3.3.1. Savoy Cabbage

The cropping system had significant e ffects on all the tested variables except for the dry biomass of the weeds and the P concentration in the marketable product (Table S1). The year e ffect was not significant for fresh marketable yield, total dry matter, mean fresh weight of corymbs, the dry biomass of the weeds, and the accumulations of N and P2O5 in marketable yield and total biomass.

In Tables 2 and 3, the within-year e ffect of the cropping system for all the tested variables on crop biomass production is shown.

**Table 2.** Least squares means and standard errors of marketable fresh yield (Y), dry matter of marketable yield (dwy), dry matter of residues (dwr), total aboveground dry matter (dwt), mean fresh weight of corymbs (MFW), Harvest Index (HI), and mean diameter of corymbs (MD) in savoy cabbage. Confidence level: 95%.


Means followed by different letters are statistically different (95% confidence interval). \* Value statistically not different from zero.

In terms of fresh marketable product, the organic conservative system (ORG+) yielded significantly less than the standard organic (ORG) and the integrated system (INT) in all three years. The worst results were achieved in 2016, when the ORG+ yielded only 2.61 Mg ha−<sup>1</sup> (a value that was even statistically not di fferent from zero) as many cabbage plants in intercropping with the living mulch of red clover were not able to reach the corymb set stage. Averaged over the first two years, ORG+ yielded

around 50% less than ORG and INT. Nevertheless, in 2015, ORG+ yielded statistically not different from INT both in terms of fresh (Y) and dry matter (dwy) of marketable yield. In 2015, ORG significantly outyielded INT, whilst in the other two years, the yields of the two systems were comparable.

The dry matter of residues was always lower in ORG+ than in ORG and INT. As a result, the total aboveground dry matter produced by savoy cabbage in the three years followed the same trend, with lowest values in ORG+. Only in 2015, we did not observe any statistical differences between ORG+ and INT, with INT performing not differently from ORG. Interestingly, the HI revealed a substantial similarity between ORG and ORG+, with the two systems showing values significantly higher than INT in two years over three (i.e., in 2015 and 2016 for ORG and in 2014 and 2015 for ORG+). Only in 2016, due to the very low yield, ORG+ was significantly lower in terms of HI with respect to ORG and INT.

The mean fresh weight of corymbs followed the same trend as the biomass of marketable yield, with ORG+ showing significantly lower values than ORG (in all three years) and INT (in 2014 and 2016). The MFW in ORG+ was clearly lower if compared to ORG and INT.

The mean diameter of corymbs had a similar trend in all the three years, with ORG+ showing on average 5 cm lower values with respect to INT and ORG.

In Table 3, the within-year main effects of the cropping system on N and P concentration and accumulation in savoy cabbage biomass are shown.

**Table 3.** Least squares means and standard errors of N concentration in marketable yield (Nconcy) and residues (Nconcr); N accumulation in marketable yield (Naccy), residues (Naccr), and total aboveground dry matter (Nacct); P concentration in marketable yield (Pconcy) and residues (Pconcr); and P2O5 accumulation in marketable yield (P2O5accy), residues (P2O5accr), and total aboveground dry matter (P2O5acct) in savoy cabbage. Confidence level: 95%.



**Table 3.** *Cont*.

Means followed by different letters are statistically different (95% confidence interval). \* Value statistically not different from zero.

N concentration in corymbs and residues was normally higher in INT than ORG, except for 2014, when we did not observe any di fference for marketable yield. The cabbage grown in ORG+ plots showed values of N concentration lower than ORG and INT only in 2015, whilst in 2014, it showed the highest value for Nconcy and, in 2016, it performed equal to the other two systems. As a result of the combination between concentration and dry matter production, N accumulation showed overall significantly lower values in ORG+ than ORG and INT. Only for Naccy in 2014, we observed comparable results among the three systems.

For P concentration, ORG+ did not show lower values compared to ORG and INT. In 2014, both Pconcy and Pconcr were higher in ORG+ than ORG and INT. Also, in 2015, Pconcr was higher in ORG+. P accumulation in the dry matter of savoy cabbage was a ffected by dry matter production values and revealed normally lower values in ORG+ than ORG and INT but with some exceptions in 2015, when P2O5accy and P2O5acct and P2O5accr were not di fferent in ORG+ and INT. Oppositely to N accumulation, in absolute terms, ORG produced slightly higher values of P accumulation in savoy cabbage total biomass with respect to INT and ORG+.
