Image analysis
To image and quantify the growth of the extraradical fungal network, we
covered the central compartment of a random subset of plates (A5:n =12, B12: n =12, Agg: n =17) with a sterile
cellophane sheet to facilitate 2D imaging (Fig. 1c). We monitored plates
for fungal growth in the focal and partner compartment and checked
weekly for fungal cross-over into the central compartment. After
approximately 20 days, the first hyphae crossed the plastic barrier to
the central compartment. We then imaged the entire fungal network in the
central compartment using a 5x objective on a Leica Wild M8 preparation
microscope, taking images with an Olympus SC180 camera.
To obtain representative images of each of the fungal strains, we
selected three spatial locations with a dimension of 5x5
mm2 (640x640 px2) across the central
fungus-only compartment in each of the treatments. The locations ranged
across the space connecting the partner compartment barrier to the
center of the central compartment (Fig. 1b). Using MATLAB, we applied
morphological operations to the images, binarized the images, removed
isolated cluster (background noise) and extracted the network skeleton
of the extraradical fungi. We calculated the mass fractal dimension
(Dm ) of every spatial area using the box-counting
technique (Hitchcocket al. 1996; Falconer 2003; Boddy & Donnelly. 2008; Boudaet al. 2016), with a square grid size ranging from 8 to 64
pixels, i.e. from 1/10 to1/80 times the total square area. We
then estimated the fractal dimension by:
Where s corresponds to the grid size and N(s) the total
number of boxes that contain fungal hyphae. We calculated the density of
the network (surface percentage) as the ratio between the surface
occupied by the network and the total square area.