Plant growth and sampling

The experiment was conducted in a greenhouse between September 2018 and April 2019, at Campinas (22°49′10.38′′S 47°04′12.88′′W), São Paulo, Brazil. The experimental design was a 5 × 2 factorial, i.e., five eucalypt species and two soil P levels, with five replicates of each species for each soil P level. The species were: Corymbia maculata, Eucalyptus acmenoides, E. globulus, E. grandis and E. tereticornis. The soil used here was a clay ferrosol (IUSS Working Group WRB, 2015), and P is referred to as the soil resin extractable-P. The soil analysis is shown in Supplementary Table S1. Before mixing 1:1 (v/v) with water-washed sand, the soil was corrected for nutrients as recommended for forest species at a medium-fertility level in São Paulo, Brazil (Raij et al., 1997). The soil P level was not correct in one treatment (4.1 mg dm^–3 available P) and corrected (12.9 mg dm^–3 available P) in another, and they are referred to from now on as low P and sufficient P, respectively. We amended the low P soil with potassium phosphate (KH₂PO₄) to reach the P level in sufficient soil.

Seeds of the five species were germinated in seedbeds with a commercial substrate mixture (Genesolo, Genfertil), and after 2 months the seedlings were transplanted to 7 L pots containing the mixture of soil and sand described previously. They were kept in a greenhouse for a period of 9 months. Irrigation was carried once a day. During the experiment, the seedlings received five applications of N (125 mg each application), using NH₄NO₃ or KNO₃.

Leaf, stem, and root samples of 8-month-old plants were harvested separately and immediately frozen in liquid nitrogen and stored at −80°C until further analysis. The roots were collected after washing out the substrate in running tap water.