3.2.3. Physiological Maturity

The content of available P at the physiological maturity of WOSR as compared to the onset of FL, as a rule, was slightly higher in 2016, and considerably lower in 2017. A significant effect of NFSs was observed only in 2016 (Table 6). A higher P content was recorded for treatments fertilized with organic N, being significant in the subsoil. The effect of N rates was non-significant, although a decreasing trend in accordance with the increasing N rates was observed. The content of available K was only affected by NFS. In 2016, as compared to FL, much lower K values were recorded in the topsoil, but an opposite trend was observed in the subsoil. In 2017, an increase in K content was noted for the subsoil.

A depletion of the available Mg content, as compared to FL, was recorded in 2016 for all soil layers. In 2017, it appeared only in the topsoil. In the subsoil, a net increase in Mg content was recorded. The impact of NFS was observed only in 2016 for the subsoil, in which a significantly higher content of Mg was recorded for the M-NFS and O-NFS. The effect of N rates was observed only for the second subsoil layer in 2016. The content of available Ca showed the highest variability. Ca status was significantly affected by both experimental factors and their interaction with a particular year. In both years, the impact of NFS was recorded for all three soil layers. As a rule, the lowest content of available Ca was found for the OM-NFS. The highest Ca values were the attribute of M-NFS. The effect of the increasing N rates was highly variable between soil layers. For the topsoil, the lowest content of Ca was generally recorded for the plot fertilized with 180 kg ha−<sup>1</sup> of N. In the subsoil, no consistent rule governing the content of available Ca was observed.

The four PCs accounted for 69.50% of the total variance in the pooled NFSs (Table A4). PC1 had the highest and most positive loadings for the Ca content in the subsoil. PC2 was associated with Kc that exerted a negative effect in its value. PC3 explained 14.16% of the total variance, and no significant loading was recorded. PC4 was controlled by the nitrate N content at BBCH 89 (N89), having a negative loading. Yield showed a positive relationship with N89 and Pa, but the latter changed in the opposite direction to that of N89 (Table S3a; Figure S3a).



**Table 6.** The content of available nutrients along the soil layers at WOSR ripening (BBCH 89), mg kg−1 soil.

\*\*\*, \*\*, \* significant at *p* < 0.001; < 0.01; < 0.05, respectively; n.s.—non significant; a, b, c significance letters, a—the highest, c—the lowest; a means within a column followed by the same letter indicate a lack of significant difference between the treatments; a, b, c—soil layers of 0.0–0.30. 0.30–0.60. 0.60–0.90 cm, respectively.

240 55.7 31.5 20.5 117.3 66.6 58.8 66.3 56.2 66.0 369.6b 122.6b 117.1 b

F-value 0.97 0.64 0.59 0.41 2.24 0.77 1.37 0.83 2.26 5.87 \*\*\* 6.13 \*\*\* 4.89 \*\* FS × N 0.31 1.49 0.53 1.98 0.54 0.76 0.40 0.57 1.37 2.99 \*\* 8.42 \*\*\* 6.41 \*\*\*

F-value 1.99 0.08 0.33 0.41 3.37 \* 5.0 \* 1.27 0.56 1.69 5.12 \*\* 16.2 \*\*\* 5.32 \*\* Nitrogen 0 63.0 34.7 25.7 105.4 78.6 63.8 66.8 64.6 74.6 379.7b 167.1b 192.9 a,b rate 60 63.1 38.2 24.6 105.3 64.5 51.1 73.8 75.4 87.4 316.4b 180.2ab 152.4 b (N) 120 48.8 26.2 19.2 104.2 90.0 53.5 79.7 73.1 111.3 595.2a 260.3a 280.5 a kg ha−1 180 54.8 24.2 18.3 110.4 74.6 52.1 71.4 64.8 76.9 316.8b 185.1ab 226.4 a,b

For the M-NFS, three PCs accounted for 75% of the total variance. PC1 was associated with seven of fourteen variables. Positive loadings were identified for P in the subsoil and Caa, but negative values for Mg and Ca for the subsoil. Variables within each subgroup were significantly and positively correlated with each other. Yield exerted a high positive loading on PC2. PC3 had the highest and most positive loading with Ka, and Kb. Yield was weakly related to other soil variables but showed an opposite direction to Pa and Kc. The latter variable was significantly and negatively correlated with Y (Table S3b; Figure S3b).

The O-NFS was associated with three PCs, which explained 73.5% of the total variance. PC1 showed high loadings of Pb, which were positive, and of Mgb, Cab, Cac, which were negative. PC2 showed high positive loading with N89 and moderate with Y (0.65) and Kc (0.63). These three variables were significantly correlated with each other, although only N89 significantly a ffected Y. The opposite direction to that set of variables was exerted by P variables, of which Pa and Pb were significantly and negatively correlated with Y. PC3 showed a high but negative loading with Ka (Table S3c; Figure S3c).

In the OM-NFS, four PCs accounted for 81.88% of the total variance. PC1 had positive loadings for Ca, but negative for P and Mgb in the subsoil. PC2 was significantly a ffected by Mga and Mgc, which had high and positive loadings, and Pa with a negative e ffect. PC3 was dominated with Ka. PC4 was associated with N89 due to its high loading and moderately a ffected by Y (0.61). Yield was closely, but not significantly, related to Ka and N89 (Table S3d; Figure S3d).
