Plant and Soil_Electronic Supplementary material What nurse shrubs can do for barren soils: rapid productivity shifts associated with a 40 year ontogenetic gradient Navarro-Cano JA*, Verdú M, García C, Goberna M
José Antonio Navarro-Cano (*corresponding author). Centro de Investigaciones sobre Desertificación (CSIC-UVEG-GV), Carretera Moncada - Náquera, Km 4.5. E-46113, Moncada, Valencia, Spain; phone: +34 963424126; Fax: fax: +34 963424160; e-mail: jose.a.navarro@uv.es

(A)

(B)

(C)

Fig. S1. (A) Overview of the study system. Serra de Crevillent (Alacant, SE Spain). Gypsum outcrops are dominated by the shrub O. tridentata. (B) This species determines a patch-gap mosaic with gaps partially covered by sealing crusts and some gypsophytes and the patches mainlyattractingnon-gypsophyte species. (C) View of the community of facilitated species below the canopy of an adult O. tridentata.

Fig. S2. Stem cross section of a 25 year old Ononis tridentata individual. (A) The four transects analysed from the pith to the cortex are shown in a whole section. (B) Detail of the annual growth rings (arrows) shaped by the light earlywood − dark latewood sequence and one of the frequent perpendicular rays (R) for lateral conduction (40×).

Fig. S3. Micro-environment characterization of Ononis patches along the studied ontogenetic gradient. Soil surface temperature (A), light intensity (B) and gravimetric humidity (C) measurements from April 2013 are shown. Analytical methods are given in the main text.

Fig. S4. Regression model showing the sigmoidal fit of Ononis Age on the litter biomass. The regression equation, explained variance and significance of the F-test are shown. Data are given on an oven-dried weight mass.

Fig. S5. Soil surface micro-nutrient values along the studied ontogenetic gradient. Regression curves showing lesser residual sum of squares are showed. Data are given on an oven-dried weight mass. Analytical methods are given in the main text.

Fig. S6. Regression models showing the quadratic fits of: (A1) the barley seedling biomass (above- + belowground biomass) as a function of the increasing chemical fertility determined by the Ononisontogenetic gradient, (A2) the barley seedling biomass as a function of the estimated Ononis age below which the soil samples used in the fertility bioassay were collected, (B1) the barley root-shoot ratio (belowground·aboveground^{-1} biomass) as a function of the increasing chemical fertility determined by the Ononis ontogenetic gradient and (B2) the barley root-shoot ratio as a function of the estimated Ononis age. The regression equations, explained variances and significance of the F-tests are shown. Each point is the average ± SE of 36 seedlings after three weeks in a growth chamber. Data are given on an oven-dried weight basis.

Fig. S7. Soil microbial productivity parameters along the studied ontogenetic gradient. Regression curves showing lesser residual sum of squares are showed. Analytical methods are given in the main text.