Tropical forests are often characterized by low soil phosphorus (P) availability, suggesting that P limits plant performance. However, how seedlings from different functional types respond to soil P availability is poorly known but important for understanding and modeling forest dynamics under changing environmental conditions.

We grew four nitrogen (N)-fixing Fabaceae and seven diverse non-N-fixing tropical dry forest tree species in a shade house under three P fertilization treatments, and evaluated carbon (C) allocation responses, P demand, P-use, investment in P acquisition traits, and correlations among P acquisition traits.

N-fixers grew larger with increasing P addition in contrast to non-N-fixers, which showed fewer responses in C allocation and P-use. Foliar P increased with P addition for both functional types, while P acquisition strategies did not vary among treatments but differed between functional types, with N-fixers showing higher root phosphatase activity (RPA) than non-fixers.

Growth responses suggest that N-fixers are limited by P, but non-fixers may be limited by other resources. However, regardless of limitation, P acquisition traits such as mycorrhizal colonization and RPA were non-plastic across a steep P gradient. Differential limitation among plant functional types has implications for forest succession and earth system models.

We initially selected seeds of 30 species that commonly occur across successional and soil fertility gradients (30 to 1272 total P mg kg^-1) in ACG and Área de Conservación Arenal Tempisque (ACAT)-Parque Nacional Palo Verde (10.35 N, 85.35 W) (Powers & Tiffin, 2010; Werden et al., 2018), and had seeds available at the time the experiment was established. Due to logistical constraints (i.e., low germination rates, pests, and high mortality), only 11 species of the original 30 were used in the experiment. We included four N-fixing species from the Papilionoideae subfamily, the most abundant Fabaceae subfamily in the area (Werden et al., 2018) and the one with the highest species richness (Koenen et al., 2021). Papilionoideae species primarily associate with arbuscular mycorrhizal fungi. We caution here that our results should not be extended beyond the Papilionoideae subfamily; however, the fixers and non-fixers used in this study do not differ significantly in distribution characteristics across successional or soil gradients compared to the regional species pool (Methods S1, Table S1, and Fig. S1). Non-fixers included one ectomycorrhizal and six arbuscular mycorrhizal species from seven different families, including one Fabaceae (Table S2 and Fig. S2). We collected most seeds in the ACG and ACAT and purchased seeds of Tabebuia impegitinosa (Mart. Ex DC.) Standl. from the Centro Agronómico Tropical de Investigación y Enseñanza (CATIE), Costa Rica.

The experiment was conducted during the rainy season from May to October in a 36 m² shade house that excluded approximately 55% of PAR (J. Powers, unpublished data). Prior to planting, all seeds were soaked in water overnight. Hymenaea courbaril L. seeds were scarified to accelerate germination (Smith-Martin et al., 2017); seeds were placed in 100°C water for 30 s and then in ice water for one minute, this process was repeated three times. Then, between 2-15 seeds (depending on seed size and availability) of each species were planted per pot (11 species x 3 treatments x 7 replicates = 231 pots). Pots were big enough to have a biomass volume ratio <1gL^-1 (Poorter et al., 2012) (10 x 10 cm and 20 cm deep containing ~2 kg of soil). Pots were randomly arranged within plastic crates to avoid confounding positioning species and treatment effects (8 pots per crate). One week after germination, seedlings were thinned to one per pot. Because we obtained low germination rates for Lonchocarpus rugosus Benth. And H. courbaril, additional seeds were added. To minimize differences in soil microbial communities, all seedlings were inoculated with microbial inoculum from three different forest soils collected across the regional P gradient (Smith-Martin et al., 2017; Waring et al., 2019). We prepared the inoculum by combining equal weights of the three soils in 1 L water, shaking them, and adding 10 mL of this solution to each pot 3 weeks after sowing. This treatment was repeated the following week to ensure microbial colonization. When seedlings were ~1 month old they were fertilized every two weeks from July to October with 10 mL of either 0.0016 g mL^-1 of phosphoric acid (H₃PO₄) (45 kg P ha^-1 yr^-1) (P++) (Waring et al., 2019), 0.00016 g mL^-1 H₃PO₄ (4.5 kg P ha^-1 yr^-1) (P+), or tap water (P0) to emulate the regional soil P gradient and P application rates used in other fertilization experiments (Siqueira & Saggin-Júnior, 2001; Alvarez-Clare et al., 2013; Waring et al., 2019). Soil moisture was measured in each pot to 5 cm depth three times over the course of the experiment with a SM150 Soil Moisture Sensor (Delta-T Devices, Cambridge, UK). When rainfall was scarce, i.e., lack of rain for >24 h, seedlings were hand-watered every other day to prevent desiccation. At the conclusion of the experiment, we measured soil pH in distilled water on air-dried soils from all pots using a 1:2.5 soil to solution ratio. Finally, to quantify the distribution of soil P among different P pools, a modified Hedley fractionation was conducted on a subset of soils from the pots at the soil laboratory at CATIE.

Harvesting and biomass

We harvested all seedlings four months after germination over a 10-day period. None of the individuals showed evidence of being pot bound. At the time of harvest, each seedling was separated into root, stem, and leaf fractions. Roots were carefully removed, gently washed in water and weighed. In the case of nodulating species, nodules were removed from the roots (before weighing), counted, and weighed. Leaves were immediately scanned to calculate leaf area using ImageJ software (Schneider et al., 2012). Leaves, stems, roots, and nodules of each seedling were dried at ~60°C for 5 days, and then weighed to calculate total dry biomass (TDB), leaf (LMF), stem (SMF), root (RMF) (Table S3), and nodule (NMF) mass fractions. Root biomass included the tap roots if they were present. Dry weights of leaves were used to calculate specific leaf area (SLA, cm² g^-1). All statistical analyses were performed using dry weights. SLA did not vary between functional types or treatments (Fig. S3, SLA p>0.05) and will not be further discussed.

Plant performance and nutrient limitation

We measured relative height growth rate (RGR), photosynthetic capacity, and water use efficiency as metrics of plant performance. Stem height was measured from the base of the stem to the apical meristem on each seedling every two weeks after the first true leaves emerged. We used the classic RGR (cmd^-1) equation: (ln[final height]–ln[initial height])/(final day–initial day) (Hunt & Cornelissen, 1997; Rees et al., 2010). Photosynthetic capacity was quantified by measuring photosynthetic responses to carbon dioxide concentration (AC_i curves). Gas exchange measurements were performed on one recently matured leaf of three seedlings per species per treatment in situ with a LI-6400XT portable photosynthesis system (LI-COR, Lincoln, NE, USA) at a constant temperature (30°C). Measurements were made at CO₂ concentrations ranging between 0 and 1200 ppm (9 points total), irradiance was 1000 μmol m^-2s^-1 (unlikely to cause photoinhibition of photosynthesis when leaves are measured below or above optimum temperatures) and relative humidity ranged between 60-80%. Measurements were made between 07:00 and 14:00 h and the ambient air temperature generally increased from 25°C to 32°C. We then obtained maximum photosynthetic rates (A_max) and rubisco maximum carboxylation rate (V_cmax) using the ‘plantecophys’ R package (Duursma et al., 2015).

Finally, we analyzed foliar nutrients, water use efficiency, and photosynthetic phosphorus-use efficiency (PPUE). We bulked all leaves per individual and transported them to the University of Minnesota (UMN) where they were ground and analyzed for total carbon (C), nitrogen (N) and C isotopic ratio (δ¹³C) on an Isoprime 100 isotope ratio mass spectrometer coupled with an Elementar Pyrocube combustion elemental analyzer (Elementar Americas Inc). Other foliar nutrients (aluminum, boron, calcium, chromium, copper, iron, phosphorus, potassium, magnesium, manganese, sodium, and nickel) were analyzed at the UMN Research Analytical Lab using standard methods in an iCAP™ 7600 ICP-OES (Inductively Coupled Argon Plasma Optical Emission Spectrometer) Analyzer (Thermo Fisher Scientific Inc. Waltham, MA, USA) following digestion in HNO₃. We calculated PPUE from A_max and foliar P concentrations (Veneklaas et al., 2012).

Belowground traits

            We quantified plant investment in P acquisition using root mass fraction (RMF) and specific root length (SRL)—a morphological response, root phosphatase enzyme activity (RPA)—a physiological response, and % arbuscular mycorrhizal colonization (AMC)—a biotic response representing degree of collaboration with symbionts. SRL was calculated after scanning all roots with a transparency scanner (Epson Perfection V800, Suwa, Japan; 1-4 images per individual). We placed the roots in a clear polycarbonate tray filled with water to ensure no root overlap and used the Rhizovision software (Seethepalli et al., 2020) to calculate root length of fine root fractions (≤2 mm diameter), i.e., excluding any tap root. After scanning, root systems were separated into two subsamples: one sample for RPA, and the other for scoring AMC. RPA was determined using para-nitrophenyl phosphate (pNPP) as an analogue substrate for phosphomonoesterase using standard methods (Methods S2; Turner et al., 2001). Finally, we scored arbuscular mycorrhizal colonization (AMC) by staining the roots with alanine blue (Koske & Gemma, 1989) and quantifying colonization microscopically using the magnified intersection method (McGonigle et al., 1990) on ~100 intersections per root fragment. We calculated colonization as the percentage of root length colonized by AM fungi (Methods S3). Ectomycorrhizal colonization was measured using the gridline intersection method (Brundrett et al., 1996). This variable was not included in any of the analyses as only one species was ectomycorrhizal, and these data are not directly comparable to AMC.

While quantifying N fixation activity was not a goal of this experiment, we assessed plant carbon construction investment in fixing structures by quantifying nodule mass and NMF (Methods S4, Table S4 and Fig S4).

Files can be accessed using excel or a csv reader software.