Fig. 3.

Lognormal distributions of nuclear volumes in uninoculated and mycorrhizal roots of wild-type M. truncatula. No statistically significant difference was observed between the distribution curves of nuclear volumes estimated with TrackMate plugin (a) in uninoculated (dark grey) and mycorrhizal (light grey) root segments. This result is indicative of the limited efficacy of this image analysis method. By contrast, the distribution curves, obtained through a bootstrap resampling method of volumes detected by 3D-Object Counter (b), revealed significantly higher values of skewness (P < 0.01), but not Kurtosis, in mycorrhizal compared to uninoculated roots. This indicates that 3D-Object Counter can be considered a more reliable method for this type of image analyses. Pairwise comparisons were performed with Tukey’s post-hoc test