1

Balogianni et al. p.

Table S1.Within-subjects Effects from Repeated Measures ANOVA for Roots and Shoots with Vegetation Type, Mowing and Nitrogen (N) as MainFactors and Their Interactions

Source of variation / NumDf / DenDf / F value / P value
Root / Time / 8 / 77 / 20.74 / 0.00*
Time*vegetation type / 8 / 77 / 1.43 / 0.20
Time*mowing / 8 / 77 / 0.81 / 0.59
Time*vegetation type*mowing / 8 / 77 / 0. 72 / 0.67
Time*N / 8 / 77 / 0.68 / 0.71
Time*vegetation type*N / 8 / 77 / 1.00 / 0.44
Time*mowing*N / 8 / 77 / 0.27 / 0.27
Time*vegetation type*mowing*N / 8 / 77 / 1.40 / 0.21
Shoot / Time / 2 / 27 / 6.87 / 0.00*
Time*vegetation type / 2 / 27 / 0.44 / 0.01*
Time*mowing / 2 / 27 / 0.16 / 0.14
Time*vegetation type*mowing / 2 / 27 / 0.00 / 0.97
Time*N / 2 / 27 / 0.42 / 0.01*
Time*vegetation type*N / 2 / 27 / 0.10 / 0.26
Time*mowing*N / 2 / 27 / 0.01 / 0.86
Time*vegetation type*mowing*N / 2 / 27 / 0.15` / 0.14

NumDF: degrees of freedom in the model; DenDF: degrees of freedom associated with the model errors. *, P <0.05

Fig. S1. Total root length (mean ± SD), per m2 of minirhizotron image in 0-10, 10-20 and 20-30 cm depths over three years, averaged across two vegetation types and four mowing and N treatments. T: time; D: depth; TD: interaction of time with depth; *: 0.01 P< 0.05; **: 0.001 P< 0.01; ***:P0.0001.

Fig. S2. Soil carbon (C) concentration (mean ± SD) (%) of soil beneath the invasive grass Agropyron cristatum (dark bars) and native grassland (light bars) in three depth classes. D: depth; DV: interaction of depth with vegetation type; **: 0.001 P< 0.01; ***:P < 0.0001.