Plasticity index to organic phosphorus (GLY)

About this dataset

Methods Species, phosphorus sources and mineral. Eight (sub)species were selected for their contrasting morphologies, phosphorus (P)-acquisition strategies and traits, and their relevance to the local context: Poaceae (*Hordeum vulgare* L., *Lolium perenne* L.), Fabaceae (*Lens culinaris* Medik., *Pisum sativum* subsp. *arvense* L., *Trifolium alexandrinum* L., *Vicia faba* L., *Lupinus albus* L.) and Hydrophyllaceae (*Phacelia tanacetifolia* Benth.). The organic P source (OP) was glycerophosphate (GLY; glycerol 2-phosphate), chosen for its ubiquity in organic inputs and soils and for its small molecular size (a 3-carbon chain carrying one phosphate group). Inorganic P (IP) was supplied as KH₂PO₄. Goethite \[FeO(OH)], representing iron oxyhydroxides, was the mineral sorbent (powder from Sigma-Aldrich and VWR, Paris, France). Preparation of goethite–P complexes. Complexes were prepared following Amadou et al. (2022a, b). Goethite (100 g) was weighed into 1 L bottles with 200 mL of 0.1 M KCl and shaken for 24 h at 25 °C to hydrate the surface sites, keeping pH at 5.5 ± 0.05 with 0.1 M HCl or NaOH. Solutions of 1 g P L⁻¹ (GLY or IP) in 0.1 M KCl at pH 5.5 were prepared, and 800 mL of each was added to the bottles. A few drops of hexanol suppressed microbial activity. Bottles were shaken for 24 h at 25 °C and centrifuged (3,000 g, 15 min). The contents were distributed into 250 mL bottles and centrifuged again, and supernatants were removed. The P-loaded goethite was resuspended in deionized water and frozen at −20 °C. Sorbed P was calculated as the initial P minus the P left in the supernatant (filtered at 0.22 µm). P was determined colorimetrically (Ohno & Zibilske 1991) after persulfate oxidation to convert OP to IP (Peters & Van Slyke 1932). Total P in a subsample of each loaded mineral (Ostrofsky 2012) differed from the supernatant-based value by less than 5.5%. Goethite was loaded to its sorption maximum; loadings are in Supplementary Table S1. Experimental setup and plant growth. Seeds were germinated beforehand under controlled conditions (8 days for *H. vulgare*, *L. culinaris* and *L. perenne*; 10 days for *P. sativum*; 12 days for *L. albus*; 14 days for *V. faba*, *P. tanacetifolia* and *T. alexandrinum*). Mesocosms were 120 × 120 mm square Petri dishes filled with about 100 g of acid-washed coarse quartz sand mixed with the P-loaded goethite. The amount of complex was calculated to supply 40 mg P per device in both treatments, enough to avoid P limitation (Gómez-Suárez et al. 2020; Amadou et al. 2022b). Substrates were autoclaved (121 °C, 2 h) to eliminate micro-organisms, so that results can be attributed to root activity. One seedling was sown per dish. Plants were watered four times a week with a P-free solution: Ca(NO₃)₂ 2 mmol L⁻¹, MgSO₄ 0.5 mmol L⁻¹, K₂SO₄ 0.7 mmol L⁻¹, KCl 0.1 mmol L⁻¹, MnSO₄, CuSO₄ and ZnSO₄ 0.5 µmol L⁻¹ each, (NH₄)₆Mo₇O₂₄ 0.01 µmol L⁻¹ and Fe-EDTA 100 µmol L⁻¹. Devices stood in a phytotron (12 h photoperiod, 25 °C, 80% relative humidity). Unplanted controls (bulk substrate) were incubated in identical devices. The design comprised 8 species × 2 P forms (GLY, IP) × 2 conditions (planted, unplanted) × 5 replicates = 160 devices. Plants were harvested 30 days after germination. Root trait measurements. At harvest, plants were separated manually from the substrate with minimal damage. Rhizosheath substrate adhering to roots (up to 2 mm after shaking) was collected and stored at −20 °C. Roots were immersed for 1 h in 0.20 mM CaCl₂ (after a quick CaCl₂ rinse), at a volume ensuring complete immersion. The solution was sampled and stored at −20 °C. Carboxylates were separated by reversed-phase liquid chromatography on a C18 column (mobile phase 93% 25 mM KH₂PO₄ at pH 2.5 and 7% methanol; 1 mL min⁻¹; 15 min; Cawthray 2003; Yacoumas et al. 2020). Total carboxylate release rate was the sum of citric, fumaric, maleic, malic and malonic acids (Supplementary Table S2). **Morphological t