates, the carbon cost of the root system was lower for nonmycorrhizal than for mycorrhizal seedlings for both species at each phosphorus concentration, except at very low phosphorus availability (Fig. 5). Mycorrhizal colonization sharply increased the net uptake rate of phosphorus. Phosphorus addition enhanced phosphorus uptake rates in mycorrhizal seedlings of D. guianensis. Despite higher costs, phosphorus acquisition efficiency was improved by mycorrhizal colonization for both species, with D. guianensis seedlings exhibiting higher phosphorus acquisition efficiencies than E. falcata.

Figure 5.

Estimates of phosphorus uptake (a,b), root carbon cost (c,d) and phosphorus acquisition efficiency (e,f) of mycorrhizal (filled bars) and nonmycorrhizal (open bars) between the two last harvests in the different phosphorus treatments in Dicorynia guianensis (a,c,e) and Eperua falcata (b,d,f). Calculations were made from mean values per treatment.


Mycorrhizal status

The range and high mean values of root mycorrhizal colonization (60–95% in D. guianensis and 45–75% in E. falcata) indicate that both species can be considered as mycotrophic species (Siqueira & Saggin-Junior, 2001). Increasing soil phosphorus availability generally leads to a reduction of mycorrhizal colonization of roots (McArthur & Knowles, 1993; Nielsen et al., 1998; Dickson et al., 1999; Mason et al., 2000). This was ascribed to a higher cost of nutrient absorption by mycorrhizal root, and/or to a lower nutrient use efficiency with increasing nutrient availability (Tuomi et al., 2001). In this study, mycorrhizal colonization of roots did not react to phosphorus addition. A similar lack of effect of phosphorus on mycorrhizal colonization has also been reported for some tropical woody species in Brazil (Siqueira et al., 1998; Siqueira & Saggin-Junior, 2001). This suggests that phosphorus availability was limiting in all P treatments. The P0 treatment might have been so poor that even mycorrhiza could not improve the phosphorus nutrition, as highlighted by the positive effect of phosphorus addition on phosphorus content for mycorrhizal seedlings. In this case, the mycorrhizal fungus would act as a parasite for the plant, as it would represent a carbon sink without bringing phosphorus back (Johnson et al., 1997).

Coils were the major mycorrhizal structures found in root cells of both species, replacing arbuscules as intracellular exchange sites. These structures are specific to Paris-type mycorrhizas (Smith & Read, 1997; Cavagnaro et al., 2003). This type is often observed in trees of the French Guyana tropical wet forest (Béreau et al., 2000). Other observed structures (extracellular mycelium and vesicles) are thought to be ineffective for phosphorus acquisition by the plant host, and arbuscules were rarely observed. Only half of the observed microscope fields contain coils in E. falcata roots compared with more than 90% for D. guianensis. This suggests that the mycorrhizal colonization was probably more ‘repressed’ in E. falcata roots than D. guianensis ones. There are no clear relationships between plant responsiveness and mycorrhizal effectiveness (Smith & Read, 1997), but lower coil frequency in E. falcata would suggest reduced transfer between the fungus and the seedling, and thus a lower contribution of the fungus to phosphorus uptake and a lower carbon cost. Both structure and function of Paris-type mycorrhizae have been less well documented than those of the Arum type ones and thus require more attention.

Phosphorus status

Seedlings of both species seemed unable to absorb phosphorus if they were not mycorrhizal. Indeed, the phosphorus concentrations in plant tissue declined similarly in both species as soon as seedling biomass increased in nonmycorrhizal seedlings. Greater seed reserves in E. falcata than in D. guianensis allowed greater biomass accumulation in nonmycorrhizal seedlings of the former species. Large-seeded tropical species were thought to be strongly dependent on mycorrhizal colonization (Janos, 1983) and seed reserve removal could reduce mycorrhizal colonization (Muthukumar & Udaiyan, 2000). However, it was recently acknowledged that large seeds were able to sustain initial seedling growth, while small-seeded species exhibited greater mycorrhizal colonization and responsiveness to symbiosis (Johnson et al., 1997; Siqueira et al., 1998; Kiers et al., 2000; Zangaro et al., 2000). In large-seeded species such as E. falcata, the benefits of mycorrhizal symbiosis should therefore remain low as long as seed reserves are not exhausted. By contrast, small-seeded species such as D. guianensis would rapidly respond to an increased access to external resources.

Mycorrhizal seedlings of both species were able to absorb phosphorus, especially when soil phosphorus availability was increased by the addition of phosphorus. A twofold increase in phosphorus content was observed in mycorrhizal seedlings of E. falcata in response to phosphorus addition while a sixfold increase was observed in D. guianensis. Indeed, E. falcata exhibited much lower rates of phosphorus uptake on a per root mass basis than D. guianensis. Dicorynia guianensis produced a dense mat of fine roots while E. falcata had a coarser, less-branched root system with beaded short roots (Béreau & Garbaye, 1994). We therefore expect important differences in specific root length between the two species, and the difference in rate of phosphorus uptake would have been less pronounced if phosphorus uptake was calculated over fine root length instead of root mass (Moyersoen et al., 1998). The possibility that E. falcata's coarse root system is less efficient than that of D. guianensis might also contribute to lower P uptake.

Phosphorus uptake in E. falcata seems driven more by growth than in D. guianensis. Indeed, phosphorus concentration remained almost constant between the second and the third harvest in E. falcata seedlings while it increased with soil availability in mycorrhizal seedlings of D. guianensis. This ‘luxury’ phosphorus consumption is often reported for seedlings of tropical trees (Thompson et al., 1992; Burslem et al., 1995; Lawrence, 2001). This is thought to enable seedlings to react to an occasional increase in light availability even on nutrient-poor soil (Aerts & Chapin, 2000).

Carbon cost

The slight decrease in root carbon cost in nonmycorrhizal seedlings of both species with increasing phosphorus availability was caused by a similar decrease in root growth rate. By contrast, root carbon cost increased with root growth rate in response to increasing phosphorus availability in mycorrhizal D. guianensis, while root respiration was not affected.

Mycorrhizal inoculation enhanced root respiration in both species while it was unaffected by the phosphorus status of the soil. Higher respiration of mycorrhizal roots has been reported (15–30%; Pearson & Jakobsen, 1993) and it has often been ascribed to higher construction cost that lead to a higher coefficient of growth respiration. However, lipid-rich fungal vesicles found in mycorrhizal roots of Citrus poorly accounted for the increase in root respiration (Peng et al., 1993). The lack of difference in root carbon content between mycorrhizal and nonmycorrhizal roots of both species suggests that differences in construction cost were negligible. A decrease in maintenance respiration of pine roots colonized by ectomycorrhizal fungi has been observed (Marshall & Perry, 1987) but others have reported higher rates of maintenance respiration in mycorrhizal roots (Peng et al., 1993). Higher root growth rate and higher respiration rate for maintenance probably account for the increase in root respiration and root carbon cost with increasing phosphorus availability especially in D. guianensis.

The carbon costs of the extraradical mycelium were not evaluated here and it was assumed that these external costs were limited. Indeed, the fraction of carbon that was incorporated into the extraradical mycelium was less than 1% of total carbon uptake in Cucumis seedlings (Jakobsen & Rosendahl, 1990; Pearson & Jakobsen, 1993). Despite a high turnover rate of extraradical mycelium (Staddon et al., 2003), carbon input into the soil as soil organic matter was not significantly affected by mycorrhizal inoculation in Plantago lanceolata (Staddon et al., 1999) and does not act as a significant substrate for microbial activity (Jakobsen, 1999).

Photosynthesis and growth response

The increase in leaf phosphorus concentration accounted for the enhanced photosynthetic rate in mycorrhizal seedlings of both species. Higher photosynthetic rates were often observed in mycorrhizal plants and were ascribed to an increased phosphorus status (Fredeen & Terry, 1988; Black et al., 2000) and sometimes to a greater sink strength of mycorrhizal roots (Wright et al., 2000).

Increased leaf area also contributed to the increase in whole-plant carbon gain, which could offset the higher cost of mycorrhizal roots and could account for greater biomass accumulation in mycorrhizal seedlings of D. guianensis. Phosphorus deficiency is known to decrease both the rate of leaf expansion and rate of leaf appearance (Fredeen et al., 1989; Rodriguez et al., 1998). The main effect of addition of phosphorus in mycorrhizal seedlings of D. guianensis was an increase in leaflet area, while the number of leaves produced remained unaffected. This increase in leaf expansion could have been driven by the increased availability of photoassimilates or by direct effects of phosphorus deficiency on cell wall properties.

This increased leaf area reflected changes in carbon allocation, with more carbon being allocated to leaves, leading to higher relative leaf mass ratio and leaf area ratio. The higher leaf mass ratio was paralleled by a lower stem mass ratio, while the root mass ratio were unaffected (not shown). The root–shoot ratio is usually lower in mycorrhizal seedlings (Smith & Gianinazzi-Pearson, 1988; Michelsen & Rosendahl, 1990; Berta et al., 1995) reducing the cost of the root system. As already reported for seedlings grown at 50% of outside irradiance (Béreau et al., 2000), this was not the case for D. guianensis. Maintenance of a high root–shoot ratio allows exploring a larger soil volume and it would contribute to the higher rate of phosphorus uptake in mycorrhizal seedlings, overcoming the lack of cost reduction.

High phosphorus concentration was required to sustain growth in seedlings of D. guianensis. This would account for the high responsiveness of D. guianensis to mycorrhizal colonization (Koide, 1991; Smith & Read, 1997). By contrast, species exhibiting an inherently low growth rate, such as E. falcata, would be able to cope with low nutrient availability and would therefore be less responsive to both mycorrhizal colonization and addition of phosphorus (Koide, 1991). A similar lack of response to the addition phosphorus has recently been reported for five ectomycorrhizal seedlings in central Africa, despite higher leaf phosphorus contents in some species (Newbery et al., 2002).


Higher uptake rates in mycorrhizal seedlings led to higher phosphorus acquisition efficiency, especially in D. guianensis, despite higher root carbon cost. Both species gained benefits from the mycorrhizal symbiosis in terms of phosphorus acquisition but the growth of E. falcata seedlings seemed unresponsive to this mycorrhizal improvement of phosphorus status, probably because of high seed reserves and low growth rate. By contrast, seedlings of D. guianensis were highly responsive to mycorrhizal inoculation, at least under favourable light conditions such as those encountered in forest gaps. Our results suggest that, in forest gaps, D. guianensis would have greater success in rich-nutrient patches while E. falcata would be more evenly distributed. This conclusion cannot be extended to the deeper shade conditions generally encountered in the understorey (1–5% full sun). Indeed, recent studies have demonstrated a strong influence of light conditions on seedling growth and mycorrhizal colonization (Torti et al., 1997; Green & Newbery, 2001), while others reported that light environment had no effect on mycorrhizal colonization of field-grown seedlings (Hurst et al., 2002). The effect of light conditions on both mycorrhizal colonization and responsiveness will therefore be further investigated in these species.

The two co-occurring Ceasalpiniaceae clearly belong to two different functional groups for phosphorus acquisition. Seedlings of D. guianensis seemed unable to grow and to acquire phosphorus without mycorrhizal colonization (obligate mycotrophic species sensuJanos, 1983). By contrast, E. falcata did not exhibit any dependency to mycorrhizal inoculation for growth, and did not respond to phosphorus addition within this range of soil availability (facultative mycotrophic species), at least at the seedling stage when seed reserves are not fully exhausted.


We thank Damien Bonal, Jean Christophe Roggy, Christopher Baraloto and Alexandre Bosc for valuable discussions. We are grateful to Pascal Imbert, Jean Yves Goret and Audin Patient for managing the seedlings and for their help during the measurements. We are also grateful to Jacqueline Marchand for chemical analyses. We acknowledge financial support from the GIP ECOFOR.