Complementary N Uptake Strategies between Tree Species in Tropical Rainforest

Within tree communities, the differential use of soil N mineral resources, a key factor in ecosystem functioning, may reflect functional complementarity, a major mechanism that could explain species coexistence in tropical rainforests. Eperua falcata and Dicorynia guianensis, two abundant species cooccurring in rainforests of French Guiana, were chosen as representative of two functional groups with complementary N uptake strategies (contrasting leaf δ 15N signatures related to the δ 15N of their soil N source, NO3 − or NH4 +). The objectives were to investigate if these strategies occurred under contrasted soil N resources in sites with distinct geological substrates representative of the coastal rainforests. Results showed that species displayed contrasting leaf δ 15N signatures on both substrates, confirming their complementary N uptake strategy. Consequently, their leaf 15N can be used to trace the presence of inorganic N-forms in soils (NH4 + and NO3 −) and thus to indicate the capacity of soils to provide each of these two N sources to the plant community.


Introduction
How species diversity is maintained in tropical rainforests is a critical issue in contemporary ecology. According to the niche complementarity hypothesis, coexisting plants are thought to complement each other in traits related to resource foraging. Consequently, resource partitioning and functional complementarity could be important in maintaining species diversity in plant communities. Mineral nitrogen is a key factor in ecosystem functioning. Within tree communities, the differential use of soil N mineral resources (nitrate, ammonium) may, among other mechanisms, reflect tree functional complementarity [1][2][3]. The spatial variability of soil nitrateto-ammonium ratio may partly result from the spatially heterogeneous occurrence of two of the major soil microbial processes, nitrification and denitrification [4]. A number of studies have linked the N uptake strategies of tropical trees to their successional status (see [5] for review). In contrast, Houlton et al. [6] and Russo et al. [7] showed that coexistence among functionally diverse species was not linked to the particular form of N available. In French Guiana, Roggy et al. [2] reported that among non-N 2 -fixing trees two groups of late-successional species with contrasting natural leaf 15 N abundance coexisted. This was interpreted as different abilities of species to utilize different soil mineral N pools. Based on these results, we carried out a preliminary study in French Guiana with two abundant cooccurring late-successional species, Eperua falcata and Dicorynia guianensis [8]. These species were chosen as representative of the two groups of non-N 2 -fixing species with contrasting leaf 15 N signatures. We demonstrated that both species exhibited complementary N uptake strategies and that their leaf 15 N were related to the 15 N signatures of their soil N source (NO 3 − or NH 4 + ). In this study, the objectives were to investigate if complementary N uptake strategies of these two tree species occurred under contrasted soil N resources in sites with distinct geological 2 International Scholarly Research Notices substrates representative of the coastal rainforest in French Guiana. Specifically, we tested the occurrence of different leaf 15 N signatures associated with the two species on volcanicplutonic rocks of the Paramaca series (high total soil nitrogen concentration) and schists of the Orapu series (low total soil nitrogen concentration). We also estimated variations in nitrification and denitrification potentials in soils and litters under the two species.

Study Sites.
The study was conducted in Crique Plomb, an experimental site of natural forest in the northern part of French Guiana (5 ∘ 0 -5 ∘ 2 N, 52 ∘ 56 -52 ∘ 57 W). This site is characterized by a clear segregation, by a small river (Crique Plomb) running along a geological boundary, between two geomorphologic domains representative of the contrast between metamorphosed sediments and a volcanic formation ( Figure 1). The northern part lies in a hilly area where the geologic substrate is pelitic schist of the Precambrian "flyschoid series" (Orapu schist of Armina series [9,10]). Physical and chemical erosion of a preexisting deep oxisol have resulted in a deep oxisol with a deep vertical drainage. The southern part is a single topographic unit where the geologic substrates are of volcanic-plutonic origins (Paramaca series). The soil cover units consist in numerous zones of ferricrete interspersed with deep nonhydromorphic soils.
Six plots of one hectare were selected in the two main geomorphologic domains, that is, three replicates per domain: plots CP3, CP2, and CP4 in the schist domain and plots CP1, CP7, and CP6 in the volcanic-plutonic domain. On each domain, plots were separated from at least 1000 m in order to integrate potential spatial variations in nutrient uptake by trees ( Figure 1).
Although the two soil types had similar pH levels, chemical analyses showed significant differences ( Table 1). The soils of the volcanic-plutonic domain had higher levels of nitrogen (0.83% versus 0.51%) and carbon (13.58% versus 8.61%) but identical C-to-N ratios (≈16).
Three angiosperm families represent around 60% of individuals in the Crique Plomb area: the Lecythidaceae, Caesalpinioideae, and Euphorbiaceae. More than 700 tree species with stems over 10 cm DBH (diameter at breast height ∼ 1.30 m) belonging to 270 genera and 68 families were recorded for 12,000 inventoried individuals. There are around 150 tree species per hectare in the plots and rare species (species with few individuals) dominate.
The climate is characterized by a clear seasonal pattern: a wet season from December to July, which is normally interrupted in February or March by a short dry period, and a long dry season from August to November with monthly precipitation of less than 100 mm. Average annual precipitation is around 3000 mm. Mean temperature is 25 ∘ C with low seasonal changes [11]. All measurements were taken at the end of the rainy season.

Tree Species, Soils, and Litter
Sampling. Dicorynia guianensis Amsh. (tribe Cassieae) and Eperua falcata Aublet (tribe Detarieae) are two cooccurring, shade hemitolerant Caesalpinioideae species belonging to the late-successional species group [12]. Bothare known as non-N 2 -fixing species with high and low leaf 15 N values, respectively [1,2]. Both exhibit a Paris-type mycorrhizal association [13,14]. In each of the six plots, at least five trees per species (representing up to 25% to 100% of the individuals per species per plot) with a DBH of 10 cm or more were randomly selected by pairs of the two species. Leaves from the upper canopy were collected using a shotgun. Only young, fully expanded leaves were selected for analysis since 15 N might be sensitive to leaf age [15]. Three samples of litter and soil were collected underneath the canopy of each individual tree. Leaves were analyzed for 15 N and total nitrogen. Soil and litter samples were analyzed for their capacities (i.e., potential rates) to express nitrification and denitrification (both processes being directly involved in the balance between NH 4 + and NO 3 − ). Soils were analyzed for NO 3 − . We will focus on potential activities rather than on actual activities because they are less fluctuating and less sensitive to short spatiotemporal variations, which allow between sites comparisons. Moreover, potential activities are not under the control of limitation or inhibition factors of the environment and represent the maximal capacity of the enzyme pool in soil to realize a biotransformation [16].

15 N Analyses.
The leaf laminae (i.e., without petioles or rachis) were oven-dried at 60 ∘ C for 48 h. Leaf laminae were then milled to a fine powder and packed in airtight containers. Nitrogen isotope ratios ( 15 N/ 14 N) and total nitrogen concentrations were measured as described by Casabianca [17], using an elemental analyzer (SCA, CNRS, Solaize, France) coupled with a mass spectrometer (Finnigan Mat, DS, Bremen, Germany). The precision of measurement was 0.2‰ (standard error of the mean, = 1000). Results are reported as percent N (leaf N (%)) and 15 where the standard is atmospheric N 2 (atom% 15 N = 0.3663, [18]).

Denitrifying Enzymatic Activity (DEA) in Soil and Litter.
The enzymatic potential of denitrification was measured according to Lensi et al. [19]. For each sample, 10 g of sieved and air-dried soil or 2 g of air-dried litter was placed in a 150 mL plasma-flask and sealed with rubber stoppers. In each flask, air was removed with a vacuum-pump and replaced with a He-C 2 H 2 mixture (90/10, vol/vol) to inhibit the N 2 Oreductase. A solution containing 100 g of N-KNO 3 , 1 mg of C-glucose, and 1 mg of C-glutamic acid/g of sample (soil or litter) was added to ensure saturation. The flasks were incubated at 28 ∘ C for 5 h and gas samples of 200 L were analyzed for N 2 O on a gas chromatograph with an electron capture detector (VARIAN 3800-CP, Les Ulis, France). The reason for these short incubations was to avoid de novo enzymatic synthesis and cell growth. DEA was expressed in ng of N- International Scholarly Research Notices

Nitrate in Soil and Nitrifying Enzymatic Activity (NEA) in
Soil and Litter. The enzymatic potential of nitrification was measured according to Lensi et al. [20]. Each sample was divided into 20 g (soil) or 2 g (litter) subsamples and placed in 150 mL plasma flasks. The first subsamples were used to estimate the initial NO   , 1984-2007). Data were ranked to avoid the assumptions of normality [21], and the differences between the means were analysed with the Mann-Whitney -test ( < 0.05) after a one-way ANOVA of Kruskal-Wallis.

Results and Discussion
Leaf samples of D. guianensis and E. falcata had consistently different 15 N isotopic signatures (average difference between the two species around 3.5‰, Figure 2), with lower values on schist than on the volcanic-plutonic domain. A similar magnitude of difference between these values was observed on both substrates. This variability of the 15 N between species may be due to spatial variability of soil 15 N, which essentially reflect the isotopic signatures of organic pools [22]. In French Guiana, the 15 N of the organic matter in soils and litters vary among forests on different substrates but remain spatially uniform for a given type of soil [12,23] and in litters and soils collected underneath E. falcata and D. guianensis [8]. These studies showed that the inorganic N assimilated by these trees came from an organic N source with a homogeneous isotopic composition. A second source of variability could come from different mycorrhizal status of trees, mycorrhized trees having access to organic soil N with a particular 15 N abundance [22]. In French Guiana, both species exhibit a Paris-type mycorrhizal association [14], which exclude differentiation of leaf 15 N due to different mycorrhizal status. A third source of variability could lie in abilities of roots to use various pools of soil inorganic N with different 15 N abundances. Thus, the foliar 15 N observed in the two species would be linked to the 15 N of the inorganic N sources they used [24], knowing that uptake by and translocation of N within the plants are not considered to cause appreciable isotopic fractionation [7]. In this study, the differences of 15 N observed in leaves of E. falcata and D. guianensis showed clearly that both species used different soil inorganic N sources with different 15 N signatures on both substrates. Differences of 15 N between the two soil N sources are due to fractionation of 15 N during enzyme-driven transformation process (e.g., nitrification [22]). Soil microbial activities were spatially partitioned between the soil (denitrification) and litter (nitrification, Table 2), confirming the results found by Schimann et al. [8] on a different type of substrate. Despite different total soil C and N concentrations in the two substrates (Table 1), soil nitrate contents and microbial activities were similar ( Table 2). Previous studies (Roggy et al. pers com) have showed that N 2 fixation by trees is twice higher on schist than on volcanic-plutonic soil (6.5 versus 3.5 kg N/ha, resp.). Additional inputs of labile organic matter by litters of N 2 -fixing trees are known to enhance microbial growth and activities [25,26] and particularly nitrification rates [4,27]. Thus, nitrification is expressed at a similar level on both substrates. Nevertheless, these processes are not sufficient to allow accumulating organic matter in soils as shown by the lower contents of C and N in schist soil than in volcanicplutonic soil. As indicated, the two species displayed different isotopic signatures whatever the geological substrates but the 15 N values were lower in the schist soils ( Figure 2). In these soils, there is probably a global decrease of 15 N in litters and soils due to the higher proportion of N 2 -fixing species in these forests which supply litters with low 15 N values [1,23,28].
On this type of substrate, NH 4 + and NO 3 − are supposed to be produced from a more 15 N-depleted organic matter as compared to the organic matter of the volcanic-plutonic soil.

Conclusions
In conclusion, our results together with those of Schimann et al. [8] showed   the presence of inorganic N-forms in soils (NH 4 + and NO 3 − ) and thus to indicate the capacity of soils to provide each of these two N sources to the plant community.