*3.1. Ryegrass Production, and Cadmium* × *Nitrogen Accumulation Relationship*

Ryegrass N and Cd accumulation are shown in Table 2. A two-way ANOVA was conducted to test the differences among biochar rates and between the two different feedstockderived biochars for a given rate. This was because the interaction *rate (R)* × *biochar (B)* was significant (*p* < 0.05) in all cases: shoots, roots, and shoots + roots (Table 2). As highlighted in the previous work of Antonangelo and Zhang [11], the ryegrass shoots, roots, and whole plant (shoots + roots) yielded better when treated with PLB presenting, on average, an increase of 64, 51, and 59% respectively when compared to those treated with SGB (Table 2). Similarly, N and Cd accumulation had an overall increase of 88, 60, and 84%, and 71, 30, and 40% respectively, when PLB was applied (Table 2). At 0.5 and 1% of PLB application, the ryegrass shoots presented the highest N and Cd accumulation (Table 2), accompanied by their highest yields. This is the first indication that Cd accumulation increases as the ryegrass yield increase with N accumulation. The approximately 35 g Cd removed ha−<sup>1</sup> in ryegrass shoots after PLB application at 0.5 and 1% (Table 2) is equivalent to a concentration of 3.4 ± 0.1 mg kg−<sup>1</sup> (dry matter, DM) in our study, which is still far below the maximum tolerable concentration of 10 mg kg−<sup>1</sup> for animal grazing [29]. The Cd accumulation in ryegrass plant parts decreased with biochar application rates regardless the feedstock from which the biochar was produced (Table 2). However, it must be pointed out that such reduction is not only a consequence of the reduced N removal and ryegrass yields at 4% of biochar amendment but also due to a direct effect of the biochar properties with the potential to immobilize Cd in the soil, amongst them surface functional groups, alkalinity, organic carbon (OC), and CEC [24].

**Table 2.** Nitrogen (N) and cadmium (Cd) accumulation in ryegrass shoots, roots, and shoots + roots as a function of biochar (B) application rates (R).


SGB: switchgrass-derived biochar. PLB: poultry litter-derived biochar. Different lowercase letters in columns (within biochar rates) and uppercase letters in rows (same biochar rate) are significantly different at *p* < 0.05 (Tukey).

> Table 3 shows the ANCOVA of linear regressions between ryegrass yield and N accumulation and between N and Cd accumulations plotted in Figure 1. Conversely to ANOVA, the interactions of *biochar* × *yield* and *biochar* × *nitrogen* from ANCOVA were not significant in ryegrass shoots and roots (*p* > 0.05), except for the whole plant when evaluating the *biochar* × *nitrogen* interaction in the relationship between N and Cd accumulations (*p* < 0.05) (Table 3). Therefore, most relationships were plotted with both biochars combined (Figure 1a–e).


**Table 3.** Analyses of covariance (ANCOVA) from linear regression models between nitrogen (N) and ryegrass yield and between cadmium (Cd) and N accumulations.

When *p* ≤ 0.05 the effect test is significant for the designated factor. When the interaction *biochar* × *yield* or *nitrogen* is *p* ≤ 0.05, results must be presented as two regression lines, one for each biochar.

Figure 1 clearly shows that N accumulation increases as a function of biochar application rates regardless of its feedstock, and such an accumulation contributes to ryegrass yield increases in shoots and roots. Concomitantly, the Cd accumulation of those plant parts follows their yields increment as well (Figure 1). This agrees with the findings of [12,15–18] who observed greater yields with increased N rates and consequently greater Cd uptake in the studied plants. In fact, N is the most limiting nutrient as far as yield increase is concerned and, in the case of our study, the same evaluations carried out for P and K showed a positive correlation with ryegrass yields but no relationship was found between P and K with Cd accumulation in any plant part (data not shown). Opposite to the increase of Cd accumulation in ryegrass plant parts, the Cd concentration in ryegrass shoots and roots, when their yields are not accounted for, decreased considerably with biochar application due to immobilization effect of phytoavailable Cd as a direct consequence of biochar properties, as observed in the previous study [11].

Curiously, the relationship between the yield of ryegrass roots and Cd accumulation was highly positive regardless of the biochar (Figure 2). There was an increment of 23 mg Cd accumulated ha−<sup>1</sup> for every kg of ryegrass roots increase (Figure 2). The maximum accumulation in the range of 80 to 160 g Cd ha−<sup>1</sup> is related to the PLB at 0.5 and 1% application rates (Table 2, and Figures 1 and 2). Feng et al. [30] also reported an increase in root yields (DM) of two ryegrass varieties when the levels of phytoavailable Cd were higher. This might be attributed to the fact that root density is higher in the presence of this toxic metal. However, Feng et al. [30] also observed that ryegrass roots although presenting higher enrichment ability for Cd, its transportability to shoots was poor. Not surprisingly, the accumulation of Cd in ryegrass roots was much higher than in shoots in the case of our study (Table 2 and Figure 1e). This larger accumulation of Cd in the roots is also a cause of the lower Cd transference to the shoots [31,32], as will be discussed further.

**Figure 1.** Relationships between ryegrass yield and N accumulation (**a**–**c**) and between nitrogen (N) and cadmium (Cd) accumulations in ryegrass shoots and roots (**d**–**f**) as functions of biochar application rates; acc.—accumulated.

**Figure 2.** Relationship between root yields and cadmium (Cd) accumulation in ryegrass roots as a function of biochar application rates. ANCOVA: analysis of covariance. When the interaction *biochar* × *Cd* is *p* ≥ 0.05, results of both biochars must be combined in a single linear regression.
