C/N ratio and nitrogen content
The lack of response of S. fuscum biomass production to high atmospheric CO2 suggested biochemical limitations of the photosynthetic machinery. To explore the nature of possible limitations we investigated changes in whole-tissue N contents and C/N ratios (Fig. 3C, D). Variation in the C/N ratio (40-89 units) was mostly explained by WT (31%, P< 0.001), atmospheric CO2(12%, P= 0.003), interaction between temperature and WT (11%,P= 0.004), and interaction between CO2 and WT (7%, P= 0.022, Table 1). Increasing the atmospheric CO2 concentration increased the tissue C/N ratio by 4 to 23 units at low WT, but had no significant effect at high WT (Fig. 3C, Table S1). Increasing the WT generally decreased the C/N ratio by 5 to 31 units. The changes in C/N ratio were primarily due to changes in tissue N content, the C content only showed minor variation (Fig. S3).
Accordingly, the total variation in N content (5.5-11.8 mg g-1) was explained by WT (31%,P< 0.001), atmospheric CO2 (11%,P= 0.013) and the interaction between WT and temperature (8%,P= 0.031, Table 1). Increasing the CO2 level decreased the N content by 7-28 % at low WT, but had no significant effect at high WT (Fig. 3D, Table S1). Increasing the WT caused a 1.1- to 1.6-fold increase in N content. Both C/N ratio and N content were strongly correlated with the D6S /D6R ratio at low WT (R2=0.68, P <0.001 and R2=0.71, P <0.001, respectively). At high WT the ratio was still correlated, but less strongly, with both the C/N ratio (R2=0.44, P <0.003) and N content (R2=0.40, P <0.005).
No major between-treatment differences were detected in whole-tissue δ15N signatures (average: -3.63 ± 0.08 ‰ SE; Fig. S3). According to the 4-way ANOVA, the only significant effect was a small (~0.45‰) increase associated with increasing the light intensity (R2=0.23, P =0.003, Table S2).