Affiliation: Department of Biology, University of Oulu

Article title: A shiftfrom arbuscular mycorrhizal to dark septate endophytic colonization in Deschampsiaflexuosa roots occurs along primary successional gradient

Journal name: Mycorrhiza

Author names: Karoliina Huusko, Ruotsalainen Anna Liisa, Markkola Annamari

Affiliation: Department of Biology, University of Oulu

E-mail address of the corresponding author:

Soil nutrient and pH analyses

Five (mineral) or ten (organic and mineral) soil samples were collected from each experimental site in August 2016. One soil sample consist of eight cores (3 cm in diameter) of soil. If organic soil layer existed, soil cores were divided to organic and mineral soils and put to own plastic bags. Thickness of organic soil layer was measured. All organic soil but only five uppermost centimeters of mineral soil were taken from each core. Organic soil layer was not present in young successional stages, totaling sample numbers to 30 for organic and 45 for mineral soil. Samples were stored in +4°C until preparation, for maximum one week.

Soluble P (phosphate) was analyzed from ammonium acetate extracts (10 ml of fresh soil and 50 ml of ammonium acetate) using a colorimetric method. pH and electrical conductivity were measured from water extracts (22ml of soil and 50 ml of Milli-Q water).

For C and Nanalyses, samples were dried for 48h in +50°C. Biggest roots and rocks were removed manually and samples ground with a ball mill. Total C and total N concentrationswere determinedfrom samples of 0,090 g DW withISO 16948:2015procedure at Ahma Group laboratory in Oulu.

Organic matter (OM%) of soil was determined using loss of ignition (550°C for 3.5 hours in a muffle furnace).

Statistics

Impacts of successional stage on soil pH, electrical conductivity, OM%, phosphate, C and N were analyzed using linear mixed models (lme) (Pinheiro et al.2016). Mineral and organic soil data were analyzed separately. In mineral soil total C and N were under detection levels (<1 mg/g and <0.3 mg/g, respectively, data not shown). Successional stage (levels for mineral soil: young, intermediate, old; levels for organic soil; intermediate, old) was used as a fixed factor, and site as a random factor. Normality and homogeneity of residuals were checked using Q–Q and scatter plots. Arcsine-square root transformations were used to normalize distribution and homogenize variation of residuals for OM % and logarithmic transformation for electrical conductivity. If ANOVA F test results were statistically significant (p < 0.05), Tukey HSD tests were conducted on all possible pairwise contrasts with the “glht” command using the “multcomp” package in R (Hothorn et al. 2008)(Tables S1 and S2).

Table S1. Mean ± SE of pH, electrical conductivity (EC, µS cm-1), organic matter (OM %), and phosphate (µg g-1) concentrations in mineral soil atthree successional stages (I, II, III).

Letters indicate differences tested by Tukey tests; numbers followed by the same letter are not statistically significant at P = 0.05. F- and P-values in boldface are significant at P = 0.05

* logarithmic transformation ; ° arcsine-square root transformation

numDF: 2; denDF: 40

Table S2. Mean ± SE of thickness (cm), pH, electrical conductivity (EC, µS cm-1), organic matter (OM %), phosphate (µg g-1), nitrogen (% in OM) and carbon (% in OM) concentrations in organic soil atintermediate and old successional stages (II, III).

F- and P-values in boldface are significant at P = 0.05.

* logarithmic transformation ; ° arcsine-square root transformation

numDF: 1; denDF: 26