SEARCH

SEARCH BY CITATION

Keywords:

  • Amazonian lowland forest;
  • detritivores;
  • elemental ratios;
  • litter chemistry;
  • litter diversity;
  • nitrogen;
  • nutrient dynamics;
  • phosphorus;
  • soil fauna

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgements
  9. References

1. Ecological stoichiometry predicts important control of the relative abundance of the key elements carbon (C), nitrogen (N) and phosphorus (P) on trophic interactions. In a nutrient-poor Amazonian lowland rain forest of French Guiana, we tested the hypothesis that decomposers exploit stoichiometrically diverse plant litter more efficiently, resulting in faster litter decomposition compared to litter with a uniform stoichiometry.

2. In a field experiment in the presence or absence of soil macrofauna, we measured litter mass loss, and N and P dynamics from all possible combinations of leaf litter from four common tree species which were distinctly separated along a C:N and along a N:P gradient.

3. Mean litter mass remaining after 204 days of field exposure varied between 25.2% and 71.3% among litter treatments. Fauna increased litter mass loss by 18%, N loss by 21% and P loss by 14%. Litter species richness had no effect on litter mass loss or nutrient dynamics. In contrast, litter mass and nutrient losses increased with increasing stoichiometric dissimilarity of litter mixtures in presence of fauna, suggesting faster decomposition of a stoichiometrically more heterogeneous litter.

4. However, the effect of stoichiometric dissimilarity was smaller than the strong C quality related litter composition effect and disappeared in the absence of fauna. Increasing proportions of litter that is relatively rich in accessible C compounds (non-structural carbohydrates, phenolics) and relatively poor in recalcitrant C (condensed tannins, lignin), correlated best with litter mass loss irrespective of fauna presence. No correlation was found for any of the nutrient related litter quality parameters and decomposition.

5.Synthesis. Our results suggest that Amazonian decomposer communities studied here are primarily limited by energy, and only secondarily by litter stoichiometry. Tropical tree species might thus influence decomposers and detritivores by the production of litter of specific C quality with potentially important feedback effects on ecosystem nutrient dynamics and availability.


Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgements
  9. References

Soils at low latitudes are typically very old, highly weathered, and thus, often extremely poor in key plant nutrients, phosphorus in particular (Irion 1978; Uehara & Gillman 1981; Vitousek & Sanford 1986). As a consequence, tropical forest plants tend to show lower tissue nutrient concentrations and wider N:P ratios than plants from higher latitudes that generally grow on more fertile soils (McGroddy, Daufresne & Hedin 2004; Reich & Oleksyn 2004). Low concentrations of foliar nutrients may even be accentuated in litter as a result of a particularly efficient resorption of limiting nutrients during leaf senescence (Vitousek 1998; Wright & Westoby 2003; Hättenschwiler et al. 2008). This leads to some of the lowest nutrient concentrations and widest litter carbon to nutrient ratios measured in plant leaf litter in tropical forests compared to other forest ecosystems (Hättenschwiler et al. 2008). Decomposer communities as the lowest heterotrophic level of the complex soil food web are directly affected by the chemical composition of plant litter. As these organisms have distinctively smaller C:nutrient ratios than the plant material they consume (Sterner & Elser 2002; Cleveland & Liptzin 2007; Martinson et al. 2008), ecological stoichiometry predicts important constraints on decomposer communities (Sterner & Elser 2002; Cherif & Loreau 2007). In accordance, substrate stoichiometry, and in particular the relative abundance of N and P, appear to dictate the stoichiometry of microbial biomass to a large extent (Cleveland & Liptzin 2007) and can control the composition and activity of microbial communities and the processes they drive (Güsewell & Gessner 2009).

Lowland tropical rain forests are among the most species-rich plant communities, with typically between 100 and 200 different tree species per hectare (Ter Steege, Sabatier & Castellanos 2000). Tree species vary tremendously in foliage stoichiometry at large regional scales (Townsend et al. 2007) and at small local scales (Hättenschwiler et al. 2008). It has also been shown that interspecific variability in litter chemistry and stoichiometry increases, compared to foliage chemistry of the same trees in a neotropical rain forest community of French Guiana (Hättenschwiler et al. 2008). When such chemically distinct litter from different species is decomposing together, decomposition rates are often different in mixtures than expected from single species litter (Wardle, Bonner & Nicholson 1997; Gartner & Cardon 2004; Hättenschwiler, Tiunov & Scheu 2005). These mixture effects have important consequences for carbon cycling and nutrient dynamics (Wardle, Bonner & Nicholson 1997; Finzi & Canham 1998; Hättenschwiler, Tiunov & Scheu 2005), but the underlying mechanisms of interactions among different litter types have rarely been identified in past studies.

If litter-feeding organisms and decomposer communities are limited by the relative availability of major elements such as C, N and P, it might be expected that their activity increases with a substrate composed of litter material of distinct C, N and P concentrations, complementing their stoichiometric needs. Alternatively or additionally, stoichiometrically heterogeneous litter might be expected to provide diverse substrates for a variety of soil organisms that were reported to differ considerably in their stoichiometry (Cleveland & Liptzin 2007; Martinson et al. 2008), resulting in more diverse soil communities with potentially more efficient substrate use overall. Consequently, substrate exploitation and ultimately litter decomposition would also be expected to be faster with increasing stoichiometric heterogeneity of litter mixtures. We tested this hypothesis with a field experiment in an Amazonian forest of French Guiana by using all possible combinations of four litter species which were distinctly separated along a C:N and along a N:P gradient. We also addressed the question to what extent litter decomposition and nutrient turnover is driven by litter detritivores such as earthworms or litter-dwelling macroarthropods, and whether their presence changes the relationship between litter diversity and decomposition compared to microbially dominated decomposition.

Materials and methods

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgements
  9. References

Study site and plant material

The experiment was set up in an undisturbed Amazonian lowland tropical rain forest at the Guyaflux experimental site located near Sinnamary, French Guiana (5°18′ N, 52°55′ W) in December 2004 (see Bonal et al. 2008 for a more detailed site description). Temperature varies only slightly during the course of the year, but relatively important intra-annual variations in rainfall are observed. This results in two distinct rainy seasons, a moderate one from December to February and a stronger one from April to July. Mean annual temperatures in 2004 and 2005 were 25.7 and 25.9 °C, and total annual rainfall was 2756 and 3072 mm, respectively. Total rainfall during the experiment (mid-November 2004 to mid-June 2005) was 2166 mm. Soils in the study area are acrisols (FAO 1998) developed over a Precambrian metamorphic formation called the Bonidoro-series (see Table 1 for detailed soil characteristics from our study site).

Table 1.   Soil characteristics of the experimental site (n = 4 blocks ± SD)
ParameterUndisturbed topsoil (top 5 cm)Homogenized soil (top 25 cm)
  1. All analyses were done at the INRA laboratory in Arras, France using standard protocols. Total P, K, Ca, Mg, Fe, Al, and Mn was measured by ICP-AES after extraction with HF and HClO4. Different superscript letters indicate significant differences between undisturbed topsoil and the homogenized soil using paired t-tests.

Loam (%)15.7 ± 1.9a16.1 ± 5.8a
Silt (%)5.3 ± 0.5a6.4 ± 1.2a
Sand (%)80.0 ± 2.2a77.5 ± 0.7a
pH4.81 ± 0.1a4.80 ± 0.01a
Total C (g kg−1)18.7 ± 0.8a16.3 ± 1.1b
Total N (g kg−1)1.24 ± 0.04a1.07 ± 0.07b
Total P (mg kg−1)24.8 ± 3.2a22.8 ± 9.0a
‘Olson P’ (mg kg−1)3.3 ± 0.5a5.5 ± 1.3b
Total K (mg kg−1)7.8 ± 3.2a7.6 ± 4.1a
Total Ca (mg kg−1)3.5 ± 0.7a3.7 ± 0.7a
Total Mg (mg kg−1)2.5 ± 0.4<2.0
Total Fe (mg kg−1)218.8 ± 74.3a339.3 ± 128.8a
Total Al (mg kg−1)300.0 ± 67.7a309.5 ± 110.9a
Total Mn (mg kg−1)38.0 ± 5.8a25.1 ± 10.0a

The forest is evergreen and composed of about 140 canopy tree species (Bonal et al. 2008). The four tree species Dicorynia guianensis (G.J.Amshoff), Eperua falcata (J.B. Aublet), Caryocar glabrum (Persoon), and Platonia insignis (Martius) were chosen for our experiment, based on the criteria of abundance (all four species are common and relatively abundant in the study area), and most importantly, of distinct C:N:P litter stoichiometries. To have a representative litter pool for each species, we collected fresh fallen leaf litter within a tree plantation located 10 km from the study site. These somewhat more than 20-year-old tree stands have been established using local seed sources and have a fully closed canopy composed of about 40 individuals of each species growing in monocultures (see Roy, Hättenschwiler & Domenach 2005 for more details). Litter was collected regularly from June 2004 to November 2004 in suspended litter traps and pooled across sampling dates. Only fresh fallen litter was selected. Leaves with obvious signs of herbivory, galls, fungal attacks, leaves that were clearly very thin (shaded canopy position), green or greenish, or having an otherwise atypical colouration were excluded. Excluded leaves typically represented < 15% of total collected leaf litter per species. Litter was air-dried immediately after collection and kept dry until used.

Experimental design

To test the hypothesis that decomposer limitation decreases, and thus, decomposition rates are increasing with increasing heterogeneity in litter stoichiometry, we exposed all possible combinations of the four litter species in the field. Besides manipulating the factor ‘litter composition’, we also included the factor ‘macrofauna presence’ in our design, to test whether or not the presence of macrofauna affects decomposition and interacts with litter composition. One replicate of each litter treatment (15 litter combinations, i.e. four single species, six 2-species, four 3-species and one 4-species combination) and two fauna treatments (with and without macrofauna), was randomly assigned to each of four blocks using a randomized complete block design (15 litter treatments × 2 fauna treatments × 4 blocks = 120 experimental units). The four blocks were established at the study site within a flat area of a homogenous aspect and tree canopy cover (leaf area index of about 6) with an average distance of about 50 m among blocks.

Treatment effects were examined using custom-made field microcosms, with a diameter of 0.15 m and a height of 0.25 m constructed from plastic cylinders. The bottom and two 0.15 × 0.15 m large openings on the sides were covered with either 10 or 0.5 mm nylon mesh depending on the fauna treatment. The large mesh-covered microcosms allowed full access for above- and below-ground animals of the macrofauna community, while the microcosms covered with the small mesh width, excluded macrofauna. To cover the top of the microcosms, we used a uniform 0.5-mm mesh across both fauna treatments, to avoid contamination with small litter particles falling from the canopy, and to keep the rate and distribution of passing rain water constant across treatments.

Each microcosm was inserted 0.15 m into the forest floor with a distance of c. 1 m among microcosms. The soil removed prior to microcosm installation (dug down to c. 25 cm) was well mixed within each block, freed of stones and roots by hand, sieved and homogenized by passing several times through a 0.005-m sieve. Equal amounts (0.00265 m3 per microcosm) of homogenized soil were put back into each microcosm. A total of 10 g of air-dried leaf litter was added into the 0.1-m tall space above the soil surface of each microcosm on 20 November 2004. Litter mixtures contained equal amounts of component species (e.g. 2.5 g per species in the 4-species mixture).

A subsample of the mixed soil added to the microcosms was taken from each block for soil analyses. In addition, we randomly sampled five soil cores from the intact topsoil (upper 5 cm) from each block prior to microcosm installation using a custom-made stainless steel cylinder (diameter of 5 cm). These additional topsoil samples were bulked within each block and processed as described above. All soil samples were sent to the INRA laboratory at Arras, France, for standard soil analyses presented in Table 1. Apart from slightly higher total C and total N, and slightly lower ‘Olson P’, concentrations in the undisturbed topsoil, the homogenized soil representing a 25-cm deep soil profile had very similar characteristics (Table 1).

Data collection and chemical analyses

Remaining litter was retrieved from all microcosms after a total of 204 days of exposure in the field on 12 June 2005. Litter was removed from all microcosms, and the soil was passed through a 0.004-m sieve to collect litter within the soil (only very small amounts of < 5% were found in the soil, if any), and plant roots that grew into the microcosms during the experiment. All litter was rinsed with water to remove soil particles, separated into species and dried at 65 °C. Based on colour, leaf structure and texture, and patterns of leaf veins, litter species identification was possible for even very small fragments. Unidentified litter material, where present, accounted for < 3% of total litter mass remaining and was attributed equally to the respective species represented in the mixture. Plant roots from individual microcosms were rinsed and dried at 65 °C.

For the determination of initial litter quality, four randomly taken subsamples out of each species’ litter pool were dried at 65 °C and ground up using a centrifugal mill (Cyclotec Sample Mill; Tecator, Höganäs, Sweden) to obtain a uniform particle size of < 1 mm. Carbon (C), nitrogen (N), phosphorus (P), water-soluble compounds (WSC), hemicelluloses, cellulose, lignin, sugar, starch and phenolics were analysed using standard methods for each parameter. Carbon and N concentrations were measured using a CN elemental analyser (Flash EA1112 Series; ThermoFinnigan, Milan, Italy). Phosphorus concentration was measured colourimetrically with an autoanalyser (Alliance Instruments, Evolution II, Frépillon, France) using the molybdenum blue method (Grimshaw, Allen & Parkinson 1989) after mineralization of a litter subsample with 36N H2SO4 and H2O2 at 360 °C. WSC, hemicelluloses, cellulose and lignin were determined according to the van Soest extraction protocol (van Soest & Wine 1967) using a fibre analyser (Fibersac 24; Ankom, Macedon, NJ, USA). Sugar and starch were determined from a 10-mg sample that was mixed with 3 mL of water and heated in a pressure cooker for 30 min. An aliquot of the dissolved sample was digested by invertase to transform sucrose into fructose and glucose, followed by complete conversion into glucose using isomerase. An aliquot of the initial sample was treated with clarase to break up starch into glucose. Final glucose concentration was measured colourimetrically at a wavelength of 340 nm by adding hexokinase and dehydrogenase. Soluble phenolics were measured from a 1-g sample to which 60 mL of water was added and shaken for 2 h. After filtering, the filtrate was diluted and phenolic concentration was measured colourimetrically with Folin-Ciocalteau reagent following the method of Marigo (1973) using gallic acid as a standard. Total phenolics concentration was measured using methanol (50%) instead of water as a solvent followed by the same procedure described above.

All litter remaining was analysed for C, N and P using the same methods as described above. These final nutrient concentrations were used to calculate net nutrient fluxes during field exposure.

Data analysis

We used single factorial analyses of variance (anova) to test for differences in species-specific initial litter quality, and post hoc pair-wise comparisons (Tukey adjusted) to identify differences among species (Table 2). The relationship between initial litter quality and total litter mass remaining at the end of the experiment was explored with linear regression analyses. Initial litter quality of litter mixtures was calculated as the average from the relative contribution of each species present. We used analysis of variance with sequential sums of squares to test for the effects of block, litter diversity (separated into a richness and species composition component) and fauna presence (all considered as random factors) on total litter mass remaining, total N remaining and total P remaining.

Table 2.   Litter quality parameters of the four studied species (n = 4 ± SE)
Parameter (in % dry mass)Platonia insignisDicorynia guianensisEperua falcataCaryocar glabrum
  1. Different superscript letters indicate significant differences among species using Tukey adjusted pair-wise contrasts (afor the highest value of each litter trait).

  2. *The values of condensed tannins (CT) are from Coq et al. (2010) using the acid butanol assay. CT was measured on leaf litter collected at the same site and from the same trees, but from a later harvest period than the leaf litter material used in this study.

Carbon (C)48.3 ± 0.1b49.8 ± 0.3a49.8 ± 0.2a48.1 ± 0.2b
Nitrogen (N)1.32 ± 0.02a1.17 ± 0.03b1.01 ± 0.03c0.83 ± 0.03d
Phosphorus (P)0.0197 ± 0.001c0.0239 ± 0.002b0.0317 ± 0.002a0.0242 ± 0.002b
C:N36.5 ± 0.6d42.7 ± 1.2c49.4 ± 1.3b58.1 ± 1.9a
N:P68.4 ± 4.9a49.7 ± 3.6b32.2 ± 2.0c35.2 ± 3.6c
Non-structural carbohydrates3.33 ± 0.32a0.24 ± 0.06c0.99 ± 0.09b1.04 ± 0.09b
Water-soluble compounds35.1 ± 1.1b26.6 ± 0.7c40.8 ± 0.7a42.5 ± 1.1a
Hemicellulose25.5 ± 0.9a8.0 ± 0.6c11.8 ± 0.3b12.1 ± 0.4b
Cellulose20.7 ± 0.4ab22.8 ± 1.0a18.1 ± 0.6b14.2 ± 0.9c
Lignin18.6 ± 0.1c42.6 ± 0.6a29.4 ± 0.5b31.3 ± 1.1b
Soluble phenolics0.8 ± 0.1b1.0 ± 0.2b2.6 ± 0.3a3.0 ± 0.2a
Total phenolics12.6 ± 0.6b5.9 ± 0.3c13.6 ± 1.1ab15.9 ± 1.0a
CT*0.4 ± 0.03d4.2 ± 0.07a3.5 ± 0.09b2.5 ± 0.05c

For a more specific test of our hypothesis that the diversity in stoichiometry rather than species richness drives litter mixture effects on decomposition, we calculated a stoichiometric dissimilarity index for each litter mixture. The calculation of stoichiometric dissimilarity is based on the functional attribute diversity (FAD2) proposed by Walker, Kinzig & Langridge (1999) for plant communities and is analogous to the functional dissimilarity used by Heemsbergen et al. (2004) to characterize litter detritivore communities. Briefly, this concept makes use of the standardized distance that separate species in attribute space, which is calculated as the Euclidean distance (ED):

  • image

where Aij and Aik are the attribute values of species j, k for attribute i, and I is the total number of attributes being considered. Before the calculation of ED, all attributes or traits considered were standardized by subtracting the mean value across all species from the mean of a given species divided by the range across all species for each individual trait (Botta-Dukát 2005). Trait standardization avoids a biased weight of certain traits related to their units or inherent variability.

In our case, the attribute space is defined by the two attributes of litter C:N ratio and N:P ratio, thus = 2. Using ED, we created a matrix for the four litter species. The index of FAD2 was then calculated as the sum of the distances between species pairs, divided by the number of species pairs included in the calculation:

  • image

where n is the number of species, and N the number of species pairs considered for the calculation (i.e. 1 for 2-species mixtures, 3 for 3-species mixtures and 6 in the case of the 4-species mixture). The higher the value of FAD2, the more dissimilar in stoichiometry is the mixture.

Simple linear regression analyses were used to test for relationships between litter mass loss, N loss and P loss (dependent variables) and the stoichiometric dissimilarity of the litter mixtures (independent variable). All data expressed in percentages were arcsine-square-root-transformed prior to statistical analysis and normal distribution and homoscedasticity of the residuals were confirmed by visual assessment using diagnostic plots. The R package (version 2.4.0; R Development Core Package 2006) and systat, version 5.2.1 (Systat Inc., Evanston, IL, USA) were used for statistical analyses.

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgements
  9. References

Initial litter quality

Clear and significant differences in all measured litter chemistry traits were observed among the four species (Table 2). Concentrations of C (F3,12 = 21.5***), N (F3,12 = 57.6***) and P (F3,12 = 7.1**) were all significantly affected by species identity. Litter N concentration was highest in P. insignis leaf litter and decreased in the order of P. insignis > D. guianensis > E. falcata > C. glabrum. Phosphorus concentration varied differently from N concentration and decreased in the order of E. falcata > C. glabrum = D. guianensis > P. insignis. The contrasting concentrations of N and P among species resulted in opposing gradients in C:N and N:P ratios, and thus, in distinct stoichiometries among litter species (Table 2). Both, C:N ratio (F3,12 = 48.0***) and N:P ratio (F3,12 = 20.4***) showed a significant species effect.

Litter from all species showed measurable concentrations of non-structural carbohydrates (NSC, Table 2) which were significantly affected by species identity (F3,12 = 70.1***). The highest NSC concentrations were measured in P. insignis with decreasing concentrations in the order of P. insignis > C. glabrum = E. falcata > D. guianensis. In C. glabrum, NSC was entirely composed of sugar without any starch (data not shown). Likewise, in P. insignis there was mostly sugar, with only 7% starch of total NSC. In contrast, starch contributed 82% and 60% to total NSC in D. guianensis and E. falcata, respectively. The carbon fractions according to the van Soest profile all showed a significant species identity effect (WSC: F3,12 = 60.8***, hemicellulose: F3,12 = 178.5***, cellulose: F3,12 = 23.3***, lignin: F3,12 = 220.4***). Total phenolics also showed a significant species effect (F3,12 = 36.5***) and varied differently among species than lignin (Table 2). For example, D. guianensis as the species with the highest lignin concentration had less than half the concentration of total phenolics compared to the species with the second lowest value. Soluble phenolics were, on average, five times lower than total phenolics, but with similar interspecific differences (Table 2, species effect: F3,12 = 36.5***). Condensed tannins (CT) in our studied species were measured by Coq et al. (2010) and showed also a significant species effect (F3,12 = 645.0***), with all species being significantly different from each other.

Fauna and litter composition effects on mass loss

Remaining litter mass after 204 days of exposure in the field was significantly affected by fauna presence and litter composition, but not by litter species richness (Table 3). The fauna effect ranged from no significant differences to 110% higher mass loss (Fig. 1), with an overall increased mass loss of 18% across all litter treatments. Fauna presence generally accentuated the differences in mass loss among the different litter treatments (Fig. 1). For example, remaining litter mass of the four litter species decomposing alone was not very different when fauna were excluded, with only D. guianensis significantly different from the other three species (Fig. 1). In the presence of fauna, however, average litter mass remaining varied more among the four single species treatments (Fig. 1). These differences in single species decomposition contributed importantly to the overall fauna and composition effect (Table 3). However, the same analysis run without the single species treatments still showed significant effects of fauna (F1,62 = 15.9***) and litter composition (F8,62 = 2.6*), with the effects of the remaining factors and interactions qualitatively unchanged compared to the full model including all litter treatments.

Table 3.   Analyses of variance to test for differences in remaining total litter mass, N and P. Indented terms show the linear contrasts for litter species richness and litter species composition that are not independent of one another (the sum of these two terms correspond to the total litter diversity effect)
ParameterSource of varianced.f.Mean squareF-valueP-value
Litter massBlock377.64.40<0.01
Fauna (F)1385.321.82<0.001
 Litter species richness (R)39.250.520.67
 Litter  composition (C)11132.87.52<0.001
F × R318.91.070.37
F × C1127.51.560.13
Residuals8417.7  
NBlock3198.65.35<0.01
Fauna (F)1889.623.95<0.001
 Litter species  richness (R)38.630.230.87
 Litter  composition (C)11105.82.85<0.01
F × R369.51.870.14
F × C1150.51.360.21
Residuals8437.2  
PBlock322.40.750.52
Fauna (F)1115.53.890.052
 Litter species  richness (R)332.81.100.35
 Litter  composition (C)1144.31.490.15
F × R320.50.690.56
F × C1166.32.230.020
Residuals8429.7  
image

Figure 1.  Average total litter mass remaining [in per cent of total initial dry mass (d.m.)] after 204 days of exposure in the field for the 15 different litter composition treatments (mean ± SE, n = 4). Litter treatments are indicated with the first letter of the component species (D = Dicorynia guianensis, E = Eperua falcata, C = Caryocar glabrum, P = Platonia insignis). Open and closed bars show data from microcosms without fauna access and with free access for macrofauna, respectively. Different letters indicate significant differences among litter treatments within each fauna treatment separately (Tukey post hoc contrasts at < 0.05) and symbols indicate significant differences between the two fauna treatments within litter treatments (**< 0.01, *< 0.05, +< 0.1).

Download figure to PowerPoint

Litter diversity effects on mass loss and nutrient dynamics

Similar to total litter mass, N and P dynamics were not affected by litter species richness, irrespective of whether fauna were present or not (Fig. 2, Table 3). Net N loss from litter correlated well with mass loss (r2 = 0.83). However, N was lost at a lower rate than total litter mass, resulting in an average 30% higher N concentration across all litter treatments compared to initial litter material (Fig. 2). In line with total litter mass, fauna presence and litter composition significantly affected N dynamics (Table 3, Fig. 2). In the presence of fauna, 21% more N was lost on average across all litter treatments. In contrast to N, net P loss from litter showed no significant correlation with litter mass loss (r2 = 0.14) and was lost at a higher rate than litter mass (Fig. 2). The presence of fauna had a marginally significant effect on P dynamics (14% higher P loss across all litter treatments compared to microcosms without fauna access), but litter composition did not significantly affect P dynamics (Table 3, Fig. 2). However, litter composition interacted significantly with fauna presence on total remaining P (Table 3), indicating that fauna influenced P dynamics differently depending on litter composition.

image

Figure 2.  Total litter mass, total nitrogen (N) and total phosphorous (P) remaining [in per cent of total initial dry mass (d.m.)] after 204 days of exposure in the field as a function of litter species number (top panel), or as a function of stoichiometric dissimilarity (bottom panel). In the top panel, each symbol represents the average of four microcosms of four single species treatments, six 2-species mixtures, four 3-species mixtures and one 4-species mixture. In the bottom panel, each symbol represents the average of four microcosms from each of the 11 different mixtures along the dissimilarity gradient (open circles: no fauna access, closed diamonds: free access for macrofauna). In the bottom panel, lines indicate fitted simple linear regressions for the two fauna treatments separately, when statistically significant. Total mass, total N and total P remaining did not significantly change along the dissimilarity gradient without fauna (mass: F1,42 = 1.7, = 0.20; N: F1,42 = 0.1, = 0.77; P: F1,42 = 1.6, = 0.21). However, in the presence of fauna, both remaining mass (F1,42 = 6.3, = 0.016) and remaining N (F1,42 = 4.5, = 0.039) decreased significantly, and remaining P marginally significantly (F1,42 = 3.8, = 0.059) with increasing stoichiometric dissimilarity.

Download figure to PowerPoint

Grouping the different treatment combinations into replicated blocks explained a significant portion of the variation in remaining litter mass and N, but not of P (Table 3). This block effect was driven entirely by the fauna treatment, for which particularly high mass losses were observed in two out of the four blocks. These two blocks were not any different with respect to decomposition processes when fauna were excluded, suggesting important spatial variability in fauna abundance and/or activity as the main driver for the block effect.

While litter species richness had no effect on remaining litter mass and nutrients, remaining litter mass and N were lower in litter mixtures that were more dissimilar in stoichiometry, but only when fauna had access (Fig. 2). It is important to note that stoichiometric dissimilarity was not correlated with the number of litter species present. For example, the most dissimilar mixture and the least dissimilar mixture are both 2-species mixtures, and the 4-species mixture had an intermediate dissimilarity index. There was also a marginally significant negative effect of increasing stoichiometric dissimilarity on P remaining in the presence of fauna. However, when macrofauna were excluded from the microcosms, there was no significant correlation between remaining litter mass and nutrients and stoichiometric dissimilarity of the litter (Fig. 2).

Predictors of decomposition

Remaining litter mass of the different litter treatments was not related to initial C:N or N:P ratios, both across the two fauna treatments and for each fauna treatment independently (Fig. 3). The same was found for remaining mass as a function of initial N concentration (no fauna: F1,13 = 2.3, = 0.16, with fauna: F1,13 = 0.05, = 0.82), or initial P concentration (no fauna: F1,13 = 1.6, = 0.23, with fauna: F1,13 = 1.4, = 0.26).

image

Figure 3.  Total litter mass remaining [in per cent of total initial dry mass (d.m.)] after 204 days of exposure in the field as a function of initial litter chemistry. Each symbol represents the average of four microcosms of 15 different litter combinations (open circles: no fauna access, closed diamonds: free access for macrofauna). Lines indicate fitted simple linear regressions for the two fauna treatments separately, when statistically significant. Total mass remaining did not correlate with initial C : N in either fauna treatment (with fauna: F1,13 = 0.1, = 0.75, no fauna: F1,13 = 2.4, = 0.15). There was also no correlation with initial N : P (with fauna: F1,13 = 0.6, = 0.47, no fauna: F1,13 = 1.6, = 0.23). In contrast, total mass remaining correlated well with initial total phenolics (with fauna: F1,13 = 23.3, < 0.001, no fauna: F1,13 = 45.2, < 0.001), and with condensed tannin (CT) when fauna were present (F1,13 = 28.7, < 0.001), but not in their absence (F1,13 = 2.3, = 0.15). Note that total phenolics did not correlate with CT (F1,13 = 3.7, = 0.08, r2 = 0.22).

Download figure to PowerPoint

In contrast, remaining litter mass correlated well with initial C quality. Across the two fauna treatments, remaining mass decreased with increasing initial concentrations of NSC, hemicellulose, soluble phenolics and total phenolics. The opposite correlation of increasing remaining mass was observed with increasing initial concentrations of CT, cellulose and lignin. However, depending on whether or not fauna were present, the strength of these correlations varied. For example, remaining mass decreased significantly with increasing initial NSC in the presence of fauna (F1,13 = 16.3, < 0.01, r2 = 0.56), but not without fauna (F1,13 = 1.9, = 0.19, r2 = 0.13). A similar pattern was observed for hemicellulose. With increasing hemicellulose concentration, mass loss was faster when fauna had access (F1,13 = 14.8, < 0.01, r2 = 0.53), but not in the absence of fauna (F1,13 = 1.8, = 0.23, r2 = 0.11). Soluble phenolics showed a contrasting pattern of no relationship with remaining litter mass in the presence of fauna (F1,13 = 0.7, = 0.43, r2 = 0.05), and a significant negative relationship without fauna (F1,13 = 8.0, = 0.015, r2 = 0.38). Some of the best predictions were found with total phenolics. Mass remaining decreased clearly with increasing concentrations of total phenolics in the litter, both with and without fauna access to the microcosms (Fig. 3). CT as another functional group of polyphenols had opposite effects compared to total phenolics. The positive correlation between remaining litter mass and initial CT concentration also depended on the fauna treatment, with a significant effect only in presence of fauna (Fig. 3). Litter mass loss was slower with increasing initial cellulose concentration, both with fauna (F1,13 = 5.4, = 0.037, r2 = 0.29) and without fauna (F1,13 = 12.3, < 0.01, r2 = 0.49). Likewise, the positive correlation between remaining litter mass and lignin was significant when fauna were present (F1,13 = 34.5, < 0.001, r2 = 0.73), as well as in its absence (F1,13 = 8.9, = 0.011, r2 = 0.41).

From all the litter chemistry parameters we measured, total litter mass remaining correlated best with lignin (positively) when fauna were present, and best with total phenolics (negatively) when there were no fauna.

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgements
  9. References

We initially hypothesized that mixing of leaf litter from different tree species varying in their concentrations of the major elements C, N and P would provide a stoichiometrically more favourable substrate for detritivore and decomposer communities, resulting in improved substrate utilization, and thus, more rapid litter decomposition compared to stoichiometrically uniform substrates of single litter species. However, the increasing number of stoichiometrically distinct litter species decomposing together explained no variation in the losses of total litter mass, N, or P, independently of whether or not soil macrofauna had access to the field microcosms. No effect of litter species richness on decomposition of litter mixtures is in line with a number of previous studies (Wardle, Bonner & Nicholson 1997; Wardle et al. 2006; Scherer-Lorenzen, Bonilla & Potvin 2007; Ball et al. 2008). Rather than species number per se, species identity and composition, i.e. the presence of litter from particular plant species or the specific combination of different litter types, appear to drive observed differences in decomposition of litter mixtures (Wardle, Bonner & Nicholson 1997; Wardle et al. 2003; Madritch & Cardinale 2007; Ball et al. 2008; Pérez-Harguindeguy et al. 2008; Hoorens, Coomes & Aerts 2010). These species identity and composition effects are related to interspecific differences in functional litter traits including litter chemistry and structural characteristics (e.g. Hoorens, Coomes & Aerts 2010). A functional approach to litter diversity such as proposed in the chemical diversity concept by Epps et al. (2007) or the functional dissimilarity sensuHeemsbergen et al. (2004), may thus provide better insight in underlying mechanisms of mixture effects than the number of species. A recent laboratory experiment showed, that chemical diversity, based on nine chemical traits of leaf litter from four alpine plant species mixed in all possible combinations, explained soil respiration and net N mineralization in litter mixtures better than species richness (Meier & Bowman 2008).

With a similar approach, the 11 different litter mixtures used here, were placed along a stoichiometric dissimilarity gradient that did not correlate with species richness. Our data showed that litter mass and nutrient losses increased with increasing stoichiometric dissimilarity when fauna had access, with up to 44% of the variation explained. In contrast to the conclusion based on litter species richness, these results support our initial hypothesis and suggest that an unaltered decomposer community more efficiently decomposes a stoichiometrically dissimilar litter substrate. However, this relationship was restricted to the presence of soil macrofauna and disappeared in their absence. This result indicates that microbial-dominated decomposition in this tropical rain forest is not sensitive to substrate stoichiometric heterogeneity, at least at the spatial and temporal scale covered by our experiment. Even more strikingly, litter mass and nutrient losses did not correlate with the average litter C:N and N:P ratios, nor with the absolute concentrations of N and P in the different litter treatments. Decomposers, thus, did not preferably exploit litter with highest total amounts of nutrients, even though the mean C:N ratio of 35 in remaining litter at the end of the experiment was still above the threshold of around 30, indicative for microbial N limitation (Kaye & Hart 1997). The average N:P even doubled from 46 to 91 at the end of the experiment, suggesting that relative P limitation should have remained high over the course of the experiment.

The lack of correlation between litter mass loss and initial litter nutrient concentrations is surprising, given that litter nutrient concentration is generally among the best predictors for litter turnover rates across various ecosystem types (Heal, Anderson & Swift 1997; Cornwell et al. 2008). Litter P, in particular, was assumed to be a limiting factor for decomposers and therefore contributing to the control of decomposition. In Costa Rican tropical forests, microbial activity responded positively to experimental P additions (Cleveland, Townsend & Schmidt 2002; Cleveland, Reed & Townsend 2006) and the abundance of litter fauna decreased significantly with decreasing soil P (McGlynn et al. 2007). In a P-limited Hawaiian montane forest, increasing litter P concentrations and increasing soil N and P availability, were both followed by increasing decomposition rates (Hobbie & Vitousek 2000). Experimental P-addition also increased litter decomposition in a Panamanian tropical forest, which however, was also positively influenced by the addition of other nutrients (Kaspari et al. 2008). Our study cannot immediately be compared to the fertilization experiments cited above where mineral P was externally supplied, increasing overall P availability to decomposers and plants much more than variation in litter internal P. The extent of ecosystem P limitation of the tropical forest of Guiana is at present not clear, and the role of soil P limitation for decomposition, in particular, remains to be shown.

The high predictive power of variable C compounds for decomposition in this nutrient-poor tropical forest was rather unexpected. In particular, increasing concentrations of labile and energy-rich compounds such as NSC that are not commonly measured in plant litter, and phenolics correlated well with increasing litter mass loss. On the other hand, CT as a distinct functional group of polyphenols, but often confounded with phenolics in the ecological literature (Coq et al. 2010), and lignin correlated negatively with litter mass loss. The negative CT effects on decomposition have been reported before for single litter species from a larger pool of 16 different tree species of the same rain forest (Coq et al. 2010). These CT effects were mostly related to litter-feeding animals as the main driver of interspecific variation in decomposition rates (Coq et al. 2010), which is supported by the strong negative relationship between litter mass loss and CT concentration in the presence of fauna in our study. Using the same functional dissimilarity approach as for litter stoichiometry, we additionally explored the impact of diversity in broad litter C fractions quantified here. None of the different combinations of ‘C traits’ used to create the dissimilarity gradients showed a positive correlation with litter mass loss, suggesting that the absolute amount of specific compounds in the different litter treatments, rather than their diversity, drives detritivores and decomposers.

Collectively, our data suggest that decomposer communities of the studied rain forest are primarily limited by the availability of easily accessible energy-rich C compounds in the litter they consume rather than its nutrient content. Although our study provides only correlative evidence for decomposer energy limitation, it is in line with recent reports of stimulated microbial degradation of recalcitrant old soil C compounds of low energy content by the addition of energy-rich fresh C (Fontaine et al. 2007; Hagedorn et al. 2008). Moreover, carbon substrate identity was found to strongly affect the decomposition of cellulose paper and to influence soil C content in a laboratory microcosm experiment (Orwin, Wardle & Greenfield 2006), suggesting that the identity of plant derived C compounds determine microbial community structure and microbial activity to a large extent.

Conclusions

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgements
  9. References

Our data suggest that decomposer communities in the studied tropical forest are not primarily limited by litter nutrients. Stoichiometric dissimilarity of litter mixtures may thus be of secondary importance and may come into play only if the basic energetic requirements of decomposers are covered. This might more readily be the case for detritivores with a larger activity range and better access to variable food sources in time and space, possibly explaining the increasing mass loss with increasing stoichiometric dissimilarity when fauna had access. Our initial hypothesis, however, needs further tests with a larger range of litter species and external resource supplies in order to assess the degree of stoichiometric control of the decomposer system in this Amazonian rain forest. If decomposer activity in the studied tropical forest would indeed mainly be controlled by energy, and thus by the quality of the carbon of their resources, as suggested by our data, tropical trees might strongly influence decomposer communities by the production of litter of specific carbon quality with potentially important feedback effects on the dynamics and availability of nutrients to plants and decomposers.

Acknowledgements

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgements
  9. References

We thank Damien Bonal and his research group at ECOFOG in Kourou for logistical support, Sandra Barantal for stimulating discussions on functional diversity of litter, Tanya Handa for helpful comments and discussions on previous versions of the manuscript, Laëtitia Brechet for her help in litter collections, Beat Aeschlimann for helping in the setup of the experiment, Bruno Buatois and Laurette Sonié for laboratory support, Anne-Marie and Vincent Domenach for their hospitality during several stays in Kourou, and the anonymous referees for their thoughtful and constructive suggestions in improving the manuscript. This research was funded through an ATIP research grant from CNRS (SDV) to S.H.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgements
  9. References
  • Ball, B.A., Hunter, M.D., Kominoski, J.S., Swan, C.M. & Bradford, M.A. (2008) Consequences of non-random species loss for decomposition dynamics: experimental evidence for additive and non-additive effects. Journal of Ecology, 96, 303313.
  • Bonal, D., Bosc, A., Ponton, S., Goret, J.Y., Burban, B., Gross, P., Bonnefond, J.M., Elbers, J., Longdoz, B., Epron, D., Guehl, J.M. & Granier, A. (2008) Impact of severe dry season on net ecosystem exchange in the neotropical rainforest of French Guiana. Global Change Biology, 14, 19171933.
  • Botta-Dukát, Z. (2005) Rao’s quadratic entropy as a measure of functional diversity based on multiple traits. Journal of Vegetation Science, 16, 533540.
  • Cherif, M. & Loreau, M. (2007) Stoichiometric constraints on resource use, competitive interactions, and elemental cycling in microbial decomposers. The American Naturalist, 169, 709724.
  • Cleveland, C.C. & Liptzin, D. (2007) C:N:P stoichiometry in soil: is there a ‘Redfield ratio’ for the microbial biomass? Biogeochemistry, 85, 235252.
  • Cleveland, C.C., Reed, S.C. & Townsend, A.R. (2006) Nutrient regulation of organic matter decomposition in a tropical rain forest. Ecology, 87, 492503.
  • Cleveland, C.C., Townsend, A.R. & Schmidt, S.K. (2002) Phosphorus limitation of microbial processes in moist tropical forests: evidence from short-term laboratory incubations and field studies. Ecosystems, 5, 680691.
  • Coq, S., Souquet, J.-M., Meudec, E., Cheynier, V. & Hättenschwiler, S. (2010) Interspecific variation in leaf litter tannins drives decomposition in a tropical rainforest of French Guiana. Ecology, 91, 20802091.
  • Cornwell, W.K., Cornelissen, J.H.C., Amatangelo, K., Dorrepaal, E., Eviner, V.T., Godoy, O. et al. (2008) Plant traits are the predominant control on litter decomposition rates within biomes worldwide. Ecology Letters, 11, 10651071.
  • Epps, K.Y., Comerford, N.B., Reeves III, J.B., Cropper Jr, W.P. & Araujo, Q.R. (2007) Chemical diversity – highlighting a species richness and ecosystem function disconnect. Oikos, 116, 18311840.
  • FAO (1998) World Reference Base for Soil Resources. World Soil Resources Report 84. FAO/ISRIC/ISSS, Rome, 91 pp.
  • Finzi, A.C. & Canham, C.D. (1998) Non-additive effects of litter mixtures on net N mineralization in a southern New England forest. Forest Ecology and Management, 105, 129136.
  • Fontaine, S., Barot, S., Barré, P., Bdioui, N., Mary, B. & Rumpel, C. (2007) Stability of organic carbon in deep soil layers controlled by fresh carbon supply. Nature, 450, 277280.
  • Gartner, T.B. & Cardon, Z.G. (2004) Decomposition dynamics in mixed-species leaf litter. Oikos, 104, 230246.
  • Grimshaw, H.M., Allen, S.E. & Parkinson, J.A. (1989) Nutrient elements. Chemical Analysis of Ecological Material (ed S.E.Allen), pp. 81151. Blackwell Scientific, Oxford, UK.
  • Güsewell, S. & Gessner, M.O. (2009) N:P ratios influence litter decomposition and colonization by fungi and bacteria in microcosms. Functional Ecology, 23, 211219.
  • Hagedorn, F., Van Hees, P.A.W., Handa, I.T. & Hättenschwiler, S. (2008) Elevated atmospheric CO2 fuels leaching of old dissolved organic matter at the alpine treeline. Global Biogeochemical Cycles, 22, GB2004, doi: 10.1029/2007GB003026.
  • Hättenschwiler, S., Tiunov, A.V. & Scheu, S. (2005) Biodiversity and litter decomposition in terrestrial ecosystems. Annual Reviews of Ecology, Evolution and Systematics, 36, 191218.
  • Hättenschwiler, S., Aeschlimann, B., Coûteaux, M.-M., Roy, J. & Bonal, D. (2008) High variation in foliage and leaf litter chemistry among 45 tree species of a neotropical rainforest community. New Phytologist, 179, 165175.
  • Heal, O.W., Anderson, J.M. & Swift, M.J. (1997) Plant litter quality and decomposition, an historical overview. Driven by Nature: Plant Litter Quality and Decomposition (eds G.Cadisch & K.E.Giller), pp. 329. CAB International, Wallingford, UK.
  • Heemsbergen, D.A., Berg, M.P., Loreau, M., Van Hal, J.R., Faber, J.H. & Verhoef, H.A. (2004) Biodiversity effects on soil processes explained by interspecific functional dissimilarity. Science, 306, 10191020.
  • Hobbie, S.E. & Vitousek, P.M. (2000) Nutrient limitation of decomposition in Hawaiian forests. Ecology, 81, 18671877.
  • Hoorens, B., Coomes, D. & Aerts, R. (2010) Neighbour identity hardly affects litter-mixture effects on decomposition rates in New Zealand forest species. Oecologia, 162, 479489.
  • Irion, G. (1978) Soil infertility in the Amazonian rain forest. Naturwissenschaften, 65, 515519.
  • Kaspari, M., Garcia, M.N., Harms, K.E., Santana, M., Wright, S.J. & Yavitt, J.B. (2008) Multiple nutrients limit litterfall and decomposition in a tropical forests. Ecology Letters, 11, 3543.
  • Kaye, J.P. & Hart, S.C. (1997) Competition for nitrogen between plants and soil microorganisms. Trends in Ecology and Evolution, 12, 139143.
  • Madritch, M.D. & Cardinale, B.J. (2007) Impacts of tree species diversity on litter decomposition in northern temperate forests of Wisconsin, USA: a multi-site experiment along a latitudinal gradient. Plant and Soil, 292, 147159.
  • Marigo, G. (1973) Sur une méthode de fractionnement et d’estimation des composées phénoliques ches le végétaux. Analysis, 2, 106110.
  • Martinson, H.M., Schneider, K., Gilbert, J., Hines, J.E., Hamback, P.A. & Fagan, W.F. (2008) Detritivory: stoichiometry of a neglected trophic level. Ecological Research, 23, 487491.
  • McGlynn, T.P., Salinas, D.J., Dunn, R.R., Wood, T.E., Lawrence, D. & Clark, D.A. (2007) Phosphorus limits tropical rain forest litter fauna. Biotropica, 39, 5053.
  • McGroddy, M.E., Daufresne, T. & Hedin, L.O. (2004) Scaling of C:N:P stoichiometry in forests worldwide, implications of terrestrial Redfield-type ratios. Ecology, 85, 23902401.
  • Meier, C.L. & Bowman, W.D. (2008) Links between plant litter chemistry, species diversity, and below-ground ecosystem function. Proceedings of the National Academy of Sciences of the United States, 105, 1978019785.
  • Orwin, K.H., Wardle, D.A. & Greenfield, L.G. (2006) Ecological consequences of carbon substrate identity and diversity in a laboratory study. Ecology, 87, 580593.
  • Pérez-Harguindeguy, N., Blundo, C.M., Gurvich, D.E., Díaz, S. & Cuevas, E. (2008) More than the sum of its parts? Asessing litter heterogeneity effects on the decomposition of litter mixtures through leaf chemistry. Plant and Soil, 303, 151159.
  • Reich, P.B. & Oleksyn, J. (2004) Global patterns of plant leaf N and P in relation to temperature and latitude. Proceedings of the National Academy of Sciences of the United States of America, 101, 1100111006.
  • Roy, J., Hättenschwiler, S. & Domenach, A.M. (2005) Tree diversity and soil biology: a new research program in French Guiana. Tree Species Effects on Soils: Implications for Global Change (eds D.Binkley & O.Menyailo), pp. 337348. Kluwer Academic, Dordrecht, The Netherlands.
  • Scherer-Lorenzen, M., Bonilla, J.L. & Potvin, C. (2007) Tree species richness affects litter production and decomposition rates in a tropical biodiversity experiment. Oikos, 116, 21082124.
  • Van Soest, P.J. & Wine, R.H. (1967) Use of detergents in the analysis of fibrous feeds. IV. Determination of plant cell-wall constituents. Journal Association Official Analytical Chemists, 50, 5055.
  • Sterner, R.W. & Elser, J.J. (2002) Ecological Stoichiometry: The Biology of Elements From Molecules to the Biosphere. Princeton University Press, Princeton, NJ, USA.
  • Ter Steege, H., Sabatier, D. & Castellanos, H. (2000) An analysis of the floristic composition and diversity of Amazonian forests including those of the Guiana shield. Journal of Tropical Ecology, 16, 801828.
  • Townsend, A.R., Cleveland, C.C., Asner, G.P. & Bustamante, M.C. (2007) Controls over foliar N:P ratios in tropical rain forests. Ecology, 88, 107118.
  • Uehara, G. & Gillman, G. (1981) The Mineralogy Chemistry, and Physics of Tropical Soils With Variable Charge Clays. Westview Press, Boulder, CO, USA.
  • Vitousek, P.M. (1998) Foliar and litter nutrients, nutrient resorption, and decomposition in Hawaiian Metrosideros polymorpha. Ecosystems, 1, 401407.
  • Vitousek, P.M. & Sanford, R.L. (1986) Nutrient cycling in moist tropical forest. Annual Reviews of Ecology and Systematics, 17, 137167.
  • Walker, B., Kinzig, A. & Langridge, J. (1999) Plant attribute diversity, resilience, and ecosystem function: the nature and significance of dominant and minor species. Ecosystems, 2, 95113.
  • Wardle, D.A., Bonner, K.I. & Nicholson, K.S. (1997) Biodiversity and plant litter: experimental evidence which does not support the view that enhanced species richness improves ecosystem function. Oikos, 79, 247258.
  • Wardle, D.A., Nilsson, M.-C., Zackrisson, O. & Gallet, C. (2003) Determinants of litter mixing effects in a Swedish boreal forest. Soil Biology and Biochemistry, 35, 827835.
  • Wardle, D.A., Yeates, G.W., Barker, G.M. & Bonner, K.I. (2006) The influence of plant litter diversity on decomposer abundance and diversity. Soil Biology and Biochemistry, 38, 10521062.
  • Wright, I.J. & Westoby, M. (2003) Nutrient concentration, resorption and lifespan: leaf traits of Australian sclerophyll species. Functional Ecology, 17, 1019.